Logic & the Workings of the Mind: The Logic of Ideas & Faculty Psychology in Early Modern Philosophy (North American Kant Society Studies in Philosophy)

LOGIC AND THE WORKINGS OF THE MIND: The Logic of Ideas and Faculty Psychology in Early Modern Philosophy Edited by Patricia A. Easton

VolumeS North American Kant Society Studies in Philosophy

Ridgeview Publishing Company

Atascadero, California

Preface The papers collected in this volume address two closely related themes: the faculty psychology and the logic of the early modern period. The themes are related because, firstly, early modern logic-especially the early modern "logic of ideas" was explicitly psychologistic. It dealt with "concepts" rather than terms, "judgments" rather than propositions, and "reasoning" rather than arguments, and it sawall of these fundamental explanatory categories as grounded in contents or operations of the nlind. And secondly, the lines of influence ran in the other direction as well. The higher cognitive faculties identified by early modern (and, indeed, by medieval and ancient) psychology were determined by logical and even granlmatical considerations. Each cognitive faculty was understood relative to the notion that reasoning consists of arguments and that judgments assert relations between concepts. The intellect was understood as the faculty for abstracting universal concepts from the deliverances of sense~ judgment, as the faculty for conlpounding and dividing concepts or as the faculty for inventing the middle term for a syllogism; and finally, reasoning was understood as the faculty for drawing inferences from previously made judgnlents. Faculty psychology cannot, therefore, be completely understood independently of traditional logic, and early modern logic certainly cannot be understood independently of faculty psychology. For most of this century both of these themes have been neglected by philosophers and historians of logic, philosophy, and psychology. The explanatory categories of traditional faculty psychology now seem naive and illfounded. And the notion that a normative discipline like logic might be grounded on purely descriptive facts of our psychology, or on the arbitrary and conventional features of the grammar of a particular natural language, is rejected as an instance of the naturalistic fallacy. The early modern period has accordingly been judged to be the dark age of logic-a time when the advances of the Middle Ages were forgotten and the entire discipline was turned down the wrong path. But, as Fred Michael observes in one of the introductory essays to this volume, although early modern logic· made virtually no contribution to the history of logic, it was a central part of early nlodern epistemology and metaphysics. One does not have to look far into the standard early modern logic textbook, with its four-part treatment of ideas or concepts, judgments, reasoning, and method, to find themes of crucial importance to early modern philosophy. It was obligatory that a textbook of early modern logic discuss the notions of conceptual clarity, distinctness and adequacy-notions that played a key role in the epistemology of Descartes, Locke, Leibniz, and Wolff, to name but a few. And in early modern logic, a discussion of general terms could no more be separated from the issues of abstraction and abstract ideas-issues that

ii LORNE FALKENSTEINIPATRlCIA EASTON were to become of central importance for later British empiricism-than a medieval treatment of the same topic could be separated from the issue of the nature of universals. Similarly, the early modern logic of propositions, because it could not be separated from the operation ofjudgment, dealt not just with the concept of relation, but with the act of relating, and referred crucially to the basis of that act in the (rationalist) analysis of concepts and the (empiricist) evidence of experience. Again, syllogistic reasoning, based as it is on categorical propositions (out of which the paradigmatic syllogistic forms are constructed), carried with it an implicit ontology of substance and property (the subject and the predicate of the categorical proposition)-an ontology that continued to dominate early modern metaphysics and epistemology long after substantial forms and real qualities had been banished from early modern philosophy of nature. Furthermore, such popular principles of early modern ontology as the notion that whatever is conceivable is a possible object of experience, are obviously parasitic on notions of logical and real possibility. And the analytic and synthetic methods discussed in the fourth part of most early logic textbooks have an obvious relation to the opposed Cartesian and Newtonian paradigms for scientific research. Given all of these avenues of possible interaction, and given that logic formed a basic part of the educational curriculum in the early modern period, it is natural to wonder whether the logics the early modern philosophers were brought up on did not influence their thought-or, alternatively, whether important aspects of their thought are not best understood as explicit reactions to those logics. But the relations go beyond those of influence between distinct disciplines. The fact is that what are today recognized as works of theoretical philosophy were regarded as "logics" in the early modern period. Locke's Essay was taught as a logic, alongside the Port Royal Logic. Hume, in the Abstract to the Treatise, billed his work as a new inductive logic, written to repair the omissions noted by Leibniz in the standard "logic" textbooks, among which Leibniz had included Locke's Essay and Malebranche's Recherche, as well as the Port Royal Logic. And Kant's Critique of Pure Reason-with its Analytic of Concepts, Analytic of Principles, Dialectic of Reason, and Doctrine of Method-is explicitly laid out on the pattern of the table of contents for a logic textbook. The relation goes the other way as well. Just as what are today regarded as works of theoretical philosophy were recognized as logics in the early modern period, so works that explicitly billed themselves as logics in the early modern period were tailored to specific positions in theoretical philosophy. This is the case with the Port Royal logic, and with the logics of Hobbes and Condi1lac, to mention a few. The point that has just been made about early modern logic can be nlade as well about its companion discipline, faculty psychology. Just as there are no clear boundary lines separating logic from epistemology or metaphysics in the early nlodern period, so there is no clear separation between any of these

Preface iii disciplines and faculty psychology. But, as Gary Hatfield argues in the second introductory essay included in this volume, this does not mean that early nlodern philosophy and logic can be regarded as "psychologistic"-as properly normative disciplines that were illegitimately grounded in a naturalistic account of the workings of the mind. There is no "naturalistic fallacy" in early modern logic or epistenlology, Hatfield observes, because, with few exceptions early modern philosophers from Descartes through to Reid had an antecedent conlmitment to the notion that the higher cognitive faculties are truthgenerating and truth-preserving. Accordingly, for them, "naturalistic" accounts of the origin of an item of knowledge in a particular cognitive faculty can also serve to validate that item. And the converse is true as well. Because early modern philosophers had an antecedent commitment to the notion that we are so constituted as to be able to discover the truth, they could base their accounts of the cognitive faculties, not on an empirical study of the workings of the mind, but on considerations of the sorts of operations that logic dictates that the mind must perform in order to discover the truth. Since logic examines argunlents, argulnents consist of propositions, propositions consist of terms, and most terms are general terms that name classes of objects of experience, the identification of the "higher" cognitive faculties of reasoning, judgnlent, and conceptualization lay ready to hand. If early modern logic and epistemology are "psychologistic," then by the same token early modern psychology is "logistic" and rests on a "normativistic fallacy." If we accept that logic, faculty psychology, and theoretical philosophy formed nlutually interdependent parts of a single enterprise in the early modern period, then the importance of paying careful attention to the former two disciplines becomes evident. To properly understand the work of the past philosophers, we need to understand the background presuppositions that influenced them and against which they reacted. Yet the background supplied to all early modern philosophers by their juvenile training in logic and their common employment of distinctions between lower (sensory) and higher (conceptualizing, judging, and reasoning) activities has been largely ignored. The purpose of the essays collected in this volume is to take some few steps to repair this omission. They demonstrate that, be it by way of unconscious influence or be it as the target of an explicit reaction, the logic and the faculty psychology of the early modern period exercised a pervasive and multi-faceted influence on the thought of the period, and that an attention to early modern logic and faculty psychology can help us better understand the nlotives and intentions of early modern philosophers. The papers have been collected into three parts. The first, introductory part contains two papers, one by Fred Michael on the logic of ideas and one by Gary Hatfield on faculty psychology. They are particularly broad-ranging and programmatic in nature, and serve to give the rationale and the justification for the entire project of this volume. The remaining papers constitute a battery of

iv LORNE FALKENSTEINIPATRICIA EASTON additional case studies that illustrate the applicability of Michael's and Hatfield's claims. The papers in Part II investigate the influence of early modern logic on, and reactions to early modern logic in, early modern philosophy. Part IIa contains papers dealing with the logic of ideas and with particular instances of the relation between that logic and early modern epistemology. Jennifer Ashworth explores the influence of the Scholastic tradition on many developnlents in early modern philosophy, reminding us that the reformation of logic was not a development based solely in the humanist movement. In pal1icular, Ashworth examines the writings of Petrus Fonseca (1528-1599) which she identifies as a significant source for the intermingling of logic and the philosophy of mind that was typical of the period. Elmar Kremer traces Arnauld:s account of the nature of ideas from the important Port Royal Logic through to a later polemical work against Malebranche, On True and False Ideas. Kremer argues that the often puzzling theme running through both works is a methodological one-one wherein the account of the nature of ideas comes to bear on issues of logic, knowledge, and theodicy. Emily Michael identifies the sources responsible for the transmission of the logic of ideas into the Scottish universities, identifying John Loudon's logic compend and a standard logic text that was developed, but never used, at St. Andrews as important ancestors to Hutcheson's popular logic textbook. Part lIb deals with judgment. A chief, if not the chief role of judgment was taken to be the assertion of relations between concepts, and it is the logic of relations that is the focus of Fred Wilson's paper. Wilson argues for the importance of distinguishing between the dominant "logic of ideas" of the early modern period and a rival "Ramist" account of logic. He shows that the dominant texts were Aristotelian not only in their logic, but also in their ontology, and that the Aristotelian substance/predicate ontology was unsuited to accommodate relations, which Aristotelian logicians tried to reduce to purely ideal phenomena-mental acts of comparison of the monadic predicates of substances. Seen fronl this perspective, the development of alternative, "bundle theories" of objects by Berkeley and Hume might be seen as opening the way for the modern, Russellian account of relations, but Wilson shows that the Ramist logic of invention already contained an account that treated the predication relation on the same level as other relations, such, as causal and part/whole relations-and that it is likely that Berkeley received instruction from Ramist logicians. Jill Buroker's paper explores another case of complementary, shifting paradigms in logic and epistemology, focusing once again on the faculty of judgment. According to Buroker, the central, revolutionary figure is Kant, who claimed that all the acts of the understanding may be reduced to judgments. Buroker argues that this is a breach with Cartesian philosophy of nlind, as reflected in logic texts such as the Port Royal Logic. These texts treated concepts, rather than judgments, as fundamental. Buroker sees Kant's

Preface v assertion of the primacy of judgment as an important step on the way to an externalist account of language. Part lIe collects papers focusing principally on reasoning and inference. Charles Echelbarger distinguishes between scholastic and modern (Cartesian and Lockean) treatments of logic and investigates which traditions can plausibly be supposed to have influenced Hume, and at what stages in his career. He argues that Hume's own view of reasoning was an explicit reaction to both scholastic and modern traditions, and grew out of a more rigorous application of new theories of the workings of the cognitive faculties to the subject matter of logic than had previously been the case. David Owen addresses Hume' s treatment of demonstration and deduction and warns against the tendency to assimilate Hume's demonstrative/probable contrast with the contemporary contrast of deductive/inductive. Finally, Patricia Kitcher's paper deals directly with the relation between faculty psychology and logic, providing an interpretation of Kant's account of self-consciousness that helps explain how logical inference occurs. The papers in Part lId deal with treatments of necessity, possibility, and impossibility as they appear in the logic and the fundamental philosophical presuppositions of early modern philosophers. Fran90is Duchesneau explores Leibniz's notion of contingent truth. In addressing the problems that arise for Leibniz's account of contingent truths under the analytic model, Duchesneau develops a solution that does not appeal to a Kantian shift in meaning. Phillip Cummins examines Hume's treatment of possible objects or states of affairs, and identifies "perspicuous" ideas as the basis for Hume' s account of psychological and ontological possibility. Cunlmins argues that Hume's methodological hostility to language-based ontology, as evidenced in his denial of the possibility of substance, for example, is one that is best explained by Hunle's reliance on the criteria for forming a perspicuous idea. As Cumnlins points out, the idea-first method that Hume adopted, while conservative, is not without its problems or metaphysical assumptions. Finally, Manfred Kuehn's paper exanlines the Wolffian notion of the proof of the possibility of a concept, and argues that this notion illuminates Kant's notion of a transcendental deduction-that Kant's Transcendental Deduction in fact sets out to provide a principium probandi possibiltatem for the categories. Kuehn defends this reading against alternative approaches to the Deduction, notably Henrich's claim that the Deduction is modeled on legal demonstrations of entitlement. Part III contains papers investigating the influence of early modern faculty psychology on early modern philosophy in general, and on logic in particular. Catherine Wilson's paper is a wide-ranging survey of thought about mental health and the proper employment of the faculties from Descartes to Kant. It investigates the degree to which the mind and its operations came to be construed in materialistic terms in the early modern period, and what this meant for conceptions of logic as a regimen for inlproving mental health and

vi LORNE FALKENSTEINIPATRlCIA EASTON the functioning of the reasoning capacity. A particular instance of this same theme is picked up by Robert Butts, who isolates a quasi-physiological account of the causes of error in Kant's remarks on the role of the sensory and intellectual faculties in generating the antinomies. Eric Palmer presents the thesis that Descartes's Rules for the Direction ofthe Mind is a treatise on logic that enlploys a decidedly naturalistic epistemology, unlike that found in his later writings. Palmer emphasizes Descartes's focus on sensation and brain physiology in his treatment of method and rules for the direction of the mind. Finally, Louis Loeb argues that Hume's formulation of the problem of induction was not intended to result in the sceptical conclusion that causal inference is not justified. Leaving aside worries about the naturalistic fallacy, Loeb argues that if we assume that Hume viewed inference as a psychological process of association, we can better search for Hunle' s theory of justification in the operations of the mind. As Hatfield points out in his introductory essay, scholars have all too often been tempted to ignore or reinterpret whatever they find naive or implausible in the works of past philosophers in accord with the principle of charity. Since early nlodern logic and faculty psychology appear fallacious or naive to us, these elements have accordingly been read out of the history of the period by "charitable" interpreters. But the principle of charity can be abused. Its proper role is to rectify minor oversights and trivial mistakes, or to interpret obscure and difficult passages in the light of clearer pronouncements made elsewhere. But when it is taken to entail that the great minds of the past cannot legitimately be supposed to have had an incorrect or ill-founded thought, or must at all costs be seen to have had some valuable contribution to make to currently raging debates, then it can do more to obscure than to reveal their actual motives and intentions. This is especially the case when what nlotivates the "charitable" omission or revision is a view special to our own time and not shared by the philosophers we are studying. Early modern logic and faculty psychology only seem fallacious or naive to us because we view them fronl a perspective that defines the purposes of the disciplines, and draws the distinction between the naturalistic and the normative in a quite different way than was done by early modern philosophers. If we accept that the purpose of the history of philosophy is not merely to comb the works of past philosophers for something that might possibly be of relevance for current concerns, but is rather to tell a story of the development and influence of ideas, and so to, as Hatfield puts it, "gain the sort of understanding that comes from uncovering the formation of our current problem space, and seeing its contingencies"-as well as get at the facts of what past philosophers really thought that lie behind the mask of "charitable" readingsthen the logic and the faculty psychology of the early modern period have to be taken seriously. This is something that has up to now not been done in any

Preface vii

systematic fashion, and it is an omission the papers in this volume aim to rectify. Lorne Falkenstein, University of Western Ontario Patricia A. Easton, The Claremont Graduate School

Why Logic Became Epistemology: Gassendi, Port Royal and the Reformation in Logic FREDERICK S. MICHAEL

I. Introduction It is quite obvious that epistemology permeates most of the logic texts written from a period beginning in the late seventeenth century and continuing into the beginning of the contemporary era in logic at the end of the nineteenth century. The model of this kind of logic appears to be the Port Royal Logic. Since this is a work suffused throughout with Cartesian doctrine, it is natural to conclude that this kind of logic is of Cartesian inspiration. Even though Descartes himself did not think of logic in this way, indeed he appears to have viewed logic, and abstract thought generally, with suspicion, the epistemological approach to logic taken in the Port Royal Logic can be seen to be a natural outgrowth of Cartesian philosophy. The problem with this judgment is that there had been an earlier logic of this same type and its author, Pierre Gassendi, not only was not Cartesian, but was Descartes's principal rival among the moderns. His Institutio Logica, published not as a separate work, but as part of the Syntagma Philosophica, which itself is available only as the first two volumes of Gassendi's posthumous Opera Omnia, was, as I will try to show, both conceptually and structurally, the Port Royal Logic's principal nlodel. Inasmuch as each of these logics has as its foundation a theory of ideas, it seems appropriate to call this kind of logic, the logic of ideas. Historians of logic do not look with much favour upon this kind of logic. In the introduction to his English translation of Gassendi's Institutio Logica, Howard Jones states that this work is "not a revolutionary logic which rejects all that the logical tradition has to offer, but a logic which Gassendi renders contemporary by selecting from that tradition only what is appropriate to seventeenth century needs."} Wilhelm Risse's assessment of the Port Royal Logic is similar. He says of this work, that it is historically one of the high points of logic, conlparable in influence to that of Aristotle, Peter of Spain, Ramus and Wolff. But he adds: "This logic is certainly not original. Its extraordinary success is due to its elegance and its pedagogically effective manner of presentation.,,2 With respect to logic after the medieval period, which includes the humanist logics of the Renaissance period in addition to the logic of ideas, William and Martha Kneale in their The Development of Logic remark that "from the 400 years between the middle of the fifteenth and the middle of the nineteenth century we have...scores of textbooks but few works that contain anything at once new and good.,,3 The logic of this same era is called by I.M. Bochenski, "classical logic" and is characterized by him a~~~~~~~t_~~~g_gll}!~}l_~w .}\l'hic_h ---

2 FREDERICK S. MICHAEL held the field in hundreds of books for nearly four hundred years,,4 but while he sees it as new, he certainly does not see it as good. This is his assessment: "Poor in content, devoid of all deep problems, permeated with a whole lot of non-logical philosophical ideas, psychologist in the worst sense-that is how we have to sum up the "classical" logics. s While I don't think that this attitude is wholly wrong, I would contend that the logic of ideas lvas revolutionary. More specifically, it was the completion of a revolution that took two hundred years to accomplish, from the sixteenth to the eighteenth century. This was the era of the religious reforn1ation, and it would be as appropriate to speak of a philosophical and scientific reformation in this era as well. It was a period of intense intellectual fernlent and upheaval, in which the medieval world view was abandoned and replaced by the modern world view. It began with an attack on medieval logic. This at first sight seems odd inasmuch as if there is one area of medieval philosophy which those involved with the history of philosophy do not think was in need of reformation, it is logic. That is no doubt at least part of the reason why the reformed logics are viewed today with so little enthusiasm. The reform of logic occurred in two phases. The first phase was largely reactive. Medieval logic was discredited by the humanists and largely abandoned. The humanists hoped to convert logic from the formal and theoretical discipline of the medieval period into a practical study, which they hoped would be an improved instrument for argumentation and disputation, and so for the discovery of truth. There was however no consensus about how this was to be accomplished. The second phase in the reformation of logic began in the early seventeenth century, with the abandonment of the view that the way to truth is via argumentation and disputation. Disputation does not lead to truth, it was held, rather the road to truth is by the way of ideas. The logic of this era is, as Bochenski says, something new. It is an important development in the history of logic. But is it also something good? Were the humanists responsible for an advance in logic? Was the epistemological turn which the logic of ideas brought about, the right turn for logic? For the most part, I would have to answer no. These developments were on the whole not good for logic~ certainly they were not good for formal logic. In the four hundred years from the end of the medieval era to the beginning of the era of contemporary logic, while there was some development in informal logic, forn1al logic was largely neglected. It was a reform of logic, a revolutionary change. But revolutions aren't always good and this one was not good for formal logic. Contemporary logicians and historians of logic have reason to be dismayed by its results. On the other hand, the situation could hardly have been more favourable for the development of epistemology, and of the theory of ideas in particular. Logic was typically the first subject in a course of university studies; and in the logic of ideas, the theory of ideas was the subject matter to which the student was

Why Logic Became Epistemology 3 first exposed. The chief focus in the logic of ideas was not on form but on content, principally on epistemological content. Yet it really was a form of logic, as I hope to make clear and the conception of logic it enlbodies is legitimate. My principal purpose in this paper is to examine the logic of ideas as it is found in Gassendi's Institutio Logica and in the Port Royal Logic, to compare these two works and to explain how this form of logic came about. But I do not think that this form of logic can be understood except in its broad intellectual context. Accordingly, it is with this that I begin.

II. From Medieval to Humanist Disputation In the twelfth century, logic was characterized as a scientia sermocinalis, a science of language, or more literally, of speech or significant utterance. This is the view of logic taken by Abelard~ it is also found in the very popular textbook of Hugh of St. Victor, the Didiscalicon. 6 But while it was formulated during the twelfth century, this conception of logic persisted throughout the Medieval period, although it was more often tacitly accepted than explicitly endorsed. But of what kind of language is logic the science? This is a question which seems not to have been explicitly addressed until the fourteenth century, although it is clear enough in practice that the grammatical rules of natural languages are not the concern of logic. Logic was rather seen to be concerned with inferential relations. The aim of logic was to develop a system of rules, distinctions and techniques which are theoretically adequate in the sense that on the basis of these one would be able to determine what follows and what does not follow from a given proposition or set of propositions expressed in a natural language such as Latin. It would also suffice for the analysis of sophisms, enabling one to lay bare the deceptive nature of these arguments. These practical developments preceded any explicit theoretical elaboration. This was finally to be given by Ockhanl. Ockham in his Summa Logicae, distinguishes between vocal and mental terms, which are the terms respectively of a vocal or mental language. Mental terms are natural signs of things~ written or spoken terms are conventional.. There are grammatical features of ternlS which belong only to spoken and written language-Ockham mentions gender and declension. These do not occur in mental propositions and logic is not concerned with them~ for logic, he argues, is not the science of written or spoken terms, but only the science of mental terms.? Now this mental language of which Ockham speaks, his language of natural signs, is a universal language, inasmuch as it is a language free of the idiosyncratic grammatical features which distinguish one natural language from another. This makes logic the science of a universal mental language. In practice, however, propositions supposed to be in this universal mental language are expressed in an idealized form of Latin. This is not good

4 FREDERICK S. MICHAEL grammatical Latin, but rather a form of Latin in which the grammatical features which are held to belong only to written or spoken Latin are either omitted or ignored, a Latin for propositions of Logic, logically perfect Latin, for short.. These refinements of Ockham were to be generally accepted. They may seem to represent an important step forward. But in retrospect it is possible to see that for the fate of medieval logic, it was an unfortunate development. It provided the Humanists with their best grounds for attacking this logic. Medieval logic proved to be vulnerable to attack for two reasons. First, the use of logically perfect Latin frequently had counterintuitive results which were easy to ridicule. For example, although "some virgins were whores," seenlS like complete nonsense, it turns out to be true in logically perfect Latin, since by the rules of ampliation it means the sanle as "Some who were virgins were (after that) whores."g This form of Latin was explicitly attacked by the Spanish humanist, Juan Luis Vives. Vives explicitly accepts the medieval conception of logic as being linguistic in nature, but differs from those he attacks in several respects. First, he sees only natural languages, such as French and Latin, as legitinlate; second, he sees logic as being language relative. So there is a logic of Spanish, a logic of Latin and of each other natural language and these are the only kinds of logics with any claim to legitimacy. The rules of grammar, logic and rhetoric, Vives argues, do not determine what is right or wrong in a language, but only reflect what really is right or wrong in the usage of that language. 9 Second, much of medieval logic developed was devised to address highly theoretical problems with no very obvious application to the practice of argumentation, and which could therefore be attacked as being useless. Such problems had to be confronted so long as logic was seen as a science or theoretical discipline, which is supposed to be applicable to all logical propositions. Thus it turns out that enormous amounts of subtlety and ingenuity had to be expended on the analysis of propositions of a kind never seen anywhere else but in a treatise on logic. As a result of humanist attack, the medieval approach to logic was largely abandoned during the sixteenth century. Between 1520 and 1540, the humanist approach to logic became the dominant approach, replacing the medieval. 10 The hunlanists had no interest in logic as a science, or theoretical discipline. They saw it instead as an art, an instrumental or practical discipline. It was seen to have an important but limited function. It is to be studied as a means of acquiring skill in disputation and practical argumentation. Whereas medieval logics are highly formal, in humanist logics the stress is on informal aspects of reasoning. Informal logic is the principal field of development during the humanist period in the history of logic. Humanist logics look very different from medieval logics. In humanist logics there is no concern with sophisms, there are no systems of subtle distinctions, no strategies for translating propositions into logically correct form. Most of them retain nothing from

Why Logic Became Epistemology 5 scholastic logic at all. Whereas there had been broad general agreement, and little debate, during the medieval period, about the nature and structure of logic, there is no such agreement among the humanists. Questions about the definition of logic, about its nature and structure, are important areas of dispute among the humanists. The humanists, in their conception of logic and their account of its variety and structure, follow a variety of models. There is no one kind of humanist logic; I can distinguish at least four kinds. Many humanist logics are no more than simplified paraphrases of Aristotle's Organon. Nothing will be said of these, as they are of no intrinsic interest. Others follow Cicero and others still, Averroes. Although a consensus among humanists about the nature of logic seemed to be developing during the seventeenth century, it was a development cut short by the emergence of a new kind of logic, the logic of ideas, the principal focus of my interest in this paper. Despite the dramatic change from medieval to humanist logic, there nlay be less to this change than meets the eye. For there is no change in fundanlental conception. Logic is still understood as a linguistic discipline. Those humanists, who like Ramus follow Cicero, define logic or dialectic as ars bene disserendi, the art of discoursing well, or as ars bene ratiocinandi, the art of reasoning well, since they see reasoning as the essential part of discoursing. 11 Philip Melanchthon defines logic or dialectic rather as ars recte, ordine et precipue docendi, the art of correct, orderly and perspicuous teaching. I2 In Melanchthon's view any transmission of truth is teaching. Teaching is a form of discourse in which the author's goal is to transmit to his audience things which, in the author's judgment, are true. Jacopo Zabarella, who follows Averroes in his conception of logic, defines logic in his highly influential De Natura Logicae in this manner. Est logica disciplina instrumentalis, quibus in omni re verum cognoscatur et a fa/so discernatur, that is, logic is an instrumental discipline by which in every thing the true is known and distinguished from the false. I3 Since an instrumental discipline is an art (as opposed to a theoretical discipline, or science), this amounts to saying that logic is the art of getting to the truth. Since this is done by argumentation, debate or discussion, in a word, discursively, logic is still conceived as a discursive or linguistic discipline. Bartholomew Keckermann complains of Zabarella's definition that it is no more true that logic is instrumental by nature than that a man is by nature a servant. This is true enough, but· nlisses Zabarella's point. Zabarella logic is instrumental not because logic functions by means of instruments, but rather because he nlaintains as do most of the other humanists, that logic is an art, not a theoretical discipline or science. Keckermann himself gives two different definitions of logic. In the first, he characterizes logic as ars dirigens mentem in cognitione rerum, the art which directs the mind in the cognition of things~ in the second, as ars recte de rebus cogitandi, the art of thinking rightly about things. I4 These two definitions, and especially the second, suggest a change in the conception of logic, from a linguistic to an

FREDERICK S. MICHAEL epistemological conception. But that is questionable. Keckermann's two definitions seem to be motivated rather by a somewhat broader view of the goal of logic, which he sees as being sonlething more than just the discovery of truth as will be explained below. Still, it is central to Keckermann's conception of logic that its principal purpose, though not its only one, is the discovery of truth and for Keckermann, as for the other humanists, truth is discovered by argumentative or discursive means. Logic, for Keckermann, as for the other hunlanists, is a discursive art and his logic turns out to be just a later stage in the development of the humanist program for logic. In the whole humanist period, there is no clear break from the conception of logic as a linguistic discipline. There is then general agreement among the humanists about the nature of logic, all seeing it as being in one way or another a discursive art or ars disserendi. They agree also in rejecting the common medieval view that logic is a science or theoretical discipline, holding instead that it is an instrunlental discipline or art. But humanists do not all agree about the goal of logic, and they have widely divergent views about the structure of logic. The early Ciceronian hunlanists see the goal of logic as being to enable us to argue, as Cicero's puts it, "with probability on either side of a question," where arguing with probability seems to amount to nothing more than arguing persuasively.Is So the goal of logic is to enable us to argue persuasively any side of any issue. Since logic is not seen as enabling us to determine on which side of a question truth lies, it is proper to characterize this approach to logic rhetorical. Ramus, on the other hand, sees logic as the means of determining truth and so does not share the rhetorical conception of logic held by his humanist predecessors. I6 Melanchthon is in general agreement with this, seeing the goal of logic as the transmission of truth, while Zabarella, following Averroes, views logic, as we have seen, as an instrument for distinguishing the true from the false with scientific knowledge as its goal. Keckernlann however denies that logic has a single goal. Logic he insists has three principal functions: to explicate, to prove and to order. I7 Its goal is not only to give us the means of proof so that we may distinguish the true from the false, but to clarify what is obscure and to order correctly. Still of these only the early Ciceronian humanists really appear to deny that truth is the principal goal of logic. Truth is reached by argumentation or disputation, it was believed. Aristotle, at the beginning of the Topics, had said that dialectic or logic is a skill which enables us to dispute on both sides of an issue and so more easily perceive the true and the false on each side and that it is the way to all the arts. That was the view of the humanists as well. Logic enables us to dispute effectively and it is by disputation that truth will be found, not scholastic disputation in an artificial language, but humanist disputation conducted in some ordinary natural language. With respect to the structure of logic, however, humanists are in disagreement. Ciceronian humanists divide logic into two parts, invention and 6

Why Logic Became Epistemology 7 judgment, as did Cicero himself. Cicero deals with invention in his Topica; he intended to write a book on judgment but never did. Invention was concerned with finding reasons to support or oppose any thesis. Such reasons were to be found by consulting the places of invention or common places. The places are a series of highly general notions which it is useful to consider when one looks for reasons to support or oppose any thesis. Cause, effect, subject, attribute, each of which have numerous subdivisions, are examples of the places. In the logic of the Ciceronian humanists, the account of invention takes the place of the Aristotelian doctrine of terms, which is concerned largely with an exposition of the predicables (genus, species and the like) and the categories or predicaments. The account of the places given in humanist logics incorporate the predicables, but not the categories, which were held to belong not to logic but to metaphysics. In the logic of the Ciceronian humanists, the emphasis was always on invention rather than judgment and the early Ciceronian humanists, Lorenzo Valla and Rudolf Agricola, give no account of judgment at all. Ramus was the first humanist logician in the Ciceronian tradition to give an account ofjudgment as well as invention and so the first to deal with the whole of logic as Cicero understood it. This is perhaps why his logic was so influential. The first version of Ramus's logic appeared in 1543 and underwent a number of revisions, attaining its mature form in the Dialectique of 1555, revised in 1576, and in the Latin Dialecticae Libri of 1556, revised in 1572. Despite variations in text, organization and terminology, the mature versions present essentially the same doctrines. Ramus holds that judgment, also known as disposition, consists of certain rules which are to enable us to judge well of things. Disposition is essentially putting things in order. There are three kinds of disposition: enunciation, syllogism and method. When ternlS are so ordered, or disposed, that one is affirmed or denied of another we have an enunciation or proposition. Syllogism is an order among propositions so that one can be seen to derive from others. The only noteworthy thing about Ramist accounts of propositional and syllogistic judgment is that it is brief and simple. There is no account of the demonstrative syllogism. Method is an order among subjects to facilitate conlprehension; this is done by proceeding always fronl what by nature is most clear and evident to what is less so. Method is disposition in its largest sense. For Ramus, there is essentially only one method, which proceeds from what is more to what is less general. Melanchthon shares the Ciceronian humanists's view with respect to the importance of the places. Melanchthon's logic appeared first in 1520 in a short handbook, the Compendaria Dialectices. This work was to be revised and amplified frequently, reaching its final state in the Erotemata Dialecticum of 1547. Although in practice, Melanchthon appears to accept the humanist division of logic into invention and judgment, his comments show that he is not comfortable about the role of invention in logic. He at first argues against the 18

8 FREDERICK S. MICHAEL view of those who claim that invention belongs to rhetoric, not to logic, but gradually comes to accept that position, calling the division of logic into invention and judgnlent, the old division. 19 But whatever his qualms about this division, he never effectively gives it up; and even when he has concluded that invention is not really part of logic, he says that invention ought to be our first concern. Nonetheless in all versions of his logic, he treats judgment first. 20 Melanchthon's logic combines the emphasis on places of the Ciceronian humanists with Aristotle. Melanchthon's logic is divided into four books. The first three books are devoted to judgment; the last book to invention. Melanchthon's treatment of judgment is largely Aristotelian. The first book deals with simple questions, questions concerning terms, such as "What is a conlet?"; complex questions, such as "Is a comet a star"; questions concerning propositions are dealt within the second and third books, the second book dealing with propositions simply, the third book dealing with the confirmation or refutation of propositions by other propositions, that is, with argument. The inquiry in the first book into simple questions, such as the question of what a thing is, proceeds by definition and division, using the categories, also known as predicaments, and the predicables, known also as antepredicaments. It also uses what Melanchthon calls nlethod, which is a short list of places, ten in all, derived from the four questions which, according to Aristotle, can be asked about anything: "Is it?"; "What is it?"; "What facts are there about it?"; and, "For what reason is it?,,21 Accordingly, in Melanchthon's logic, there are two lists of places, a prelinlinary list under the heading "method" and a much more extensive and systematically developed one in Book IV. The places, in the fourth book, are said to be the means by which propositions are confirmed or refuted. The places are held to apply to all arguments, including the demonstrative and the fallacious. So Book IV contains an account of fallacious reasoning and a brief treatment of demonstrative arguments. Probably the most distinctive feature of Melanchthon's logic is its two lists of places. They are evidence of Melanchthon's efforts to combine Aristotle and Cicero. For Melanchthon, judgment is more important than invention, but the places of invention still playa prominent role in his logic. Zabarella follows Averroes in dividing logic into universal and particular logic. 22 Universal logic is that part of logic which applies to any subject matter; in particular logic, universal logic is applied to particular kinds of subject matter. Universal and particular logic canle to be known as formal and material logic respectively. The studies of terms, propositions and fornlal argument all belong to universal logic. Particular logic deals with demonstrative arguments, dialectical arguments, and arguments that are sophistical or fallacious. In addition, Zabarella claims, following Averroes, particular logic also includes rhetorical and poetic arguments; they are part of logic, Zabarella argues, because they are applications of universal logic to a particular kind of subject matter. 23

Why Logic Became Epistemology 9 In Keckernlann's Systema Logica of 1503, Melanchthon is combined with elements taken from Zabarella. Logic is divided according to the acts of the mind, a division which is not original with Keckermann, but which goes back to Albertus Magnus and Thomas Aquinas. Many humanist logics use it. The first act is apprehension of simple terms, that is, the formation of simple concepts. However there is no account of the formation of concepts to be found in this or any other humanist logic. The second act is the formation of whole propositions and sentences; and the third is discursus, or discursive reasoning, which is the act of the human nlind advancing from known to unknown by putting propositions together in an appropriate manner. Discursus is of two kinds, inference and order. 24 Although Keckermann shares the views of the Ciceronian humanists on the importance of the places, he is sharply critical of Ramus, accusing him of plagiarisnl, taking his criticisms of Aristotle from the writings of J. L. Vives. 25 The accusation is probably unjust, as the criticisnls in question are of a sort quite typical of humanists of that period. With respect to the places, it is Melanchthon that Keckermann follows. As in Melanchthon's logic, there are two lists of places. The first is in Book I, devoted to simple terms, following the treatment of predicables and categories. The other is in Book III, in connection with Keckermann's account of what he calls the notional or topical syllogism. In Book III, concerned with argumentation and with order, Keckermann divides the syllogism, but not the whole of logic, into the general and the particular. 26 After discussing the form of the syllogism in general, he identifies three particular kinds of syllogism, the notional or topical syllogism, the demonstrative syllogism and the sophistical or fallacious syllogism. He believes the notional syllogism to be most general in its application and so treats it first. In his account of demonstration and the importance of demonstrative reasoning, Keckermann in large measure follows Zabarella. Keckermann's logic is an early example of a sort of humanist logic which became quite common during the seventeenth century, one incorporating a great deal of traditional Aristotelian logic, with the stress on places of the Ciceronian humanists and with a Zabarella inspired treatment of demonstrative reasoning. It is also a logic which is in some respects similar to Gassendi's logic. The structure of humanist logics was becoming more conlplex, but this process was stopped by succeeding developments.

III. Gassendi, Port Royal and the Logic of Ideas In 1624, Pierre Gassendi published the first book of his Paradoxical Exercises Against the Aristotelians. There were to be seven books in all, but only one other was written. Book II was concerned with logic. It was written at about the same time as Book I, but published only posthumously in Gassendi's Opera Omnia of 1658. He contends that artificial logic, !o~~_a~ ~ ~1!bie~~ ~f~tugy, ]s__

10 FREDERICK S. MICHAEL as a whole useless?? It is useless for providing definitions expressing the true nature of things; for no truth about the nature of things is known by us. It is useless for finding an appropriate division of a subject matter into parts; for that one needs a knowledge of the subject matter in question. For instance, it takes a knowledge of geolnetry, to know how the subject matter of geometry ought to be divided. It is useless for distinguishing the true from the false; to know what is true and what is false in geometry, it is geometry which one must know, not logic or dialectic. Considering in turn each of the functions which Cicero says that logic performs, Gassendi argues that logic is useless or unnecessary for performing that function. It appears that he intends this to apply not only to Aristotelian logic, but to Ciceronian logic as well, the logic of the humanists as much as that of the medieval period. Gassendi later takes back this view that logic is useless, not in my view because he thinks that any of the logics with which he is fanliliar do lead to truth, but because he has discovered in the opinions of the ancient philosopher Epicurus, the principles of a logic that would lead to truth, for Gassendi did share the view of most of the humanists, that the goal of logic is in fact truth. 28 In the period since humanist logic had become dominant, it had become painfully obvious that humanist disputation was no more a sure road to truth than scholastic disputation had been. In his writing against the Aristotelians, Gassendi was writing as a sceptic, for whom real truth about the nature of things cannot be obtained. But others, who were not sceptics shared Gassendi's scepticism about the powers of disputation, whether scholastic or humanist. Descartes certainly did not think that disputation was the way to discover truth, and, as is well known, he proposed a road to truth of his own. Hobbes was contemptuous of disputation and, in later years, Locke would be so as well. Gassendi, in claiming that logic was useless as a road to truth was addressing a large and receptive audience. If logic was to be of use in the discovery of truth, it would have to be in some way other than as the result of argumentation or disputation. If logic was to be made into a useful subject of study, further reform would be needed. If truth is the goal of logic, and disputation or argunlentation does not lead to truth, how may it be reached? Gassendi's response is that truth is known by a sign. 29 A sign is something perceptible by sense, by perception of which what is hidden fronl sense may be known. Of course, we have no guarantee that things grasped in this way are as we suppose them to be. Knowledge by signs is not infallible, but, Gassendi believes, this is the only kind of knowledge we have. There are two kinds of signs. There are empirical signs, which are signs of things temporarily hidden, and in this way, smoke is a sign of fire. There are also indicative signs, which make known to us things naturally hidden, and it is by an indicative sign, perspiration, that we know of pores in the skin. It is also by indicative signs that we know of atoms, the void, and other things not directly perceptible by sense, if perceptible at all. Ideas, insofar as they are

Why Logic Became Epistemology 11 capable of leading us to truth, are indicative signs. Truth, insofar as it can be attained, can only be reached, by way of ideas. The principles on which Gassendi based his logic come from his reconstruction of the Epicurean canonic. The canonic is a series of rules derived from Epicurus, and other ancient sources, concerned with knowledge, nleaning and truth. 30 The first set of canons or rules, as represented by Gassendi, are the canons of sense. According to these, sense is never mistaken, inasmuch as mistake is only possible when something is affirmed or denied, and the testimony of the senses is no more than a report; it does not affirm or deny anything. An opinion is sonlething added to sense which is capable of truth and falsity. An opinion is true if supported directly or indirectly by sense, and false, if directly or indirectly opposed by sense. So all our knowledge of truth derives ultimately from the senses. What has been perceived by sense produces an image in the mind. These images provide a conceptual framework for the apprehension of future perceptions, for which reason they may be called anticipations of perception or prenotions. Epicurus calls them prolepses. Even in antiquity, prolepses were called ideas and this is the way in which Gassendi interprets them. "Idea" is the term I will use here. There is a group of four canons concerned with ideas. The first asserts that ideas are images which are derived directly from sense impression, or else are formed by increase or dinlinution, as we acquire the idea of a giant or pygmy, or by composition, as we obtain the idea of a golden nl0untain, or by analogy, as we form the idea of a town we have never seen from one we have. The second holds that the idea is the very notion of a thing, fundamental to its definition and that without an idea, we cannot inquire about, think about or even name anything. The last two canons concerning ideas spell out the connection of ideas with logic. The third asserts that the idea is what is fundamental in all reasoning, and so in logic. It is what we consider when we infer that one thing is the same as another or different, joined to another or separated from it. Finally, the fourth canon affirms that what is not directly evident to sense, must be demonstrated from the idea of something which is evident to sense, as the observation of perspiration on the skin enables us to infer the existence of pores. The logic, which Gassendi constructs using these principles, is the Institutio Logica, a late work, first published in the posthunl0us Opera Omnia of 1658. In this work, Gassendi defines logic as ars bene cogitandi, the art of thinking well. 31 Thinking well involves the following four skills: imagining well, that is, forming correct images or ideas of things; posing propositions well, that is, advancing propositions that are correct; inferring well; and ordering well. (PGI, 91; Jones, 1-2) Accordingly, logic is divided into four parts: the first concerns simple apprehension by nleans of images or ideas; the second, propositions; the third, syllogism; and the fourth, method. Each of the parts of Gassendi's logic is directed at pronloting the acquisition of certain skills. Part I is c~J!~1"n-~g __

12 FREDERICK S. MICHAEL with the skill of acquiring concepts and of judging those that have been acquired~ forming propositions and judging them is the concern of Parts II~ Part III is concerned with the skills of forming and judging arguments~ and Part IV is concerned with the skill by which propositions are effectively ordered. Each part consists of a series of canons or rules, each of which is followed with some commentary. The rules do not always take the form of precepts or injunctions~ often, as Gassendi himself points out, they are in the form of statements or theses proposed for our consideration and are called rules just because they are intended to be used by the mind as a guide to thinking well. Gassendi's definition of logic as ars bene cogitandi recalls Keckermann's ars recte de rebus cogitandi. Also the four parts into which he divides logic appear to come from the division of logic according to the three acts of the mind, except that argumentation and method, which are both forms of discursus, the third act of the mind, in Keckermann and other similar logics, are seen as separate parts of logic by Gassendi. But when we conlpare the parts of Gassendi' s logic with those of humanist logics, particularly the first part, dramatic differences become apparent. Humanist logics sometimes begin with a treatment of the predicables and categories, sometimes with an account of the places of invention and sometimes with both. In Gassendi's logic there is no account of the categories, none of the predicables and there is no consideration of the places of invention until the end of the Part II. The content of Part I of Gassendi's logic is entirely different from that of any humanist logic. Humanist logics begin by considering what kinds of things, or conceptions of things, there are, or by giving a more or less systematically arranged list of conceptions which it may be useful to consider when one is looking for reasons to support or oppose a thesis. Gassendi begins instead with an account of how ideas may be acquired and how to determine whether we have proper ideas of things, skills with respect to which previous logics have nothing to say. Logic is no longer conceived principally as a linguistic art or science, as it had been in medieval and humanist logics~ it is conceived epistenlologically. What we have in Gassendi is a new conception of logic, and in consequence of this, a new form of logic as well, the logic of ideas. This same conception and form of logic was soon to be adopted also by Arnauld and Nicole in their extremely influential Port Royal Logic. This form, the logic of ideas, was to become the dominant form of logic for the next two hundred years. The authors of the Port Royal Logic agree with Gassendi that the source of all knowledge is perception, that the theory of ideas is fundamental to logic and agree with him on the structure that logic should have. But they agree with him on little else. The title of the Port Royal Logic is Logic, or the Art of Thinking. In the second discourse, which is not in the first edition of the logic, the authors explain that they chose not to call logic the art of thinking well, as Gassendi had, because it takes no art to think badly. But the art of thinking well (I 'art de bien penser) is what they had originally called logic in the earliest version of ------------------------ -

-

-

-

-

-

-

-

-

-

-

--

Why Logic Became Epistemology 13 their work, which exists in manuscript only.32 The Port Royal Logic is divided into four parts: conception, judgment, reasoning and method. Conception is the simple view of things, without making any judgment about them. Judgment is the joining together of ideas. Reasoning forms one judgment from a plurality of others; and method is concerned with the ordering of thoughts. These are the same divisions that Gassendi recognizes, although in somewhat different terminology; and the description of the divisions is pluch the same. 33 Also, as in Gassendi, little attention is given to the categories or places of invention. There is a brief treatment of the categories in Part I and the subject is said to be of little use. The places are not discussed until part III and they are said to be of little value as well. Part I, as in Gassendi, is devoted to the theory of ideas. The account of ideas in the Port Royal Logic, although similar in structure to that in Gassendi's logic is in general opposed to it in content. Whereas the principles of the treatment of ideas in Gassendi's logic come from Epicurus, the principles of the treatment of ideas in the Port Royal Logic come from Descartes. The positions taken with respect to the nature of ideas in the Port Royal Logic often are in direct contradiction to that found in the Institutio Logica. This is not coincidental. Much in the Port Royal Logic is polemical in character, Arnauld was a major controversialist and Gassendi was one of the favourite targets of his invective. The authors of the Port Royal Logic sought not only to defend a Cartesian conception of knowledge and ideas, but also to discredit what Arnauld saw as the ridiculous, impious, and dangerous doctrines of Gassendi, particularly the doctrines that ideas are corporeal and that all knowledge derives from sensation. There is one other contrast between the Institutio Logica and the Port Royal Logic which it would be useful to recognize. Gassendi's objective is to compose a logic that is pure; pure, of course, according to Gassendi' s conception of what the nature of logic is. As a result it is concise and alnl0st austere, at least by comparison with the Port Royal Logic. This latter logic makes no attempt to remain true to its basic conception. In addition to the material that would be found in a pure logic of ideas, the Port Royal Logic incorporates large quantities of traditional logic. Moreover later editions of the logic include considerable material taken from the Port Royal Grammar of Arnauld and Lancelot. It also contains some writings on geometry by Pascal and Arnauld and even some of Arnauld's theological opinions. It is a very rich work, one which provides a good picture of seventeenth century thought in philosophy and other fields; and so apart from its value as logic, it is useful as an introduction t6 seventeenth century thought and to Cartesian thought in particular. This was a major reason for its success. Part I of the Institutio Logica consists of a group of 18 canons and has the following structure. The first three are concerned with the nature, origin and the formation of particular ideas, the next three with general ideas. There is then a set of four concerned with the perfection of ideas, then four concerned

14 FREDERICK S. MICHAEL with how to avoid being led into error and a final group of four concerned with definition and other matters. The Port Royal Logic adopts the main innovation of Gassendi which is to base logic on a theory of ideas and the presentation of the theory of ideas is structurally similar to that of Gassendi: there are analogues of Gassendi's account of singular ideas, general ideas, the perfection of ideas, the errors we can make with respect to ideas and of the conformity of definitions to our ideas. But the view taken of these subjects is in general very different from that of Gassendi. Part I of Gassendi's Institutio Logica is entitled, "Of the sinlp1e imagination ofthings.,,34 Simple imagination, Gassendi explains, is a form of simple appre.hension, the apprehension of a thing without anything affirmed or denied of the thing apprehended. Simple imagination is simple apprehension by means of corporeal images. As there can be falsehood only when sonlething is affirmed or denied, the simple imagination of things cannot be false. It is the simple imagination of a thing that Gassendi calls (in Canon I) its idea. The idea of a thing, Gassendi tells us is genuine, legitimate and true, when it confornls to the thing itself. This is not the thing as it really is in itself; fOI, Gassendi holds, its real nature is not known to us. What we want are ideas that will correctly represent things as our senses apprehend them. An idea is an image, because it is in the corporeal imagination and derives ultimately from sensations with which it may be compared, but it is at the same time a perceptible sign of something not itself perceptible, the thing as it is in itself. Also in order for our ideas to represent things correctly, we want them to be as clear and distinct as possible and the more frequent, recent and striking are the experiences we have of a thing, the more clear and distinct is the idea we form of it likely to be. A Cartesian clear and distinct idea guarantees that the idea we have of a thing represents that thing as it really is in itself; a Gassendist clear and distinct idea is obviously something quite different. The authors of the Port Royal Logic attack the view that ideas are images of things, claiming that there are ideas of things of which no image is possible, as, for example, the idea of a chiliagon. (PRL, Pt. I ch. I, 27-37) The correct view they clainl is that of Descartes, according to which ideas are what is in the nlind when we can truthfully say that we conceive of a thing, in whatever way we conceive it. They also reject the view that ideas are corporeal, going so far as to argue that no ideas, not even ideas of sensible things or qualities are corporeal and thus it is not necessary to have a body to feel the heat of hell. 35 Gassendi, following Epicurus, traces the origin of all of our ideas to sense. In Canons II and III, Gassendi incorporates in his logic the first of the Epicurean canons of the Anticipatio. Canon II holds that the original source of all our ideas is our senses. The mind is a tabula rasa on which nothing has been engraved or depicted, Gassendi says and those who hold that there are ideas that are naturally imprinted or innate, not acquired by sense, do not at all prove what they say. In support of the view that all our ideas originate in

Why Logic Became Epistemology 15 sensation, Gassendi offers the following inductive argument: blind men have no idea of colour and deaf men no idea of sound; so if there could be a person with no senses at all, that person would have no ideas either. (Jones, 4, 84-85; PG-I, 92b) Similar arguments are given by Locke, Hume and other empiricists. In addition to ideas acquired directly by sensation, ideas can be formed from ideas we already have by increase or diminution, as when from the idea of a person of normal size, we form the idea of a pygmy or giant; by composition, as when from the ideas of gold and a mountain, we form the idea of a golden mountain; and by analogy or similitude, as when by analogy with a city we have seen, we form the idea of one we have not seen. (Jones, 5, 85; PG-I, 92b93a) Ideas of incorporeals, such as God, according to Gassendi, are always analogical. Thus we form the idea of God from the image of some such thing as a grand old man or a blinding light. We do not of course believe that he is any of these things, only that in some respects he is like them. In the Port Royal Logic, Gassendi's view that all our ideas originate in sense and that ideas not directly derived from sense are formed by composition, increase and diminution and analogy are briefly described and then ridiculed. This is the way they argue: No proposition is more clear than "I think, therefore I am." This means that we must have clear ideas of thought and of being. But if Gassendi's position is true, these ideas must have been acquired by sense. But from what sense? And if they have been formed from ideas we already have, from what ideas have they been formed and how? If there is no answer, it is obvious that the theory advanced must be false. (PRL, 34-35) Yet at least with regard to being, Gassendi does have an answer; the idea of being is reached by abstraction. How abstraction works will be explained below. The Port Royal Logic not only denies that all of our ideas originate in sense, it inclines to the opposite extreme in suggesting that perhaps none of our ideas come from sense, sense supplying only the occasion for the mind to form ideas on its own. (PRL, 37) Also attacked is Gassendi's clainl that God, whom . we do not sense, can be represented analogically by the image of a grand old man, on the grounds that to conceive God in this way would nlake many of the beliefs we have about him false. (PRL, 35-36) Here having an idea of God as a grand old man appears to be misrepresented as the believing that God is a grand old man. What is perhaps the best known feature of the Port Royal Logic, its distinction between the extension and the comprehension of an idea, is actually developed from a distinction made by Gassendi. Gassendi is a nominalist. He holds that whatever exists is singular and consequently any idea derived directly from sense will be singular as well. The mind when it has many similar ideas can form general ideas from them in two ways, by collection and by abstraction. (Jones, 6, 86-87; PG-I, 93) The first way is by putting together the ideas of many sinlilar things and so forming the idea of the collection to which they belong; the second way is by abstraction, whereby we compare a group of

16 FREDERICK S. MICHAEL similar ideas, determine what features they have in common and, disregarding differences between them, construct a separate idea of the common features. The idea is general, since it represents the features which a group of singular ideas shares. Once general ideas have been formed, others still more general can be formed from these in the same two ways; and by proceeding in this manner, we ultimately reach the most general idea of all, that of being. Irrational aninlals, Gassendi believes, can form ideas of collections of things; they cannot, however, form general ideas by abstraction. In explaining general ideas, the authors of the Port Royal Logic, like Gassendi, observe that all that exists is singular, but there are general ideas formed by abstraction. They do not see abstraction as resulting from a comparison between similar ideas as Gassendi does; they see it instead as a matter of selective attention, the mind paying attention to some components of an idea, while disregarding others. So we can obtain the general idea of a triangle from an equilateral triangle, if we pay attention to its being a closed figure with three sides, while disregarding the equality of its sides. (PRL, 49f.) In his treatment of the formation of general ideas, Gassendi combines two conceptions of universals. One is an extreme nominalist conception which sees universals as being, or more properly, as signifying, nothing but a collection of singulars. The other is a more moderate nominalism which sees a universal as something constructed by the mind to apply generally to all the individuals in a particular collection. But what is in the mind is always something singular~ it only signifies universally. A general idea then is sonlething singular in the mind which signifies universally in two ways: by signifying a collection of singulars; and also by signifying the characteristics these singulars have in common. In the Port Royal Logic, the collection of singulars to which an idea applies is called its extension, and the group of characteristics in virtue of which it applies to these singulars is called its comprehension. (PRL, Pt.I, ch.V) In the Port Royal Logic, extension and comprehension are characterized as follows. The extension of an idea is the collection of things or species to which the idea applies; and its comprehension consists of its essential attributes together with all that these attributes imply. (pRL, 51) Extension and comprehension, the Port Royal Logic later shows, can be used in the clarification of propositions and logical relations. Thus although the source of the distinction is in Gassendi, it is in the Port Royal Logic that its logical significance was first developed and for that reason the Port Royal Logic deserves the principal credit for its introduction. Having completed his account of the acquisition and fornlation of ideas, Gassendi turns, in Canons VII through X, to the question of how to determine whether we have imagined well, that is, whether the ideas we have of things represent them properly. In general, Gassendi holds, the closer the idea of a thing conforms to the thing itself as we experience it, the more perfect it is. In

Why Logic Became Epistemology 17 place of the perfection of ideas, the Port Royal Logic considers its Cartesian counterpart, the clearness and distinctness of ideas. Ideas are said to be clear if and only if they are distinct. A clear idea of a thing is an idea of a thing as it actually is. Ideas of sensible qualities, since they are not of things as they really are, are obscure and confused. (PRL, 70-72) This means that none of what Gassendi considers ideas would be clear and distinct, since even the most perfect are of things as they are experienced, not of things as they actually are. The importance of having clear ideas is that clear ideas are always correct~ only ideas which are obscure and confused lead us into error. Gassendi of course denies that there are any ideas so clear as to ensure truth. Canons XI through XIV of Part I of the Institutio Logica concern the steps that must be taken to avoid error due to ideas which misrepresent things. (Jones, 12-14, 93-95~ PG-I, 96-97a) Though sense itself cannot be mistaken, it can deceive us. We avoid deception basically by being aware of the conditions under which sense deception commonly occurs and then take steps to determine if our perceptions are veridical. Thus, if a stick which appears bent when partly immersed in water, to determine if it is straight or bent we take it out of the water~ when a tower seen at a distance appears to be round, we determine if it is by going up to the tower~ and so on. Other common sources of error are temperament, passion, custom, prejudiced opinion, relying on the authority of unreliable witnesses, ambiguity and figurative language. Whenever an idea can lead us into error, this is always attributed, in the Port Royal Logic, to its being obscure and confused. Accordingly, to avoid error, we have to make sure to base our beliefs on ideas that are clear. But clear ideas need not be perfect, as the Port Royal Logic understands this. The idea we have of God, for instance, is clear, because it is correct as far as it goes, but it is imperfect because incolll..plete. (PRL, 71-72) The idea of a thing, Gassendi holds, determines our definition of it, as the second of the Epicurean canon concerned with ideas asserts, and also its division into species, parts and attributes. So the more perfect our ideas are, the more perfect our definitions and divisions will be. The idea of a thing, Gassendi adds, includes the relations it bears to other things. The authors of the Port Royal Logic accept all of this, but add a significant innovation, introducing the notion of real definitions (definition de chose), definitions which, unlike nominal definitions, can be false. (PRL, 8Off, 154ff.) In both of these logics, propositions are said to be concerned with the relation of ideas. What Gassendi says is that the mind, in attending to the various ideas it has, unites in an affirmation those that agree with one another and separates by negation those that do not agree~ the Port Royal Logic says sOInething very similar. Gassendi's account of argument begins with the claim that in a syllogism, the agreement or disagreement between the two ideas conjoined in its conclusion is shown by a thir~ ~~a~ ~~~e~s~
18 FREDERICK S. MICHAEL term of the two premises, which either agrees with the other two ideas or agrees with one and disagrees with the other. Key to the revolution in logic which Gassendi engineered, and to the logic of ideas in general, is a determination to base logic on concrete perceptions. Of course, the account of perception in the Port Royal Logic is very different from that given by Gassendi~ and the account of perception given by Leclerc in his logic is different again. In the two hundred years in which the logic of ideas was donlinant, it was in fact the theory of perception that was the principal focus of attention; it is this which was principally developed. And given the fact that philosophy was at the centre of the curriculum at most institutions of higher learning during this period and that a course of studies in philosophy usually began with logic, the theory of perception would be the first thing most students would learn. The extraordinary attention paid to the theory of perception during this period is thus easy to understand. But confining logic to concrete perceptions severely limited its prospects for development, just as insisting that all mathematicians confine themselves to purely finitary mathematics would be a serious impediment to development in mathematics. The humanist concentration on the informal aspects of logic was not harmful to the development of logic proper. It fulfilled a need. In the medieval period, informal aspects of logic had been largely neglected. It was the rigorous scientific ideal for logic that was pursued during this period. Formal and informal logic need not be seen as rivals, as they certainly were during the humanist period. Both types of logic are important; they complement one another. The logic of ideas, while not in itself illegitimate, was a barrier to development in logic. This was a revolution in logic that was not to the benefit of logic. But without it the momentous developments in epistemology of the next two centuries would probably not have taken place. NOTES 1. Howard Jones, Pierre Gassendi's Institutio Logica (1658) (Assen, The Netherlands: Van Gorcum, 1981) p. LXVII. This work will henceforward be referred to as "Jones," followed by page number(s). 2. Wilhelm Risse, Die Logik der Neuzeit (Stuttgart-Bad Carstatt: Friedrich Frommann Verlag, 1964) vol II, p. 79. 3. William and Martha Kneale, The Development ofLogic (Oxford, 1962) p. 298. 4. 1. M. Bochenski, A History of Formal Logic, translation by Ivo Thomas (Notre Dame, Indiana: Notre Dame University Press, 1961) p. 254. 5. 1. M. Bochenski, A History of Formal Logic, translation by Ivo Thomas (Notre Dame, Indiana: Notre Dame University Press, 1961) p. 258. 6. See The Didiscalicon of Hugh of St. Victor, translated by Jerome Taylor (New York: Columbia University Press, 1961). 7. See Ockham's Summa Logicae, Pt. I, ch. 3. Part I of the Summa Logicae is translated by Michael 1. Loux as Ockham's Theory of Terms (Notre Dame, Indiana: Notre Dame University Press, 1974). See pp. 52-54.

Why Logic Became Epistemology

19

8. For a discussion of ampliation, see Alexander Broadie, The Circle of John Mair (Oxford, 1985) pp. 76-80. For the humanist view on this subject, see the letter of Thomas More to Martin Dorp, 166-181 in Juan Luis Vives, Against the PseudoDialecticians, translation by Rita Guerlac (Dordrecht, Holland: Reidel, n.d.). 9. See Juan Luis Vives, Against the Pseudo-Dialecticians, translation by Rita Guerlac (Dordrecht, Holland: Reidel, n.d.). The whole of this work is devoted to the subject. But see in particular, pp. 55-57. 10. Until about 1520, most of the works published in logic are scholastic. After 1540, few editions of scholastic works in logic are published. Most are humanistic. This is documented in W. Risse Bibliographia Logica (Hildesheim: DIms, 1965) volume 1. 11. Cicero himself had said (in his Topica, ch II) only that logic or dialectic was a ratio disserendi, an account of discoursing. 12. Melanchthon, Erotemata Dialectices in his Opera Omnia (Halis-Saxonum, 1846) vol. XIII, p. 513. 13. Zabarella, Opera Logica (Cologne,1597) p. 32. 14. Keckennann, B., Opera Omnia (Geneva, 1614) pp. 549, 550. 15. Rudolf Agricola says this. See his De Inventione Dialectica, Libri Tres (1528) p. 155. See also W. Risse, Die Logik der Neuzeit (Stuttgart-Bad Cannstatt, 1964) vol.l, p. 17. This comprehensive work contains numerous selections from humanist logics, including selections from most of the logics considered here. 16. See P. Ramus, Dialectique (Paris, 1555) p. 2. 17. Keckennann, B., Opera Omnia (Geneva, 1614) p. 550. 18. CicerQ makes this division in his Topica, ch.2 19. Melanchthon, Erotemata Dialectices in his Opera Omnia (Halis-Saxonum, 1846) vol. XIII pp. 641-42. 20. In the first version of his logic (1520), Melanchthon barely sees a significant distinction between logic and rhetoric and he defines logic or dialectic as an artificium disserendi, a theory of discoursing. Its principle divisions are said to be finitio, divisio et argumentatio. By 1528, he characterizes logic as ars difiniendi,dividiendi et argumentandi and in considering whether invention should be considered part of logic or rhetoric, he concludes that either it is part of logic alone or of logic and rhetoric both. By 1534, he changes his mind about this, concluding that invention definitely is part of rhetoric but that it mayor may not be part of logic. Finally, while he does not address this question directly in 1547, he appears to dismiss the distinction between invention and judgment as the old division. Yet the structure of his book, in all of its versions, appears to reflect this division; and Melanchthon never ceases to think that the places are important, holding that they should be our first concern, useful as a guide both to what is to be investigated and to what is to be chosen. For this reason, he devotes the last book of his logic to a detailed consideration of the places, even when he does not consider them part of logic proper. The effect of this is to produce a logic which, at least in practice, incorporates both the Aristotelian and the Ciceronian conception of logic. 21. Melanchthon, Erotemata Dialectices in his Opera Omnia (Halis-Saxonum, 1846) vol. XIII, pp. 574-78. 22. Zabarella, Opera Logica (Cologne,1597) p.53. 23. Zabarella, Opera Logica (Cologne,1597) p. 65. 24. Keckennann, B., Opera Omnia (Geneva, 1614) p. 551. 25, Keckennann, B., Opera Omnia (Geneva, 1614) pp. 128-29. 26. Keckennann, B., Opera Omnia (Geneva, 1614) p. 737.

20

FREDERICK S. MICHAEL

27. See The Selected Words of Pierre Gassendi, edited and translated by Craig Brush (New York and London: Johnson Reprint Corporation, 1972) pp. 30-42. There is a complete translation of the Paradoxical Exercises into French with the Latin text on facing pages. This is the Dissertations en Forme de Paradoxes Contre Les Aristoteliciens, translated by Bernard Rochot (Paris: Vrin, 1959). The Latin text is found in volume III of Gassendi's Opera Omnia (Lyons, 1658). 28. See P. Gassendi, The Selected Words ofPierre Gassendi, edited and translated by C. Brush (New York and London: Johnson Reprint Corporation, 1972) p. 284. 29. See P. Gassendi, The Selected -fVords of Pierre Gassendi, edited and translated by C. Brush (New York and London: Johnson Reprint Corporation, 1972) pp. 326ff. 30. This set of rules can be found in Gassendi,s Opera Omnia, vol I, pp. 53-56. Volume I will in the future be referred to as "PG-I". The canons of sense are: 1. Sensus nunquam fallitur; ac proinde est omnis Sensio, omnisque Phantasia, seu apparentiae preceptio, vera. 2. Opinio est consequens sensum, sensionique super-addiecta, in quam veritas et falsitas cadit. 3. Opinio ilia vera est, cui vel sufJragatur, vel non refragatur sensus evidentia. 4. Opinio iUa falsa est, cui refragatur, vel non sufJragatur sensus evidentia. The canons of the anticipatio are: 1. Omnis, quae in mente est Anticipatio, seu Prenotio dependet a sensibus, idque vel incursione, vel proportione, vel similitudine, vel compositione. 2.Anticipatio est ipsa rei notio, et quasi definitio, sine qua quicquam quaerere,dubitare, opinari, imo et nominare non licet. 3. Est Anticipatio in omni ratiocinatione principium; quasi nempe id,ad quod attendentes, inferemus unum esse idem, aut diversum,' coniunctum aut disiunctum ab alio. 4. Quod inevidens est, ex rei evidentis anticipatione demonstrari debet. 31. See PG-I, p. 91 and Jones, p. 1. This reference is to Jones's edition of the Institutio Logica, not to his translation. There is a French translation of the Institutio Logica by Franvois Bernier; it occupies volume I of his Abrege de la Philosophie de Gassendi (Lyons, 1684). 32. See the critical edition of the Port-Royal Logic by H. Brekle (Stuttgart-Bad Canstatt, 1967), vol. ill, p. 12. This work is referred to in the future as "PRL". 33. PRL, pp. 23-25. This and all later citations are from the first edition first published in 1662, which resembles the Institutio Logica most closely in structure. In later editions, much new material was added and the original structure is somewhat altered. For instance, Part II in the first edition begins with a general discussion of propositions; in later editions, this is preceded by two chapters on grammatical topics, Part IV, in the first edition, begins with a chapter distinguishing two kinds of method; in later editions, this is preceded by a chapter on science in general. No revisions in text were made after the fifth edition (1683) and it is in this final fonn, that the Port-Royal Logic is best known. 34. For Pt.I of the Institutio Logica, see Jones, pp. 3-20 (Latin), pp. 80-101 (English); PG-I, pp. 92-99. 35. This is not in the first edition. In the fifth edition, it is in Part I, chapter IX (a revised version VITI of the first edition).

The Workings of the Intellect: Mind and Psychology* GARY HATFIELD

The narrative structures within which we describe the origin and development of early modern philosophy at the. same time reveal something about what we find interesting and valuable in that philosophy. In recent decades, the older trend of characterizing early modern philosophy as a triumphant "Age of Reason" has given way to the organizing theme of a skeptical crisis and the responses to it. According to the earlier story, in the seventeenth century Reason cast off the yoke of Church authority and Aristotelian orthodoxy; newly-freed thinkers revitalized philosophy, created the "new science," and pushed on toward Enlightenment.} Now, however, it is more popular to speak of a skeptical crisis in the sixteenth and seventeenth centuries, which set the philosophical task of "refuting the skeptic" for subsequent generations. Lacking a compelling response to skepticism, philosophers were forced to retreat, and they proposed ever narrower "limits to knowledge" until Kant took a last stand on the redoubt of transcendental idealism. 2 Both descriptive stories portray "epistemology" or theory of knowledge-often allied with a concern for method-as the defining preoccupation of early modern philosophers from Descartes through Kant. In describing this "epistemological turn," story tellers from Thomas Reid through Richard Rorty have given pride of place to the "theory of ideas,,,3 though others have properly recognized the role of metaphysical concepts, including the concepts of "substance" and of "necessary connections" between properties or events. 4 There can be no doubt that these elements-anti-Aristotelianism, skepticism, method, knowledge, substance, and necessity-must all be found in any account of early modern "metaphysics and epistemology," as we often, but anachronistically, label the theoretical (as opposed to practical) philosophy of the seventeenth century. I wish to show that they can be combined into yet a third narrative, one that begins by taking seriously seventeenth-century conceptions of the topics and methods central to the rise of a "new" philosophy. In this revisionist story, differing approaches to the central subject matter of early modern metaphysics-knowledge of substances through their essences and causal powers-arise as a result of disagreements about the powers of the human cognitive faculties. 5 Methodological writings are seen as attempts to direct readers in the proper use of their cognitive faculties. The early modern rejection of the Aristotelian theory of cognition ranks equally in importance with rejection of Aristotelian doctrines about nature. Skepticism is more often than not a tool to be used in teaching the reader the proper use of the cognitive faculties, or indeed in convincing the reader of the existence or inexistence of certain cognitive faculties or powers. Instead of early modern "epistemology" or "theory of knowledge," one speaks, along with seventeenth-century writers, of

22 GARY HATFIELD theories of the cognitive faculties or knowing power. The early modern rejection of Aristotelian logic can then be seen as reflecting a negative assessment of the fit between the syllogism and logic considered as an art of reasoning that refines the use of the cognitive faculties. References to "reason" and "the senses," which, in the traditional historiography, are typically understood as shorthand for "a priori propositions" and "empirical evidence," can now be seen as references to cognitive faculties. When described in its own terms, the development of philosophy from Descartes to Kant may be seen as a series of claims about the power of the intellect to know the essences of things, with resulting consequences for ontology and for the role of sensory cognition in natural philosophy. Thus, Descartes employed skepticism as an artifice in order to bring his readers to an awareness that (as he claimed) the faculty of the intellect, contrary to Aristotelian doctrine, can be exercised independently of sensory images and their content. Having revealed the power of the intellect to operate independently of the senses in grasping the "cogito" reasoning, he next exercises this power in contemplating God-without, of course, the aid of sensory content-and then in discerning the foundations for a new natural philosophy; only subsequently do the senses play an essential role in the investigation of nature. Spinoza and Leibniz each looked to pure intellect to achieve his own revised metaphysical picture. When Locke tried to follow, he became convinced that the power of the intellect or understanding is more· restricted than either the Aristotelians or Descartes had claimed: in particular, he found that the understanding cannot discover real essences within sensory experience, and also that it cannot achieve any content independently of sensory experience (or reflection thereupon). Berkeley nlounted a direct attack on the use of the intellect to know matter, denying the very intelligibility of material substance as understood in Cartesian metaphysics, but he affirmed the power of the intellect to know spiritual or immaterial substances. Hume continued Locke's inquiry into limits on the powers of the human understanding, arriving at the conclusion that it is unable to know the substances and causal powers of traditional metaphysics. Hume held that the operation of the understanding is limited to two separate domains: reasoning about relations of ideas, where intuitive and demonstrative knowledge can be obtained, but without thereby achieving any knowledge of facts, or reasoning about facts known through the senses, which provides no rational insight into the substances and causal connections of traditional metaphysics, with the consequence that the understanding is here limited to charting successions of sensory perceptions. Kant entered his critical period when he realized that human cognizers do not have available the "real use" of the intellect or understanding to know an intelligible world of substances; at the center of his critical (theoretical) philosophy was his new theory of the human understanding as a faculty limited to synthesizing the materials of sensory representation but unable to penetrate to things in themselves, with the

The Workings ofthe Intellect: Mind and Psychology 23 consequence that knowledge of necessary connections could be attained only within the bounds of transcendental idealism. 6 It is not my intention to put fonvard this revised narrative as a single nlaster story for early modern philosophy. Indeed, beyond the three narrative themes sketched so far, others might be suggested in which differing subsets of philosophers would play greater or lesser roles; these include the story of the changing relations among metaphysics, theology, religion, and science (here Malebranche would enter prominently), and the relation of metaphysics. and theory of mind to moral and political philosophy. My aim is to illustrate the· force of one particular revised narrative by using it in a comparison of three conceptions of the intellect, conceptions respectively held by sonle scholastic Aristotelians, Descartes, and Locke. These three examples are not intended to yield an exhaustive taxonomy of early modern theories of the intellect, nor have they been chosen for what they might contribute directly to a present-day theory of the intellect. Rather, discussion of these conceptions will demonstrate the central role played by the theory of the intellect (and other cognitive faculties) in three prominent theoretical philosophies of the early modern period, and it will clarify the point of some early modern disputes. It will also offer an opportunity to locate early modern discussions of the cognitive faculties with respect to recent understandings of psychology, epistemology, logic, mind, and their relations. The early modern discussions are not easily fit into the modern categories of epistemology and psychology. Reflection on this fact may help us see some problems in recent conceptions of naturalism as applied to philosophy and psychology. In this way, contextually sensitive historical reflection contributes directly to contemporary understanding.

1. Three Conceptions of Intellect Theories or conceptions of the intellect are indicators or even determiners of the scope and limits ascribed to theoretical philosophy by their holders. If one thinks that the intellect has access to eternal Forms or that it can discern the essences of things, one might well have great hopes for the discipline of metaphysics and related theoretical pursuits in natural philosophy. Conversely, if one holds that the power of the intellect is limited, that essences are hidden and unknowable, then one will, by traditional standards, have a nlodest conception of what can be done in metaphysics and natural philosophy, though one might also be led to revise the aims of those disciplines or to propose a new vision of the proper content of natural philosophy, as did Locke, Hume, and Kant. As the early modern period began, Aristotle's theory of intellect was predominant. His De anima analyzed the powers of psyche or soul, understood as an animating principle possessed of vegetative, sensitive, and (in humans) rational powers. It devoted greatest attention to the cogn!!i.y~jJ~F~~_~f_th~_

24 GARY HATFIELD soul, especially the senses and intellect. Aristotle's doctrine of the intellect had taken on a particular fascination for late antique and Arabic commentators, and parts of Book III, chs. 4-5-especially where he said that there is an element of thought that is capable of "making all things" and another capable of "becoming all things" 7-were extensively elaborated. Interpreters dubbed the first power the "active intellect" and the second the "patient" or "passive" intellect. They offered diverse theories of the natures of these intellectual powers, including the theory that there is one active intellect for all human beings. Although the latter position did have some adherents in the Latin West, the orthodox view attributed individual active and patient intellects to . individual human beings. 8 As a background to early modern philosophy, my interest here is in late scholastic Aristotelian theories of cognition, rather than in the interpretation of Aristotle per se. Much of the De anima is organized as a theory of cognitive faculties. Late scholastic Aristotelian theories (following Aristotle) strictly separated the sensitive and intellectual powers of the soul. According to such theories, the sensory power always relies on corporeal organs, but the intellect (it was usually held) does not, it being an in1material power of the forn1 of the human body. The primary function of the Aristotelian intellect is to abstract essences or COll1mon natures from the in1ages received by the senses. In accordance with the dictum that "there is nothing in the intellect that was not first in the senses," this act of abstraction depends on sensory images or "phantasms" for its operation. Central interpreters of Aristotle-from Thomas Aquinas to such late scholastics as Suarez, the Coimbra Commentators, Rubio, and the textbook author Eustace of St. Paul-all held that there is "no thought without an image," that is, that each act of intellection requires a material image drawn from the senses and actually present in the imagination or "phantasia."9 Aristotelian theories of cognition describe a chain of events starting from external objects and ultimately resulting in the reception of an "intelligible species" in the patient intellect. External objects produce "intentional species" in the medium between them and the cognizer~ an oak tree thus produces species of brown bark and green leaves. These species are received by the senses and conveyed to the imagination. Then the intellect, perhaps operating over several species received across time, abstracts the essence or common nature of the oak tree. Systematic knowledge, or scientia, is of the common nature or the universal, not of the particular. 10 Beyond this general description, a further and misleading tenet is often ascribed to late scholastic Aristotelian theories of cognition: viz., that the process by which the common nature is "abstracted" amounts to an "absorption" of the species from the senses and imagination into the intellect. On this interpretation, it is as if, as the term "abstraction" itself might suggest, the intellect simply received the "form" in the species separated from all material conditions. 11 Intellection would simply be a kind of dematerialization,

The Workings ofthe Intellect: Mind and Psychology 25 or an extraction of a form from the still-material representations of the senses and its transferral to the patient intellect as an intelligible species (a conception that is indeed suggested by the common turn of phrase that the active intellect "illuminates" the phantasm). There would be no need to explain how intelligible species are "created" by the active intellect~ the latter's agency would simply be that of preparing the form in the material phantasm ~or transfer to the patient intellect. None of the interpreters of Aristotle cited above held that the intellect absorbs a form from the imagination or phantasia. Consonant with the principle that a "lower being" such as matter cannot act on a "higher being" such as the immaterial intellect,12 these authors all affirmed that intelligible species are produced in the patient intellect by the causal power of the active intellect, which can "make all things"~ the material phantasm serves as a "material," "instrumental," or "partial" causal factor. "Abstraction," therefore, should not be equated with "extraction." Aquinas put the point as follows: Phantasms, since they are likenesses of individuals, and exist in corporeal organs, do not have the same mode of existence as does the human intellect (as is obvious from what has been said), and therefore are not able to make an impression on the patient intellect by their own power. This is done by the power of the active intellect, which, by turning toward the phantasms, produces in the patient intellect a certain likeness that represents, as regards specific nature only, that of which the phantasms are phantasms. And it is in this way that the intelligible species is said to be abstracted from the phantasms; not as though a fonn, numerically the same as the one that existed before in the phantasms, should subsequently come to be in the patient intellect, in the way a body is taken up from one place and transferred to another. 13

The position that the corporeal phantasm, being material, cannot of itself be received into or affect the immaterial intellect was accepted by each of the other authors. 14 More generally, these authors saw the active intellect's ability to "make all things" as playing an important explanatory role: it explains how the intellect can abstract common natures from imperfect sensory images. Without . adopting a doctrine of innate ideas, and while affirming that the patient intellect is a tabula rasa, these authors could hold that the active intellect brings something to the creation of intelligible species. 15 As Aquinas put it, the light of the human intellect is a "participating likeness" of the "uncreated" (divine) light that contains the eternal types. 16 Far from simply absorbing its content from "phantasms," the intellect has a dispositional capacity to create intelligible species that reflect the eternal types ~ but (so they argued, appealing to introspection, among other considerations), it cannot do so without the presence of an appropriate phantasm. The essential role assigned to corporeal phantasms in the operation of the intellect placed limits on the cognition of immaterial entities such as God and

26 GARY HATFIELD the soul. There are no sensible species, and hence no phantasms, of such entities. Consequently, those who accepted this account of the intellect held that in this life human beings can at best achieve a confused intellectual cognition of God or the SOUl, by reasoning fronl creation to creator or from the soul's bodily operations to its nature and powers. Francisco Toledo, whom Descartes would later remember from his school days, contended that an embodied intellect "cannot naturally possess clear and distinct cognition of immaterial substance" ~ Aquinas, the Coimbrans, Rubio, and Eustace said similar things. 17 Authors in this tradition developed elaborate analyses of how God and the immaterial soul can be known, given that their theory of intellection precluded clear and evident cognition of them. The doctrine of analogy is one instance of such analysis. According to a prominent form of Aristotelianism, then, systematic knowledge or scientia is of universals or common natures, cognized by means of intelligible species which themselves can be formed only with the aid of sensory images. The ability of the intellect to form representations of the essences of things cannot be explained by its simply "taking up" the content provided by the senses, or even by its sifting through and comparing sensory images. The intellect is an immaterial power that cannot be affected by the inherently corporeal activity of the senses, but which is able to make intelligible species with the cooperation of sensory images. This ability was taken to reflect a similarity between the human active intellect and the divine intellect, containing the eternal types. The things best known by the human intellect are the substantial forms or common natures of corporeal things. Immaterial entities are cognized only confusedly in this life. Descartes, who was well-schooled in this tradition, turned nearly every tenet of this theory of cognition on its head. In particular, he held that the intellect can operate independently of the senses and imagination, and that in so doing it can achieve "clear and distinct" cognition of God, the soul, and matter. Whereas sense and intellect were markedly distinct faculties for the Aristotelians, with the intellect depending on sense, for Descartes intellect was the only essential cognitive faculty, sense and imagination being "modes" of intellection, arising from mind-body union. I8 Intellect can operate independently of the senses-when it is known as "pure intellect"-but·· sense perception (in humans) is an operation of the intellect (broadly construed). Thus, beyond his notorious rejection of Aristotelian physics, Descartes also rejected the Aristotelian theory of cognition, including especially the view that intellectual cognition requires sensory images. I believe that this rejection was first consolidated in 1629 or 1630, simultaneous with Descartes's discovery of his mature metaphysics. I9 His new theory of cognition became an essential bridge to his metaphysics, in that he appealed to the deliverances of the intellect, given independently of the senses, to convince his readers of important new metaphysical doctrines, including his assertion that the essence_ of --------------------------------------------------------

_

The Workings ofthe Intellect: Mind and Psychology 27 matter is extension Descartes's concern with method, which has often been linked to "epistemology," in fact reflects his efforts to train his audience in the proper use of their cognitive faculties. Descartes crafted the primary statement of his metaphysics, in the Meditations on First Philosophy, as a tool for bringing his readers to a discovery that the pure intellect is a faculty best exercised independently of sensory content. In adopting the meditational structure, Descartes chose to pattern his work after devotional literature or spiritual exercises, a literary genre that paradigmatically employed a theory of the faculties to order the meditator's search for God. 21 In Descartes's hands, the structure of this devotional genre was turned toward a cognitive end: that of attaining knowledge of first principles through proper use of the intellect. 22 In order to reach the cognitive states toward which Descartes was leading them, Aristotelians such as those canvassed earlier would have needed to be convinced that there can be thought without a phantasm, or at least they would have needed to be induced to have such thoughts. To this end, Descartes begins his meditations with a skeptical purging ofthe senses (and even the evident cognitions of arithmetic and geometry), resulting in the discovery that only the thinking "I" itself cannot be doubted. He then explores the nature of this "I," finding that it consists in thinking alone. In the midst of this exploration, Descartes has the meditator reflect on the prospect of using the faculty of imagination-a faculty essential to all human intellectual cognition according to the Aristotelians-to know the soul. Part way through the Second Meditation, while still contemplating the "I," the meditator has the following insight: 20

It would indeed be a case of fictitious invention if I used my imagination to establish that I was something or other; for imagining is simply contemplating the shape or image of a corporeal thing. Yet now I know for certain both that I exist and at the same time that all such images and, in general, everything relating to the nature of body, could be mere dreams. Once this point has been grasped, to say "I will use my imagination to get to know more distinctly what I am" would seem to be as silly as saying "1 am now awake, and see some truth; but since my vision is not yet clear enough, 1 will deliberately fall asleep so that my dreams may provide a truer and clearer representation." 1 thus realize that none of the things that the imagination enables me to grasp is at all relevant to this knowledge of myself which 1 possess, and that the mind must therefore be most carefully diverted from such things if it is to perceive its own nature as distinctly as possible. (CSM IT: 19, AT Vll: 28)

He then proceeds to list the activities of thought that belong to himself as a thing that thinks: doubting, understanding, affirming, denying, willing, and seeming to imagine and to sense. Notoriously, the meditator then notices that corporeal things still seem better known than "this puzzling 'I' which cannot be pictured in the imagination." (CSM II: 20, AT VII: 29) So he begins to contemplate wax as an instance of body, thereby discovering in himself a

28 GARY HATFIELD faculty distinct fronl the imagination and able to grasp the infinity of shapes that melted wax can take. The meditator then reflects that this faculty is implicated in every act of cognition, even those that are described as sinlple acts of seeing. In the Second Meditation he simply characterizes this faculty as "the mind alone," and its operation as a "purely mental scrutiny." (CSM II: 21, AT VII: 31) At the beginning of the Sixth Meditation he again distinguishes the faculty that can grasp many geometrical figures from the faculty of imagination. Here he puts a name to this faculty: it is "intellectio pura," i. e., the "pure intellect" (or "pure understanding," in the words of Cottinghanl et al.). (CSM II: 50-51, AT VII: 72-73) Pure intellect is, by Descartes's lights, one of two faculties essential to nlind (the other being will), and it is the faculty by which the essences of mind and matter are discerned, and by which God is • known. 23 Descartes's conception of the intellect, then, is absolutely central to his philosophy. Just as in the Aristotelian framework, the question arises of how Descartes could account for the intellect's ability to grasp the essences of things, and for him the question seems all the more pressing, since he alleged that the intellect can do so independently of sensory contact. This question is a correlate to one later posed by Kant, who asked how the understanding could ever cognize objects, as regards their substance and causal connections, independently of the senses (which, by themselves, he considered inadequate for the task). (CPR A85-941B118-127) Platonist philosophers had maintained that the human intellect attains knowledge of the essences of things via cognitive access to eternal Forms, or to archetypes in the mind God, or else to copies of those archetypes implanted in human n1inds. They posited a "preformation-system of pure reason," in Kant's words, among eternal Forms or essences, the things in the world that participate in them, and the objects of hunlan intellection. 24 Descartes, however, rejected this conception of the link between mind and world. In connection with his doctrine of the creation of the eternal truths, he forsook the claim that the so-called "eternal truths" pertaining to created things reflect the basic structure of the divine understanding. Rather, these truths are created, just as are the things. 25 The access that the human mind has to these created essences is still explained by a "pre-established harmony," enacted by God's will, between created substances and their essences as known by pure intellect. Descartes retains a divine role in explaining the functioning of pure intellect, without needing to claim that the human intellect, and the knowledge of natural things gained by it, reflect the divine understanding. 26 In this doctrine the relations among essences, minds, and things become tightly bound, and hence the theory of intellectual cognition itself beconles a part of metaphysics. In comparison with the Aristotelian and Cartesian conceptions, Locke attributed to the human mind a weak intellectual candle. Although showing signs of nostalgia for knowledge of real essences, Locke grudgingly admitted

The Workings ofthe Intellect: Mind and Psychology 29 that such knowledge is beyond our ken. He came to this conclusion in a work entitled An Essay Concerning Human Understanding, a title in which the word "understanding" is not a gerund referring to the activity of understanding, but a count noun referring to the faculty of understanding. 27 Yet curiously, despite this fact, and unlike our Aristotelians and Descartes, Locke does not layout in a systematic fashion his conception of this faculty and its relation to the other faculties. He uses the word "understanding" in Descartes's broad sense, to denote the "perceptive power" of the mind, as distinct from will. (E ILxxi.5-6) He does no more than list a number of faculties exhibiting this power, including perception, contemplation, memory, discerning, comparing, composition, enlarging, and abstraction (but no separate faculty of pure understanding, in Descartes's narrow sense). (E II.ix-xi) This lack of a systematic theory of the metaphysics of the faculties and their powers is perhaps consonant with Locke's belief that the power of the understanding itself is limited and so is not able to deternline its own nature-any more than it can, more generally, deternline the natures of mind or matter. (E ILxxiii) Thus, Locke's restriction of his inquiry to the "plain, Historical method," a method of observation based in experience, even though coming at the beginning of his Essay, reflects an important conclusion of that work: that human knowledge can be based only on experience, not on purely intellectual cognition of the sort claimed by Descartes. To that extent, his "empiricism" reflects a direct and substantive disagreement with both Descartes and the Aristotelians concerning the power of the hunlan intellect. A principal aim of Locke's Essay was to discern the bounds of the understanding's power, to learn the "Extent of its Tether." (E Li.4) Sonle of his most vigorously argued conclusions pertain to what the understanding can't do, or doesn't possess. Thus, he argues, contra Descartes and others, that the understanding possesses no innate ideas and knows no principles innately. (E I.ii) The content of thought must come from the senses or from reflection on the operation of the mind in connection with sensory materials: from either "external" or "internal" sensation. (E ILi.2-4) Human cognition is limited to sensory ideas, or images. 28 But, contrary to the Aristotelians, Locke does not find that the understanding, in operating upon sensory inlages, has the power to extract the "common natures" or essences of things. (Locke was in any case strongly dubious of the existence of Aristotelian "substantial forms," a notion that he found unintelligible-E III.vi.l0.) In the end, he decided that knowledge of real essences of substances is beyond us. (E IILvi.6, 9) In his view, "abstraction" yields general ideas that can denote many particulars, but we achieve general ideas only of what he termed simple or mixed nlodes, or nominal essences-general ideas either of a single type of simple sensory idea such as a color, or such as are produced through a combination of such simple ideas (E III.iii-vi)-but not of the real essences of substances. 29 Further, "intuitive" and "demonstrative" knowledge, to which Locke attributed the

30 GARY HATFIELD highest degree of certainty, are found only in clearly perceived relations among ideas. (E rVOiii.1-5) Since we have no idea of the real essences of substances, we are unable to achieve intuitive knowledge of the relation between property and essence-the best we can do is to achieve intuitive certainty with respect to "visible connections" among some of the primary qualities of things, such as the connection according to which figure presupposes extension. (E IV.iii.14) Locke's Essay is an intricate web of argument and assertion, comprising other factors besides the theory of the faculties, including ordinary appeals to cognitive virtues such as clarity (appeals that can be assessed for themselves without the need to draw upon a theory of the faculties). Still, appeals to the powers and limits of human cognitive faculties play an important role, even in those parts of the work that are not specifically directed toward an analysis of cognition itself. In particular, Locke repeatedly invokes limitations on "Our Faculties" in explaining the failure to know real essences. (E III.vi.9) There are at least three aspects of this failure. First, there is a failure to know the corpuscular constitution of things (on the assumption that the "real essences" of bodies are corpuscular),30 which may in part be due to remediable causes, such as lack of experiments, but in other cases is due to a lack of sensory acuity for perceiving the minute constitution of bodies, or (Locke speculates) perhaps even a lack of the appropriate kind of sense organ. (E IV.iii.23-25) Second, even if we could perceive the "real essence," we are very limited in our cognitive ability to grasp any connection between that essence and the properties that flow from it (E III.vi.19), as regards both primary and secondary qualities. (E IV.iii.12, 29) Third, "we may be convinced that the Ideas, we can attain to by our Faculties, are very disproportionate to Things themselves, when a positive clear distinct one of Substance it self, which is the Foundation of all the rest, is concealed from us." (E IV.iii.23) Having limited the contents of cognition to simple sensory ideas and their combination, and having restricted the cognitive powers to those that perceive, store, compare, and combine such ideas, Locke found that the human mind is incapable of grasping real essences, either of minds or of bodies. He did make one seemingly metaphysically ambitious claim, to demonstrate the existence of a supreme intelligence, creator of the world; but in this demonstration all cognitive access to God comes via inference from created things, reasoning by analogy with the actions and attributes of human minds. (E IV.x) Locke in effect held that the human intellect lacks the cognitive resources to succeed at the tasks of traditional metaphysics. This being the case, he, by contrast with the Aristotelians and Descartes, had no need to explain how the understanding can grasp the essences of things. Long before Kant, then, the Lockean intellect had already forsaken any bid to know the "things in themselves" (substances as they are in themselves). Kant presented a fuller range of arguments for a more definitive version of this conclusion, and he constructed an account of knowledge in which our

The Workings ofthe Intellect: Mind and Psychology 31 knowledge of nature meets the criterion of scientia as an organized body of necessary and universal propositions. Locke, by contrast, has the knower still trying to grasp the real essences of mind-independent objects, and simply coming up short. Although Kant admired Locke's analysis of the faculties of cognition, he felt that Locke had misunderstood the role of the faculties in metaphysical cognition, and had pursued the investigation incorrectly, by making it empirical. (CPR A86-87/B 119) Kant also limited the materials upon which the understanding can operate to the representations of sensibility, but he attributed a set of categories to the understanding that rendered such representations into cognition of a law-governed world of nature, ordered in space and time. He gave up claims to know the intelligible world of things in themselves, in order to gain title to knowing an ideal but comprehensible world of nature. 31 2.

Mind and Psychology

Philosophers of the early modern period, whether conceiving of themselves as metaphysicians or as inquirers into the grounds and limits of human knowledge, proffered theories of the cognitive faculties. These were theories of the senses, imagination, and intellect, among others. Viewed from the standpoint of the twentieth century-and especially that of our middle decades-this penchant for investigating the mind has seemed like an embarrassment to philosophy, like an early version of the fallacy of "psychologism."32 Consequently, many recent philosophers have deemed it best to ignore or minimize the allegedly outdated "faculty psychology" of the early moderns. This charge of psychologism provides an interesting lesson in the ironies of anachronism. The indictment of "psychologism" relies on an assimilation of early modern theories of cognition to recent conceptions of mind, psychology, epistemology, and their relations. It thereby misreads the substantive positions of the early modern authors, and then, on the grounds of this misreading, charges those same authors with errors they did not commit, while at the same tinle failing to detect their real mistakes, or at least our real differences with them. Psychologism is a species of the naturalistic fallacy. The alleged "fallacy" lies in the move from fact to norm, from descriptions of how things are-for exanlple, with patterns of human behavior, or with habits of hunlan thoughtto conclusions about how things ought to be. Thus, even if most people lie, that doesn't make lying morally correct. Moral philosophy and epistemology respectively speak to how we ought to behave or what constitutes good warrant for belief, in spite of what empirical study may show about actual behavior or belief formation. The contention that the psychologistic inference from actual pattern of thought to norm for thought is a "fallacy" assumes a particular philosophical position. It assumes that our innate patterns of thought do not in fact reflect and

32 GARY HATFIELD thereby manifest norms for good thinking. By the late nineteenth century this assumption may have possessed good philosophical warrant. Of interest here is the fact that the early modern authors discussed herein, including the Aristotelians, Descartes, and Locke, all rejected this assumption. According to the Aristotelians, the natural human faculties by themselves tend toward true cognition. Logic, in their view, was an artificial system for aiding and improving cognition. It systematized the norms jm_plicit in actual human reasoning, and provided aids for avoiding error. 33 Similarly, Descartes considered the deliverances of pure intellect to directly present the truth. He took the "impulses" of the will to affirm clear and distinct intellectual perceptions as the . sure sign of the truth of those perceptions. He held that the "natural" intellectthe intellect we have by nature-sets a· norm for good thinking, because its proper use cannot fail but to achieve truth. 34 Within such a framework, the move from mental fact to cognitive norm is warranted. Locke, too, accepted the workings of the "discerning faculties" as constitutive of right thinking, (E IV.i.2) though he made the weakest claims for the scope of the truth-discerning power of the human intellect. Perhaps because the Aristotelians and Descartes each made such strong claims for the power of the intellect, they both attempted to explain why the deliverances of the intellect could be trusted: Aquinas appealed to the "participation" of the hunlan intellect in the "uncreated light" of the divine intellect, and Descartes to God-installed innate ideas and faculties ofjudgment. Given that early modern authors investigated mental faculties in connection with method, metaphysics, and the theory of soul, shall we conclude that they were engaged in psychology? Was their investigation naturalistic, and if not, what was it? Supernaturalistic? And if we reject Descartes's claims for the intellect, is that because we think he was a bad psychologist, or is it because we have more substantive disagreements with him over the powers of human cognition, and the existence of substances constituted with intelligible essences? These questions, like the charge of psychologism, invite us to reflect on the fit between (on the one hand) our conceptions of the natural, the psychological, and the mental, and (on the other) the corresponding early modern conceptions. Let us begin with psychology. The name derives fronl the account of the soul, "logon peri tes psyches," as pursued by Aristotle~ in the Middle Ages this discipline was most known under the latinate label "de anima," but from the sixteenth century on it was sometimes latinized as "psychologia. ,,35 The subject matter of "de anima" psychology, determined as it was by the Aristotelian conception of soul, included the nutritive, motive, sensory, and rational faculties of animate or ensouled beings. Nonetheless, in the textbooks and De anima commentaries of the early modern period, as in Aristotle's own text, greater attention was given to the cognitive faculties, sensitive and rational, than to the others. The material conditions of the operations of the senses were charted,

The Workings o/the Intellect: Mind and Psychology

33

cerebral anatomy was discussed, and some mention was made of the cognitive division of labor among the external and internal senses, the estimative power, and the active and patient intellects. 36 Within the Aristotelian curriculum, the theory of the soul fell under the rubric of physics, or natural philosophy. The soul was considered part of nature. 37 Only in the discussion of the immaterial intellect was there a tendency to consider supranatural explanatory agencies, as in the doctrine of the unity of the active intellect. The Aristotelians discussed above rejected this doctrine, affirming that the active intellect is a natural, if immaterial, power of the human soul, where the latter is regarded as the form of a corporeal substance, the human being. Already we can tell that our categories "natural," "physical," and "psychological" do not easily map the Aristotelian position, in which an immaterial power is considered to be part of nature, and indeed, to form a portion of the subject matter of physics, understood as the science of all natural things. Perhaps even more seemingly odd, Antoine Le Grand, a dualist follower of Descartes, ranged the theory of mind or soul under the heading of physics. And, looking further ahead, the eighteenth-century systematist Christian Wolff placed the soul, considered as an inlnlaterial substance, within the natural world, and Kant put the discipline of psychology under the discipline of physics, or, in his terms, under "physiologia" (the logos of physis).38 If naturalism as applied to the mind is the doctrine that we should explain mental activity by appeal only to natural agencies, then by their own lights these Aristotelians and substance dualists both count as naturalists. Yet these same groups also regarded the "natural" mind as an instrument for discerning truth; hence, "naturalistic" description of that mind could at the same time serve as the basis for an analysis of the conditions for knowledge. Kant developed a sharp distinction between empirical psychology (part of physiologia) and the transcendental philosophical investigation of the knowing faculties. By the middle decades of our own century, it was usual to relegate psychology to the "logical space of causes," by contrast with that of "reasons." Scientific psychology, insofar as it concerned itself with the mental at all, came to be viewed as descriptive of the causal mechanisms of cognition, not of its norms. Yet the "common wisdom" that septic boundaries must be observed between epistemology and psychology on pain of psychologistic fallacy is now being challenged by some attempts to "naturalize" epistemology. Is naturalized epistemology a return to the early modern project of charting the cognitive faculties? The answer must be "yes and no." Both base the investigation of the faculties on experience, though the early moderns gave greater weight to ordinary first-person reports of cognitive experience than do today's experimentalists. Both consider the actual operating characteristics of the nlind to be relevant to determining the limits of human knowledge, as in a recent philosophical attempt to argue that with our cognitive resources it may be impossible for us to solve the mind-body problem. 39 But there is divergence

34 GARY HATFIELD over the central question of defining epistemic norms. As we have seen, early modern theorists held that well-functioning natural mental faculties exhibit norms for good thinking. Recent naturalists are split on this question. Some see our natural faculties as shaped by natural selection to track the truth, much as, in the earlier theories, God forged a harmony between the faculties and their objects. 40 The operation of our faculties can thus be expected to exhibit epistemic norms (though these are, of course, open to refinement). But others see a different, and more limited role for naturalistic explanation in epistemology. They take epistemic norms or standards as given by acknowledged cognitive achievements-say, those of the sciences-and endeavor to under41 stand naturalistically the processes by which such achievements occur. There is, then, an analogy between recent investigations of the role of cognitive faculties in human knowledge and the early modern investigations. Both look to the natural capacities of the mind for insight into human knowledge, which seenlS a reasonable strategy if it is not pursued with a predetermined conclusion (e. g., one of the reductionisms). But the commonalities between now and then turn out to be quite linlited, and these limitations can help us to see the need to consider again the framework within which we now discuss mind, cognition, and psychology. Our seventeenth-century authors placed great weight on the investigation of the cognitive faculties because they believed that the hunlan mind has a fixed cognitive structure, and that study of the noetic powers manifested within this structure reveals, in the case of metaphysical optimists such as the Aristotelians and Descartes, the possibility of the cognition of natural essences, or, in the case of pessimists such as Locke, the limits to our cognitive domain. In either event, the early moderns held that the very mechanisms of belief fixation are given with the architecture of the mind. The plausible boundaries of a "fixed cognitive architecture" are not as extensive today. Some cognitive capacities, especially sensory capacities, are relatively fixed: visual acuity, stereoscopic depth perception, perhaps even color similarity metrics. But this is not so for belief fixation. Even those who give great weight to evolution in shaping the mind must admit that a principal biological fact about human beings is that they possess general learning mechanisms capable of acquiring markedly distinct theoretical concepts and general conceptual schemes. The range of this diversity must be at least as broad as the historically actual diversity of human thought. Thus, whereas Descartes could hope to discover the fundamental concepts of physics through proper reflection on innate ideas, scientists today have no such hope. Belief fixation is highly sensitive to conceptual structure and background beliefs. Conceptual structure and background beliefs depend on culturally transmitted learning.. A physicist today who is seeking to determine the basic categories of physics brings to bear his or her understanding of post-Newtonian physics. Many of these concepts had not been envisioned during_the tiIl1~_~fJ?_~~~artes_.- __ -- -

The Workings ofthe Intellect: Mind and Psychology

35

But if belief formation is deeply culturally conditioned, then basic cognition is deeply culturally conditioned. As post-Kantian developments in geometry reveal, what can at one time seem so patently manifest that one is tempted to say that it is constitutive of our cognitive faculties and hence must permanently limit the range of scientific theories, can later be recognized as a contingent and falsifiable hypothesis that has become deeply entrenched in a cultural tradition. If belief fixation is a central feature of human mentality, and if it is, to a significant extent, culturally constituted, then the human mind is a culturally constituted thing. Should it therefore be seen as at least partly standing outside nature? That depends on whether one posits a nature/culture demarcation. If culture is held to be naturally conditioned but itself not part of "human nature" (except for the necessity of having a culture!), then the culturally constituted part of mind stands outside nature. ("Natural" as applied to human beings is here narrowly construed to extend no further than to what is "biologically fixed.") By contrast, if "the natural" is given broad boundaries so as to include all that might be contrasted with "the supernatural," then nature includes human culture, and the mind is wholly part of nature. But if the mind as culturally constituted is part of nature, and if cognitive frameworks vary significantly across cultures, then naturalism cannot promise to achieve the same kind of generality that the seventeenth century wanted from its own "naturalism": insight into the permanent structure of cognition. Thus, under either the broad or narrow conception of nature, naturalism ultimately undermines any hope for the kind of finality with respect to human cognitive structure that had been the goal for Descartes and Locke. Historical reflection might then suggest that we rethink the rhetoric of epistemology and cognitive theory, and move beyond the early modern project of seeking to dissect the faculties of higher cognition once and for all. Reflection on the differences between our conception of psychology and Descartes's understanding of his project reveals that our major differences with him do not pertain to the relevance of psychology to epistemology and metaphysics. Rather, we disagree with his nletaphysics of intellect: we reject his attribution to the mind of "noetic powers" for grasping essences by pure intellect. The Aristotelian and Cartesian conceptions of intellect were laid to rest through the work of Locke, Hume, Kant, and others. As the a priori powers granted to the intellect were proscribed, the sharp distinction between empirical psychology as a descriptive discipline and epistemology as a normative discipline came into being, and with it first arose the framework for leveling the charge of "psychologisnl." As a consequence of these developments, it can now seenl that talk of cognitive faculties could not be anything but a misapplication of our kind of empirical psychology; by contrast, in Descartes's time "psychology" or the study of nlind might well have included investigation of the noetic powers. Philosophical progress is often reflected in changes in the

36 GARY HATFIELD problem space, and those very same changes may in fact serve to mask the developments that brought them about. 3. Historiography, Philosophy, and Interpretation The investigation of the cognitive faculties, their powers and limits, was a central focus of early modern theoretical philosophy. Not only Descartes and Locke, who are discussed here, but Hobbes, Berkeley, and Kant made the faculties central to their discussions of the possibility for and limits to human knowledge. In all of these discussions, the fortunes of metaphysics are directly linked to an investigation of the mind's powers. Descartes sought to open up a new metaphysics, whereas Locke and Kant were coming to grips with the failure to know the real essences of mind-independent substances. In either event, discussions of the mind's real capacities contributed to metaphysical work. In highlighting the theme of the cognitive faculties I have sought to draw attention to an important but relatively neglected factor in the history early modern philosophy. This theme is intended to complement, not to replace, other themes. Indeed, with respect to the two themes mentioned at the beginning of this essay, attending to the role of the cognitive faculties can deepen our understanding of the ways in which early modern philosophy was part of an "Age of Reason," or rose to meet a skeptical challenge: reason was conceived as a faculty of mind (or as an activity of the faculty of intellect), and skeptical writings typically were organized as challenges to the faculties of sense and intellect. Reflection on the latter fact may help interpreters to see more clearly the uses to which skeptical arguments were put by Descartes and others. More broadly, attention to controversies about the cognitive faculties can sharpen our understanding of a core substantive disagreement between "rationalist" and "empiricist": a disagreement about the power of the intellect to know the essences of things. If the cognitive faculties were so important, why have they been neglected in recent discussions? Curiously, much of what early modern writers took to be central to their work has been excised from it out of a "principle of charity." In the middle decades of this century, philosophical interpreters of past texts adopted the strategy of looking for what was "still of philosophical interest" in them, which meant what might still stand as a candidate solution to a philosophical problem of current interest. These same interpreters were wellsteeped in the notion of the "psychologistic fallacy." Further, they were far removed from the notion that the mind might possess special powers or capacities for perceiving essences. Hence, when they read the work of a Descartes or Locke or Kant, the immediate response was either to ignore talk of faculties and cognitive powers, or to translate it into something that seemed more respectable. A striking instance of this may be found in Strawson's

The Workings ofthe Intellect: Mind and Psychology

37

Bounds of Sense, in which he sought to untangle "what remains fruitful and interesting" from "what no longer appears acceptable, or even promising," in Kant's work. He thus replaced the "imaginary subject of transcendental psychology"-including its reference to a "nlanifol~ of intuition'~ and its app~al to an activity of "synthesis" to explain the unIty of consciousness-wrth philosophical analysis of "ordinary reports of what we see, feel, hear, etc." and of the "rules etnbodied in concepts of objects" as exemplified in "the general coherence and consistency of our ordinary descriptions of what we see, hear, feel, etc.,,42 Here, talk in the "psychological idiom" is translated into the mid. twentieth century idiom of "philosophical analysis" in order to preserve what is "fruitful and interesting" in Kant. The spirit of this interpretive tack received an extreme expression in Donald Davidson's work on radical interpretation (though he cannot be held responsible for excesses in practice), in which Davidson concluded that our most effective strategy for making sense of others' utterances is that of interpreting them so that agreement is optimized. 43 In any event, under the principle of "charity," when Descartes and the others were talking about cognitive faculties, they either were talking isolated nonsense or were engaged in (allegedly bad) empirical psychology~ either way, those parts of the text can be treated as philosophically irrelevant in themselves. In contrast with the method of sifting through the detritus of past philosophy for salvage, in this paper I have adopted the strategy of starting from and working with the categories used by past authors, in order to achieve an understanding of their philosophical projects and of the (alleged) force of their philosophical argunlents and conclusions as they saw it. This means taking their claims at face value and seeking to understand whence they expected the force of the claims to come. In pursuing this strategy, one is of course "charitable" in that one avoids easy attributions of silly mistakes or blunders to past authors-though such attributions are not ruled out. In Descartes's case, beginning with his claims and conceptions nleans taking seriously his injunctions to "meditate with" him, and his assertion that he could intellectually perceive the extension of the geometers, or the idea of an infinitely perfect being. In tenth-grade geometry I was told to imagine planes without thickness and lines without width. I thought I was doing it, but now I believe not. I certainly am unable to find in myself the pure intellectual cognition of a triangle of determinate shape, untinged by sensory qualities. Further, though I know what it is to be intuitively certain of something, I don't believe that such certainties can of themselves reveal the contours of mind-independent reality, the essences of substances. I thus reject both Descartes's conception of the intellect's power and some of his assertions about its consciously accessible deliverances. The fact that I think Descartes was wrong does not seem a good reason to allege that he was really saying something else. In order to learn from-or even to learn about-the history of philosophy, we nlust come to understand the historical development of philosophy in its

38 GARY HATFIELD own terms. For the early modern period, this means acknowledging the centrality of the theory of the cognitive faculties in the philosophical work of the time. Rather than ignoring talk of the faculties in classical texts, we should come to understand the role the faculties played. If the role is one that we now reject, then we should seek the philosophical reasons that led us to reject it. In the course of doing this, we may learn the answers to (or at least learn to ask) questions such as the following: How have the relations among logic, mind, and psychology changed in the past three hundred years? How did philosophers come to adopt the notion of a "psychologistic fallacy"? What is the origin of our current notions of the relations between the natural and the mental? What can we say now about our ability to discern the truth? Is it a simple biological capacity, or, at least for truths as complex as those of the natural sciences, does this ability depend on cultural processes that are underdetermined biologically? If we interpret past authors so as to have them (as much as possible) say only things that we might consider saying now, we shall surely do little more than find our own reflection in their texts. We certainly won't gain the sort of understanding that comes from uncovering the formation of our current problem space and seeing its contingencies. Questions about the deep conceptual changes will go unasked, because the changes will be masked against the foreground of "charitable" renderings. But contextually guided study of early modern philosophy can help bring such questions to light. I am therefore suggesting that the philosophical works of the early modern period are of interest in their own right (sans a strong principle of charity) for what they can reveal to us about the structure of philosophy itself. The "principle of charity" turns out to be a stultifying principle of interpretation for the history of philosophy. I propose that we reject it, or at the least supplement it with the practice of reading texts in the intellectual context of their time, using that context to make interpretive sense of conceptions that are prima facie foreign to us now. In this way, we may truly come to learn about other philosophies, which is a necessary condition for learning from them. At the same time, we will come to see that there is much to be learned about the implicit and explicit conceptions of mind, cognition, and logic in the philosophical texts of the early modern period, and about the heritage of those conceptions in the philosophical common sense of today. NOTES *Thanks to Lanier Anderson, Alan Kors, Holly Pittman, and Alison Simmons for their comments on earlier versions of this paper. 1. The emphasis on the free use of reason arose early: Johann Jakob Brucker, Historia critica philosophiae, 4 vols. (Leipzig: Breitkopf, 1742-44) vol. IV. It structured Friederich Ueberweg, History of Philosophy, from Thales to the Present Time, George S. Morris, trans., 2 vols. (London: Hodder and Stoughton, 1880) vol. IT, though he

The Workings ofthe Intellect: Mind and Psychology

39

incorporated the skeptical theme, as well, dividing early nlodern philosophy into three periods described as (1) "transition to independent investigation," (2) "empiricism, dogmatism, and skepticism," and (3) "criticism and speculation." It is reflected in Ernst Cassirer, Philosophy of the Enlightenment, Fritz C. A. Koeln and James P. Pettegrove, trans. (Princeton: Princeton University Press, 1951) ch. 1, sec. 1, and Peter A. Schouls, Descartes and the Enlightenment (Kingston and Montreal: McGill-Queens University Press, 1989). Alfred North Wllitehead, Science and the Modern World (New York: Macmillan, 1925) does not follow the theme of throwing off authority, but he characterizes the seventeenth century as a "Century of Genius" that yields eighteenth-century Enlightenment (chs. 3,4). 2. Richard H. Popkin, History of Scepticism from Erasmus to Spinoza (Berkeley, Los Angeles, London: University of California Press, 1979) has made the skeptical theme prominent in recent years; Ueberweg's second period of modem philosophy ranked skepticism together with empiricism and dogmatism as "rival systems" to which "criticism" was a response (History, vol. II); Immanuel Kant, Critique ofPure Reason, Nonnan Kemp Smith, trans. (New York: St. Martin's, 1965) proposed a similar tripartite division (A761/B789) among other analyses of philosophy's history (A85256/B880-84); "A" and "B" refer to the pagination of the first and second editions, respectively, of Kant's Kritik der reinen Vernunft (Riga: Hartnoch, 1781, 1787) hereafter cited as "CPR" plus page numbers. E. M. Curley, Descartes against the Skeptics (Cambridge, Mass.: Harvard University Press, 1978) endorses a portion of this picture by maintaining that Descartes's mature philosophy was directly motivated by the threat of pyrrhonian skepticism (p. 38). 3. Thomas Reid, Inquiry into the Human Mind, ch. 1, secs. 3- 7, in his Works, William Hamilton, ed., 2 vols. (Edinburgh: James Thin, 1895) vol. I, pp. 99-103 (Reid of course did not use the term "epistemology," and his remarks on the theory of ideas were part of an analysis of knowledge of the human mind itself, and its cognitive capacities); Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University Press, 1978) who cites Reid, among others. The historiography of an epistemological turn, with central emphasis on the theory of ideas, is found in recent general histories of philosophy, e. g., Roger Scruton, From Descartes to Wittgenstein: A Short History of Modern Philosophy (New York: Harper Colophon, 1982); John Cottingham, The Rationalists (Oxford: Oxford University Press, 1988) pp. 4-11; R. S. Woolhouse, The Empiricists (Oxford: Oxford University Press, 1988) ch. 1. 4. Louis E. Loeb, From Descartes to Hume: Continental Metaphysics and the Development ofModern Philosophy (Ithaca: Cornell University Press, 1981). 5. I have sketched this story-line for the history of modem philosophy in my The Natural and the Normative: Theories of Spatial Perception from Kant to Helmholtz (Cambridge, Mass.: The MIT Press, 1990) chs. 2, 6. John W. Yolton, Perceptual Acquaintance from Descartes to Reid (Minneapolis: University of Minnesota Press) appreciates the significance of the faculties in early modem philosophy, but assimilates concern with the faculties directly to a present-day conception of "psychology" (pp. 16, 39,105). 6. Support for various descriptive claims made here may be found in my Natural and Normative, chs. 2-3. 7. Aristotle, De anima, in his Complete Works, Jonathan Barnes, ed., 2 vols. (Princeton: Princeton University Press, 1984) vol. I, 430a14-15.

40

GARY HATFIELD

8. Giacomo Zabarella, Commentarii in III. Aristotelis libros De anima (Frankfurt am Main: Zetner, 1606) "Liber de mente agente," ch. 13 (cols. 935-7) held that God performs the function of the active intellect for all humans. The orthodox view was held by Thomas Aquinas, Summa theologiae (Cambridge: Blackfriars, 1964-81) 1.76.2; 79.45, hereafter "ST"; Aquinas, Questions on the Soul, J. H. Robb, trans. (Milwaukee: Marquette University Press, 1984) qus. 3-5; John Duns Scotus, De anima, quo 13, in his Opera omnia, L. Wadding, ed., 26 vols. (Paris: Vives, 1891-95) vol. ill, p. 546; Francisco Toledo, Commentaria una cum quaestionibus in tres libros Aristotelis De anima (Kaln: Birckmann, 1594) n.l, quo 2 (fol. 40vb-48vb); Francisco Suarez, De anima (hereafter, "DA") IV.8.4-8, in his Opera omnia, M. Andre, ed., 26 vols. (Paris: Vives, 1856-78) vol. ill, pp. 741a-43b; Coimbra College, Commentarii in tres libros De anima (Kaln: Zetner, n.d., ca. 1600) ill.5, quo 1, art. 1-2 (pp. 369-374); Antonio Rubio, Commentarii in libros Aristotelis Stagyritae philosophorum principis, De anima (Lyon: Joannes Pillehotte, 1620) "Tractatus de natura, et ratione atque officio intellectus agens," quo 4 (pp. 652-53); works entitled "commentaries" on De anima will subsequently be referred to as "CDA." On Avicenna, Averroes, and the late Greek and Arabic background to the view that human intellection depends upon a single active intellect, see Herbert A. Davidson, Alfarabi, Avicenna, and Averroes, on Intellect (New York, Oxford: Oxford University Press, 1992). 9. Aquinas, ST 1.84.6-7,1.87.1; Suarez, DA IV.7.3 (p. 739); Coimbra College, CDA ill.5, quo 3, art. 2, (pp. 383-4) ill.8, quo 8, art. 2 (pp. 453-5); Rubio, CDA, "Tractatus de intellectu agente," quo 2-3 (pp. 637-46) "Tractatus de natura, actu et obiecto intellectus possibilis," quo 7 (pp. 692-3); and Eustace of St. Paul, Summa philosophiae quadripartita, 4 parts (Kaln: Philip Albert, 1638) pt. ill, "Physica" (hereafter, "SP-P") m.4, disp. 2, qus. 4-5, 7, 10 (pp. 287-9,290-3, 298). 10. For a survey of late Aristotelian theories of sensory and intellectual cognition, see my "Cognitive Faculties," in the Cambridge History of Seventeenth Century Philosophy, Michael Ayers and Daniel Garber, eds. (Cambridge: Cambridge University Press, in press). Leen Spruit, Species Intelligibilis: From Perception to Knowledge, 2 vols. (Leiden, New York: E. J. Brill, 1994-95) vol. II, has just published a detailed study of intellectual cognition in later scholasticism. 11. Yolton, Perceptual Acquaintance, pp. 6-10, where "absorption" is used to characterize some scholastic accounts of sensory perception, but also fits his account of intellectual abstraction and the production of intelligible species. Also D. W. Hamlyn, Sensation and Perception: A History of the Philosophy of Perception (London: Routledge & Kegan Paul, 1961) p. 48; Brian E. O'Neil, Epistemological Direct Realism in Descartes's Philosophy (Albuquerque: University of New Mexico Press, 1974) pp. 48-49. But others have avoided this reading, at least of Aquinas's position: Sheldon M. Cohen, "St. Thomas Aquinas on the Immaterial Reception of Sensible Forms," Philosophical Review 91 (1982) pp. 193-209, on p. 199; Paul Hoffman, "St. Thomas Aquinas on the Halfway State of Sensible Being," Philosophical Review 99 (1990) pp. 73-82, on p. 75 n8. 12. Aquinas cites this principle, attributing it to Aristotle himself: "Aristotle held that the intellect does have an operation in which the body does not communicate. Now, nothing corporeal can make an impression on an incorporeal thing. And therefore in I\order to cause an intellectual operation, according to Aristotle, the mere impression raused by sensible bodies does not suffice, but something more noble is required, for

The Workings ofthe Intellect: Mind and Psychology

41

the active is superior to the passive, as he says himself." (ST 1.84.6; my revisions to the translation) The other authors cited in n. 9 also held this principle. Davidson, Alfarabi, Avicenna, and Averroes, on Intellect, discusses this and related principles in late Greek and Arabic commentators. 13. Aquinas, ST 1.85.1, ad 3, in which the final quoted sentence reads: "Et per hunc modum dicitur abstrahi species intelligibilis a phantasmatibus; non quod aliqua eadem numero forma quae prius fuit in phantasmatibus, postmodum fiat in intellectu possibili, ad modum quo corpus accipitur ab uno loco, et transfertur ad alterum." (Translation altered from Blackfriars; see also Aquinas, Summa Theologica, English Dominicans, trans., 19 vols., London: Thomas Baker/Burns, Oates & Washbourne, 1911-22.) 14. While agreeing that a species conj oined with matter-even if conj oined nonstandardly, being a "form without matter"-cannot by itself affect the immaterial intellect, these authors characterized the causal role of corporeal phantasms in the production of intelligible species differently: Suarez maintained that the phantasm does not affect the possible intellect by "influx," but "materially" or by "exemplar," mediated by the fact that imagination and intellect are powers of the same soul (De anima, IV.2.10-12, vol. III, p. 719a-b); the Coimbran text discussed ways in which the active intellect might be taken as both a "partial" and an "instrumental" cause, and said that the phantasm "cooperates" to "excite" the active intellect to produce the species (CDA III. 5, quo 6, pp. 407-9); Rubio designated the phantasm an "instrumental" cause ("elevated" by another power) and the active intellect the "principal" or "primary" cause of the production of an immaterial intellectual species in the patient intellect (CDA III.4-5, "Tractatus de intellectu agente," quo 3, pp. 646-52); Eustace described the phantasm as a "material" or "dispositive" as opposed to "efficient" cause (SP-P m.4, disp. 2, quo 7, pp. 292-3). 15. Thomas Aquinas, ST 1.79.2; 84.3-5; Suarez, DA IV.2.7-18; 7.3; 8.7-8; Coimbra College, CDA m.4-5, quo 1, art. 2, "nuda tabula" (pp. 372, 374); Rubio, CDA m.4-5, "Tractatus de intellectu agente," qus. 1-3; Eustace of St. Paul, SP-P ill.4, disp. 2, quo 7, "tabula rasa" (p. 291); the active intellect "makes" (jabricare) intelligible species (pp. 291-2). 16. Thomas Aquinas, ST 1. 84.5: "Et sic necesse est dicere quod anima humana omnia cognoscat in rationibus aeternis, per quarum participationem omnia cognoscimus. Ipsum enim lumen intellectuale, quod est in nobis, nihil est aliud quam quaedam participata similitudo luminis increati, in quo continentur rationes aeternae"; he explicitly distinguishes this position from Platonism and other positions in which the eternal types are beheld by the human intellect independently of the senses, or are known innately. See also ST 1.79.3-4; Aquinas, Questions on the Soul, quo 5, resp. and ad 6; Aquinas, Truth, R. W. Mulligan, trans., 3 vols. (Chicago: Regnery, 1952-54) quo 10, art. 6. 17. Toledo, CDA III.7, quo 23, concl. 3: "Intellectus in corpore non potest habere naturaliter claram & distinctam cognitionem substantiae immaterialis" (fo1. 168ra); also, concl. 4: "Substantiae immateriales a nobis confusem in hoc statu cognoscuntur" (fol. 168rb). Aquinas, ST 1.87.3; 1.88; Coimbra College, CDA ill.5, quo 5, art. 2 (pp. 402-3); III. 8, quo 7, art. 2 (p. 449); quo 8, art. 2 (pp. 453-55); Rubio, CDA ill.4-5, "Tractatus de intellectu possibili," qus. 5-6 (pp. 680-89); and Eustace of St. Paul, SP-P m.4, disp. 2, qus. 4-5, 7 (pp. 287- 89,290-93).

42

GARY HATFIELD

18. Descartes, Meditations on First Philosophy, in The Philosophical Writings of Descartes, John Cottingham, Robert Stoothoft~ and Dugald Murdoch, trans., 2 vols. (Cambridge: Calnbridge University Press, 1984-85) II:51, 54; Principles of First Philosophy, pt. 1, art. 32 (1:204) where Descartes lists pure understanding (or pure intellection) imagination, and sense percepton as modes of "perception" or of the "operation of the intellect"; hereafter, vols. I and II of the Cottingham et al. translation is abbreviated "CSM" (plus volume and page number). 19. In 1630 Descartes wrote to Mersenne that he had worked on metaphysics intensely during his first nine months in the Netherlands (a period ending in 1629): to Mersenne, 16 April 1630, in his Oeuvres, Charles Adam and Paul Tannery, eds., rev. ed., 11 vols. (Paris: VrinlCNRS, 1964-1976) vol. I, p. 144 (hereafter, the Oeuvres are referred to as "AT," followed by volume and page numbers); translation in Philosophical Writings of Descartes, vol. ill, John Cottingham, Robert Stoothoff, Dugald Murdoch, and Anthony Kenny, trans. (Cambridge: Cambridge University Press, 1991) p. 22; hereafter, vol. ill is abbreviated "CSMK." In 1637 he reported that "eight years ago" he had written "in Latin the beginnings of a treatise of metaphysics," in which, among other things, he argued for a soul-body distinction (to Mersenne, 27 February 1637, AT 1: 350; CSMK, p. 53). On Descartes's "metaphysical turn," see my "Reason, Nature, and God in Descartes," Essays on the Philosophy and Science o/Rene Descartes, Stephen Voss, ed. (New York: Oxford University Press, 1993) pp. 259-287, and the literature cited therein. 20. Descartes asserts that matter's essence is extension in the opening paragraphs of the Fifth Meditation; he draws a clear distinction between intellectual and imaginal cognition of extension at the start of the Sixth Meditation. (CSM II: pp. 44, 50- 51, AT VII: pp. 63, 72-3) 21. Ignatius of Loyola, The Spiritual Exercises, with the Directory to the Spiritual Exercises of his followers, W. H. Longridge, trans., 4th ed. (London: Mowbray, 1950) First Week, First Exercise, pp. 52-57, and Directory, ch. 14, sees. 2-3; Francis de Sales, An Introduction to a Devoute Life, 1. Yakesley, trans. (Douai: Heighman, 1613) pt. 2, pp. 138-143. 22. On Descartes's use of the meditative genre, see the first three essays in Essays on Descartes' Meditations, Amelie O. Rorty, ed. (Berkeley and Los Angeles: University of California Press, 1986) and Berel Lang, Anatomy of Philosophical Style (Oxford: Basil Blackwell, 1990) ch. 3. 23. In the Sixth Meditation, Descartes explains that the senses should not be used for making judgments "about the essential nature of the bodies located outside us"; rather, such judgments should be left to the "intellect," or the "mind alone," operating independently of the body (CSM II: pp. 57-58, AT VII: pp. 82-83). In Meditations Three, Five, and Six he uses the intellect (ostensibly) to know God and the essences of matter and mind. 24. Kant, CPR B167 (Kant here makes no mention of Platonism, but see also A31314fB370). On Platonist theories of cognition in the early modem period, see my "Cognitive FacuIties." The harmony is not "preformed" if it is established via the causal agency of the Forms themselves, being "seen" by the human intellect; it is preformed on a "reminiscence" reading of Plato. 25. Descartes, letters to Mersenne in the 1630s (CS:MK, pp. 23-26, AT I: pp. 145, 149-53); Fifth and Sixth Sets of Replies (CSM II: pp. 261, 291, 293-4; AT VII: pp. 380, 432, 435-6).

The Workings ofthe Intellect: Mind and Psychology

43

26. For further discussion, see Emile Brehier, "The Creation of the Eternal Truths in Descartes's System," Descartes: A Collection of Critical Essays, Willis Doney, ed. (Notre Dame: University of Notre Dame Press, 1968) pp. 192-208; and my "Reason, Nature, and God." 27. John Locke, An Essay Concerning Human Understanding, Peter H. Nidditch, ed. (Oxford: Oxford University Press, 1975) epistle, p. 6. Hereafter, the Essay will be cited as "E," followed by book, chapter, and section numbers. 28. On Locke as an "imagist," that is, as someone who took the content of thought to limited to sensory images and their combination (together with refections on the mind's own operations) see Michael Ayers, Locke, vol. 1: Epistemology (London and New York: Routledge, 1991) pt. I, ch. 5. 29. For a comparison of Locke's position to Aristotelian and Cartesian (among other) conceptions of substance and our cognitive grasp of it, see Ayers, Locke, vol. II: Ontology, p1. I, chs. 2, 6. 30. The relations among the concept of substance, that of real essence, and the corpuscular theory of matter in Locke's writing is a matter of some interpretive delicacy; for an overview, see Edwin McCann, "Locke's Philosophy of Body," Cambridge Companion to Locke, Vere Chappell, ed. (Cambridge: Cambridge University Press, 1994) pp. 56-88, especially sec. 4. 31. CPR, A256-57fB312-313; Kant, Prolegomena to Any Future Metaphysics That Will Be Able to Come Forward as Science, Gary Hatfield, ed. and trans. (Cambridge: Cambridge University Press, forthcoming) sec. 34. 32. Richard Rorty, in his Philosophy and the Mirror, reconstructs the history of modem philosophy as part of a narrative within which, from the time of Locke through Kant to the present day, philosophy's (alleged) claim to intellectual authority has rested on a confusion between epistemology and psychology, which he compares to the "naturalistic fallacy" in ethics (p. 141); hence, though he did not use the term "psychologism," his charge fits the classical meaning .of that term, according to which psychologism is the attempt to base epistemology on psychology. 1. E. Erdmann gave this meaning to the term in introducing it, Grundriss der Geschichte der Philosophie, 2d ed., 2 vols. (Berlin: Hertz, 1870) vol. II, p. 636; see also John Dewey, "Psychologism," Dictionary of Philosophy and Psychology, James Mark Baldwin, ed., 3 vols. (New York: Macmillan, 1901-05) vol. II, p. 382. Rorty reviews earlier instances of this charge against early modem philosophy by T. H. Green and Wilfrid Sellars, Philosophy and the Mirror, pp. 140-43. 33. Francisco Toledo, Commentaria, una cum quaestionibus, in universam Aristotelis Logicam (Kaln: Birckmann, 1596) pref., quo 1 (pp. 3-7); Coimbra College, Commentarii collegii conimbricensis e soecietate iesu, in universam Dialecticam Aristotelis (Lyon: Horation Cardon, 1607) proem, quo 4, art. 2 (pp. 57-61); Antoniao Rubio, Logica mexicana, sive comentarii in universam Aristotelis Logicam, 2 parts (Kaln: Birckmann, 1605) proem, quo 1, p1. I (cols. 1-11); Eustace of S1. Paul, Summa philosophiae, p1. I, "Dialecticae sive logicae," proem, quo 4 (pp. 10-11). It was common to describe the operations of the "natural light" of the human intellect as instantiating "natural logic," by contrast with the "artificial logic" developed by Aristotle and others; Toledo declines to adopt this terminology, refusing to call these natural operations in themselves a "logic." (p. 5) 34. Descartes, Meditations, IV: "since my understanding comes from God, everything that I understand I understand correctly, and any error here is impossible"

44

GARY HATFIELD

(CSM II: p. 40, AT VTI: p. 58); clear and distinct perceptions of the intellect produce a "great inclination in the will," and as long as one assents only to such clear and distinct intellectual perceptions, one will not fall into error. (CSM II: p. 41, AT VTI: p. 59) See also Principles, 1. pp. 30-42. 35. The earliest free-standing work entitled "psychology" was by RUdolph Goclenius, Psychologia: hoc est, de hominis perfectione, animo (Marburg: Paul Egenolph, 1594) which focused more on problems concerning the infusion of the soul into the embryo at conception than on the discussions of the cognitive faculties that characterized the De anima literature; the latter sort of discussion occurred in Johann Conrad Dannhauer, Collegium psychologicum, in quo maxime controversae quaest.iones, circa libros tres Aristotelis De anima, proponuntur, ventilantur, explicantur. (Argentoranti: Josias Staedel, 1630) On the origin of the tenns "psychologia" and "psychology," Francois H. Lapointe, "Who Originated the Term 'Psychology'?," Journal of the History of the Behavioral Sciences 8 (1972) pp. 328-35; on early psychology, Paul Mengal, "Naissances de la psychologie: la Nature et l'Esprit," Revue de Synthese, 115 (1994) pp. 355-373, and my "Psychology as a Natural Science in the Eighteenth Century," Revue de Synthese, 115 (1994) pp. 375-391. 36. Toledo, CDA, devoted fo1. 65rb-73vb to the vegetative soul, 73vb-129ra to the sensitive, 129ra-169ra to the intellect, and 169rb-179rb to appetite, will, and motion; Coimbra College, CDA, devoted pp. 148-61 to the vegetative soul, 160-361 to the sensitive, 360-469 to the intellect, 460-98 to appetite, will, and motion, with separate treatises on the separated soul (pp. 499-596) and on additional problems pertaining to the five senses (pp. 597-619); Rubio, CDA, devoted pp. 278-305 to the vegetative soul, 305-632 to the sensitive, 633-735 to the rational, and 735-57 to appetite, will, and motion, adding a treatise on the separated soul (758-94). The coverage was slightly more balanced in the textbooks: e. g., Eustace of St. Paul, SP-P ("Physica") devoted 197-228 to the vegetative soul, 228-77 to the sensitive, including motion, and 278-308 to the rational soul, including will. 37. Toledo, CDA, proem, quo 2 (fo1. 4) subsumed the soul in all of its operations under physics; Coimbra College, CDA, proem, quo 1, art. 2 (pp. 7-8) and Rubio, CDA, proem, quo 1 (pp. 10-11) subsumed the study of embodied souls under physics, and separated souls under metaphysics. Eustace of St. Paul, SP-P, treated "de anima" topics in the part entitled "Physica," per the norm. 38. Antoine Le Grand, Institutio philosophiae secundum principia de Renati Descartes (London: J. Martyn, 1678) praecognoscenda, art. 7, 15, 16. Christian Wolff, Psychologia rationalis (Frankfurt am Main and Leipzig: Libraria Rengeriana, 1640) sec. 69; Cosmologia generalis (Frankfurt am Main and Leipzig: Libraria Rengeriana, 1637) sec. 509; Wolffs follower Alexander Baumgarten, Metaphysica, 7th ed. (Halle: Hemmerde, 1779) sec. 351, 402, explicitly placed monads or simple substances, including spirits, within cosmology. Kant, CPR A846-47/B874-75; in the Prolegomena, sec. 15, Kant places psychology under "universal natural science." 39. Colin McGinn, "Can We Solve the Mind-Body Problem?" Mind 98 (1989) pp. 349-366. 40. W. V. O. Quine, "Natural Kinds," in his Ontological Relativity and Other Essays (New York: Columbia University Press, 1969) pp. 114-138, on pp. 125-28. 41. Miriam Solomon, "Scientific Rationality and Human Reasoning," Philosophy of Science 59 (1992) pp. 439-455, on pp. 442-43.

The Workings ofthe Intellect: Mind and Psychology 45 42. Strawson, The Bounds of Sense: An Essay on Kant's Critique of Pure Reason (London:~ethuen, 1966)pp. 16,32. 43. Donald Davidson, "On the Very Idea of a Conceptual Scheme," Proceedings and Addresses of the Americal Philosophical Association 47 (1974) pp. 5-20, on p. 19.

Petrus Fonseca on Objective Concepts and the Analogy of Being E. JENNIFER ASHWORTH

Petrus Fonseca was a Portuguese Jesuit who lived from 1528 to 1599. He was one of those responsible for drawing up the Jesuit Ratio Studiorum which set the curriculum for Jesuit schools across Europe, and he was also responsible for initiating the production of the Coimbra commentaries on Aristotle, or Conimbricenses, which served as texts for many schools and universities in the seventeenth century.l He was himself the author of two popular texts, an introduction to logic, and a commentary on Aristotle's Metaphysics. 2 His logic text was one of two alternatives prescribed by the Ratio Studiorum of 1599, and may have been used at La Fleche;3 his Aletaphysics commentary was used at many Jesuit schools, and may also have been used at La Fleche. 4 In short, Fonseca was a leading figure in the Scholastic Aristotelian tradition of the late sixteenth century, a tradition which lies behind many of the developments in early modern philosophy, and which in many ways is more important than the humanist tradition represented by Petrus Ramus. 5 I have chosen to discuss Fonseca on objective concepts and the analogy of being both because an exanlination of these issues will help us to understand how logic came to be bound up with the philosophy of mind and because the history of how these issues were treated helps solve a small problem about Descartes's sources. My paper has four parts. I shall begin by giving a historical outline of treatments of analogy and their relevance to Descartes. 6 Secondly, I shall discuss late medieval theories of signification, particularly as they appear in Fonseca, in order to show how logicians turned away from spoken language to inner, nlentallanguage. Thirdly, I shall explain how it was that analogy, as a theory of one kind of language use, was particularly bound up with the discussion of concepts. Finally, I shall look at the distinctions Fonseca made while discussing the concepts associated with analogical terms. 1 Historical Outline: From Scotus to Descartes In Meditation 3, Descartes uses a distinction between formal and objective reality with respect to ideas in order to prove the existence of God. In the secondary literature this distinction is invariably linked with a distinction between fornlal and objective concepts found in Suarez, whose Metaphysical Disputations (published in 1597) was cited by Descartes on one occasion. 7 However, as the literature acknowledges, it is not clear where the distinction originated, or how Descartes came to know of it. The earliest paper I know of, published by Dalbiez in 1929, looked in two directions. 8 Dalbiez quite accurately traced the distinction back to Duns Scotus and his discussion of the kind of being creatures had in God's mind prior to creation,9 but Dalbiez

48 E. JENNIFER ASHWORTH thought it improbable that Descartes would have read Duns Scotus. He then suggested that the notion is more likely to have come from Suarez and another near-contemporary, Vasquez, both of whom used the notion in a theological dispute about the views of the fourteenth-century theologian Durandus of Saint Pour9ain (d. 1334) on the nature of truth. 10 Little new light has been shed since 1929. 11 In recent papers, Norman Wells still privileges both Suarez and the debate about Scotus on divine ideas. I2 In a paper entitled "Meaning and Objective Being: Descartes and His Sources," Calvin Normore first discusses Duns Scotus and William of Alnwick on objective being in the context of God's ideas; and he then shows how the notion was used by Peter Aureol, William Ockham, and Walter Chatton in a variety of contexts. However, Normore acknowledges that there is a gap between about 1340 and the beginning of the seventeenth century. In his conclusion, he writes that his examination "suggests a Descartes firmly rooted in a Scholastic tradition which is deeply in debt to Duns Scotus and closely allied with fourteenth-century developments in epistemology and in the theory of meaning. This makes the problem of Descartes' immediate sources and the question of his originality even more puzzling. ,,13 My own recent work on analogy as a theory of one kind of language use shows that at least one historical path between Scotus and the early seventeenth century can be traced through the Thomistic tradition, though we must renlember that late medieval and Renaissance Thomism enlbraced a variety of different approaches and doctrines. What Thomists had in common was a kind of moderate realism with respect to common natures that differentiated them from the nominalists on the one hand and the Scotists on the other. Nonetheless, Thomists embraced many theses put forward by nominalists, especially Pierre d'Ai1ly (d. 1420/1); and much of their agenda had been set by Duns Scotus rather than by Aquinas himself. I shall begin by mentioning two important early fourteenth-century sources. The distinction between formal and objective concepts is used in the discussion of analogical terms by Hervaeus Natalis (d. 1323) in his Quolibeta, and by Peter Aureol (d. 1322) in his commentary on the Sentences of Peter Lombard. I4 Aureol was widely read, and in the early fifteenth century Johannes Capreolus (d. 1444) in his Defensiones theologiae divi Thomae Aquinatis, made very heavy use of him. Long passages from Aureol, including his criticisms of Hervaeus, appear in the sections on analogy, and in these same sections Capreo1us makes important use of the distinction between formal and objective concepts. I5 Hervaeus Natalis was also well known. His works were widely disseminated in manuscript form, and his Quolibeta had more than one early printed edition. I6 The views of Aureol, Hervaeus, and Capreolus on analogy are reflected in varying degrees in the commentaries on the Metaphysics by Dominic of Flanders (d. 1479) and Paulus Barbo Soncinas (d. 1495), though only the latter used the distinction between formal and objective concepts. I7 In

Petrus Fonseca on Objective Concepts and the Analogy ofBeing 49 turn, material drawn from Capreolus, Soncinas, and possibly Dominic of Flanders, lies behind the short treatise On the Analogy of Names which Thomas de Vio, Cardinal Cajetan published in 1498 as a supplement to his commentary on Aristotle's Categories. I8 In this work, formal and objective concepts are discussed at length. In the sixteenth century, Domingo de Soto's commentary on the Categories cites Capreolus and Cajetan in the discussion of analogy.I9 At the end of the sixteenth century and the beginning of the seventeenth century, all these authors and themes turn up in the treatments of analogy by Suarez in his Disputationes Metaphysicae,2o by Sebastian de Couto, the author of the Coimbra commentary on Aristotle's Organon,21 and by Antonius Rubius in a treatise on analogy he inserted into his commentary on Aristotle's Organon.22 Nor was Fonseca any exception. Scotus, Aureol, Hervaeus, Capreolus, Dominic of Flanders, Soncinas, Cajetan were all cited in his Metaphysics commentary,23 and the discussion of formal and objective concepts is firlnly tied to the discussion of analogy. There are of course other areas in which the distinction between formal and objective concepts was enlployed in logic texts belonging to the Thomistic tradition. One of these areas, which needs to be mentioned here because of its implications for the doctrine of analogy, was the discussion of the problem of universals. Logicians in the Thomistic tradition wanted to avoid nominalism by arguing that universal terms and concepts do have an objective (in the modern sense) basis in the world. At the same time, they also wanted to avoid both Platonisnl, the view that universal ternlS correspond to separate common natures, and Scotism, the view that universal terms correspond to common natures which have their own lesser existence and lesser unity within the real individuals which they in part constitute. Toletus, the other logician recornmended by the Ratio Studiorum of 1599, gave an account which is fairly standard. 24 He argued that a universal taken materially is a nature, while a universal taken formally (in formali) is a being of reason (ens rationis) because community and universality are the work of the intellect. They are not real properties of real individuals. A being of reason is something that exists objective in the intellect, but not everything that has esse obiective in the intellect is a being of reason, for anything, including real individuals, that is an object of cognition has esse obiective in the intellect. There are two groups of objects that have only esse obiective. Pure figments, such as chimeras, have esse obiective directly, while those beings of reason that are founded on real things and can be predicated of real things (as Fonseca added25 ), have esse obiective indirectly and as a consequent. Toletus then went on to introduce the fully fledged distinction between the formal and the objective intentio or conceptus, still in the context of his discussion of universals. Similar accounts can be found in the logic texts of Domingo de Soto and Sebastian de CoutO. 26 It is interesting to note that the scholastic authors that Descartes told Mersenne he remembered were the Coimbrans, Toletus, and Rubius. 27 Two of

50 E. JENNIFER ASHWORTH these discussed formal and objective concepts in the context of analogy, and the one exception, Toletus, made the distinction in his discussion of universals, as did the Coimbrans. Moreover, they are all authors more likely to be found in the classroom than Suarez. Fonseca is not in the list, but as I remarked at the beginning, he was closely associated with the Conimbricenses, and his views are explicitly discussed by the Coimbra Categories commentary and by Rubius. Thus there is a clear relationship between Descartes and the Thomistic textbook tradition. 2. Signification and Mental Language

So far as individual words were concerned, the central claim was that the spoken word is a sign, and the central semantic notion was that of signification. This notion of the spoken word as sign is linked with two traditions, the Augustinian tradition which was appealed to particularly by theologians, and the Aristotelian tradition found especially in logic and metaphysics. The two traditions meet in Fonseca. In his logic text he quotes Augustine's definition, "A sign is something which is itself sensed and which indicates to the mind something beyond the sign itself,"28 and he also refers to On interpretation 16a3-4, where Aristotle said that spoken words were signs of concepts?9 In his own main definition, "to signify is nothing other than to represent something to a cognitive power" ("significare nihil aliud est, quam potentiae cognoscenti, aliquid repraesentare"), he paraphrases the words of the popular fourteenthcentury nominalist, Pierre d' Ailly. 30 While these definitions are conlpatible with paying attention to the hearer's understanding in a particular context, the role of the speaker was nornlally privileged, and the significative power of utterances was seen as a function of the speaker's cognition without reference to the context. Thus in his Metaphysics commentary Fonseca emphasizes that the purpose of spoken language is to make known the concepts of the speaker. 31 One can relate this emphasis on the speaker's concepts to a big change which took place in the early part of the fourteenth century. Ockham insisted that the concept itself must be regarded as a sign. 32 This notion (while foreign to Augustine) was not new, but Ockham made it central, and in so doing, made mental language rather than spoken language the paradigm of signification. The doctrine that there is a language of thought that is naturally significant and common to all human beings is present at least from Augustine on, but it was only fully developed in the fourteenth century, first by Ockham and then by Pierre d' Ailly. Mental propositions were thought of as having syntactic stmcture33 and mental terms were thought of as having supposition, so that the notion of a language system was internalized, and the very way that thought was conceived of changed. As a corollary, the place of grammar in the study of logic and philosophy of language was devalued. By the later sixteenth century,

Petrus Fonseca on Objective Concepts and the Analogy ofBeing 51 interest in the structure of an inner language had disappeared, but the doctrine remained. Fonseca is no exception. In his discussion of whether dialectic is a linguistic science (scientia sermocinalis), he remarked that dialectic was primarily concerned with the sermo mentalis, and unlike grammar and rhetoric, which deal with conventional language, dialectic concerns itself with the naturally significant sermo. 34 Once the concept had come to be regarded as a sign, we find in some early sixteenth-century authors, especially Domingo de Soto, a careful classification of both signs and types of signification. 35 In relation to the speaker, spoken words were said to be instrunlental signs, because of their causal properties, and lllental terms were said to be formal signs, because they represented by their very nature. In relation to the things signified, spoken words were said to be conventional signs, and mental ternlS were said to be natural signs. Fonseca, of course, made the same distinctions. 36 The doctrine that spoken words and concepts are signs raises obvious questions about such logical terlllS as "all," and "none," about non-referring terms such as "chimera," and about supposed analogical terms such as "being." I shall ignore the first two cases, though I shall return to the third, and for the moment I shall confine myself to terms such as "cow" or "dog" which pick out ordinary physical objects belonging to natural kinds. There are three important areas of discussion here: the word-concept-thing relation, the word-signijicatum relation, and the question of whether words can signify in a way that concepts do not. The precise nature of the word-concept-thing relation was the focus of the long-standing debate whether spoken words signify concepts or things. 37 All the participants in the debate agreed that concepts play an essential role in the significative process, for we cannot refer to objects we do not know; and they also agreed that words are typically used to pick out things in the world. Fonseca reminds us that names are not imposed principally (praecipue) to signify concepts, but the thing itself. 38 If I say "Some dogs are running," what I say is true (if at all) of individual dogs and not of my concepts. What people disagreed on was how the role of concepts in the significative process was to be described. Some authors. held that words primarily signified or made known concepts and only secondarily signified things. Others, following Ockhanl, held that words signified things alone while being subordinated to concepts. Fonseca took up the issue in his Institutiones and argued that while concepts signified things immediately, with no intervening sign, words signified things only mediately, through the intervening concepts of which the words were the proximate signs. 39 He also pointed out that when one hears a word, one need not form two concepts, one of the concept of the thing, and the other of the thing signified, for the mind can go directly to the thing. This account was complicated by the doctrine of COlllmon natures which have somehow to be involved in the referential process. Whenever a general

52 E. JENNIFER ASHWORTH term is used, the mind must apprehend the particular thing referred to as belonging to a genus or species, but, for a moderate Thomist, the common nature is neither the universal concept nor a separate Platonic idea. It is at this point that we find discussion of a word's significatum. Fonseca said that a word has both an immediate significate and mediate significates, and the immediate significate is a common nature. 40 He discusses these distinctions in the context of supposition theory. In "Man is a species of animal," the word "man" is said to have simple supposition in that it stands for its immediate significate, the common nature man. In other contexts, such as "A man is running," the word has personal supposition, that is, it stands for its mediate significates, the individuals. In yet other contexts, such as "Man is animal," Fonseca postulated an intermediate kind of supposition, called absolute supposition, by virtue of which the subject stands for a common nature which is not thought of as abstracted and separate from individuals. 41 In his Metaphysics commentary Fonseca emphasizes that all univocal terms in normal contexts have a double supposition in that they stand both for individuals and for the common nature taken in the imprecise manner just described. 42 The last problem concerns the nature of conceptual representation: when a spoken term signifies a thing, how determinate does the intervening concept have to be? This issue was introduced for Fonseca by Scotus's argument that our words can signify more clearly than our concepts warrant,43 Fonseca takes it up in his Metaphysics commentary in a question about the word "God.,,44 Can this single word signify God simply, in the absence of a simple conception of the divine essence? Fonseca argued that we cannot separate the signification of words from the signification of concepts, for the significatio of a name is nothing other than the representation of the thing signified by means of a concept. If a word signifies simply then we have a concept which represents its object in a simple manner. He then claimed that because the word "God" is a simple term, we do have a sinlple concept of God, rather than a complex concept that essentially involves the operations of negating and relating. This does not mean that we grasp God's essence. Even though the concept signified by the word "God" is proper to God, in the sense of successfully picking out God rather than some other object, it does not represent him distinctly but confuse, in an undiscriminating manner, because we do not conceive the divine attributes through it, nor did we acquire it as a result of perceiving divine attributes. In some cases an ordinary undiscrinlinating concept can be replaced by a full concept. For instance, we can start with an undiscriminating concept of man, and later, when we come to grasp a quidditative definition, we can form a simple concept of the whole essence. However, only the blessed in heaven can know God by a simple concept through which all the divine perfections are known clare ac perspicue, and we cannot arrive at such a concept of God naturally. For those who wonder how we even begin to make successful reference to God, Fonseca explains that when we he3-~_~~~~~iptiye_

Petrus Fonseca on Objective Concepts and the Analogy ofBeing 53 phrases such as "simply infinite being" or "first cause," we can without any difficulty form a simple concept of the underlying substance (ipsius laten/is substantiae), or as we might say, the whatever it is that is being picked out by the speakers.

3 Analogy and Concepts After this overview, it is time at last to turn to analogical terms, which were not thought to function in the same way as the terms "rnan" and "ox." l~ogicians and theologians together had developed a theory whereby words could be divided into three sorts, independently of context. 45 Some were univocal (always used with the same sense), some were purely equivocal (used with totally different senses), and some were analogical (used with related senses). These divisions have thirteenth-century roots, and were produced in response to three problems. In logic, there is the problem of Aristotle's distinction between equivocal and univocal terms at the beginning of the Categories. Equivocal terms (e. g., "bank") are those which can be used in two quite different senses, and it seems natural to extend the notion of an equivocal term to cover those terms that are used in different but related senses. In metaphysics, there is the problem of how to speak of being (ens), given that the being of a substance is so different from the being of an accident. In theology, there is the problem of religious language. How can words normally used of humans, such as "just" or "good," be meaningfully used of God, when God is so different from human beings? In each context, analogy seemed to provide an answer, and "ens" soon became the main example of an analogical term in all three contexts, with the result that we find considerable overlap between discussions in logic, in metaphysics, and in theology. Fonseca himself pays attention to the theological problem in his commentary on the Metaphysics, and I have already noted that his views were taken up in two later commentaries on Aristotle's logic. The term "analogy" itself had two senses. In the original, Greek, sense, it involved a comparison of two proportions. Thus "principle" was said to be an analogical term when said of a point and of a spring of water because a point is to a line as a spring is to a river. This type of analogy caIne to be called the analogy of proportion, proportionality, or (in the hands of Cajetan), proper proportionality. In the second sense, it involved a relation between two things (or one pair of things and a third), of which one thing is secondary and the other primary. Thus "healthy" was said to be an analogical term when said of a dog and its food because while the dog has health directly, its food is healthy only as contributing to or causing the health of the dog. This second type of analogy became known as the analogy of attribution. One of the main subjects of debate was how to classify types of analogy, and how to apply the various types to the different metaphysical and theological cases. Although in one of his writings (De veritate 2.11) Aquinas said that

E. JENNIFER ASHWORTH religious language must be interpreted by means of the analogy of proportionality, in other writings he appealed to the analogy of attribution. During the fourteenth and fifteenth centuries most logicians and theologians, including Capreolus, appealed to the analogy of attribution (if they discussed the topic at all). At the end of the fifteenth century, however, Cajetan argued that the analogy of proportionality was the only true analogy, and that whereas the analogy of attribution only involved extrinsic properties, the analogy of proper proportionality involved intrinsic properties. Cajetan's view was not always accepted, but there was much discussion of the issue in later scholastics, including Soto and Suarez. Fonseca presented a complex theory, for while he fully recognized the legitimacy of the analogy of proportionality, he also argued that it could be combined with the analogy of attribution. 46 Some terms are analogical by the analogy of proportionality alone, as in the case of "principium." This term is not applied to subordinate principles by virtue of their relationship to some one chief principle, and the relevant form is intrinsic to all analogates. Other terms are analogical by the analogy of attribution alone, and here the form referred to is intrinsic to the principal analogate alone. There are, however, terms which combine the two types of analogy. These include metaphorical terms. Thus "ridere" as said of fields and fortunes picks out a similarity of proportions but it also implies a reference to a smiling face as a principal analogate, because imitation, and hence some attribution, is involved. Indeed, an intrinsic form is also involved, for laughing does in some sense (suo modo) inhere in the flowering field,47 just as "pedalitas" inheres in the foot of a cOUCh. 48 However, Fonseca's chief example of the corrlbination of two types of analogy is "ens." On the one hand, it is clear that the form of entitas is intrinsic to all analogates, God and creatures, substance and accidents, and that each is related to its own esse as the others are. On the other hand, it is also clear that attribution is involved. What belongs to God through his essence and to creatures through participation, involves attribution, and since creatures are entia only through participation, "ens" must be said of them through attribution to God. There is no implication, as there was with Cajetan, that two different senses of "ens" must be involved. Finally he argues that the analogy of attribution precedes the analogy of proportionality, for the proportionality between God and creatures with respect to being is founded on God's having bestowed being on creatures. While Fonseca does not explicitly reject Cajetan's use of the analogy of proper proportionality, it plays very little part in his own doctrine. It is there, but only as an appendage to the analogy of attribution, which is fundamental. The other main subject of debate in treatments of analogy was very closely related to philosophy of mind, and it springs mainly from the work of early fourteenth-century philosophers, particularly Duns Scotus. In the Categories, Aristotle (in Latin translation) had said that the difference between a univocal term and an equivocal term was that the latter was subordinated to more than 54

----------- -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

Petrus Fonseca on Objective Concepts and the Analogy ofBeing 55 one ratio substantiae. 49 In the early fourteenth century the ratio substantiae soon came to be identified with a concept, and the question then was, how many concepts are involved when an analogical term is used? There were three vie~s among those logicians and theologians who accepted analogy. 50 The nominalists held that analogical terms were straightfonvardly equivocal terms subordinated to two distinct concepts but the Thomists were split. Analogical terms could be viewed as subordinated to an ordered cluster of concepts (possibly but not necessarily described as a disjunction of concepts) ~ or they could be subordinated to a single concept which represents in a prior and a posterior manner (per prius et posterius). Scotus rejected all these possibilities, arguing that "ens" was a univocal term subordinated to a single univocal concept. Even for those within the Thonlistic tradition, Scotus's arguments about the univocity of "ens" had to be taken seriously. On the one hand, the word does not seem to be straightfonvardly equivocal, in the sense of being subordinated to more than one concept, for we at least have the illusion of being able to grasp "ens" as a general term. As Scotus pointed out, in an argument reproduced by all who considered the issue, we can grasp that something is a being while doubting whether it is a substance or an accident, and this surely involves having a relatively simple concept of being at our disposal. 51 On the other hand, there does not seem to be any common nature involved, and in the absence of a common nature, Thomists thought that to call the term "univocal" was inappropriate. What was needed was a way of allowing the concept to enjoy some kind of unity, while allowing the word to have a significate that was not a simple common nature. For many thinkers from the early fourteenth century onward, the distinction between formal and objective concepts provided the answer. Before going any further, we should consider the word "conceptus." It is the past participle of the verb "concipere," and as such, it obviously has a dual meaning. It can refer to the result of mental conceiving, i. e., to whatever it,is that is formed in the mind, or it can refer to the object conceived, by picking out one of its passive modes. 52 In this second sense, the conceptus, the thing conceived, need not be a mental item, and it is for this reason that authors often specified that they were speaking of the conceptus mentis or conceptus mentalis, phrases which sound strange to the modern ear. 53 In the light of the distinction between two ways of taking "conceptus," we can look at the formal and objective concept. In Fonseca's definitions, which were perfectly standard, the conceptus formalis is said to be an actual similitude of the thing understood, produced by the intellect. 54 Here we should note that the notion of similitude is not that of a pictorial likeness (though it can be). Any kind of sharing of form will count. The conceptus obiectivus is the thing which is understood according to the form or nature conceived by the formal concept, 55

56 E. JENNIFER ASHWORTH or the thing which is understood, insofar as it is represented by the formal concept. 56 The obvious question then was, what is the object represented, and does it have only objective being or the being of being thought (esse objective or esse cognitum), or does it have some kind of real being as well? Fonseca's answer is related to his theory of conlmon natures as general objects whose communitas exists only obiective in intellectu, though they are truly said of real things. 57 In the case of "ens," for which there is no common nature, he introduces a theory of simple entities (entitates simplices) which include God (in whonl, we should renlember, there is no difference between abstract and concrete, essence and .existence), the supreme genera, differentiae, matter, and substantial form. 58 In his discussion of these entities, he appeals to his doctrine of absolute supposition. If I say "Substantia est ens," there is a double reference: in one way, I am speaking of being as such without any connotation of separation or abstraction, but at the same time, I am speaking of the secondary significates, substance, quantity and so on. 59 4 Fonseca on Formal and Objective Concepts We are now ready to look at Fonseca's distinctions more closely. He began by arguing that the formal concept of being can be one of three types: fully distinct (distinctus), fully undiscriminating (confusus) , or partly undiscriminating, partly distinct. 60 A fully distinct concept of being is one which represents all simple entities "determinate et expresse," and only God can have such a concept. A fully undiscriminating concept is one which represents all simple entities of this sort "confuse et indeterminate," and an example is provided by the subject and predicate terms in the mental proposition "Ens est id quod est." The concept of substance as a supreme genus is an example of the third type since it represents being by virtue of representing substance clearly and distinctly (expresse et determinate) and the other categories implicitly and indeterminately (implicite et indeterminate), as all having some relation (proportio) to substance. He mentioned that the same distinctions can be applied to an objective concept viewed as the immediate significate of a formal concept. Fonseca then introduces a further distinction between two ways of being one, and two ways of being distinct. 61 These ways were in re, that is, in reality, when viewed as a thing, and in ratione, which might be translated here as "with respect to content" or "with respect to understanding." The formal concept of man is one in re because it is one quality of the nlind~ it is also one in ratione because it directly (prorsus) represents one human nature, which is multiplied in individuals. The corresponding objective concept is one in ratione but not one in re, given that there is no Platonic Idea or common man. Similar distinctions can be applied to ways of being separate (praecisus).62 A formal

Petrus Fonseca on Objective Concepts and the Analogy ofBeing 57 concept is separate in re when it is a separate quality of mind, in the way that our concept of man is separate from our concepts of Plato and of Socrates. Plato took the objective concept of man also to be separate in this way, but Fonseca does not, since common natures cannot exist apart from individuals. A concept is separate in ratione when it has been abstracted, that is, when it includes nothing proper to the concepts from which it was separated. Thus the formal concept of man is separate because it represents nothing proper or peculiar to any given individual, and the objective concept or quiddity of man is separate because it contains nothing proper or peculiar to any actual individual. In the light of these distinctions, Fonseca set out to answer the main questions that arose in the context of analogical concepts, whether formal or objective. These were first, whether such concepts could be viewed as truly one, and second, whether they could be viewed as separate (praecisum) fronl the proper concepts of their dividing members. 63 In his discussion of the formal concept of being, Fonseca concluded that while a fully distinct formal concept of being could not be one quality, presumably since each part would constitute a single concept, any partly or wholly undiscriminating formal concept is one in re, for it can be the subject of a mental proposition. 64 The fully undiscriminating concept of being is also separate in re, i. e., separate from the concepts of the dividing menlbers, for any concept which gives a precise representation of what is undiscriminatingly represented here will be a separate quality of the nlind. However, Fonseca goes on, even if a fully undiscriminating concept of being is simply one and separate from other concepts, it cannot be called one concept in a simple sense, for it is one only in the way that an equivocal name is one word. 65 The notion of being one concept carries with it reference to separateness in ratione, which is equivalent to abstraction, and the concept of ens is not abstracted from its dividing members. The formal concept of being cannot be either one or separate in ratione, that is with respect to its content, what it is of or about, unless it is so merely secundum quid. Strictly speaking, it is an ambiguous or multiple concept (conceptus multiplex). Fonseca did not enlarge on this point, but it is a body blow for any theory that thought constitutes an ideal language in the sense of being perfectly clear and precise. Fonseca's conclusion that the formal concept of being is a multiple concept followed directly from his belief that what it represents, the objective concept of being, is not truly one or separate either in re or in ratione. 66 As in the univocal case, the objective analogical concept can be neither one nor separate in re, for that would be Platonism. Moreover, even the wholly undiscriminating concept can be called one and separate in. ratione only secundum quid, or in a relative sense, for otherwise it would be univocal. To be separate in ratione is to be abstracted, and only common natures can be abstracted. While anything proper to Socrates and Plato can be excluded from the essence of man, the sanle is not true of substance "and other simple entitates" where ens is concerned. There is always implicit inclusion of these dividing members.

58

E. JENNIFER ASHWORTH At this point, while denying that the objective concept of being is truly separate from its members, Fonseca differentiated himself from those who denied that the objective concept was separate on the grounds that it is one only by the unity of a disjunction, and hence was constituted by a sequence of objective concepts. 67 Disjunction, said Fonseca, is not in the things but in the indeterminacy of the intellect, and when we say that sonlething is a being using an undiscriminating formal concept we do not think of it as a substance or as a quality or as any particular entity. Since we cannot conceive a disjunction without conceiving its parts, disjunction cannot be a feature of what is represented. Fonseca also explicitly rejected the view that the objective concept is both one and separate in ratione, but is not univocal because it is unequally participated, being said in a prior and a posterior way of God and creatures, substance and accidents. 68 If one asks how unequal participation might occur, it turns out that there are four possibilities, none of which is viable. First, one could appeal to unequal participation according to the ratio (or nature in a loose sense) participated in. This leads to inconsistency, because one will have to explain the unequal participation in one ratio by saying that God has one ratio insofar as he is being, substance has another and so on, with the result that there is no one ratio. Second, one could appeal to unequal participation according to the esse of the participators. This would lead us to say that a genus is analogical because it has unequally perfect existence in the different species, but a genus is not analogical. Third, one could appeal to unequal participation according to degree, but this would produce the absurd result that such ordinary terms as "white" were analogical. Finally, one could appeal to unequal participation according to diverse nl0des of predication, for instance, accidental and essential predication. However, this kind of inequality is irrelevant because ens is not merely intrinsic but essential to all. Conclusion I don't want to claim that I can point to precise passages in Fonseca which have influenced Descartes, or Mersenne, or Arnauld, or any other early modern philosopher. On the other hand, I do want to claim that this is the style of discussion, and these are the types of distinctions, with which early modern philosophers, at least up to and including Locke, would have been familiar through the scholastic texts by which they were educated. NOTES 1. On Fonseca's life and works, see Charles H. Lohr, Latin Aristotle Commentaries: II. Renaissance Authors (Florence: Olschki, 1988), pp. 150-51; and John P. Doyle,

Petrus Fonseca on Objective Concepts and the Analogy ofBeing

59

"Peter Fonseca,"Routledge Encyclopedia of Philosophy, E. 1. Craig, ed. (Routledge: forthcoming). 2. Pedro da Fonseca, Institui90es Dialecticas. Instilutionum Dialecticarum Libri Octo, 2 volumes, Joaquim Ferreira Gomes, ed. and trans. (Coimbra: Universidade de Coimbra, 1964), cited as Instil. Dial.; Petrus Fonseca, Commentariorvm In Metaphysicorvm Aristotelis Stagiritae Libros (2 volumes), (Cologne, 1615; reprinted Hildesheim: Georg Olms, 1964), cited as In Met. Volume 1 contains Tomus I-II and has numbered columns; volume 2 contains Tomus ill-IV, and has numbered pages. 3. Timothy 1. Cronin, Objective Being in Descartes and in Suarez (Roma: Gregorian University Press, 1966), p. 34 4. Cronin suggests, pp. 32-33, that Fonseca's commentary was normally used in Jesuit schools. 5. Useful background is provided by Peter Dear, Mersenne and the Learning of the Schools (Ithaca and London: Cornell University Press, 1988). However, Dear overemphasizes the influence of Ramist-style humanism on Fonseca's logic: see pp. 19-21. For an alternative assessment of Fonseca, see E. 1. Ashworth, "Changes in Logic Textbooks from 1500 to 1650: The New Aristotelianism," Aristoteli~mus und Renaissance: In Memoriam Charles B. Schmitt, Eckhard Kessler, Charles H. Lohr and Walter Spam, eds. (Wiesbaden: Otto Harrassowitz, 1988), pp. 82-84. 6. Jean-Luc Marion, Sur la theologie blanche de Descartes (Paris: Presses Universitaires de France, 1981), devotes much of his book to the theme of Descartes's reaction to Suarez's doctrine of the analogy of being. In his discussion, he notes the relation between the analogy of being and objective concepts (e. g., p. 119), and he also mentions Fonseca briefly (p. 123). However, the nature and scope of our investigations is quite different. 7. Descartes, Replies IV, AT VII 235. For discussion see Roger Ariew, "Descartes and scholasticism: the intellectual background to Descartes' thought," The Cambridge Companion to Descartes, John Cottingham, ed. (Cambridge: Cambridge University Press, 1992), pp. 58-90, esp. p. 79. 8. R. Dalbiez, "Les sources scolastiques de la theorie cartesienne de l'etre objectif it propos du 'Descartes' de M. Gilson," Revue d'histoire de la philosophie 3 (1929), pp. 464-472. 9. For Fonseca on God's ideas, including reference to fonnal and objective concepts, see In Met., III, pp. 280b-296b, esp. 286a-288b (Lib. VII, cap. VIII, q. 2). 10. Dalbiez, pp. 468-470. 11. Cronin, p. 206, opts for Scotus and Suarez as Descartes's sources. One useful source is Gabriel Nuchelmans, Judgment and Proposition from Descartes to Kant (Amsterdam, Oxford, New York: North-Holland Publishing Company, 1983). He traces the early history of the phrases "esse obiective" and "conceptus obiectivus" in Hervaeus, Aureol, and Durandus, pp. 17-26. In discussing Descartes, he says (p. 41), "it remains difficult to single out any individual sources. His debt is of a very general nature and could have come from any work belonging to a certain climate of thought. There can be little doubt, however, that one of the main detenninants of this climate was the objective-existence theory as it had been developed by such thinkers as Durandus and Aureolus." 12. Nonnan 1. Wells, "Objective Reality of Ideas in Descartes, Caterus, and Suarez," Journal ofthe History ofPhilosophy 28 (1990), pp. 33-61, esp. pp. 49-50. See also Nonnan 1. Wells, "Objective Being: Descartes and His Sources," The Modern

60

E. JENNIFER ASHWORTH

Schoolman 45 (1967), pp. 49-61; id., "Objective Reality of Ideas in Arnauld, Descartes, and Suarez," The Great Arnauld and Some ofHis Philosophical Correspondents, Elmar 1. Kremer, ed. (Toronto, Buffalo, London: University of Toronto Press, 1994), pp. 138183. 13. Calvin Normore, "Meaning and Objective Being: Descartes and His Sources," Essays on Descartes' "Meditations," Amelie Oksenberg Rorty, ed. (Berkeley, Los Angeles, London: University of California Press, 1986), pp. 223-241, esp. p. 240. 14. Hervaeus Natalis, Quolibeta (Venetiis, 1513; reprinted Ridgewood, N.J.: Gregg Press, 1966), ff. 43ra-46vb, esp. f. 43rb; Peter Aureol, Scnptum Super Pn'mum Sententiarum,2 vols., Eligius M. Buytaert, ed. (St.Bonaventure, N.Y.: The Franciscan Institute; Louvain, Belgium: E. Nauwelaerts; Paderbom, Germany: F. Schoningh, 1956), pp. 471-523, esp. p. 483. For a fuller discussion of these authors, see E. J. Ashworth, "Analogical Concepts: The Fourteenth-Century Background to Cajetan," Dialogue 31 (1992), pp. 399-413. 15. Johannes Capreolus, Defensiones theologiae divi Thomae Aquinatis. Vol. I, Ceslai Paban and Thomas Pegues, eds. (Turonibus: A. Cattier 1900-1908; reprinted, FrankfurtJMain: Minerva GmbH, 1967), pp. 117a-144a. 16. For more infonnation, see Michael Tavuzzi, "Hervaeus Natalis and the Philosophical Logic of the Thomism of the Renaissance," Doctor Communis 45/2 (1992),pp.132-152. . 17. Dominic of Flanders, Quaestiones super XII Libros Methaphysicorum (Venice, 1499; reprinted, Frankfurt: Minerva G.M.B.H., 1967), sigs. h 6 vb-i 8 vb; Paulus Soncinas, Quaestiones Metaphysicales Acutissimae (Venice, 1588; reprinted, Frankfurt: Minerva G.M.B.H., 1967), pp. la-l1a, esp. p. 2b. Dominic of Flanders refers explicitly to Hervaeus, e. g., sig. i 7ra, and Soncinas quotes Aureol, e. g., p. 2a. For a fuller discussion of these authors, see E. J. Ashworth, "Suarez on the Analogy of Being: Some Historical Background," Vivarium 33 (1995), pp. 50-75; and Michael Tavuzzi, "Some Renaissance Thomist Divisions of Analogy," Angelicum 70 (1993), pp. 93-122. 18. See Bruno Pinchard, Metaphysique et semantique. Autour de Cajetan. Etude [texteJ et traduction du tiDe Nominum Analogia" (Paris: J. Vrin, 1987). 19. Dorningo de Soto, Liber Praedicamentorum in In Porphyrii Isagogen, Aristotelis Categorias, librosque de Demonstratione Absolutissima Commentaria (Venice, 1587; reprinted Frankfurt: Minerva, 1967), pp. 112a-121a, 129a-136a. For references to Capreolus and Cajetan, see, e. g., p. 130b. For discussion, see E. 1. Ashworth, "Domingo de Soto (1494-1560) on Analogy and Equivocation," Proceedings of the Third Pamplona Conference on the History of Logic, Ignacio Angelelli and Maria Cerezo, eds. (New York, Berlin: Walter de Gruyter, 1996) pp. 117-131. 20. Francisco Suarez, Disputationes Metaphysicae in Opera Omnia, vols. 25 and 26 (Paris, 1866; reprinted Hildesheim: Georg Olms, 1965), see especially disputations 2, 28, and 32. For instance, d.2.9 (v.25, p. 68a) includes references to Cajetan, Capreolus, Dominic of Flanders, Fonseca, Hervaeus, Scotus, Soncinas, and Soto. 21. [Sebastian de Couto], Commentarii Collegii Conimbricensis e Societate Jesu. In universam dialecticam Aristotelis (Cologne, 1607; reprinted HildesheimlNew York: Georg Olms, 1976), cols.305-327. For instance, Cajetan, Fonseca, Soto, and Suarez are cited in col. 320 and Hervaeus in col. 321. 22. Antonius Rubius, Tractatus de Nominum Analogia in Commentarii in Universam Aristotelis Dialecticam (Coloniae Agrippinae, 1609), pp. 148-183. Cajetan,

Petrus Fonseca on Objective Concepts and the Analogy ofBeing

61

Fonseca and Soto are cited on p. 158a, Capreolus, Dominic of Flanders and Soncinas on p. 160b, Aureol on p. 167a and Suarez on p. 173b. 23. Fonseca, In Met. I, cols. 689-736 (Lib.4, cap. 2, qq. 1,2). See, e. g., col. 713 for Capreolus, Cajetan, Dominic of Flanders, Hervaeus and Soncinas, and col. 727 for Aureol. 24. Franciscus Toletus, Introductio in universam Aristotelis logicam in Opera omnia philosophica I (Cologne, 1615/16; reprinted Hildesheim: Georg Olms, 1985), pp. 23a-33b, especially p. 28a-b (on the universal as an ens rationis), pp. 30b-31a (on esse obiective in intellectu), p. 32a-b (on formal and objective concepts). 25. Fonseca, In Met. II, cols. 1036-1037. Cf. [Sebastian de Couto], col. 141. For the full discussion of universals, see Fonseca, In Met. II, cols. 947-1050 (Lib. 5, cap. 28, qq. 1-10). For entia rationis, see Fonseca, In Met. II, cols. 462-469 (Lib. 5, cap. 7, q.6). 26. Soto, Liber Praedicabilium in In Porphyrii Isagogen, etc., pp. 28a-45a, especially p. 30b (formal and objective concepts), p. 39a (entia rationis); [Sebastian de Couto], cols. 77-161, especially cols. 151 and 155-158 (entia rationis) and cols. 140141 (concepts). The phrase "conceptus obiectivus" is not used. On cols. 151-152 there is a discussion of whether entia rationis need an efficient cause. 27. Descartes, AT ill 185. 28. Fonseca, Instil. Dial., p. 36. The editor gives a reference to the definition in Augustine's De dialectica rather than to De Doctrina Christiana 2.1, the more usual source. 29. Fonseca, Instil. Dial., p. 32. 30. Fonseca, Instil. Dial., p. 34. For Pierre d'Ailly, see Paul Vincent Spade, Peter of Ailly: Concepts and Insolubles. An Annotated Translation (Dordrecht, Boston, London: Reidel, 1980), p. 16. 31. Fonseca, In Met., I, col. 408 (Lib. 2, cap. 1, q.2): "quia ea de caussa sunt nomina rebus imposita, ut homines suos conceptus sibi mutuo aperirent, & communicarent." 32. See in particular Joel Biard, Logique et theorie du signe au XIVe siecle (Paris: 1. Vrin, 1989). 33. See E. J. Ashworth, "The Structure of Mental Language: Some Problems Discussed by Early Sixteenth Century Logicians," Vivarium 20 (1982), pp. 59-83; reprinted in ead., Studies in Post-Medieval Semantics (London: Variorum Reprints, 1985). 34. Fonseca, In Met., I, col. 492 (Lib. 2, cap. 3, q. 1): "Quia igitur Dialectica de sermone agit, primo quidem de mentali, quo quisque secum quasi loquitur: secundaria vero ratione de vocali & scripto, non in realibus, sed in sermocinalibus numeranda erit. quanquam etiam dicitur rationalis, non quia de ente rationis agat (isto enim modo potius dicendae essent rationales, Grammatica et Rhetorica, quae agunt de solo sennone significativo ex instituto, quam Dialectica, quae primo agit de sennone significativo naturaliter) sed quia dirigit operationes intellectus nostri, qui propter naturalem discurrendi facultatem ratio appellatur." 35. See E. 1. Ashworth, "Domingo de Soto (1494-1560) and the Doctrine of Signs," De Ortu Grammaticae. Studies in medieval grammar and linguistic theory in memory of Jan Pinborg, G.L. Bursill-Hall, Sten Ebbesen and Konrad Koerner, eds. (Amsterdam: John Benjamins, 1990), pp. 35-48; and ead., "The Doctrine of Signs in Some Early Sixteenth-Century Spanish Logicians," Estudios de Historia de la Logica. Actas del II Simposio de Historia de la Logica: Universidad de Navarra Pamplona 25-27 de Mayo

62

E. JENNIFER ASHWORTH

de 1987, Ignacio Angelelli and Angel d'Ors, eds. (Pamplona: Ediciones EUNATE, 1990) pp. 13-38. 36. Fonseca, Instil. Dial., pp. 34-38. 37. E. J. Ashworth, "'Do \Vords Signify Ideas or Things?' The Scholastic Sources of Locke's Theory of Language," Journal of the History ofPhilosophy 19 (1981), pp. 299326; ead., "Locke on Language," Canadian Journal ofPhilosophy 14 (1984), pp. 45-73; ead., "Jacobus Naveros (fl. ca. 1533) on the Question: 'Do Spoken Words Signify Concepts or Things?'," Logos and Pragma: Essays on the Philosophy of Language in Honour ofProfessor Gabriel Nuchelmans, L. M. de Rijk and H. A. G. Braakhuis, eds. (Artistarium Supplementa 3. Nijmegen: Ingenium Publishers, 1987), pp. 189-214. The first two papers are reprinted in ead., Studies in Post-Medieval Semantics. 38. Fonseca, In Met., I, col. 409 (Lib. 2, cap. 1, q. 2). 39. Fonseca, Instil. Dial., pp. 38-40. 40. Fonseca, Instil. Dial., pp. 690-696. 41. For discussion, see E. J. Ashworth, Language and Logic in the Post-Medieval Period (Dordrecht, Boston: Reidel, 1974), pp. 87-88. 42. Fonseca, In Met., I, col. 722 (Lib. 4, cap. 2, q. 2). 43. For discussion, see E. J. Ashworth, '''Can I speak more clearly than I understand?' A problem of religious language in Henry of Ghent, Duns Scotus and Ockham," Historiographia Linguistica 7 (1980), pp. 29-38. 44. Fonseca, In Met., I, cols. 407-411 (Lib. 2, cap. 1, q. 2, sect. 4). 45. For discussion and references, see E. J. Ashworth, "Analogy and Equivocation in Thirteenth-Century Logic: Aquinas in Context," Mediaeval Studies 54 (1992), pp. 94135; ead., "Equivocation and Analogy in Fourteenth Century Logic: ockham, Burley and Buridan," Historia Philosophiae Medii Aevi: Studien zur Geschichte der Philosophie des Mittelalters, Burkhard Mojsisch and Olaf Pluta, eds. (Amsterdam, Philadelphia: B. R. Gruner, 1991), vol. 1, pp.23-43. 46. Fonseca, In Met., I, cols. 701-710. 47. Fonseca, In Met., I, col. 703. 48. Fonseca, In Met., I, col. 708. 49. "Aequivoca dicuntur quorum nomen solum commune est, secundum nomen vero substantiae ratio diversa, ut animal homo et quod pingitur. Univoca vero dicuntur quorum et nomen commune est et secundum nomen eadem substantiae ratio, ut animal homo atque bos..." Aristotle, Categories (1 a1-2) in Aristoteles Latinus I 1-5. Categoriae vel Praedicamenta, L. Minio-Paluello, ed. (Bruges, Paris: Desclee de Brouwer, 1961), p.5. 50. See Ashworth, "Analogical Concepts: The Fourteenth-Century Background to Cajetan," and ead., "Suarez on the Analogy of Being: Some Historical Background." 51. John Duns Scotus, Ordinatio I d.3 pars 1 q. 1-2, 27-29, in Opera Omnia (Vatican City: 1950-), Vol. ill, pp. 18-19. Fonseca reproduces the argument, In Met., I, col. 691. 52. Nuchelmans, p. 22, cites a passage from Aureol: "conceptus autem obiectivus non est aliud quam res apparens obiective per actuffi intellectus; qui dicitur conceptus, quia intrinsece includit ipsum concipi passivum." 53. See Hervaeus Natalis, f. 43 va; [Sebastian de Couto], col. 322; Cajetan, p. 123 par. 36. 54. Fonseca, In Met., I, col. 710: "Conceptus fonnalis nihil est aliud, quam actualis similitudo rei, quae intelligitur, ab intellectu ad earn exprimendam producta...Est et alia

Petrus Fonseca on Objective Concepts and the Analogy ofBeing

63

ratio, cur dicatur fonnalis: nempe, quia repraesentat rem sub ea fonna seu natura secundum quam intelligitur." 55. Fonseca, In Met., I, col. 711: "Conceptus obiectivus est res, quae intelligitur, secundum earn fonnam, naturamve quae per formalem concipitur." 56. Fonseca, In Alet., I, col. 712: "... conceptus obiectivus respondet conceptui fonnali, cum sit nihil aliud, quam res quae intelligitur, quatenus per formalem repraesentatur..." 57. Fonseca, In Met., II, cols. 1037-1038 (Lib. 5, cap. 28, q. 9). 58. Fonseca, In Met., I, cols. 714-715,724. 59. Fonseca, In Met., I, col. 722. 60. Fonseca, In Met., I, cols. 714-715. He admits that he is drawing on material from Caj etan, but says that he hopes to present it in a more understandable manner. [Sebastian de Couto], col. 322, compares Fonseca and Cajetan. 61. Fonseca, In Met., I, col. 715. 62. See Fonseca, In Met., I, cols. 715-716. 63. See Fonseca, In Met., I, cols. 711-714, for the posing of the questions and an outline of the three main positions. 64. Fonseca, In Met., I, cols. 718-720. 65. It is on this point that Fonseca differentiates himself from holders of the "common view" (col. 711), and aligns himself with Cajetan (col. 714). 66. Fonseca, In Met., I, cols. 716-718. 67. He refers to Hervaeus Natalis, Dominic of Flanders, and Soncinas in stating the view, col. 713. He refutes it cols. 725-726. 6S. The view is stated Fonseca, In Met., I, cols. 712-713 and refuted cols. 724-725.

Arnauld on the Nature of Ideas as a Topic in Logic: The Port-Royal Logic and On True and False Ideas ELMAR J. KREMER

Antoine Arnauld's treatment of the nature of ideas in On True and False Ideas has long puzzled philosophers in two ways. First, with this treatise Arnauld began his attack on Malebranche in an unexpected way, for what he had promised was a criticism of Malebranche's Treatise of Nature and Grace, and it seemed odd to his contemporaries that he should begin by attacking the account of ideas given by Malebranche in the Search after Truth. Thus when Arnauld sent a draft of Ideas to his old friend and collaborator Pierre Nicole for comnlent, Nicole counselled against beginning his attack on Malebranche in this way: ...although I admired almost everything in the treatise, I am not entirely in agreenlent with the idea of making it a preliminary to your dispute with the Reverend Father. I do not find in it any principles necessary in order to destroy the foundations of the Treatise of Nature and Grace. Your subject lllatter has a small audience and can easily become confused. Why not begin with a subject more worthy of you, more theological, more popular, and one with a large audience?l

Malebranche hinlself complained that Ideas had no more connection with the Treatise ofNature and Grace than the Turkish war in Hungary.2 He had earlier complained that Arnauld had acted deviously, attempting to discredit his theodicy and his views on the distribution of grace by means of a hostile attack on an abstruse topic, 3 and there may be some truth in this complaint, for Arnauld says in a letter to Quesnel: In the meantime, it [Ideas] will be useful, if I am not mistaken, to

diminish the inflated opinion which many have of the reliability of our friend's mind: and people might expect him to have been mistaken in the matter of grace, if it can be shown that he wandered strangely astray in the questions of metaphysics, which he has always claimed as his strong point. 4

However, Arnauld never withdrew his claim that their disagreement about the nature of ideas had a bearing on their theological differences. 5 Yet commentators continue to be puzzled as to exactly what that bearing is. Thus Steven Nadler raises the question, "What binds together the mainly [sic] philosophical issues...dealt with in VFI [On True and False Ideas] and the Recherche, with those problems treated in Malebranche's Traite [de fa nature et de la grace] and Arnauld's Reflexions?" He says, "I can here only make several suggestions, since neither thinker really tells us what the connection might be." His isuggestions are that both sets of issues have to do with orthodoxy, understood 'las agreement with the views of the Fathers of the ~~u!~h~ !~a! fO! j\.!~~u!d~ 111 _

66 ELMAR J. KREMER least, both debates "can be seen as parts of a disagreement about how best to dignify and do justice to the true conception of God," and, thirdly, that Arnauld 6 thought his position on both issues best fosters piety and avoids pride. Nadler is looking for what On True and False Ideas and Arnauld's criticisms of Malebranche's theodicy "have in common.,,7 In contrast, I shall argue that the connection between the two projects is not doctrinal but rather methodological, that Arnauld intended what he says about the nature of ideas to forearm his readers against n1isguided ways of thinking that led Malebranche into theological error. The content of Arnauld's position on the nature of ideas has been no less puzzling. Indeed, Nadler notes that "Arnauld's...theory of ideas...has been subject to misunderstanding ever since the publication of VFI in 1683.,,8 Thomas Reid, in his survey of positions on the nature of ideas, says that Arnauld does not have a consistent position but rather wavers between the position that ideas are, quite simply, perceptions, and the "common," or Cartesian, position, according to which ideas are a sort of intermediate between perceptions and their real objects. 9 Early in the twentieth century A. O. Lovejoy, Morris Ginsberg and R. W. Church engaged in a debate about whether Arnauld, in Ideas, was a "direct realist" or a "representationalist" with regard to the perception of material things. 10 More recent commentators include John Yolton, who argues that Arnauld's account is an important step toward the "deontologizing of ideas";11 Nadler, who claims that Arnauld, in addition to attempting to refute Malebranche's theory, offers a solution to "the problem intentionality";12 and Richard Watson, who says that Arnauld's theory is a failed attempt to give an ontological explanation, in terms of a metaphysical system, of how we know objects by way of ideas. 13 This disagreement about the point of Arnauld's account of ideas is obviously connected to the first puzzle, about why he began his attack on Malebranche's theodicy by criticising his account of the nature of ideas. My proposal is that we can go a long way to clearing up both questions by considering Ideas as a continuation of the discussion in Chapter One of the Port-Royal Logic, which bears the title "Ideas according to their Nature and Origin." Let me first give a brief outline of my proposal, and then develop each of its steps in more detail. One of the main purposes of logic, according to the authors of the Port-Royal Logic, is "to discover and explain more easily the error or defect that can occur in the operations of our mind." (Logique, 38)14 The opening chapter of the work provides an account of the nature of ideas and then uses it to forearm its readers against two philosophical errors. Against this background, it would be natural for Arnauld, faced with what he took to be important errors in Malebranche' s theodicy, to initiate his attack with a "logical" investigation of the nature of ideas. My hypothesis is that Arnauld indeed thought of the main topic of Ideas in this way.15 I shall attempt to confirm this hypothesis by showing that the points about the nature of ideas

Arnauld on Ideas as a Topic in Logic

67

which he emphasizes are precisely what he would need to forearnl his readers against what he took to be the central mistakes of Malebranche's theodicy. What Arnauld emphasizes in his discussion of the nature of ideas is, in brief, that our mental operations are "essentially representative" of their objects. (Ideas, 20; OA, Vol. XXXVIII, 199)16 Arnauld took this to imply that our cognitive acts ("perceptions" in Arnauld's use of the term) determine not only what objects they represent or exhibit to the mind, but also what properties the objects are represented as having. He emphasized this point because he thought it would forearm his readers against the mistake which he says lies at the origin of Malebranche's theodicy, the mistake of transgressing the limits of human knowledge of God. I: The Nature of Ideas in the Port-Royal Logic17 For Arnauld and Nicole, "Logic is the art of using one's reason well in the knowledge of things, both in order to learn for oneself and in order to instruct others." (Logique, 37) Logic cannot teach us to think; that we are able to do naturally. But the logician can carry out reflections on what we are enabled by nature to do, reflections which enable us to do well what we can do naturally without them. Such reflections are of use in three ways: First, to be sure that we are using our reason well, because the consideration of the rule redoubles our attention. Second, to discover and explain more easily the error or defect that can occur in the operations of our mind... Third, to make us know the nature of our mind better by reflection upon its operations. This is something more excellent in itself, if it were considered only as a matter of speculation, than the knowledge of all corporeal things, which are infinitely lower than spiritual ones. (Logique, 38)

But knowledge of our mental operations, in so far as it is part of logic, is not a matter of speculation. Rather it is practical knowledge of mental operations, knowledge sought for the sake of the first two uses. Reflection upon the nature of the mind, as the study of an immaterial thing, could also be part of the science of metaphysics. Thus, in "Discours II" at the beginning of the Logic, we read that the work contains a good deal of material from other sciences, including metaphysics. Among the metaphysical topics that also appear as parts of logic are "the origin of ideas" and "the separation of spiritual ideas from bodily images." (Logique, 39)18 r-rhe discussion of the nature of ideas in Book I, Chapter I of the Logic begins with a refutation of the clainl that ideas are the same thing as "images painted in our brain." (Logique, 40) This position is rejected on several grounds: that we can think of shapes we cannot imagine, that we can think of our thinking but cannot imagine it, and that the difference between affirmative and negative judgments does not correspond to any difference in images. The

68 ELMAR J. KREMER authors then present their own account of the nature of ideas: "Thus when we speak of ideas, we are not referring to images painted in the imagination, but to anything that is in our mind when we can truthfully say that we conceive a thing, in whatever manner we conceive it." (Logique, 41) The authors assume that, given this account of the nature of ideas, it is obvious that if a person uses words understanding what he is saying, then he has ideas corresponding to what the words signify. Using this assumption, they give two examples of errors, both taken from Hobbes,. that we are armed against by a correct understanding of the nature of ideas: first, the mistake of thinking that we have no idea of God because we have no image of God, and, second, the theory that reasoning is merely a manipulation of words, and that, consequently, "our soul is nothing other than a motion in certain parts of the organic body." (Logique, 42) The last part of this discussion illustrates the second of the ways in which logic was said to be useful. But uncovering particular errors in judgment is not the only way in which reflection on ideas is supposed to contribute to the improvement of our mental operations and to the avoidance of error. On the contrary, "Since we cannot have any knowledge of what is outside us except by means of ideas that are in us, the reflections we are able to make on our ideas are perhaps the most important part of Logic, because they are the foundation of all the rest." (Logique, 39) The proposition that we have ideas without any corresponding images in the brain is foundational, for example, to the clainl, in the opening chapter of Part IV, "On Method," that what is known by the mind is more certain than what is known by the senses. The title of the chapter lays down two principles: "That the things we know by the mind are more certain than what we know by the senses; that there are things that the human mind is incapable of knowing." (Logique, 291) The second principle gives rise to a limitation on human knowledge that the authors consider particularly important: "The greatest limitation on the extent of the sciences is not to seek anything that is above us and that we cannot reasonably hope to comprehend. Of this sort are all questions about the power of God, which it is ridiculous to attempt to enclose in the narrow bounds of our mind..." (Logique, 295) These two principles mark out "the limits of our mind [les bornes de notre esprit]": larger than Hobbes would allow, smaller than would be claimed by anyone who thought he could understand the extraordinary effects of God's power. It is worth noting that what the Logic has to say about the nature of ideas is quite thin, and would be disappointing to anyone who looked there for a systematic metaphysical or epistemological discussion of ideas. But it is not the purpose of the Logic to provide such a discussion. Its purpose is rather practical, and it says no nlore about the nature of ideas than is needed for its practical purposes of the work: to help us use our cognitive faculties well and to avoid error. The sanle, I shall argue, is true of On True and False Ideas.

Arnauld on Ideas as a Topic in Logic

69

II: The Nature of Ideas in On True and False Ideas The doctrine on the nature of ideas presented in On True and False Ideas is consistent with what was said in the Logic: that an idea is anything that is in our mind when we conceive a thing, in whatever way we conceive it. What is new is an account of how what we "conceive" is "in our mind." In order to specify how this takes place, Arnauld recalls the notion of the objective being of the thing known in the mental act through which it is known: "I say that a thing is objectively in my mind when I conceive of it." (Ideas, 19~ OA, Vol. XXXVIII, 198) As far as I know this notion, which occurs frequently in Ideas and in later items in the controversy with Malebranche, does not appear in any of Arnauld's writings between the Fourth Objections (in 1641) and Ideas (in 1683). Arnauld's return to this notion pinpoints both his central disagreement with Malebranche, and his central, new claim about the nature of ideas. In his reply to Ideas, Malebranche says that the main issue between himself and Arnauld regarding "the nature of ideas" is whether our cognitive acts are "essentially representative of their objects," and Arnauld seems to accept this description of their disagreement. 19 Whereas Arnauld held that our perceptions, or, as I shall say, our cognitive acts, are essentially representative of their objects, Malebranche says that what represents material things to the mind is the divine idea of bodies, and what represents God to the mind is God himself. We can begin to see what this issue amounts to if we recall that in seventeenthcentury philosophical French, the verb "representer" was used to translate both the Latin verbs "exhibeo" and "repraesento,,,20 and rephrase the issue as follows: Whether our cognitive acts essentially exhibit their objects to us, or whether they are exhibited to us by something else. We will not be far from the thought of Arnauld and Malebranche if we think that to exhibit (represent) an object to the mind is to hold the object up before the mind, or, as Arnauld and Malebranche say, to make the object "present to" the nlind. Arnauld, then, maintains that all our cognitive acts exhibit their objects to us. Malebranche, by contrast, nlaintains that whenever the object of cognition is something other than mind and mental acts, the object is exhibited to us by something distinct from ourselves and our mental acts: a divine idea if the object is a body or bodies, and God himself if the object is God. As Malebranche would put it, material things are represented to us by the divine idea of material things, and God is represented to us by God Himself. To this picture, we need to add an important qualification. The proposition that a mental act, or an item of some other sort, exhibits an object to us, nlust not be understood as ascribing efficient causality to whatever does the exhibiting. Arnauld puts this point by saying that to represent an object to the mind is not a case of efficient causation, but rather of formal causation, and Malebranche eventually accepted this clarification. Arnauld's position can be put as follows: That person S has a cognition of object 0 is in every case a fact

70 ELMARJ. KREMER about person Sand S' s cognitive acts alone, and not about anything existing apart from S. In other words, representing a given object to the mind" is, according to Arnauld, an intrinsic property of a cognitive act?1 It is also an essential property of a mental act A cognitive act could not have a different object while remaining the same act. By contrast, according to Malebranche, where 0 is something other than mind and its acts, that S has a cognition of 0 is a fact about S, S's mental acts and an independently existing item I, which contains the perfection of 0 formally or eminently, and exhibits 0 to S. To repeat a point made above, if 0 is a body or bodies, I is the divine idea of bodies, while if 0 is God, I is God himself. Arnauld points out that Malebranche is not consistent in the way he describes our cognition of God: Malebranche often speaks of our "idea of God" even though it is his official position· that we should speak of an idea of X only if the X is exhibited to us by something distinct from X, and hence that there is no idea of God. Arnauld emphasizes that he uses "idea" and "perception" to refer to the same thing, which is a mental act. But he adds that "idea" refers to a mental act considered precisely as having something exist in it objectively. (Ideas, 19-21; OA, Vol. XXXVIII, 198-200) In fact he speaks of both perceptions and ideas as representations of objects and as items in which things exist objectively. He says that an idea not only represents or makes known its object, but it also represents various properties of its object. These can be represented as connected with the object in several different ways: A property may be represented in an idea as the essence or part of the essence of the idea's object (such as the property of being extended or the property of engaging in thought), or again it may be represented as a property that it is possible that the object may have because of its essence (such as the property of having such and such a shape, or the property of thinking about geometry), or again it may be represented as a property common to the object and to other objects that are essentially different from it (such as the property of being created). In all such cases, the property is represented as "compatible with the object (qui peut convenir a cet objet)." (Ideas, 129; OA, Vol. XXXVIII, 306) Because our ideas contain some of the properties of the things they represent, Arnauld says, we can discover how those properties are related to one another by reflecting upon the ideas. 22 Arnauld emphasizes that it is in the modes of our own mind that we can discover the connections among properties of things: .. .in addition to the reflection which can be called virtual and which is found in all our perceptions, there is another, more explicit, in which we examine our perception by another perception, as everyone can easily verify. This occurs especially in the sciences which are formed only by the reflections that men make upon their own perceptions, as when a geometer, having conceived a triangle as a figure bounded by three straight lines, has found, by examining

Arnauld on Ideas as a Topic in Logic

71

his perception of that figure, that it had to have three angles and that the three angles had to be equal to two right angles... ...WE ALSO SEE how we ought to interpret the statement that it is in the idea of each thing that we see its properties. Surely nothing is more useless for that purpose than the representative being, distinct from perceptions, which our soul is supposed [by Malebranche] to need in order to conceive of numbers and extension. (Ideas, 25-26, 28; OA, Vol. XXXVIII, 204-205, 207)

lIe emphasizes the same point in a letter to Leibniz: ...from the obscurity and difficulty of knowing how things are in God's knowledge, and what sort of connection obtains among them there, whether it is an intrinsic or extrinsic connection, so to speak, all I want to conclude is that it is not in God, who, as far as we are concerned, dwells in an inaccessible light, that we ought to seek the true notions, individual or specific, of the things we know, but rather in the ideas of them that we find in ourselves. 23

This conclusion supports and underscores the doctrine of the limits of human knowledge which had been presented in the Logic. Arnauld applies the doctrine to the idea of God, and the application is inlportant for his criticism of Malebranche's theodicy. In the controversy with Malebranche, the Defense de Mr. Arnauld contre la Reponse au livre Des vrayes et des fausses idees, published in 1684, Arnauld says that by reflecting on our idea of God "we recognize...what He is, insofar as the feebleness of our nature permits." (OA Vol. XXXVIII, 543) In this way he thinks that we can "see that he is eternal, all knowing, omnipotent, the source of all truth and beauty, creator of all things..." Yet our idea of God is abstract. 24 In particular, although our idea of God includes the property of being all knowing and the property of having a will that cannot be resisted, these properties are present only abstractly in our idea. We cannot understand, we have no idea, of how, specifically, it is that God knows or wills. Arnauld's clearest statement of the last point, that we have no specific idea of how it is that God knows or wills, occurs in his letter to Leibniz of May 13, 1686: What do we know at present [in this life] about God's knowledge? We know that he knows all things, and that he knows them all by a single and entirely simple act, which is his essence. When I say we know this, I mean that we are sure that it must be so. But do we comprehend it? Ought we not recognize that, however sure we are that it is so, it is impossible for us to conceive how it can be so? Can we conceive that God's knowledge is his very essence, which is entirely necessary and immutable, and yet that he has knowledge of an infinity of things which he might not have had, because the things might not have been? It is the same with respect to his will, which is also his very essence, in which there is nothing that is not

72

ELMARJ. KREMER necessary. And yet he wills and has willed from all eternity things that he might not have willed. 25

Arnauld, of course, held that we do have specific, clear and distinct ideas of our own cognition and volition. These ideas are of two types: First, there are ideas of knowing and willing which are not distinct from the knowing and willing themselves. Arnauld held that every mental act is conscia sui, and as such can be called an idea of itself as well as of its original object. (Ideas, 25~ OA, Vol. XXXVIII, 204) Every act of mind, in addition to exhibiting some26 thing other than itself to the mind, also exhibits itself to the mind. Second, there are the ideas of our mental acts which arise when "we examine our perception by another perception." (Ideas, 26-8~ OA, Vol. XXXVIII, 204-7) The content of both sorts of ideas is the same. But how is their content related to the idea of God, which also contains the property of having knowledge and will? Arnauld's position seems to be that our ideas of God's knowing and willing are abstract versions of our ideas of human knowing and willing. God's cognition, like that of human beings, can be said to consist of perceptions in which things exist objectively. (Ideas, 66-67~ OA, Vol. XXXVIII, 145-46) Again, within God, as within human beings, Arnauld distinguishes between willing to permit X to happen, and willing that X happen?7 However, knowing and willing can correctly be attributed to God only if we abstract from everything in human knowing and willing that implies finitude or imperfection. As Arnauld's letter to Leibniz, cited just above, indicates, this means that when God has cognition of a contingent occurrence outside of himself, this does not involve an act of cognition within God distinct from his essence. Again, when God efficaciously wills that something occur in the created order, he does not bring about the effect by eliciting an act of volition distinct from his essence. Arnauld thought that Malebranche fell into confusion about God's knowledge and volition, that he attempted to exceed the limits to theological knowledge set by our idea of God, and in the process fell into a grossly anthropomorphic account of God's knowledge and volition. This confusion, in Arnauld's view, contributed mightily to the mistakes in his theodicy. My hypothesis is that he thought that the careful reader of Ideas would be forearmed against this confusion and these mistakes by gaining a firm grasp on several important principles: (1) that any object of our cognition is represented to us by our own perceptions (that is, by our own ideas); (2) that it is by reflecting on our own ideas that we should attempt to discover the modal connections among the properties of things~ and (3) that if we reflect on the idea of God we will find that although it contains the properties of omniscience and having an all-powerful will, these properties are contained there only in an abstract way. Before turning to Arnauld's attack on Malebranche's theodicy, let me point out that Ideas resembles the Logic in that what it has to say about the nature of ideas is quite thin. Although Arnauld returns, in Ideas, to the notion of object-

Arnauld on Ideas as a Topic in Logic

73

ive existence, he says very little about its relation to other sorts of existence. Again, he does not discuss the relation between his claim that objects exist objectively in the nlind, and the claim that the properties of objects exist objectively in the mind. Similarly, the entire work assumes some sort of mindbody dualism, but he does not explain his dualism or offer any argument in its favor. These are a few of the points one might have expected to see discussed if one assumed that Arnauld's purpose was to develop a metaphysical or epistemological theory. But he does not take them up. Nor does he argue extensively in favor of his own chief claims about the nature of ideas. Instead he says that they should be obvious to anyone who takes the trouble to reflect on his own mind. His attitude is intelligible if we see him as assembling points about the nature of ideas which he thinks will help his readers avoid mistakes in thinking about other topics, but not if we see him as attempting to develop a systematic position on a theoretical problem of metaphysics or epistemology. III: Arnauld's Critique of Malebranche's Theodicy of Nature and Grace28 The main purpose of Malebranche's theodicy is to reconcile the proposition that not all men are saved, with the proposition, "God wills that all men be saved." (First Epistle to Timothy, 2,4)29 There were at least two such reconciliations already in the field: The Augustinian view that "all men" in the text from I Timothy means "men of all nations and races," the point of the text being to emphasize the universalist view of Christianity as against Judaisnl;30 and the Thomistic position that the text refers only to God's "antecedent will" as contrasted with his "consequent will.,,31 The Augustinian position had been taken up by Jansen, and defended by Arnauld. 32 Malebranche rejected it out of hand. 33 He accepted something like the Thomistic position, but wanted to show that the will whereby God wills that all men be saved was a "sincere will," and not, as Aquinas would have it, a nlere "velleity.,,34 To this end, Malebranche thought he needed, and could provide, an explanation of why God did not save all men, even though he sincerely wanted to. The core of Malebranche's explanation, as Arnauld sees it,35 is a claim about God's volitions: that with the rare exception of miracles, God's wisdom pernlits him to act with regard to creatures only by general volitions and not by particular volitions. This means, to begin with, that everything that happens in the created world is covered by a general law. (In the material world everything that happens is covered by the basic laws of nlotion, of which Malebranche thought there were just two.) But the principle that God acts only by general, and not by particular volitions also nleans, at least according to Arnauld, that the content of God's volition-what God wills-·does not by itself determine any particular effect of the volition. Arnauld expounds the point as follows: We can say, according to the new system, that [in the material world] God does not will anything properly, positively and directly, except that all motion takes place or tends to take place in a

74

ELMARJ. KREMER straight line, and that in collisions the motion is communicated (les mouvements se communiquent) in proportion to the colliding bodies. 6 (Traite,243i

We can say that God wills the particular effects that take place in the created world-but only indirectly. That is, he has general volitions .like those just mentioned, and these issue in particular effects upon the occurrence of antecedent, created occasional causes. Thus Malebranche says, "In order that the general cause act by laws or general volitions, and that his action be regular [reglee], constant and uniform, it is absolutely necessary that there be some occasional cause which determines the efficacy of the laws..." (Traite, 67)37 All disorders and evils in the natural order and in the order of grace are the direct responsibility of occasional causes, and can be attributed only improperly and indirectly to God. Applying these principles to the distribution of grace, Malebranche argues that the occasional cause of the distribution of grace is the intercession of Jesus Christ, acting in his human nature. So the general volition by which God distributes grace is the volition that people are to receive exactly such grace as Jesus, acting in his human nature, requests for them. The fact that not all people receive the grace that would bring about their salvation is explained by the fact that Jesus requests grace for people only on the basis of his limited knowledge of their inner lives and their need for grace. 38 Arnauld thought that the views about Jesus to which Malebranche was led by his "system" were heretical, and that he had been led into them by a mistaken philosophy: Recall the time when you had not fallen into that new philosophy, that the universal cause ought not act by particular volitions, either in the order of nature or in the order of grace. Judge for yourself what you would have said at that time about the strange thoughts into which you have been thrown by that false principle, if someone else had proposed them to you.. .! limit myself to two...

One is that the soul of Jesus Christ, although personally united to the Word, is so little dependent on it that it has no power other than that of an occasional cause, and that it exercises that power only by...desires that it has from itself, without the Word forming any of them in it and determining it to have them. The other is, that the same soul, although wholly united to the eternal wisdom, is so little enlightened thereby that it does not know the secrets of hearts, no matter how much it needs to know them, according to you, in order to act wisely in the distribution of grace. (Reflexions, 844-45)39 This is not the place to go into the many and varied arguments that Arnauld mounted against both the philosophical and theological parts of Malebranche' s position. But in addition to these arguments, he gives an account of the genesis of Malebranche's position. Part of this account is contained in the above passage: that Malebranche was led into novel theological views he would otherwise have rejected,40 by the philosophical principle mentioned at the

Arnauld on Ideas as a Topic in Logic 75 beginning of the passage. The rest of the account is that Malebranche espoused the philosophical principle on the basis of no more than "comparisons [of God] \vith men, which on the one hand prove quite the contrary of what he claims, and which on the other cannot prove anything about God~ and popular thoughts, unworthy, it seems to me, of a philosopher, which reduce the vast extent of the plans of God...to a single end." (Reflexions, 188) Arnauld singles out, in the First Discourse of Malebranche's Traite, five "reasons" for the philosophical principle that is basic to his entire system. The first two are comparisons of God's actions with those of human beings, which Arnauld thinks are improper because they attribute to God the imperfections of human ways of acting~ the other three are popular notions that ignore the complexity of the ends for which God brings things about in the created order. The first reason begins with the observation that "An excellent workman proportions his action to his work: he does not accomplish in complex ways what he could carry out by simpler ways~ he never acts without a goal, and he never expends useless efforts." It continues with the claim that "God finds in the infinite treasures of his wisdom an infinity of possible worlds...which he could establish," and concludes that God "is determined to create the one that could be produced and conserved by the simplest laws, or which would be the most perfect in relation to the simplicity of the ways needed for its production and conservation." (Quoted by Arnauld, in Reflexions, 188, from Traite, Premier Discours, # 13.) Arnauld says that this does not establish Malebranche's principle for two reasons: First, even for a human workman, wisdom dictates that he first decide on what he wants to make and then consider the simplest way to accomplish it. Second, the only "vay that consideration of the means to achieve a goal enters into the choice of a goal by a wise human workman is that such a workman limits his goals to those he has the means to carry it out. But no such consideration could affect God's choice. Hence, says Arnauld, "these comparisons of earthly workmen with the infinite and all powerful Workman prove nothing..." (Reflexions, 190) The second reason is "once again a comparison with men." (Reflexions, 190) It begins with the claim that someone who first makes something and then destroys it, thereby shows a lack of intelligence or of constancy: "He who builds a house and then tears down a wing in order to rebuild it, shows his ignorance~ he who plants a vine and tears· it out as soon as it has taken root, shows his fickleness (legerete)~ because he who wills and then no longer wills lacks either intelligence or strength of mind." But God cannot act "by caprice or ignorance." Hence, when fruit that is nearly ripe is destroyed by hail, God does not bring this about "by particular volitions," but rather by general volitions or general laws. (Quoted from Traite, Premier Discours, #19, at Reflexions, 190.) Once again, Arnauld says that what Malebranche claims does not hold true in general even for human beings, who often enough build something for temporary use and then tear it down, without thereby showing

76 ELMARJ. KREMER any deficiency in intelligence or any fickleness. But it is even more obviously false in the case of God. For if a fruit is killed by hail before it is ripe, then evidently God, by one particular volition, willed that the fruit should attain a particular degree of maturity, and, by another particular volition, willed that it should killed by hail, and there is not the slightest indication of contradiction here, no sign that God "wills and then no longer wills." The third reason begins \vith the claim that each seed of a plant contains in miniature the full blown plant, and hence that it is clear that God wills in a very true sense (veut en un sens tres-veritable) that every seed should develop into a mature plant. 41 Yet many seeds are prevented from developing by "hail or some other untoward accident which is a necessary consequence· of the laws of nature." (Quoted from Traite, Premier Discours, #23, at Rejlexions, 197.) This would never happen, says Malebranche, if "God were not a general cause who must not act by particular volitions." His thought seems to be that the volitions by which God actually causes things to happen in the created order, what he elsewhere calls God's "practical volitions," are a subset of God's "true volitions,"42 and that God's true volitions can be frustrated by the laws of nature only if God's practical volitions are necessarily general and not particular. 43 Arnauld's basic response to this argument is that there is no reason to think that God truly wants every seed to develop into a mature plant. On the contrary, God evidently makes seeds to serve a variety of purposes, including that of nourishing animals. Thus Malebranche's argument is nothing but a popular consideration, which ignores the multiplicity of ends that may be served by something God brings about. The fourth reason begins with the claim that God could without doubt have made the world more perfect than it is; he could have done this, for example, by making the rain fall more often on ploughed fields and less often in the sea. But in order to make the world more perfect, Malebranche concludes, God would have had to change "the simplicity of his ways," that is, he would either have had to act by particular volitions or by a larger number of general volitions. 44 To this Arnauld responds in part as he had to the third reason: "Rain is useful for many other things than to make the earth fruitful." But he adds, "We do not even need that reply, because there is an infinity of things, as the Author must agree, that God has nlade...without our being able to discern his plan in making them." (Rejlexions, 201) The fifth and last reason has to do with "monsters," that is, malformed animals. Malebranche claims that God, "in a true sense would like [souhaite] all of his creatures to be perfect. ..he does not love monsters." Hence "if he could,by ways just as simple, make and conserve a more perfect world, he would not have established laws of which so many. monsters are necessary consequences." (Quoted from Traite, Premier Discours, #22, at Rejlexions, 202.) By "ways just as simple," he seems to mean "by general volitions that are just as few in number." It is strange, says Arnauld, that Malebranche did not -------------- -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

_

Arnauld on Ideas as a Topic in Logic 77 realize that he undercuts his own argument by his very next statement: "But it would have been unworthy of his wisdom, to multiply his volitions in order to avoid certain particular deformities, WHICH INDEED PRODUCE A SORT OF BEAUTY IN THE UNIVERSE." (Arnauld's emphasis) Arnauld's point is the Augustinian one, that the beauty and good order of the world as a whole is a sufficient explanation of the existence of monsters and other things that we find hideous. Hence the existence of monsters is no evidence for Malebranche's claim that God must produce the world by a small number of general volitions or general laws. Arnauld's criticism of the five "reasons" is not so much a cool and respectful refutation of an argument as an attempt to show that Malebranche's "new system" is based on a mass of confusion. Clearly Arnauld thinks that Malebranche's theodicy is not just wrong, but wrongheaded. 45 My hypothesis is that Arnauld thought Malebranche fell into this confusion, though he knew enough to have avoided it, because he refused to limit his philosophical claims about God to those that he could square with the content of his own idea of God; and that he was led to transgress the limits of human theological knowledge in this way by his failure to recognize that God is represented or exhibited to him by his own finite idea of God. If he had recognized this, then, in Arnauld's view, he would have avoided the sort of crude anthropomorphic speculation about God that led him to his philosophical principle and beyond it to a novel and mistaken theology. I must admit that there is little direct textual evidence to support my hypothesis. But perhaps there is at least one bit of textual support. At the outset of On True and False Ideas, Arnauld ties the work together with his anticipated critique of Malebranche's theodicy as follows: Our friend warned us, in the second edition of his Treatise ofNature and Grace, that in order to understand it well it would be a propos to know what principles are set forth in the book The Search after Truth. He emphasized in particular his doctrine on the nature of ideas, i. e., his opinion that we see all things in God. Therefore I set about studying that doctrine, and after careful consideration found so little appearance of truth, to put it mildly, in everything that our friend teaches on the subj ect, that I thought I could do no better than to begin by showing him that there is more reason than he thinks to distrust the heap of speculations which seemed certain to him, so that by this lesson he would be disposed to seek an understanding of the mysteries of grace in the light of the saints rather than in his own thoughts. (Ideas, 2; OA, Vol. XXXVllI, 180)

The "heap of speculations" to which Arnauld here refers is, of course, Malebranche's account of ideas in Book III, Part II of the Search. We have now seen that Arnauld also thought that Malebranche's account of grace was a heap of misguided speculations. It is not, I think, a very great leap to the conclusion that Arnauld thought the heap of speculations about grace was a result of the

78 ELMAR J. KREMER heap of speculations about ideas, and that the way to avoid both was by developing a correct account of the nature of ideas. NOTES 1. Nicole aArnauld, 2 juin 1682. Reproduced by Andre Robinet in Correspondance acts et documents, 1638-1689, recueillis et presentes par Andre Robinet, volume xvm of Oeuvres Completes de Malebranche (Paris: 1. Vrin) p. 238. These Oeuvres were published between 1958 and 1969 under Robinet's direction, and will be cited hereinafter as OC. Volume xvrn will be cited as Documents. All translations in this paper are my own. 2. " ... son livre des vraies et des fausses idees n'ayant effectivement pas plus de rapport au Traite de la Nature et de la Grace que la guerre du Turc en Hongrie ..." (Troisieme lettre contre Defense de M..Arnauld, #5, OC, vol. VI) 3. Reponse au livre de Mr. Arnaud, Des Vrayes et des fausses idees, OC, vol. VI, p. 18. 4. Arnauld to Quesnel, 18 October, 1682, in Documents, OC, vol. XVIII, 241-42. 5. See Arnauld's impassioned conunents about the importance of the questions about ideas if we are to speak worthily of God, at the beginning of the eighth of his 1685 Lettres de Monsieur Arnauld au Reverend Pere Malebranche sur les idees generales, la grace & retendue intelligible, Oeuvres de Messire Antoine Arnauld, edited by G. Du Pac De Bellegarde and 1. Hautefage, 43 volumes (Paris-Lausanne: Sigismond D'Arnay et Cie, 1775-1783) vol. XXXIX, p. 119. This set will be cited hereinafter as OA. See also the list of Arnauld's main disagreements with Malebranche at the end of Quatres lettres de Monsieur Arnauld au Pere Malebranche, written in 1694, the year of Arnauld's death, OA, vol. XL, pp. 108-9. 6. Steven Nadler, Arnauld and the Cartesian Philosophy of Ideas (Princeton: Princeton University Press, 1989) pp. 182-83. Hereinafter cited as Nadler. 7. Nadler, p. 179. 8. Nadler, p. 101. 9. Thomas Reid, Essays on the Intellectual Powers ofMan , in Philosophical Works, with notes and supplementary discussion by Sir William Hamilton, 2 vols. (Hildesheim: Georg Olms Verlagsbuchhandlung, 1967) vol. I, p. 297. 10. The debate is summarized in Monte Cook, "Arnauld's Alleged Representationalism," Journal of the History of Philosophy, 12(1974) pp. 53-64, and more recently in Nadler, p. 104ff. 11. Perceptual Acquaintance from Descartes to Reid (Minneapolis: University of Minnesota Press, 1984) p. 221. 12. Nadler, p. 143. 13. Richard A. Watson, "Arnauld, Malebranche, and the Ontology of Idea," The Great Arnauld and Some of His Philosophical Correspondents, Elmar 1. Kremer, ed. (Toronto: University of Toronto Press, 1994) pp. 129-137. 14. Antoine Arnauld et Pierre Nicole, La Logique ou 1~rt de Penser, Edition critique par Pierre Clair et Fran90is Girbal, Seconde edition revue (Paris: 1. Vrin, 1993). All references to La Logique are to this edition and will be given in parentheses in the main text.

Arnauld on Ideas as a Topic in Logic

79

15. Ideas, like Chapter One of the Logic, also provides a discussion of the origin of ideas, but Arnauld makes it clear that this topic is less important for his attack on Malebranche than that of the nature of ideas. 16. Throughout this paper I quote Antoine Arnauld, On True and False Ideas, New Objections to Descartes's Meditations and Descartes's Replies, translated, with an introduction by Elrnar 1. Kremer (Lewiston: The Edwin Mellen Press, 1990). References to the page numbers of this translation, followed by those in volume 38 of the Oeuvres de Messire Antoine Arnauld will be given in parentheses in the main text. 17. Regarding the authorship of the Logic, I accept the judgment of the editors of the critical edition: "On ne sait pas d'une maniere certaine ce qui revient a Arnauld et a Nicole dans la composition de I 'Art de Penser." (Logic, note 1, p. 335) I shall assume that Arnauld shares the views expressed in the work. Indeed, he quotes the work as if it were his own in his Defense... contre la Reponse au Livre des vraies et des fausses Idees, OA, vol. XXXvm:, p. 367. (The historical and critical preface to volume XLI of OA misidentifies the location of this text.) I would add, however, that Nicole seems to have been responsible for the additions to the later editions of the work. This point is made by Racine, and confirmed by the editors OA. (Racine's comment is cited by Clair & Girbal in note 1 of their edition. See also OA, vol. 41, p. iv.) This may help to explain the fact that the doctrine about the nature of ideas emphasized by Arnauld in Ideas-that an idea is a cognitive act considered as that in which the thing known exists objectively-does not appear in the Logic. Not only was the material added to the Logic in 1683 most likely the work of Nicole rather than Arnauld, but, as we have seen, Nicole had counselled Arnauld not to begin his attack on Malebranche by discussing the nature of ideas. It is not, then, entirely surprising that the doctrine Arnauld worked out for the first time in Ideas does not show up in the 1683 edition of the Logic. 18. Hence the fact that Arnauld elsewhere characterizes the questions dealt with in On True and False Ideas as "metaphysical" (note 4 above), does not count against rny proposal to consider it as a treatise in logic. Nor does the fact that Arnauld's attack on Malebranche's theodicy is primarily theological count against my hypothesis, for the authors of the Port-Royal Logic, from the very beginning, thought it would be useful in theology as well as philosophy. The point is underscored in the fifth edition, which appeared in 1683, the same year as Ideas: The "Avertissement sur cette nouvelle edition" points out although that the 1683 additions have to do largely with theological difficulties raised by certain Protestant Ministers regarding the eucharist, "they are no less proper and natural to Logic." (p. 12) 19. See Malebranche, Reponse au livre de Mr. Arnauld, Des vrayes et des fausses idees, in OC, vol. V, p. 50, and Arnauld, Defense... contre la Reponse au Livre des vraies & des fausses Idees, OA, vol. XXXVIII, p. 381 ff. Nadler (p. 82) comments on this point. 20. See for example the thirteenth paragraph of the Third Meditation, which includes an interplay of "repraesento" and "exhibeo." Here is a literal translation: To be sure, in so far as those ideas are only certain modes of thinking, I do not recognize any inequality among them and they all seem to proceed from me in the same way; but, in so far as one represents (repraesentat) one thing, and another another it seems that they are quite different from one another. For without doubt those which exhibit (exhibent) substance to me are something greater, and, so to speak, contain in then1selves more objective

80

ELMAR J. KREMER

reality, than those which represent (repraesentant) only modes, or accidents; and on the other hand that [idea] through which I understand a supreme God, eternal, infinite, omniscient, omnipotent, and creator of all things outside of himself, has more objective reality in itself, that those through which finite substances are exhibited (exhibenter)." (Adam & Tannery, vol. Vll, p. 40. Emphasis mine). The two occurrences of "exhibeo" as well as the two of "repraesento" were translated by "representer" in the 1647 French translation by the Duc de Luynes. Similarly, the new English translation by Cottingham, et al., like that of Haldane & Ross, translates both "exhibeo" and "repraesento" by the English "to represent." Michelle Beyssade, in her recent French translation of the original Latin text, translates "repraesento" by "representer" and "exhibeo" by "donner a voir" (Descartes, Meditations metaphysiques, Texte Latin et traduction du Duc du Luynes; Presentation et traduction de Michelle Beyssade (Librairie Generale Francaise, 1990) pp. 100-3). It may be wondered whether Descartes is using the two tenns to draw a technical distinction, whether "repraesento" goes with accidents and "exhibeo" with substance. But other texts in the Meditations seem to rule this out. I take it that Descartes is using the two different words to express the same concept. 21. Cf. Ideas, p. 21; OA, vol. XXXVIII, p. 200. 22. In order to discover how the properties are related we must make sure not to confuse one idea with another or one property with another in a single idea. It seems to be Arnauld's view that confused ideas are always the product of precipitous judgments. On this point see Alan Nelson, "The Falsity in Sensory Ideas: Descartes and Arnauld," in: Interpreting Arnauld, Elmar 1. Kremer, ed., (Toronto: University of Toronto Press, 1996). 23. Arnauld to Leibniz, 13 May 1686, in Leibniz, Discours de Metaphysique et Correspondance avec Arnauld, edited by Georges Le Roy (Paris: 1. Vrin, 1957) p. 99. The Amauld-Leibniz correspondence took place while Arnauld was in the midst of his controversy with Malebranche, and there are many echoes of that controversy in this letter to Leibniz. 24. " .. .in this life the mind knows God only abstractly..." Humanae Libertatis Notio, in Causa Arnaldina, edited by P. Quesnel (1699, apud Hoyoux, Leodici Eburonium) p. 104. A French translation of this work by Quesnel is contained in OA, 10, and the above quotations found on p. 618. 25. This passage occurs one page earlier than that quoted at note 23 above. 26. "I know myself then, in knowing other things." (Ideas, p. 6; OA, vol. XXXVIII, p. 184.) 27. See for example, OA, XXXIX, p. 281. 28. Arnauld criticizes Malebranche's theodicy primarily in the Reflexions philosophiques et theologiques sur Ie nouveau systeme de la nature et de la grace, published in two instalments, in 1685 and 1686, and I shall be concerned mainly with this work. But the picture is in fact more con1plicated on both sides. The work in which Malebranche first developed his theodicy, and which Arnauld pron1ised to criticize, was the Traite de fa nature et de la grace, first published in 1680. But Malebranche had anticipated some of the themes of that work in the fifteenth Eclaircissement to the Recherche de la verite, published in the third edition, in 1678, and expounded his theodicy again in the Meditations Chretiennes et metaphysiques, first published in

Arnauld on Ideas as a Topic in Logic

81

1683. Arnauld, on the other hand, began his direct criticism of this theodicy in 1685 with his Dissertation sur les miracles de I'ancienne loi, a work followed shortly by the first volume of the Reflexions. The Reflexions are found in volume XXXIX of OA, and will be cited hereafter by page numbers given in parentheses in the main text. 29. I translate Malebranche's French, "Dieu veut que tous les hommes soient sauvez." See Malebranche, Traite de la Nature et de la grace, Troisieme Eclaircissement, #21, OC, vol. V. All further references to the Traite are to this edition and will be given in the main text in parentheses. 30. This interpretation receives some support from the fact that Paul goes on to say, "For this I was appointed a preacher and apostle...a teacher of the Gentiles in faith and truth." (I Timothy, 2, 7) 31. On these two traditions see Jean Laporte, La Doctrine de Port-Royal, Les Verites sur la grace (Paris: Vrin, 1923) pp. 250-51. 32. The fifth of the five condemned Jansenist propositions is "It is Semipelagian to say that Christ died or shed his blood for each and every man [Christum pro omnibus omnino hominibus mortuum esse aut sanguinem fudisse]." The text can be found in Hem. Denzinger & Carolus Rahner, Enchiridion Symbolorum (Rome: Herder, 1960) p. 361. 33. The introduction to the edition of the Traite in volume V of OC reproduces a passage from Malebranche's biographer, P. Andre, in which a letter is quoted as follows: "If people in this century did not so obstinately maintain that God does not have a sincere will to save all men, it would not be so necessary to establish the principles which serve to destroy that unhappy opinion.. .I protest before God that that was the principal motive that made me [qui m'a presse] to write" [the Traite] (p. xliv). 34. See Arnauld's comments on this point in Reflexions, p. 198. 35. I set aside the question of whether Arnauld's account of Malebranche's position is fair and accurate. On this question, see two long and excellent introductions to the Traite: one by Ginette Dreyfus, to a presentation of the original, 1680, version of the Traite (Paris, 1958)-not to be confused with volume V of the Ouevres Completes de Malebranche, also edited by Drefus-and one by Patrick Riley, to his English translation, Treatise on Nature and Grace (Oxford: Clarendon Press, 1992). 36. The reference to willing "properly, positively and directly" is taken from Meditations Chretiennes, VII, which can be found in OC, volume X. 37. Cf. Traite, Premier Eclaircissement, #4, "When we see that an effect is produced immediately after the action of an occasional cause, we should judge that the effect was produced by the efficacy of a general will." (p. 149) 38. See the brief "abrege" of Malebranche's "new system regarding grace" given by Arnauld in Reflexions, pp. 673-74. 39. Arnauld held to the traditional view that Jesus, in his human nature, had habitual knowledge of everything known to God. He cites Aquinas, Summa Theologiae, Part ill, Question 9. 40. Arnauld says repeatedly that Malebranche violates his own principle, "that in matters of theology, novelty is a sign of error, and it is right to discount [mepriser] opinions for the sole reason that they are new and without foundation in the tradition." (Reflexions, p. 171) In the very title of the work, and repeatedly in the text, he refers to Malebranche's position on nature and grace as "Ie nouveau systeme," and the complaint rehearsed at the end of each of the three books into which the work is divided. (Reflexions, pp. 414, 643, 848)

82

ELMAR 1. KREMER

41. Malebranche, like Leibniz, held the "homunculus" theory of plant and animal reproduction. 42. For the expression "practical volitions," see the Troisieme Eclaircissement to the Traite, # 5 and 6, pp. 179-80. 43. Arnauld makes some interesting conunents about what Malebranche might have meant by "willing in a very true sense," comparing Malebranche's terminology to the Thomistic distinction between God's antecedent and consequent will. 44. The fourth "reason" (Reflexions, p. 199) is taken from the Traite, Premier Discours, # 14. But the source of the quoted matter is not identified by Arnauld, who says, instead, "Voyez 1. Discours, Partie I. art. 2. art. 38 & dans l'Eclaircissement art. 7. 15.16.17." 45. Arnauld's account of Malebranche's five "reasons" resembles Chapter IV of On True and False Ideas, in which he tries to show "that what the author of The Search after Truth says about the nature of ideas.. .is based on fantasies which we retain from the prejudices of childhood." (12~ OA, vol. XXXVllI, p. 190)

Francis Hutcheson's Logicae Compendium and the Glasgolv School of Logic l EMILY MICHAEL

1. Introduction I wish to show that a distinctive epistemological logic, a particular strain of a modern logic tradition, here called the logic of ideas, took root at the University of Glasgow during the late seventeenth century, culminating in Francis Hutcheson's popular Logicae Compendium. 2 Hutcheson's logic text was published posthumously in 1756, completing thereby a set of textbooks by Hutcheson for two years of the arts curriculum regularly taught at the Scottish universities. Hutcheson's logic text was used in the logic course of Glasgow professor James Clow for the next thirty years,3 and after Clow, it was used by his successor as professor of logic, George Jardin, at least during the early years of his tenure. Hutcheson's text, thereby, shaped and crystallized, with its publication, the distinctive epistemological approach to logic of the school of Glasgow that I have characterized as a strain of the modern logic of ideas. In the early seventeenth century, the curriculum at the Scottish Universities (i. e., at the Universities of Glasgow, Edinburgh, Aberdeen, and St.Andrews) was a four-year course: humanities (Latin and Greek) in the first year~ and three years of philosophy courses, logic, metaphysics, moral philosophy, and natural philosophy.4 Human cognition, during this period, was discussed in the natural philosophy course as a subject of psychology, i. e., the science of the vital and cognitive operations of living things, a study shaped, in particular, by Aristotle's De Anima and Parva Naturalia. At this time, nletaphysics, concerned, in particular, with ontology, had Aristotle's Metaphysics as its chief source~ and logic, with Aristotle's Organon as its chief source, was viewed as a linguistic study of terms and their logical relations. The place of human cognition in the Scottish curriculum changed over the seventeenth century. 5 In the early eighteenth century, cognition remained, as previously, a subject in the natural philosophy course. But this course was viewed as the study of the nature of bodies (somatology), and examined the fundamental principles, structure, and activities of bodies, inanimate bodies and living things, and in this context considered the human soul's cognitive powers and the corporeal organs empowered by the soul. In addition, the metaphysics course changed its character with the addition to it of a study of the nature of spiritual substances (pneumatology), considering, in particular, God and the human mind~ metaphysics, thereby, examined the contents and activities of the human mind, and so too examined human cognition. Finally, the logic course also changed its character. The Scottish schools, influenced by humanistic and early modern developments, claimed a new interest in teaching students to think clearly and

84 EMILY MICHAEL write effectively. At the University of Glasgow, this promoted the development of a distinctive epistemological approach to logic. Logic was viewed as the study of the nature, origin, and logical relations of ideas, and so involved, once again, a study of human cognition. Hutcheson's logic, and with this, the Glaswegian strain of what is called above the logic of ideas developed in this context. The follo\ving is a study of the source and nature of Hutcheson's logic of ideas. But first, it nlight well be asked: what is a logic of ideas? At St. Andre\vs professor William Barron, in his lectures on logic which were published posthumously in 1806,6 develops what is here called a logic of ideas. Barron explains: As the object of logic is to teach the best use of our rational faculty, both in investigating and in communicating truth, the theory of it and the materials of which that theory consists, are deduced from this end. That theory, accordingly, consists of two parts; the nature of ideas, which are the materials on which we reason, and the nature of the faculties or operations of the mind which are concerned in reasoning. Before we can reason, we must have ideas, and before we can reason right, we must understand what kind of operation reasoning is. The explication of both comprehends the whole of logic, which is of any use. (B 367)

Barron is committed to the view that the whole of logic consists in the explication of "the nature of ideas, and the operations employed about them." (B 367) He claims three operations are relevant to the investigation of truth, viz., perception by which ideas are obtained, judgement of in particular "the agreement or disagreement of any two ideas," and reasoning, the comparing of two ideas by means of one or more intermediate ideas. Here we can identify as a fundamental thesis indicative of a logic of ideas: Tl: Logic consists in the explication of the nature of ideas and the operations fundamental to acquiring and effectively employing them. The goal of this logic is the practical one of directing thought to acquire knowledge, and this is seen as involving a study of the elements and the nature of reasoned thought, in particular, of its ideas and intellective operations. Barron's late eighteenth century analysis is clearly and quite exclusively influenced by Locke's theory of ideas. But this is not true of Hutcheson's account; here Locke's imprint is unquestionable, but there are non-Lockean elements as well. Comnlentators have commonly but mistakenly assumed that Hutcheson's theory of perception is that of Locke, and this has led to confusion in interpretation of his moral and aesthetic sense theories. An aim of the following study is to explain the rationale and nature of non-Lockean elements in Hutcheson's Logicae Compendium. In what follows, I will examine the original source of the logic of ideas, and the proximate source and distinctive nature of Hutcheson's logic!~~~_£~3J~}_

Hutcheson's Logicae Compendium and the Glasgow School 85 will illustrate Hutcheson's divergence from Locke's account of ideas by considering several representative issues.

2. The Origin of the Logic of Ideas Henry Aldrich, whose Artis Logicae Compendium was published in 1691,7 shows no knowledge of Locke's recently published Essay On Human Understanding. But Aldrich, in some concluding remarks about logic traditions, vetus et nova, distinguishes, based he says on Gassendi's De Origine & Varietate Logicae, a new and recent logic tradition from the more traditional approach that Aldrich himself takes (and that he claims to share with, for example, Sanderson and Keckermann).8 Candidates for progenitor of this recent logic tradition are Descartes and Bacon, discussed by Gassendi, and Gassendi himself. But Descartes, Aldrich states, has no logic~ and Bacon deals exclusively with experimental methods, not logic proper. This leaves Gassendi's Institutio Logica as the source of the recent modern logic, and Aldrich further explains that the subsequent Ars Cogitandi is the Cartesian version of this modern logic. 9 In both of these seminal texts, consistent with Tl, a logic is built upon a discussion of the nature of ideas ("the materials on which we reason"), and deals essentially with "the operations employed about them." Gassendi, seeking to explain the logic implicit in Epicurus' s account, maintains that, for Epicurus, sense ideas are the basis of all reasoning and therefore oflogic. 1o He thereby develops an empiricist logic, in which all ideas, indeed all thought, has its origin in sense experience. Gassendi defines logic as ars bene cogitandi, the art of thinking well,ll and he sees the goal of logic as the practical one of directing thought to acquire knowledge. For Gassendi, thinking well involves four skills: imagining well, that is, forming correct images or ideas of things~ posing propositions well, that is, advancing propositions that are correct~ inferring well; and ordering welt and accordingly, Gassendi's logic is divided into four parts, considering in turn, imagination, judgement, reasoning and method. In the Port Royal Logic, which appeared four years after Gassendi's Institutio Logicae, logic was defined as the art of thinking. The Port Royal Logic has the same structure as Gassendi's logic, dividing logic into essentially the same four parts, but the theory of ideas in this logic is Cartesian and antiempiricist. As a result of the influence of the Port Royal Logic, and to a lesser extent, of Gassendi's logic as well, a new genre of logic emerged, one in which the theory of ideas was of central concern. Accordingly, when Locke's Essay appeared, it was seen by himself and others as a contribution to logic. Locke's Essay was built on the theory of ideas in Gassendi's logic, and in Hutcheson's logic text we find a logic of this genre. 12 We turn now to the proximate source of Hutcheson's Logicae Compendium.

86

EMILY MICHAEL

3. The Context and Nature of Hutcheson's Logicae Compendium

The late seventeenth century was an unsettled time in the history of logic, and we find in Scotland the variety of logics noted by Aldrich, traditional Aristotelian and humanist logics and syncretist combinations of these, along with the new logic of ideas. I3 A dramatic turning point in Scottish university education occurred in 1695, when a Parliamentary Commission was appointed by the new Presbyterian government to investigate the curriculum of the Scottish universities and to make recommendations for their improvemen1. 14 The Commission mandated that Scottish regents compose common philosophy texts for the four Scottish universities, an aim of which, our evidence suggests, was to guarantee that teachers did not lead their very young charges astray.IS S1. Andrews was responsible for the logic text,16 and an epistemological approach to logic introduced in the St. Andrews text culminated in the logic of ideas of Hutcheson's influential Logicae Compendium. In both works, as was usual in logics of ideas, we find, to explain the logical relations among ideas, such traditional topics as predicables, contrary and contradictory propositions, the figures, moods and basic rules of syllogistic. But in these two Scottish works, we also find a comnl0n, distinctive account of ideas, as will be shown below. Here Hutcheson does not present such original Hutchesonian features as reflex sensation and concomitant ideas found in the discussion of ideas in his metaphysics compend. 17 His account of ideas in his logic compend instead closely corresponds, in structure, content, even terminology, to that of the S1. Andrews text. But the evidence seems to indicate that the original manuscript for the St. Andrews text, a handwritten work, currently in the University of Edinburgh archive collection, was circulated to the Scottish universities at the time,18 but that it was never published, and never actually used in teaching. But then how did the S1. Andrews account of ideas come to Hutcheson, who in 1695 was one year old? In fact, the account of ideas found in the S1. Andrews logic text is found at the University of Glasgow in lectures of John Loudon,19 who taught at St. Andrews when their logic text was written. The common view among contemporary commentators is that Hutcheson was Gershom Carmichael's student,20 but I have argued in a recent paper21 that there is considerable and convincing evidence that Francis Hutcheson was John Loudon's student in logic at the University of Glasgow in 1711-1712, and Loudon's logic lectures of that year are available as a handwritten logic compend. 22 The epistemological logic here presented is strikingly similar to the S1. Andrews logic text and, in turn, is even more similar, e. g., in its terminology, to Hutcheson's logic text. 23 From extant notes, it seems probable that Loudon's logic lectures changed little throughout his fifty-one years of teaching at Glasgow (1699-1750), and I believe that Hutcheson, in his logic, builds on

Hutcheson's Logicae Compendium and the Glasgow School 87 this existing precedent. But why did Hutcheson write this derivative Latin logic text? When the Foulis brothers, Hutcheson's proteges, started a printing firm in Glasgow in 1741, "the state of printing was low...in Scotland,,24 and textbooks were largely unavailable. 25 Hutcheson's Latin moral and metaphysics compends, published in 1742, filled a gap in these subjects, and a year later the Foulis brothers were appointed printers for the University of Glasgow. Perhaps this motivated Hutcheson to compose a logic book to thereby provide the sorely needed texts for two complete years of Glasgow university courses as then taught. 26 Texts were later composed for the other two years (for the beginning year on Humanities, which was the study of Latin and sonle Greek grammar and literature, and for the final year of natural philosophy), and all these texts saw numerous editions. His logic text enabled Hutcheson to foster the continued learning of Latin, one of his stated goals, and also to propagate the distinctive logic of ideas he was taught in his own student days at Glasgow. But why was the latter important to him? I would suggest that the St. Andrews regents and Loudon had, in their logics, an agenda of their own, that was shared by Hutcheson. The distinctive role of Scottish logic is laid out by 8t. Andrews regents in 1695: Man's greatest concern is that of his beatitude, but it is easy to err; it is in our greatest interest to attend diligently so as to observe the causes by which we are seduced into error and to find the remedies of error. This is presented in fine works of ingenious philosophers, who have collected observations and precepts, which can at all moments help us; these are called the observations and rules of a system of logic. 27

The logic of ideas was perceived as a way of learning to think clearly and avoid error, a principal aim of which, among the Scottish Calvinists, was salvation; and the 1695 Commission was concerned to sanitize Scottish education with this aim in mind?8 But, as Locke's new logic was seized upon as a useful analysis of ideas, it was also attacked as way of seducing the unsuspecting into error. Locke separated logic, i. e., the investigation of ideas simpliciter, from psychology and nletaphysics, from considerations about the nature of the mind; but critics of Locke complained that this leads Locke to omit consideration of pure activities of the human mind, activities that do not depend upon the body and that are essential for the mind's immortality. That this consideration was important to Hutcheson is clearly indicated in the arguments for human immortality in his· metaphysics compend, a text based on his own course.29 Locke's approach to ideas can lead to the view that all ideas depend upon corporeal activities, that bodies think and that minds are corporeal and mortal. To avoid this problenl, the Scots considered, in their logic compends, the nature of the faculties of the mind and the mind's purely intellective activities, considerations only lightly touched upon by the Port Royal Logicians, treated

88 EMILY MICHAEL by Gassendi only in his natural philosophy, and not considered by Locke at all. All three Scottish logics, in their discussion of simple apprehension, distinguish the powers of sensation and imagination, which depend upon the body, and the mind's purely intellective powers, which employ no corporeal organ and permit an immortal nlind, correcting thereby this Lockean omission. 30 The introduction into logic of these psychological considerations was a further step in the gradual process of developing, starting from Gassendi' s initial logic of ideas, a more thoroughgoing epistemological logic. In what follovvs, I will examine the common account, in the three Scottish logic compends, of the corporeal and the purely intellective faculties of human cognition. In conclusion, I seek to show that this common Scottish approach inspired differences, on several basic points, in Hutcheson's and Locke's accounts. 4. Corporeal and Purely Intellective Faculties of Cognition All three Scottish works claim, as Barron later does, that logic has the practical role of directing the operations of the mind in acquiring knowledge or investigating truth. All take apprehension, judgement, and discursus to be the mental operations that logic directs. Hutcheson adds an alternative distinction between apprehension and two kinds of judgement, noetic, comparing two ideas, and dianoetic, comparing two ideas by means of a third, as does Loudon. His text is divided into four nlajor sections, entitled On Apprehension, On Noetic Judgement, On Discursus (which is dianoetic judgement), and, finally, On Method, added, as by Loudon, as a appendix?l All three texts describe judgement as passing sentence (ferens sententiam) on two ideas compared with each other,32 and all three, after laying out like rules for deduction, examine seriatim dilemma, enthymeme, induction, example, sorites, and epichirema. We will focus upon the first and fundamental cognitive activity, viz., apprehension, also called by all three, idea or perception. All three works identify apprehension as a simple and bare perception or representation of a thing, without any mental judgement, without affirming or denying anything of it. There are three sorts of ideas, three kinds of perceptions or apprehensions, those of sensation, imagination, and pure intellection. The former two are corporeal perceptions or corporeal ideas. The St. Andrews text explains the first sort of corporeal idea as "a perception of a material thing caused by the occasion of a change which a thing present at this time induces in our body, e. g., as when...a part of the body is moved or agitated, the mind, by that motion, then perceives a thought which is said to be a sensation..." (St. A 25) In the words of both Loudon and Hutcheson, "the perception of a corporeal (material) thing affecting corporeal organs" is a sensation. (L 1.1; H I.I.i) A sensation is corporeal in three ways: 1) A present corporeal object is the exciting cause of a sensation, which itself being twofold, consists of 2) a corporeal impression; and 3) a corporeal image.

Hutcheson's Logicae Compendium and the Glasgow School 89 The view of all three Scottish logics is that a present corporeal object raises a sense idea by affecting, in tum, an external sense organ, nerves and brain. This physical event results in a brain trace, a corporeal impression, and a concomitant sense idea, which is itself a representation or image of the external exciting cause. The St. Andrews text explains the second sort of corporeal idea or perception: "The perception of a material thing, which image is said be an image remaining in the brain, as distinguished from sensation, as e. g., when at night we think of the sun~ when we represent a man, a house, a clock first seen by our senses, now not affecting them, this thought is said to be imagination." . (St. A 25-26) And Loudon and Hutcheson both say, in the same words, "the idea of a corporeal (material) thing while it affects no organ" is imagination. (L 1.1 ~ H Ll.i) Loudon elaborates: "If we are said to perceive a corporeal thing, let us say a cube, and we understand this to be produced by an act, where this thing affects no sense organ...and makes no impression on the body, but if its origin...will be the perception of a corporeal thing by sensation, of which this is an imitation, we say it is by imagination...as an absent friend." (L 1.1) This sort of idea, we are told, is corporeal as the perception of a corporeal object associated with a remaining brain trace~ as such, it is a weaker idea of a previous vivid sense perception. So Loudon, characterizing imagination, explains that "...a preceding sensation...will be a vivid impression in a corporeal organ, and, with the repetition of these same objects of sensation, there will be a/aint trace in the organ, which will be depicted in the brain as an image of the object. .." (L 1.1) And I-Iutcheson sinlilarly says, "Imagination recalls weak ideas of things previously perceived by sensation. Nor is the mind able to form images, except those where all the elements were perceived by sensation." (H Ll.ii) We might see here common ground with Hume's distinction between impressions and ideas. But unlike Locke and Hume, the three texts we are examining distinguish a noncorporeal sort of idea or perception. So the St. Andrews text explains: "Pure intellection is the perception of a thing which is without any material impression." (St.A 26) And Loudon similarly says: "Pure intellection is said to be perception of an object without simulacra, images or corporeal impressions~" (L 1.1) as does Hutcheson: "a pure intellection is 'an idea acquired or understood without the power of any corporeal sense'." (H Ll.i) All three claim that ideas of the brain nlust be imagistic. And these imagistic ideas are perceptions that are initially raised by and that represent particular present or absent external corporeal objects. But human beings also have ideas that are not imagistic and that represent no individual corporeal object, and such non-imagistic ideas cannot be associated with any particular physical impression or brain trace. These non-imagistic ideas or perceptions, Cartesians and Gassendists agree, are acts of an incorporeal entity, the human

EMILY MICHAEL mind itself, and, associated with no present external exciting cause, no brain trace, no image, these ideas are purely intellective. So Loudon, distinguishing two sorts of pure intellections, first, spiritual entities, and, second, all universals, says that "we can perceive...that which cannot be produced by the power of sensation or imagination, as God, the human mind, ideas of all universals~ all these nl0des of perception are said to be pure intellection." (L 1. 1) The St. Andrews text similarly explains that only by intellection, can "we conceive God, angels, the human mind, virtue, truth and so forth." (St. A 26) In these Scottish logic texts, pure intellection supplements a variety of conceptions, inherited from Locke, such as sensation and reflection, 'primary and secondary qualities, simple and complex ideas. I will illustrate this by considering Hutcheson's account of each of these. (1) Hutcheson on Sensation and Reflection: Hutcheson, as Locke, claims that all our ideas have their source in sensation and reflection, but Hutcheson distinguishes reflection of two sorts. First, he explains: "The sensation is also internal that provides purely intellectual perceptions; which power is called consciousness or reflection. This is aware of all actions, passions or modes of the soul; to wit, judging, reasoning, certainty, doubt, joy, sorrow, desire, aversion, love and hate, virtue, vice." (H I.I.iii) This reflective awareness of what my own mind does and undergoes, by means of which I directly apprehend my inferring, doubting, .desiring, fearing, hoping, loving, is like Locke's conception of reflection. But Hutcheson, unlike Locke, analyzes these internal sense perceptions as ideas, which, involving no corporeal impression, no corporeal image, no corporeal exciting cause, are purely intellectual. Second, Hutcheson also sees reflection as the mind's power to call up and to review ideas it already has. By internal sensation, the mind apprehends its activities, its passions, and its external and internal sense ideas. This latter perception, i. e., my internal sense awareness of my ideas, may, in turn, become an object of reflection~ and these internal sense ideas may be compared and contrasted, and other ideas may be fornled on this basis. This is a power by which we can form, in addition to ideas we get from external and internal sensation, ideas of these ideas, that is, higher order ideas, the universal concepts formed by abstraction. By external sensation, I can perceive blue and red, and by internal sensation I can perceive my perceiving; but only by abstractive reflection can I apprehend the universal ideas of color, of sensible quality, of vision, of sense perception, of mind, of spiritual entity. Here reflection, a distinctive power of the immaterial human intellect, is, as in Gassendi's view, the power by which we form abstract general ideas, all of which are purely intellective. Further, Hutcheson takes the view that from my corporeal and intellective ideas, I can, by inference, formulate additional ideas of pure intellection, e. g., ideas of theoretical entities and of such spiritual objects as God and angels. 90

Hutcheson's Logicae Compendium and the Glasgow School 91 Finally, in his metaphysics compend, and elsewhere, Hutcheson adds anothe sort of reflection, fundamental to the awareness of relational qualities that are the objects of aesthetic and moral perception, but consideration of this is beyond the scope of the present project. 33 (2) Hutcheson on Primary and Secondary Qualities: Hutcheson, as Locke, distinguishes, among ideas of external sensation, those of primary and of secondary qualities. He explains, as do the 81. Andrews text and Loudon, that the power in bodies to excite ideas of color, sound, odor, taste, heat and cold are called secondary qualities, each of which is the proper sensible of but one sense. Further, primary qualities are primary in the sense that they, in combination, "have the power of exciting ideas of secondary qualities, to which there is nothing similar in the body itself." (H I.l.i) Only primary qualities are real or true. But while Hutcheson says that primary qualities, extension, figure, situation, nl0tion and rest, number, and duration, are perceived by many senses, he also tells us that accurate ideas of primary qualities are not sensible ideas. He says: "Also referred to pure intellection are accurate and abstract ideas of primary qualities." (H I.1.iii) Accurate ideas of primary qualities have their source in sense experience, but they are in fact abstract ideas of pure intellection. This claim can be better understood through a passage from a 1727 letter of Hutcheson: "Figure and bounded color are not to me the same. Figure accompanies bounded color, but the same or perfectly like idea may arise by touch, without any idea of color, along with the ideas of hard, cold, smooth.,,34 Hutcheson here claims that the real quality of squareness is not like a visible square or a tactile square. We can form a more accurate idea of squareness common to both sensations. This common idea is an idea of pure intellection. Hutcheson, in his letter of 1727, employs this non-Lockean analysis of primary qualities to counter Locke's response to the Molyneux problem: Mssrs. Locke and Molyneux are both wrong about the cube and sphere proposed to a blind man restored to sight. He would not at first view know the sphere from a shaded plain surface by a view from above~ but a side view would discover the equal unifonn round relievo in one, and the cubic in the other. We can all by touch, with our eyes shut, judge what the visible extension of a body felt shall be when we shall open our eyes; but cannot by feeling judge what the colour shall be when we shall see it; which shews visible and tangible extension to be really the same idea, or to have one idea common, viz., the extension; though the purely tangible and visible perceptions are quite disparate. 35

Hutcheson therefore, unlike Locke, claims that no sensible ideas are adequate ideas of objective qualities, but we are able to abstract from them intellectual ideas that are a true or correct understanding of those objective qualities. (3) Hutcheson on Simple and Complex Ideas: Hutcheson first presents what Loudon calls the old notion of simple and complex. Loudon explains that an -------------------------------

92 EMILY MICHAEL apprehension is incomplex (or simple) when we perceive things, however man there may be, without any order or relation to one another, as when we conceive a glass and water. Apprehension is complex when we instead perceive them in relation to each other, as when we perceive water contained in a glass. Hutcheson similarly says that apprehension "is either inconlplex, as, calamus (pen), or complex as, calamus in manu (pen in hand)." (H 1.1.i) In Hutcheson's account oftlie new modern notion of simple and complex, a simple idea is unifornl, and not composed of dissimilar parts; a complex idea is composed of dissimilar parts into which it can be resolved. Here an experienced secondary quality, as, for example, the idea of blue or smooth, is said, as by Locke, to be simple. But, Hutcheson also claims: "The idea of being is the most simple." (H 1.2.iii) Loudon explains this claim further. A simple idea, he says, is analyzed by some as a single uniform perception of a thing divisible into no parts, and from this viewpoint such sensibilia as blue and smooth are simple ideas. But the preferable analysis is: a simple idea is one that does not involve other simpler ideas into which it can be resolved as if into parts. Loudon here distinguishes two senses of simple that are compounded in Hutcheson's account, viz., uniform; and not resolvable into parts. From Loudon's viewpoint, the ideas of sensation are not completely simple. Loudon says that in comparing a color and a taste, there is something in which they agree and something in which they differ, and this shows that each is resolvable into simpler ideas. So, for example, if \tve compare blue with sweet, we find these are alike as sensible qualities, but they also differ for one is a color, the other a savor. Blue is therefore resolvable into the simpler ideas of color and sensible quality. Loudon, as Hutcheson, here asserts that being, which is not further resolvable into simpler parts, is the simplest idea. Resolvability, as here understood by Loudon and Hutcheson, requires further explanation. Hutcheson explains that the complexity of an idea is relative to its conlprehension, which he defines as "the collection of all simple ideas conjoined in the complex, as in animal; body, living, sensitive." (H 1.3.i) The comprehension of idea X is the set of all simpler ideas into which X can be resolved as if into parts. 36 That is, it is all the ideas that can be abstracted fronl X, by comparison and contrast of X with other ideas. This presumes a particular view of abstraction as a matter of perceiving the sinlilarities and differences between ideas, and this view of abstraction is indeed explicitly proposed by both Loudon and Hutcheson. 37 From this viewpoint, once again a departure from Locke, such abstract ideas of pure intellection as being, power, thought, will always be simpler than any idea of a sensible quale. To conclude, Hutcheson's logic text introduces epistemological considerations in the first and foundational section on simple apprehension. That Hutcheson's moral and aesthetic theories were influential is well-known, but virtually unknown is Hutcheson's seminal role in the epistemological logic

Hutcheson's Logicae Compendium and the Glasgow School 93 tradition that developed during the Scottish enlightenment. Hutcheson's logic text continued and integrated a distinctive logic of ideas tradition at the University of Glasgow, and, further, was a principle instrument in the persistence of this tradition well into the eighteenth century. As such, it seems appropriate to conclude that Hutcheson's Logicae Compendium, though largely neglected, merits, as much as his influential aesthetic and moral theories, recognition and further attention. NOTES 1. Research for this paper was partially funded by a grant from the Research Foundation of the City University of New York. All translations are my own unless otherwise noted. 2. Francis Hutcheson's Logicae Compendium (Glasgow: R.& A. Foulis, 1756) was published by his son. (Hereafter noted as H). Hutcheson never taught logic at the University of Glasgow, but his name lent prestige to the posthumously published text. Why or when he wrote the text is not clear. 3. For a transcription of Clow's English lectures on Hutcheson's Latin logic book, see A System of Logic by James Clow A. M., Professor of Logic in the University of Glasgow, transcribed by John Campbell (Glasgow, 1773), University of Edinburgh Library (EUL). 4. A professorial system, in which each year of the curriculum was taught by a professor of Humanities, Logic, Moral Philosophy, and Natural Philosophy respectively, was introduced at Edinburgh in 1708, at Glasgow in 1727, at St. Andrews in 1747, at Aberdeen, Marischal College in 1753, and at Aberdeen, King's College in 1799. Before this, Scottish regents carried a single group of students through all three years of the philosophy classes. Most schools appointed professors of Humanities and Mathematics prior to the above dates. 5. Evidence of the teachings of Scottish regents is provided in published philosophical theses, written by the regent of the graduating class, and in manuscript copies of student lecture notes. Scottish regents of this period dictated lectures to their students. 6. William Barron, Lectures on Belles Lettres and Logic (London, 1806). (Hereafter noted as B) William Barron was professor at St. Andrews from 1778 till 1803. 7. He~ry Aldrich, Artis Logicae Compendium (Oxford, 1691). 8. Aldrich, (Artis Logicae Compendium, "Conclusio," p. 52) says: "Summam ejus exhibet Gassendus cap. 8. De Origine & Varietate Logicae..." He also here mentions "Keckermannus tract. 2. Praecog. Log. c. 2 & 39." See p. 54ff. for his discussion of the new logics of Gassendi and the Ars Cogitandi. 9. See in Collection of papers illustrative of the history and constitution of the University ofEdinburgh, 1611-1742 (2 vols., Dc. 1. 4 (2), EUL, p. 13) in remarks by regents of the University of Edinburgh, the mention of use of both the Ars Cogitandi and Gassendi. See, for example, note 16. 10. See Canon 7 of the Epicurean Canonic in Pierre Gassendi, Opera Omnia (Lyons, 1658), vol. I, p. 55. It begins: Est Anticipatio in omni Ratiocinatione principium. Anticipations (of perception) are interpreted by Gassendi as images. 11. Gassendi, vol. I, p. 91.

94

EMILY MICHAEL

12. For discussion of the influence of Gassendi's logic on Locke's theory of ideas, see E. & F. Michael, "The Theory of Ideas in Gassendi and Locke," Journal of the History ofIdeas, 51 (1990), pp. 379-399. 13. Indeed, logics of ideas appeared in Scotland as early as 1679 in lectures of, for example, Alexander Cockburn of St, Andrews, (Df. 9. 138 (3), EUL); and thereafter. Andrew Massie of Edinburgh (e. g., La. III. 154, EUL, 1680-81), and John Law of Glasgow (e. g., Dc. 8. 18, EUL, 1692). 14. Prior Commissions of the General Assembly of 1642, 1664, 1672, and 1683 had also recommended uniform printed courses, but these reconunendations had little influence until 1695. (See Evidence oral and documentary taken and received by the Commissioners appointed by His Majesty George IV, July 23, 1826.. for Visiting the Universities of Scotland, 4 vols. on Edinburgh, Glasgow, 8t. Andrews, and Aberdeen respectively'(London, 1837)). 15. See report (Evidence, vol. ill, p. 215) dated on April 24 th, 1695, "The which day, it being represented to the Committee for visiting the University of 8t. Andrews, that there are dangerous and pernicious tenets treated in the Dictates and Theses of some of the Masters of this University, which do much tend to Atheism, to the everting of all natural Religion, and consequently to the ruining of the Christian Religion, as likewise to Pyrrhonisme," that the rector investigate theses and dictates for evidence of "whatever error or dangerous principles are taught. " 16. The 8t. Andrews logic text exists as a manuscript in Edinburgh University Library (shelf mark: Dc. 5. 17). (It is hereafter noted as St. A). That this manuscript is the logic contributed by 8t. Andrews to the project of preparing a uniform course in philosophy for the four Scottish universities is convincingly shown by Christine King in the last few (unnumbered) pages of her dissertation, Philosophy and Science in the Arts Curriculum of the Scottish Universities in the 17th Century (Edinburgh: University of Edinburgh, 1974). For example, passages cited in discussion of the St. Andrews text found in University of Edinburgh records correspond exactly to passages in this manuscript, page by page. The criticisms presented by the University of Edinburgh can be found in Edinburgh University Papers, Dc. 1. 4 (2), EUL, p. 13, Feb. 1697. Here we find such criticisms as: In the introduction, the three ordinary figures of syllogism are omitted and only Gassendi's figure is presented; "the definitions of genus and species given by Ars Cogitandi need not be so positively approved seeing they are liable to some considerable difficultys." 17. F. Hutcheson, Synopsis Metaphysicae, Ontologiam et Pneumatologiam Complectens, 1742, part IT, ch. 1, sec. iii. 18. The 1695 commission mandated that copies of the 8t. Andrews logic be transmitted to the other universities for their correction, and then be used by all regents the following year. (Collection ofPapers, Dc. 1.4 (2), EUL, p. 119). 19. John Loudon was a regent at the University of Glasgow, 1699-1726, and Professor of Logic from the inception of this position in 1727 until his death in 1750. 20. Carmichael was appointed regent at the University of Glasgow in 1694 and then Professor of Moral Philosophy at Glasgow at the inception of the professorial system in 1727. He died in 1729, and Hutcheson was appointed to replace hinl. Carmichael, in lectures on pneumatology (the study of spriritual substances), develops an account of human cognition that is like the accounts of Loudon and Hutcheson here presented. But Carmichael, in his Breviuscula Introductio ad Logicam (1720), claims that logic is not the correct locus for discussion of the psychological faculties involved in apprehension

Hutcheson's Logicae Compendium and the Glasgow School

95

(i. e., sensation, imagination, and pure intellection), and so, in this logic text, he does not adopt the distinctive Scottish logic of ideas approach that we are here examining. Still, Carmichael's logic text is in the modem logic of ideas tradition. lIe states, in a report to the 1695 parliamentary visitation commission (Glasgow University Archives, 43170), that he uses the Port Royal logic in teaching his logic course. Further, in his Breviuscula, Carmichael divides his text into the same four parts as his contemporaries, and defines in much the same way the mental operations of apprehension, judgement, reasoning, and method. In addition, Carmichael develops such Port Royal notions as extension and comprehension, which are also presented in Hutcheson's text. See also Carmichael's logic lectures of 1697 (National Library of Scotland, Ms. 2741) and of 1708 (Glasgow University Library (GUL), Ms. Mu. 67; Ms. Gen. 255). 21. "Francis Hutcheson's Confusing University Career," Notes and Queries, New Series 42 (1995), pp. 56-59. 22. These lectures of John Loudon survive in a notebook in Glasgow University Library, Ms. Gen. 406. (hereafter noted as L) Loudon, in his introductory comments, objects to the dangers of Descartes's skepticism, and sides with Gassendi and with De Vries, who follows Gassendi in his attack on Descartes. 23. For a similar logic of ideas approach, see also, for example, S1. Andrews regent Colin Vilant, "Course on Logic," S1. Andrews (St. Salvatore), 1724, (Dc. 7. 91, EUL). Vilant's handwritten "Course on Logic" is said, on the title page, to be "after the same plan with Mr. Taylor's." Thomas Taylor taught at St. Leonard's, S1. Andrews when the St. Andrews text was written. Student notes of Taylor's lectures of 1698-99 (transcribed by James Goodsire; LF. 1117. c. 99 (1475), 81. Andrews U. Library) are very similar to the 1712 logic lectures of Glasgow regent John Loudon. 24. "A Slight Sketch of the Origin of the Glasgow Press and Academy of the Fine Arts" (EUL, Ms. 363). 25. That is, textbooks were unavailable for the basic university courses, viz., a year of humanities (i. e., Greek and Latin), and three years of philosophy, logic, metaphysics, ethics, and natural philosophy. 26. Another possibility is that Hutcheson originally composed a logic text early in his career, when he taught a logic course as head of a dissenting academy in Dublin. 27. St. Andrews logic text, Preface. 28. John Tran (regent at Glasgow), in a statement for the 1695 Commissioners, entitled "Method of Teaching in the Philosophy Class," says: "After the Public Examination, I begin with them their course of Logic, and finding no presented course I could ever yet satisfy myself .with, some of them being altogether out of the road of the philosophy now most valued; others of a later date industriously stuffed with doctrines of a very dangerous tendency (such as what Le Clerk teaches De Scientia Fide et Opinione in se collatis) and examples, apt enough to make a bad impression on the minds of youth, I dictate to them a short system much the same with what was ordered by the visitation, with some remarks on what I think may be improved or amended." (Glasgow University Archives, Ms. 43228). Gershom Carmichael reports of his teaching of logic at the University of Glasgow, that he first dictates a short compend, and then uses the Ars Cogitandi: "The places of that book that favour Popery are already noticed, and shortly are judiciously obviated.. .in notes that are printed with the book; besides I took further notice of some of them, in my marginal animadversions." (Glasgow University Archives, 43170). See also Munimenta Alme Universitatis Glasguensis, ed. C. lImes, 4 vols. (Glasgow, 1854), in the common statement of Scottish masters (vol. 2,

96

EMILY MICHAEL

p. 530): "Nor when we have seriously perpended the severall courses of philosophy which ar extant can we find any that wee can recomend as sufficient to be taught, for many of then1 ar wryte by popish professors and therein they cunningly insinuate the heretical tenets mixing them with their philosophy which ar not so easyly discerneable by the youth. Nor are the courses of philosophy wryte and calculate to be taught the youth..." 29. Hutcheson, Synopsis Metaphysicae, Pt. IT, ch. 3 and 4. 30. Following Gassendi's strategy, these Scots presume that purely intellective activities require an incorporeal mind, and that the mind, as incorporeal, must be immortal. 31. G. Carmichael, in lectures of 1708 (Ms. Mu. 67, GUL), adopting a similar approach towards the standard fourfold division of logic, distinguishes two maj or divisions in logic, apprehension and judgement. Discursus is treated as a part of the major section on judgement; and method is treated as an appendix. 32. See Jardine: "Judgement is called passing sentence...for in the compend [it is] expressed thus Ferens Sententiam, taken from a law court." He explains that as in a law court, there are three parties, defender, pursuer and judge, and it is the office of the judge to consider the sentence of both parties, and so to weigh them both in his mind, so this is the office of judgement. "For in judgement, the mind compares the subject with the predicate and judges of their agreement or disagreement." (GUL, Ms. Gen. 166, 1783, p. 33). (The compend referred to here is Hutcheson's Logicae Compendium). 33. For a discussion of this sort of reflection, see my paper "Francis Hutcheson on Aesthetic Perception and Aesthetic Pleasure," British Journal ofAesthetics, 24 (1984), pp.241-255. 34. This letter (dated 6 September 1727, from Hutcheson to Mr. William Mace) appeared in 1788 in the European Magazine and London Review, pp. 158-160. It is reprinted in part (with an introductory discussion by David Bennan). ("Francis Hutcheson on Berkeley and the Molyneux Problem," Proceedings of the . R . oyal Irish Academy, 74, p. 264) 35. Bennan, p.264. 36. Comprehension and extension are discussed in Scottish lectures in logic beginning in the late eighteenth century. See, for example, Herbert Kennedy (Dc. 8. 132, 1687, EUL); John Law (Dc. 8. 18, 1691, EUL). Gershom Carmichael, in his Breviuscula provides a lengthy discussion of comprehension and extension. He similarly states that a simple idea is one that cannot be resolved into different ideas, e. g., being, power, thought; a complex idea con1prehends several different ideas into which it can be resolved, as spirit, which is a thing capable of thinking. 37. Hutcheson, for example, says: "Postquam varias res, complexas excitantes ideas, mens observavit, easque vidit sibi invicem in quibusdam qualitatibus similes, in aliis dissimiles; abstrahendo se ab iis quibus differunt, retinendo vero ideas earum qualitatum in quibus sunt similes, easque nomine quodam signato denotando, facit ideam universalem." (H 1.3. i)

The Priority of Thought to Language in Cartesian Philosophy JILL VANCE BUROKER

Among its many successes, 20 th-century philosophy can count overturning the classical view of the relation between thought and language. The tradition lasting from the Greeks up to the 20 th-century took thought to be prior to language, and language to be incidental to the thinking process. This view is now known as "internalism." Beginning with Frege and Wittgenstein, however, the order of priority is reversed. In the recent model, called "externalism," language is essential to human thought in the strong sense that cognitive thoughts are partly constituted by linguistic practices. 1 As one would expect, these two models present rather different accounts of the thinking process and the way language signifies. In this paper I want to explore the Cartesian view of thought and language. In addition to Descartes's works, I shall draw from the Port-Royal Logic of Antoine Arnauld and Pierre Nicole, since it contains the most developed account of language and logic in the nlodern period. First I shall discuss the way Descartes's dualism informs the POI1-Royal view of the priority of thought to language. Then I shall mention the major difficulties in this theory. Finally, I shall make a few comments about Kant's role in the shift to externalism, and close with some remarks on the way externalism attempts to resolve the difficulties in the internalist view. 1. Cartesian dualism and the nature of thought

The modern account of thought and language has its basis in Descartes's dualistic metaphysics, according to which lllinds and bodies are really distinct substances. Despite the independence of minds from bodies, a human being is the exceptional case in which a mind and a body are connected to form a composite substance. One hallmark of substance dualisnl is the contrast between the public nature of bodies and the private nature of minds. Since bodies are extended in space they are in principle perceivable by the senses, and so they are public objects. Minds, of course, share none of these features. Not being spatially extended, minds cannot come under the purview of the senses. In consequence, they are not open to inspection by others, and so the only mind each of us can know directly is our own. Given this dualism, thinking is necessarily non-corporeal. Although humans have sensory states such as sensations, appetites and emotions, these experiences arise only by virtue of the connection between the mind and the body. In these cases the process culminating in the sensory state has a physiological basis: either contact with a body external to the perceiver's body, or, as in appetites and emotions, some process within the perceiver's own body.

98 JILL VANCE BUROKER Minds that have no connection to a body would presumably never enjoy or suffer from such sensory experiences. The key to Descartes's notion of thought is his theory of ideas, and his distinction between the formal and objective reality of an idea. The formal reality of an idea is its existence as a state of a mind. But it is also essential to an idea that it represent something or present some content to the mind. Descartes calls this aspect the objective reality of the idea. The primary function of Cartesian ideas is cognitive: to enable the thinker to know the truth. Knowing occurs when the thinker's grasp of this objective content corresponds in appropriate ways to a real state of affairs. The key point for us is that the representation relation between ideas and reality is a natural and intrinsic feature of the idea. It depends not at all on what thinkers do with these ideas, or even whether they are properly understood. To understand the Cartesian view of the relation between thought and language, we need to note four main features of this storehouse of ideas. First, this collection of ideas is fixed and non-subjective: there is only one such storehouse of cognitive contents, and it is not dependent on whether or how it is grasped by thinkers. I believe that for Descartes, the objective content of ideas is derived from the true and immutable natures of things. 2 Second, this storehouse is intersubjectively accessible. Because God is not a deceiver, he has given all minds the same innate capacity to apprehend these natures and their interrelations. 3 Third, the veridical grasp of these ideas is purely intellectual. As readers of Descartes are aware, one of his revolutionary views was to deny that sense experience yields more than an obscure and confused picture of reality.4 An accurate representation of reality requires a "clear and distinct" apprehension by the intellect alone. Fourth and finally, this intellectual grasp is intuitive or direct-it is not nlediated by anything. It does not depend on having a body, or on one's relations to other minds or bodies, or on any use to which one may put these ideas. Most important, of course, it does not depend on having or using language. 5 Descartes was representative of thinkers of his time in classifying mental operations into four kinds. These operations, arranged in order of increasing complexity, are conceiving, judging, reasoning, and ordering. This classification scheme provides the organization of topics in the Port-Royal Logic. Conceiving is the simplest operation, and consists in a passive, instantaneous grasp of an idea. When the perception is clear and distinct, it is the act of intellectual intuition described above. Despite the instantaneous nature of conception, the content of the thought is always complex. In fact, in Descartes's account there is no clear separation between a complex idea and a proposition: for example, apprehending the nature of a right triangle includes recognizing that it has "the properties which license the inference that its three angles equal no more than two right angles."6 Descartes drew a sharp line, theoretically, between conceiving and judging, since judging includes, in addition to the

The Priority of Thought to Language in Cartesian Philosophy 99 conception, an act of the will in which the thinker affirms or denies that the conceived proposition is true. In general, however, it is possible to grasp a proposition without judging its truth, and so conceiving is logically prior to judging. And despite Descartes's stated view that only judgments are formally true or false, it appears that the propositional ideas being judged are really the bearers of truth values. At least this is implied by Descartes's notion that in clear and distinct perception one can apprehend the truth of a proposition, since this intuitive apprehension is logically independent ofjudgment. In Descartes's model of the mind, successive thinking does not enter the picture until the third stage, that of reasoning or deduction. Descartes defines deduction at Rule Three in the Rules for the Direction of the Mind as "the inference of something as following necessarily from some other propositions which are known with certainty." In deducing "we are aware of a movement or a sort of sequence...and.. .immediate self-evidence is not required for deduction, as it is for intuition... ,,7 Deduction is not as certain as intuition because it depends on memory. But even here Descartes does not make a sharp separation, since in principle deductions can be reduced to intuitions, with the right amount of practice (see Rule 11). In spite of the complexity of the intuited idea, then, our most certain grasp of it is instantaneous. Put another way, Descartes's model of judgment makes no room for the notion of constructed or synthetic thought. Now this is one place where the authors of the Port-Royal Logic differ from Descartes. While agreeing that judging involves an act of the will, Arnauld and Nicole maintain that this act just is identical to forming the proposition judged: "After conceiving things by our ideas, we compare these ideas and, finding that some belong together and others do not, we unite or separate them. This is called affirming or denying, and in general judging."8 The part of a judgment that represents the act of willing is the copula, expressed linguistically by the verb. The copula has two functions in judgment: it connects the subject and the predicate, and it signifies affirmation or denial. So although Arnauld and Nicole make it impossible to distinguish merely apprehending a proposition from judging it, they recognize that propositional thought is successive or constructed. Since forming a proposition requires one to compare ideas, merely apprehending a proposition cannot be instantaneous as it is for Descartes.

II. The Dependence of Language on Thought Descartes's dualism leads directly to the traditional view that language depends on thought. As opposed to the purely mental nature of thought, language is inherently physical: "Words," say Arnauld and Nicole, "are distinct and articulated sounds that people have made into signs to indicate what takes place in the mind.,,9 This is a succinct statenlent of the classical conception of language. First, words are taken to be sounds, so language is essentially

100 JILL VANCE BUROKER corporeal. Second, the fundamental use of these sounds is to express interior thoughts to others. And third, words express thought by being conventional signs of ideas. Let us now look at each of these features in turn. 1) Words are articulated sounds. As I remarked earlier, the Cartesians analyzed sensations as mental states caused by bodies coming into contact with the perceiver's body. Now insofar as language is identified with sound, beings who communicate linguistically must have the appropriate physical organs. A disembodied Cartesian mind could think, but it would have no way to express itself or communicate with other minds. Of course, not just any sound emitted by a human being counts as a word. A word is an articulated sound, that is, a discrete unit which, when combined with others in appropriate ways, produces a sentence. Furthermore, the words of natural languages allow an infinity of possible combinations: there is in principle no upper bound to the con1plexity of the resulting expressions. In part V of the Discourse on the Method Descartes recognizes that although we can devise machines to produce sounds resembling words, these sounds are not part of a language precisely because the possible combinations in response to a situation are limited by rn.echanical constraints. Nor, says Descartes, should we think that "the beasts speak, although we do not understand their language. For if that were true, then since they have many organs that correspond to ours, they could make themselves understood by us as well as by their fellows."lo In other words, every natural language is in principle translatable into every other natural language. This is undoubtedly because language expresses ideas, and there is only one storehouse of meaningful ideas. The unit of linguistic meaning for Cartesians, then, is the word, which is an element of speech and is essentially corporeal. 2) Language expresses antecedent thought. Since nlinds can think independently of relations to bodies or other minds, language is merely an external vehicle for making public one's private thought. And since thinking is non-corporeal and language is corporeal in the way described above, thought must itself be non-linguistic. Descartes does admit in the Principles that once we have learned to speak, we may not be conscious of non-verbal thinking: ... because of the use of language, we tie all our concepts to the words used to express them; and when we store the concepts in our memory we always simultaneously store the corresponding words. Later on we find the words easier to recall than the things; and because of this it is very seldom that our concept of a thing is so distinct that we can separate it totally from our concept of the words involved. ll (Principles I, art. 74)

Similarly, Arnauld and Nicole explain that a logic text must be concerned with language "because we can make our thoughts known to others only by accompanying them with external signs..." And they remark that "this habit is so strong that even when we think to ourselves, things are presented to the mind only in the words in which we usually clothe them in speaking to others... ,,12 So although we think linguistically in practice, words are only the external

The Priority ofThought to Language in Cartesian Philosophy

101

"clothing" of thoughts. The thought is intrinsically heterogeneous from its verbal expression. 3) Language is a conventional sign of thought. Like their predecessors, the Cartesians explain how language works in terms of conventional signification. This account is part of a more general theory which recognized three kinds of signs: natural, conventional, and divine. In all three cases, the sign exists insofar as it causes or prompts an idea in the mind of the person recognizing its significance. In chapter 4 of Part I of the Logic, Arnauld and Nicole explain that signs are perceivable objects that signify other things by exciting an idea of the signified thing in the mind of the observer. Hence the signification relation is a causal-psychological relation between an object and an associated idea: something is a sign only if it triggers the appropriate idea in the perceiver. 13 For natural signs the relation has four parts. To cite a standard example, smoke is a sign of fire when the person who perceives the smoke is immediately brought to think of the fire causing it. First there is the thing being signified, in this case the fire. Next is the sign, the object doing the signifying, here the smoke. The last two elements are mental states of the person for whom the sign has significance. For the perceiver must first perceive the sign itself as an object, and this perception must then prompt in the perceiver a second idea of the thing being signified. For linguistic signification the model is slightly more complex, because, strictly speaking, language signifies the speaker's ideas. Thus the model for linguistic signification has five parts. First is the speaker's idea, which is expressed by, second, the words, which, third, the listener perceives under normal conditions. This perception then prompts the listener to form, fourth, the appropriate idea associated with the sound. And this idea, of course, represents, fifth, the same things represented in the speaker's mind. In the narrow sense, words signify ideas in the speaker's mind. But Port-Royal uses "signify" more widely since they often say the words signify the things one thinks of. In the traditional account of language, the relation between words and things was generally construed as a form of naming. 14 Now it is clear from this model that communicating can succeed only if the speaker and hearer associate the same sound with the same idea. Thus linguistic communities are defined by the mental dictionaries they share. Speakers who want to express a thought must first consult their dictionaries. Hearers must do the same to decode the sounds. Neither Descartes nor Arnauld and Nicole explain how we learn these associations, but presumably it would be through usage, as we develop our linguistic abilities. In this theory of meaning the connection between words and ideas is arbitrary. That is, the association is not natural or logical, but is subject to our will. This explains why different natural languages associate the same thoughts with different sounds, and why, within a language, speakers can invent new words and change the meanings of sounds already in use. This freedom is possible only because sounds are in themselves indifferent to the ideas they are

102 JILL VANCE BUROKER used to express. But as Arnauld and Nicole point out, we should be careful about exactly what feature of language is arbitrary. What is arbitrary about words is how they are associated with thoughts, not the content of the thoughts they express .15 The conventional nature of this connection between thought and language leads to a particular account of a word's sense. Since words and ideas are heterogeneous items, words have neither a natural nor a logical relation to the ideas they signify. The sound, as we saw earlier, merely prompts the idea based on a given association. On this view, the relation between words and ideas is opaque rather than transparent. Hence language can in no sense be said to constitute thought. But in spite of the external connection between words and ideas, language can signify thought because both are articulated systenls. Ideas have a natural order and structure independent of their linguistic expression. So the syntax of the sentence must bear some correlation to the relations among ideas. But this correlation is at best an isomorphism, since speech and ideas are heterogeneous items. III. Difficulties and Solutions Now as contemporary philosophy has shown, this traditional view of thought and language is beset by serious problems. Here I will briefly mention five of them and then sketch Kant's role in the developments leading to the externalist view of thought and language. 16 These are the main problems with the classical view that thought is prior to language: First, there is no plausible account of autonomous cognitive thought or, as it has been dubbed in contemporary literature, "mentalese." Even worse, it is not clear what such an account would be like. Given the constraints on the Cartesian theory, mentalese can neither consist of sensory images nor be dependent on language, since both require that the mind be connected to a body. There is in fact no good candidate for purely mental activities detached from linguistic or other behavioral practices. This leads directly to the second problem: since we do not know what "mentalese" consists of, surely we cannot know what it means to grasp it. Recall that for the Cartesians this grasp is in principle both instantaneous and complex. One symptom of this problem is Descartes's recourse to the visual metaphors of "clear and distinct" perception and "the natural light of reason." In a way intellectual intuition resembles a mystical experience. Another symptom of the problem is the nunlber of debates in the modern period over whether in fact we even have ideas of certain things: consider Berkeley on the idea of matter, and Hume on the "fictitious" ideas of power and the mind. It has always struck me as bizarre that although the contents of consciousness are supposed to be transparent to the subject, it is possible to disagree about whether in fact one has a certain idea. Once we give up the notion that thought

The Priority of Thought to Language in Cartesian Philosophy 103 is autonomous and fundamental, however, this second problem will disappear along with the first. Third, the traditional theory cannot explain how ideas are "about" the world or the connection between thought and truth. As we have seen, the representational nature of mental states is just a primitive, intrinsic feature. For Descartes the guarantor of this connection is, of course, God. Now perhaps it has always seemed plausible that there is such a thing as mentalese about the world, because of the analogy with sensations. Sensations, of course, both have a presentational content and are caused by objects in the world. But that model cannot work for the relation between pure thought and reality, since real natures cannot cause our thoughts in any direct sense. The fourth problem concerns the connections between thought, truth, and judgment. For Descartes judging is an act of the will in which one commits oneself to the truth or falsity of the proposition apprehended. Now as has been pointed out, this amounts to a transitive model in which judging consists of the judger, the act, and the thought being judged. 17 On this view, to judge the proposition p one must apprehend not only p, but also that p is true or false. The problem is how to classify one's apprehension of the truth value ofp. This thought, that p is true (or false), cannot be merely apprehended; one must commit oneself to its truth. So one must both apprehend and judge that p is true or false before one can judge p, which results in an infinite regress. 18 This problem follows from two related views, namely that the thought is grasped independently of being judged, and second, that truth-values attach to the thought rather than to its assertion. The last problem concerns the idea of the mental dictionary, both the possibility of one in general, as well as an account of how thinkers acquire it. The most important developnlent on this topic was Wittgenstein's private language argument, which showed that the associationist model of meaning cannot give a coherent account of the right association between language and independently existing thoughts. 19 Since we have no means of identifying thoughts independently of linguistic acts or other activities, the whole notion of a set of associations between thoughts and words collapses. The process of ostensive definition cannot solve the problem either, since at best it directly associates words with existing things. And as Wittgenstein argued, for any object pointed at there are too nlany possibilities to determine the thought to be associated with the thing. The externalist response to these difficulties has been to reject the view that cognitive thought is independent of language. While Frege undoubtedly took the first major step in this direction, Kant's theory of judgment in the Critique ofPure Reason prepared the ground for this shift. Of course Kant is still in the Cartesian tradition of taking thought to be fundamental, and in fact he has almost nothing to say about language. But in spite of his traditional stance, Kant's Critical theory makes it possible to reverse the priority.

104 JILL VANCE BUROKER In The Critique of Pure Reason Kant simultaneously rejects two basic features of the classical view: first, that the intellect can instantaneously grasp complex representations, and second, that conceiving is prior to judging. Human intuition is passive for Kant, but it is possible only through the sensibility. In contrast to this sensible receptivity, the understanding acts to connect the data given through the senses. Understanding is a synthetic activity in which one constructs a complex unified representation using concepts. }~0\X/ it is an axiom of Kant's theory of synthesis that we grasp as complex only what we have combined: Combination does not. ..1ie in the objects, and cannot be borrowed from them, and so, through perception, first taken up into the understanding. On the contrary, it is an affair of the understanding alone, which itself is nothing but the faculty of combining a priori, and of bringing the manifold of given representations under the unity of apperception. 20 (B134-35)

This conceptual activity of the understanding, which I(ant calls "synthesis," just is the act of judging. Kant says explicitly: "Now we can reduce all acts of the understanding to judgments, and the understanding may therefore be represented as a faculty ofjudgment." (A69fB94) Whereas Descartes sharply divided the passive intuition of the intellect from the voluntary act of judging, Kant declares that the understanding has one and only one function, which is to judge. In a key move that paves the way for Frege's context principle-"It is only in the context of a proposition that words have any meaning,,21-Kant analyzes concepts as "predicates of possible judgments." (A69/B94) Conceiving is not independent of judging, and concepts are not prior in the order of explanation. In short, Kant has reversed the relation between concept and judgnlent: for him all conceptualizing is judging. Once thinking is construed as the synthetic activity of judging, we are only one step from reversing the priority of thought and language. All that remains is to identify judging with the speech act of assertion. On the externalist view, the unit of significance is the entire speech act. Speech acts are utterances in which a speaker performs some act in a social context, against a background of conventional rules, including linguistic rules. 22 The utterance typically consists of some propositional content expressed in a certain way. 'The way the content is expressed is called the illocutionary force of the act. So, for example, assertions have one kind of i110cutionary force, questions another, promises a third, and so on. In this model the thought or the proposition cannot be grasped independently of performing sonle speech act. Now this does not mean that one cannot merely apprehend a thought. But rather than being a pre-requisite of asserting, considering a proposition has a different illocutionary force from asserting or judging. In sharp contrast to the Cartesians, speech act theorists take the unit of significance to be the entire utterance, not any bit of language. Despite the dependence of thought on language, however, externalism does not reject the possibility of interior monologue or unexpressed thought. It is just

The Priority of Thought to Language in Cartesian Philosophy 105 that such "private" thinking depends on having mastered the conventions of public language. It should be fairly clear how this model solves some of the problems left to us by Cartesians. Since articulated and cognitive thinking depends on speaking, the "form of thought" is language. Grasping the thought just consists in performing and recognizing speech acts. The relations of thought to the world and of thought to truth are also less problematic on this view. Thinking is about the world insofar as we refer to objects and say things about them. Similarly, if truth-values are features of utterances rather than propositions, then the relation between thought and truth becomes less mysterious. To think of the truth value of a proposition is really to abstract from the general truth-conditions of the entire act of assertion. And finally, the problem of the association between meanings and words disappears once we do away with independent mental entities. We learn the meanings of words by learning how to use them, by learning how to perform various speech acts. Although language is essential to thought for adult humans, there is no reason thought could not take other forms. Animals and infants do not use language, so their thinking must be non-linguistic. What this consists of is anyone's guess, although obviously it would be closely connected to perceptual, affective and motor states. But even their sense perceptions of objects would have to fall short of the full-blown judgnlental forms Kant argues for in the transcendental deduction. 23 What the failure of Cartesianisnl has taught us is the impossibility of cognitive thought without some form of embodiment.

NOTES 1. Undoubtedly the first major paper arguing for this approach is Hilary Putnam's "The Meaning of 'Meaning,'" reprinted in Philosophical Papers: Mind, Language and Reality, vol. IT (Cambridge: Canlbridge University Press, 1975) pp. 215-271. A second major work contributing to the development of externalism is Saul Kripke's Naming and Necessity (Cambridge: Harvard University Press, 1972). 2. Briefly, this is because for Descartes all knowledge concerns existing things, which are either minds or bodies, or common notions and eternal truths. The latter group includes such propositions as "nothing comes from nothing," "two things equal to a third are equal to each other," and "nothingness possesses no attributes." These propositions function roughly as axioms and rules of inference in his system. The simple natures are ideas of things and their properties. In this group Descartes distinguishes between the real natures-thinking and extension-and the common (or transcendental) natures, which include substance or existence, duration, order, and number. See Rule 12 of Rules for the Direction of the Mind, in The Philosophical Writings of Descartes, translated by John Cottingham, Robert Stoothoff, Dugald Murdoch, and Anthony Kenny, 3 vols. (Cambridge: Cambridge University Press, 1985-91) vol. I, pp. 44-45~ and arts. 48-49 and 52 of Part I of Principles ofPhilosophy, in Philosophical Writings, vol. I, pp. 208-210.

106

JILL VANCE BUROKER

3. This does not imply that all thinkers can grasp these truths with the same ease, or that human minds, being finite, are capable of fonning an adequate idea of the nature of God, an infinite being. 4. I explain my view of Descartes's rationale for labelling sensory states as obscure and confused in "Descartes on Sensible Qualities," Journal ofthe History ofPhilosophy 29 (1 991) pp. 585-611. 5. As I shall show below, for the Cartesians language requires having a body. 6. Fifth Meditation, in Philosophical Writings, vol. II, p. 47. 7. Philosophical Writings, vol. I, p. 15. 8. Antoine Arnauld and Pierre Nicole, Logic or the Art of Thinking: the Port-Royal Logic, translated and edited by Jill Vance Buroker (Carubridge: Cambridge University Press, 1996) Part II, chap. 3, p. 82. 9. Arnauld and Nicole, Part II, chap. 1, p. 74. 10. Discourse on the Method, in Philosophical Writings, vol. I, p. 141. 11. Principles ofPhilosophy, in Philosophical Writings, vol. I, p. 220. 12. Arnauld and Nicole, from the introduction to Part I, pp. 23-24. 13. Moreover, as the authors make clear further on in this same chapter, the sign must signify sonlething other than itself, or itself in another state: "Although something in a given state cannot be a sign of itself in the same state, since every sign requires a distinction between the thing representing and the thing represented, it is certainly possible for something in a given state to represent itself in another state, just as it is possible for someone in his room to represent himself preaching. Hence the mere difference in state is enough to distinguish the symbol from the thing symbolized. In other words, the same thing can in a particular state be the symbol and in another state be the thing symbolized." (Logic, p. 36) Although the reason is not stated, I assume that the thing being signified must differ in some way from the sign to avoid trivializing the notion of signification: otherwise everything would be a sign of itself. On this analysis, ideas would not themselves be signs, contrary to the treatment by such thinkers as Locke. 14. For a more detailed view of the Port-Royal theory of language, see my "The Port-Royal Semantics of Tenns," Synthese 96 (1993) pp. 455-475; and "Judgment and Predication in the Port-Royal Logic," in The Great Arnauld and Some of His Philosophical Correspondents, Elmar 1. Kremer, ed. (Toronto: University of Toronto Press, 1994) pp. 3-27. 15. See Part I, chap. 1 ofArnauld and Nicole, p. 28. 16. These problems are suggested by Michael DUffiL~ett's discussion of Frege's theory of sense in "Frege's Myth of the Third Realm," in Frege and Other Philosophers (Oxford: Clarendon Press, 1991) pp. 249-262. 17. See David Bell, Frege's Theory ofJudgement (Oxford: Clarendon Press, 1979) p. 121. 18. Here's another way to see the regress: To judge p is to recognize first that p is true; let "p is true" = q; so before one can judge p one must judge q. But then to judge q one must recognize first that q is true; let "q is true" = r. So before judging q one must judge r, and so on. This regress was suggested to me by Bell's argument that thoughts naturally occur assertively in Frege 's Theory ofJudgement, pp. 127-128. 19. Ludwig Wittgenstein, Philosophical Investigations, translated by G. E. M. Anscombe (New York: The Macmillan Co., 1958). See especially sections 243-315. 20. Immanuel Kant, The Critique of Pure Reason, translated by Norman Kemp Smith (London: S1. Martin's Press, 1970) p. 154. As is customary, passages are identified by the standard AlB pagination, referring to their location in the 1781 and 1787 editions.

The Priority o/Thought to Language in Cartesian Philosophy

107

21. Gottlob Frege, The Foundations ofArithmetic, translated by 1. L. Austin Oxford: Basil Blackwell, 1959) p. 73 e . 22. An excellent introduction to speech act theory is available in John R. Searle, Speech Acts (Cambridge: Cambridge University Press, 1969). 23. For Kant, all perceptual states involve synthesis, but they are not all governed by concepts of the understanding. Kant believes that animals and infants lack intellectual capacities. Thus, in contrast to adult human sense perceptions, their sensory states would not incorporate judgments. A helpful discussion of Kant's view is Steve Naragon's "Kant on Descartes and the Brutes," Kantstudien 81 (1990) pp. 1-23.

Berkeley's Metaphysics and Ramist Logic FRED WILSON

"What is truth?" asked Pilate and did not wait for an answer. Aristotle would have told hinl that "To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, and of what is not that it is not, is true..." (Metaphysics, lOllb 27-9)1 But Aquinas tells us that: ...a house is said to be true that expresses the likeness of the fonn in the architect's mind; and words are said to be true so far as they are the signs of truth in the intellect. In the same way natural things are said to be true in so far as they express the likeness of the species that are in the divine mind. For a stone is called true, which possesses the nature proper to a stone, according to the preconception in the divine intellect. Thus, then, truth resides primarily in the intellect, and secondarily in things according as they are related to the intellect as their principle. 2

Saying of what is, that it is, is a rather complicated thing, in other words, in spite of its apparent simplicity. For, what is, depends upon one's ontology. According to Aquinas, the truth of a thing is a standard to which the thing bears a likeness. This standard is the form or species of the thing in the divine intellect. The person who knows abstracts the species or standard of its truth from the thing judged about, and compares the object to this standard. When... [the mind] judges that a thing corresponds to the fonn which it apprehends about that thing, then first it knows and expresses truth. This it does by composing and dividing: for in every proposition it either applies to, or removes from the thing signified by the subject, some fonn signified by the predicate...(ST, FP, Q 16, Art. 2)

When the mind is thus judging the thing, the form of the thing known is in the mind. ... since everything is true according as it has the fonn proper to its nature, the intellect, in so far as it is knowing, must be true, so far as it has the likeness of the thing known, this being its fonn, as knowing. For this reason truth is defined by the confonnity of intellect and thing; and hence to know this confonnity is to know truth. (ST, FP, Q 16, Art. 2)

As for the individual of which forms are predicated, this, e. g., this man ".. .is said to be a suppositum, because he underlies [supponitur] whatever belongs to man and receives its predication. (ST, TP, Q2, Art. 3) This subject is made actual by virtue of its having accidents predicated of it: "...a subject is compared to its accidents as potentiality to actuality~ for a subject is in some sense made actual by its accidents. (ST, FP, Q3, Art. 6) The substance is, however, an individual thing that is distinct from those properties or accidents:

110

FRED WILSON Although the universal and particular exist in every genus, nevertheless' in a certain special way, the individual belongs to the genus of substance. For substance is individualized by itself~ whereas the accidents are individualized by the subject, which is the substance; since this particular whiteness is called "this," because it exists in this particular subject. (ST, FP, Q 29, Art. 1)

The unity of the individual derives from the substance. The collection of properties or accidents which characterize the individual are unified into the set of characteristics of a particular individual precisely because each of them inheres in the substance. The unity of any ordinary thing, e. g., a die, derives from the substance in which all the properties of the thing inhere. As for Pilate's question, What is truth?, in order to answer that we must lay out our ontology. In the case of the traditional Aristotelian ontology that we have been looking at, a substance is true, true to its own self if you wish, provided that the properties or accidents that make the substance actual are those required, or of the sorts required by the species or fornl of the substance. This doctrine of truth depends upon each substance having a certain form or nature or essence, or, what is the same, from the side of mind, exhibiting an intelligible species. In the 17th century a different ontology developed, one that rejected the intelligible species of Aristotle and Aquinas. It retained, however, the doctrine of substance. This entailed a rejection of the traditional doctrine in which a substance was said to be true to the extent that it conformed to the intelligible species that defined its essence and gave it being. But the retention of the doctrine of substance still tied truth to the substance ontology. To speak the truth is to say of what is, that it is. And to say of what is, that it is, is to predicate a property of a substance. Now, what this means is that the relation of predication occupies a special place in any account of logic and language prior at least to the early 20 th century, and is contrasted to other, "ordinary" if you wish, relations. Logic and grammar as accounts of language deal with propositions. These are either true or false, as Aristotle said. When we say of what is that it is, when we say, for example, that "0 is S" then what we are saying, on the traditional view, is that the substance, referred to by the subject ternl "0", has the property referred to by the predicate term "S." If a is in fact S, then we are saying of what is, that it is, or, in other words, we are saying what is true. And if 0 is after all not S, then we are saying of what is not that it is, or, in other words, we are saying what is false. On the view that substances are the basic entities in one's ontology, as Aquinas, following Aristotle, indicates, then all other, "ordinary," relations are among the properties that are predicates of substances. This view is standard enough to constitute the pattern laid out in the comnlon logic texts of the age. Let us look, to take one example, at the Monilio Logica of Franco Burgersdijck, a widely used text in the early modern period. Burgersdijck was professor of philosophy at Leyden, and his Monilio Logica was republished in

Berkeley's Metaphysics and Ramist Logic 111 his native Holland several times, as well as being published in England, including a translation of 1697. 3 In spite of its condemnation by Locke, it continued to be used. Anlong other things it discusses relations, presenting a view that was uncontroversial in the early modern period, putting them in a context of an ontology of substances. Burgersdijck in fact discusses relations in two contexts. One is in the context of the categories or predicaments and the other is in the context of the . 4 tOPICS.

For the mediaevals, the entities about which one could talk were divided into the categories. These categories or predicaments are those of substance, quality, quantity, relation, action, etc. Burgersdijck gives the traditional list which derives, of course, like most everything else in logic both mediaeval and modern, from Aristotle. A sentence or proposition makes a predication of an entity in a category when it attributes a certain property to it. A predication is legitimate, or, as we would say, "well formed," no matter the category to which the subject of predication belonged, provided that the property predicated of that subject fell vvithin one of the predicables: relative to the subject, it had to be a genus, species, difference, property or accident. The list of predicables defines the various sorts of judgment that can be made, and that can enter into discourse and into the reasoning that appears in syllogisms. In Burgersdijck the discussion of the predicables follows immediately upon the discussion of the predicaments or categories. This discussion is immediately followed by a discussion of the various topics, that is, the places fronl which arguments nlay be drawn for use as premises in disputations and arguments. These places include those of property and accident, whole and part, cause and effect (which includes matter and form, efficient cause, and end), subject and adjunct, similarity (what Burgersdijck calls "convenience"), difference ("diversity or distinction"), opposition, and order. In fact, as the text proposes, there is no real separation of the predicables from the topics, so that the various predicables themselves become topical places, or, what in effect amounts to the same, the topics come to define different sorts of legitimate predication. Once this is recognized, we see that Burgersdijck has in effect extended, whether he knew it or not, the list of legitimate forms of predication. Bringing predicates together as cause and effect, as in "fire causes heat" for example, becomes one form of predication, alongside the predications allowed by the predicables, for example, that of a genus as in "man is animal." For Burgersdijck, then, the list of legitimate forms of judgments has been extended beyond that of the mediaevals: the forms of legitimate judgment now include not only that defined by the relation of copulation, but also those defined by such relations as cause and effect, whole and part, similarity or resemblance, and difference. Here we find a difference between Burgersdijck's treatment of relations and that deriving from Aristotle. For the latter, relations are one of the categories~ they are among the things predicated of substances. But in Burgersdijck~!~~-__

112 FRED WILSON tions are discussed in two ways, once in the context of the list of categories, and once in the list of topics. Relations are at once in a special category of entity and are also ways of relating such entities to each other in judgments. Burgersdijck describes the category of relation in this way-which of course goes all the way back to the first chapter of Aristotle's Categories: "Those things are said to be related, which in Respect of what they are, are said to be others, that is, of others or in any other Manner or Respect are referred to another." (ML19) According to Burgersdijck every relation involves two entities, that from which the relation originates and that in which it terminates. "In every Relation," he tells us, "are required Subject and Ternl...That is called the Subject to which the Relation is attributed; or that which is referred to some, other thing...That, the Term to which the Subject is referred." (I\1L21) The subject is the relate and the term the correlate. Relational predications thus presuppose qualities in both the subject and the term. The quality may either be a property, that is, a quality had universally by subjects of the relevant sort, or an accident, that is, a property that is not present universally. In the latter case Burgersdijck speaks of the relation having a foundation. "Some Relations are supposed, supposing the Subject and Term: Others besides these do require a Foundation...And a Foundation is that by whose Means the Relation acrews to the Subject." (ML22) He gives the following examples: When an Egg is said to be like an Egg, the Similitude between these two Eggs arises in each as soon as they begin to exist; nor is there any thing required towards their Relation, besides the Existence of two Eggs. But the Relation of Servant does not presently arise in the Subject so soon as the Tenn exists; but it behoves that something else also do intercede upon which this Relation is founded: For a Servant is therefore the Servant of one, because by him he has been either saved or purchased, &c. (ML22)

The distinction between property and accident was to become increasingly unimportant, so for the sake of convenience let us speak of those qualities in a thing by which the relation "acrews to the subject" as the foundation. Now, in the basic instance the subjects and terms of relations are substances. The substances are related to each other by virtue of the foundations that are present in them. As the examples make clear, the qualities that are the foundations of the relation are themselves non-relational. Moreover, each of the substances that are the subject and the term is, as Burgersdijck puts it, following the ancient formula, "a Being subsisting of it self, and subject to Accidents." (ML8) Unlike accidents, substances do not depend for their existence on something else: "To subsist by it self is nothing else but not to be in any thing as a Subject. .." (ML8) Now, an accident is "a being inherent in a Substance"; it "cannot exist without a Subject;" nor can it "pass from one Subject to another." (MLIO) Thus, an accident cannot exist apart from the substance in which it is present. So substances can exist apart from other substances. If a substance ceases to exist, so do its accidents; but since substan~e~

_

Berkeley's Metaphysics and Ramist Logic 113 subsist by themselves, a substance can exist unchanged, that is, remain the same in its being, even if other substances were to cease to exist. Since it is not predicated of other things, the non-existence of other things does not affect its existence or being. What is distinctive of relation, then, is not that it introduces an entity over and above substances and the qualities that are present in them as properties and accidents. What is important about relation is that by virtue of the qualities in them one substance is referred to another. In a relational judgment, that is, one of the judgemental forms given in the list of topical relations, one substance becomes the subject of the judgments, another the term. Here, the substance that is the subject becomes so through a non-relational quality that has been picked out as the foundation of the judgement in that substance~ and the substance that is the term becomes so through a non-relational quality that has been picked out as the foundation of the judgment in that substance. Thus, substances become subjects and terms, and qualities become foundations, through the fact that they appear in a relational judgment. Relations do not constitute an ontological union among substances; such union as occurs is a union in the judgment. We find this view of relations repeated in Locke, but in the context of a systenlatic account of the operations of the mind. 5 Thus he tells us that: The nature therefore of Relation, consists in the referring, or comparing two things, one to another; from which comparison, one or both comes to be denominated. And if either of those things be removed, or cease to be, the Relation ceases, and the Denomination consequent to it, though the other receive in it self no alteration at all, v. g. Cajus, whom I consider to say as a Father, ceases to be so to morrow, only by the death of his Son, without any alteration made in himself. (Essay, II, xxv, 5)

A little later Locke adds that, "...there can be no Relation, but betwixt two Things, considered as two Things. There must always be in relation two Ideas, or Things, either in themselves really separate, or considered as distinct, and then a ground or occasion for their comparison. (Essay, II, xxv, 6) Here we have the same doctrine of relations that one finds in Burgersdijck: relational judgments are about facts that are essentially non-relational, facts in which a non-relational quality is present in a substance, while the bringing together of these two into a unity of the subject and term is provided by the mind referring the one to the other in a mental act ofcomparison. This view can be represented as follows. 6 Consider the relational statement that (@) aisRtob On the Lockean account, such a statement has a two-fold analysis. On the one hand, there are the objective truth conditions, the objective facts concerning a and b that determine whether the statement is (objectively) true or false. On the other hand, there is the subjective mental state that the use of the relational

114 FRED WILSON statement expresses. As for the former, the objective facts represented by (@) are non-relational: (#) a is rl and b is r2 The non-relational properties rl and r2 are the (objective) foundations of the relation. Both (+) a is rl and (++) b is r2 will be true. What must be noted is that a and b are independent of each other in the sense that, even if one ceases to exist, this will not affect the being of the other. Thus, for example, if a were to cease to exist, the relational fact (#) would no longer obtain. Moreover, (+) would also cease to obtain. Nonetheless, the being of b would be unaffected. For, (++) would still be true: the nonexistence of a will imply that (+) could no longer be true, but this does not affect the predications that would continue to be made of b, including the predication represented by (++). As for the subjective state that the use of (@) expresses, this is ajudgement of comparison. The two substances a and bare entities capable of subsisting by themselves on this account of relations. On the usual account, with which we are more familiar, deriving from Russell, this is not so. On this latter account, (@) will be represented by a primitive relational predicate: (*)

R(a, b)

This in turn means that we will have both (&) a is (R to b) (&&) b is (R-ed by a) However, if a ceases to exist, then not only will (*) cease to be true, and not only will (&) cease to be true, but in addition (&&) will also cease to be true. In other words, upon Russell's ontology of relations, if one of the relata in a relational fact were to cease to exist, the being of the other relatum would change. Upon Russell's account, then, the two relata are not independent of each other as they are upon the traditional account for relations defended alike by Burgersdijck and Locke. Locke locates this traditional account of relations in the context of his more general account of thought and language. Human beings are sociable animals, and language is the nledium by which they communicate. Articulate sounds express our ideas, and enable us to communicate those ideas to others. (Essay, III, i, 1-2) Various complex ideas, such as the mixed modes, consist of complex ideas whose unity is not a matter of objective connection but of the mind putting them together. Nobody can doubt but that these ideas of mixed modes are made by a voluntary collection of ideas put together in the mind, independent from any original patterns in nature, who will but reflect that this sort of complex ideas may be made, abstracted, and have names

Berkeley's Metaphysics and Ramist Logic

115

given them, and so a species be constituted, before anyone individual of that species ever existed. (Essay, III, v, 5)

The examples that he gives are the relations of sacrilege and adultery; he could equally have used sons or fathers. There are three aspects to the creation of these ideas, and their representation in language, according to Locke: ...we must consider wherein this making of these complex ideas consists; and that is not in the making any new idea, but putting together those which the mind had before. Wherein the mind does these three things: First, it chooses a certain number: Secondly, it gives them cOlmexion, and makes them into one idea: Thirdly, it ties them together by a name. If we examine how the mind proceeds in these, and what liberty it takes in them, we shall easily observe how these essences of the species of mixed modes are the workmanship of the mind; and consequently, that the species themselves are of men's making. (Essay, III, v, 4)

Indeed, it is the habit formed by the word that provides the cement as it were that keeps the parts of the idea together over time in a single, unchanging complex. (Essay, III, v, 10) In any case, the point is that the unity of the complex is a unity that derives from an act of the mind. The word that we use not only expresses the idea but expresses, too, the act of unifying that binds the simple ideas together in the mixed mode or other abstract idea. Communication thus consists in grasping not only the simple ideas that make up our complex ideas, but in also grasping the· act of the mind that unifies these ideas into wholes. The relational acts in my mind must correspond to the relational acts in your mind, if our communication is to be successful. There is one exception, one case where the unity of the complex derives from an objective ontological basis, and not simply from the act of mind that unifies the ideas into a whole. That, of course, is the unity that is observed in perception and derives from a substance. Whatever ideas we have, the agreement we find they have with others, will still be knowledge. If those ideas be abstract, it will be general knowledge. But, to nlake it real concerning substances, the ideas must be taken from the real existence of things. Whatever simple ideas have been found to co-exist in any substance, these we may with confidence join together again, and so make abstract ideas of substances. For whatever have once had an union in nature, may be united again. (Essay, IV, iv, 12)

The Lockean account of language and relations is thus very much of a piece with that found in Burgersdijck. What we find in Berkeley, however, is a radical change: so far as material things are concerned, substance disappears: As to what philosophers say of subject and mode, that seems very groundless and unintelligible. For instance, in this proposition "a die is hard, extended, and square," they will have it that the word die denotes a subject or substance, distinct from the hardness, extension, and figure which are predicated of it, and in which they exist.

116

FRED WILSON This I cannot comprehend: to me a die seems to be nothing distinct from those things which are tenned its modes or accidents. And, to say a die is hard, extended, and square is not to attribute those qualities to a subject distinct from and supporting them, but only an explication of the meaning of the word die. 7

We have here a new ontology, in which predication no longer represents the relation between a property and a substance but rather the relation between a property and a whole of which it is a part. To say of what is, that it is, is to attribute a property to a whole of which it is indeed a part~ while to say of what is not, that it is, is to attribute a property to a whole of which it is not a part. It is the origins of this radical new ontology that I wish to explore in this essay. In fact, of course, there are several things that are in the background to this amazing innovation in the history of philosophy. What I want to argue is that we can see in the logic of Petrus Ramus one of the things that we can reasonably suppose to have contributed to showing Berkeley the way to this breakthrough. The immediate background is of course Locke. Locke retains a substance ontology for ordinary things. Berkeley exorcises substances in this context. What leads him to see that he can do this? What leads him to see that one need not give predication a special place in accounting for the structure of the world? What I want to suggest is that there is in the logic of Ramus a view of logic and of language, a view of discourse, that carefully avoids tying logic and language to an ontology of substances. Berkeley himself, of course, places his new account of the nature of predication within the context of a view of the world in which properties are the basic characters or words of a natural language, the language of God: Hence it is evident, that those things which, under the notion of a cause co-operating or concurring to the production of effects, are altogether inexplicable, and run us into great absurdities, may be very naturally explained, and have a proper and obvious use assigned them, when they are considered only as marks or signs for our information. And it is the searching after, and endeavouring to understand those signs (this language, if I may so call it) instituted by the author of nature, that ought to be the employment of the natural philosopher, and not the pretending to explain things by corporeal causes; which doctrine seems to have too much estranged the minds of men from that active principle, that supreme and wise spirit, "in whom we live, move, and have our being. (PP66)

Winkler has argued that Berkeley "transforms the natural world from a system of bodies with powers and operations into a system of inert signs-a text-with no existence apart from the spirits who transmit and receive it."g As he puts it a little later, "...experience is a text, authored by God in the language of ideas for the sake of our well-being."9 There is an important sense in which Winkler's characterization of Berkeley's world as a text is true and important.

Berkeley's Metaphysics and Ramist Logic 117 Berkeley's metaphysics did indeed involve a radical re-conceptualization of the universe. But Winkler nonetheless has not got it exactly right either. Already the new science had given up much of Aristotle's metaphysics. In particular, it had given up the unanalysable powers that provided the explanations of ordinary events. Thus, Robert Boyle, in his A Free Enquiry into the Vulgarly Receiv'd Notion of Nature, 10 attacked specifically the appeal to natures or forms as something that at once was no more than an appeal to ignorance-the sceptical arguments made acceptance of any such metaphysical entities unreasonable-and at the same time something that interfered with the progress of experimental science. The "Nature...is so dark and odd a thing, that 'tis hard to know what to make of it, it being scarce, if at all, intelligibly propos'd, by them that lay the most weight upon it." (FE129) At the same time, Boyle "observe[s] divers Phaenomena, which do not agree with the Notion or Representation of Nature..." (FE136) Boyle cites the phenomenon of a vacuum as contrary to many of the things that people have said about Nature as a causal and explanatory force. That is, when one appeals to such an entity, one not only settles for explanations that are bad because they are obscure but settles for explanations that are in fact wrong, misdescribing the phenomena in question. He also cites (FE145-50) the alleged explanations of motions in terms of the qualities gravity and levity. Bodies of the former sort, that is, with that sort of "Innate Appetite," (FE147) move in straight lines towards the centre of the Earth, while bodies of the other sort move in straight lines away from the centre and towards the heavens. Boyle points out how this doctrine makes very little sense with regard to the motion of a pendulum. (FE148-l49) Boyle also argues in detail (sect. viii) that various propositions that are supposed to be established in regard to the notion of "nature," e. g., "nature does nothing in vain," either explain too much or too little, and in fact in general can be understood, when taken as scientific and referring to patterns of behaviour of objects in the world of sense experience, as making assertions compatible with the mechanical philosophy. 11 Or, to take another example, E. Halley argues against those who would explain gravity in terms of "a certain Sympathetical attraction between the Earth and its Parts, whereby they have, as it were, a desire to be united..." This, he says, ".. .is so far from explaining the Modus, that it is little more, than to tell us in other terms, that heavy bodies descend, because they descend. I2 In this respect Winkler is incorrect: the world had lost its active dispositions. The world had becolue inert. However, if the world had lost active forms, and all these features of the traditional Aristotelian metaphysics, it continued to be a world of substances. The new science had not (yet) attacked the substance nletaphysics. Thus, while Locke provided the metaphysical/epistemological framework in which the empirical methodology of the new science was put to work, he nonetheless retains substances as entities that we must hypothesize, though we know not

118 FRED WILSON what they are in themselves. They are simply those things, "we know not what," that underlie and support qualities: "...our idea of substance is...obscure...It is but a supposed I know not what, to support those ideas we call accidents." (Essay, II, xxiii, 15) Although the idea is obscure, we can use it to think about substances as the support of qualities, and to judge that the latter exist is to judge that they are supported by a substance. This much of the traditional doctrine of substance remains. However, if there are substances, what we in fact perceive are collections of qualities: The mind being, as I have declared, furnished with a great number of the simple ideas, conveyed in by the senses, as they are found in exterior things, or by reflection on its own operations, takes notice also, that a certain number of these simple ideas go constantly together; which being presunled to belong to one thing, and words being suited to common apprehensions, and made use of for quick dispatch, are called, so united in one subject, by one name: Which, by inadvertency, we are apt afterward to talk of, and consider as one simple idea, which indeed is a complication of many ideas together; because, as I have said, not imagining how these simple ideas can subsist by themselves, we accustom ourselves to suppose some substratum wherein they do subsist, and from which they do result, which therefore we call substance. (Essay, II, xxiii, 1)

Accidents may well be predicated of an "I know not what" but so far as perception is concerned, what is given to us is simply a collection. For Locke, events are in principle connected by the relation of causality. But in fact we do not know this relation: ...whatever change is observed, the mind must collect a power somewhere able to make that change, as well as a possibility in the thing itself to receive it. But yet, if we will consider it attentively, bodies, by our senses, do not afford us so clear and distinct an idea of active power, as we have from reflection on the operations of our minds. (Essay, II, xxi, 4)

For most qualities of things, there simply are no forms that could provide a connection any stronger than matter-of-fact regularity: the necessary connections are not given to us in ordinary experience; we therefore have no· idea of such connections; and in the absence of such an idea, such necessary connections are simply inconceivable. The reason why the one ["primary qualities"] are ordinarily taken for real qualities, and the other only for bare powers, seems to be, because the ideas we have of distinct colours, sounds, &c. containing nothing at all in them of bulk, figure, or motion, we are not apt to think them the effects of these primary qualities, which appear not, to our senses, to operate in their production; and with which they have not any apparent congruity, or conceivable connexion. (Essay, II, viii, 25)

Berkeley's Metaphysics and Ramist Logic 119 Or, as he puts it elsewhere, the complex idea we have of any substance, "...cannot be the real essence...for then the properties we discover in that body would depend on that complex idea, and be deducible from it, and their necessary connexion with it be known." (Essay, II, viii, 25) There are no perceivable necessary connections among the properties that things are presented as having in the world of ordinary experience. Words that philosophers use that purport to refer to real essences are in fact meaningless; sounds without referents: ...when I am told that some thing besides the figure, size, and posture of the solid parts of that body, is its essence, some thing called substantial form, of that I confess, I have no idea at all, but only of the sound form, which is far enough from an idea of its real essence, or constitution. (Essay, IT, xxxi, 6)

In the absence of real essences, Aristotelian forms or necessary connections, we are reduced to regularity. But this is, after all, all that we need for our practical purposes in our ordinary life in our ordinary world: we seem able to get on in life with just fine with just regularity: Tell a country gentlewoman that the wind is south-west, and the weather lowering, and like to rain, and she will easily understand it is not safe for her to go abroad thin clad, in such a day, after a fever: She clearly sees the probable connexion of all these, viz. south-west wind, and clouds, rain, wetting, taking cold, relapse, and danger of death, without tying them together in those artificial and cumbersome fetters of several syllogisms, that clog and hinder the mind, which proceeds from one part to another quicker and clearer without them; and the probability which she easily perceives in things thus in their native state would be quite lost, if this argument were managed learnedly, and proposed in mode and figure. (Essay, IV, xvii, 4)

So, for the most part, what we have are not casual relations but merely constant conjunctions. In these constant conjunctions, some qualities are in effect signs for other qualities. For Locke there are the metaphysical relations on the one hand-the relations of substance to quality, and of causation-and, on the other hand, there is what we perceive-collections and constant conjunctions. What Berkeley did was reject the doctrine that the world of ordinary things is a world of inert substances. The world of ordinary things is, rather, a world of qualities tied together by various relations. Ordinary things are qualities tied together by the part-whole relation, and predication is a matter of a quality being predicated of the whole of which it is a part. The whole, so far as it is perceived, is a conjunction of qualities. As he put it in his Commonplace Book,13 with specific reference to the relations of space and extension which were supposed by the Cartesians to constitute the essences of things, uniting them into wholes: "We think by the meer act of vision we perceive distance from us, yet we do not, also that we perceive solids yet we do not, also [planes],

120 FRED WILSON yet we do not. Why may I not add? We think we see extension by meer vision, yet we do not." (PC#215) Or, as he puts it right at the beginning of the Principles: It is evident to anyone who takes a survey of the objects of human

knowledge, that they are either ideas actually imprinted on the senses, or else such as are perceived by attending to the passions and operations of the mind, or lastly, ideas formed by help of memory and imagination, either compounding, dividing, or barely representing those originally perceived in the aforesaid ways...And as several of these are observed to accompany each other, they come to be marked by one name, and so to be reputed as one thing. Thus, for example, a certain colour, taste, smell, figure, and consistence having been observed to go together, are accounted one distinct thing, signified by the name apple. Other collections of ideas constitute a stone, a tree, a book, and the like sensible things...(PPl)

Ordinary things are tied together by the relation of causality. But these causal relations are conjunctions. As Locke argued, so did Berkeley: we are not acquainted with necessary connections: All our ideas, sensations, or the things which we perceive, by whatsoever names they may be distinguished, are visibly inactive; there is nothing of power or agency included in them. So that one idea or object of thought camlot produce, or make any alteration in another. To be satisfied of the truth of this, there is nothing else requisite but a bare observation of our ideas. For since they and every part of them exist only in the mind, it follows that there is nothing in them but what is perceived. But whoever shall attend to his ideas, whether of sense or reflection, will not perceive in them any power or activity; there is therefore no such thing contained in them. (PP25)

SO far as concerns the world of ordinary things, causation is constant conjunction: There are certain general laws that run through the whole chain of natural effects: these are learned by the observation and study of nature, and are by men applied as well to the framing artificial things for the use and ornament of life, as to the explaining the various phenomena: which explication consists only in showing the conformity any particular phenomenon hath to the general laws of nature, or which is the same thing, in discovering the uniformity there is in the production of natural effects; as will be evident to whoever shall attend to the several instances, wherein philosophers pretend to account for appearances. (PP62)

For Locke, words are signs of ideas, which are in turn the representatives of entities. For Berkeley, in contrast, words are directly the signs of entities. These entities in turn are the signs of other entities. The connection between words and ideas-as in Locke-or-as in Berkeley-words and entities is conventional~ the connection is not intrinsic but rather is established by human artifice.

Berkeley's Metaphysics and Ramist Logic 121 For Berkeley the same holds for the way in which entities are signs of other entities. There is no intrinsic connection: there are simply conjunctions. The connection is, however, established by artifice: the artifice of the Great Artificer, God. The structure of the world is established by the activity of spirits: We perceive a continual succession of ideas, some are anew excited, others are changed or totally disappear. There is therefore some cause of these ideas whereon they depend, and which produces and changes them. That this cause cannot be any quality or idea or combination of ideas, is clear from the preceding section. It must therefore be a substance; but it has been shown that there is no corporeal or material substance: it remains therefore that the cause of ideas is an incorporeal active substance or spirit. (PP26)

What holds for causal relations also holds for the relations that bind qualities into ordinary things, and, indeed, all other relations. Relations as such are not perceived by sense. As Berkeley put it, "we know and have a notion of relations between things or ideas, which relations are distinct from the ideas or things related, inasmuch as the latter may be perceived by us without our perceiving the former." (FP, 89) Relations are not perceived by sense because they involve activity: "...all relations including an act of the mind, we cannot so properly be said to have an idea, but rather a notion of the relations or habitudes between things." (FP, 142) The structures of the world, including the structures of qualities that constitute ordinary things, are thus a matter of the relating activities of active substances. 14 Qualities are the words of the language of God, and the conjunctions that we perceive and learn are the syntax of God's language. This syntax is of course the relational structure of the world. Thus, the structure of the world is provided by the activity of the deity, or, sometimes, by the activity oflesser spirits. In Burgersdijck's logic, there is a double account of relations, one of relations considered objectively, and one of relations considered as forms of judgment. The former takes them to be predicated of substances. The latter takes them to be forms of mental activity. What Berkeley does with regard to ordinary ("material") things, the objects of ordinary perceptual experience, is radically alter the ontology by eliminating the substance. The role of substance in accounting for the unity of ordinary perceptual objects is replaced by the part-whole relation. Now, this relation, that of part to whole, is one of the relational forms of judgment, alongside causation, etc., that Burgersdijck considers among the topics where arguments are to be found. Thus, in Berkeley, predication is assimilated to the other relations, and loses the special status that it had in Aristotle, Aquinas, Descartes, Locke, Burgersdijck, etc. In Berkeley, predication reflects an act of the mind that has no objective basis in ordinary perceptual objects, or, at least, no more objective basis than any other relation, e. g., the relation of causality;15 predication is, rather, an act of the

122 FRED WILSON mind that unifies sense qualities into wholes. Except, of course, there is in a wayan objective basis for this relation, as for the other structural relations among the sense contents of the world, e. g., the relation of causation: the objective basis is the structuring activities of God. Berkeley in this way as it were re-animates the world. Where Descartes and Locke, and, in general, the new ·science, had removed activity from the world, Berkeley restored it. Berkeley's world as a linguistic text has become once again, as it was for Aristotle, a world that is filled with and moved by activity, the primitive unanalysable activity of substances. But if Berkeley has succeeded in re-filling the world with activity, he has done so only by giving up the substance analysis of ordinary perceptual objects. In this respect he has broken away from the tradition in a 'Nay in which neither Descartes nor Locke nor the scientists of the Royal Society were able to break away from the tradition. One understands the language of God, the natural language of sense contents, provided that the acts in one's own mind mirror the acts by which God structures his language. I get it right, for example, if the acts by which I join up ideas into causal sequences correspond to the way in which God structures the regular connections among sense contents. And I get it right if the acts of unification by which I join the simple sense contents that I receive into things correspond to the acts by which God unifies those ideas. There are no doubt many strands to the story of how Berkeley was able to achieve this break. What I hope to do, as I have said, is trace what I think is one of these strands. It is an important strand, however, because it led to later developments of attempts to account for structure. The strand that I want to exanrine lies in a logical tradition different from that of Burgersdijck. In this other logical tradition, the role of substance is not central. I refer to the account of logic developed by Petrus Ramus, Pierre de la Ramee. Ramus developed certain ideas that had already been presented by Rudolphus Agricola. 16 This was the re-orientation of the traditional doctrine of topical places. In the earlier tradition of logic, one that would be more familiar to us from our own texts of formal logic, one proceeded from concepts through judgments to syllogisms. The discussion of concepts included the discussion of the categories and the predicables, as well as the different sorts of supposition of terms. The discussion then passed on to judgnlents or propositions, which included such things as affirmative and negative, universal and particular and what we call the immediate inferences. The discussion then turned to reasoning, which emphasized the categorical syllogisms of Aristotle, but included hypothetical and disjunctive syllogisms, together with sorites, induction, example, and enthymeme taken as special cases of syllogism. Scientific nlethod as we know it, that is, the method of empirical science, of course had no place here. What followed the discussion of the traditional forms of argument was their use, on the one hand, in science in the sense of scientia, that is, demonstrative science where the premises are self-evident, allA1h~_!lS~---

Berkeley's Metaphysics and Ramist Logic 123 of the forms of argument, on the other hand, in ordinary arguments with only probable premises. It was in the latter context that the doctrine of topical places was located: the places were storehouses for different sorts of premises that could be used in the sorts of discourse that was relevant in legal and political contexts. This organization of the traditional logic can be found, for example, in Robert Sanderson's Logicae Artis Compendium. 17 Sanderson had been appointed Reader in logic at Lincoln College, Oxford in 1608. He was also Regius Professor of Theology prior to the deposition of Charles I. A royalist, he was removed from the chair during the interregnum, only to be reinstated after the restoration. He career ended with a brief episcopate at Lincoln. Sanderson first published his logic in 1618. It is an excellent text, and reasonably proved very popular, subsequently going through many editions. We would recognize most of the questions discussed by Sanderson. What we do not find in our own logics is the orientation towards the standard substance metaphysics deriving form Aristotle. In our own logic texts we would not find a discussion of the categories or predicaments. Nor would we find an orientation of the reference of terms to the members of the first of these categories, substance, which Sanderson defines in the usual way: "Substantia est Ens per se subsistens." (Logicae, 17) Nor would we find the other categories listed, e. g., quality, about which Sanderson asserts that, "Qualitas est forma accidentalis, a qua Substantia denominatur Qualis: ut Albedo. Quale est, quod a Qualitate denonlinative dicitur: ut Album." (Logicae, 21) Nor, finally, would we find these things discussed in a context where it has been laid down that the legitimate forms of predication for the entities in any category are defined by the list of predicables. Then there are relations. We would of course find these discussed in our own logics, but not in the way that Sanderson does, nanlely, in conformity with the Aristotelian metaphysics in which substances and properties are the only entities. Relations are properties in the other categories that are referred one to another: "Veteres Logici Relata definienbant esse, ea, quae altgerius esse dicuntur, aut alio quopian modo ad aliud referentur: ut Scientia et Scibilie." (Logicae, 24) Sanderson classifies relations in much the same way that Burgersdijck does. Where Sanderson differs from Burgersdijck is where he locates the topics. Sanderson locates them in the context of probable, or non-demonstrative syllogisms. They are places where we store premises for arguments for use in legal and political oratory and for sermons. If we wish to advance an argument in the context of such oratory, then we turn to our places and search for an argument that will fulfil our purposes. For Sanderson, the topics have nothing to do with demonstrative syllogisms. Burgersdijck locates the topics much earlier in his discussion, treating them as in effect forms of relational judgment, as we have seen. Burgersdijck retains

124 FRED WILSON the distinction between the topics, on the one hand, where arguments are to be found, and, on the other hand, the categories and predicables, which define the various forms of judgnlent. But at the same time he blurs them, as Sanderson does not for Burgersdijck, the categories and the predicables, while logically different from the topical places, are also, like the topics, places, that is, places where arguments are to be located. Burgersdijck is in effect trying to have it both ways. He wants to defend a logic, which is, like that of Sanderson, Aristotelian in the strong sense of reflecting the Aristotelian ontology of substances as independent entities, and accidents as non-relational entities dependent upon substances. But he also wants a logic that will be practical in its orientation, helping students not only to evaluate propositions for well-formedness and argunlents for validity but to find arguments that they can use in their own discourses. For the latter, the difference between topical places, on the one hand, and the categories and predicables, on the other, tends to disappear, as indeed we have seen happen in Burgersdijck. It is the logic of Ramus that is pulling the resisting Burgersdijck in this direction, away from the Aristotelian logic like that of Sanderson. The logic of Ramus in one way or another incorporates much of the traditional doctrine. But this material is covered only in the context of a logic that is oriented towards its practical uses in the discourse of politics and law. He takes up Agricola's distinction between invention and judgment, but where Agricola dealt only with the subject of invention, Ramus goes on to write a logic that incorporates both aspects. I8 The result is a logic that must have been like a breath of fresh air. Not only does the logic have a clear appeal to practical purposes, compared to a logic such as that of Sanderson, it has the sort of appeal that informal logic has in our own day. But it is also remarkably well-written, with examples that have a wide literary appeal. Sanderson's book is reasonably well-written, but like most logic texts in our own day is rather dull reading. If there are any examples, they are pedestrian. Burgersdijck is even worse: it is not even well-written. Ramus's Dialectique is well-organized for its purposes, well-written, and filled with lively and interesting literary examples that provide useful samples for the poet and orator. For our purposes, however, the important point is that it separates logic from its former role as a language that reflects perspicuously the ontological structure of the world as determined by the substance metaphysics deriving from Aristotle. Ramus aims to give to students a presentation of logic that will be useful for the purposes of legal and political oratory and for sermons. For these purposes, the doctrine of substances need play little role. Thus, when Ramus provides examples of cause, and, in particular, final cause, he cites as explanatory of certain events, first, the end of marriage, and second, the end of taking up arms against Caesar in civil war. In neither case does the subject of the discourse seenl to be a substance. (Dialectique, 64-65) This orator's logic allows for judgments in which we have connections among things that are not substance~__ }Jlp.S-,_ In

_

Berkeley's Metaphysics and Ramist Logic 125 Ramus's logic we are no longer committed to the view that when we say of what is, that it is, that what we are doing is attributing a property to a substance. Ramus, in other words, provides a logic in which discourse is no longer tied to an Aristotelian account of the world. The point of the logic is not to fit an ontology that is no longer substantialist; it is to provide a logic that will be useful to orators. But in producing the latter sort of logic, the result is an account of logic and language that has little place for the traditional doctrine of substance. So the result is in effect an account of logic and language that will fit a world that, like Berkeley's, argues against construing ordinary things as substances. We know that Hume did read Berkeley. Did Berkeley read Ramus? This I do not know. I know of no reference in Berkeley to Ramus. It is clear, however, that Ramus's logic was widely known, as is clear from the history given in Walter 1. Ong's Ramus and Talon Inventory.!9 Logic as a two part art was first clearly laid out in the Dialectique published in Paris in 1555. This was followed by several other editions, including of course the Latin. 20 The first English language edition of Ramus was The Logicke of..Pfeter} Ramus, translated by Rolland M'Kilwien, published in London, 1574?! There was also 1. Piscator's commentary on Ramus, In P. Rami Dialecticam animadversiones (London: H. Bynneman, 1581). William Temple made his name as a Ramist with his P. Rami dialecticae libri duo (Cambridge: Thomas Thomas, 1584). Ong notes that "there seem to be more English translations of Ramus' Dialectica than translations in any other language.,,22 These include editions of 1656, 1658, 1685, and 1699. There were also Latin editions published in England. Thus, there was John Seton's popular Dialectica (annotated by Peter Carter, London, 1611). There was published at Cambridge in 1672 a P. Rami dialecticae libra duo with commentary by Guilielmi Amesii. George Downame's scholastic elaboration of Ramus's logic was published at Cambridge in 1699. But this is not yet to connect Berkeley with Ramus. We know that Oliver Goldsmith attended Trinity College Dublin, and complained of the "cold logic of Burgersdiscius" and of "the dreary subtleties of Smiglecius." Burgersdijck we know about; it was indeed a logic that was influenced by Ramus, though it still retained the crucial Aristotelian notions of substance, categories, and predicables. The Logica of Martin Smiglecki (1638), was essentially Aristotelian, being distinguished only by its mode of presentation, which was in the form of disputations. So Smiglecki and Burgersdijck were used to teach logic at Trinity College Dublin about the time Berkeley was a student. 23 Provost Narcissus Marsh published his Institutiones logicae (Dublin: S. Helsharn, 1681), and this, referred to as the "Provost's logic," was well known at Trinity College Dublin. It is not, however, a Ramist text. It gives the older division of logic into terms, propositions and arguments. In this section there is a discussion of the predicables followed by the usual Aristotelian categories, beginning with substance, which is defined to be "Ens per se Existens."

126 FRED WILSON (Institutiones logicae, 39) In the context of propositions, Marsh includes a section on supposition. The section on argumentation includes a section on demonstration. This is restricted to scientia: "certa et evidens rei cognitio." (Institutiones logicae, 205) It is in this context that Marsh discusses causes. He then turns to dialectical syllogisms, and here he includes the traditional material on the topics. This is decidedly not Ramist. At the time of Berkeley, then, there seems to be little official use of Ramus, and no clearly Ramist text, despite the fact, to which the publication record testifies, of the continued popularity of Ramist texts in English. There is, however, one text that TI1USt be mentioned in this context. This is the already-mentioned annotated edition of Ramus's Dialectica published in 1584 by William Temple. Temple was a tutor in logic at King's College, Cambridge, and made a name for himself as a defender of Ramus against the latter's critics. He responded to criticisms of Ramus by Oxford's Everard Digby (Temple's reply, 1580) and Johannes Piscator (from Strassbourg). The latter (1581) was appended to a second reply to Digby, and was thought to be of sufficient merit that it was reprinted at Frankfurt in 1584. Meanwhile, in 1582 he had written a second reply to Piscator. The point of this discussion of Temple is simply that this logician became fourth Provost of Trinity College Dublin in 1609, and retained the position until his death in 1627. He was an able administrator and left a deep imprint on Trinity College. It is hard to believe that the work of such a vigorous and able defender of Ramus would not have been remembered in the College of which he had been an impressive Provost. Other Provosts were also Ramist, including the seventh, William Chappell. 24 It is more than likely, then, that Ran1ist views were well known at Trinity College Dublin when Berkeley was a student there. None of this is proof that Berkeley ever read or heard of Ramus while at Trinity College Dublin, but leads me to suspect that he did. However, there is one more fact that is relevant. 25 Berkeley was a student at Kilkenny College before going up to Trinity College Dublin, and there is considerable circumstantial evidence that the headmaster at the time, Edward Hinton, was a Ramist. Again, this is no proof that Berkeley had ever encountered Ramus's views on logic and language. But it is hard to conceive that it is not so: the circumstantial evidence from Kilkenny College and Trinity College Dublin seems strongly in its favour. What, however, is in Ramus's logic? The Dialectique begins by defining logic as "art de bien disputer." (61) Its practical orientation is thus established immediately. But Ramus intends this in full generality. Disputes are resolved by discovering the truth. Logic or dialectic is thus the means to discover and display the truth; it is the root of all knowledge: "devons-nous apprendre la Dialectique pour bien disputer a cause qu'elle nous declaire la verite..." (61) Logic does this whether we are considering matters that are scientific, or matters that are contingent, or, what is the same, matters of opinion. Dialectic or logic is "rart de cognoistre, c'est-a-

Berkeley's Metaphysics and Ramist Logic 127 dire Dialectique ou Logique, est une et mesme doctrine pur apercevoir toutes choses." (62) Ramus, following Agricola, divides logic into two parts. The parts of dialectic, or logic, according to Ramus, are invention and judgment. Of these he says that "La premiere declaire les parties separees dont toute sentence est composee. La deuxiesme monstre les manieres et especes de les disposer..." (Dialectique, 63) The first part is described by Ramus as "la doctrine des lieux Topiques," (Dialectique, 63) but insofar as it gives the parts of sentences that are to be admitted into the discourse of reason, of logic, it also covers what would be for the traditional logic the various categories. Indeed, Ramus makes this connection himself. (63) Inartificial arguments derive their worth from authority; the places for inartificial arguments are such things as the law and testimony. (968) But it is the places for artificial arguments that interests us. These are arguments that logic alone helps us construct. What we begin with is a concept. We then wish to discover concepts that can be related to it in propositions which will subsequently be used in syllogisms to bring out in discourse the truth of things. The concept from which we begin and the concept discovered in the topical place will be united to form a proposition. The list of topics therefore gives a list of acceptable sentence forms, the list, in other words, of "well formed" sentences. Although these places occupy in logic the role of Aristotelian categories, together with the predicables, it turns out that neither the traditional categories nor the predicables appear in Ramus's list. What he gives instead are four quite different species of categories (which are themselves subdivided as the discussion proceeds). These are: Causes et Effectz; Subjectz et Adjoinctz; Opposez; Comparez. (64) For our purposes we should note, first, that the list contains only relations. These relations are the forms of acceptable sentences. These relations may be either necessary or contingent; if the former, then we have demonstrative science, and if the latter then we have probabilities. .Ramus indicates that his opposition of subject and adjoint is the same as that of Aristotle who "en plusieurs lieux oppose Ie subjet et I' accident," (Dialectique, 74) though he is also careful to distinguish essential attributes of the subject, where the connection is necessary; and mere accidents, where the connection is contingent. (74) We should therefore note, second, that, where the Aristotelian logic gives a special place to the relation of predication, contrasting it to other relations, Ramus instead includes the relation of subject and adjoint parallel to other relations, including the relation of cause to effect. In this respect, Ramist logic separates itself from all other logics in divorcing itself from the substance metaphysics. At the same time, it provides the form that a logic might reasonably take if predication is to be a relation alongside other relations such as cause and effect, similarity and difference.

128 FRED WILSON Finally, we should note that the apparatus of species and genus also loses its special place in logic and is re-located in the topical place "comparez." (Dialectique, 80) Ramus first discusses "comparison de quantite," (80) and under this head refers to the relations of equal, (80) more, (82) and less. (83) He then turns to "comparison de qualite," (85) and the allocation of things to species and genera is made a matter of the relations of similarity (85) and dissimilarity. (87) In this respect, Ramist logic separates itself from other logics of the age in divorcing itself from a metaphysics in which the relations of species and genus provide the necessary connections among the terms of scientific syllogisms. The Cartesians, especially in the Port Royal Logic, were to make a similar break with the traditional Aristotelian reliance upon species and genus, and in this they were followed by Locke. This leads to similarities between the accounts of inference that we find in Ramus, on the one hand, and in the Cartesians and Locke on the other. At the same time, however, it must be emphasized that both the Cartesians and Locke retained the core of the substance metaphysics: the basic entities are substances. This means that the accounts of logic found in the Cartesians and Locke do not yet make the radical break with the tradition that marks both Ramus's logic and Berkeley's metaphysics. To see this latter point more clearly, we must tum to Ramus's discussion of argumentation or inference. This occurs in the second part of logic. This second part is called "judgment," and is contrasted to invention. Judgment "...monstre les voyes et moyens de bien juger par certain regles de disposition..." (Dialectique, 115) Judgment in tum has three sub-divisions: "Enonciation, Syllogisme, et Method." (115) The traditional logics had three parts, term, proposition and syllogism. "Enonciation" covers roughly the same material as was covered in the traditional logics under the heading of proposition, while "syllogisme" covers the inferences in the traditional way, omitting, however, much of the material of mood and figure and all of the traditional rules. Ramus's view seemed to be that one would pick up the relevant norms more quickly by studying effective examples than by internalizing a set of formal rules whose use depends upon a rather artificial categorization in terms of mood and figure. This disdain for the traditional apparatus was to find itself repeated in the Cartesians and in Locke. "Enonciations" or propositions are divided into simple or complex. The latter include disjunctive, conjunctive and conditional propositions. As examples of simple propositions, Ramus gives (Dialectique, 115): "Le feu brule" "Le feu est chault" "Le feu n'est eau." The first derives from the topical place cause and effect, the second from subject and adjoint, and the third from opposition. Thus, if we have, for example, burning as the object of our interest, then we can locate an argument to be joined to it in a proposition in the topical place cause. This argument is fire, and we form the proposition, "fire bums."

Berkeley's Metaphysics and Ramist Logic

129

Similarly, if we have fire as the object of interest, then we can locate an argument with which to join it in the place adjoint; this argument is hot, and yields the proposition, "fire is hot." The connection may be necessary or contingent. Ramus has pointed out, in his peroration to the discussion in the first of judgment, (Dialectique, 123) that scientific propositions are self-evident. In this case the connection between the terms the proposition will be necessary. Judgnlents that are contingent and not necessary are called "opinion." (124) In this case the connection between the terms of the judgment will only be probable. In general, the places where probable arguments are to be located are those of inartificial arguments, that is, law or testimony divine or human. In the case of necessary judgments, the places to be searched are those of artificial arguments, that is, those of: Causes et Effectz, Subjectz et Adjoinctz, Opposez, Comparez. The justification for asserting the proposition is, in the case of inartificial arguments, the authority, and, in the case of artificial arguments, self-evidence. It should be noted, however, that one can continue to use the loci or places of artificial arguments even if not all connections are necessary or self-evident. It can be held both that causal connections are contingent, and that terms are connected to each other by these relations. The authority will be that of experience, and the tie that will lead the mind from one ternl to another located in the topical place will not be a necessary connection but a psychological habit created by the experience that is the authority that testifies to the truth of the connection. Ranlus next passes on to syllogism: "Syllogisme est disposition par laquelle la question disposee avecques 1'argument est necessariement conclue..." (Dialectique,125) We need a syllogism when we have an "enonciation doubteuse" and "...pour la preuve d'icelle, est besoin de quelque moyen et tiers..." (125) This notion that syllogisms involve the interpolation of third terms between extremes in order to justify the linking of those extremes is commonplace. The extremes in the syllogism All M is P AllSisM All S is P are the subject and predicate of the conclusion. For the Aristotelians, in a scientific syllogism, these are respectively the species and genus, and the middle that links them is the specific difference. In other words, for the Aristotelians, the syllogism exhibits the real definition of the subject. Since understanding a thing involves the grasp of its real definition, to grasp a scientific syllogism and to understand a thing amount to the same: syllogism provides insight into the ontological structure of reality. Once logic is released from the ontology of forms understood in species-genus terms, inference can be based on other relations than these traditional connections. In particular, for

130 FRED WILSON example, as Ramus points out, (Dialectique, 129) Aristotle never treats an argument such as: Octave est heriter de Cesar Je suis Octave Je suis donques heriter de Cesar as a syllogism-science does not deal with singulars. It was called instead an "exposition." (129) But for Ramus, this form is one among the several forms of syllogism. No longer bound by the ontology of forms, relations other than those of genera and species can yield valid arguments. Ramus's point was later taken up by other philosophers. Thus, Descartes distinguishes his own method from that of the "dialecticians": ...when they expound the fonns of the syllogisms, they presuppose that the terms or subject-matter of the syllogisms are known; similarly, we are making it a prerequisite here that the problem under investigation is perfectly understood. But we do not distinguish, as they do, a middle term and two extreme terms. 26 To be sure, his method is like that of the dialecticians in proceeding in terms of self-evident inferences. As he puts it in his Rule Five: The whole method consists entirely in the ordering and arranging of the objects on which we must concentrate our mind's eye if we are to discover some truth. We shall be following this method exactly if we first reduce complicated and obscure propositions step by step to simpler ones, and then, starting with the intuition of the simplest ones of all, try to ascend through the same steps to a knowledge of all the rest. (CSM I: 20) In fact, these inferences are based on relations among ideas: This common idea is carried over from one subject to the other solely by means of a simple comparison, which enables us to state that the thing we are seeking is in this or that respect similar to, or identical with, or equal to, some given thing. Accordingly, in all reasoning it is only by means of comparison that we attain an exact knowledge of the truth. Consider, for example, the inference: all A is B, all B is C, therefore all A is C. In this case the thing sought and the thing given, A and C, are compared with respect to their both being B, etc. But, as we have frequently insisted, the syllogistic forms are of no help in grasping the truth of things. (CSM I: 57) There are relations other than those of species and difference that validate inferences, including the relations that justify A == B, B == C, :. A == C As Descartes puts it, inference rests on a comparison of our ideas, and the discovery therein of relations that connect those ideas. We must: think of all knowledge whatever-save knowledge obtained through simple and pure intuition of a single, solitary thing-as resulting from a comparison between two or more things. In fact the business of human reason consists almost entirely in preparing for this opera-

Berkeley's Metaphysics and Ramist Logic

131

tion. For when the operation is straightforward and simple, we have no need of a technique to help us intuit the truth which the . comparison yields; all we need is the light of nature. (CSM I: 57)

Inferences of the sort just indicated are based on judgments of comparison. These judgments, when linked systematically, yield knowledge of connections among essences or notions or forms. It is the inferences that are essential to knowledge~ the claim that kno"vledge is not obtained until the argument has been put in strict syllogistic form is sinlply not true. What the defenders of syllogistic do, that is, the dialecticians, is take the inferential knowledge and rearrange it into syllogistic form. But that form does nothing to help us to discover what the links are in the chain that leads to the solution of the research problem that we have posed to ourselves: .. ,on the basis of their method, dialecticians are unable to fonnulate a syllogism with a true conclusion unless they are already in possession of the substance of the conclusion, i. e., unless they have previous knowledge of the very truth deduced in the syllogism. ~t is obvious therefore that they themselves can learn nothing new from such forms of reasoning, and hence that ordinary dialectic is of no use whatever to those who wish to investigate the truth of things. Its sole advantage is that it sometimes enables us to explain to others arguments which are already known. It should therefore be transferred from philosophy to rhetoric. (CSM 1: 36-37)

As the Port Royal Logic puts the point,27 "we must discover the content of an argument before we can arrange its prenlises." (Pt. III, ch. 17)28 Locke makes much the same point. The "original way of knowledge" is "by the visible agreement of ideas." (Essay, IV, xvii, 4) "To infer is nothing but, by virtue of one proposition laid down as true, to draw in another as true, i. e., to see or suppose such a connexion of the two ideas of the inferred proposition..." (Essay, IV, xvii, 4) Syllogism presupposes that we have grasped the inferential structure, the connections among our ideas, rather than itself being essential to such knowledge. As we have already noted, any country gentlewoman: ... clearly sees the probable connexion of all [her ideas regarding the weather] ...without tying them together in those artificial and cumbersome fetters of several syllogisms, that clog and hinder the mind, which proceeds from one part to another quicker and clearer without them; and the probability which she easily perceives in things thus in their native state would be quite lost, if this argument were managed learnedly, and proposed in mode and figure. For it very often confounds the connexion: And, I think, every one will perceive in mathematical demonstrations, that the knowledge gained thereby comes shortest and clearest without syllogism. (Essay, IV, xvii, 4)

Ramus is thus followed by Descartes and Locke in giving up the idea that the only relations that can legitimate inferences, syllogisms if you wish, are

132 FRED WILSON those of species and genera. However, while Descartes and Locke abandon the insistence that only species-genera relations give scientific inferences, their notions of logic and language are still bound up with a substance metaphysics. It is precisely this that is also given up by Ramus: connections of qualities to the things of which they are predicated, that is, the relations of adjoints to subjects, do not hold a special place but. are simply relations among relations alongside those of, say, causation. Walter 1. Ong has been highly critical of Ramus's logic,z9 We must distinguish, according to Ong, reasoning, which is "the drawing of consequences from one or more propositions,"3o from understanding. Reasoning is where formal logic has its home. Formal logic constructs its rules with regard to the extensions of terms, but even this abstraction from content is not a total abstraction from matter: its nlovement reflects, and therefore presupposes, an ontological structure in the reality that it aims to illuminate. "Logic is a study of the reflection of [the] material world-the world with which man is directly confronted-in the structures of the mind.,,31 Thus, "the presence of discursive reason in the human intellectual apparatus is due to the material component in man's cognitional make-up and in the make-up of the reality he is faced with.,,32 Reason brings to reality the light of the intellect,. but its role is as it were the transmission of light from one part of reality to another. It cannot do this task of transmitting light unless it has ready for use some already illuminated piece of reality. It is the understanding that provides this initial light. Any knowledge that we have by way of logic, then, presupposes knowledge by way of understanding. The upshot of logical inference is understanding, but that is an understanding that logic can achieve only by taking for granted that understanding has already been achieved. And so, as Ong points out, "Aquinas takes a rather dim view of discursive reason or ratiocination as compared to sheer understanding, which in its pure state would be intuitive.,,33 With this, it is unlikely that anyone would disagree: the inferences of logic transmit truth, they do not justify their premises as true. There must therefore be something, call it the "understanding" if you will, by which we come to know the truth of our premises. Ramus, in particular, would not disagree. Inference is located within the part of logic called judgment, the second part of logic. The premises are, in contrast, discovered by reference to the topical places, which occur in the first part of logic, the part called invention. Ong, however, criticizes Ramus and Ramists for their ".. .insistence that it is reasoning or ratiocination, not understanding, which differentiates men from animals, and on the insistence...that God prescribes 'method', which is conceived of as a kind of protracted ratiocinative process... ,,34 According to Ong, then, Ramus must omit from logic whatever it is that is required if we are to recognize the role of understanding in logic. Since the understanding yields prenlises, that which RanlUS omits when he ignores the understanding must be onlitted from the list of topics. Whatever Ramus puts

Berkeley's Metaphysics and Ramist Logic 133 into his list of topics or categories, it does not include a list of those things that must be there if we are to recognize correctly the role of understanding in grasping the ontological structure of reality. But what is missing? As we have seen, it is precisely the traditional Aristotelian categories and predicables, precisely those things that make the traditional logic of Sanderson or Burgersdijck or Aquinas fit so neatly a world the ontology of which is given by the metaphysics of substances and intelligible forms. Ong's criticism of Ramist logic, then, amounts to a statement that this logic is inappropriate to a world whose ontology is that of the Aristotelian nletaphysics of forms and substances. However, this statement is a criticism only if the ontology of the world is indeed of that sort. Since Ong does not defend his statement, his running stream of contemptuous dismissals of Ramus is little more than a begging of the ontological question. His complaint about Ramus and the "decay of dialogue" is really a complaint with regard to the decay of the Aristotelian metaphysics. But he offers no defence of that metaphysics. In fact, in many respects, the criticism that Ong makes of Ramus applies also to Descartes, Arnauld and Locke, all of whom insist that, while syllogism is the interpolation of middle terms, the logical structure of the relations does not have to be that of real definitions, in terms of species and genus, and, moreover, for Locke, not even a necessary connection. Nonetheless, there remains for Descartes, Arnauld and Locke, the substance metaphysics. Thus, while the Port Royal Logic rejects the apparatus of predicables and real definitions, and most of the traditional categories, it nonetheless includes a categorial scheme based on the Cartesian metaphysics, and therefore presupposes an ontology of substances and properties. That is, it presupposes an ontology in which the relation of predication occupies a special place, different from that of all other relations. It is in precisely this respect that RanlUS develops a view of language that is similar indeed to that of Berkeley. It is a view of language and logic in which predication is treated as the relation of causation is treated rather than given a special place. Given that we take causation to be a relation that is, on the one side, regularity, but, on the other side, a connection established by the activity of the mind, then we should say that same thing about subjects and adjoints. There will be, on the one side, a collection of properties including the adjoint, and, on the other side, a connection of these into a unified whole established by the activity of the mind. The structure of contents into things will parallel the structure of things into causal sequences. But this, of course, is precisely what Berkeley argues. The question that I raised was what enabled Berkeley to nlake the important breakthrough that permitted philosophers for the first tinle to conceive of things as collections of qualities rather than as substances with properties. There are probably many strands that met to account for this radical change in the way of conceptualizing things in one's ontology. But what I anl suggesting is that

134 FRED WILSON Ramus'ss logic is one such strand. Unlike the traditional logic, Ramus did not wed his to the ontology of substances. At the same time he clearly put predication as a relation among relations, and permitted philosophers for the first time to think of predication in ways that differed from the tradition. In particular, predication could be assimilated to the ways in which one thought about causation, or fatherhood, or sacrilege. Berkeley could therefore take Locke's account of the latter relations and apply it to the relation that unified sense contents into things. The result is the new account of predication, the new ontology of things. What I have argued, then, is that the structure of Ramus's logic can reasonably be seen as one of the strands that enabled Berkeley to come to his rejection of the old substance metaphysics. I have little doubt that the influence of Ramus was so pervasive, even where he was rejected, as in Burgersdijck, that it was in fact a contributing factor for Berkeley's great achievement in ending the hegemony of the old Aristotelian ontology that insisted that ordinary things are substances: thanks to Berkeley, but behind him thanks to Ramus, things could now be conceived as bundles of properties. NOTES 1. Aristotle, The Complete Works, revised Oxford translation, ed. 1. Barnes (Princeton: Princeton University Press, 1984). 2. Thomas Aquinas, Summa Theologica, trans. by the Fathers of the Dominican Province (London: Benziger Bros., 1947). References are to "ST" to the First Part (FP) or First Part of the Second Part (FS) or Third Part (TP), by Question (Q), Article (A), Objection (Obj) by number, and, where appropriate, Reply to Objection by number. Present reference is at (ST, First Part, Q 16, Art. 1) 3. Franco Burgersdijck, Monilio Logica, or An Abstract and translation of Burgersdiscius His Logick, by a Gentleman (London: Richard Cumberland, 1697). Hereafter abbreviated as "ML" followed by page number(s). 4. I discuss this in greater detail in F. Wilson, "The History of Relations from Burgersdijck to Bradley," forthcoming in a volume on the ontology of relations, edited by K. Barber. 5. John Locke, Essay concerning Human Understanding, Fourth Edition, ed. A. C. Fraser (Oxford: Oxford University Press, 1894; reprinted New York: Dover, 1959). References are to book, chapter and section in the form "Essay II, xxv,S" to refer to Book II, Chapter 25, Section 5. 6. For a clear exposition of the history of this doctrine of relations from the preSocratics through to Peirce, see 1. Weinberg, Abstraction, Relation and Induction (Madison, Wise.: University of Wisconsin Press, 1964). 7. G. Berkeley, The Principles ofPhilosophy, in his Works, vol. II, ed. A. A. Luce and T. E. Jessop (Edinburgh: Nelson, 1947). References are abbreviated by "PP" followed by numbered paragraph. Present reference is to PP49. 8. Kenneth Winkler, Berkeley: An Interpretation (Oxford: Oxford University Press, 1989), p. 1.

Berkeley's Metaphysics and Ramist Logic

135

9. K. Winkler, Berkeley: An Interpretation (Oxford: Oxford University Press, 1989) p.231. 10. Robert Boyle, A Free Enquiry into the Vulgarly Receiv'd Notion of Nature (London, 1686). Abbreviated by FE, followed by page number(s). 11. Boyle does not exclude the use of final causes in science. Indeed, in his Disquisition about the Final Causes of Natural Things (1688), he allows that the best explanation for certain phenomena is to be found in terms of the final causes for which the Creator intended them. But the teleology here is external, rather than internal, by reference to metaphysical natures. On Boyle's view, the hypothesis of God is an hypothesis about the causes of natural phenomena, and God is a scientific entity alongside atoms and other minute parts and mechanisms to which our experiments and observations enable us to make inferences. 12. E. Halley, "A Discourse concerning Gravity," Philosophical Transactions of the Royal So'ciety, Jan., Feb., 1686, p. 5. 13. Berkeley, Philosophical Commentaries, in his Works, vol. I. References are abbreviated to "PC" followed by numbered entry. 14. I have explored some other aspects of this account of Berkeley on the structure of the world in F. Wilson, "On the Hausmans' 'A New Approach'," in Robert Muehlmann, ed., Berkeley's Metaphysics: Structural, Interpretative, and Critical Essays (University Park, Pennsylvania: Pennsylvania State University Press, 1995) pp. 67-89. 15. Except for the structuring activities of God. 16. Rudolphus Agricola, De inventione dialectica libri tres (Tubingen: M. Niemeyer, 1992). 17. Robert Sanderson's Logicae Artis Compendium, (abbreviated by Logicae, followed by page number(s)), in vol. VI of Robert Sanderson, The Works, ed. William Jacobson, in 6 volumes (Oxford: Oxford University Press, 1854). 18. References will be to Pierre de la Ramee, Dialectique (1555), Edition critique avec introduction, notes et commentaires de Michel Dassonville (Geneve: Librairie Droz, 1964). 19. Walter 1. Ong, Ramus and Talon Inventory (Cambridge, Mass.: Harvard University Press, 1958). 20. W. 1. Ong, Ramus and Talon Inventory, (Cambridge, Mass.: Harvard University Press, 1958) pp.179-186. 21. W. 1. Ong, Ramus and Talon Inventory, (Cambridge, Mass.: Harvard University Press, 1958) p. 182. 22. W. 1. Ong, Ramus and Talon Inventory, (Cambridge, Mass.: Harvard University Press, 1958) p. 184. 23. See Constantia Maxwell, A History of Trinity College Dublin, 1591-1892 (Dublin: Trinity College University Press, 1946) p. 149. 24. See 1. W. Stubbs, The History of the University of Dublin (Dublin: Hodges, Figgis & Co., 1889), p. 146. 25. Professor Steven Daniel has informed me of this fact. 26. R. Descartes, Rules for the Direction ofthe Mind, in vol. I of The Philosophical Works of Descartes, trans. J. Cottingham, R. Stoothoff, and D. Murdoch (Cambridge: Cambridge University Press, 1985), 2 vols. Abbreviated references refer to "CSM" followed by volume number and page number(s). Present reference is to CSM vol. I, page 51 (CSM I: 51).

136

FRED WILSON

27. Antoine Arnauld, Art of Thinking, trans. J. Dickoff and P. James (Indianapolis: Bobbs-Merrill, 1964); La Logique de Port-Royal (Paris: Hachette, 1854). 28. 11 est "necessaire que la matiere soit trouvee pour la disposer..." (p. 210) 29. W. J. Ong, Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958). 30. W. J. Ong, Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958) p. 74. 31. W. J. Ong, Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958) p. 74. 32. W. J. Ong, Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958) pp. 73-74. 33. Ong refers to 1. Peghaire, Intellectus et Ratio selon S. Thomas D 'Aquin (Ottawa: Inst. d'Etudes Medievales, 1936). See his Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958) p. 74. 34. W. 1. Ong, Ramus, Method, and the Decay of Logic (Cambridge, Mass.: Harvard University Press, 1958) p. 74.

Burne and the Logicians CHARLES ECHELBARGER

Part I Due to the luedieval universities' requirement of courses on logic, nearly everyone who received a higher education (until the Renaissance interregnum in logic) absorbed a certain psychological theory of human intellectual powers which was presented in the context of those courses. Aristotle's definition of man as the rational animal came to have a specific connotation due to the familiarity of the psychological theory presupposed in standard logic texts. The human intellect was pictured therein as structured specifically for the task of syllogistic deductive reasoning and as having specific powers corresponding to the main ingredients of syllogistic reasoning. Syllogistic reasoning, accordingly, depends on a specific act of reason, the deducing of a proposition from others. Propositions depend on another sort of act, the act ofjudgment by which simple terms or concepts are combined as subject and predicate. The intellectual faculty directly related to simple concepts was referred to as "the understanding." This psychological theory can be traced back to ancient origins, but the principal figure among these sources is Aristotle. Aristotle conceived of logic as the instrunlent of theoretical science, the perfection of which is scientific demonstration or, to use the scholastic term, scientia. (See Posterior Analytics, Book I, chapters 1-13) Scientific knowledge, according to Aristotle, is of necessarily true propositions concerning that which exists independently of the knower. Since nlany true propositions are not evident to our senses, attempts to establish them as scientific knowledge must be justified by the strongest possible type of reasoning: deductive reasoning from necessarily true premises~ otherwise, the conclusion might be no more than contingently true. In or~er that the chain of reasoning which justifies conclusions be not infinite, there must be some premises which serve as first principles, i. e., those which express real definitions of the species with which the science is concerned. These principles need not be and cannot be justified by any reasoning. Each of the special sciences rests on its own set of first principles. Since only individual substances are presented to sense perception, it is by an intellectual operation that universals come to be in the soul. There must be an intellectual operation by which first principles are conceived through definitions of universals and there nlust be an intellectual operation by which conclusions may be drawn from first principles serving as premises. Each syllogism must contain exactly three terms, one occurring twice in the premises and referring to the cause of the conclusion's truth. The intellect is potentially capable of such a perfect performance. Hence, it must be naturally endowed with powers

138 CHARLES ECHELBARGER by which it can be achieved. Logic itself might be, as Aristotle said, an art rather than a science, but the intellectual powers of the soul on which this art depends can be inferred from its achievements and its errors. We are capable of thinking as well as of perceiving and imagining. We are capable of thinking true as well as false propositions. We are capable of supporting true opinions by means of good reasoning as well as by means of poor reasoning; of syllogistic reasoning in general (valid or invalid) as well as of scientific demonstration and scientific knowledge, which is the full actualization of the intellective soul's potentiality. The intellect must be so constituted as to be capable of these acts. When scholastic logic came under attack by such reformers as the Renaissance humanists, by Bacon and by Descartes, what was called for by most philosophers was not an elimination of logic from the curriculum of higher education but an improvement of its role as an instrument in matters of probability and experimental inquiry. Neither Bacon nor Descartes nor Locke denied the achievements of scholastic logicians in the area of deductive logic, and they conceded the value of syllogistic reasoning for explaining or communicating what one already knows. They hoped for a logic of discovery, which syllogistic logic by itself cannot be. It is significant that perhaps the most widely admired early modern logic textbook, The Art of Thinking l by Pierre Nicole and Antoine Arnauld, continues with nluch the same division of logic in accordance with the old scholastic distinction of three intellectual faculties. It has been shown by Ashworth, 2 that conservatism in the teaching of logic prevailed in English universities well into the seventeenth century, particularly at Oxford. Virulent critic of scholastic logic that he was, even Locke was thoroughly familiar with contemporary neoscholastic logic manuals and even assigned them to be mastered by his own students. Ashworth has also established (Ashworth, 1984) that much of Locke's Essay Concerning Human Understanding is far more intelligible when we read it as, in part, designed to compete with traditional logic books. Indeed, the Essay was often adopted as a logic text in some universities even before Locke's death. Often, it was either assigned instead of or together with The Art of Thinking. Nevertheless, these changes in the way logic was taught and in the texts used for teaching it hardly affected associated presentations of traditional psychological theories concerning the nature of the intellect. Indeed, loyalty to the Aristotelian conception of scientific knowledge and the allied conceptions of the intellect continued well into the seventeenth and eighteenth centuries in Britain. Glanville, in his The Vanity of Dogmatizing, asserted in 1661 that deduction from "foundation propositions" is "the very essentials of our rationality.,,3 In 1696, in his Method to Science, Sergeant claimed that "the deducing evidently new knowledges out of antecedent ones" is part of "our very essence and Rational Nature.,,4 Sergeant also wrote that "...our notions being clear'd, First Principles established, the true form of a syllogism manifested, p~£~E1i~dle_t~r!!1~ fO!l:rrd~ ~lld__

Hume and the Logicians 139 the necessity of the consequence evidenced, all those conclusions may be d.educed with Demonstrative Evidence... ,,5 In recent years, some philosophers6 have compared book I of Rume's Treatise of Human Nature 7 with the Port-Royal Logic, as The Art of Thinking came to be called. Most of these comparisons focus on what there may have been in the Port-Royal treatment of inductive reasoning or its treatment of ideas which influenced Runle. Recent studies of Rume's correspondence during these years indicate that Hume clainled to have had the basic plan 9f the Treatise in mind well before he began to write it. I have done some investigations in the hope of learning what works on logic, other than the Port-Royal Logic, may have influenced Runle's thinking as he wrote the Treatise, especially book I. Since Hume mentions no other logic books in any of his writings, one must be content with circumstantial evidence. Unfortunately, my investigations have not led to results that give nle any confidence as to which specific logic texts (other than the Port-Royal Logic) nlay have directly influenced the young Hume. In the second part of this paper I will explain why, for now, I do not have a persuasive answer this question. That is not to say that I am doubtful as to whether he had any familiarity with any of them. We know that Hume was an academically outstanding student. It would be highly implausible to suppose that he did not even consult library copies of the most widely used and respected neoscholastic logic texts when he studied for the logic course he took as an undergraduate. Anyone who has read the Treatise is familiar with the contenlptuous references Hunle makes to "schoolmen" and "the schools" and "peripatetics" and "school metaphysics." I cannot believe that he based these references or this contempt entirely on hearsay, having had no significant firsthand acquaintance with any of the standard popular neoscholastic logic texts. I am confident, in light of my investigations, that there are some features of Rume's theory of the understanding which are evidently reactions to standard doctrine in logic texts. In the third part of my paper, I will analyze these features and I will try to explain why they are of major importance to Hume. II Two lines of investigation suggest themselves. (1) What were the most popular logic texts in Britain in the eighteenth century? (2) Which logic texts were most commonly used by lecturers on logic at Edinburgh when Hume was an undergraduate there? Attempts to answer these questions are complicated by the fact that the sorts of logic texts assigned were not of a single type, as they had been in earlier centuries. This diversity was directly related to the fact that Scottish universities at this time were revising their curricula. This was especially true at the University of Edinburgh. The main reasons for revision were the relatively recent events of the scientific revolution, the "New Philo-

140 CHARLES ECI-IELBARGER sophy" that was emerging and political events associated with changes in religious climate, especially the ascent of Protestant religion in Britain, specifically Presbyterianism in Scotland. These led to large-scale reforms in university curricula, including the nature of the instruction and type of text to be used in logic courses. Hume had the unusual experience of attending university at a time when debates among faculty, university officials and government officials had caused the nature of the courses, including logic, to be in flux. Some faculty continued to use distinctly Aristotelian scholastic texts. In most cases, they were texts written by contemporary scholars who attempted some awareness of the "new philosophy" and were stylistically somewhat superior to the manuals that originated in the Middle Ages. Switching to one or another of the "modern" logics was also popular. Parts of Locke's Essay were often adopted as a logic text. Others adopted the PortRoyal Logic. Still others may have used Jean Le Clerc's Logica sive Ars Ratiocinandi,8 a text which blends traditional approaches with modern, especially Lockean, epistemological themes. Some used a conlbination of traditional and new approaches. This diversity of instructional approaches was due to the fact that university and government officials simply could not agree, even after years of discussion aimed at creating a uniform set of new courses, on which works on logic ought to be assigned. Thanks to the research of Mossner and other biographers, it is reasonably certain that young David Hume heard lectures on logic in 1723 in the required Semi course on logic and metaphysics at the University of Edinburgh. These lectures were given, in Latin, by a Professor Colin Drummond who also taught Greek and classical literature. I have found and studied some student dictates on the course which were made in the very year that Hume was enrolled in it and the lecturer referred to in the dictates was Professor Drummond. 9 Professor David Fate Norton has also provided me with a copy of the compendium of lecture notes, Compendium Logicae, (dated 1725) from Drummond's logic course of that year. These dictates and this compendium provide good evidence that Hume was not ignorant of the traditional scholastic treatment of logic. Drummond's lectures were typical scholastic treatments of logic. They present standard explanations of such nlatters as the distinction between the three operations of the mind, types of signs, simple and complex terms, concrete and abstract terms, the various types of propositions, species, genera, categories and predicables. I have tried compiling circumstantial evidence as to formative influences on Hume's views on the nature of logic by comparing Drummond's lectures with widely popular neoscholastic texts from this period. The most conspicuous of these were by the Dutch neoscholastic Franco Burgersdijk (1590-1635).10

Hume and the Logicians 141 Hume may have been familiar with texts like Burgersdijk's, but my examination of some of these has not, so far, provided any strong evidence that he would have gained this familiarity through any assignments made by Drummond in such texts in conjunction with his course on logic. Although the overall perspective on the nature and contents of logic in Burgersdijk's text, for example, are similar to those of Drummond's lectures, the style, organization and manner of explanation found in Burgersdijk's text are obviously different from what one finds in Drummond's lectures. One might conlment reproachfully that this result is exactly what one might have expected. One might say that, even if Hume was in the environs of such traditional influences, they must have been minor in comparison to those of Cartesian and Lockean presentations of logic. This line of reproach may seem even more powerful in view of the fact that reforms made in the organization and curriculunl of the University of Edinburgh in 1708 gave powerful impetus to the teaching of such modern authors as Descartes, Newton and Locke. Hume attached great importance to these thinkers and, in his own work, Hume seems to ignore almost completely the subject of traditional logic. Moreover, Hume was a contributing member of a new library which had been founded in 1724 by Janles Steuart, Professor at Edinburgh, and was called the Physiological Library.l1 One had to contribute books or an annual fee in order to be a borrowing member. Besides a list of members, there is a record of the holdings of the Steuart Library during the time that Hume was a member. As one would expect, the record shows that a large part of the holdings were scientific works by such authors as Newton, Huygens and Boyle but there were also some works on logic. On the list, there is of course, the Port-Royal Logic but there is also Isaac Watts's Logick, or The Right Use of Reason (first published in London, 1724). Of the nlodern logic texts which were widely used in eighteenth century Britain, Watts's book had few more popular competitors. Another modern logic text which should be mentioned in this context is 1. P. de Crousasz's A New Treatise of the Art of Thinking. 12 This book was first published in French in 1712. It was so appealing to British readers that it was translated into English and published in 1724. Hume was still an Edinburgh student at that time. Crousasz's book was an attempt to improve upon the PortRoyal Logic by responding to new themes in the works of Locke and Malebranche and by incorporating themes on allied subjects such as epistemology and semantics into logic. Peter Jones argues cogently that Crousasz's book was quite likely known either by Hume or by others with whom Hunle was closely associated. 13 Hume gives no hint in any of his writings that he knew Crousasz's A New Treatise of the Art of Thinking but anyone who was already familiar, as Hume may have been, with the Port-Royal Logic, would have found it readily accessible and very useful. When one compares the contents of Book I of the Treatise with Crousasz's treatments of perceptions, ideas, the acts of the understanding, the imagination, the passions and skepticism, it is

142 CHARLES ECHELBARGER difficult t.o doubt that Hume knew this work, even if his own views on these subjects are often different from those of Crousasz. 14 These facts strongly support the contention that modern logic texts greatly influenced Hume. But these facts must be set beside others. Hume did not begin actually to compose the Treatise until more than ten years after his 1723 undergraduate logic course with Drummond. It is reasonable to assume that he at least knew of the Port-Royal Logic while he was an undergraduate, but there is no clear evidence that he studied it during those years. The university reforms of 1708 had been in the works for a number of years. Before the end of the seventeenth century, a Parliamentary commission was created by the Scottish government to recommend the contents and structures of uniform courses which would be required of all undergraduates. Many texts were vigorously rejected on all sides and the Port-Royal Logic was one of the rejects. There is also considerable evidence that, while Hume was an undergraduate, traditional approaches to logic were still firmly in place in most Scottish Universities. In his letters and journals of 1654, Robert Baillie, principal of the University of Glasgow, writes that the works of Keckerman, Burgersdijk and Scheibler are used in the Scottish universities. He wrote this to the Dutch scholar, Gisbert Voel. As late as 1717, at Glasgow, records show that John Law began logic by teaching Burgersdijk. Library acquisitions lists through the end of the seventeenth century at Edinburgh show that copies of neoscholastic logic texts were still purchased. 15 Students rarely purchased copies of textbooks, relying heavily instead on library holdings for their studies. In light of these facts, it becomes even more understandable that courses such as Drummond's 1723 logic course would have still been scholastic in tone. The tide in logic books and logic instruction during the eighteenth century at Edinburgh may have been turning from Aristotle to Locke via Descartes, but it did not turn all at once. III Two pieces of textual evidence concerning Hume's early acquaintance with works on logic are from Hume's own writings. The first is the footnote in the section of book I of the Treatise dealing with belief. This evidence will be dealt with at length after I have considered the second. The second is what he says in the Abstract about logic and about book I of the Treatise. In his anonymously published piece of self-advertising, An Abstract of a Treatise of Human Nature,16 Hume tried to convey to the public a general sense of the contents of the Treatise by pointing out some similarities between the Treatise and other recently published works that dealt with subjects discussed in the Treatise. According to the author of the Abstract, Book I of the Treatise contains "the Logic of this author." One should note what Hume says in the same place about other contemporary works specifically on logic, works to which he wanted

Hume and the Logicians 143 Book I of the Treatise to be compared. I will quote at length from the Abstract, since the point at issue is crucial. Besides the satisfaction of being acquainted with what most nearly concerns us, it may be safely affinned that almost all the sciences are comprehended in the science of human nature and are dependent on it. The sole end uf logic is to explain the principles and operations of our reasoning faculty, and the nature of ideas; morals and criticism regard our tastes and sentiments; and politics consider men as united in society, and dependent on each other. This treatise, therefore, of human nature seems intended for a system of the sciences. The author has fInished what regards logic and has laid the foundation of the other parts in his account of the passions.

The celebrated Monsieur Leibniz has observed it to be a defect in the common systems of logic that they are very copious when they explain the operations of the understanding in forming demonstrations, but are too concise when they treat of probabilities and those other measures of evidence on which life and action intirely (sic) depend and which are our guides even in most of our philosophical speculations. In this censure, he comprehends the Essay on Human Understanding, Le (sic) Recherche de la Verite (sic) and L'Art de Penser. The author of the Treatise of Human Nature seems to have been sensible of this defect in these philosophers and has endeavored, as much as he can, to supply it. (T646-647) I think we may safely infer that, by the time he began to write the Treatise, Hume had a substantial acquaintance with the "common systems of Logic." From a reading of Book I of the Treatise, it is also clear that he did not believe the only defect of the common systems of logic was that they lacked an adequate treatment of probabilities and other measures of evidence. His footnote to the section on belief straightforwardly attacks all previous systems of logic as having radically misconceived the nature of the understanding itself, as having misconceived its "principles and operations of reasoning." Book I of the Treatise is about the understanding. In this book, there are also discussions of judgment and reasoning. The latter acts do not receive separate treatment outside the book on the understanding. Hume's long footnote to book I's section on belief explains this arrangement: We may here take occasion to observe a very remarkable error, which, being frequently inculcated in the schools, has become a kind of established maxim, and is universally received by all logicians. This error consists of the vulgar division of the acts of the understanding into conception, judgement, and reasoning and in the defInitions we give of them. Conception is defmed to be the simple survey of one or more ideas; Judgement to be the separating or uniting of different ideas; Reasoning to be the separating or uniting of different ideas by the interposition of others which show the relation they bear to each other. But these distinctions and defmitions are faulty in very considerable articles. For, fIrst, 'tis far from

144

CHARLES ECHELBARGER being true that in every judgement which we form we unite two different ideas: since in that proposition God is, or indeed any other which regards existence, the idea of existence is no distinct idea which we unite with that of the object and is capable of forming a compound idea by the union. Secondly, as we can thus form a proposition which contains only one idea, so we may exert our reasoning without employing more than two ideas and without having recourse to a third to serve as a medium betwixt them. We infer a cause immediately from its effect; and this inference is not only a true species of reasoning but the strongest of all others and more convincing than when we interpose another idea to connect the two extremes. What we may in general affirm concerning these three acts of the understanding is that, taking them in a proper light, they all resolve themselves into the first and are nothing but particular ways of conceiving our objects. Whether we consider a single object or several; whether we dwell on these objects or run from them to others; and in whatever form or order we survey them, the act of the mind exceeds not a simple conception; and the only renlarkable difference which occurs on this occasion is when we join belief to the conception, and are perswaded of the truth of what we conceive. (T96-7, n.l, my emphasis)

Note, first, Hume's apparent inlplication that treatments of the understanding by modern logicians are just as guilty of the "remarkable error" as are those of the scholastics who first inculcated it. In fact, one finds such texts as those by Arnauld and Watts treating the understanding in exactly the way that Hume refers to as a "renlarkable error." But one must also compare what is said on this same topic by authors of traditional approaches to logic such as Burgersdijk. In the Monitio Logica, we find Burgersdijk saying such things as the following: Logic is the art of making instruments and therewith directing the understanding in the Knowledge of Things...A Theme is whatsoever may be propos'd to the Understanding to be known...Themes Simple are those which are understood without a Composition or Complexion on Notions, as Man, runs, etc. Compos'd such as are understood by two or more Notions by an Affirmation or Negation join'd together, as Man runs...The judgment of the Mind is said to be true when it composes those things which really are conjoyn'd or divides those things which really are diverse. 17

In the aforementioned student dictate made for Drummond's 1723 logic course at Edinburgh, we read: And now we talk about systematic logic which henceforth can be defined as the art or practical science directing the operation of the mind to truth. Three operations of the mind are laid out. Apprehension, judgment and discursive reasoning. Apprehension is the naked mental representation of something without affirming or denying, as when we consider the sun, or anything else you may wish...but only by contemplating and as if by intuiting.

Hume and the Logicians

145

Judgment is the assertion of the mind about its own ideas compared among themselves as when the mind compares the earth with the· idea of round and either affinns or denies that the earth is round. Discursive reasoning is the operation of the mind by which we elicit one judgment from another, as when we judge that which is best is above all things, loving God is best, and hence we infer that loving God is above all things. 18

Compare these with the corresponding part of Watt's Logick: The Object of Perception is that which is represented in the Idea, that which is the Archetype or Pattern according to which the idea is formed: and thus Judgements, Propositions, Reasonings, and long discourses may all become the objects of perception. Every object of our ideas is called a Theme, whether it be a Being or Not Being...When the mind has got acquaintance with things by framing ideas of them it proceeds to the next Operation, and that is, to compare these ideas together and to join them by Affirmation, or to disj oin them by Negation, according as we find them to agree or disagree. This act of Mind is called Judgment. 19

Then, consider this from the introduction to the Port-Royal Logic: Logic is the art of directing Reason to a knowledge of things for the instruction of both ourselves and others. This art consists in man's reflecting on the mind's four principal operations-conceiving, judging, reasoning and ordering. To conceive a thing is simply to view that thing as it presents itself to the mind...To judge is to join two ideas, affirming or denying the one of the other...To reason is to form one judgment from several others. (PRL 29-30)

Examples of this sort could be multiplied at great length by way of illustrating the point that Runle was quite accurate in taking all logicians, the moderns and those of the Schools as the target of his criticism insofar as he was concerned with the way that their texts portray the faculties and operations of the mind's "purely rational" aspect. The moderns may have rightly turned their attention to giving better treatments of inductive reasoning and probability but, on the fundanlental operations involved in one of the most conceptually basic types of thought and reasoning, syllogistic deduction, they had made little or no change. I take I-Iume to be arguing that, far from being of no great importance in the era of modern logic and the new philosophy, it is precisely in the area of syllogistic deduction where drastic change is most seriously required. Even the best texts of the time leave in place an unexamined, apparently ad hoc, metaphysical picture of the mind as having one part that is specialized for deductive reasoning and another that is specialized for inductive reasoning. If traditional logicians failed to adequately treat probabilistic reasoning, modern logicians as well as traditionalists have failed to accurately portray the workings of the mind where deduction in concerned. I believe Hume concluded that what is called for is a unified conception of the mind's (deductive and

146 CHARLES ECHELBARGER inductive) reasoning capacities and, I believe, that conception is one of the main goals of Hume's new science of Human Nature. In a publication of a few years ago, I have attempted to reconstruct what I believe to be Hunle's theory of the psychological realities of categorical syllogistic reasoning which would be consistent with the agendas of a radically empiricist and naturalistic conception of the mind. 20 If Hume finds the standard tripartite division of conception, judgment and reasoning to be remarkably erroneous, we may find his remedy for the error to be equally remarkable, even if not necessarily erroneous. It involves what at first seems to be his using the same definition of the understanding typical of both traditional scholastic logic and modern logic books. The understanding, he says, is the faculty of conception. But Hume also makes the understanding take over the roles of judgment and reasoning, a change in characterization of the understanding that would have been repugnant to both traditional and modern logicians. What does seenl clear, then, is that, though Hunle retained the old functional characterization of the understanding, he expanded the Understanding's function, thereby simplifying its nature. But there is also a simplification in the nature of the objects of the understanding in Hume's new approach. The object of the understanding whether in simple conception, judgment or reasoning is just one or more ideas, "surveyed" in some "form or order." Ideas qua objects distinct from propositional contexts, ideas qua propositions, and ideas qua arguments are to be thought of as being essentially the same sort of object. In fact, Hume's argument that there are single-idea propositions, propositions not consisting of at least two distinct ideas joined together, suggests that he may have thought that all ideas, even ideas of substances are propositional. We cannot sinlply think of God or of a tree without thinking of theexistence-of-God or of a-tree's-existing. To think of any thing (object or state of affairs) is, as he says, to think of it as existing. Hume says: There is no impression nor idea of any kind of which we have any consciousness or memory that is not conceiv' d as existent;... [therefore] the idea of existence must either be deriv' d from a distinct impression, conjoin'd with every perception or object of our thought, or must be the very same with the idea of perception or object. [There are no] two distinct impressions, which are inseparably conjoin'd. (T66) ...Whoever opposes this, must necessarily point out that distinct impression, from which the idea of entity is deriv' d, and must prove, that this idea is inseparable from every perception we believe to be existent. (T67)

Nor, he says, will it "serve us in any stead" to claim that a distinction of reason may be made here as in other cases where there is no real difference.

Hume and the Logicians

147

That kind of distinction is founded in the different resemblances which the same simple idea may have to several different ideas. But no object can be presented resembling some object with respect to its existence and different from others in the same particular, since every object, that is presented, must necessarily be existent. (T67)

He also argues in the Appendix to the Treatise that, if there were a simple idea such as that of reality or existence, then it could be conjoined to the simple conception of an object and the result of that annexation would be a belief concerning that simple conception. That suggestion he finds absurd since it would mean that "it would be in a man's power to believe what he pleas'd." (T624)

So, while many propositions may contain at least two terms united together by the understanding, the psychologically and epistemologically simplest of propositions will be single ideas of the existence of some kind of individual object and these will not have to be formed by any act of uniting ideas other than the ideas composing the idea of the object itself. Once one has formed the complex idea of an object, one need do nothing more to have an existential proposition in mind. Moreover, it seenlS obvious that another reason why Hume found the traditional notion of judgment erroneous is that, typically, it is said (in both modern and traditional texts) to be either the act of affirming one idea of another or of denying one idea of another, depending on whether the proposition is affirmative or negative in quality. It seen1S to me that one clear reason why Hume protested against the traditional characterization of the understanding in the part of the Treatise that deals with the nature of belief is that the traditional view of judgment makes judgment tantamount to belief or disbelief. Hume's point is probably that believing and disbelieving are mental states having objects or contents which have a definite truth-value. So, the joining or separating of ideas cannot itself be what constitutes believing or disbelieving. Granted, the uniting and separating of ideas are real and indispensable functions of the understanding, but they are not the same as affirmation or denial. One would have to (dis)believe something before there was anything to (dis)believe. The plain fact is that we understand many things that we neither believe nor disbelieve. This point is obvious in the case of propositions which are about matters with which we are not well acquainted. I neither believe nor disbelieve that all senators from Georgia are at least six feet tall, but in order to believe, disbelieve or suspend judgment about the matter, I must at least understand the proposition in question. Insofar as logicians fail to preselVe this point, their conception of the nature of the understanding is erroneous. The standard view of judgment to which Hume refers confuses believing with merely predicating. Indeed, Hume's argument that there are single-idea propositions also implies that propositions do not even have to involve predicating one idea of another distinct idea. Hume's insight into the. nature of the understanding, I submit, is

148 CHARLES ECHELBARGER that the act of the understanding is to form and use that which is capable of bearing a truth-value, i. e., propositions. But, insofar as one does that, one is only conceiving something which is itself either true or false, not judging something to be either true or false. Judgnlent or belief, presupposes both the act of the understanding and the object of the understanding. What the logicians call simple apprehension, judgment and reason are different ways in which the understanding functions either to form or manipulate propositions. Treating arguments as a type of compound idea is not a difficult suggestion. Something sinlilar is quite familiar to contemporary logicians. Every argument may be treated as a conditional proposition formed by conjoining the premises as the antecedent and putting the conclusion as the consequent. In whatever sense a conditional proposition is a single idea, so mayan argument be considered a single idea. In this sense, treating a deductive argument as a proposition also fits well with Hume's related remarks on belief and with what is involved in the appraisal of an argument as valid or invalid. Much as we saw in the case of fornling propositions, we ought not to describe the formation of an argunlent as consisting either of the assertion of the premises or of the conclusion as true. To simply form an argument, we need not believe or assert the premises or the conclusion to be true. We are "simply surveying" two ideas as in a "form" and "order" such that, if the one is true, the other must also be true. To infer a proposition from other propositions, to draw a proposition from others as a conclusion, is a different sort of act which goes beyond the proper business of the understanding. Inferring or concluding is more than mere formation or survey of propositions and their purely logical relations (i. e., as "relations of ideas"). Inferring or concluding, Hume would maintain, involves belief, and thus, the passionate side of our nature. In an earlier work, (see Echelbarger, 1988) I have tried to show how Hume's theory of the understanding enables him to deal with the psychological reality of syllogistic reasoning. His theory of the understanding contrasts sharply with both the traditional and modern logicians's reliance on a theory which represents the intellect as essentially structured with three distinct po\vers, naturally disposed to three distinct types of act. Hume' s alternative to this traditional psychological theory provides him with what he thought necessary to explain how human beings are capable of acquiring the ability for syllogistic reasoning rather than having it in their very nature. A theory of the understanding, for Hume, is not to be identified with a theory of how human beings think syllogistically. It nlust be broad enough and flexible enough to also provide the foundations of an account of how the ability for such thinking as well as a wide variety of other cognitive abilities, including, of course, abilities for inductive and causal reasoning, are acquired by human beings. Since there is only one fundamental act of the understanding, according to Hume, namely conception, the way is clear to develop a psychological theory of the understanding which explains the capacity for syllogistic and other kinds of

Hume and the Logicians 149 reasoning as acquired~ a theory which shows how abilities for conceiving increasingly complex types of propositions develop gradually as experience molds us and the growing records of memory allow the imagination to construct increasingly complex representations of the past, the future and presently unobserved places or objects~ a theory which treats the process of acquiring deductive reasoning capacities in such a way that it nlight plausibly be claimed to take place only in conjunction with the acquisition of language and with socialization. I suspect that it is exactly this sort of theory which the epistemological works of modern philosophers like Hobbes and Locke had encouraged Hume to believe in. Neither traditional nor modern logicians provided a conception of human intellectual faculties and operations which was consistent with this sort of theory.21 Perhaps the authors of standard logic texts, sincerely wanting to satisfy the pedagogical concerns of the academic community, could not have provided anything that would have satisfied Hume, given the kinds of ontological commitments concerning the intellectual and sensitive parts of the soul to which most modernists and traditionalists were still loyal in Hume's day. If I am right, Hume rejected the psychological theory implicit in both traditional and modern logic texts because he had, even as a student, rejected this distinction of intellectual versus sensitive parts of the soul as marking something original in human nature, and his reason for rejecting this ontological commitment was that it is incompatible with his naturalistic agenda for a new science of human nature. NOTES 1. All references to this work are to the Dickoff and James translation, A. Arnauld, and P. Nicole, The Art of Thinking; trans. 1. Dickoff, P. James (Indianapolis, BobbsMerrill, 1964). Abbreviated as "PRL" followed by page number(s). 2. E. 1. Ashworth, "Locke On Language," Canadian Journal of Philosophy, vol. XIV, no. 1, 1984. Abbreviated (Ashworth, 1984). 3. Joseph Glanvill, The Vanity of Dogmatizing (London, 1661; reprinted Sussex: Harvester, Hove 1970, ed. by S. Medcalf) pp. 97-98. 4. John Sergeant, The Method To Science (London, 1696) preface. 5.1. Sergeant, The Method To Science (London, 1696) preface. 6. See Hendel's introductory essay in A. Arnauld and P. Nicole, The Art of Thinking; trans. 1. Dickoff, P. James (Indianapolis, Babbs-Merrill, 1964). 7. All references to Hume's Treatise are to David Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge. 2nd edition text revised and variant readings by P. H. Nidditch (Oxford: Clarendon Press, 1978). Abbreviated hereafer as T followed by page number(s). 8. An edition of LeClerc's Logic was published in London in 1692.

150

CHARLES ECHELBARGER

9. Student dictates on Professor Colin Drummond's Latin lectures on logic, Institutiones Logicae, Ms 3938, National Library of Scotland. Abbreviated in the text as Institutiones Logicae. 10. See especially Franco Burgersdijk, Monitio Logica or, An Abstract... of Logic, trans. by a Gentleman [from the Latin of 1626] (London, 1697). Burgersdijk's texts are not the only examples of popular neoscholastic texts in late seventeenth- and early eighteenth-century British universities. Some others whose works ran into many editions during this period were Phillipe Du Trieu (1614-1748), Bartholomeo Keckermann (1599-1641), Christopher Scheibler (1611-1685) and Martin Smiglecius (1618-1658). See Ashworth (1984) for references on these. Like Burgersdijk, these men were "modem" Aristotelians, meaning that they tried to keep the essentials of Aristotelian logic, ethics, and metaphysics without Aristotelian natural philosophy. They were also Protestants, which made them attractive as textbook authors in late seventeenth and early eighteenth century Scotland. Still another widely used neoscholastic logic text was Robert Sanderson's Artis Logicae Compendium. This text went through many editions. One was published at Oxford in 1696. 11. See Edinburgh University manuscript Ms. De.10. 127. 12.1. -Po de Crousasz, A New Treatise ofthe Art of Thinking (London, 1724). 13. Peter Jones, Hume's Sentiments, Their Ciceronian and French Context (Edinburgh: Edinburgh University Press, 1982). 14. When Hume resided in La Fleche and had begun to write the Treatise, he obtained privileges to borrow books from the library of the Jesuit College there. Given the Cartesian tone of Crousasz's A New Treatise of the Art of Thinking and its evident similarity to the Port-Royal Logic, it is reasonable to assume that the College library would have had a copy of Crousasz's work. If Hume already knew the book in its English translation, study of the French version would have been a useful way of sharpening his skills in French, as task which, he told a friend in correspondence, required some efforts on his part once he settled in France. Another reason Hume would have found the book attractive is that, in it, Crousasz constantly quotes Cicero, one of Hume's favorite authors. 15. See Christine Shepherd-King, Logic Teaching in Seventeenth Century Scottish Universities. Unpublished Ph. D. dissertation, University of Edinburgh, (1974), page 105 and passim. Also, see, John Veitch, "Philosophy In the Scottish Universities," Mind 2 (1877), pp. 74-91. Veitch notes that, even though texts on logic were actually written at the directions of the commissioners by various Scottish universities, the attempt at putting a uniform curriculum into place failed: "After 1701, nothing more is heard of the project; and it had no practical effect on the course of philosophical teaching in the universities." (Veitch, p. 91). In any event, the text on logic" Veitch says, was "based chiefly on the Logic of Port-Royal." (p. 91) Shepherd-King thinks it was more likely to have been patterned after Burgersdijk but it was never used. 16. David Hume, An Abstract oj a Treatise ojHuman Nature, ed. by 1. M. Keynes and P. Sraffa (Hamden: Archon Books, 1965). 17. Franco Burgersdijk, Monitio Logica or, An Abstract... oj Logic, trans. by a Gentleman [from the Latin of 1626] (London, 1697) p. 1. 18. Student dictates on Professor Colin Drummond's Latin lectures on logic, Proemium, Institutiones Logicae, Ms 3938, National Library of Scotland, pp. 3-5. My translation. I have omitted the phrase "nihil deiis determinante" from the original since it seems to make no sense in the context. Some of the words in the original seem to have been

Hume and the Logicians 151 omitted or abbreviated. Some ungrammatical features of the original may be due to haste or imperfect grammatical skill of the author. The actual text reads as follows: "Et horam systenla dicitur Logica quae perinde potest. Ars seu scientia practica mentis operationes. Aprehensio, Judicum et Discursis. Apprehensio est nuda rei alicujus in mente representatione affirmatio aut negatione ut consideramus solem aut aliud quod vis nihil deiis detnninante sed tantum contemplante & quasiintuente. Judicum est sententia mentis de suis ideis inter se comparatis ut cum mens habet idea terreis afffinnat vel negat terreim esse rotundam. Discursis est mentis Operatio qua unum judicum ex allis elicimus ut· cum judicamus id quod optimum est esse supra omnia amandum Deum esse optimum & inde inferimus Deum esse supra omnia amandum." 19. Isaac Watts, Logick or The Right Use ofReason. (London, 1726; reprinted, New York: Garland, 1964) pp. 142-144. 20. Charles Echelbarger, "Hume On Deduction," Philosophy Research Archives, vol. XIII, 1987-88. Abbreviated (Echelbarger, 1988) 21. A possible exception to this generalization is Hobbes. Hobbes might, in some sense, be considered a logician. In his Logica Sive Computatio, he develops a theory of how capacities for reasoning are developed in human beings as a function of experience, socialization and language acquisition. Almost all of Hobbes's psychological theory on this matter is concerned with syllogistic reasoning. In know of no evidence that Hume was directly influenced by Hobbes's views on logic. It is, in any case, implausible to classify Hobbes as a logician in the sense that, say, Burgersdijk or Watts were considered to be logicians. To my knowledge, Hobbes's Logica was never used in any university as a textbook.

Burne on Demonstration DAVID OWEN

I. Introduction It is very conlmon to take Hume's use of the term "demonstration" and its cognates as meaning "deduction." For instance, one recent author takes Hume's contrast of causal inference with demonstration to "prejudge the question of whether causal inference can be recast as sound deductive argument."l Such a reading of "demonstrative" as "deductive" of course does no harm if nothing is meant by "deductive" other than some unspecified mode of reasoning. 2 But ordinarily, I would hazard, when interpreters of I-Iunle talk about "demonstration" as "deduction" they have in mind the standard view of deduction as a forulally valid argument, according to the rules either of syllogism or propositional or predicate logic. So Hume's "demonstrative/probable" contrast becomes our "deductive/inductive" contrast. I have argued elsewhere against the assimilation of the former contrast to the latter and here want only to concentrate on special problems with respect to Hume on demonstration. 3 Let us suppose that when Hume's use of "demonstrative" is glossed as "deductive," those giving this gloss really mean "deductive" in the sense of "formally valid according to the rules of syllogism or lllodern logic." A consequence of this view is that when Hume says, as he frequently does, 4 that there are no demonstrative argunlents with conclusions that are possibly false, he means there are no deductively valid arguments with contingent conclusions. It appears, then, that those who interpret Hume this way "impute to him an error unbelievably gross and often repeated."s In order to avoid saddling Hume with this apparent disastrous Inistake, Stove plausibly suggests that what Hume really meant by "demonstrative arguments" was not "deductively valid arguments simpliciter" but rather "deductively valid arguments with necessarily true (or a priori) true premises." It follows, of course, that such arguments will have necessarily true conclusions. Let us call Hume's claim that if a proposition is possibly false, it is non-demonstrable, criterion 1) of non-demonstrability. On the view of demonstrations as deductively valid arguments with necessarily true conclusions, criterion 1) is saved fronl obvious falsehood. Stove's characterization of demonstrative arguments seems to have beconle the standard view of those unhappy with simply equating demonstration with deduction. 6 It has much to be said for it in general, as it seenlS to be the nlodern equivalent of the scholastic notion of demonstration as a syllogistic argument from axioms, first principles or maxims. Though this interpretation saves Hume's criterion 1) from falsehood, it does not, apparently, save it from triviality.7 Hume states and use~_~~i_s_ ~~~t~~~
_

154 DAVID OWEN of non-demonstrability nlany times. For instance he says "To form a clear idea of any thing, is an undeniable argument for its possibility, and is alone a refutation of any pretended demonstration against it."(T89) And again "whatever we conceive is possible, at least in a metaphysical sense: but wherever a demonstration takes place, the contrary is impossible, and implies a contradiction." (A650) But if by "demonstrable" he means "conclusion of a logically valid argument with necessarily tnle premises," his criterion 1) is the merest logical consequence of this account of demonstration. Any deductively valid argument with necessarily true premises will have a necessarily true conclusion. So if a proposition is not necessarily true, it is not demonstrable. But could Hume's oft-repeated dictum really be such a trivial consequence of what he nleant by demonstration? One possible answer is that Hume did in fact think that his criterion 1) was trivial. What was controversial, it might be held, was his further claim that all matters of fact, especially those concerning cause and effect, were nonnecessary. But it follows from the controversial claim plus the trivial criterion 1) that all nlatters of fact are non-demonstrable. So on this view it would have to be said that this last claim, that all matters of fact are non-demonstrable, was held by Hume to be controversial. There is some textual evidence against this. In III.l.i, he says '''tis allowed on all hands, that no matter of fact is capable of being demonstrated." (T463) This is not decisive, of course, because it may be that Hume simply means, at this late stage of the Treatise, that it is or should be allowed by all those who read Book 1. Another, more likely, story is as follows. Suppose that the criterion 1) of non-dern.onstrability is trivial. That criterion argues from non-possibility to non-demonstrability. What Hunle takes to be substantial and controversial, it might be held, is another criterion: the criterion 2) of possibility: a proposition is possibly false (or true) if it can be conceived to be false (or true). So Hume is really working with two criteria, and usually applies them together. If a proposition is possibly false, it is non -demonstrable~ if it is conceivably false, it is possibly false~ so if it is conceivably false, it is non-demonstrable. Let us call this last criterion, from conceivably false to non-demonstrable, criterion 3) of conceivability. This criterion 3) of conceivability is substantive, and though it presupposes the trivial non-demonstrabilty criterion 1) it is not exhausted by it. The relevant three criteria are: 1) If a proposition is possibly false, it is non-demonstrable. 2) If a proposition is conceivably false, it is possibly false. 3) If a proposition is conceivably false, it is non-demonstrable, where 3) follows from 1) and 2). Now it may well be that our original criterion 1) is indeed trivial, and that Hume was really interested in criterion 3) which rests on the allegedly controversial 2). The textual evidence is inconclusive, though by and large consistent with this view. Hume often cites conceivability

Hume on Demonstration 155 of the contrary as evidence that a proposition is not demonstrable. Consider the following passage from the Abstract: Such an inference, were it possible, would amount to a demonstration, as being founded merely on the comparison of ideas. But no inference from cause to effect amounts to a demonstration. Of which there is this evident proof. The mind can always conceive any effect to follow from any cause, and indeed any event to follow upon another: whatever we conceive is possible, at least in a metaphysical sense: but wherever a demonstration takes place, the contrary is impossible, and implies a contradiction. There is no demonstration, therefore, for any conjunction of cause and effect. And this is a principle, which is generally allowed by philosophers. (A650-51)

When Hume says "this is a principle which is generally allowed by philosophers," it is utterly unclear which principle he means. The principle might mean "whatever we conceive is possible, at least in a metaphysical sense," i. e., criterion 2) of possibility. But this is the allegedly controversial criterion. Why would Hume say it was generally allowed by philosophers? One plausible answer is that it was not controversial, and was arguably held by Descartes, Leibniz and Arnauld. It might mean "wherever a demonstration takes place, the contrary is impossible, and implies a contradiction," an augmented version of criterion 1) of non-demonstrability. Grammatically, it is most likely to mean "There is no demonstration, therefore, for any conjunction of cause and effect." If it does mean the latter, it follows that Hume thought that what apparently was one of his most important positions concerning causation was already "generally allowed by philosophers." It may well be, then, that Stove's account of demonstration in Hume does entail that one of Hume's more famous claims about demonstrability is trivial. But the evidence is unclear, and in any case there are other problems. Consider this. Any argument with a necessarily true conclusion (in the relevant sense of "necessarily") is deductively valid. But any deductively valid argument with necessarily true premises is a demonstration, according to the conception we are considering. It follows then, that any set of necessarily true sentences constitutes a demonstration, as a necessarily true sentence follows from anything. One might object that this point depends on including "tautologous" in the extension of "necessary" and that this would be anachronistic. The point is well made. But anachronism runs deeper in the account of Hume's view of demonstration as a deductively valid argument with necessarily true premises. I want to argue that any account of Hume's notion of demonstration that includes the notion of "deductively valid," where that notion is construed formally either in the modern or the syllogistic sense, is equally anachronistic. Suppose demonstrations are deductively valid arguments with necessarily true premises. We would then expect Hume to at least acknowledge the class of deductive arguments with contingent premises. But Hume rarely talks of deduction and its cognates at all. And where he does, he is using "deduction" in

156

DAVID OWEN

its standard 18th century use of "argument.,,8 To understand the irrelevance of a formal notion of deductive validity to Hume's use of "deduction" and "demonstration," we must look at some historical background. II. The Background The background to understanding Hume's account of demonstration is the rejection of the value of formal syllogistic by the Renaissance humanists, who considered reliance on such a model to be a main reason for the barrenness of scholasticism. Descartes followed them in this rejection of syllogism, and developed the theme by also blanling reliance on syllogism for the lack of scientific progress. The thought was that because of its formal nature, syllogistic reasoning could at best reconstruct arguments for propositions already known to be tme. It could never discover new ones, nor help us in developing a methodology for such discoveries. Descartes developed a new, more naturalistic and psychologically accurate account of reasoning and his development of a method for the discovery of new truths was based on it. Descartes's rejection of formalism and his new account of reasoning based on ideas and their relation to one another was enormously influential. Although its clearest statement is in the posthumously published Regulae, and there is no direct evidence that Locke ever read that work,9 the Cartesian view of reasoning construed non-formally, and the rejection of syllogism, is evident in Locke's work, especially the Essay.IO But to discuss Locke on reasoning and demonstration, we must start with his account of knowledge. Locke's account of intuitive and demonstrative knowledge is well known. For our purposes, it is important to emphasize, firstly, that the crucial relations in both sorts of knowledge are relations among ideas, and not formal relations among propositions,I1 and secondly, that the intuitive knowledge is immediate while demonstrative knowledge involves reasoning. Hume shared both these views with Locke, although unlike Locke he allowed a single idea to have propositional content. But for Locke, since the crucial ingredients for knowledge are ideas and their relations, then it is fair to assume that if we are to understand Locke on reasoning we must look for structures of ideas. It follows, as we shall see, that the notion of formal deductive validity has nothing to do with it. For Locke, knowledge is "the perception of the connexion and agreement, or disagreement and repugnancy, of any of our Ideas." (Essay 4.1.2) Intuitive knowledge is had when "the Mind perceives the Agreement or Disagreement of two Ideas immediately by themselves, without the intervention of any other." (Essay 4.2.1) This intuitive knowledge is immediate: the agreement of ideas is apprehended immediately, without the intervention of any intermediate ideas. Hume uses a similar notion of intuition (e. g., T70). Moreover, such intuition is the basis of demonstrative knowledge as well: "Certainty depends so wholly on

Hume on Demonstration 157 this intuition that in the next degree of Knowledge, which I call Demonstration, this intuition is necessary in all the Connexions of the intermediate Ideas, without which we cannot attain Knowledge and Certainty." (Essay 4.2.1) Sometimes the mind cannot immediately apprehend the agreement or disagreement of any two ideas because "the Mind cannot so bring its Ideas together as by their immediate Comparison and as it were Juxta-position or application one to another." In order "to perceive their Agreement or Disagreement, it is fain, by the Intervention of other Ideas (one or more, as it happens) to discover the Agreement or Disagreement which it searches; and this is that which we call Reasoning." (Essay 4.2.2) Demonstrative reasoning, then is a means of apprehending the agreement or disagreement of two ideas by the intervention of other ideas. These intermediate ideas are called "proofs." Demonstrative reasoning, then, is a chain of ideas, and only derivitively a chain of propositions, and the link between any two ideas in the chain is intuitive: Now, in every step Reason makes in demonstrative Knowledge, there is an intuitive Knowledge of that Agreement or Disagreement it seeks with the next intermediate idea which it uses as a Proof: For if it were it not so, that yet would need a Proof. Since without the Perception of such Agreement or Disagreement, there is no Knowledge produced. (Essay 4.2.7)

This conception of demonstrative reasoning, and indeed reasoning in general, is foreign to both the scholastic/Aristotelian syllogistic reasoning that preceded it and the more modern notion of deductive reasoning that succeeded it. Although syllogistic reasoning is a logic of terms, each piece of syllogistic reasoning can be perspicaciously rendered as a set of propositions. So too with formally valid deductive reasoning. This is because what makes either syllogistic or modern deductive reasoning good reasoning is formal validity, which depends on the structure of the propositions and the overall structure of the piece of reasoning. The content of the propositions is irrelevant: the goodness, i. e., validity, of the reasoning can be displayed in formal terms. Lockean reasoning is quite different. It is a chain of ideas such that the first idea in the chain is seen to be related to the last idea in the chain, not directly, but via intermediate ideas. The link between each adjacent idea is intuitive, and whether two ideas can be so related depends on the content of the ideas, and not any formal structure. Consider Locke's example of demonstrative reasoning in Essay 4.2.2. One wants "to know the Agreement or Disagreement in bigness between the three Angles of a Triangle, and two right ones, . . " but one does not immediately perceive the relation. What is one to do? "In this Case the Mind is fain to find out some other Angles, to which the three Angles of a Triangle have an Equality; and finding those equal to two right ones, comes to know their Equality to two right ones." That is to say, one starts with the idea of the three angles of a triangle, one finds an idea of some other angles intuitively known to be equal to those three angles, and if those other angles are intuitively known to

158 DAVID OWEN be equal to two right angles, one's demonstration is complete. The proposition to be demonstrated is that the three angles of a triangle are equal to two right angles. But one doesn't formally deduce this proposition from premises. Rather, one starts with one idea that makes up part of this proposition (three angles of a triangle) and shows it to stand in the relation of equality to another idea (two right angles) by interposing the relevant intermediate ideas. 12 There is nothing wrong with calling this process a deduction, as Locke sometimes does, as long as one remembers that "deduction" in this context has nothing to do with formal validity. The term is simply serving as a synonym for "argument" or "inference"; no formal structure is implied. III. Hume In section I, problems with construing Hume's account of demonstration as formally valid deductive arguments with necessarily true premises were raised. It was pointed out that there was no evidence that Hume used "deductive" to mean "valid according to formal logic." In section II, it was argued that Locke's account of demonstration, following Descartes, had nothing to do with formal validity but was rather to be construed in terms of how ideas are related to other ideas, via intermediate ideas or "proofs." And the link between any two ideas in the chain of ideas that constitutes a demonstration must be intuitive. I want to suggest that Hume took over from Locke and Descartes this view of demonstrative reasoning as a relation of ideas, and shared with them the rejection of any formal account of reasoning. Intuition is the direct awareness that two ideas stand in a certain relation. And since for Hume ideas can sometimes be propositional, if two such ideas are intuitively related, we will have an intuitive inference. Demonstrative reasoning is the process whereby we beconle aware that one idea stands in a relation to another, not directly but via a chain of one or more intermediate ideas such that the relation between each idea in the chain and its neighbor is intuitively known. And again, if the two ideas that stand in this indirect, demonstrative relation are themselves propositional, we will have a demonstrative inference from one proposition ·to another. But this latter case is just a subclass of the more general characterization of demonstration as showing that two ideas are suitably related. 13 If this is correct, no account of Hume on demonstrative reasoning that characterizes it in terms of deductive validity, where "deductive" is characterized in formal terms, can be correct. It remains true, of course, that if one uses "deduction" only to mean "argument" or "inference," with no formal implications, as Locke and Hume did, then demonstrations are deductive. But that characterization is now completely vacuous, saying merely that demonstrative arguments are arguments. The textual support for almost any substantial account of Hume's view of demonstrations is thin, and the best argument for my view is both its historical

Hume on Demonstration 159 plausibility (I take this case to have already been made out in Section II) and the way it handles problems that would otherwise arise. I will return to this latter strategy shortly. But let us examine what textual evidence there is. Firstly, while considering "propositions, that are prov'd by intuition or demonstration," Hume says "the person who assents, not only conceives the ideas according to the proposition, but is necessarily determin'd to conceive them in that particular manner, either immediately or by the interposition of other ideas." (T95) This seems to be a clear enunciation of the thought that intuition is a matter of seeing immediately that two ideas are related (intuition) or seeing that relation indirectly via "the interposition of other ideas" (demonstration). Hume here contrasts demonstrative with causal reasoning, which in an important footnote on the following two pages is characterized as often functioning "without employing more than two ideas, and without having recourse to a third to serve as a medium betwixt them." The relation between two such ideas in causal reasoning Hume will of course shortly explain in terms of the mechanism of the association of ideas. Secondly, there is the very general point about Hume's frequent use of the locution "relation of ideas" to characterize both the process and the object of intuition and demonstration. The official distinction between "relations of ideas" and "matters of fact," with their respective associated modes of reasoning (denlonstrative and probable), comes in the Enquiry Concerning Human Understanding (EHU 25), but Hume sometimes uses the locution "relations of ideas" as limited to matters of intuition and demonstration in the Treatise as wel1. 14 Intuition and demonstration are well characterized as relations of ideas, because nothing more is needed other than the relevant ideas (two in the case of intuition, more than two in the case of demonstration), to know that they stand in that relation. In the Treatise, instead of the clear distinction between matters of fact and relations of ideas, we get the contrast between two classes of philosophical relations. One class, "Resemblance, contrariety, degrees in quality, and proportions in quantity or number" (T70) are "such as depend entirely on the ideas, which we compare together." (T69) The other class, identity, relations of time and place and causation, are "such as may be chang'd without any change in the ideas." The thinking behind this distinction is that relations of the first class require nothing further than the ideas themselves in order for us to know that they stand in that relation. That is why they can be the objects of intuition and demonstration. But the second class of relations require further ideas or experience. Judgments that a relation of identity holds between two ideas require us to observe that an idea remains the same over time and, as it were, turns into the other idea. Is Judgments that a spatial or temporal relation holds require us to observe the intervening distance, which of course can change without any change in the ideas related. Similarly with causation; merely holding two ideas in our mind is not enough to tell whether or not they stand in the causal relation. Our awareness of the first two of these relations is a matter

160 DAVID OWEN of "perception rather than reasoning" because both objects or ideas "are present to the senses." As a result, "we ought not to receive as reasoning any of the observations we make concerning identity and the relations of time and place; since in none of them the nlind can go beyond what is inlmediately present to the senses" (T73) or memory. 16 Let us return to the four relations of the first class that are susceptible to intuition and demonstration. I-Iume says that "[t]hree of these relations are discoverable at first sight, and fall more properly under the province of intuition than demonstration." (T70) Judgments concerning these three relations, resemblance, contrariety and degrees in quality, "we always pronounce at first sight, without any enquiry or reasoning." (T70) That leaves the relation of proportions in quantity or number as the only relation susceptible of demonstration. The classification of relations of degrees of quality as intuitive seems odd. Why should judgements of degrees of "colour, taste, heat, cold," (T70) seemingly paradigmatic matters of fact, be intuitive and necessary? And so too the limitation of demonstration to relations of proportions of quantity or number. I submit we cannot understand Hume here if we treat demonstration as arguing deductively from necessarily true prenlises. On that view, there can be no constraint on the content of conclusions of demonstrations. Given that Hume allows four sorts of relations to be "objects of kno,vledge and certainty," (T70) no conception of demonstration as valid deduction can plausibly explain Hume's limitation of demonstration to relations of proportions of quantity and number. 17 Why does Hume allow judgements of degrees in quality to be intuitive? When we have several ideas of heat, say, "'tis easy to decide, that any of them is superior or inferior to another, when their difference is considerable. And this decision we always pronounce at first sight, without any enquiry or reasoning." (T70) The claim is, I take it, that when comparing two ideas of heat, say the idea received when a hand is put in water at 50 degrees F as opposed to the idea received when the other hand is put in water at 100 degrees F, it is easy to tell which is hotter than the other. And this judgenlent is intuitive, with the accompanying certainty and necessity. No other ideas are needed, and the relation will remain the same as long as the ideas remain the same. This very idea of heat is hotter than that very idea of heat, and it is inconceivable that the first idea could be colder than the other idea as long as the two ideas rernain the same. Notice what does not follow from this. Hume is not saying that when I judge one object to be warmer than another, this judgement has intuitive certainty. The certainty is limited to the ideas. Nor is he claiming that all such comparison of ideas has intuitive certainty. If the ideas are very close in the relevant degree, two adjacent shades of red on the spectrunl, for example, no such intuitive certainty is forthcoming: "it be impossible to judge exactly of the degree of any quality. . . when the difference betwixt them is very small." (T70) And further, he never suggests that we can -------------------------

Hume on Demonstration 161 judge with intuitive certainty when two colours, say, have exactly the sanle degree of saturation or are of exactly the same hue. Proportions of quantity and number have, at least with respect to "algebra and arithmetic," (T71) a feature that other relations of ideas lack, and this explains why "we can judge of the equality and proportion of nunlbers," (T71) even when the difference between the ideas of two numbers is small. This feature also explains why demonstration is limited to algebra and arithmetic, which are "the only sciences, in which we can carry on a chain of reasoning to any degree of intricacy, and yet preserve a perfect exactness and certainty." (T71) This feature is the availability "of a precise standard, by which we can judge of the equality and proportion of numbers." (T71) The standard is the unit, so that "when two nUlTlbers are so combin'd, so that the one has always an unite answering to every unite of the other, we pronounce them equal." (T71) The availability of the unit as a precise standard makes demonstrations possible in arithmetic. In cases of "very short numbers," such as I + 1 = 2, or where "the difference is very great and remarkable," such as 6728 and 57, we can make intuitive judgements of, respectively, equality and "superiority or inferiority betwixt" (T70) the numbers. "In all other cases we must settle the proportions with sonle liberty [i. e., guess], or proceed in a more artificial manner [i. e., reason]." (T70)18 The availability of the unit as a precise standard allows the "artificial" technique of demonstrative reasoning to occur. Suppose we want a demonstration of the equality of 3467 with 2895 + 572. It is not intuitively certain to most people, nor do we "perceive an impossibility of falling into any considerable error." (T70) There is of course a formally valid deductive argunlent with the proposition "3467 = 2895 + 572" as its conclusion, but that is not the way Hume thought (nor is it the way) we reason. Rather, the idea of 3467 is seen to stand in the relation of equality to the idea of 2895 + 572, because we can match every unit in 3467 with every unit in 2895 + 572. This matching is a matter of providing the relevant intermediate ideas. The complete chain of ideas would be something like this. We can intuitively judge that 3467 is equal to 3466 + 1, which is in turn equal to 3465 + 2. This chain ends with the intuitive judgement that 2896 + 571 is equal to 2895 + 572. But because we have a precise standard, we are assured that this "chain of reasoning" preserves "a perfect exactness and certainty," and thus we have demonstrative certaint'"j that 3467 = 2895 + 572. Geometers think that a similar "precise standard" is available in geometry: the infinitesimal point. Hume thought such a geometrical standard was inconceivable; hence the lack of demonstrations in geometry.19 The lack of a precise standard also explains why the relations of resemblance and degrees in quality do not admit of demonstration. There is nothing in our ideas of colour and heat that provides for such a standard. We try to remedy this by the use of instruments and scales of measurement; that is, we use instruments to map degrees of quality onto numbers, which are then

162 DAVID OWEN susceptible of demonstration. But reliance on the mapping is very much the result of probable reasoning. 20 IV. An Alternative I have outlined a non-formal interpretation of Hume's account of demonstration, derived from Locke and based on perceiving, directly or indirectly, the relation of ideas. Hume' s account is looser than Locke's, as· he allows some ideas to be propositional in nature. That some ideas are propositional in nature is required, given Hume's account of belief: Thus when we affirm, that God is existent, we simply fonn the idea of such a being, as he is represented to us.... When I think of God, when I think of him as existent, and when I believe him to be existent, my idea of him neither increases nor diminishes. (T94)

Many of Hume' s examples of demonstration concern the denial of the possibility of the demonstration of existential propositions, or matters of fact, so it will not be surprising if in those instances he speaks of demonstrating, or rather the impossibility of demonstrating, that two propositional ideas stand in the relevant relation. But concerning demonstrations that Hume thinks can actually occur, i. e., those concerning quantity or number, there is no need to consider the relevant ideas as propositional in nature. Rather, propositions concerning quantity or nunlber are better thought of as being made up of at least two ideas, and demonstrating the proposition is a matter of finding the relevant intermediate ideas that are "interposed." The latter interpretation certainly makes more sense of such mathematical demonstrations as we considered in the last section. It would be odd to treat "3467 = 2895 + 572" as a propositional idea rather than as a proposition constructed from the comparison of the idea of 3467 with the idea of 2895 + 572 in respect to equality?l If I am right, Humean demonstration is largely a matter of content, rather than form. Since the real "work" in demonstration is to be explained in terms of ideas and their relations to one another, this is hardly surprising. Apart from its relation to its corresponding impression, there is little to say about an idea apart from its content. 22 I have claimed that Hume's account of demonstration does not rely on any formal notion of deduction. Formal accounts of deductive validity are usually considered to be syntactic, but when we teach formal logic, we often introduce the notion of deductive validity, not syntactically, but in a semantic and modal fashion, in some formulation such as the following: An argument is deductively valid just in case it is not possible for the premises to be true and the conclusion/alse. Now, if Hume' s demonstrations are a matter of content, that really amounts to calling them semantic. So one might argue that I have been too swift in denying that Humean demonstrations have anything to do with deductive validity~ In ruling out the relevance of fornlal validity, I have left open the possibility that Hunlean demonstrations might rely on a semantic and modal

Hume on Demonstration 163 notion of validity, as outlined above. Moreover, it might be said, Hume has the resources for such an account, as the conceivability criterion for nondemonstrability shows, and in fact appears to rely on such a notion when he denies that Adam, prior to experience, could denl0nstratively infer from the motion of one billiard ball the subsequent motion of a second ball. 23 In response to this, the first thing to notice is that though Hume is happy to talk about demonstrative arguments (e. g., at T127, and TI72), when he gets down to details, he almost always speaks of whether a proposition is demonstrable simpliciter, and not whether it is demonstrable given that other propositions are true. That is the point of the conceivability criterion. We can .tell whether a proposition is demonstrable or not just by considering whether its opposite is conceivable or implies a contradiction. That is to say, with an important exception which we will consider in a moment, Hume never considers whether a proposition is demonstrable, given that something else is true, but only whether it is demonstrable simpliciter. So no account of demonstration in terms of deductive validity, even informally conceived, will adequately account for this feature of his account. But the account of demonstration in ternlS of the interposition of intermediate ideas does account for this. That is one of its strengths. The point is that any conception of deductive validity isolates a feature of inference that makes it necessary, irrespective of the necessity or otherwise of the propositions involved in the inference. But Hume's account of demonstrative inference, like Locke's, locates the necessity entirely in the relation two ideas bear to one another. "3467 = 2895 + 572" is a demonstrable proposition because one idea, 3467, can be shown to stand in the relevant relation to another idea, 2895 + 572, by interposing other ideas the first of which will be intuitively related to 3467 and the last of which will be intuitively related to 2895 + 572. The necessity of the demonstrative inference is entirely exhausted by the relation intuitively perceived to hold between adjacent ideas in the chain. Now let us consider the apparent exceptions. There are instances in all of the Treatise, the Abstract, and the Enquiry Concerning Human Understanding, and most significantly, they occur at identical places in the argumentation. When Hume begins to investigate the nature of causal or probable inference, he first considers whether, prior to experience, a person, upon observing one event, might reason demonstratively to predict the occurrence of another event caused by the first. It is during this consideration that the relevant examples occur. After denying that such reasoning is possible in these circumstances, Hume goes on to consider how we might reason after experience. At this point we get into what has been called the argument about induction, a topic beyond the scope of this paper. 24 Here are the examples, taken from Hume's discussion of what might happen prior to experience. 1) There is no object, which implies the existence of any other if we consider these objects in themselves, and never look beyond the ideas which we fonn of them. Such an inference wou'd amount to

164

DAVID OWEN knowledge, and wou'd imply the absolute contradiction and impossibility of conceiving any thing different. (T86-87) 2) Were a man, such as Adam, created in the full vigour of understanding, without experience, he would never be able to infer motion in the second ball from the motion and impulse of the first. ...Such an inference, were it possible, would amount to a demonstration, as being founded merely on the comparison of ideas. But no inference from cause to effect amounts to a demonstration. Of which there is this evident proof ...(A650) 3) We fancy, that were we brought on a sudden into this world, we could at first have inferred that one Billiard-ball would communicate motion to another upon impulse.... [But] The mind can never possibly find the effect in the supposed cause, by the most accurate scrutiny and examination and examination. For the effect is totally different· from the cause, and consequently can never be discovered in it. (EHU28-29)

In each of these cases, we have Hume considering the possibility of inference, prior to experience, from an observed event to its alleged effect, prior to the effect actually occurring and being observed. Since this occurs prior to any experience (except the experience of the alleged cause), the only sort of reasoning possible is demonstrative. Since we only have an idea of the alleged cause, the idea of the alleged effect must "be discovered in" the idea of the alleged cause, and if this were possible, the relation between the two ideas would be intuitive or demonstrative. And even if we did have an idea of the alleged effect, any inference to it from the idea of the alleged cause must be based on "the comparison of ideas," not experience, and hence be intuitive or deIllonstrative. But we can certainly conceive the one to occur without the other, so any such demonstrative inference is impossible. If one is wedded to the notion of deductive validity, whether formal or not, it is certainly tempting to reconstruct Hume's reasoning in these examples as follows. Upon observing the motion of one billiard ball, I may certainly conceive that its attendant effect will not occur: [M]ay I not conceive, that a hundred different events might as well follow from that cause? May not both these balls remain at absolute rest? May not the first ball return in a straight line, or leap off from the second in any line or direction? All these suppositions are consistent and conceivable. (EHU30-31) Consider these two propositions: A) Billiard ball 1 is moving towards billiard ball 2, and makes contact. B) Billiard ball 2 moves off at a proportionate angle. So it looks as if Hume is saying that since conception implies possibility, it is possible for proposition A to be true and B to be false. But this is just our semantic and modal notion of deductive validity. So when Hume is denying that such inferences are demonstrative, he is really denying that they are deductively valid in that sense.

Hume on Demonstration 165 This tempting line of thought runs into the same problems that pronlpted Stove to argue that by "demonstration," Hume meant, not "deduction simpliciter," but "deduction from necessary or a priori true premises." Suppose "demonstration" means "deduction simpliciter." Then we can make sense of the above passages as being concerned with the alleged demonstrative (i. e., deductive) inference from a proposition about the existence of one object to a proposition about the existence of another. But the cost of this interpretation has been pointed out by Stove: given Hume's conceivability criterion, Hume must be guilty of committing the howler of believing that there are no deductively valid arguments with contingent conclusions. Now suppose "demonstrative" means "deduction from necessarily true premises." Then it is utterly unclear what Hume is up to in the above passages. All he has to do is point out that the existence of either object is contingent, and that is the end of the matter. But that doesn't seem to be his concern: he certainly appears to be concerned with the inference from one proposition to another. The only remaining alternative, if we stick with an account of demonstration that involves deduction, is to say that Hume is inconsistent~ sometimes he requires necessary prenlises, and sometimes he doesn't. We can do better. Consider "idea 1" and "idea 2." According to my interpretation of Humean denlonstration, the proposition formed from the relation of idea 1 to idea 2 is demonstrable if other ideas can be interposed between them in a chain such that each idea is linked to its neighbour in a way that we can intuitively perceive. Furthermore, Hume thinks he has come up with a test that will tell us in advance of trying whether such a chain is possible. If we can conceive of one idea without the other, no such chain is possible, and we know in advance that the proposition formed from the two ideas is not demonstrable. Usually, ideas themselves are not propositional in nature, but go to make up propositions. But in questions concerning existence, an idea, of God or of a certain sort of event, is itself propositional. When we conceive of God's existing, as opposed to just thinking about God, we "make no addition to or alteration on our first idea." (T94) Hunle's theory of belief is not the subject of our present concerns (though it is clear it is crucial for a thorough study of probable or causal inference), except insofar as it compels us to see that ideas of existent entities are propositional for Hume. So if idea I is an idea of the alleged cause, while idea 2 is an idea of the alleged effect, and Hume wants to raise the question of whether there is a demonstrable relation between them, he does it in just the way he discusses all other demonstrations: can we can conceive of one idea without the other? If so, no demonstration is possible. Since it is an inference between two propositions, we are tempted to see it as a question of deductive validity, but then we misunderstand Hume and are compelled to ascribe either stupidity or inconsistency to him. But Hume looks at it in just the same way as he looks at all other cases of demonstration: is one idea suitably related to another? It is

166 DAVID OWEN just that in this case, the ideas, since they involve existence, are propositional in nature. Once one sees that, for Hume, all questions of demonstration are questions of whether one idea is suitably related to another, irrespective of whether these ideas are propositional or not, it becomes obvious how Hume would treat that class of cases we would describe as deductively valid arguments with contingent conclusions. Such an argument would be reconstructed by Hume in the following way: consider the complex idea made up of the premises, and the idea nlade up of the conclusion. Are these ideas related so that there can be a demonstrative connection between them? We should note that Hume is committed to saying that unless the connection is immediately seen (and so the relation is more properly called intuitive than demonstrative), we know in advance that the connection can't be demonstrative unless we are dealing with relations of quantity or number. I conclude then that even an informal account of deduction is misplaced as an ingredient in Hume's account of demonstration. But consideration of it has at least allowed us to properly consider several troublesome examples, and show how my interpretation can handle them in a way consistent with Hume's overall account of demonstration. And, it seems to me, it has also allowed us to show that no account of demonstration in ternlS of deductive validity can account for such cases, or at least account for them in a way that is consistent with Hume's other uses and examples of demonstrative reasoning.

v. Other Problems My interpretation of Hume's account of demonstration as involving the awareness of two ideas standing in a certain relation via one or more intermediate ideas is historically plausible, and allows us to make sense of his troublesome classification of philosophical relations. Since being aware that ideas stand in a certain relation involves being aware of their content, no formal account of demonstration can accurately reflect Hume's view. This interpretation allows us to understand why Hume limits demonstration to arithmetic and algebra~ since his account of demonstration depends on the content of ideas, there is no reason in principle why demonstration shouldn't be limited to matters with a certain content. Passmore reports having difficulty persuading his colleagues that Hume limited syllogisms to mathematics. 25 If syllogisms are formally construed, as they must be, his difficulty is not surprising. But if we drop talk of syllogisms, as Runle did, and speak instead of demonstrations construed in terms of relations of ideas, Hume' s restriction has far more plausibility. I now want to briefly discuss what other problems this account of Humean demonstrations can help us solve. 1) The Demonstrative/Probable Contrast: If I am right about demonstration, equating Hume's demonstrative/probable contrast with our deductive/

-----------------------------------

-

--

Hume on Demonstration 167 inductive contrast is a non-starter. But the account I have argued against is more subtle than that. According to Stove, Hume's demonstrative/probable contrast is to be cashed out as follows. Demonstrative arguments are deductively valid arguments with necessarily true premises; probable arguments are deductively valid arguments with at least one contingent premise. According to this view, the reason that probable arguments are unreasonable, according to Hume, is that a certain premise (the uniformity principle) is not available to us. This view makes it virtually impossible both to understand many of the details of Hume's negative arguments about probable reasoning and to reconstruct Hume's own positive account. In order to understand Hume on probable arguments (arguably the most discussed feature of his philosophy), one must understand his demonstrative/probable contrast. If I am right about Hume and demonstration, then another way must be found to understand his views on probable arguments. One plausible start is this: just as demonstrative arguments are chains of ideas, so probable arguments are chains of ideas. But each such chain in a probable argument, if it be based on reason, will require a link between an idea about what has happened in the past and an idea about will happen in the future. Hence the need for the Principle of the Uniformity of Nature. But no such principle is available to us. Another way must be found to groilnd probable arguments, and Hume gives us such an account in terms of the association of ideas. 26 2) Conceivability and Possibility: In Section I, we discussed the nondemonstrability criterion 1) and the criterion of possibility 2). The former cites possible falseness as a grounds for non-demonstrability; the latter cites conceivable falseness as a grounds for possible falseness. An account of demonstrations as deductively valid arguments with necessarily true premises makes sense of criterion 1); if a conclusion is possibly false, it is not necessary and hence can't follow from necessarily true premises. But it leaves criterion 2), and consequently the combined criterion 3), completely unmotivated. Why should conceivability be a criterion of possibility and hence non-demonstrability? This has rightly troubled commentators. 27 But if Hume's account of intuition and demonstration is as I have sketched it, criteria 2) and 3) are perfectly natural extensions of the account. If two ideas are seen to stand in a certain relation either directly or via a chain of intermediate ideas, then they must stand in that relation as long as the ideas remain the same. That is how Hume characterizes philosophical relations that can be known either intuitively or demonstratively. (T69) But then it is inconceivable that the ideas could not stand in that relation; if they were not so related they would not remain the same ideas. It is not possible for them not to stand in that relation while remaining the ideas they are. Hume's criteria 2) and 3) seem baftling and misguided if we characterize demonstration formally, but they are obvious and well-motivated if we correctly understand his account of demonstration.

DAVID OWEN 3) Implying a Contradiction: Hume occasionally says things such as "wherever a demonstration takes place, the contrary is impossible, and implies a contradiction." (A650)28 As far as I can tell, he never says that the contrary is a contradiction, but always that it implies one. 29 The force of "implies" here is difficult to unravel. Hume held the demonstrative/probable contrast to be exhaustive,30 and I am sure all sides would admit that he could not have meant implication here to be taken as probable inference. But nor could it be demonstrative inference, if demonstration is taken to be a deductively valid argument from necessarily true premises. A contradiction can't deductively follow from necessarily true premises. This problenl has remained unnoticed, I suppose, because while in the grip of the notion of a deductively valid argument, we have just taken it for granted that Hume meant something like what we learn in first year logic. If you take the conclusion of a valid argument, negate it, and add it to the original premises, you can derive a contradiction. But since the original argument is valid, then either its conclusion is true, or at least one of its premises is false. In either case, at least one of the premises of the derivative argument, vvhose conclusion is a contradiction, will be false. So it is not possible for Hume to mean what we might have taken him to mean, if a demonstration is a deductively valid argunlent with necessarily true premises. The point is that the denial of an intuitive or demonstrative truth involves a change of one or both of the ideas so related. So one is holding that an idea both is and is not something: that is the contradiction. It helps to remember that Hume classified contrariety as one of the four relations "which depending solely upon ideas, can be the objects of knowledge and certainty," (T70) that is, can be known either intuitively or demonstratively. Hume appears to say (T70) that contrariety is more a nlatter of intuition than demonstration. Demonstration is limited to algebra and arithmetic, "the only sciences, in which we can carry on a chain of reasoning to any degree of intricacy, and yet preserve a perfect exactness and certainty." (T71) Since we have a "precise standard," the unit, that makes this possible, there should be no reason why we couldn't have mathematical demonstrations that result in contradictions as conclusions. But there appears to be a problem, which we can put in terms of the contrapositive of Criterion 3): if a proposition is demonstrable, it is not conceivably false. But if the denial of a demonstrative truth implies a contradiction, surely we must be able to at least conceive of its being false, if only in order to show that it leads to a contradiction. If we couldn't even conceive of its being false, then presumably the contradiction would be on the surface, and not merely implied. That is to say, if Hume's criterion 3) is correct, it looks as if any hypothetical reasoning from the denial of a demonstratively true proposition is at best otiose and at worst impossible. 31 Hume has a way out, and it turns on the relationship between a proposition's being inconceivable and appearing inconceivable. Hume is happy to admit that the latter entails the former: "The plain consequence is, that whatever appears impos168

Hume on Demonstration 169 sible and contradictory upon the comparison of these ideas, must be really impossible and contradictory, without any farther excuse or evasion." (T29)32 But nowhere, as far as I can tell, does Hume commit himself to the claim that being inconceivable entails appearing inconceivable. Hume says that the proposition " ... the square of the hypothenuse is equal to the squares of the other two sides, cannot be known, let the terms be ever so exactly defined, without a train of reasoning and enquiry." (EHU163). If the proposition cannot be known without reasoning, its negation cannot be known to be false without reasoning. Thus both the proposition and its denial must at least appear to be conceivable, pending the results of demonstrations. Shortly after the passage just quoted, Hume says: "[in] the sciences, properly so called [i. e., the sciences of quantity and number], [e]very proposition, which is not true, is there confused and unintelligible. That the cube root of 64 is equal to the half of 10, is a false proposition, and can never be distinctly conceived." (EHU164, emphasis mine) A complex proposition, at least in the sciences of quantity or number, mayor may not be demonstratively true. Until its truth or falsity is demonstrated, either the proposition or its negation may appear to be conceivable. But it will turn out that at least one of these alternatives really is inconceivable. 33 We have already reconstructed a Humean demonstration of the equality of 3467 with 2895 + 572. How could Hume construct a demonstration to show that the contrary implies a contradiction? Suppose that 3467 is not equal to 2895 + 572. Then, as a matter of intuition, 3467 is not equal to 2896 + 571. By a series of intuitive steps, it follows that 3467 is not equal to 3466 + 1, and finally that 3467 is not equal to 3467. But we know that intuitively to be a contradiction. So the contrary of the equality of3467 with 2895 + 572 implies a contradiction, and we can have demonstrative knowledge that the equality holds. Since no other fields have the "precise standard" required for demonstration, such demonstrations of contradictions are limited to algebra and arithmetic. In any other area, a contradiction is known either intuitively or not at all. There can be no demonstrations of matters of fact. 4) Demonstration and Intuition: Hume carefully distinguishes between intuition and demonstration at T69-73 (and the similar distinction between perception and probable reasoning at T73-74). Demonstration involves reasoning, while intuition does not. But Hume frequently talks of intuition and demonstration in the same breath. For instance, he claims to have proved "that the necessity of a cause to every beginning of existence is not founded on any arguments either demonstrative or intuitive." (TI72) It is odd to describe some arguments as intuitive, as the point of distinguishing intuition from demonstration was to distinguish those things known directly with certainty from those that required argument or reasoning. And when arguing against certain rationalist moralists, he says "If you assert, that vice and virtue consist in relations susceptible of certainty and demonstration, you must confine yourself

170 DAVID OWEN to those four relations, which alone admit of that degree of evidence." (T463) But Hume had explicitly limited demonstrations to only one of those four relations; the others were susceptible to intuition only. These problems turn out to be trivial. If demonstrations consist of chains of intuitions, each step in the chain can loosely be called an intuitive argument. In the first Enquiry, he even speaks of one proposition being intuitively inferred from another: "You say that the one proposition is an inference from the other. But you must confess, that the inference is not intuitive; neither is it demonstrative." (EHU37) Such an intuitive inference would be one that requires no intermediate steps. For instance, in our earlier example, we could intuitively infer that 3467 is not equal to 2896 + 571 from our supposition that 3467 is not equal to 2895 + 572. But it took many steps to reach the contradiction that 3467 is not equal to 3467, so that equality of 3467 with 2895 + 572 is known demonstratively, not intuitively. As to the problem of limiting yourself to the four relations if you want to show that morality is susceptible to demonstration, what Hume says is strictly speaking true. If you want to show that morality is demonstrative, you must first show that it involves only those relations that have the relevant degree of certainty, i. e., those that can't be changed without any change in the ideas. But once you have done this, you have the further task of showing that for each of the relations you need, there is a precise enough standard to allow demonstrations as well as intuitions. Hume thinks he has shown in advance that this cannot be done, and doesn't need to rehearse the arguments again here. Hume's arguments against the possibility of a demonstrative morality in Book III of the Treatise are not then simply a matter of begging the question, as has sometimes been alleged, but the result of a hard fought battle in Book I concerning the nature and scope of demonstration. In general, when Hume runs intuition and demonstration together, he is considering only those areas where the relevant standard of equality is not available, and where, strictly speaking, only intuitive judgements, not demonstrative proofs, are available. 5) Belief in the Results of Demonstrative and Probable Reasoning: As Hume is notorious for demanding of probable reasoning that it meet certain standards if it is to be considered as "based on reason," it is worth considering whether he raises a similar question about demonstration. In fact, he does raise the question and considers it to admit of an easy answer that has nothing to do with deductive validity. When considering belief, he says: I therefore ask, Wherein consists the difference betwixt believing and disbelieving any proposition? The answer is easy with regard to propositions, that are prov'd by intuition or demonstration. In that case, the person who assents, not only conceives the ideas according to the proposition, but is necessarily determin'd to conceive them in that particular manner, either immediately or by the interposition of other ideas. Whatever is absurd is unintelligible; nor is it possible

Hume on Demonstration for the imagination to conceive any thing contrary to demonstration. (T95)

171

a

In intuition, where the ideas are seen to be related directly, and in demonstration, where the "interposition of other ideas" is required, we are "necessarily determin'd to conceive them in that particular manner." If the ideas were not conceived to be related in that way, they wouldn't be the same ideas. It is "not possible for the imagination to conceive" them in any other way. The content of the ideas determines what they are, and relations of ideas that rely solely on that content cannot be altered by the imagination. Any such alteration would change the ideas themselves. Propositions known intuitively or denl0nstratively are necessary and known with certainty because they depend solely on the ideas that make them up. Since they depend on nothing else, nothing else can make them false. 34 There remain other issues and problems. If all ideas are separable, as Hume appears to claim (T79), how can any ideas stand in an intuitive or demonstrative relation? If demonstrative reasoning is abstract, how does his account of demonstration cohere with his account of abstract ideas? If the necessity of causal reasoning turns out to be the same as the necessity of demonstrations, as Hume apears to say on T166, can there really be any substantial difference between demonstrative and probable reasoning? Whatever the answer to these questions turns out to be, I hope at least to have shown that we must abandon our notion of formal deductive validity if we are to understand Hume on demonstration, and a fortiori on demonstrative and probable reasoning. 35 NOTES 1. Armette C. Baier, A Progress of Sentiments: Reflections on Hume's Treatise (Carrlbridge: Harvard, 1991) p. 63. The slide between "demonstration" and "deduction" occurs often between pp. 63-68. I use this example only to illustrate how even a careful author of a brilliant book on Hume can make this slip in spite of the fact that the same author warns us explicitly (p. 302, note 8) not to confuse "demonstration" with "deduction." See also Kretschmer's review of her book in Mind 102 (1993) pp. 340-348 for repeated reading of "demonstrative" as "deductive." 2. In the late seventeenth and early 18th century, "deduction" was in fact used to mean simply "argument" or "inference," with none of the overtones of formal validity that it has today. Locke and Hume used the term this way. See John Yolton, Locke and the Compass ofHuman Understanding (Cambridge: C. U. P., 1970), and my "Locke on Reason, Probable Reasoning and Opinion," The Locke Newsletter 24 (1993) pp. 35-79. 3. "Hume's Doubts About Probable Reasoning: Was Locke the Target?," Hume and Hume's Connexions, M. A. Stewart and John P. Wright, eds. (Edinburgh University Press, Penn State University Press, 1994). 4. All references to Hume's Treatise are to Nidditch's correction of the Selby-Bigge edition (Oxford: Clarendon Press 1987) in the fonn "T256." References to Hume's Abstract are to the version contained in the aforementioned edition of the Treatise, and

172

DAVID OWEN

are of the fonn "A650." References to Hume's Enquiry Concerning Human Understanding are to the Nidditch correction of the Selby-Bigge edition (Oxford: Clarendon Press, 1975) and are in the form "EHUI26." See, for example, T89 and A650, quoted below. Other examples occur at T79-80, T87, T95, TIll, TI61-162. Curiously, Hume doesn't make this claim at T463-464, where we would expect him to during his discussion of the possibility of a demonstrative morality. See also A651 and A652-653, and EHU26, EHU35, and EHUI63-64. 5. David Stove, Probability and Hume's Inductive Scepticism (Oxford: Clarendon Press, 1973) p. 36. 6. Mackie explicitly adopts Stove's view in "A Defence of Induction," in Perception and Identity, G. F. Macdonald, ed. (London: MacMillan, 1979). T. L. Beauchamp and i\. Rosenberg take a similar view in Hume and the Problem of Causation (Oxford: Clarendon Press, 1981). For more details, see my "Hume's Doubts About Probable Reasoning." (see, fn. 3) 7. As has been argued by Christopher Belshaw, "Hume and Demonstrative Knowledge," Hume Studies 15 (1989) pp. 141-162. 8. See note 2 above. It is nonetheless of interest to contemplate what someone who holds a Humean theory of demonstration would say about the class of arguments that we would describe as deductive with contingent premises. We will return to this problem below. 9. But Locke was in Holland when a Dutch translation was published. 10. See John Passmore, "Descartes, the British Empiricists, and Fonnal Logic," The Philosophical Review 62 (1953) and my "Locke on Reason, Probable Reasoning and Opinion" for substantial discussion of this point. All references to Locke are to An Essay Concerning Human Understanding, P. H. Nidditch ed. (Oxford: Clarendon Press, 1975)and are in the form "Essay 4.2.7," where 4.2.7 designates book 4, chapter 2, section 7. 11. Even in the chapter on truth, Essay 4.5, where Locke most frequently talks about propositions, he reduces propositions, at least mental propositions, to ideas and their relations. Thus although any two adjacent ideas in a chain of reasoning will make up a proposition, the work is done at the level of ideas. Locke followed the traditional distinction between terms and propositions, re-expressed in terms of ideas and propositions, and was pretty careful not to allow a single idea to have propositional content. But Burne, like Descartes, often allowed an idea to be propositional in nature. This is required of him, of course, if his theory of belief as a more lively idea is to have any chance of getting off the ground at all. 12. See Locke's extensive, detailed discussion of an example that makes just this point in Essay 4.17.4, p. 672line19 to p. 673 line 31. 13. That the ideas involved in Humean reasoning may be propositional creates great difficulty in interpreting particular examples in Hume's texts. I deal with some particularly troublesome cases below, especially in Section IV. 14. See T89, T452, T458, T496. 15. T200. I don't pretend that this is a full and accurate account of Hume's views on identity. 16. For the justification of addition of memory here, see T89, EHU26. 17. There is an implausible explanation. Proponents of the view I am arguing against could reconstruct Hume's thinking as follows. Only relations of proportions of quantity or number are transitive. The other three relations in this class are not. So

Hume on Demonstration 173 although all four relations can be the objects of single, intuitive judgements, only relations of proportions of quantity or number are transitive. Thus we can have demonstrations of the form "A is bigger than B, B is bigger than C, therefore A is bigger than C" but not of the form "A is hotter than B, B is hotter than C, therefore A is hotter than C." 18. "Artificial," like "oblique" and "indirect," is sometimes a hint that Hume is about to talk of reasoning. See Tl 04 and Tl97. It seems "artificial manner" here means something more like "an artful manner," as in "the art of reasoning," rather than "a fake or non-genuine manner." The Concise Oxford Dictionary lists "real, but made by art" as the third meaning of "artificial." And the OED lists as the first meaning of artificial "real, but not natural." There are two subcategories here: artificial in result as well as process, and natural products or results artificially produced. They quote "Harrison's machine for the production of artificial ice" as an example; there is nothing in the nature of the ice produced to make it artificial (it is made of water, not fish oil); strictly speaking it is the production that is artificial, not the water. Hume's distinction between the natural and artificial virtues is very likely to be using "artificial" in this sense; there is nothing artificial about the virtue except the story about how it comes to be. 19. This is not the place to discuss Hume's rejection of demonstrative geometry in the Treatise. His argument against a standard of equality in geometry comes mainly in T45-48, where he rejects such a standard as "useless" and "imaginary." He considered the topic important enough to return to it in the "Appendix." (T637-38) 20. I leave aside detailed consideration of the relation of contrariety. I think it is likely that Hume did not mention relations of contrariety as susceptible to demonstrative reasoning because he thought that all such relations could be seen to hold intuitively: "No one can once doubt but existence and non-existence destroy each other, and are perfectly incompatible and contrary." (T70) But see section IV below. I also leave aside further discussion of the enormously complex question of mixed mathematics, raised by the issue of mapping degrees of quality onto numbers. Hume returns briefly to the subject in Section IV, part 1, and Section Xll, part 3 of the Enquiry Concerning Human Understanding. 21. Although distinguishing between propositional and nonpropositional ideas is important for modem understanding, there is a certain degree of artificiality about this distinction with respect to Hume. As the footnote at T96 explicitly states, not all judgements (the mental equivalent of verbal propositions) involve the relating of "two different ideas." Hume's theory ofbelief commits him to the view that some judgements involve just one idea. And even when the proposition does involve the relating of two ideas, the resulting relation is, after all, just a complex idea. (Tl3) 22. This is not to say that the question of the content of an idea isn't problematic. For an excellent recent discussion, see Marjorie Grene, "The Objects of Hume's Treatise," Hume Studies 22 (1994) pp. 163-177. 23. A650-5l; EHU28-30. 24. For an attempt to relate the issues raised in this paper to "the problem of induction," see my "Hume's Doubts About Probable Reasoning." (see fn. 3) 25. Hume's Intentions (London: Duckworth, 1968) p. 41. See also his "Descartes, the British Empiricists, and Formal Logic," p. 548, where he attributes to Hume, surely mistakenly, the view that Locke's demonstrations were syllogisms. He cites EHU163 in support, the most important passage in the Enquiry concerning the limitation of demonstration to relations of quantity or number. The main point of the passage is that

174

DAVID OWEN

since "the component parts of quantity and number are entirely similar, their relations become intricate and involved; and nothing can be more curious, as well as useful, than to trace, by a variety of mediums, their equality or inequality." Perhaps the tenn "mediums" suggested "middle terms" to Passmore, and hence the interpretation in terms of syllogisms. But if I am right, "mediums" means here, as elsewhere, "intermediate ideas." The main point is that "the sciences of quantity and number" have the exact standard that n1akes demonstration, or chains of reasoning, possible. The reference to "pretended syllogistical reasonings" later in the paragraph is to the futile attempt to apply demonstrative reasoning to any "branch of learning, except the sciences of quantity and number." The reference to Locke's claim that "where there is no property, there can be no injustice" is meant to show that even if that claim is intuitively certain (it is, after all, "a more imperfect definition") it and other clain1s like it cannot ground a den10nstrative morality. The relevant standard of equality that makes demonstration possible is just not available outside of relations of quantity and nUluber. 26. For an attempt to work out of such an account, see my "Hume's Doubts About Probable Reasoning." 27. See especially Barry Stroud, Hume (London: Routledge, 1977) and Saul Kripke, Naming and Necessity (Oxford: Blackwell's, 1980). 28. The contrary of anything intuitively known would be contradictory, and could be seen to be so, without the need of any implication. 29. See also T87; A653; EHU26; EHU35; EHU164. 30. Among other places, see T463 where he says "the operations of human understanding divide themselves into two kinds...nor is there any third operation of the understanding." 31. Demonstrations of contradictions must be hypothetical, that is the premises cannot be known to be true. Interestingly, Hume explicitly allows hypothetical arguments concerning matters of fact or "reasonings upon a supposition; there being in them, neither any present impression, nor belief of a real existence." (T83) The problem we are currently concerned with doesn't, of course, arise for hypothetical probable or causal arguments. 32. I take it that this does not commit Hume to the claim that we never make mistakes about a relation of ideas being impossible. He is not ruling out carelessness or inattention. 33. This point may well answer the powerful objections put forward by Stroud and Kripke (see note 27) concerning the link between conceivability and possibility. But the issue is extremely difficult and complex. It is important to note that the epistemological slack between apparent and actual conceivability is limited to the sciences of quantity and number, that is to the sphere where demonstration is possible. Where demonstration is impossible, apparent conceivability really is conceivability. In other words, in Hume's system, apart from the sciences of quantity and number, contradictions are either immediately apparent or not available. 34. The issue of belief in the results of probable reasoning is far more complex. For further details of one aspect of this question, see my "Philosophy and the Good Life: Hume's Defence of Probable Reasoning," Dialogue 35: 2 (1996). 35. Earlier and shorter versions of this paper were given at Colloquia at the University of North Carolina at Chapel Hill and Stanford University. A version was also read at the Hume Society Conference in Rome, June, 1994. Queries from members of the audience at these occasions were most helpful.

Kant on Logic and Self-Consciousness PATRICIA KITCHER

I

One serious barrier to our understanding of Kant's views on logic are his persistent and-to contemporary readers-jarring references to consciousness. Consider, for example,his elucidation [Erklarung] of judgment: "A judgment is the representation of the unity of consciousness of various representations, or the representation of their relation insofar as they constitute a concept." 1 Or his account of conceptual clarification: "When I nlake a concept distinct, however, my cognition does not grow at al1. ..The content remains the same, only the form is altered, in that I learn to distinguish better, or to cognize with clearer consciousness, what lay in the given concept already." (Young, 1992: 569, AA IX: 64) Or his view of concept formation, from his handwritten logic notes [these are all Kant's notes, arranged as they appear in the manuscript]: The logical question is not: how we attain concepts, but which acts of the understanding fonn a concept. .. (Nr. 2857, AA XVI: 548) (attention: relation to consciousness) Logical source of concepts: 1. through comparison [how in one consciousness] how they relate to each other in one consciousness. (Comparison with each other) 2. through reflection (with the same consciousness): [how they compare with each other in one consciousness as identical or not] how different can be conceived in one consciousness. 3. through abstraction: [through] in this case omitting that in which they differ. (Nm. 2876, 2878, 2879, cf. 2865, AA XVI: 555-57; 552 <My additions are in comer brackets; Kant uses parentheses; the editor's interpolations are in square brackets. »2

The reaction of contemporary readers to such passages is typified by the comments of Robert Hartman and Wolfgang Schwarz, who provided the first English translation of the Jasche Logic. In a note to the account of judgment, they explain apologetically that "Here we have a definition of Relation (as unity of consciousness of several presentations), something that logic textbooks in Kant's time-as again in ours-have shied away from." (1974, 106n.) My goal in this paper is to argue for the opposite assessment of Kant's attempts to forge a link between logic and consciousness. Far from being an embarrassment, I believe that this project has the potential to deepen our understanding of both logic and consciousness. The illumination that might come to logic is of an unusual sort. I have no reason to believe that Kant produced a particularly elegant system of logic. This was not, I think, his goal.

176 PATRICIA KITCHER What interested Kant was a different, but nonetheless important set of issues. He wanted to determine how it is possible for us to use logic. How are we able to make concepts? How can we make judgments and inferences? How can we apply the tools oflogic to the world we encounter? Hence, I see Kant's logical investigations as fitting very straightforwardly into the general framework of transcendental logic. The question he was asking is: What are the [or some of the] necessary conditions for the possibility of logical cognition? Although Kant drew a careful distinction between general and transcendental logic (A55ff.1B79ff.), characterizing the latter in solely in terms of its a priori subject, in actual arguments, his claim was always that a priori elements are necessary, because failing these we would be incapable of any kind of cognition at all. So a crucial part of his transcendental project was to discover the necessary conditions for the possibility of cognition; he then needed to argue that among these necessary conditions were certain a priori features. For this paper, I will not be concerned with the a priori, but with necessary conditions more generally. My focus will be on the prima facie unusual condition of consciousness or self-consciousness. Kant believed that consciousness or self-consciousness had various roles to play in various kinds of cognition. In particular, as we will see, he believed that consciousness was a necessary condition for cognition generally, for clear concepts, for conceptual clarification, and for inference. This is why his work might also shed much needed light the subject of consciousness. For one obvious question to ask is: What are the functions of consciousness or self-consciousness? Kant's many claims about consciousness and self-consciousness are often regarded as less than limpid. To bring needed clarity to these discussions, I am going to adopt a simple functionalist model of self-consciousness recently proposed by Robert Van Gulick. 3 Van Gulick's own functionalist account has not generated much enthusiasm, because it does not seem to capture the interest or importance that we attach to this notion. My strategy is to clarify Kant's reflections on logical cognition and use them to develop a novel and hopefully interesting account of self-consciousness by infusing Kant's rich account with Van Gulick's clarity. II

In this section, my purpose is two-fold. I want to present Kant's most general argument for the necessity of self-consciousness in cognition and also to argue for the fruitfulness of Van Gulick's model as a regimentation of Kant's sometimes contradictory and confusing claims about self-consciousness. Kant scholars should be chafing at my references to "consciousness" or "selfconsciousness," as ifthese were the same thing. A standard criticism of Kant's general claim that any cognition requires self-consciousness is that he seems to

Kant on Logic and Self-Consciousness 177 slide from claims about consciousness to claims about self-consciousness without recognizing that he needs an argument that consciousness requires selfconsciousness. 4 One of the great strengths of Van Gulick's model of selfconsciousness is that it permits a clear and satisfying resolution of this longstanding problem in Kant scholarship. On Van Gulick's model, we will be able to understand exactly why Kant believed that cognition required selfconsciousness and the relation between this doctrine and his equally important claim that cognition requires a single or unified consciousness. Kant often expressed these two claims in a single phrase: "The synthetic proposition that all the variety of empirical consciousness [consciousness of particular representations] must be combined in one single self-consciousness, is the absolutely first and synthetic principle of our thought in general. (Al17a, my underscoring) To appreciate his argument about self-consciousness, it will be helpful to rehearse the familiar considerations that support his claims for synthetic unity. In a sense, this central doctrine follows directly from another, equally famous central doctrine: knowledge requires both intuitions, including at least information provided by the senses, and concepts, which at least classify and organize sensory information. (A5lIB75, cf. Anthropology VII: 138) But knowledge does not just require sensibility and understanding; it also requires their "union," because the information captured in conceptual form must be based on and reflective of (though not perhaps exhausted by) the information provided by the senses. Hence understanding "combines" or "synthesizes" sensory information in a further representation to produce knowledge. If we take "synthetically unified in one consciousness" to indicate just this condition required by cognition, then what Kant means by the phrase is that representations are connected by operations of synthesis-either directly or through chains of syntheses-to each other: some are synthesized together to produce others; some are the synthetic products of others. If this is what Kant means by a single or unified consciousness, then he has a powerful argument for part of his absolutely first, synthetic principle: Cognition requires that diverse mental states be united in one consciousness. But Kant does not just conclude that all representations must belong to a single consciousness, he also claims that they involve self-consciousness or, combining both ideas, that they belong to a single self-consciousness. It is here that Van Gulick's functionalist model of self-consciousness can provide needed clarification of Kant's thinking. Van Gulick shares with Kant three crucial assumptions. First, the notion of self-consciousness must be prised away from its traditional dependence on the admittedly disastrous method of introspection. Second, functionalism is a promising and, at least for Kant, the only possible non-introspective means of characterizing our thought processes. 5 Finally, Van Gulick and Kant agree that mental states typically carry information about the outer or inner environment.

PATRICIAKITCHER As Van Gulick observes, many inanimate objects also have states or properties that carry information, for example, the rings of a tree. In these cases, however, the information has no relevance to the object~ it cannot be used by the object. By contrast, Van Gulick asks us to consider the following. When rats are nauseated after eating a new food, they will refuse that food in the future. Presumably, this is because one of their inner states carries information about the new food and its link to nausea or just to something bad. In this case, the information stored in one of the rat's inner states is available to the rat for some use or other. Van Gulick's proposal is that in such cases, we describe the rat as being "self-conscious," meaning simply that inforn1ation stored in one of the rat's states is used by it. Van Gulick acknowledges that such examples may not provide very rich or interesting cases of self-consciousness. Nonetheless, I think that they are clear cases. The rat is "conscious" or "aware" of his own mental state in the straightforward functionalist sense that it is properties of that state that account for other states of the rat and, ultimately, his behavior. 6 Second, what he is conscious of is a state of himself-even if he does not recognize it as such. Kant comes very close to this model of self-consciousness in his most explicit discussion of conscious versus unconscious mental states, which occurs in the Anthropology. While noting that it seems contradictory to speak of cognitive states of which the subject is unconscious, he observes that in some cases, the description is quite tempting. Consider seeing a man in a far off meadow. One might not be able to tell exactly what his eyes, nose, mouth, and so forth look like, even though one must be conscious of those features in some sense, because one is, after all, seeing the man: 178

For if I want to say that I do not at all have the idea of him in my intuition because I am not conscious of perceiving these parts of his head...then I could not say that I see a man, since the whole idea (the head or the man) is composed of these partial ideas. (AA VJI: 135)

Notice that what forces Kant to conclude that these ideas are conscious in a sense-he describes then1 as "mediately conscious" or "obscure"-is precisely that information from these states makes a contribution to other cognitive states and so is used by the systenl. On the other hand, in Jasche's rendering of his Logic Lectures, which Kant commissioned, it is claimed that "Consciousness is really a representation that another representation is in n1e." (Young, 1992: 544) In his own handwritten logic notes, Kant explains that "Consciousness is an [idea] (clear) representation of his own idea" (AA XVI: 79). He goes on to claim: This [consciousness] provides the ground, so that the thinking power of the soul can come to exercise its full measure. For because the soul knows what it conceives and how it conceives, it can manipulate representations in various ways, comparing them, abstracting, combining, and through this correcting some of them, which one

Kant on Logic and Self-Consciousness

179

calls reflection [nachdenken]. Which animals cannot do because they are not conscious." (AA XVI: 79-80)

In a note written on the next page one year later, he also claims that "To be conscious of a representation is to know that one has this representation; that is, to separate this representation from others." (AA XVI: 80) Finally, on the subsequent page, an even later note asserts that "An obscure idea is one that we are not conscious of, or that we cannot differentiate from others. We thus do not even know that we have it." (AA XVI: 81) In these passages, Kant seems to be offering two related criteria for consciousness of a representation: the representation must be differentiable from others and be recognized as the subject's own, and the former property is necessary for the latter. These criteria are, however, inconsistent with the claim in the Anthropology that we are conscious of obscure representations. More importantly, they are inconsistent with many passages in the Critique where he claims that consciousness or self-consciousness is necessary for knowledge, and then describes the consciousness involved as either "indistinct" or "obscure," or not a consciousness of my representations "as such." (A104, A116a\B132, B133-34) And, most importantly, they are inconsistent with his most explicit discussion of the relationship between obscurity and consciousness inthe Critique: "Clearness is not, as the logicians assert, the consciousness of a representation. A certain degree of consciousness, though it be insufficient for recollection, must be met with even in many obscure representations... (B415a) Like his predecessors and successors, Kant struggled, and failed, to find a coherent account of consciousness. So my claim is not that Kant had Van Gulick's clear functional sense of "self-consciousness"-a system's using information contained in one of its own states-firmly in mind when analyzing the prerequisites of cognition. If he had, his discussions would be much clearer. It is, rather, (1) that this notion is very close to some of his most important claims about consciousness and self-consciousness, particularly in the Critique, (2) that there is no way of interpreting all of Kant's claims about consciousness that renders them consistent, (3) that using Van Gulick's' functionalist notion renders Kant's arguments for the necessity of consciousness or self-consciousness for various sorts of cognition sound, and (4) that with this regimentation, Kant's reflections about the necessity of self-consciousness offer an alternative vision of self-consciousness. At this point, I have argued only for (1) and (2); the remainder of the paper will constitute an extended argument for (3) and (4). To begin to understand Kant's arguments for the necessity of selfconsciousness, consider again his reasons for concluding that cognition requires that different representations be united in a single consciousness. Even the simplest cognition requires intuitions and concepts. Specifically, information from intuitions must be ordered and combined in concepts, so intuitions and concepts must belong to a single synthetically unified consciousness. Notice, however, what the operation of synthesis involves. Information from

180 PATRICIAKITCHER intuitive states is synthesized in a later intuitive or conceptual state, or information fronl intuitive and conceptual states or information from several conceptual states is synthesized in a later conceptual state. Hence, each and every act of synthesis is at the same time an act involving self-consciousness, in Van G-ulick's sense, since it is information from earlier states of the unified consciousness that is used in the synthesis. Kant does not slide from claims about consciousness to claims about selfconsciousness. Rather, he maintains indifferently that diverse mental states or representations must belong to one single consciousness or to a self-conscious.. ness, because exactly the same considerations about the prerequisites of cognition and representation support both doctrines. By employing Van Gulick's functionalist model of self-consciousness, we can finally understand the logical relations between these two crucial doctrines. On that model, the difference between them is not one of content, but merely of emphasis. III. In the Critique, Kant argues that consciousness or self-consciousness is necessary for "recognition" in a concept and for objective judgment. Although these well-known discussions are inlportant, my focus will be on two cognitive tasks that are both less discussed and more directly linked to logic: conceptual clarification in this section and inference in the next section. A note to the Paralogisms chapter partially cited above offers a fairly extensive account of the relation between clear and obscure concepts and consciousness: Clearness is not, as the logicians assert, the consciousness of a representation. A certain degree of consciousness, though it be insufficient for recollection, must be met with even in many obscure representations, since in the absence of all consciousness we should make no distinction between different combinations of obscure representations...But a representation is clear, when the consciousness suffices for the consciousness of the distinction of the representation from others. If it suffices for distinguishing, but not for consciousness of the distinction, the representation must still be entitled obscure. There are therefore infinitely many degrees of consciousness, down to its complete vanishing. (B414-15)

In Jasche's version of the Logic Lectures, clarity and distinctness are explicated further, in terms of marks: A mark is that in a thing which constitutes a part ofthe cognition of it, or-what is the same-a partial representation, insofar as it is considered the ground ofthe whole representation. All our concepts are marks...(Young, 1992: 564) ... logical distinctness.. .is to be called complete distinctness insofar as all the marks, which, taken together, make up the whole concept have come to clarity. (Young, 1992: 567-68)

In the Anthropology, Kant distinguishes clarity from distinctness:

Kant on Logic and Self-Consciousness

181

When we can distinguish one object from another in our ideas, we have a CLEAR consciousness of them. But if the composition of our ideas is also clear, our consciousness of them is DISTINCT. Only when our ideas are distinct does a collection of them become knowledge. (AA VII: 137-38)

A mark is a representation of a characteristic or property that has been abstracted from individuals and through which individuals can be known. But any mark only amounts to a partial cognition. My son Andrew is a teenager, but he also has other representable characteristics through which I can understand him. The concept of a teenager itself contains other marks or concepts. So teenagers are hunlan beings of a certain age. The distinction between obscure and clear concepts is well-drawn. An obscure concept can be used to classify objects. So, for example, among my son's friends, I can distinguish the awesome from the ordinary, even though I am unclear what marks the "awesome." By contrast, in the case of a clear concept, we are conscious of the various marks of the concept~ the marks are distinguished one from another and so can be used (separately) to distinguish classes of objects, and the marks are recognized as part of the concepts of which they are marks. To understand Kant's complex views on distinctness, we need to consider the difference between two kinds of marks. A coordinate mark is an immediate mark of a concept, but a subordinate mark is the mark of one of its other nlarks. So "hunlan being" is an immediate mark of "teenager," but "human being" itself has marks, "rational" and "animal," so these would be subordinate marks of "teenager." In Jasche's Logic, distinctness is characterized as a "higher degree of clarity." (Young, 1992: 567) The passage goes on to differentiate extensively from intensively complete distinctness. A concept is extensively completely distinct if all of its coordinate marks are clear. It would be intensively completely distinct if all of its subordinate marks were clear. This, however, is impossible for empirical concepts. Even if we can ascend to the highest genus, we can never complete the serires of lower species. (Young, 1992: 565)7 Nonetheless, a concept is distinct (though not completely so), if its marks are distinguished and ordered with the subordinate under the coordinate. (AA VII: 138) Although I believe that this account of the relations among obscure, clear, and distinct concepts and consciousness is correct as far as it goes, it does not go nearly far enough. For it presents Kant's various claims about the consciousness involved in using the different types of concepts only in terms of an unanalyzed notion of consciousness. The "obscure," the "clear," and the "distinct" are differentiated, because through "consciousness," the composition of concepts is articulated, distinguished, or recognized. Further, we have been given no indication of why it is necessary to assume that we have these different types of concept or why some of them require a higher degree of consciousness. To remedy both defects, I will again appeal to Van Gulick's

182 PATRICIAKITCHER model of self-consciousness, for, as will be evident, this captures much of what Kant wants to say, while steering away from any unwelcome implications of introspection. Van Gulick's model clarifies Kant's position on obscure concepts. Users of obscure concepts are conscious of their own representations, and hence selfconscious, because they can differentiate objects from each other-say the "awesome" from the rest of the universe---:-only because they are able to use information stored in one of their own states. Now consider Kant's claim that in the case of clear or distinct concepts, subjects are conscious of the distinction among their concepts, or conscious of the structural composition of their concepts. On Van Gulick's model, this would mean that information about the structure of their concepts is available to cognizers for use in classifying and inferring. In the case of the concept of "teenager," for example, cognizers do not just use information from the the states containing information about "human being," "aged 13-19," and "teenager." They also possess information (contained perhaps in another state or simply in the relations among these states) about the relation between the information contained in the two mark states and the information contained in conceptual state of which they are marks. This information might indicate that one of the nlark concepts can be applied to any object to which the concept of which it is a mark is applicable. (Young, 1992: 618) So, for example, once I have classified my son as a teenager, I can infer that he is a human being and also rational, and hence that it is possible to reason with him. On this gloss, the advantage of clear or distinct concepts is manifest. Since in these cases cognizers possess information about the relations among the information represented in different states, bringing an object under a clear concept enables the system to gain a significant amount of knowledge about it and the more disntinct the concept, the greater the gain. As Kant put it, "...the more distinct our cognition of a thing is, the stronger and more effective it can be too." (Young, 1992: 569) One problem with Kant's claims about being conscious of the distinctions among our concepts, or conscious of their composition, is that they suggest some type of introspective perusal and reporting of mental contents. This cannot be what he intends. Kant's rejection of introspection was consistent, principled, and even passionate. Besides making the standard objection that without noticing it, introspectionists "imagined discoveries are [only] of those things that they themselves have placed within their [minds]," he believed that continuous introspection was a fairly sure road to lunacy. (AA VII: 133) His principled objection rested squarely on his basic epistemological principles: Any awareness-even of one's own states-must reflect a form of sensibility; any judgment based on materials gathered by sensibility nlust reflect the necessary rules of the understanding. So, despite their seeming immediacy, the data of introspection enjoy no special status of direct or certain knowledge.

Kant on Logic and Self-Consciousness 183 To appreciate the ability of Van Gulick's model to clarify Kant's views about concepts and consciousness while honoring his rejection of introspection, we need to bring the details into sharper focus. In the simple case of the nauseated rat, the argument was straightforward. The rat is self-conscious, because he is conscious of one of his states in the following sense: information about the nausea-producing food is stored in one of his states and available to him. The claim that he is conscious in this sense is defensible because, otherwise, there is no accounting for his avoiding the food. Hungry rats are not picky eaters. In the case of clear and distinct concepts, cognizers are conscious of the composition or structure of a concept in the sense that information about the relations among the informational contents of several of their states, including the state representing the concept and the states representing its marks (other concepts) is available to the cognitive system. What phenomena stand behind Kant's claim that cognizers are conscious of the composition of their concepts in this sense? One obvious phenomenon is inference. Consider again the contrast between the obscure concept "awesome" and the clear concept "teenager." In the latter case, a cognizer can move from classifying someone as a teenager to classifying him as "aged 13-19." It is this move, from one classification to another, that is explained by assuming that the system is conscious of the structure of its concepts in the sense that it possesses information about the relations among the informational contents of its own representational states. With obscure concepts, no such move is possible. The second phenomenon, which is prominent in the Logic Lectures and probably stands behind his discussions in the first critique as well, is conceptual clarification. As Kant explains, this can happen in one of two ways. A scientist can make a distinct concept by proceeding "synthetically," by beginning with the parts and forging a whole. (Young, 1992: 568-69) So, for example, a chemist might explicitly define an "acid" as a "proton donor." Henceforth, his usage will be explicable only by assuming that information about the relation among the representational states with these informational contents is stored in him and available for use in reasoning. Clear concepts can also be produced by analysis, by distinguishing "better, or [cognizing] ...with a clearer consciousness what lay in the given concept already." (Young, 1992: 569) Since introspection is not an option, how is Kant conceiving of this process? I assume that he envisions something like the following. To clarify my concept of "awesome," I contemplate some paradigms and some foils. At this point I might note that individuals and objects that are "awesome" have other properties in common: the individuals are physically attractive, popular with their peers, and skilled at tasks teenagers regard as important; the objects are exemplary versions of things teenagers regard as important. In this way, I am making the marks that were perhaps implicit in my use explicit or I am discovering the marks that should govern my usage,

184 PATRICIAKITCHER because these properties are common to those I call "awesome." The key point is that this process is not carried out by "looking" at my mental states, but by looking at the world and the classifications I make or can make of its objects. I believe that this is Kant's understanding of the process of conceptual clarification, because of his explicit description of how philosophers make "given concepts distinct" in Jasche's version of the Logic Lectures. Most interpreters have fastened on the memorable contrast between m.aking a distinct concept and making concepts distinct (Young, 1992: 568) But notice how these processes become intertwined in practice: The philosopher only makes concepts distinct. Sometimes one proceeds synthetically even when the concept that one wants to make distinct in this way is already given. This is often the case with· propositions based on experience, in case one is not yet satisfied with the marks already thought in a given concept. (Young, 1992: 569)

So, for example, I might discover in my own usage and that of my circle that the classification "teenager" is often linked with the classification "irrational." Recognizing that teenagers are human beings and hence rational, or perhaps encountering a teenager with detectable levels of common sense, I make the necessary adjustments. I cease to infer "irrational" immediately upon hearing sonleone classed as a "teenager" and modify my behavior so that I now treat teenagers with the more appropriate Kantian attitude of respect. At this point, three levels of self-consciousness, or information contained in the system and used by it, are in play: information contained in conceptual states, information contained in the links between conceptual states (or a separate state representing those links), and information contained in whatever states or arrangement of states enable the system to evaluate and alter the links between its conceptual states. Van Gulick's model provides a helpful clarification of Kant's claims about our consciousness of the composition of our concepts, because it captures. his reasons for thinking that we must be so conscious in some sense, and to a higher degree than we are with obscure concepts, and also his reasons for thinking that conceptual clarification requires an even higher degree of consciousness, without mnning afoul of his complete rejection of the method of introspection. Alternatively, given Van Gulick's model of self-consciousness and the practices of conceptual clarification and certain kinds of inferences, we can see that Kant is justified in clainling that although all concept use requires unity of consciousness and self-consciousness, the use of clear concepts requires these features to a higher degree than obscure ones. Before examining reason and inference in somewhat greater detail in section IV, I should just note that on this interpretation, conceptual clarification has a very different role than it has played in twentieth century philosophy. The point of clarifying concepts is not to gain a priori, certain, or infallible knowledge. Kant is quite clear in both the Logic Lectures and the Methodology

Kant on Logic and Self-Consciousness 185 section of the Critique (A7271B755) that definitions are impermanent. Rather, the point of clarifying our concepts is to integrate· our knowledge claims with each other, and so to permit quick (but not infallible) moves from one classification to another.

IV. Reason has an odd place in the Critique ofPure Reason. Because Kant wished to stress the unfortunate role of reason in unleashing metaphysical speculation, the essential positive contribution of pure reason to cognition in general, and science in particular, has only recently begun to be appreciated. (Kitcher, 1986; Guyer, 1990) When Kant turned to reason and inference in the Dialectic, he made no claims about the necessity of consciousness or self-consciousness. Nonetheless, from his general description of inference in, for example Jasche's version of the Logic Lectures, it is clear that reasoning involves a unified consciousness: "By inferring is to be understood that function of thought whereby one judgment is derived from another. An inference is thus in general the derivation of one judgment from the other." (Young, 1992: 609, cf. A3031B359) What Kant does not mention, presumably because it is so obvious, is that an inferred judgment is possible only if it belongs, with the judgment from which it is inferred, to a unified consciousness. According to Jasche, Kant began his lectures on logic by explaining the relation of the laws of logic to the understanding. Like everything in nature, the understanding operates according to rules and logic makes those rules explicit (and systematizes them). Since logic makes them explicit, a unified consciousness that can be said to possess a faculty of reason must contain and have available for use particular inferential patterns, such as disjunctive syllogism. Hence, cognizers must be conscious of inferential patterns or rules, in the same sense and for the same reason that they must be conscious of prior representations, concepts, and the composition of some concepts. We can explain their cognitive activities only by assuming that this information is available to them. But some judging and inferring require an even higher degree of selfconsciousness. To see why, consider the following conundrum. In Kant's view, the laws of logic, plus the categorial principles are the laws of thought. Indeed, they are the necessary rules without which no thinking is possible. If these rules are both necessary for thinking and descriptively adequate of our actual mental processes, however, then what is the point of writing texts in either logic or epistemology? Seemingly, it cannot be to improve our thinking, since that is already and of necessity in perfect accord with both the laws of logic and the categorial principles. Yet the point of such texts is manifestly to improve the quality of our thinking.

186 PATRICIAKITCHER In Kant's view, cognizers originally use logical rules, but only implicitly. Through studying those rules, we recognize that some of them always take us from true premises to true conclusions, while others do not. The truth presenJing rules (as we would now describe them) are reduced to the smallest number of different forms, and contrasted with less trustworthy patterns of reasoning. Cognizers can then learn these explicit rules, although eventually they become automated and are rarely thought of as such. Nevertheless, because we can think these rules "for themselves... apart from their application" in logic, as in ethics, we can act not just according to rules, but according to the concept of rules. Kant regarded the laws of Aristotelian logic as necessary, because coherent thinking requires the use of some laws and Aristotelian logic provided the best systematization of inference. Due to various impedinlents, such as lack of attention, haste, and prejudice, people make inferential errors. So, for example, cognizers might try to reason fronl a major premise, such as "all things are alive or dead," which fails to include the realm of the non-living, which is within the sphere of "thing." When taught that in a disjunctive syllogism, it is crucial that all of the disjuncts be so related that they exhaust the sphere of a given concept, they will be less likely to commit such errors. On Kant's view, the conundrum vanishes, because logic can be both grounded in inferential practice and normative for it. So much, Kant believed, would be conceded by any [then] contemporary logicians. His distinctive contribution came from his transcendental project of examining the necessary conditions for the possibility of cognition. (A7831B811) The rules of logic may be necessary for thinking, but that conclusion simply raises another, deeper question: How can we have any assurances that the laws of logic apply to the world we encounter? In the Aesthetic and the Axioms of Intuition, Kant tried to explain how the wonderfully useful system of mathematics is applicable to the world we sense. One project of the Dialectic is just the logical analog of the earlier attempt to vindicate the "possible application [of mathematical concepts] to empirical intuitions." (B147) To see the problem, consider his textbook example. (A3211B378) Suppose that we wish to establish Caius's mortality not by empirical means, but mediately, through an inference of reason. We must seek "a concept (in this case "man") that contains the condition under which the predicate (general term for what is asserted [,morta!']) of this judgment is given." (A3221B378) That is, we must seek some concept, that is such that the marks of this concept-the properties universally associated with it (say M}, M 2 , M3) provide a sufficient condition for the property of being mortal. Using, as he sometinles does, Euler circles to illustrate this point, we can prove that Caius © is in the realnl of the mortal only by finding a concept to apply to him that is

Kant on Logic and Self-Consciousness 187 itself in the realm of the mortal, because it is universally associated with propel1ies (M}, M 2 , M 3) that are sufficient conditions for mortality:

Imortal I [ MIM2M3 => mortal

I human being=>MIM2M31

In this example, we can see the principle of categorial inference noted above clearly in action: the inference works only because "human being" is a restriction or specification of "mortal." Hence Kant concludes that if reasoning is going to be possible, empirical concepts must be orderable in hierarchies, they must be related as genus and species. That is, it is only because "human being" is a species of "mortal" that we can make the desired inference to Caius 's mortality. The obvious objection to Kant's proposal is that he assumes that there is one set of sufficient conditions for any property. Conceding the difficulty with creating a perfect hierarchy of all concepts showing their relative positions as genera and species, however, he is still right that, for any particular inference, the concepts involved need to bear this relation to each other. Hence, he is justified in concluding that beyond rules of inference, reason contains a nurrtber of ideas (A327/B383-84), principles (A308/B365), or maxims (A666/B694), homogeneity, specificity, and continuity, which "guide the empirical employment of reason." (A663/B691) Although not expressed in terms of a requirement of self-consciousness, but in the even odder language of a "demand of reason," the point is by now familiar. In addition to all the other levels of selfconsciousness needed for cognition, cognizers capable of inference must also have and use principles about the systematic unity of their concepts. On Van Gulick's model of self-consciousness, logic and epistemology have the. following impact on our information-processing. The structures of some concepts, inferential principles, and principles of systematicity must already be contained in some states or arrangements of states of cognizers, since otherwise there would be no accounting for their ability to use this information in making judgments and inferences. After learning logic and epistemology, this inform-

188 PATRICIA KITCHER ation contained in the system and used by it also becomes available for use in more sophisticated inferences. Previously, cognizers could just move from a state representing the fact that all X's are a, b, or c to asserting major premises "X's are a's or X's are b's or X's are c's." .After tutelage, they can separately access "X's are a's", "X's are b's", "X's are c's," infer to these from its assertion of the disjunctive major premise, note that these are not of the standard "a or not-a" form, and then attempt to determine whether these properties are nonetheless exhaustive. In the case of epistemological instruction, about, for example, the importance of hierarchically related concepts, prior to instruction, cognizers will tend to discard concepts that do not fit into their hierarchies~ afterwards, they may infer from the prevalence of a concept that it may fit into hierarchies of concepts that they have not yet mastered. In his paper, Van a-ulick refers to the "increasing transparency of thought" as we consider more advanced cognitive creatures. Although I am not sure exactly what he has in mind, the framework of transcendental idealism provides a straightfonvard warrant for.the epithet "transparent": I take Kant's deepest epistemological insight to be that all our cognitions are mediated by our representations of the information that we acquire from interactions with the outer and inner world and by our ways of processing that information. Since our intuitions, concepts, judgments, and inferences are reflective of our own mental structures, and since our thinking becomes more sophisticated by making its own implicit rules explicit, there is a clear sense in which sophisticated cognizers are transparent to themselves and a clear sense in which self-conscious thought is truly a consciousness of self. V.

With the suggestion that philosophy-logic and epistemology-provide a nleans of increasing self-consciousness, we have I believe, rejoined our pretheoretic understanding of the interest and important of self-consciousness. I will conclude with just a few summary remarks. On this reading, there is nothing particularly strange about Kant's claims that various kinds of logical activities require consciousness or self-consciousness. For this position is a reflection of two different and sound realizations. The first is that logical activity draws on a great deal of complex information that, roughly, enables us to apply rules to conceptual structures that are wellconstructed to enable those rules to apply. (I should note as an aside that this suggests that the many attempts to design logic programs without taking any account of how the system constructs concepts are not going to be very good models of our logical abilities.) The second is the fundanlental doctrine of transcendental idealism just noted-that any information that is going to be used by the system in performing various cognitive tasks nlust be contained in the system. Hence Kant's link between logic and consciousness is not strange,

Kant on Logic and Self-Consciousness 189 but potentially very informative, because he is trying to determine the kinds of information that must contained in a system that is able to perfornl logical inference. I believe that Kant's work is also informative about self-consciousness, because he offers a very different focus for our thinking about this most treacherous topic. Partly due to his rejection of introspection, and partly due to his positive views on the constructive nature of cognition, Kant did not regard self-consciousness as a faculty for revealing what our minds or selves are like so much as an integral part of a functioning cognitive system. To borrow a term from Coleridge, Kant's work shifts our understanding of self-consciousness away from a passive (and usually inaccurate) observation of what happens in the mind, to the "shaping" powers of the mind thenlselves. 8

NOTES 1. All references to the various versions of Kant's Lectures on Logic will be to Michael Young (ed. and trans.), Lectures on Logic, Cambridge Edition of the Works of Immanuel Kant, (New York: Cambridge University Press, 1992). Abbreviated references will appear hereafter in the text as (Young, 1992: page nUlnber(s)). The present passage occurs on p. 597, AA IX: 101. References to and citations from the Critique of Pure Reason will be to Norman Kemp Smith, Kant's Critique of Pure Reason (New York: S1. Martin's, 1968) and will be indicated in the text in the usual "A" and "B" notation. Most other references to Kant's works will be to Kant, hnmanuel, Kants gesammelte Schriften, Akademie A usgabe, edited by the Koniglichen Preussischen Akademie der Wissenschaften, 29 vols. Berlin and Leibzig: Walter de Gruyter and predecessors, 1902-, and will be indicated in the text by AA, the volume number, and page reference. 2. Throughout I am translating "in einem Bewusstsehn" as "in one consciousness." 3. Van Gulick, Robert, "A Functionalist Plea for Self-Consciousness," Philosophical Review 97 (1988). 4. Jonathan Bennett, Kant's Analytic, (Cambridge: Cambridge University Press, 1966), pp. 104-107. 5. On introspection see Van Gulick (1988, p. 163). Kant is extremely critical of introspection in the Anthropology (AA VII: 132-33). Kant's functionalism is evident form the way he introduces faculties. See, for example, his famous contrast between sensibility and the understanding at A51/B75. The opening sentences of Bonnet's (1755) Essai de Psychologie provide a clear recognition that 18th century theorists of mind had little choice but to be functionalists: "We know the soul only through its faculties; we know these faculties only by their effects" (Charles Bonnet, Essai de Psychologie, 1755, reprinted by New York: Georg alms Verlag, 1978.) 6. Dennett, Daniel C. Content and Consciousness, (New York: Routledge & Kegan Paul, 1969) 118-19. 7. Notice that this claim from the Logic Lectures appears to be inconsistent with Kant's claim in the Dialectic that both the principle of genera and the principle of

190

PATRICIAKITCHER

specification are both regulative ideals. (A652ff./B680ff.) Given this inconsistency, I take the Critique as authoritative, although I will not try to argue the point here. 8. This paper is a preliminary version of a larger project. I presented a preliminary version of another part of the same progject at the 8th International Kant Congress, which appears in the Proceedings of that conference. Besides being parts of a large project, this paper overlaps that one at pages 187-90 and 193-94.

Leibniz and the Model for Contingent Truths FRANyOIS DUCHESNEAU

I shall first remind you of the classical Leibnizian definitions and principles involved in the notion of contingent truth by quoting from the Monadologie. Article 33 states: There are two kinds of truths, truths of reasoning and truths of fact. Truths of reasoning are necessary, and their opposite is impossible. Truths of fact are contingent, and their opposite is possible. When a truth is necessary, the reason for it can be found by analysis, resolving it into more simple ideas and truths, although we cannot know the reasons in most cases. 1

In the preceding articles (§31 and §32), Leibniz had stated that two principles rule over our reasonings, the principle of contradiction and the principle of sufficient reason. Once our arguments have been found to abide by the principle of contradiction, there is still a need to show that there is a reason for the existence of a fact, as well as for the truth of a proposition. And, in the case of propositional truths, Leibniz had added that most often those reasons may not be known to us. Evidently, in truths of reason, we can hope for an analytic unfolding of such reasons, in terms of more simple truths, down to primitives. Now article 36 specifies the situation for contingent truths: But a sufficient reason must also be found in contingent truths or truths offact, that is to say, in the sequence of things distributed through the universe of creatures, whose analysis into particular reasons could proceed into unlimited detail because of the immense variety of things in nature and the division of bodies into the infinite. (L646; GP VI: 612-613)

From these statements, a problem comes immediately to nlind: how can we acknowledge that contingent truths, or truths of fact, are analytically true, as all truths are supposed to be, if we are faced with the onus of an infinite analysis which we cannot achieve. The problem is further increased by the logical requirements placed on the principle of sufficient reason when applied to propositions. As we know, Leibnizian logic builds up fronl the inesse principle: in omni propositione vera, prredicatum inest subjecto. So the sufficient reason for any truth is presumed to derive from the inner connection of notions. As Leibniz writes to Arnauld on 4/14 July 1686: "There must always be some foundation for the connection between the terms of a proposition, and this must be found in their concepts." (L337~ GP II: 56) The problem I shall address is therefore that of the analytic model for contingent truths according to Leibniz: how are we to conceive of the inner connection of concepts if such an analysis is to remain an endless task? Indeed, a way out would be to deny that there is any analytic model for contingent truths, and to fall back on a synthetic connection of notions, based on sense experience or on the formal conditions

192

FRANCOIS DUCHESNEAU

for this experience. But could we not make sense of Leibniz's position without .resorting to that Kantian shift in meaning? I shall first briefly recall some standard features of contingent truths for Leibniz, before turning to Leibniz' s attempts to spell out the analytic pattern for that type of truth.

I A set of criteria are jointly used by Leibniz to distinguish contingent from necessary truths. 1) Their contrary is possible. 2) They relate to "the sequence of things distributed through the universe of creatures.,,2 They therefore bear on existents. 3) Their analytic unfolding would proceed ad infinitum, if we applied to thenl the way of proof by means of substituting equivalents. 4) Their proof strictly requires the principle of sufficient reason. 5) The concepts signified by the subject terms are complete notions by contrast with the abstract terms in truths of reason. Every logical enunciation is governed by the principle prcedicatum inest subjecto. As Couturat explains: "The reason for this is that all truth is determined by the logical nature of its terms: truth is in a way inscribed in them in advance, and it suffices to analyze them fully to unveil it.,,3 This boils down to postulating that the conceptual connections in contingent truths can get an a priori explanation, that is an explanation based solely on the concepts involved, independently of the empirical occurrences that would exemplify such connections in nlatters of fact. Furthermore, the reason for those empirical relationships could not be inferred from mere facts: propositional connections need to be provided. In principle, the analysis of concepts could account a priori for those connections, if proceeding ad infinitum. To complete this characterization, it should be recalled that Leibniz mentions primitive truths of fact based on "inner experiences which are immediate with the immediacy offeeling.,,4 These are propositions instancing both sides of the cogito experience: on the one hand, sum res cogitans; on the other hand, sunt cogitata: my thinking relates to a manifold of objects. Following the analogy of primitive truths of reason which reduce to identities, those primitive truths of fact own a certainty that is not grounded in proofs.. As for derivative truths of fact, these cannot not be reduced to prinlitives, as derivative truths of reason cannot be reduced to mere formal identities. The analysis of derivative rational propositions relies on chains of definitions which involve the possibility of connecting complex terms with one another. In substituting definitional equivalents, identity serves as a formal operative mode for derivation. But, in truths of fact, the operative mode is provided by a sufficient reason or a sequence of sufficient reasons for propositions which are more or less directly connected with immediate inner experience. If all logical connections in truths of fact are empirically entailed in the cogito, they need to

Leibniz and the Model for Contingent Truths

193

be accounted for by an adequate analysis of the conceptual contents wherein those connections are involved. To solve the very complicated question of the status of truths of fact, Leibniz resorts to: (1) logico-metaphysical reasons; (2) epistemological reasons linked with his notion of scientific knowledge. Some texts develop quite neatly the logico-metaphysical arguments: for instance, sections 8 and 13 of the Discours de metaphysique and the letter to Arnauld of 4/14 July 1686. The individual notion of Alexander is the sufficient reason for all these propositions that assert a predicate of the subject Alexander. And each of those predicates is necessarily implied in the corresponding essence. To borrow an expression of Claude Troisfontaines: "The concept of an individual is the integral of all its predicates.,,5 Contemplating the unfolded notion, God would discover a priori in it that Alexander will defeat Darius and Porus, etc. In the same instance, our knowledge cannot but develop a posteriori through nominal definitions based on extrinsic denominations. But, "when we consider the connection of things,"6 the requirement of a sufficient reason for all intrinsic denominations of the temporal subject would constrain us to admit that an a priori proof for the unfolding of predicates is at least possible-a proof that would be akin to a real definition. Technically, the same rule of implication applies both to necessary and contingent truths. The difference can only come about in the way we draw the inference concerning the analysis or explanation of terms. In the style of the Discours de metaphysique, the inference for contingent truths ties up with an hypothesis: the necessary connection of notions is postulated, depending on God's free decree which actualized the best of possible worlds. But, if we abstract from the metaphysical style, the propositional connection that makes Ccesar will cross the Rubicon into a contingent truth relies on the connection between the enunciation and the determinations in the complete individual notion of Cresar. This notion is an actualized possible. As a possible, it will not exclude an infinity of other possibles which, one by one, would entail contradictory statements to those that can be deduced from the notion. As an actualized possible, the notion with its entire categorical and even conditional implications, requires a determinating reason. Leibniz's prime strategy consists in locating this determinating reason both inside the notion, to cope with the rule of true enunciation for any possible subject, and outside the notion, to account for the actualization of the given subject. From the metaphysical viewpoint, the reason for this contingent truth derives from the original free decision to create the universe, on the basis of which the true enunciations relative to Cresar will be certain, that is to say necessary ex hypothesi: this is the external reason. From the logical perspective, the notion of Cresar as (possibly) existing, according to the expression of the letter to Arnauld, "includes sub ratione possibilitatis existences or truths of

194

FRANCOIS DUCHESNEAU

fact or the decrees of God upon which facts depend.,,7 This is the corresponding inner reason. Now, what exactly do the notions whose internal connections form contingent truths imply? If we consider a sphere sub ratione generalitatis seu essentice, seu notionis specificce sive incompletce, 8 that is according to abstract notions of modes without existential substrate, we may draw therefrom truths of reason about geonletrical properties of this solid, but not truths of another type which would bear on particular circumstances about a sphere given hie et nunc. The generic notion of the sphere does not enable us to infer the specific dimensions of a given sphere. A fortiori it is impossible to assert that the individual agent that I am will necessarily accomplish the travel I am planning. The inference seems unwarranted from the notion of a finite rational agent to an act or event individualized in time and space. The individual notion requires an analysis that proceeds to the infinite. For us, the starting point is unavoidably an empirical, therefore partial, meaning of the notion, as set up on the basis of phenomenal marks. The analytic resolution of this notion which ties up with an empirical instantiation would require distinctions ad infinitum to discriminate between the elenlents of the notion and those of other possible subjects: For the concept of myself in particular and of every other individual substance is infinitely more inclusive and more difficult to grasp than is a specific concept such as that of a sphere, which is incomplete and does not include all the circumstances necessary practically to arrive at a particular sphere. To understand what this 1 is, it is not enough that I sense myself as a substance that thinks; I must also distinctly conceive that which distinguishes me from all other possible spirits. But of this I have only a confused experience. Although it is easy, therefore, to judge that the length of its diameter is not included in the concept of a sphere in general, it is not so easy to judge with certainty-though it can be judged with enough probability-whether the voyage that I plan to make is included in the concept of me. Otherwise it would be as easy to be a prophet as to be a geometrician. 9

It is worth noting that a probable inference would make it possible to deduce contingent truths in a way. God's understanding possesses an adequate representation of the "complete" notions. These notions would unfold the possibility relations behind individuals, namely the possibility of their causes, which is the infinite manifold of determinations the individual results from in the possible world he belongs to. The integral representation of the essence, namely of that system of determinations, involves the connection of the individual with all other individuals compatible with an ordering law which expresses concretely the universal order. This order in turn involves the irepresentation of a possible divine decree, as it shows a teleological reason Iidetermining a supreme and wise volition. The epistemological transcription Iwould provide something like the following: the full and comprehensive notion

Leibniz and the Model for Contingent Truths 195 at the basis of any contingent truth unfolds analytically in the shape of a law of causal determination~ this law would account for the infinitely diversified object by showing how it fits into a system of individuals structured according to an optimally ordered plan. For our finite understandings, truths of fact are presumed to be analytic and to meet the rule prcedicatum inest subjecto. In all cases, including truths of reason, the connection of terms in the true proposition depends upon the working out of a concept. The connection is analytic once an adequate concept about the structure of the object has been reached: Now I do not demand any further connection here than what is really found a parte rei between the tenus of a true proposition, and it is only in this sense that I say that the concept of an individual substance includes all its events and all its denominations, even those which are commonly called extrinsic, that is those which pertain to it only by virtue of the general connection of things and from the fact that it expresses the whole universe in its own way. For there must always be some foundation for the connection between the terms of a proposition, and this must be found in their concepts. (L337; GP II: 56) 0

In truths of reason, the adequate representation of the connection depends on an a priori construct wherein the relationships of the conceptual ingredients are given fronl the start. The concepts have referents we may account for by real definition-at least in principle-and their inner connections can be reached by analysis without our being compelled to rely on extrinsic reasons. By contrast, according to the Monadologie, the concept of an actualized possible refers to an "analysis into particular reasons [which] could proceed to unlimited detail."lo But for us the connection of terms in the true contingent proposition refers to an analysis that is presumed to have been achieved in an adequate representation of the object~ and such an analysis can only come about a posteriori. The true enunciations bearing on existents and events can only provide a necessity ex hypothesi. The proof of these enunciations is derived analytically from concepts under the sole condition that a sufficient reason for the connection is postulated. The connection appears to be synthetic in virtue of the necessarily inadequate knowledge we get about individuals from nominal definitions. But the principle of reason as it applies to the propositional form of contingent truths points to the admission of an analytic connection underpinning the synthetic one. How can Leibniz justify this substituting of a presumed analytic connection for the synthetic connection that we draw de facto? The question is still left quite open among Leibniz scholars and no interpretation of Leibniz's texts yields a simple answer. I will tentatively sketch my answer. Leibniz conceives the building of science according to a pattern of demonstration which is constructed by a stepwise substituting of definitional equivalents. This supposes that we can rewrite enunciations about facts by

196 FRANCOIS DUCHESNEAU substituting functionally equivalent enunciations, which at the same time may fit the analytic standards of necessary truths. One of the principal problems in Leibnizian methodology is to find appropriate forms and means for this conformity operation. The principle for such a transcription is consistently reaffirmed throughout Book IV of the Nouveaux essais sur I'entendement humain. I shall briefly quote passages that bear witness to that move: The linking of phenomena which warrant the truths offact about sensible things outside us is itself verified by means of truths of reason, just as optical appearances are explained by geometry. (NE 4.2.14; A VI vi: 374-75) I do believe that we shall never advance as far as one might wish; yet it seems to me that considerable progress will eventually be made in explaining various phenomena. That is because the great number of experiments which are within our reach can supply us with more than sufficient data, so that all we lack is the art of employing them; and I am not without hope that the small beginnings of that will be extended, now that the infinitesimal calculus has given us the means for allying geometry with natural science and now that dynamics has supplied us with the general laws of nature. (NE 4.3.26; A VI vi: 389) And the truth about contingent singular things is grounded in success, whereby sensory phenomena are linked together in just the way required by truths of the intellect. (NE 4.4.5; A VI vi: 392)

In my study of Leibniz's methodology, I have tried to consider the various strategies used to submit truths of fact to developments that may render them more conformable to the analytic rationality of necessary truths. 11 A significant Leibnizian trend consists in setting up a category between truths of fact and truths of reason, the category of hypothetical truths, truths necessary ex hypothesi. 12 The challenge is to conceive methods that may relate contingent and necessary truths in the practice of rational knowledge. A glimpse at the methodological scheme would reveal a kind of standard pattern. The relationship between irrational and rational numbers is a nlatter of construction, giving way to "approaching definitions" (definitions approchees), and hence, to distinct symbolic notions which involve more or less basic analyses and more or less comprehensive systematizations. The progression in analytic representation and comprehensiveness marks the transition from explicit truths of reason to implicit ones, wherein we are faced with an infinite complexity akin to that of actual existents. Would not an extension of the same scheme provide for the linkage between contingent and necessary truths in terms of analysis? This typically Leibnizian project does not make it possible to find in Leibniz an appropriate match for Kant's distinctions between analytic and synthetic judgments. Because they have underestimated the special constructive activity of reason intervening in both types of truths, many classical interpretations of

Leibniz and the Model for Contingent Truths 197 Leibniz's epistemology are to be faulted, starting with Russell's and Couturat's.13 II

The analytic model for contingent truths was one of Leibniz's major concerns during the period 1679-1686, which came to an end with the Discours de n1etaphysique and with the Generales inquisitiones de analysi notionum et veritatum. F. E. Andrews has suggested that in that period Leibniz's views on the analysis of truth evolved significantly fronl a rationalist to a nlore relativistic and empiricist position. 14 My interpretation is quite different since I do not believe that Leibniz changed his epistemological tenets to a considerable degree in the time sequence considered. However, during that period, the mathematical analogies about the relationship between finite and infinite quantities become more and more meaningful for unveiling the type of a priori proof we can attain regarding contingent truths. Indeed, the a priori proofs we may attain in that way concerning the analytic validity of contingent truths should not be confused with the a priori knowledge that would result from a deduction grounded in primitives. In a fragment De fibertate edited by Foucher de Careil,15 Leibniz reaffirms that "God alone knows the contingent truths a priori and sees their infallibility otherwise than by experience." (L264; F de C: 181) For our finite understanding, demonstrating that such and such predicates are inherent in the subject would imply an infinite analytic process, since the nexus terminorum in the subject's complete notion forms a series whose law exceeds our comprehension. However, Leibniz seems to hint at some a priori proof for contingent propositions when he concedes that we learn these truths either by experience or by reason: "We know by experience when we perceive a thing distinctly enough by our senses; by reason, however,when we use the general principle that nothing happens without a reason, or that the predicate always inheres in the subject by virtue of some reason." (L265; F de C: 182) There would be an a priori proof by means of the principle of sufficient reason, which would hypothesize a lawful connection of elements in the infinite series of predicates for a given substance. The fragment Cum animadverterem edited by Gerhardt recommends practicing hypothetical inferences and assuming provisional principles so as to demonstrate more complex propositions. (L225-27; GP VII: 299-301) In factual propositions, the conceptual connection would express a certainty pertaining to God's rational decree. Hence our ability to account for such propositions by reference to a "reason that inclines but does not necessitate" (ratio incfinans, non vero necessitans). (L227; GP VII: 301) The full knowledge of this reason would depend on a deductive analysis of the notions involved: it would match God's a priori intuition of the same. But we must be content with postulating such a reason and drawing formulas conformable with the indications of an a

198 FRANCOIS DUCHESNEAU posteriori analysis. In that context, the axiom of sufficient reason would afford an a priori justification for contingent truths: "It is certain, therefore, that all truths, even highly contingent ones, have a proof a priori or some reason why they are rather than are not." (L226~ GP VII: 301) Drawing his examples from physics, Leibniz mentions that without resorting to that axiom "an argument [cannot] be carried from causes to effects or from effects to causes." (L227~ GP VII: 301) If factual propositions may not undergo an integral a priori analysis, they can be given some analytic expression that is so ordered as to warrant their objectivity and fruitfulness. Thus, Archimedes has justified the laws of statics by resorting to a corollary of the principle of sufficient reason, stating a continuous homology between determinants and consequents in the analysis of phenomena. Thus, the principle of sufficient reason would afford sonle a priori proof of the order involved in a infinite analytic progression whose end result is evidenced through experience. This interpretation is confirmed by the fragment Quemadmodum in loco sabuloso edited by Couturat. 16 The axiom "Nothing happens which we could not account why it so happened rather than otherwise" serves as an a priori proof for a sequence where the antecedent consists in two mechanical apparatus in equilibrium, and where the consequent that is inferred consists in a similar state of equilibrium in the issuing effects. In a true proposition, as the predicate is contained in the subject, so the consequent is contained in the antecedent: "And so there must be some connection between the notions of terms, namely a foundation a parte rei wherefrom the proposition can be accounted for, or an a priori proof can be discovered."l? Leibniz seems to distinguish between: (1) the connection of terms as it derives from the nature of the subjects, which only an adequate analysis of the complete notions could unveil, and (2) the formal structure, the order relation according to which our understanding apprehends the implication of the predicate in the subject, or of the consequent in the antecedent. In the case of truths of fact, the axioms derived from the principle of sufficient reason exemplify this imposition of an a priori intelligibility that may rule over the a posteriori unfolding of phenomenal connections. Other texts in the corpus determine the norms for the analysis of truths of fact. The prime norm is indeed the principle of reason as it implies that any contingent truth owns an a priori proof, even if this proof is perceived only as a formal proof. But this formal proof is required to ground the "reality" of the analytic models that we try to develop a posteriori. In this context, Leibniz resorts to mathematical analogies to specify the relationship. between contingent and necessary truths: in my opinion, these analogies are of foremost importance since they suggest how to devise the a posteriori analytic nlodels needed to explain contingent truths. The main analogy refers to the relation between rational and irrational ("surd") numbers, between commensurable and infinite quantities, between finite and infinitesimal algorithms. The Specimen inventorum de admirandis

Leibniz and the Model for Contingent Truths 199 naturce generalis arcanis combines the assertion that every true proposition which is not per se nota, comprises an a priori proof in virtue of the principle of reason,18 with the following argument: The difference is essential between necessary or eternal truths and truths of fact or contingent: they differ so to say like rational and surd numbers. For necessary buths can be resolved into identicals, like commensurable quantities into a common measure~ but in contingent truths, like in surd numbers, resolution develops to the infinite and never ends~ hence the certainty and perfect reason of contingent truths can be known to God alone who encompasses the infinite with one intuition. Once this arcane is known, the difficulty pertaining to the absolute necessity of things is lifted away, and the distance shows up between infallible and necessary. 19

In a fragment probably preliminary to the Generales inquisitiones, which Couturat has edited under the title Verites necessaires et contingentes, (VE III: 455-463; C16-24) Leibniz acknowledges that the a priori science of complex terms would result from a proper understanding of the non complex terms. In the case of contingent truths, finite understandings lack this a priori knowledge, which needs to be substituted with a posteriori experience. Nevertheless, Leibniz does not deny the possibility that there may be a formal apprehension of the dependence of contingent truths on a system of a priori proofs, symbolized by the principle of reason. At this point, he draws on the analogy of rational and irrational numbers, but with a significant proviso: In a way contingent truths relate to necessary ones as the surd ratios in incommensurable numbers relate to the expressible ratios in commensurable numbers. As we can indeed show that a lesser number is contained in a greater one, by resolving both down to the greatest common measure, so we demonstrate essential propositions or truths by setting a resolution that gets to terms which are evidently by definition common to both terms. But, as a greater number contains an incommensurable one though we can never get to a common measure, even if resolution is carried on so to say to infinity, in the same way, we never get to a demonstration though we resolve [contingent] notions as much as possible. The only difference is that, with surd numbers, we can nevertheless set up demonstrations by showing that the error is less than anything assignable, but, with contingent truths, not even that has been granted the created mind. But I believe that I have thus exposed the arcane which had kept me perplexed for so long, since I did not understand how the predicate could inhere in the subject, while the proposition would not be rendered necessary on that account. But the knowledge of geometry and the analysis of the infinite brought me the light required for understanding that even notions could thus be resolved ad injinitum. 20

Leibniz here does not do away with the notion that there can be some a priori proof of contingent truths. Indeed, we may not produce a denlonstration by

200 FRANCOIS DUCHESNEAU adequate analysis for these truths the way we may set up calculi about incommensurable quantities. Nevertheless, the analogy with mathematical relations entailing an infinite progression suffices to warrant that such notions could be resolved ad infinitum, which means that they possess some kind of a priori proof. The difference with propositions implying irrational numbers is that with the latter we can demonstrate that the error occasioned by the transition to the limit is less than any assignable quantity; and so, we may anticipate provisionally the termination of the demonstrative process. In any event, when dealing with notions which entail an infinite analysis, we may not reach a demonstration stricto sensu, except if the notion is analogically considered as analyzable in a finite number of steps. Indeed, this cannot be the case with contingent truths, for which there is no demonstration possible based on strict conceptual implications: "Hence it shows up why we cannot discover any demonstration of a contingent proposition, the farthest we keep resolving the notions." (VE III: 457; C19) The Generales inquisitiones de analysi notionum et veritatum (1686), a complex and somewhat heterogenous treatise, tends to present analysis as a process based on the resolution of notions down to primitive terms. These are conceived as notions for which any further analytic regression is impossible or methodologically unwarranted: such are the concepts involved in the notions of terminus, ens, eXistens, individuum, and ego. Leibniz even goes to the point of admitting provisional primitive terms that provide theoretical foundations for the rational transposition of empirical properties, for instance, concepts which fall under the categories of extensum, durans, intensum, etc., and help schematize the rational ingredients in phenomena. When he addresses the question of contingent truths, Leibniz presupposes at once that their inner rationality is to be interpreted in light of the inesse principle and the principle of substitution of definitional equivalents salva veritate. Furthermore, reduction is achieved when analysis reaches either primitive terms or their methodological equivalents. And when an infinite progression is involved in unfolding the antecedent for a given consequent, it may suffice that we be certain that there would not be any latent impossibility in the infinite sequence considered from the viewpoint of its generic law. This serial order would form a general relationship enacted in the conceptual contents down to an infinite detail. And if we could reach near enough to that general relationship, this would provide us with a means to analytically resolve propositions whose inner connection involves infinity. From this it follows that for us to be certain of a truth, either the analysis must be continued to primary truths-at least, either to that which has already been treated by such a method, or to that which is agreed to be true-or it must be proved that from the very progression of the analysis, i. e., from a certain general relation between preceding analyses, and the one which follows, nothing of this kind will ever appear, however long the analysis is continued.

Leibniz and the Model for Contingent Truths

201

This is particularly worth noting, for in this way we can often be freed from a long continuation of an analysis. It can also happen that the analysis of letters itself contains something about the analyses of those which follow, much as the analysis of "true" here. It can also be doubted whether every analysis must necessarily end with primary truths, i. e., with what is unanalysable-especially in the case of contingent truths, as there is no time for these to be reduced to identical propositions. 21

Contingent truths may sometimes be analyzed by means of a serial law which reveals at once that an ultimate reduction is quite impossible, and that a determinating reason can be inferred for the derivation of the subject's essential predicates. This strategy can be justified on the basis that there needs to be an a priori proof for factual propositions. In fact, Leibniz insists that resorting to experience in this instance is unavoidable, and that the unfolding of sufficient reason for contingent truths requires that some conceptual elements be founded a posteriori, while others are provided a priori, nanlely logical principles and transpositive definitions: For to prove that a complex term is true is to reduce it to other true complex tenns, and these finally into complex tenns which are primary truths-that is into axioms (i. e., propositions which are known through themselves), definitions of non complex tenus which have been proved to be true, and the data of experience. 22

There is every reason to believe that the analytic status of contingent truths is akin to hypotheses of rational connection that would be expressed in an infinite series of requisites, under the rule of a global law. In the Generales inquisitiones, two groups of articles concern themselves with the distinction of necessary and contingent truths. The first group results in article 74, the second in articles 134 to 136. In the first group, Leibniz's thesis is to the effect that an infinite resolution ought to take place in some propositions, which we nevertheless consider as conditionally true. The question relates to the possibility of subnlitting the analytic progression to a rule of that kind. As stated in article 65: But if we say that the continuation of an analysis ad infinitum is possible, then it can at any rate be noticed whether one's progress in analysis can be reduced to some rule; from this a similar rule of progression will appear in the proof of complex tenns, which have as ingredients non complex temlS which can be analysed ad infinitum. 23

Let us consider once more the generic structure of existential propositions. These are defined as non-necessary, since they cannot be given an adequate resolution through the substitution of finite equivalents. Their analysis would unfold ad infinitum, because their requisites depend upon complete notions of individuals, be these substances or phenomena, and because those notions would in turn integrate the entire system of actualized compossibles. However this system inlplies a general sufficient reason for its integration. According to

202 FRAN<;OIS DUCHESNEAU Leibniz's methodological hypothesis, there could be a tentative access to this integrative reason through exhaustion of the individual requisites by means of a kind of transition to the limit. This technique of the transition to the limit borrowed from mathematics would afford an analogy for the manner of integrating an infinite analysis. The case at hand is that of Peter the denier. The proposition Peter denies presupposes a specific place and time for the corresponding truth to happen, and this involves the infinity of Peter's coexistents. Indeed, we may consider the proposition indefinitely abstracting from the very circumstances of time and place: in this second case as well the truth of the proposition must proceed from the concept of Peter, this time a vague Peter about whom we may be uncertain whether he denies or does not. Evidently that abstractive strategy falls short of accounting for the infinite elements contained in the complete notion of Peter. But, by proceeding in that fashion, underlines Leibniz, even though it is not possible to get an integral demonstration, "one always approaches it more and more, so that the difference is less than any given difference.,,24 Indeed, Leibniz resorts here to the metaphor concerning the summation of a sequence of infinitesimals by means of differentials. But is there something more to it than a mere metaphor? The same analogy is to be found in article 66, but it is suggested that we get nearer to a demonstration whenever we can set a progression law such that it would exclude any unresolved ingredient that contradicts it. How is this possible, but because the law would imply a continuous scheme and that it expresses an architectonic and infinite determination of the real? Leibniz has probably in mind something like his own principle of continuity, which forms a methodological presupposition in his physics. The law conceived this way ought to define the subject's requisites down to an infinite detail. This is the message the article considered seems to convey: But if, when the analysis of the predicate and the subject has been continued, a coincidence can never be proved, but it does at least appear from the continued analysis (and the progression and its rule which arise from it) that a contradiction will never arise, then the proposition is possible. But if, in analysing it, it appears from the rule of progression that the reduction has reached a point at which the difference between what should coincide is less than any given difference, then it will have been proved that the proposition is true. If, on the other hand, it appears from the progression, that nothing of this sort will ever arise, then it has been proved to be false .. 25

In articles 132-136, a set of correlative distinctions underlines the limits of validity pertaining to the a priori proof as well as to the analogy between determination laws and the algorithmic determination of unassignable quantities. The a priori proof is presented here as a capacity for furthering analysis according to a continuous scheme in infinite development. Leibniz presumes that contingent truths fit that structure of determination under the notion of a serial law, even if it cannot be strictly evidenced through

Leibniz and the Model for Contingent Truths 203 substitution of definitional equivalents. Only in God can this serial law be perfectly apprehended. So article 134 states: A true contingent proposition cannot be reduced to identical propositions, but is proved by showing that if the analysis is continued further and further, it constantly approaches identical propositions, but never reaches them. Therefore it is God alone, who grasps the entire infinite in his mind, who knows all contingent truths with certainty 26

As a consequence, Leibniz insists on both aspects of the analogy that the Specimen inventorum and the fragment on Necessary and Contingent Truths had separately highlighted: on the one hand, we can achieve some incomplete, or abstract, or provisional, or "analogical" demonstrations for truths of fact as for incommensurable quantities; on the other hand, since it is not possible to resolve terms integrally because of the infinite regress in conceptual ingredients, a strict reduction to identicals is unavailable in either case, if the objects to analyze are considered in their concreteness. Article 136 develops this twofold implication. Leibniz shows that we can rely on the a priori proof afforded by the principle of reason to work out architectonic principles, such as that of continuity, to account for the laws ruling over the domain of truths of fact. The model for those laws is provided by the mathematics of the infinite, which reduces discontinuities according to an optimal pattern of continuous propensities. Indeed, the use of architectonic principles, like that of continuity, is grounded on the notion that there is an a priori proof for every true proposition, even if contingent. But those principles are based on the postulated existence of laws of infinite series. These laws unveil themselves in contingent truths as well as in certain mathematical truths, whose status is provisionally necessary for finite understandings. Such a projection expresses the integrative rationality of various elements in the form of an hypothesis conformable to sufficient reason: But a difficulty stands before us. We can prove that some linenamely an asymptote-constantly approaches another, and (also in the case of asymptotes) we can prove that two quantities are equal, by showing what will be the case if the progression is continued as far as one pleases; so human beings also will be able to comprehend contingent truths with certainty. But it must be replied that there is indeed a likeness here, but there is not a complete agreement. Further, there can be relations which, however far an analysis is continued, will never reveal themselves sufficiently for certainty, and are seen perfectly only by him whose intellect is infinite. It is true that, as with asymptotes and incommensurables, so with contingent truths we can see many things with certainty, from the very principle that every truth must be capable of proof. Consequently, if all things are alike on each side in our hypotheses, there can be no difference in the conclusions-and other things of this sort, which are true both in the case of necessary and contingent propositions,

204

FRANCOIS DUCHESNEAU since they are reflexive. But we can no more give the full reason for contingent truths than we can follow asymptotes and run through infinite progressions of numbers. 27

I

In his Analysis und Synthesis bei Leibniz,28 Martin Schneider has tended to relativize Leibniz's serial law model for contingent truths. He believes that Leibniz's theory of knowledge needs to get beyond an analytic conception 0 judgments. Contingent truths would require a synthetic foundation, dependent upon the actualized unity of the object in experience, which is itself grounded in the reflexive experience of the self. According to Leibniz, the analytic determination in contingent truths would conform to a progression law analogous to that of mathematical series. But is this solution adequate? It would never reach a full analytic determination of the various states pertaining to contingent realities. With mathematical series, it suffices to know part of the sequence, i. e., a finite set of sequential elements to infer the progression law that applies beyond the scope of the portion analyzed. With the series of states in contingent realities, the law would only be determined by the totality of the elements comprised. Hence, an indefinite carrying over for our understanding. 29 Schneider insists that experience is called upon to compensate for the impossible unfolding of a law that would make for the entirety of states in a contingent reality. Experience warrants the objectivity of ideas of sensible qualities, which form primitive concepts for us; but it is also called upon to verify the possibility of contingent realities, whether in the descriptive analysis of particular phenomena or in the causal inference concerning sequences of phenomena. Indeed, as Schneider finds out, in the causal explanation about phenomena, Leibniz substitutes an indirect analysis (per circuitum) to the direct analysis which is out of reach for us: qualities, at once sensible and intelligible, embodying geometrical and mechanical properties are taken to express in an analogous orderly fashion the original sensible qualities in their effective complexity.30 But would this strategy mean a true replacement of empirical knowledge? To succeed one should arrive at properties from which all other properties could be deduced. Such definitions can only be provisional and must be revised as experience evolves. This is indeed the case with our definitions of gold: these are founded on sets of features within the compass of our empirical knowledge. If we succeeded in counterfeiting gold and if counterfeited gold possessed all those features identified through our discriminating experiences, we would need to discover further discriminating marks.31 For sure, Schneider acknowledges that in mathematics, setting progression laws. that involve the whole serial determinations is the way to prove that the corresponding object is possible. But, for him, when dealing with contingent realities, the model of converging mathematical series would fail, for we would need first to unfold the infinitely actualized structure underpinning those \realities.

Leibniz and the Model for Contingent Truths 205 In light of the principal Leibnizian texts I have analyzed, Schneider's sceptical arguments seem to carry too much. He onlits two major aspects of Leibniz's theory: these should be highlighted. The first is the doctrine of expression which nlakes it possible to transcribe the connection of phenomena in nlathematical language. Indeed, sensible qualities as such remain irreducible to rational determinations. But Leibniz believes that we can build intelligible models for phenomena by relying on the mathematical (geometrical) features correlative of phenomenal properties. We nlay thus apply the expressive resources of appropriate algorithms to the connection evidenced by phenomena; and this expression can make it possible to anticipate on discovering further phenomenal sequences. The second element to take account of is the method of hypotheses Leibniz resorts to for explaining phenomena. The notion of an hypothesis is not that of an a priori conjecture, representing an abstract and, as it were, arbitrary model. For Leibniz, an hypothesis must be justified as anticipating a system of determining reasons for the phenomenon or the set of phenomena to explain. How do we judge an anticipation valid? The criteria involved combine the following methodological requirements: (1) analytic precision of the model chosen: hence the need to resort to an expressive system that fits the logical requirement of the mathesis; (2) empirical corroboration of the connections of phenomena symbolized, at the initial stage, as well as subsequently when hypothetico-deductive inferences are experimentally controlled; (3) compatibility of the system of sufficient reasons with the architectonic principles that guide theory building and indicate the type of order contingent realities are supposed to conform to. These principles represent the inaccessible law that rules over the unfolding of infinitely diversified states. They also provide the rational frame for those models we resort to so as to mean the inner and irreducibly implicit rationality of contingent truths.

III. Conclusion The analogy with the laws of converging series which Leibniz borrows from the mathematics of the infinite to represent the structure of contingent truths needs not be the deceitful paralogism that has sometimes been denounced. Leibniz shows that we may rely on the a priori proof of contingent truths to work out some law projections specifying series of determinations. Even though these laws are ultimately grounded a posteriori, they can deploy ex hypothesi an analytic scheme that may account for analogical connections down to an infinite detail. Contrary to M. Schneider, I do not hold that Leibniz went astray in drawing an unwarranted analogy between resolutions by means of the infinitesinlal calculus and the proof structure for contingent truths. By so doing, Leibniz was attenlpting to establish how this type of truths and the propositions used to signify them could be conceived to obey a system of

206 FRANCOIS DUCHESNEAU analytic sufficient reasons ad infinitum: for this purpose, analogous models were available and could be resorted to, which underlined the analytic intelligibility of contingent truths dans leur ordre, so to say. This flexibility in the analytic model for contingent truths has been unfortunately lost sight of only to be replaced by Kant's more rigid distinctions about analytic and synthetic judgments. There might be considerable benefit to reap from reexploring Leibniz's pre-Kantian schemes, and their close linkage with the scientific methodology he was promoting. NOTES 1. G. W. Leibniz, Philosophical Papers and Letters (abbreviation: L followed by page number(s», ed. by L. E. Loemker (Dordrecht: Reidel, 1969), p. 646; G. W. Leibniz, Die philosophischen Schriften (abbreviation: GP, followed by volume and page nUlYtber(s)), hrsg. von C. 1. Gerhardt (Hildesheim: G. alms, 1965), VI: 612. 2. G. W. Leibniz, Monadologie, §36 (L p. 646); GP VI p. 612: "la suite des choses n~pandues par I'univers des creatures." 3. Louis Couturat, La logique de Leibniz (Hildesheim: G. alms, 1969), p. 209: "La raison en est que toute verite est detenninee par la nature logique de ses tennes, qu' elle y est en quelque sorte inscrite d' avance et qu'il suffit de les analyser it fond pour l'y decouvrir." 4. G. W. Leibniz, New Essays on Human Understanding (abbreviation: NE followed by book, chapter, and section nUlYlber(s», ed. by P. Remnant and 1. Bennett (Cambridge: Cambridge University Press, 1981), 4.2.1, p. 367; G. W. Leibniz, Samtliche Schriften und Briefe (abbreviation: A) (Berlin: Akademie-Verlag, 1923-... ), VI, vi, p. 367. 5. Claude Troisfontaines, "L'approche logique de la substance et Ie principe des indiscemabIes," Leibniz: Questions de logique, hrsg von A. Heinekamp, Studia Leibnitiana, Sonderheft 15 (Stuttgart: F. Steiner, 1988), pp. 94-106: here, p. 97: "[Le] concept [de l'individu est] l'integrale de tous ses predicats." 6. G. W. Leibniz, Discours de metaphysique, §8 (L p. 308; GP IV p. 433). 7. G. W. Leibniz, Letter to Arnauld, 4/14 July 1686 (GP II p. 52): " ... enveloppe sub ratione possibilitatis les existences ou verites de fait ou decrets de Dieu, dont les faits dependent. " 8. Leibniz, Discours de metaphysique, §13 (GP IV p. 438). 9. Leibniz, Letter to Arnauld 4/14 July 1686 (L p. 334; GP II pp. 52-53). 10. Leibniz, Monadologie, §36 (L p. 646; GP VI pp. 612-613). 11. Franyois Duchesneau, Leibniz et la methode de La science (Paris: Presses Universitaires de France, 1993); Franyois Duchesneau, La dynamique de Leibniz (Paris, Vrin,1994). 12. On the notion of hypothetically necessary truths in Leibniz, cf. Robert McRae, Leibniz. Perception, Apperception, and Thought (Toronto: University of Toronto Press, 1976), especially, pp. 121-25. Account should also be taken of Hide Ishiguro, "Leibniz on Hypothetical Truths," Leibniz: Critical and Interpretive Essays, Michael Hooker, ed. (Manchester: Manchester University Press, 1982), pp. 90-101.

Leibniz and the Model for Contingent Truths

207

13. Bertrand Russell, A Critical Exposition of the Philosophy of Leibniz, 2nd ed. (London: George Allen and Unwin, 1937\ Couturat, La logique de Leibniz. These works were originally published in 1900 and 1901. 14. F. E. Andrews, "Leibniz's Logic within his Philosophical System," Dionysius 8 (1983) pp. 74-127. If her selection of texts is most relevant, her views about the successive stages remain questionable. 15. G. W. Leibniz, Nouvelles lettres et opuscules inedits de Leibniz (abbreviation: F de C followed by page number(s)), ed. par L. A. Foucher de Careil (Hildesheim: G. OlIns, 1971), pp. 178-85 (L pp. 263-66). 16. G. W. Leibniz, Opuscules et fragments inedits (abbreviation: C), ed. par L. Couturat (Hildesheim: G. Olms, 1988), pp. 401-3. 17. C p. 402: "Ac proinde necesse est quandam inter notiones tenninorum esse connexionem, sive fundamentum a parte rei ex quo ratio propositionis reddi, seu probatio a priori inveniri possit." In this regard, Couturat refers also to Generales inquisitiones, §§ 132-36 (C pp. 388-89); to the fragment on Necessary and Contingent Truths (C pp. 16-24); and to the fragment Primce veritates (C pp. 518-23; L pp. 267270). 18. G. W. Leibniz, Vorausedition (for A VI, iv)(abbreviation VE followed by page number(s)) of III, p. 482 (GP VII p. 309): "Et principium reddendre rationis quod scilicet omnis propositio vera, qure per se nota non est, probationem recipit a priori, sive quod omnis veritatis reddi ratio potest, vel ut vulgo ajunt, quod nihil fit sine causa. Hoc principio non indiget Arithmetica et Geometria, sed indiget Physica et Mechanica, eoque usus est Archimedes. 19. VE III pp. 482-83 (GP VII, p. 309): "Essentiale est discrimen inter Veritates necessarias sive retemas, et veritates facti sive contingentes, differuntque inter se propemodum ut numeri rationales et surdi. Nam veritates necessarire resolvi possunt in identicas, ut quantitates commensurabiles in communem mensuram,. sed in veritatibus contingentibus, ut in numeris surdis, resolutio procedit in infinitum, nec unquam tenninatur; itaque certitudo et perfecta ratio veritatum contingentium soli DEO nota est, qui infinitum uno intuitu complectitur. Atque hoc arcana cognito tollitur difficultas de absoluta omnium rerum necessitate, et apparet quid inter infallibile et necessarium intersit." 20. VE III p. 456 (C pp.17-18): "Itaque Veritates contingentes ad necessarias quodammodo se habent ut rationes surdre, numerum <scilicet> incommensurabilium, ad rationes effabiles numerorum commensurabilium. Ut enim ostendi potest Numerum minorem alteri majori inesse, resolvandum utrumque ad maximam communem mensuram, ita et propositiones essentiales seu veritates demonstrantur, resolutione instituta donec perveniatur ad tenninos quos utrique tennino communes esse, ex definitionibus constat. At quemadmodum Numerus major alterum incommensurabilem continet quidem, licet resolutione utcunque in infinitum continuata, nunquam ad communem mensuram perveniatur, ita in contingente veritate, nunquam pervenitur ad demonstrationem quantumcunque notiones resolvas. Hoc solum interest, quod in rationibus surdis nihilominus demonstrationes instituere possumus, ostendendo errorem esse minorem quovis assignabili, at in Veritatibus contingentibus ne hoc quidem concessum est menti creatre. Atque ita arcanum aliquod a me evolutum puto, quod me ipsum diu perplexum habuit; non intelligentem, quomodo prredicatum subjecto inesse posset, nec tamen propositio fieret necessaria. Sed cognitio rerum Geometricarum atque

208

FRANCOIS DUCHESNEAU

analysis infinitorum hanc mihi lucem accendere, ut intelligerem, etiam notiones in infinitum resolubiles esse." 21. G. W. Leibniz, Generales inquisitiones, §56, Logical Papers (abbreviation P followed by page nUIuber(s)), ed. by G. H. R. Parkinson (Oxford: Clarendon Press 1966), pp. 60-61; VE III p. 1969 (C p. 371): "Hinc sequitur, ut certi simus veritatis, vel continuandam esse resolutionem usque ad prima vera , vel demonstrandum esse ex ipsa progressione resolutionis, seu ex relatione quadam generali inter resolutiones prrecedentes et sequentem, nunquam tale quid occursurum, utcunque resolutio continuetur. Hoc valde memorabile est, ita enim srepe a longa continuatione liberari possumus. Et fieri potest, ut resolutio ipsa literarum aliquid circa resolutiones sequentium contineat, ut hie resolutio Veri. Dubitari enim potest an omnem resolutionem finiri necesse sit in primo vera seu irresolubilia inprimis in propositionibus contingentibus, ut sci!. ad identieas reduci non vacet." 22. Leibniz, Generales inquisitiones, §61, P p. 62; VE VIII p. 1970 (C p. 372): "Naln probare verum esse tenninum complexum est eum reducere in alios terminos complexos veros, et hos tandem in tenninos complexos primo veros, hoc est axiomata (seu propositiones per se notas), definitiones tenninorum incomplexorum quos probatum est esse veros; et experimenta." 23. Leibniz, Generales inquisitiones, §65, P p. 63; VE VITI p. 1971 (C pp. 373-74): "Quodsi dicamus possibilem esse continuationem resolutionis in infinitum, tunc saltern observari potest, progressus in resolvendo an ad aliquam regulam reduci possit, unde et in tenninorum eomplexorum, quos incomplexi in infinitum resolubiles ingrediuntur, probatione talis prodibit regula progressionis." 24. Leibniz, Generales inquisitiones, §74, P p. 67; VE VIII p. 1974 (C p. 377): "[.. .] attarnen semper magis magisque acceditur, ut differentia sit minor quavis data." 25. Leibniz, Generales inquisitiones, §66, P pp. 63-64; VE VIII p. 1972 (C p. 374): "Quodsi jam continuata resolutione prredicati et continuata resolutione subjecti nunquarn quidem demonstrari possit coineidentia, sed ex eontinuata resolutione et inde nata progressione ejusque regula saltern appareat nunquam orituram contradictionem, propositio est possibilis. Quodsi appareat ex regula progressionis in resolvendo eo rem reduci, ut differentia inter ea qure coincidere debent, sit minus qualibet data, demonstratunl erit propositionem esse veram." 26. Leibniz, Generales inquisitiones, §134, P p. 77; VE VIII pp. 1985-86 (C p. 388): "Propositio vera contingens non potest reduci ad identicas, probatur tamen, ostendenrlo continua magis magisque resolutione, accedi quidem perpetuo ad identicas, nunquam tamen ad eas perveniri. Unde solius DEI est, qui totum infinitum Mente complectitur, nosse certitudinem contingentium veritatum." 27. Leibniz, Generales inquisitiones, §136, P, pp. 77-78; VE, VIII p. 1986 (C pp. 388-89): "At difficultas obstat: possumus nos demonstare lineam aliquam alteri perpetuo accedere, licet Asymptotam, et duas quantitates requales esse, etiam in asymptotis, ostendendo progressione utcunque continuata, quid sit futurum. Itaque et homines poterunt assequi certitudinem contingentium veritatum; sed respondendum est, similitudinem quidem esse, omnimodam eonvenientiam non esse. Et posse esse respectus, qui utcunque continuata resolutione, nunquam se, quantum ad certitudinem satis est, detegant, et non nisi ab eo perfecte perspiciantur, cujus intellectus est infinitus. Sane ut de asymptotis <et incommensurabilibus>, ita et de contingentibus multa certo perspicere possumus, ex hoc ipso principio quod veritatem omnem oportet

Leibniz and the Model for Contingent Truths 209 probari posse, unde si omnia utrobique se habeant eodem modo in Hypothesibus, nulla potest esse differentia in conclusionibus, et alia hujusmodi, qure tam in necessariis quam contingentibus vera sunt; sunt enim reflexiva. At ipsam contingentium rationem non magis possumus, quam asymptotas perpetuo persequi et numerorum progressiones infinitas percurrere." 28. Martin Schneider, Analysis und Synthesis bei Leibniz, PhD (Bonn, 1974). 29. Martin Schneider, Analysis und Synthesis bei Leibniz, PhD (Bonn, 1974) p. 156: "Solche Ordnungsgesetze aber lassen sich [... J gerade nur aufgrund der vollstandigen Kenntnis der gesamten Fo1ge von inhaltlichen Konstituentien, nicht aber bereits (wie mathematische Bildungsgesetze) aufgrund der Kenntnis nur eines Teils der Glieder der unendlichen Folge. Daher sind solche Ordnungsgesetze nur von Gott angebbar und k6nnen somit nie efIektiv (d. h. vom Menschen) angegeben bzw. aufgestelt werden." 30. Concerning this point, Schneider refers to C pp. 190,360, and GP VII p. 293. 31. Cf. NE 3.4.16 (A VI vi pp. 299-300); 3.6.17 (A VI vi pp. 311-312); 4.6.4 (A VI vi 403-4).

Burne on Possible Objects and Impossible Ideas PHILLIP D. CUMMINS

a. Possible Objects Hume frequently invokes possible objects or states of affairs in order to deny the necessity, the intuitive certainty, or the demonstrability of universal principles. An example is his treatment of the causal maxim, whatever begins to exist has a cause. Contrary to what is commonly assumed, he claimed, it is not a necessary truth and so can be neither intuitively nor demonstratively certain. He wrote: We can never demonstrate the necessity of a cause to every new existence, or new modification of existence, without shewing at the same time the impossibility there is, that any thing can ever begin to exist without some productive principle; and where the latter proposition cannot be prov'd, we must despair of ever being able to prove the former. Now that the latter proposition is utterly incapable of a demonstrative proof, we may satisfy ourselves by considering, that as all distinct ideas are separable from each other, and as the ideas of cause and effect are evidently distinct, 'twill be easy for us to conceive an object to be non-existent this moment, and existent the next, without conjoining to it the distinct idea of a cause or productive principle. The separation, therefore, of the idea of a cause from that of a beginning of existence, is plainly possible for the imagination; and consequently the actual separation of these obj ects is so far possible, that it implies no contradiction nor absurdity; and is therefore incapable of being refuted by any reasoning from mere ideas; without which 'tis impossible to demonstrate the necessity of a cause. 1

The causal maxim makes a claim about all things that begin to exist. Hume argues against its necessity so universalized by showing that a falsifying counter-example, an object which begins to exist, but has no cause, is possible. To establish that possibility, he first makes two general points, arguing that if an idea is distinct from another, it is separable from it, and that the idea of a cause is distinct from the idea of an effect. 2 He next appeals to a situation in thought. One can entertain the idea of a new existent without conjoining to it the distinct idea of a second entity which is its cause. The possibility of this psychological situation demonstrates, Hume claimed, the possibility of the state of affairs it represents. Since the state of affairs, if actual, would be a counterexample to the causal maxim, Hume inferred the latter is not necessary. Since a demonstration of the causal maxim, if successful, would prove the impossibility of any such factual situation, Hume concluded such a demonstration is impossible. 3

212 PHILLIP D. CUMMINS The argument under consideration relies on a principle that Hume sometimes asserts with reference to the represented situation, claiming that what is possible in thought is possible in fact and reality. Stated with reference to thought and ideas, this is the thesis that if a psychological structure involving a perspicuous idea is possible, the situation represented by it or its component idea is possible in fact and reality. Here and throughout this essay I shall use "perspicuous" as a surrogate for Hume's favored adjectives: "clear," "distinct," and "adequate.,,4 To simplify, we may say that for Hume if a perspicuous idea of an object is possible, the object represented by that idea is possible. Let us call this the Perspicuous-Idea-to-Object Possibility Principle, PPP for short. Hume frequently used arguments based on PPP, especially in developing his account of causal inferences. One of his first conclusions about such inferences is that they are not a priori insights regarding a connection between two ideas. In supporting it he wrote, 'Tis easy to observe, that in tracing this relation, the inference we draw from cause to effect, is not derived merely from a survey of these particular objects, and from such a penetration into their essences as may discover the dependence of the one upon the other. There is no object, which implies the existence of any other if we consider these objects in themselves, and never look beyond the ideas which we form of them. Such an inference would amount to knowledge, and would imply the absolute contradiction and impossibility of conceiving any thing different. But as all distinct ideas are separable, 'tis evident there can be no impossibility of that kind. (T86-7)

A more familiar example is Hume's argument concerning the implications of experienced constant conjunctions among pairs of objects or events. He denied reason can infer causal connections from such constant conjunctions, since that would require intuitive or demonstrative knowledge of the principle "that those instances, of which we have had no experience, must resemble those, of which we have had experience, and that the course of nature continues always uniformly the same." (T89) However, there can be no such knowledge of this principle, since, We can at least conceive a change in the course of nature; which sufficiently proves, that such a change is not absolutely impossible. To form a clear idea of any thing, is an undeniable argument for its possibility, and is alone a refutation of any pretended demonstration against it. (T89i

Thus, arguments based on PPP permeate Hume's celebrated or infamous account of causation and causal inference. The topic was different, but the style of arguing was the same, when earlier in the Treatise, as part of his account of the composition of extension, space, and time, Hume dismissed, sight unseen, all purported demonstrations of the infinite divisibility of extension. He first claimed to derive from an impression of sensation a perspicuous idea of an indivisible part of extension. (T27) In the

Hume on Possible Objects and Impossible Ideas 213 next section he appealed to that idea in attempting to demonstrate that the infinite division of a finite extension is impossible. Just as he would later dismiss all attempted demonstrations of the causal maxim before considering them in detail, so at the end of Section Two he supported his position by offering a wholesale refutation of all purported demonstrations of infinite divisibility, some of which were soon to be criticized in detail in Section Four. He based the wholesale refutation on PPP, writing, 'Tis an establish'd maxim in metaphysics, That whatever the mind clearly conceives includes the idea of possible existence, or in other words, that nothing we imagine is absolutely impossible. We can fonn the idea of a golden mountain, and from thence conclude that such a mountain may actually exist. (T32)6

Hume, of course, used his metaphysical maxim not to authorize a search for a golden mountain, but to assure us of the possibility of ultimate parts of extension. Not surprisingly, then, Hume proceeded to insist that a perspicuous idea of an indivisible part of extension is possible. He was entitled to recall the actual idea of an indivisible from the preceding section and infer possibility from actuality~ instead, surprisingly, he employed a different and decidedly inferior argument. One component of it is worth our attention. Hume's first step in the new argument was to claim an idea of extension. For evidence he appealed to language, writing, "Now 'tis certain we have an idea of extension~ for othenvise why do we talk and reason concerning it?" (T32) Unfortunately, the underlying presupposition of the argument embedded in this rhetorical question was not one Hume could afford to hold, since it would warrant a telling objection to his proof of the impossibility of infinite divisibility. Earlier in the same section Hume had appealed to his supposedly perspicuous idea of an indivisible part of extension in order to argue that an infinite number of such parts would together yield an infinitely large extension. This result was taken to demonstrate the impossibility of the infinite divisibility of any finite extension. 7 But suppose we define "infinitesimal" as a magnitude which when multiplied by infinity yields a finite magnitude as product. We can then at least formulate an alternative to Hume's position, asking whether a finite extension can be composed of infinitesimals. We can then argue, following Hume, that since we can talk and reason about infinitesimals, we must have an idea of an infinitesimal, which proves by PPP that infinitesimals are possible, in which case Hume's purported demonstration of finite divisibility must itself be fallacious. Fortunately for Hume, we have good· reason to believe that in appealing to language to establish an idea of extension, he was merely being careless. A word deemed meaningful does not imply a perspicuous idea. In Section Three of Book One, Part Two, that is, in the very next section, Hume presented his own positive account of extension and claimed as a consequence of it that a vacuum, i. e., an extension which is not a body, is impossible. In Section Five he considered the objection, probably from Locke, that the intelligibility of the

214 PHILLIP D. CUMMINS dispute about whether or not a vacuum is possible proves decisively that it is, since it establishes the existence and thus the possibility of the idea of a vacuum, which in turn implies the possibility of a vacuum. 8 This is Hume's reply: The frequent disputes concerning a vacuum, or extension without matter, prove not the reality of the idea, upon which the dispute turns~ there being nothing more common, than to see men deceive themselves in this particular; especially when by means of any close relation, there is another idea presented, which may be the occasion of their mistake. (T62)

Although, given the obscurity of Hume's expression "close relation", his argument is not altogether obvious, it nonetheless suggests that for him although a spoken or written expression can be regarded as intelligible, it may still lack genuine representational meaning. Finding such meaning seems to involve isolating a perspicuous idea which is associated with the expression. The question is: what kind of investigation can establish the existence or nonexistence of a disputed idea? We have several. striking examples of how I-Iume answered this question. b. Impossible Ideas Two Humean constraints on the occurrence of ideas are familiar. A complex idea can occur only if each of its component simple ideas can occur. A simple idea can occur only if there has been prior experience of the object of that idea, that is, for I-Iume, if the subject has had the corresponding simple impression either in isolation or in combination with other simple inlpressions. (T4-5) We nlay sayan idea is psychologically impossible if the derivation condition for it or one of its component simple ideas cannot be satisfied. Psychological impossibility has various degrees. One who prior to a given time has never tasted pineapple cannot at that time have an idea of its savor, assuming the latter's simplicity. This is just a matter of how that person's experiences have gone. It is superficial in comparison with the situation of a blind person,who is not prevented from having an idea of purple merely by lack of exposure to purple objects, but by a permanent disabling condition. Further along the continuum is a situation invoked in one of Hume's arguments that we can have no idea of substance per se. Following Locke he allowed that one may call one's idea of a plum an idea of substance, since it is of a collection of qualities which we name by a noun and consider one thing. (T16) However, he flatly denied having an idea of an underlying unitary substrate in which the qualities inhere. One of his arguments for this claim is the following: As every idea is derived from a precedent impression, had we any idea of the substance of our minds, we must also have an impression of it; which is very difficult, if not impossible, to be conceived. For

Hume on Possible Objects and Impossible Ideas

215

how can an impression represent a substance, otherwise than by resembling it? And how can an impression resemble a substance, since, according to this philosophy, it is not a substance, and has none of the peculiar qualities or characteristics of a substance? (T232-3)

Hume is not just claiming no substance is an impression or is represented by an impression, he is denying the possibility of an impression which represents a substance. Deep conditions for representation cannot be satisfied given what an impression is and what a substance is supposed to be. Good argument or not, it illustrates a further step along the continuum of psychological impossibiiity. A situation this far or further along will shortly be considered as part of an examination of a trio of argunlents against the existence of abstract general ideas. In "Of Abstract Ideas, (Tl7) Hume reported he would follow Berkeley's lead on the question whether abstract ideas are "general or particular in the mind's conception o/them." This is a problem about how non-particular objects are represented in thought. It encompasses two questions: First, when one thinks of a quality or kind, what is the nature of one's object of thought, insofar as one has a distinct object of thought? Second, what is the nature of the representing entity (one's psychological state) when the object of thought is non-particular and how does it represent that object? Following the bad examples of Locke and Berkeley, Hume used "idea" for both the object of thought and the representing entity. Hence, when he states that Berkeley's new answer is "all general ideas are nothing but particular ones, annexed to a certain term, which gives them a more extensive signification," (T17) it is difficult to ascertain which of the two questions is being addressed. Despite this, Hume begins his defense of Berkeley's answer by asking how one's idea of, say, man, can represent a kind when the members of that kind are not completely uniform, but differ from one another in a variety of ways. One must in some way abstract from these differences, since as Hume put it, "...An object ceases not to be of any particular species on account of every small alteration in its extension, duration and other properties." He continued, "The abstract idea of a man represents nlen of all sizes and all qualities~ which 'tis concluded it cannot do, but either by representing at once all possible sizes and all possible qualities, or by representing no particular one at all." (T17-18) Here, surely, Hume is concerned about one's actual object of thought when one thinks of a kind and its relation to the many members of that kind, not the relation of a psychological state to a single abstract object of thought. He next insists that the received opinion is the second of the two options just specified. According to it, the objects of abstract thought are general in nature. One abstracts from the differences in detail among kind members to secure ideas which "represent no particular degree either of quantity or quality." (TI8) This position is widely accepted, Hume claimed, because, supposedly, its rival implies an impossible "infinite capacity in the mind" and so cannot be true. An

216 PHILLIP D. CUM1v1INS infinite capacity is required, presumably, because, on a theory which denies thoughts about a general object, that is, about what is common to all members of a kind, a person thinking of that kind must consider individually each of its possible members. Hume's attack on the received position was two-pronged. He tried to cut away its nlain prop by answering the above criticism of its rival, 9 but first he assaulted it directly, arguing that it is "utterly impossible" to conceive any quantity or quality without forming a precise "notion" of its degrees. This is to claim an idea of a general quantity or quality is impossible. I shall next examine Hume's direct assault, which consists of three argunlents. Each of those arguments appeals to something other than ideas in order to reach a conclusion about ideas. The third or last-stated reveals this pattern of development Inost strikingly. It begins: Everything in nature is individual. It is utterly absurd to suppose there is "a triangle really existing which has no precise proportion of sides and angles." (T19) Hume next asserted, "If this therefore be absurd in fact and reality, it must also be absurd in idea; since nothing of which we can form a clear and distinct idea is absurd and impossible." (T19-20) Note that he first claimed to know the situation in fact and reality: the existence of an indeterminate individual is impossible. From this he drew a conclusion regarding thought: One cannot have a perspicuous idea of an indeterminate, purely general thing. One interested in historical connections might note the similarity between this argument and an argument occurring in an unpublished draft of Berkeley's Introduction to his Principles ofHuman Knowledge. The latter is: It is, I think, a receiv'd axiom that an impossibility cannot be

conceiv'd. For what created intelligence will pretend to conceive, that which God cannot cause to be? Now it is on all hands agreed, that nothing abstract or general can be made to really exist. Whence it should seenl to follow, that it cannot have so much as an ideal existence in the understanding. 10

One, however, whose interest lies in the connections among Hume's arguments will notice that the major premise in his third argument is PPP, since "Nothing of which one can form a perspicuous idea is impossible" is logically equivalent to "Anything of which one can form a perspicuous idea is possible." Yet the argument here differs in form from those considered earlier. They were modus ponens; this is modus tallens. Beyond labels what is new is the direction in which Hume argued. Unlike the cases examined before, he here specified a state or situation as impossible in fact and reality, then concluded it is impossible to form a perspicuous idea of and so genuinely represent that state or situation. Given that in "Of Abstract Ideas" Hume was arguing that when one .thinks of an abstract object one cannot perspicuously represent a truly general object, the direction of his argument is not surprising. 11 It is very important to contrast the impossibility of conceiving a nonindividual entity with what was earlier called psychological impossibility. It is, according to Hume, merely a matter of fact that a causal condition for having a

Hume on Possible Objects and Impossible Ideas 217 simple idea is having a prior corresponding simple impression. Hence, even if it is categorically impossible to have an impression of a substance or a substance as an impression, it is for Hume only a contingent truth that it is psychologically impossible to have a simple idea of substance. In contrast, PPP is a maxim of metaphysics.. The possibility of a state of affairs which can be perspicuously conceived is what Hume called absolute possibility and the impossibility of perspicuously conceiving an absolute impossibility is itself absolute. Even if one could form a simple idea without having a prior corresponding impression, one could not perspicuously conceive an absolute impossibility. This point is especially worth making since at least one of Hume's remaining two arguments against general ideas is formulated in terms of psychological impossibility. Before considering those arguments, I want to formulate an objection to Hume's use of the premise that only individuals exist. It is that he has totally ignored the possibility that even though in one sense only individuals really exist, they may possess or even be partially constituted by qualities some or all of which are general in their nature. This concession-only determinate individuals can really be said to exist, and rebuttal-nevertheless, general qualities exist, is founded on a distinction between ontologically dependent and ontologically independent entities. This distinction permits one to affirm nonautonomous beings which exist, but only as constituents or properties of other, self-subsistent, beings, which alone are counted as really existing. Taking it seriously, one might hold, for example, that although all lines are individual and determinate, sound philosophical analysis requires acknowledgement of general or indeterminate qualities, such as length. One could then conclude that abstract ideas represent such dependent entities and thus are distinct from ideas representing individuals. This undiscussed alternative also imperils Hume's second argument, which employs the familiar Humean claim that ideas at bottom are derived from and exactly copy impressions, differing from them only in force and vivacity. This premise allows Hume to move from an alleged truth about impressions (primary objects of experience) to a conclusion about ideas. The truth alleged is that "...no object can appear to the senses~ or in other words, ...no impression can become present to the mind, without being determined in its degrees both of quantity and quality. (T19) This, Hume held, is a necessary truth~ to deny it is to be guilty of contradiction. 12 The argument continues, "Now since all ideas are deriv'd from impressions and are nothing but copies and representations of them, whatever is true of the one must be acknowledg'd concerning the other." Hume concluded, "An idea is a weaker impression~ and as a strong impression must necessarily have a determinate quantity and quality, the case must be the same with its copy or representative." (TI9) Since the idea of an indeterminate quality, say length in general, is denied by appealing to the derivation principle, it is psychologically impossible. Unlike the cases of pineapple

PHILLIP D. CUMMINS deprivation and blindness from birth, but like the substance example, Hume here denies the possibility of a certain category of impression. An indeterminate impression is said to be impossible in principle. This conclusion is unsecured, of course, since Hume's argument, as it stands, fails to consider the supposition introduced above that although the object (impression) itself is necessarily determinate, it has or is partially constituted by dependent indeterminate qualities which can be represented by ideas. In fairness to Hume, though, it should be noted that this key supposition had already been rejected in his first argument, if only by fiat. Hume initiated that first argument by announcing several principles concerning difference, distinguishability, and separability in thought. He wrote, 218

We have observ'd, that whatever objects are different are distinguishable, and that whatever objects are distinguishable are separable by the thought and imagination. And we may here add, that these propositions are equally true in the inverse, and that whatever objects are separable are also distinguishable, and that whatever objects are distinguishable are also different. For how is it possible we can separate what is not distinguishable, or distinguish what is not different? (TI8)

Hume here asserted two biconditionals and implied a third. I call them the Difference Principles. They are, first, an object is different from another if and only if it is distinguishable from the other, second, an object is distinguishable from another if and only if it is separable in thought and imagination from the other, and, third, an object is different from another if and only if it is separable in thought and imagination from the other. 13 Since for one object to be separable in thought and imagination from another undoubtedly is for the idea of the former to be separable from the idea of the latter, distinguishability surely is to be understood with reference to ideas. Hume's first Difference Principle thus amounts to or implies the claim that an object is different from another if and only if any idea which perspicuously represents the former is distinct from any idea which perspicuously represents the latter. The way in which Hume formulated these principles and thereby opened them to a telling objection may reflect his deep commitment to a principle to which he would soon appeal in rejecting infinite divisibility: "Existence in itself belongs only to unity."14 As formulated, the second and third biconditionals invite the objection that they fail for every entity which is composed of parts. A whole is different and distinguishable from each one of its parts, yet strictly speaking cannot be separated in thought or reality from any of those parts. If J comprises K and L, L is both different and distinguishable from J, yet one's proper idea of J includes an idea of L. Having an idea of J without thereby having an idea of L is impossible, which is to say J cannot be conceived separately from L. This objection could have been circumvented had Hume stated his principles so as to exempt wholes in relation to their parts. He was well aware, as Donald Baxter has emphasized, that a whole is inseparable from

Hume on Possible Objects and Impossible Ideas 219 its parts in the sense that its existence and continued existence presupposes the existence or continued existence of its parts. IS Yet Hume acknowledged no problem for his Difference Principles and made no appropriate modifications of them. My explanation is that when thinking metaphysically Hume habitually thought in terms of true unities and principles that hold for them even if inapplicable to ontologically suspect wholes. If this is so, Hume's Difference Principles are principles about simples and the simple ideas representing them. I return to Hume's first argument, which continues, 'Tis evident at first sight, that the precise length of a line is not different nor distinguishable from the line itself; nor the precise degree of any quality from the quality. These ideas, therefore, admit no more of separation than they do of distinction and difference. They are consequently conjoined with each other in the conception; and the general idea of a line, notwithstanding all our abstractions and refinements, has in its appearance in the mind a precise degree of quantity and quality; however it may be made to represent others, which have different degrees of both. (T18-9)16

In language we distinguish between a line, its precise length, and length in general, and formulate claims or questions about them. One such question is Hume's question about abstract ideas. Thus one can be said to have an idea of a line, an idea of its precise length, and an idea of length in general. But do we have a distinct idea of each? Can each be perspicuously represented as an entity or possible existent? One alternative is that we have three ideas, each of which perspicuously represents one of these objects. A rival alternative is that there is only one idea, which, however, can function in different ways and, so, for purposes of discourse, can represent the line-an individual, its precise length-a determinate quality, and length in general-an abstract or determinable quality. As I interpret Hume in the above passage, he rejects the first alternative and embraces the second, making two ontological claims: The first is that the precise length of a line is not different from the line itself and so cannot be distinguished from it. The second, which is more relevant to Hume's concern with whether abstract ideas are general in nature, is that the precise degree of any quality (e. g., the precise length of the line, that is, given the first claim, the line itself) is neither different nor distinguishable from the quality (length in general). There are not three different entities (line, precise length, length in" general) to be distinguished and, thus, perspicuously represented by different ideas. Observing the line cannot yield two perspicuous ideas, an idea of a line, and a second idea of its precise length, let alone a third idea of length in general. Why not? Because there is nothing there beyond one entity, an ontological simple, as it were. I7 Since there is but one idea, one cannot perspicuously isolate in thought either the precise length of the line or length in general. Hume implicitly admitted in the last passage quoted that we are entitled to say in some sense that lines do differ from one another in length and qualities

220 PHILLIP D. CUMMINS do differ from one another in their precise degrees. This acknowledgement was not as such an admission of inconsistency; it merely established the minimal condition for there being a problem about abstraction and indicated the need for Hume to explain how there can be distinctions at the level of language which have a foundation neither in ideas nor in the objects of those ideas. It set him the task of showing how we are able in some way to abstract qualities from their instances and general types of qualities from their precise degrees. For him to succeed, he was not required to embrace the first alternative; he merely needed to show that the linguistic distinctions we do draw can still be drawn on the second alternative. That becomes Hunle's task in the later passage (T24-5) on distinctions of reason, which might better be called distinctions of discourse. I shall conclude this examination of Hume' s first argument against general ideas by asking whether the impossibility claimed for a perspicuous idea of a length, as distinct from the line, is absolute impossibility or merely contingent psychological impossibility. It seems quite clear that Hume is not appealing to what he regarded as a contingent matter in insisting upon the ontological sinlplicity of the line. His is a categorical clainl. Not only is he committed to the thesis that only individuals exist, a key premise of the third argument, but also to the denial of non-autonomous entities such as specific lengths and length in general. I-lis argument is thus an attempt to show the impossibility of perspicuous ideas of specific qualities and of indeterminate qualities on the ground that there is and can be nothing to be so represented. Once again the impossibility of an entity or state of affairs is the ground for denying the possibility of an idea of that entity or state of affairs. However, there is still the question of the status of the Difference Principles, which bridge the gap between the ontological and psychological situations. Are they on a par with PPP or are they offered as contingent psychological principles conlparable to the derivation principle for simple idea? I have no firm answer. Admittedly, the Difference Principles look like they are stating factual conditions for separating two ideas. Still, they seem to function more like metaphysical principles. Moreover, the question, how can one thing be different from another if is impossible to distinguish it from the other, looks like a metaphysical, not a psychological question. It is quite possible that Hume regarded them as on the same footing with PPP and merely disguised them as factual principles. He was to show throughout Book One of the Treatise that he had no qualms about using apparently factual causal claims to do epistemological and metaphysical work. I8 c. Language, Ideas, and Method In the remainder of this essay I shall offer some reflections on the argument patterns under examination. The first is that there is a kind of ontology founded upon features of linguistic structures to which Hume is d~~E~ElI_I!~~tile~_~Q~~---

Hume on Possible Objects and Ilnpossible Ideas 221 philosophers want to use syntactical or semantical distinctions at the level of intelligible language as evidence of category distinctions that hold for all actual and possible things. For such philosophers the intelligibility and difference in meaning of the expressions "the line", "the line's exact length" and "length" reveal an ontological distinction between individuals, determinate qualities, and determinable qualities. Not so for Hume. If it implies anything at all, the meaningfulness of two or more linguistic expressions of different types proves, at most, that there is at least one perspicuous idea associated with the expressions. Mere linguistic diversity does not imply diversity at the level of perspicuous ideas. The case of substance can be used to illustrate Hume's methodological hostility to language-based ontology. His contemporary, Thomas Reid, argued that the subject-predicate form of sentences in all languages showed that the distinction between substance and attribute is well-founded. Reid also endorsed the traditional distinction between ontologically autonomous entities, substances, which can exist by themselves, and dependent entities, attributes, which can exist only as attributes of substances. 19 In contrast, Hume insisted that endorsing substance requires that one first exhibit a perspicuous idea of substance, a task he found to be impossible or nearly so. Moreover, he invoked PPP and the Difference Principles to reject the notion of dependent entities and with it any hope of providing a relative idea of substance as a substitute for a perspicuous idea of substance. To the definition of "substance" as "something which may exist by itself," a definition that might be thought to provide a relative idea, Hume objected "that this definition agrees to every thing, that can possibly be conceiv'd~ and never will serve to distinguish substance from accident, or the soul from its perceptions." He continued: For thus I reason. Whatever is clearly conceiv'd may exist; and whatever is clearly conceived, after any manner, may exist after the same manner. This is one principle, which has been already acknowledg'd. Again, every thing, which is different, is distinguishable, and every thing which is distinguishable, is separable by the imagination. This is another principle. My conclusion from both is, that since all our perceptions are different from each other, and from every thing else in the universe, they are also distinct and separable, and may be consider'd as separately existent, and may exist separately, and have no need of anything else to support their existence. They are, therefore, substances, as far as this definition explains a substance. (T233)

Here the full linkage between PPP and the Difference Principles is made apparent for the sake of showing that the definition of substance as that which can exist alone is satisfied by anything of which one can have a clear and distinct idea. The reverse side of this coin is that the expression "dependent entity", meaning "necessarily dependent entity", is inconsistent .if an entity is understood to be something that can be represented by a perspicuous idea.

222 PI-IILLIP D. CUMMINS Let us examine Hume's methodology further. It would appear that he wants to deny that a difference· in a meaningful linguistic structure is sufficient to secure an ontological distinction. A difference at the level of perspicuous idea must also be established. This commitment suggests that for Hume the proper philosophical methodology is to survey first one's perspicuous ideas, then draw conclusions about what is possible in fact and reality. Before this suggestion is pursued further, a caveat must be entered. One must be careful not to think that for Hume, as for Locke, meaning is completely fixed at the level of ideas and that words merely acquire meaning by being associated with ideas. Quite the contrary, whereas for Locke there are general ideas prior to and independently of language, for Hume an idea can become general only when associated with terms applied commonly to many individuals on the basis of their resemblances to one another. This is no subsidiary position taken by Hume~ it is central to his alternative solution to the problem which gave rise to the established position on abstract ideas. One does not need an idea representing a general entity comnlon to many individuals because when associated with a name an idea of an individual thing can "be applied beyond its nature" to represent all possible degrees of quantity and quality found among members of the kind to which it belongs. Hume explained this as follows: When we have found a resemblance among several objects, that often occur to us, we apply the same name to all of them, whatever differences we may observe in the degrees of their quantity and quality, and whatever other differences may appear among them. After we have acquired a custom of this kind, the hearing of that name revives the idea of one of these objects, and makes the imagination conceive it with all its particular circumstances and proportions. But as the same word is suppos'd to have been frequently applied to other individuals, that are different in many respects from that idea, which is immediately present to the mind; the word not being able to revive the idea of all these individuals, only touches the soul, if I may be allow'd so to speak, and revives the custom, which we have acquir'd in surveying them. (T20)

The custom permits the mind to survey and employ as needed ideas of other individuals to which the word applies. Hence, the idea actually occurring in thought can be applied beyond its nature and nlay be said to represent a kind or quality, though not perspicuously. Since this position implies that one cannot think of kinds and qualities in general independently of language, one cannot simply equate thinking and having ideas. One therefore needs to emphasize that Hume does not link the thinkable to the possible; only perspicuous ideas, are so linked. Let us return to that linkage, this time with an eye on Hume's method of philosophizing. As has been shown, usually to challenge the epistemological status of some principle or other, he more than once claimed to have an adequate or perspicuous idea,

Hume on Possible Objects and Impossible Ideas 223 invoked PPP, and concluded that a certain state of affairs was possible. 20 In such cases the direction of inference was from possible idea of object to possible object. However, the arguments regarding general ideas revealed an alternative pattern of argument, since in at least one of them Hume began by insisting that something, e.g., a non-individual entity, was impossible, invoked PPP, and concluded that there can be no perspicuous idea of such an entity. This argument, formally considered, is just as valid as the arguments from the possibility of the idea to the possibility of what it represents. However, from the methodological standpoint one might clainl something is amiss. Is there not something unsound in sometimes using actual or possible ideas as touchstones for drawing conclusions about what is possible in fact and reality and at other times using claims about what is impossible in fact and reality to rule out the very possibility of certain types of ideas?21 Must one not stick with one's introspective procedures or abandon them altogether? Or, if that is too stringent a requirement, must not one at least always begin either with arguments from ideas or with arguments from objects? If one ·decides Hume cannot argue in both directions and remain methodologically sound, then obviously the next question is to which direction should he restrict himself? The case for always starting with actual or possible ideas is initially attractive. Such a procedure seems experimental and modest. In contrast, starting with pronouncements about what is impossible seems far more dogmatic. For example, had Hume begun his treatment of the composition of extension with the proclamation that infinitesimals are inlpossible, so cannot be represented by a perspicuous idea, one would feel strongly tempted to insist he started at the wrong end. Still, beginning with ideas also has its problems. Such a procedure might be disabling and even, perhaps, misleading. It would seem to be disabling because while actually having a perspicuous idea of something shows the possibility of both the idea and, given PPP, its object, lack of an idea leaves open all the interesting questions, since the idea in question may still be possible. Further, the method of finding ideas and inferring possibilities can work to block uncritical assertions about impossibilities, but can do little else. The idea-first method is also likely to mislead because its empirical character encourages the belief that it is free from metaphysical presuppositions when it is not. Using it, one might claim to be modestly examining one's own ideas in order to block metaphysical assertions. This is to forget that PPP, one's bridge from psychological premises to metaphysical conclusions, is itself a metaphysical principle. This was not lost on Hume, who introduced it as an established maxim of metaphysics. One might do better, as a Humean, to begin with one or more fundamental metaphysical principle and augment them with the psychological principles needed to warrant inferences from perspicuous ideas to possible objects. One could then proceed to examine ideas. What might those fundamental metaphysical principles be? Consider as candidates two theses which together

224

PHILLIP D. CUMMINS

constitute what I call the Doctrine of Radical Contingency. The first is that whatever exists exists contingently. We shall not be further concerned with it. The second is that nothing that exists is necessarily connected to any other existent. 22 Hume sometimes asserted it as a general ontological or metaphysical claim, as when he asserted, in a passage quoted earlier, "There is no object, which implies the existence of any other if we consider these objects in themselves..." (T86) If this holds at all, it holds for all objects. Consequently, it holds for all immediate objects of experience (impressions) and implies that none is strictly inseparable from any other, even if they are never experienced apart. Ideas, which represent actual and possible impressions, are equally separable from one another, even if they are never thought apart from one another. Just because an idea is itself an existent, a perception, the separability thesis applies to it. Hume reached the same result by means of his copy principle, but also recognized the direct metaphysical route. He wrote: Nor will this liberty of the fancy appear strange, when we consider, that all our ideas are copy'd from our impressions, and that there are not any two impressions which are perfectly inseparable. Not to mention, that this is an evident consequence of the division of ideas into simple and complex. Where-ever the imagination perceives a difference among ideas, it can easily produce a separation. (TID.)

Since simple ideas presuppose their objects as prior objects of experience and· every actual object of experience exists and so is possible, we have some basis for holding PPP. On this line of thought, Hume's fundamental commitment regarding contingent connections has one negative consequence: metaphysically (J~r)~IIC:I­ ent entities are impossible. To say that something is a genuine entity is to deny that it is absolutely impossible for it to exist in the absence of some other entity. This perhaps explains Hume's confident and vehement denial of a distinction between a line and its length. To grant the distinction in fact and reality, for Hume, would be to grant that the specific length of a line could exist by itself in the absence of any or all lines. If, as I suspect, this for Hume would be absurd or impossible, it is no wonder he denied the distinction. NOTES 1. Hume, David, A Treatise of Human Nature, Second revised edition, edited by L.A. Selby-Bigge and P. H. Nidditch, (Oxford, Clarendon Press, 1978), Book I, Part iii, Section 3, pp. 79-80. Hereafter most references to the Treatise will be given in the text, abbreviated by T followed by page number(s). 2. Hume is careless in asserting the distinctness and thus the separability of cause and effect. Later (T I.iii.3, p. 82) he concedes that cause and effect are correlatives, so that if something is a cause, something else must be its effect. The converse also holds. As correlatives, cause and effect are clearly inseparable. Hume was quick to note that it does not follow from this that any given thing must be an effect. Hence, in saying the ideas of cause and effect are distinct, Hume should be taken to mean that the idea of any

Hume on Possible Objects and Impossible Ideas 225 given thing is distinct from the idea of whatever is taken to be its cause or, for that matter, its effect. 3. The then widely-held doctrine of intuitive and demonstrative knowledge is being exploited by Hume. To demonstrate is to make evident a necessary truth which is not self-evident by deriving it from other necessary truths which themselves are either selfevident or have been derived from yet other necessary truths. A truth of intuition is selfevident. Thus demonstrative knowledge presupposes intuitive knowledge and both yield or concern necessary truths only. In denying a proposition's necessity, Hume can thereby deny its demonstrability in the special sense under consideration. 4. A key statement in Hume's theory of knowledge occurs at T Lii.2, p. 29, where he wrote, "Wherever ideas are adequate representations of objects, the relations, contradictions and agreements of the ideas are all applicable to the objects; and this we may in general observe to be the foundation of all human knowledge." 5. Emphasis added. I take "is alone" to mean "is by itself." 6. Hume on p. 32 seems to introduce the additional principle that if one cannot form a clear idea of something, then the thing in question is impossible. By contraposition one can put this positively as if something is possible, then it is possible to have an idea of it. This could be called the Object-to-Perspicuous-Idea Possibility Principle. Some care is needed here, since here is what he wrote: "We can form no idea of a mountain without a valley, and therefore regard it as impossible." I shall not consider this principle further in this paper. 7. Hume regarded his argument as a demonstration, as is clear from his curt dismissal of "obstinate defenders of infinite divisibility" who would prefer to classify it as "a difficulty." (See, T Lii.2, pp. 31-2) Hume's treatment of "difficulties" may have been inspired by Bayle's insinuation that each of the contending parties in the debate over the composition of extension regard the arguments against their opponents' positions as rigorous refutations of them, but the arguments against their own positions as mere difficulties. See Pierre Bayle, Historical and Critical Dictionary----Selections, translated and edited by R. H. Popkin (Indianapolis, Hackett Press, 1991), "Zeno of Elea", Remark G, pp. 59-362. 8. See T Lid, pp. 33-4, for Hume's doctrine of qualitied points, Lii.4, pp. 39-40 for his resulting denial of a vacuum, and Lii.5, p. 54, for the objection. For the connection to Locke, see Essay concerning Human Understanding, Book Two, Chapter Thirteen, Section 23. 9. Hume described his attempt to answer the criticism by saying he would show "we can at once form a notion of all possible degrees of quantity and quality, in such a manner at least, as, however imperfect, may serve all the purposes of reflexion and conversation." (T I.I.7, p. 18) This account was developed at pp. 20-4. 10. George Berkeley's Manuscript Introduction, edited by Bertil Belfrage (Oxford: Doxa, 1987) p. 75. 11. Having argued that even though in some as yet unspecified way it can represent kinds and common qualities, an abstract idea strictly speaking can only be of what is individual and determinate, Hume proceeded to argue that the idea itself must be individual and determinate. He secured this peculiar conclusion by asserting a connection between what an idea is and what it is of. His puzzling premise was "To form an idea of an object, and to form an idea simply is the same thing; the reference of the idea to an object being an extraneous denomination, of which in itself it bears no mark or character." (T I.I.7, p. 20) One of the most difficult tasks in the interpretation of

226

PHILLIP D. CUMMINS

Hume is comprehending his theory of how ideas represent what they represent. This passage only adds to the puzzlement. Compare T I.iv.5, pp. 239-240. He could have saved space, I think, by simply arguing that since ideas exist and everything which exists must be individual and determinate, they, too, must be individual and determinate.

12. At T 1.1.7, p. 19, Hume develops the following puzzling argument: "The confusion, in which impressions are sometimes involv'd, proceeds only from their faintness and unsteadiness, not from any capacity in the mind to receive any impression, which in its real existence has no particular degree nor proportion. That is a contradiction in terms; and even implies the flattest of all contradictions, viz. that 'tis possible for the same thing both to be and not to be. I am not at all clear about how a contradiction follows from the supposition that an indeterminate or general impression exists." 13. Since Hume used "whatever" to formulate his principles, as in "whatever objects are different are distinguishable," representin~ them as biconditionals is slightly misleading, but significantly assists exposition. 14. Hume, T I.ii.2, p. 30. The principle was invoked to secure the reality of compositional simples, indivisible and thus ultimate parts of extended things. 15. See, for example, Hume, T Liv.6, pp. 255-6. On Hume's view of wholes, see Donald L. M. Baxter, "Hume on Virtue, Beauty, Composites, and Secondary Qualities," Pacific Philosophical Quarterly 71 (1990) pp. 103-118, especially 108-110. His "Hume on Infinite Divisibility", History ofPhilosophy Quarterly, 5 (1988) pp. 133-140 should also be consulted. 16. When Hume wrote the line and its determinate length are conjoined in the idea, he could not be implying the very distinctness he has just denied. "Fused" would have better made his point. 17. In my "Hume on Qualities," Hume Studies, 22 (April 1996), pp. 49-83, I explain ontological simplicity in considerable detail and argue that Hunle was not inconsistent in combining (1) his analysis of things as collections of qualities, (2) his doctrine that extended things are ultimately composed of qualitied points, and (3) his doctrine of ontological simples-his denying ontological status to qualities. The present essay reciprocally explores in depth some topics merely touched upon in that essay. 18. Recall that Hume's assessment of the inference from constant conjunction to causal connection was formulated as the claim that imagination not reason detennines the mind to have the idea of an effect upon experience of its cause. See Hume, T I.iii.6, pp. 88-9. See, too, I.iv.2, where Hume presents his account of the evidence for belief in anextemal world as an answer to the question whether such belief is due to sense, reason, or imagination. Note, especially, pp. 187-8. 19. Thomas Reid, Essays on the Intellectual Powers of Man (1785), (Cambridge, Mass., MIT Press, 1969): Essay One, Chapter Two, pp. 35-7; Essay Two, Chapter Nineteen, pp. 276-8; Essay Six, Chapter Six, pp. 650-2. 20. Both Reid and Kant can profitably be considered as responding, though quite differently, to this pattern of reasoning in Hume. Reid rejected the principle that whatever is conceivable is possible in his essay on conception in Essays and proceeded to reaffirm the causal maxim and the substance principle as first principles of necessary truth. Kant's response was more complex. One could say he accepted Hume's position as it applies to the noumenal, i. e., with reference to absolute possibility and impos-

Hume on Possible Objects and Impossible Ideas

227

sibility, but introduced a further concept of possibility, possible as realizible in a possible experience, and further conditions for such possibility in addition to mere conceivability through a perspicuous concept. The causal maxim and the substance principle were then held to be valid as principles governing all, but only possible experiences. 21. Besides the methodological issues being raised, there are further complications about possible objects and impossible ideas. First, there is the issue of whether Hume argues frorn the impossibility of having a perspicuous idea to the impossibility of what it represents. We know more about his denials of allegedly necessary truths than we do about his positive theory of how intuitive and demonstrative knowledge is secured. See T Lii.2, p. 29, and Liii.l, pp. 69-70. Compare David Hume, Enquiry concerning Human Understanding, third revised edition, edited by L. A. Selby-Bigge and P. H. Nidditch, (Oxford, Clarendon Press, 1975), Section Four, Part One, pp. 25-6. Second, Hume ties his account of belief in cases of intuitive and demonstrative knowledge to the inlpossibility of conceiving the contrary of what is being so known. On the latter, see Burne, T Liii.7, p. 95. 22. I here assert, but shall not argue, that in his treatment of causation Hume did not allow any intelligible sense of objective necessary connection between any cause and its effect and did not posit any such connections. He did, as in discussing liberty and necessity, sometimes use "necessity" in connection with causation, but in a suitably cleansed sense of necessity. See Hume, T II.iii.1, p. 400, for example.

The Wolffian Background of Kant's Transcendental Deduction MANFRED KUEHN

Christian Wolff (1679-1754) was the most influential philosopher in Germany before Kant. Although his influence began to wane after the middle of the century under the influence of especially British philosophers, his works still relnained important. In particular, his works on logic continued to be relevant. Indeed, one might argue that the philosophical developments leading to Kant, and then again away from Kant, cannot be properly understood without a knowledge of Wolff's Logic and Ontology. Since Wolff is usually regarded simply as a follower of Leibniz, it is also thought that Leibniz's theories provide us with all that is needed to understand Wolff. However, this is a mistake. Though the phrase "Leibniz-Wolffian philosophy," which gained currency during the eighteenth century, refers to what was understood as a more or less unified movement, this is fair neither to Leibniz nor to Wolff. Leibniz should not be held accountable for the confusions of Wolff and his followers, and the Wolffian theories should not be measured by whether they develop adequately Leibniz's many "tantalizing hints and suggestions."l Wolff himself insisted rather strenuously that his position was very different fronl that of Leibniz, and I believe he was right. He was much more of an empiricist than Leibniz ever wished to be. In characterizing his approach, he pointed out: When I base cognition on experience...then I am most careful that I do not surreptitiously introduce [erschleiche] anything...However, I have also found that this carefulness is very difficult and that it is almost easier to acquire a skill in demonstration than this carefulness.. .! make inferences from reality to possibility.. .in this way I keep my concepts pure so that nothing can sneak in whose possibility has not been cognised...and in this way I provide the foundation of absolutely reliable inferences in the sciences. 2

Philosophy, investigating why things are the way they are and in this way going far beyond experience, must be careful never to lose itself in mere possibilities. His philosophy was meant to be a marriage of reason and experience (connubium rationis et experientiae), and even if this was not necessarily a marriage of equal partners, reason and experience were still partners for Wolff. 3 Nor is it always clear which of these partners was more dominant in his vievi. This is important to remember. For, when Wolff defines philosophy as a "science of all possible objects, how and why they are possible," or as "the science of the possibles insofar as they can be," and when he thus claims that existence is nothing but the "complement of possibility," he does not mean that we can dispense with experience. 4 Experience, or historical knowledge, as he

MANFRED KUEHN also calls it, remains the foundation of all philosophizing. "Experience establishes those things from which the reason can be given for other things that are and occur, or can occur.,,5 Though he defines a thing as anything that exists or might exist and identifies "reality," "possibility," and "what does not involve contradiction," he does not believe that we can start our inquiry from just anything that does not involve contradiction. His "ontology ... is an analysis of the logical possibilities for the existence of real entities." It may well be true that his philosophy is not about "existing reality," as it had been for almost any philosopher before him," but it is nevertheless about possibles in so far as they can be or actually are. 6 This empirical bent is also predominant in Wolff's logical theory. His German Logic, entitled Rational Thoughts on the Faculties of the Human Understanding and their Proper Use in the Cognition of Truth, appeared in 1713. It begins with sensation, and it ends with a reminder that we should make compatible the "use of the senses and the understanding."? Although in the Preface to the first edition he calls attention to the fact that Leibniz's "Meditations on Knowledge, Truth, and Ideas" of 1684 gave him "unexpectedly a great light" about the subjects discussed in the Logic, he clearly is also indebted to Locke's On the Conduct of the Human Understanding of 1706. Logic is for Wolff a science of the human mind, and in particular of the faculty of understanding. However, it is not just a theoretical investigation of this faculty, it is also a discipline that has a practical application, and one of its main goals is to show us the proper use of its principles in the obtaining the truth. It is meant to investigate what are the conditions of knowledge, and we best acquire it. Though Wolff's Logic is organized in accordance with the "three operation of the mind," or the "tres operationes mentis," namely concept, judgment and inference, it is perhaps less mentalistic than the Logic textbooks of many of his predecessors. Indeed, the words "mind," "subject," "self," and even "apperception" do not playa large role in it. It is not so much a logic of the workings of the mind as it is a logic of the preconditions of knowledge and truth. As such a logic, it cannot do without referring to mental operations and the faculties, but it is not primarily about them. Wolff's philosophy has been called a philosophy without a subject. And its strenuous insistence on being per se and "objectivity" separates it as much from Leibniz as anything else. It was this kind of logical theory that forms the background of Kant's own Transcendental Logic. He was steeped in Wolffian logic. During his high school days he was already instructed in it, and at the University of Konigsberg it formed an important part of the curriculum. 8 And when he taught his own logic courses, he did so on the basis of a Wolffian textbook, Georg Friedrich Meier's Auszug aus der Vernunftlehre of 1752. It should therefore not come as a surprise that his Critique of Pure Reason still exhibits some traces of this influence. Kant's terminology is almost entirely Wolffian. Though he himself 230

The Wolfjian Background ofKant's Transcendental Deduction 231 invented a number of his most important technical terms for his own purposes, most of his philosophical dictionary comes from Wolffian philosophy. Athough the theories offered by Kant in these sections are fundamentally different from Wolff's and represent as radical a departure from Wolffian logic and ontology as one might wish, they remain indebted to them in surprisingly many of the details. As the devil is admittedly in the details, one may well wonder what consequences this has for Kantian philosophizing, and, to change the metaphor, whether Kant's Critique does not prove once again that it is impossible to fill new wine into old wine skins. On"the other hand, it might well be the case that the Wolffian background can help us in understanding what Kant was up to in certain parts of the Transcendental Logic. And that is what I wish to argue here with regard to one of the most central concerns of this part of the Critique, namely the issue concerning "objective validity" and "objective reality" as it gets raised in the Transcendental Deduction. I shall suggest that the Deduction itself can profitably be seen as being modeled after a standard kind of argument found in Wolffian textbooks. It is an argument that provides us with a proof of the possibility of a concept or an object, or with a principium probandi possibilitatem. By examining the connection of this aspect of Kant's theory of concepts with that of the Wolffians, I hope to be able to make clearer what Kant took himself to prove in the Transcendental Deduction. I am not sure whether this will make the Transcendental Deduction more or less interesting. Nor do I know whether it will make a significant difference to the present understanding of the function that transcendental arguments can or cannot fulfill, but I do hope that it will at least clarify the history of Kantian philosophy. To do so, I shall first briefly characterize the notion of a proof of the possibility of a concept as an important aspect of Wolff's theory of concepts. Secondly, I shall show how the Wolffian approach was put to new use by Johann Heinrich Larrlbert and Johann Nicolaus Tetens in their discussions of basic concepts, a priori cognition, and fundamental principles. Especially Lambert's view is highly significant because during the sixties Kant was essentially a follower of Larrlbert. Thirdly, I shall try to show that Kant's Transcendental Deduction must be understood an attempt at providing the principium probandi possibilitatem of the categories in the tradition of Wolff, Lambert, and Tetens. And finally I shall make some brief suggestions as to what difference my observations might make to Kant scholarship, suggesting, among other things, that this Wolffian aspect of the Deduction dovetails nicely with Kant's more Humean concerns.

II Wolff's Logic begins with a chapter entitled "Of the Concepts of Objects." In it, he defines "concept" as "any representation of a thing in our thoughts,"

232 MANFRED KUEHN differentiating it thus from "perception" which for him is a mental states that is characterized by consciousness of something as present to us and "thought" which he defines as mental states by means of which we are conscious. A concept is a particular kind of thought, i. e. a thought that has an object. As a thought, it is characterized by consciousness. It is different from a perception because the object of the concept need not be present to us. Concepts are the most fundamental of the mental operations. Following Leibniz, he then differentiates between clear and obscure, distinct and indistinct, adequate and inadequate, complete and incomplete concepts, and he discusses in great detail how we obtain concepts, how we should analyze them, and how we might make them clear and distinct. For Wolff, there are basically three ways to obtain concepts. Some are derived directly from sensation, others we derive from the concepts of particular things by abstraction, and still others we derive from the other two kinds of concepts by means of arbitrary considerations that we might engage in, or that we might find useful in certain contexts. The latter are complex concepts. They are conlposed of two or more concepts that are themselves more simple. They are our own creations, as it were, and they include what Locke called complex ideas. However, since Wolff is not as confident as Locke that the ideas that "enter by the senses" are "simple and unmixed," not all complex ideas are by that very fact also our own creations. He gives the example of the concept of "gratitude," which he views as a combination of the concepts of "pleasure" and "debt to a benefactor." After having discussed these distinctions, he raises a point that is important for our purposes, namely he asks "How do we know whether a concept is possible?" This is a general question that is meant to apply to all concepts. Wolff seems to think that there might be representations that seem to be concepts, but really are not. We nlight perhaps say that there are words to which no concept corresponds. Thus the concept of a "square circle" nlight appear to be just as legitimate as the concepts of "square" and "circle," yet it clearly is not because its two components contradict each other and thus fail to express a coherent thought. Wolffs answer to the question "How do we know whether a concept is possible?" understandably goes in a somewhat different direction, however. Rather than asking whether a concept expresses a coherent thought, Wolff asks whether a concept has a possible object. And his answer is: if it has a possible object, then it is a possible concept, if it does not have a possible object, then it is an impossible concept. Thus for Wolff conceptual questions cannot be isolated from ontological questions. Questions about concepts necessarily lead to questions about objects; and it appears to be at the level of objects that conceptual questions are decided. Leaving aside the question what precisely is the difference between "having a possible object" and "expressing a coherent thought," we might ask "How do we know whether a concept has a possible object?" Wolffs answer is that this

The WolfJian Background ofKant's Transcendental Deduction 233 depends on which of the three kinds of concepts is in question. We can be sure that a concept is possible, if it is directly derived from sensation. As he puts it: "When we have been led by the senses to a concept, we cannot doubt that it is possible. For who would doubt that something is possible that we have really encountered? Therefore concepts of this sort lay a secure foundation for correct cognition.,,9 If a concept has an actual object, then it must also have a possible concept. We can also be sure in the case of general concepts: "Since the general concepts do not contain anything that cannot be found in the particular ones, they must also be possible, if they have been abstracted from possible concepts."lO So one way to determine whether concepts are possible is to compare them with the experiences which gave rise to them in the first place. If the concepts are directly or indirectly based on sensation, this process is relatively unproblematic for Wolff. Concepts based on experienced objects need no further justification than these experienced objects. However, if we have formed a complex concept by more or less arbitrary considerations, we cannot always be so sure whether it has a possible object or whether it has no corresponding object. "For this reason we must prove in such cases that the obtained concepts comprise something that is possible. Nor is it sufficient that the determinations are by themselves possible. It is also required that they can co-exist with the others."ll To show that a concept is possible is also to show that it is real because what is possible is also real for Wolff. There are two ways in which we can show that a complex concept comprises something that is possible. Either we show it by experience or by proof. Thus he claims that we can know that "gratitude" is possible by simple introspection. If we do find this emotion actually within us, then its concept is also possible. Indeed, Wolff seems to think that direct consultation of experience allows us to know the large majority of concepts as possible concepts. But, however that may be, the first way of showing that a concept is possible is no different from showing how experiential concepts are possible. Doubts about such concepts can be answered by referring to experienced objects. However, in some cases experience is insufficient, and we must prove that the concept is possible. Again there are two ways: We can prove whether a concept is possible either by showing how such a thing can come into existence, or even by showing whether something follows from it of which we already know that it is possible. For if we know how something can originate, then we may no longer doubt whether it can exist. If only impossible things flow from a concept, then the concept itself cannot be possible either. However, if only possible things flow from it, then it must also be possible. For if something flows from another thing, then it can only be because the other thing exists. 12

Wolff refers to several examples: Euclid proves that an equilateral triangle is possible by showing that it is possible to construct one on any given straight line; a machine can be shown possible, by showing how it is made; virtue can

234

MANFRED KUEHN

be shown to be possible, if I can show how it originates in the soul; I show that a figure consisting of two angles and two straight lines is impossible because it implies that two straight lines can intersect at two points, etc., etc. He tells us that he has proven "the possibility of most moral rules in accordance with these principles. ,,13 Wolff's notion of the proof of the possibility of a concept thus leads to the notion of another kind of proof, namely a proof of the possibility of an object or thing. The conception of this latter proof is explicated in his Ontology. There he differentiates between six different "principia probandi possibilitatem." One ofthese is again an a posteriori principle of proving possibility, while the other five are for him a priori principia probandi possibilitatem. The a posteriori principle allows him to prove that an object is possible on the basis of its existence or on the basis of the existence of something on which it depends. This a posteriori principle of proving possibility can only prove that something is possible not why it is possible. For this reason it is for Wolff the least interesting. It is not really a philosophical principle, but merely an empirical one. The other five principles of proving possibility are a priori. In accordance with the first, we must try to prove that some characteristic of an object depends on another one of its characteristics. Put differently, we must prove that when we assert one characteristic of a thing we also assert the other. Thus when we say that a triangle has angles of equal size we also assert that all its sides are equally long. If we succeed in proving that there exists such an dependency, then we have shown that an object having these two properties is possible. The second a priori principle involves showing that two properties of a thing, even though not standing in a relation of dependence can be combined. If two properties of a thing are non-contradictory, they are compatible and an object having both is possible. The third one simply states that if something has been proved then it is possible. The fourth one maintains that it if we show how something is produced then we have shown that it is possible. Thus if we have constructed a geometrical figure, then we have proved it to be possible. The fifth principle amounts to the claim that a thing is possible if its attributes are such that they allow claims whose truths have already been shown either by proof or by experience. 14 However, within the context of logic, it is not really necessary to go into a detailed proof of the possibility objects themselves. What are possible objects is proved in other disciplines, most importantly in the ontology. Logic is concerned more with the possibility of concepts than with the possibility of objects, and while Wolff has a tendency to confuse the two at times, he also insists that it is important to keep them separate. Wolff's demand for arguments proving the possibility of complex and arbitrary concepts is of course closely connected with the idea that there are simple concepts into which all complex concepts can be analyzed. And one might expect that, in what some would consider good Leibnizian fashion, he goes on to reduce all concepts to these simple ones. However, Wolff prefaces

The WoljJian Background ofKant's Transcendental Deduction 235 the entire discussion with a very skeptical note about our ability to get to these simple concepts. He believes that it is neither necessary nor often possible that: ...we can bring the analysis to an end, that is, that we can reach those concepts that are in and for themselves no longer analyzable because they no longer contain many disparate characteristics. We may be content, when we have brought the analysis so far that we can satisfy our aim, i.e. when we have a concept that allows us to make ourselves understood to others or one that allows us to prove something. 15

Mentioning Euclid who simply assumed that the concepts of "identity," "larger," and "smaller" needed no further analysis, he suggests that this lack of analysis in Euclid is not a shortcoming, even though he also finds it necessary to point out that he himself has analyzed them further in his own Principles of Arithmeti c. Wolff s theory of concepts, and especially his doctrine of proving the possibility of concepts, appears to have survived more or less intact until the early sixties of the eighteenth century. It remained an essential part, and as far as I can tell, most logic textbooks follow him more or less closely. Thus Johann Christoph Gottsched agreed that notiones arbitrariae must be shown to be possible by argument or experience, and to prove that they are possible meant for him showing either that they do not contain a contradiction or that they really exist. 16 Hermann Samuel Reimarus agrees with Wolff that with regard to complex concepts everything depends on whether the "arbitrary conjunction is free from internal and external contradiction." 17 He also agrees that the possibility of concepts must be proved by recourse to their objects, offering an account of the one a posteriori and the five a priori principia probandi possibilitatem that is very similar to that of Wolff, even though it is offered in the context of his logic rather than in an ontology. Even the textbook by Meier . on which Kant based his Logic lectures, takes over this doctrine of Wolff, nlore or less without change. Thus we find: A learned concept which has been created by arbitrary conjunction must be proved or disproved. We can achieve this either by (i) experience, if we show that their concepts are real (wuerklich) or not real, or (ii) by reason, either directly or indirectly by showing that and how their 0 bj ects can become real or that they cannot become real. 18

Kant's reflections on this paragraph are highly interesting, showing that this part of the Wolffian theory of concepts must have been significant in his own development of the doctrine of the distinction between "synthetic" and "analytic." Thus we find him observing that: "...all arbitrary concepts are synthetic and all synthetic concepts are arbitrary" in a relatively early reflection. And in a somewhat later reflection Kant observes that "empirical concepts have also been made because they determine and constitute the object

236 MANFRED KUEHN by perceptions that I arbitrarily collect, but because the thing itself is given we can analyze it.,,19 III However, before we discuss Kant, it is best to take a brief look at two of his most important contemporaries who significantly contributed to the discussion of the problem of the possibility of basic concepts and principles before he did, namely Larrlbert and Tetens. Lanlbert is usually characterized as 11lore or less simply following Locke. This is clearly historical nonsense. He himself sees his relation to Locke more clearly, saying: Locke made in his work on the human understanding the anatomy of our concepts his main business, and he tried in to make distinct both the simple concepts and their modifications and connection as far as possible. We would therefore here have largely to repeat his work, if our goal were the same. Yet Locke is content to build his entire work on experience, and he therefore proceeds entirely a posteriori, taking things as they are. We therefore call his system an anatomy of concepts and cognition.. .It is not enough to collect simple concepts, we must also see how we can achieve universal possibility in regard to their composition. 20

Lambert followed Wolff in holding that complex concepts need a proof of their possibility. He also took over most of the details of Wolff s method of proving it. However, he is most interesting because he raises the problem of the possibility of concepts in the very context of the problem of the possibility of knowledge a priori. 21 Distinguishing very much like Kant between concepts and judgments a priori and a posteriori, he defines an a priori concept as one in which we do not have to consult experience "immediately." The same holds for judgments. If we need "immediate experience" for a concept or a judgnlent then it is only a posteriori. Thus for Lambert, just as for Wolff, "a priori" and "a posteriori" are merely relative concepts. They indicate "a certain order." Indeed, Lambert does not seem to think that there are concepts or judgments that are absolutely a priori, saying, among other things, that he will admit that we can call in the strictest sense "a priori" only concepts which owe nothing at all to experience, but that it is an entirely different and superfluous question to ask whether we find such concepts in our experience. However, he insists that we can call "a priori" anything we can know prior to a particular experience. This notion of the a priori is very important for Lambert because he believes that knowledge a priori, even if only a priori in a relative sense, is superior to knowledge a posteriori. It allows us to make more general claims about things. Furthermore, if we can say something about a certain subject without having to nlake experiments and observations, we are not as likely to commit errors. For this reason he thinks that it is very important

The Wolfjian Background ofKant's Transcendental Deduction

237

...to investigate...whether and in how far we can make any cognition scientific merely on the basis of its concept and therefore a priori. If we find that this is possible, then \ve can extend the concept of scientific knowledge that we have so far only assumed in so far as it is concerned with bringing experiences into connection with one another and to derive one from the other. 22

Now, since Lambert holds that our concepts are the more a priori the further removed from experience, the question concerning the difference between a priori concepts and a posteriori soon turns into a question about how we obtain the concepts that are the farthest removed from experience, and, in effect, he asks: How are a priori concepts possible? Lambert raises in this way the question of the principia probandi possibilitatem in a new context, namely in the context of the possibility of a priori knowledge. His answer is also quite interesting. Starting in the familiar way, he argues that if concepts derive from experience (or even if they could derive from it) then they are possible. 23 Secondly, he agrees with Wolff that universal concepts, in so far as they are based on abstractions, are also unproblematic. They also can be shown to be possible by means of experience. Only the complex concepts, which' Larrtbert now calls hypothetical concepts, are problematic. Their possibility needs to be proved. But, and here Lambert seems to differ fronl Wolff, there cannot be any direct proof of the possibility of these concepts that is a priori in any strong sense. Any direct proof of the possibility of complex concepts will, he claims, quickly reduce to an appeal to experience, and thus not allow us to go very far in making a priori claims. Accordingly, Lambert chooses the indirect route. He concentrates on simple or basic concepts. As he puts it: "Since complex concepts can be analyzed into more simple ones which constitute their characteristics, we can think of completely simple concepts which cannot be analyzed any further, but which can be determined and exhibited by means of their relations to other concepts. Such concepts form the basis of all our knowledge, and they may therefore be called basic concepts in the truest sense." Thus, whereas Wolff was very skeptical about our ability to obtain simple concepts, Lambert is confident that the answer to his question will come from simple concepts. One of the reasons for his confidence in this regard is clearly his greater reliance on Locke. Indeed, Lambert thinks that sensation provides to us a number of basic concepts that we only have to collect. His (rather open-ended) list basic concepts consists of 1. extension, 2. solidity, 3. motion, 4. existence, 5. duration and succession, 6. unity, 7. consciousness, 8. the power to move, 9. willing?4 Now, these basic concepts have for Lambert two important characteristics: (i) "the mere representation of a simple concept shows at the same time its possibility," and (ii) the simple concepts are entirely a priori and independent of experience. His reasons are as follows: First, to show that a concept is possible we must show that it is not self-contradictory. Well, simple concepts are just that: simple. Therefore there is nothing in them that could contradict

238 MANFRED KUEHN anything else. A simple concept can therefore not fail to be possible, and an impossible concept cannot be simple. As soon as we understand a simple and truly basic concept we have also proved that it is possible. Second, since the possibility of a basic concept is proved by its very representation, it becomes thereby entirely independent of experience. Although we have it from experience, experience is merely the occasion of our consciousness of it, as it were. Once we are conscious of it we do not need to get the ground of its possibility from experience because the possibility exists with its mere representation. Therefore it becomes independent of experience. And this is a requisite to cognition a priori in the strictest sense...Thus our cognition becomes in the strictest sense a priori. 25

As soon as we "can be assured of its possibility without experience, we can view a concept as subsisting by itself and as a priori.,,26 And such concepts allow us to develop an a priori science. There appears to be nothing in this that would have been fundamentally troubling to a Wolffian. Yet, because Lambert's simple or basic concepts are identical with Locke's simple ideas of sensation, which are simply given, and they are not themselves the result of an intricate analysis of the confused representations of the senses, he has a much easier task in establishing a completely a priori science, independent of experience. In a way, Lambert actually undernlines Wolff's connubium of reason and the senses. Reason by means of certain a priori concepts can be divorced from sensation and become independent. However, not everyone was satisfied that Lambert had actually accomplished his goal, and that he had put enough distance between himself and Locke. Especially Tetens criticized him in this way. Although he admitted that Lambert had "collected" and "exactly determined" many of the basic concepts, he found that his predecessor had left unsolved some of the most important problems connected with them. In particular, he found lacking the very thing that would allow us to decide between Leibniz and Clarke and which would make: ...evident what and how much the understanding really possesses in the aforementioned concepts. Mr. Lambert assumes the very same notions as simple, which Locke also assumed as simple. Are they the ones? So others also accept them as the simple ones? Has it already been decided that they are more than confused appearances of the understanding ... ?-that they are real ideas, which correspond to objects? Has it already been made evident from the nature of the understanding in how far they are real? And has it been made evident in how far the axioms and postulates that are based on them, and whether the theory that is built on them is transcendent and applicable even where we think about beings that are entirely beyond the circle of the sensation from which these concepts derive?27

The WoljJian Background ofKant's Transcendental Deduction 239 Tetens asks for the realization of the basic concepts, i.e., he wants to know how far the basic concepts correspond to real objects. His own answer, which amounted to the claim that they are basic concepts of any possible understanding and not just our own, and that they are therefore real concepts took another step away from Wolff. Tetens was no longer satisfied with merely logical possibility. "Reality" for him was more than "possibility." He sought to give a proof that would prove that our basic concepts correspond to objects that are more "weighty" than objects of sensation or mere appearances of objects.28 IV This is the status questionem concerning the possibility of a priori basic concepts inherited by Kant. Indeed, the early Kant, that is, the Kant during the period from about 1764 to 1769 was essentially a follower of Lambert. His claims about the precedence of analysis over synthesis, his rather obscure view of how analysis leads to simple a priori concepts, as well as many of his other theories can be shown to be dependent on Lambert, and perhaps even to amount to ultimately rather trivial modifications of the latter's view. Even in the Inaugural Dissertation he has no great difficulties with the possibility of his basic concepts and their real use. However, all this has changed in the Critique of Pure Reason. In this work Kant no longer speaks of simple or ·basic concepts, but of categories or "pure concepts of the understanding," and in it he is clearly convinced that he has an obligation to provide a proof of the possibility of these concepts because they are neither directly experiential nor derived from experience by abstraction. Furthermore, his reflections from the period during which he wrote the Critique of Pure Reason show not only that he was well aware of the piece of Wolffian logic just summarized, but also that he understood that his concepts of the understanding needed a justification in a similar way as arbitrary concepts. It is therefore more than reasonable to assume that the Transcendental Deduction, which clearly represents at least in some sense a justification of the categories, can be understood as Kant's attempt at formulating a "principium probandi possibilitatem" for the categories. If this is true, then we should expect that what gets "deduced" or justified in the Transcendental Deduction are these categories, and that Kant answers one .of the demands of Wolffian logic in this passage. This would mean that the conclusion of the Transcendental Deduction should be something like: "the categories are possible concepts," or "the pure a priori enlployment of the categories is possible." And this is precisely what we actually find. At the very end of the A Deduction Kant says: "Pure concepts of the understanding are thus a priori possible, and, in relation to experience, are indeed necessary; and this only because our knowledge has to deal solely with appearances ..." (A130).29 And at the end of the B-Deduction, in a section entitled "Brief Outline of this Deduction" Kant

240 MANFRED KUEHN tells us that "this deduction is the exposition of the pure concepts of the understanding, and therewith of all a priori knowledge, as principles of the possibility of experience" (B 168). If this were not enough to show that Kant's Transcendental Deduction is essentially a proof of the possibility of concepts, we might take a closer look at A86/B118 where Kant actually tells us with regard to the categories that what he would like to find is "the principle of their possibility," and, that this principle cannot be found by an empirical derivation in the Lockean fashion, but only by a transcendental deduction. Accordingly, I would hold that the conclusion of the Transcendental Deduction can only be that the categories are possible concepts. If successful, Kant's Transcendental Deduction has fulfilled the demand of the justification of the categories that a Wolffian might formulate. However, matters might appear to be more complicated, and they have appeared more complicated to commentators. Kant also describes the outcome of the Deduction in another, and apparently different, way. Thus he says that he in the Deduction has made the "objective validity (objektive Gultigkeit) of the pure a priori concepts ... intelligible," and that he has determined "their origin and truth" (A 128). Now, as if this were not enough, Kant also uses the term "objective reality" or "objective RealiUit" to describe what he is after in a Transcendental Deduction. Indeed, the phrases "objective validity" and "objective reality" are more frequent than "possible" or "possibility." Thus "objective validity" appears 30 times and "objective reality" appears 38 times in the second edition, while "possible" or "possibility" does not appear as frequently in connection with the categories themselves. Yet this does not mean much. First, "objective validity" means for Kant nothing but "(empirical) reality," or better "real possibility." "Objective validity" means the same thing as "real possibility. ,,30 Both are opposed to "merely logical possibility." Kant, unlike Wolff, differentiates sharply between merely logical possibility and real possibility. Merely logical possibility has to do only with thinking, and its only criterion is the principle of contradiction. Real possibility, or objective validity, requires, Kant says "something more." I must be able to prove the possibility of its object either by experience or a priori. And only when I have proved its possibility can I be sure that I have cognised something. 31 In other places Kant claims that the only concepts that have "objective reality" are those which apply to "possible things" (A2201B268), thus making clear however much he differs from Wolff in other respects, he is close to him in accepting the necessity of proving any concept he uses as well as in thinking that the possibility of a concept is proved by proving that its object is possible. Kant's very distinction between real possibility and nlerely logical possibility reaffirms his indebtedness to Wolffian philosophy. But however that may be, when Kant says at A 130 that the categories are a priori possible, he does not mean to say, of course, that he has proved that they are logically possible, but rather that they are really possible. For the former he would not have needed as elaborate a proof as he

The Wolffian Background ofKant's Transcendental Deduction

241

actually puts fonvard. He is using "possible" as a short expression for "really possible." Secondly, "objective reality" is for Kant just another way of talking about what is really possible, i. e., "what has application to objects which can be given us in intuition" or what is necessary for the moral law. 32 Whenever Kant uses "possible" with regard to the categories, he means "really possible" or "objectively real" and "objectively valid," but there really is no fundamental difference between the three ways of talking about the result of the Transcendental Deduction-at least not in Kant. 33 Since there are many inlportant philosophical scholars of Kant \vho believe that there is a significant difference between "objective reality" and "objective validity," and that he has proved the one, but not the other, this is perhaps not a trivial point. 34 There are many other passages in the Deduction itself and in other parts of the work that make it abundantly clear that this is what Kant was up to. I shall not bore you with listing all or even most of them. Rather, I would like to say just a little-far too little~about what consequence an interpretation of the Deduction as an a priori principium probandi possibilitatem has for our understanding of its "proof structure." I believe it should be immediately clear that the possibility of experience is not a conclusion of the conclusion of the Deduction. Just to remind ourselves, Wolff argued that to show that a concept is possible we must show that it comprises or refers to a possible object, and to do so we must show either (i) how the object of a concept can come into existence, or (ii) whether something follows from the object of which we already know that it is possible. I submit that Kant is trying in the Deduction to fulfill both these tasks. he gives both a genetic account of how knowledge can originate from the basic faculties he has identified, and he tries to show how something follows from the object of the categories of which we already know that it is possible. The former is usually called the subjective deduction, the latter is called the objective deduction. I shall concentrate on the objective deduction. It appears to me that Kant intends to show in the Deduction that these categories or pure concepts are possible because their object makes possible something that we already know to be possible. He first proves that the categories have an object. In one sense, of course, this is a trivial task. Categories, for Kant, are concepts of "an object in general" (BI27). We need them, at least according to arguments presented earlier in the Critique, to think any object whatsoever, and without them we cannot think at all. As such concepts, they are not abstracted from experience, but are concepts we have a priori.35 So it is obvious that we must use them to think possible objects. Yet, this conclusion is far from sufficient for Kant because this proves only the logical possibility of the categories. In particular, it does not address "the very thing that would allow us to decide between Leibniz and Clarke and which would make evident what and how much the understanding really possesses in the aforementioned concepts" that Tetens asked for. Nor does it accomplish the second task he formulated, namely to show whether these categories are

242 MANFRED KUEHN "transcendent and applicable even where we think about beings that are entirely beyond the circle of the sensation." Kant must engage in a more difficult task, namely, he must show that the categories have objects that are really possible, or objects that are empirically real, or, as Kant himself puts it, "if we seek to discover how pure concepts of the understanding are possible, we must inquire what are the a priori conditions upon which the possibility of experience rests" (A95f.). If we find that the categories are necessary for the possibility of objects of experience and therefore also for the possibility of experience, we have justified the categories as objectively valid. His argument proceeds roughly as follows: (1) The categories are necessary for thinking any kind of object. (2) Possible experience must be an experience of objects. (3) The objects of possible experience are necessarily also objects of thought (they cannot simply be sensed).

Therefore (4) The objects of possible experience presuppose the categories.

Therefore, (5) The categories have possible objects.

However, (6) We can only think objects by means of the categories in so far as something (material) has been given to us in intuition.

Therefore, (7) We calUlot think by means of the categories any other possible obj ects than those of possible experience.

and, therefore (8) The categories are objectively valid (or really possible) only with regard to possible experience.

(1) is proved for Kant by the Metaphysical Deduction, while (2), (3), and (6) are, of course, not simply asserted, but argued for. The details of Kant's arguments are admittedly messy. It is not always clear what Kant means and how he takes himself to have proved particular points. However, the general strategy should now be clearer. Kant shows that the categories are possible concepts because they make experience possible. They are not merely empty words or functions of the understanding which That Kant is consciously following this approach should be clear from such passages as the following: The Transcendental Deduction of a priori concepts has thus a principle according to which the whole inquiry must be directed, namely, that they must be recognized as a priori conditions of the possibility of experience, whether of the intuitions which are to be nlet with in it or of the thought. Concepts which yield the objective ground of the possibility of experience are for this reason necessary (A94=B126).42

The proof does not move fronl asserting the existence of the categories, or something still more basic than the categories, to prov~~~!~~_~~s~!l~i!i~ _~f_

The Wolffian Background ofKant's Transcendental Deduction 243 experience. The Deduction shows that the categories, which, he claims, may be assumed as given because of the Metaphysical Deduction, are necessarily presupposed in any possible experience (which is also presupposed as given). And the use of the categories is shown to be justified in so far as it is related to this experience. That this is a fair sketch of Kant's argument can be seen from the way in which he contrasts ideas and concepts of the understanding later, saying that ideas cannot form "the basis of any objectively valid synthetic judgment," while "through concepts of understanding reason does, indeed, establish secure principles, not however directly from concepts alone, but always only indirectly through relation of these concepts to something altogether contingent, namely possible experience" (A737=B705). Thus he can say that any principle of the understanding has "the peculiar character that it makes possible the very experience which is its own ground of proof' (A737=B705). And this implies not only that experience would not be possible without these categories, but also that we cannot establish the objective validity of the categories without relating them to possible experience, which, per se, is "something entirely contingent." Someone might argue this account of the role of possible experience in the Transcendental Deduction is misleading, since it suggests that the possibility of the categories is based on an a posteriori or experiential principium probandi possibilitatis. This would mean that the Transcendental Deduction is an empirical argument, something that Kant insists it is not. He explicitly argues that the justification of the categories cannot be accomplished by a merely empirical or a posteriori argument, but it must be done by an a priori argument. Yet, the obvious response is that an argument about possible experience does not make use of premises that thenlselves are empirical. As soon as we talk about possible things qua possible things, we no longer are engaged in an empirical inquiry, or an inquiry that relies on empirical premises in a way detrimental to the purpose of a Transcendental Argument (at least not necessarily so). This is clearly what Lambert thought.36 This is also seems to be implied by some of Wolffs principia probandi possibilitatem a priori, especially the fifth, I think. And there are passages that suggest that this is what Kant had in mind. Indeed, Kant's Deduction clearly is not meant to establish simply that our human cognition is such that it needs categories, but that any sensitive cognition must have such categories. He is proving not something that applies especially to our experience, but to any other experience that is not based on spontaneous thinking. Kant also tries to show how knowledge is possible, or how our faculties (and in particular the understanding and its categories) actually make knowledge possible. We may see this as another way of trying to prove the categories to be really possible by showing how something that we already know to be possible becomes possible. This aspect of Kant's enterprise is usually referred to as his Subjective Deduction, and it is what gets pursued in

I

244 MANFRED KUEHN greater detail in the chapters of the Transcendental Analytic that follow the Deduction. While Kant clearly thought that these parts of less importance, it has been argued recently that they are actually much more significant than Kant (and many commentators) have been willing to admit. Perhaps this is so, but I doubt it. Kant is primarily concerned not with psychology, but with logic. I will not specifically discuss his faculty psychology and its consequences for his transcendental logic here. Nevertheless, it should be clear that much of what I have discussed here has already been "psychological" in a broad sense. For Kant, just as it was for Wolff, logic is concerned with specific operations of the mind, nanlely with those of thinking. For Kant, concepts, judgments, and 'inferences, whether a priori or empirical, are mental operations, and in the discipline of logic "reason has to deal with itself alone" (B x). This is not new in Kant. What does appear to be new in Kant is that logic in his philosophy becomes not just the science of the workings of the nlind, but also the science of the workings of the universe. It is because we must perceive and conceive the world in accordance with the laws that constitute our mind that we can know the world by knowing ourselves. When we talk of "object," "objective" and "objectivity" we 'talk primarily of achievements of the knowing subject, not of things that exist by themselves. This conclusion would have been even more abhorrent to Wolff than it was to most of his contemporaries. For Wolff, the knowing subject qua subject appeared to be irrelevant. In fact, the term "subject" appeared hardly at all in his work. What is important is being qua being possible and capable of existence. Kant is really nluch closer to Tetens in this regard. However, as far as his method of proving the real possibility of the basic concepts of the hunlan mind is concerned he is close to Wolff. And some of the same ambiguities that haunt the Wolffian account also haunt the Kantian account. Though he is nlainly concerned with the possibility of concepts, he is almost inevitably led to a discussion of the possibility of objects. He also cannot isolate conceptual from ontological questions. Just as in Wolff, questions about concepts necessarily lead to questions about objects; and it appears to be at the level of objects that the conceptual questions really get decided. Yet, the objects Kant is concerned with are much less "weighty" than those Wolff had in mind. The possibility "transcendent use" that Tetens wanted to establish has been denied. It is perhaps somewhat ironic that Kant used essentially Wolffian strategies in order to argue for his anti-Wolffian conclusions.

v What are some of the conclusions that we can draw from this Wolffian interpretation of the Transcendental Deduction, or what difference does it make ~o the more recent Kant scholarship in English? First of all, this Wolffian ~nterpretation shows that Dieter Henrich, who has recently argued that a franscendental Deduction "cannot assume the shape of a rigorous and exhaust-

The WolfJian Background ofKant's Transcendental Deduction 245 ive reasoning," is mistaken. 37 Kant's Deduction was meant to constitute a rigorous argument. Kant's model in the Deduction was not primarily the socalled Deduktionsschriften or "deduction writings," which were produced by such eminent German jurists as J. S. Putter "to justify controversial legal claims between the nunlerous rulers of the independent territories, city republics, and other constituents of the Holy Roman Empire."38 The deduction writers had to decide whether an acquired right was real or whether it was only presumed to be real, and they had to trace the possession someone claimed "back to its origin." Therefore, the "process through which a possession or a usage is accounted for by explaining its origin, such the rightfulness of the possession or usage becomes apparent, defines the deduction. Only with regard to acquired rights can a deduction be given. This implies that by definition a deduction must refer to an origin."39 More importantly, however, since these juridical deductions must always remain "loose argumentation," and since Henrich believes he has shown that Transcendental Deduction is such a kind of a deduction, he claims that it may be embedded in more rigorous reasoning, but that it is, taken by itself, neither rigorous nor logically binding. 40 The question "How is knowledge possible?" does not just aim at a sufficient condition of our claim to knowledge, but at the "real origin of our claim and with that the source of its legitimacy.,,41 This is clearly wrong. The term "deduction" clearly comes from this tradition, but the substance of the Deduction has little, if nothing, to do with it. Henrich is right, of course, when he claims that there exists a certain similarity between the Deduktionsschriften and Kant's Transcendental Deduction. Just as the former are nleant to show that certain legal claims are legitimate, so the Deduction proves that certain concepts are legitimate. Both kinds of deduction are concerned with legitimacy or right, not with simple possession. However, this is where the comparison ends. The Wolffian model identified in this paper shows that Kant was aiming at something much deeper and much more rigorous than Henrich suggests. 42 We may not think that Kant succeeded, and that his approach is wrong in that respect, but this does not give us the right to say that he meant to do something that was never his intention to do. Secondly, it proves once more (if this needs more proof) that Kant's goal could not possibly have been to give a complete answer to the faceless "epistemological skeptic," postulated by some recent Anglo-American philosophers as much as by such figures as Aenesidemus Schulze fairly soon after the Critique appeared. The possibility of experience is not what is proved in the Deduction. Nor is it necessary for this part of the deduction that Kant show how experience is possible. The possibility of experience is simply used to prove the objective validity or necessity of the categories in accordance with the Wolffian approach of justifying concepts in accordance with their principle of proving possibility. It is what we already know to be true, but which can only be true, if we have the categories. To answer the skeptic about objective knowledge, Kant

MANFRED KUEHN would have had to answer the question whether experience is possible in a different way. But he did not do that. And if we construe his answer to the question of how knowledge is possible in such a way that it answers a question that it was never designed to answer, we will find what many a recent commentator has found, namely that "the conclusion of a transcendental argument. ..is not something that can be demonstrated by such an argument." However, to say it again, this hardly constitutes a criticism of Kant. Thirdly, the Wolffian interpretation shows that the Transcendental Deduction contains an anti-skeptical argument. However, this is not an argument against a skepticism about objective knowledge, but an argument against the skeptic about a priori concepts or basic categories of human thinking that are completely independent of experience. It is an argument against what has been called "meaning empiricism." As such, it is also an argument against Hume, who denied the possibility of any a priori concepts. Yet, it is at the very same time a defense of Hume' s principle "not to carry the use of reason dogmatically beyond the field of all possible experience," as I have tried to show elsewhere. 43 The Transcendental Deduction is just as much driven by Wolffian concerns as it is by Humean concerns. It also lends some support to such interpretations as those of Jay Rosenberg which argue that it is best reconstructed as "a piece of practical reasoning, a justificatory argunlent legitimizing a conceptual practice.,,44 If we intend to make the Deduction, fruitful for present-day concerns, this seems to be the most promising route to take. However, if we are more concerned with the historical investigation of Kant's critical theory, then the Wolffian background of Kant's Transcendental Deduction may perhaps serve to call renewed attention to the fact how close Kant remains tied to Wolffianism in the details of his theory. Hans Poser has argued that Kant differs radically from Wolff in his view of modality, and that, "whereas Wolff begins from ontological modalities, it is Kant's goal to answer the question concerning epistemological modalities," and that this represents "the first and fundamental transformation of traditional modalities.,,45 Wolff is clearly not unconcerned about epistemological modalities, and Kant's search for the conditions of the possibility of experience is more indebted to Wolff than it might appear at first. 246

NOTES 1. I take this phrase from Benson Mates, The Philosophy of Leibniz, Metaphysics and Language (Oxford: Oxford University Press, 1986), p. 12. I would like to add that W01ff, though short on "tantalizing hints and suggestions," should not be held responsible for Leibniz's confusions either. For a discussion of the differences between the two see especially Charles A. COIT, "Christian Wolff and Leibniz," Journal of the History of Ideas 36 (1975) 241-262, and Jean Ecole. "Cosmologie Wolffienne et dynamique Leibnizienne. Essai sur les rapports de Wolff avec Leibniz." Les etudes

The Wolfjian Background ofKant's Transcendental Deduction

247

philosophiques 19 (1964), pp. 2-10. Wolff appears to have been rather uncomfortable with Leibniz's theory of irreducible simple elements, conceived as "monads," i.e. spiritual entities, and his so-called inesse principle. In so far as the "praedicatum inest subfecto principle" involves the rejection of the distinction between essential and nonessential constituents of things, Wolff cannot follow it. Indeed, for him this distinction is of fundamental importance. He is also somewhat reserved about Leibniz's theory of monads, and his view that the principle of sufficient reason is a basic principle. Wolff believes it can be derived from the principle of contradiction. 2. Christian Wolff, Ausfuhrliche Nachricht von seinen eigenen Schriften, die er in deutscher Sprache von den verschiedenen Theilen der Welt- Weisheit ans Licht gestellet (Frankfurt, 1726) 28. I quote in accordance with Hans Ltithje, "Christian Wolffs Philosophiebegriff," Kant-Studien 30 (1926), pp. 39-66, p. 55. 3. See also C. A. Van Peursen, "Christian Wolff s Philosophy of Contingent Reality," Journal of the History of Philosophy 25 (1987), pp. 69-82; Jean Ecole, "En que1 sens peut-on dire que Wolff est rationaliste" Studia Leibnitiana 11 (1979), pp. 4561, and Hans Werner Arndt, "Rationalismus und Empirismus in der Erkenntnislehre Christian Wolffs," Christian Wolff, 1764-1754, ed. W. Schneiders (Hamburg, 1983), pp.31-47. 4. See Christian Wolff, Vernunftige Gedanken von den Kraften des menschlichen Verstandes und Ihrem Gebrauche in Erkenntnis der Wahrheit, ed. Hans Werner Arndt, vol I of Gesammelte Werke, ed. 1. Ecole, 1. E. Hofmann, M. Thomann, H. W. Arndt (Hildesheim: Georg Olms, 1965), p. 115; Christian Wolff, Preliminary Discourse on Philosophy in General, tr. R. 1. Blackwell (Indianapolis: Bobbs-Merrill, 1963), p. 3; and Christian Wolff, Philosophia prima sive Ontologia (1730) ## 134, 174, 574 (pp. 115,143,574). 5. Wolff, Preliminary Discourse, p. 6. Anything that excludes logical contradiction is not just possible, but it is a possible thing. It is for this reason that any "notio" or concept corresponds with a possible thing. "a thing that cannot be reasonably argued cannot possibly exist," or perhaps better is an Unding or "no thing at all." In some ways, his theory of objects is very similar to Meinong's Gegenstandstheorie that also seems to be concerned with an investigation of the logical content of objects without any prejudice in favor of existing objects. 6. Van Peursen, "Wolffs Philosophy," p. 73 7. Wolff, German Logic, p. 248. 8. See Riccardo Pozzo, Kant und das Problem einer Einleitung in die Logik. Ein Beitrag zur Rekonstruktion der historischen Hintergriinde von Kants Logik-Kolleg. FrankfurtlMain, 1988. Pozzo rightfully calls attention to the fact that this does not mean Kant did not know Aristotelian logic or was not aware of the Thomasian conception of logic. 9. Wolff, German Logic, p. 139. The word translated as "cognition" is "Erkantnij3." It is important to understand that the secure "foundation" of knowledge consists in concepts of sensation. 10. Wolff, German Logic, p. 139f. 11. Wolff, German Logic, p. 140. 12. Wolff, German Logic, p. 141. 13. Wolff, German Logic, p. 141. 14. See also Gottfried Wilhelm Leibniz, Philosophical Papers and Letters, ed. Leroy E. Loemker (Dordrecht: Kluwer, 1989), pp. 291-295, "Meditations on

248

MANFRED KUEHN

Knowledge, Truth, and Ideas." Wolffs six principles are clearly derived from Leibniz's considerations as found on p. 293: "An idea is true when the concept is possible; it is false when it implies a contradiction. Now we know the possibility of a thing either a priori or a posteriori. We know it a priori when we resolve the concept into its necessary elements or into concepts whose possibility is known, and we know there is nothing incompatible in them ... We know an idea a posteriori when we experience the actual existence of the thing ... Whenever our knowledge is adequate, we have a priori knowledge of a possibility, for if we have carried out the analysis to the end and no contradiction has appeared, the concept is absolutely possible." 15. Wolff, German Logic, pp. 131f. Actually a very similar remark can be found in "Meditations on Knowledge, Truth, and Ideas" by Leibniz: "Whether men will ever be able to carry out a perfect analysis of concepts, that is, to reduce their thoughts to the first possibles or to irreducible concepts, or (what is the same thing) to the absolute attributes of God themselves or the first causes and final ends of things, I shall not now venture to decide. For the most part we are content to learn the reality of certain concepts by experience and then compose other concepts from them after the pattern of nature" (293). 16. Johann Christoph Gottsched, Erste Grande der Weltweisheit, p. 119 17. Hennann Samuel Reimarus, Vernunftlehre (Hamburg, 1766; reprint Munchen: Carl Hanser Verlag, 1979), 90. 18. AA16, 569. 19. AA 16 571f. Adickes dates this very late. I doubt that he is correct in such a late dating. 20. Johann Heinrich Lambert, Neues Organon oder Gedanken uber die Erforschung des Wahren und dessen Unterscheidung vom lrrtum und Schein. (Leipzig, 1764; ed. Gunter Schenk, Berlin: Akamie Verlag, 1990), 365 Anyone who is familiar with Kant's pronouncements on Locke, knows this characterization of Locke. Kant uses very much the same language to describe his relation to Locke. This is perhaps not surprising. What might be surprising is that Kant at times describes his relation to Lambert in the same way. 21. Lambert, 1. H., Neues Organon, 34-40, and 320ff. 22. Ibid.,. 323. 23. To use his words: "Experiential concepts are in themselves possible concepts, and in fact such as we have from immediate sensation."(Ibid., 323). 24. Ibid., 368-369. 25. Ibid., 327. 26. Ibid., 329. 27. Johann Nicolaus Tetens, Ober die allgemeine spekulativische Philosophie (Biitzow and Weimar, 1775), 84f. 28. For a discussion of how Tetens meant to show this, see Manfred Kuehn, Scottish Common Sense in Germany, 1768-1800: A Contribution to the History of Critical Philosophy. With a Preface by Lewis White Beck (KingstonIMontreal: McGill/Queen's U. P.,1987), and "Hume, Tetens, and Kant," HumeStudies 15 (1989) 365-375. 29. Kemp Smith left out the "only," which appears to me an important qualification. 30. See B xxvi n., B 44, and B 52, for instance. 31. B xxvi n. 32. See A151 and AA 5, 47, for instance.

The Wolffian Background ofKant's Transcendental Deduction

249

33. I believe that I am here in agreement with Guenter Zoeller, Theoretische Gegenstandsbeziehung bei Kant (BerlinlNew York: Walter de Gruyter, 1984). Though I am here not interested in the subtleties of the grammar of "objective Realittit" and "objective Gultigkeit," I would wish he had said more about the relation of both tenns to "possibility." 34. See, for instance, Henry E. Allison. Kant's Transcendental Idealism. An Interpretation and Defense (New Haven and London: Yale University Press, 1986 ), pp. 133-172. This observation by itself does not necessarily touch the substance of his argument, however. It may still be that Kant did not accomplish the tasks that Allison ascribes to him, using the tenns in a somewhat anachronistic fashion. 35. It is really quite peculiar that most interpretations slip so quickly from seeing the deduction as the justification these a priori concepts to seeing it as the justification of systematic knowledge of experience. Karl Ameriks tries to show in his "Kant's Transcendental Deduction as a Regressive Argument," Kant- Studien 69 (1978), 273287 that the line of interpretation which has Kant justify knowledge in general is implausible. I do accept his criticism of what he calls the "received interpretation," and my paper is intended to make the same point in a different way. But I am not sure whether his positive account successfully captures Kant's argument. In other words, I am not persuaded that the Deduction is simply a regressive argument. "Possible experience" seems to playa more complicated role than Ameriks' account allows. 36. Lambert, 1. H., Neues Organon, 331,369. 37. Dieter Henrich, "Kant's Notion of a Deduction and the Methodological Background of the First Critique" Kant's Transcendental Deductions, The Three Cn'tiques and the Opus Postumum, ed. E. Forster (Stanford: Stanford University Press, 1989), pp. 29-46. 38. Ibid., p. 32. 39. Ibid., p. 35. And this is, according to Henrich, the reason why the deduction and the origin of knowledge are so closely connected for Kant. 40. Ibid., p. 46. 41. Ibid., p. 35. 42. Henrich is misled by an ultimately rather superficial similarity between juridical justification and Kant's conception of a transcendental justification into making too strong a claim, claiming that whatever holds of the one must also hold of the other. However, there are fundamental differences. The juridical deductions are necessarily empirical, Kant's Deduction is meant to be entirely different from an empirical or "physiological" deduction. It is meant to be a priori, necessary, and apodictic. Nothing loose here! Henrich is wrong on at least two other counts: (i) it is not our claim to knowledge in general that is in question for Kant, but our claim to certain concepts; and (ii) the question is not primarily one of origin, but one of legitimacy, or, as Kant would put it, possibility. The question is whether the categories are possible concepts. 43. Manfred Kuehn, "Kant's Conception of Hume's Problem," Journal of the History of Philosophy 21 (1983), pp. 175-193. (reprinted in David Hume: Critical Assessments, ed. S. Tweyman. London: Routledge, 1995) and "Kant's Transcendental Deduction: A Limited Defense of Hume," in New Essays on Kant. Ed. Bernard den Ouden (New YorklBern: Peter Lang Publishing, Inc., 1988), pp. 47-72. 44. Jay F. Rosenberg, "Transcendental Arguments Revisited," The Journal of Philosophy 75 (1975), pp. 611-624,623. See also Moltke S. Gram's reply in the same issue of the journal, pp. 624-626 and Barry Stroud's "Transcendental Arguments and

250

MANFRED KUEHN

'Epistemological Naturalism.'" Philosophical Studies 31 (1977), pp. 105-115 and Rosenberg's reply in the same journal pp. 117-121. Kant could not, of course, have bought into Rosenberg's naturalism. It is not concerned with one practice among other, but with the only possible practice. 45. Poser, Hans. "Mogliche Erkenntnis und Erkenntnis der Moglichkeit. Die Transfonnation der Modalkategorien der Wolffschen Schule in Kants kritischer Philosophie." Grazer Philosophische Studien 20 (1983) 129-147, 133f. Bemward Grunewald, ModaliUit und empirisches Denken. Eine kritische Auseinandersetzung mit der Kantischen Modaltheorie (Hamburg: Felix Meiner, 1986), on the other hand, has only three incidental references to WoItl and only one to Leibniz.

Between Medicina Mentis and Medical Materialism CATHERINE WILSON

At the close of the sixth memoire of his Rapports du Physique et du Moral, (1803) Pierre-lean-George Cabanis (1757-1807) remarks: We must learn to know the physical man if we are to study fruitfully the moral man, in order to learn how to govern the customary actions of the mind and the will through the customary actions of the organs and of the temperament. And the more one progresses with this sort of improvement, which has no limit, the more one senses how important the study which occupies us is; so that it will be one of the greatest moments for astonishment in our descendants to learn how, among people who passed for enlightened, and who really were in many respects, this study hardly figured in the most sophisticated systems and in the most respected educational institutions. 1

It appears at first a simple matter to date the thoughts expressed here to the last quarter of the eighteenth century or the beginning of the nineteenth. The disparaging remarks about systems and educational institutions, the references to enlightenment, posterity, and progress are typical of a period influenced by Locke, Condillac, Rousseau, and Condorcet. Just as typical is the reference to "physical man" and the suggestion that moral-intellectual improvement will follow from the detailed study of the body, its organs and its organization. These improvements are not keyed, as they are in earlier writers bent on the improvement of the human race, to the development or application of any organon, that is, any system for manipulating terms, concepts, experiences or ideas, but to applied knowledge of the functioning of bodily organs. We can see this as the natural development of La Mettrie's idea that it is only those physicians who were incidentally philosophers, rather than mere philosophers, who have the right to speak about the soul, for they have "traveled through and illuminated the labyrinth of man~ they alone have laid bare to us those springs [of life] hidden under the external integument which conceals so many wonders from our eyes.,,2 The notion that medical approaches rival approaches directed to the mindthat medical materialism interferes with the idea of a medicina fnentis in eighteenth-century thought-is a curious one, and the relations between philosophy and medicine in the period from Ramus to Kant are complex. Here and there the concepts run parallel: "We are ignorant," Leibniz says, "of the medicine of bodies and of minds. We treat the former as does an agent for the sake of gain~ we treat the latter as a boy does his lesson-as nothing, for he learns it in the hope of forgetting it.,,3 In La Mettrie, medicine strikes at philosophy under the pretext of being a kind of philosophy. The question

252 CATHERINE WILSON whether the idea of a medicina mentis is at odds with medicine-taken in its broadest connotation-has many faces. Contemporary philosophers, for example, struggle with the relation between descriptive and normative accounts of knowledge-acquisition, and the question whether there can be a pure epistemology or a theory of scientific knowledge-acquisition, distinct fronl the empirical study of human information-processing. In other regions of inquiry, it is unclear to what extent cognitive disorders can be helped or managed by remoulding the subject's habitual patterns of thinking without direct intervention into the organism's physiology. These remarks are I hope sufficient to motivate attention to physiology and medicine in a volume dedicated to logic and the workings of the mind. My purpose here is to show how the notion of universal human reason, though assaulted by the new enlpiricism and by medical approaches to the study of the mind, survives this assault by demarcating new boundaries between philosophy and other subjects, with Kant playing a major role in this exercise in purification. To begin with, it is essential to remember the degree to which the idea that it is the body which thinks was familiar to the canonical seventeenth-century philosophers, even to those whose names are associated with such immaterialist excesses as monads, intelligible extension, and vision in God. Moreover, skepticism about the traditional regimens for mental improvement is evident even in an age addicted to novel schemes and systems for the amelioration of human reason. On reflection, one may come to doubt that the main ideas expressed by Cabanis above really required the intermediate efforts of the figures mentioned as preparation. Expressions of contempt for logic, for "grammar," and for the conceptual analysis of the Schools are fanliliar to us from Campanella, van Helmont, Bacon, and Descartes himself. Descartes in particular was frustrated by educational institutions, concerned with progress and posterity, and saw medical knowledge as the key to moral-intellectual improvement. It is towards the close of Discourse VI on the last page of this treatise that we find the following words: "I will say only that I have resolved to devote the rest of my life to nothing other than trying to acquire some knowledge of nature fronl which we may derive rules in medicine which are more reliable than those we have had up till now.,,4 To nothing other? We might consider this a rather temporary statement of Descartes's aims were it not for the fact that he says explicitly in the Preface to the French edition of the Principles of Philosophy that the principal benefit of philosophy "depends on those parts of it which can only be learnt last of all," to wit, medicine, nlechanics and nlorals rather than nletaphysics and physics, 5 leading us directly to the hypothesis that the project of pure enquiry is actually the improvement of medical practice and morals-taking "morals" in the seventeenth-century sense not as the solution to moral dilemmas but as psychological health and content-

\ment.

6

Between Medicina Mentis and Medical Materialism 253 It is only recently that physiology and medicine in their relation to Descartes's overall project beyond their role as expressions of a mechanistic approach to phenomena either irrelevant to or at best complementary to an immaterialist account of the mind have begun to receive the attention they deserve. 7 And with good reason: efforts to present Descartes as an extremely discreet materialist have never been successful. A review of the evidence shows that Descartes-however duplicitous and masked in his broader intelle'ctual commitments-presented himself in his treatises, essays, and letters as implacably opposed to the thesis that thinking, unlike sensation and even perception and dreaming, depends upon physical organs. Challenging those philosophers who "take the view that the formation of thoughts is due to the combined activity of parts of the brain," he argues that they have no positive argument for their conclusion and simply lack experience of disembodied thought: in this they are like men whose legs have been shackled from infancy who conclude that the shackles are part of their legs and are needed for walking. 8 The mind joined to the body uses it as its instrument, and the instrument may be impaired by wine and other corporeal substances, but the mind itself does not weaken when the body does and does not require a suitably developed body to think: the fetus-presunlably from the moment of conception-thinks-presumably like an adult. 9 Nevertheless, Cartesianism is ambivalent about the place of the body in philosophy, and one expression of this ambivalence is a bifurcated appeal to ideas about medicine and therapy. On one hand, Descartes launched an influential program of renlediation supposed to operate on the purely intellectual level. Philosophy and philosophical method are seen as elements of a medicina mentis, a therapy for an underfunctioning mind, but Descartes also laid the groundwork, not only in his nlany speculative writings in sensory physiology, theory of generation, and anatomy, but even in those writings of his which are designated as philosophy for the medical approach to mentality so contested from Locke onwards. These vectors point, I shall argue, in opposing directions, despite the common conception of philosophy as ameliorative. The idea of a medicina mentis is linked to ideas about the uniqueness of the human mind in animated nature, and its essential sameness in every human. It carries with it the implication that every mind is improvable insofar as it is willing to submit itself to an intellectual regimen. Medical nlaterialism, by contrast, leads to the fragmentation of the faculty of reason. It leads to a picture of the mind in which the cognitive faculties in humans are not unique but continuous with those of animals and represent in turn only a refinement of perception. At the same time medical, materialism confronted the universalism of the Cartesian idea of the res cogitans by discovering a multiplicity of minds: adolescent minds, the minds of infants, the minds of women, of geniuses and madmen, each characteristic and distinctive and arising out of the organic foundation of tissues and textures. To be sure in Cabanis, as the best expositor of this multi-

254 CATHERINE WILSON plication and differentiation of minds and bodies, the healthy male mind is unmarked in the linguist's sense: it is the standard against which all other minds become describable without being describable in itself, for it has no peculiarities. I shall discuss later some consequences of this residue of the generalized Cartesian mind at the heart of Cabanis's theory. In the last section of the paper, I want to discuss the evident tension between what might be called pure philosophical views of the mind and material or anthropological views as these make themselves evident in the work of Kant. Part I: Method as Medicina Mentis The subject of Descartes's Rules is often taken to be a disembodied being who employs his intuitio mentis rather than corporeal vision to grasp the formal nature of a problem and the elements that must be conlbined or recombined to effect a solution. The clear and distinct perception that enables the meditator of the Meditations to reach such exceedingly salutary conclusions as the existence of God and the immortality of his soul is an example of the proper use of intellectual intuition, which, unlike the sensory faculty, is particularly suited to the conceptual examination of incorporeals. Yet this intellectual vision is to be explicated by comparison with ordinary vision-hence the terms "clear and distinct"-and, in the Rules, Descartes even recommends tasks requiring good visual acuity such as embroidery as preparation for conceptual work. Although we do not have to my knowledge a definitive study of the reception of either Descartes's Discourse on Method or of the Rules, first published in Dutch in 1684, in Latin in 1701, we can safely assume that Descartes's methodological writings influenced later conceptions of medicina mentis--in Tschirnhaus, for example, whose logic of invention demands the proper ordering of definitions, axioms, and theorems, as well as the decomposition and recombination of elements-and in Leibniz and Wolff. lo These logics of invention, which optimistically aim for mathematical demonstrative certainty and at the same time for application to scientific problems, seem to fall between the theory of the syllogism and the method of analysis and synthesis actually applied to physical science bye. g. Bacon. As several papers in this volume have made apparent, school-logic had great staying power in t~e seventeenth- and eighteenth-century curriculum, despite the contempt routinely shown by philosophers. We may wonder at the ancient notion that abstract and formal studies, especially mathematics, possess therapeutic power unrelated to their capacity to produce novel results, that they are a form of mental exercise or mental discipline parallel to physical exercise (another form of exertion that creates no external product) and medical regimens, but this idea still exercises an influence in the modern curriculum. The idea is that it is the studies themselves, not through the content that is absorbed but through the very process of disciplined attention to natural or

Between Medicina Mentis and Medical Materialism 255 artificial language and the syntactical rule-governed manipulations of formal studies, which confer a benefit. It is widely acknowledged that instruction is difficult, that learning is for most pupils distasteful, yet the notion that such regimens produce an irreversible improvement in the epistemological capabilities of the learner is firmly rooted. In this they reserrlble those regimens in the history of medicine (restrictions on diet, giving and withholding of liquids, forced movement and enforced immobility) whose relation to actual outcomes does not explain the seriousness with which they are prescribed and followed. Now, an important difference between Descartes and Leibniz is that Leibniz thought that traditional logic, as the study of terms in combination, could be converted into an engine of invention and discovery, while Descartes appears to have thought that one should ignore traditional logic, try to determine what one was doing in mathematics which led to an increase in knowledge, and import that method into knowledge-seeking generally. Notoriously "what one is doing" in mathematics was understood by Descartes in terms of the subjective experience of the mathematician and characterized in terms of mental vision, analysis, stepwise movement, enunleration, and so on, rather than in terms of the theory of proof. 11 Yet the idea of a regimen and a discipline is not lost. The attention and repetition of method are opposed to the natural idleness, willfulness, impulsiveness, and inattention of the mind. By such means is the mind to be brought to exercise its faculties properly in judging, deciding, and knowing. As observed, the immaterialist strain in Descartes appears to anchor his conception of method even while the method is explicable only by analogy, so that the very terminology betrays an interweaving of mental and physical: (mental vision: but not casual sense-perception~ clear and distinct perception: but not the confusion of ordinary visual apprehension). The thing that knows, that comes to an explicit understanding of what it is doing in generating mathematical knowledge-the knowledge which will be turned to physics, to optics, and via physics and optics to an understanding of the mechanisms of the body and thence the passions and thence, turning back on the body, the improvement of morals,-cannot be the body. The presence of what might be called a "metaphysical interlude" in Descartes's otherwise physical investigations should not surprise us. This metaphysical interlude was not only a screen for an evolving science of nature which would have brought down on itself the usual criticisms of vanity, uncertainty and impiety and a device for better infiltration of the Protestant and Catholic educational establishments. Nor did it simply provide needed theological and metaphysical foundations for the new science by proving the existence of a veracious God. Although dualism cleared away the world of intermediary entities which Inade a rational and practical science of nature impossible, the foundationalist interpretation of Descartes is difficult to sustain, given Descartes's rather pragmatic attitude towards foundations, as exemplified in his account of a true and certain science

256 CATHERINE WILSON based upon imaginary foundations in the Principles. 12 All the same, Descartes seems to be tenacious about dualism in a way in which he is not tenacious about the afterlife or our religious duties. I3 . And this is perhaps because it seemed to him that for the mind to achieve an understanding of its own body, an understanding that permitted it to alter and improve its condition, it could not be that body, and, conversely, that bodies could not understand mathematics, mechanics and optics, in order to achieve that ameliorative understanding. Bodies cannot perform mathematical operations, nor comprehend their own interior workings. Hence Descartes's position exhibits an official dualism within a broader program of material technology. The Passions of the Soul is sometimes read as a sort of corrective to an untenable dualisnl undertaken in response to the criticisms of Princess Elizabeth. If the above analysis is right, this cannot be the correct interpretation of that work. A better way to see it is as the continuation by other means of Descartes's long-announced programme. The purpose of positing a separate soul-substance, as he had already made plain in the closing passages of the Treatise ofMan of c. 1629, was to delimit the function of the soul and so to make as many "external movements" and internal processes including appetites and passions as possible intelligible in mechanical terms and thus theoretically susceptible to the intervention and direction of the physician. 14 As he told us in the Discourse on Method: Even the mind depends so much on the temperament and disposition of the bodily organs that if it is possible to find some n1eans of making men in general wiser and more skillful than they have been up to now, I believe we must look for it in medicine. ...Intending as I did to devote my life to the pursuit of such indispensable knowledge, I discovered a path..." 15 Later in the Passions, Descartes appears to have abandoned hope of acquiring

the appropriate sort of micromechanical knowledge and to settle not for the improvement of health, but the agent-initiated improvement of morals. 16 This improvement depends nevertheless on reconceiving the passions as scientifically intelligible, inevitable, and even beneficial to the creature which experiences them. Part II. Materialism and Mentality For La Mettrie, Descartes had been merely insincere and was to be considered the true founder of the view that thinking is a function of matter. But the evolution of the doctrine of thinking matter required drastic revisions to Cartesian ontology, revisions which marked in some ways reversions to earlier naturalisms. 17 It is well known that the mechanism of the seventeenth century, which imagined matter as corpuscular, inactive, possessed of primary qualities alone and needing to be set into motion by a divine or human mind, ceded to the picture of matter as locus of active forces, parts of which, when suitably

Between Medicina Mentis and Medical Materialism 257 organized give rise to vitality and mentality without animation by a distinct soul-substance. As early as 1704, John Toland's Letters to Serena announce definitely that "Motion is essential to matter, that is to say, as inseparable from its Nature as Impenetrability or Extension.. .1 deny that Matter is or ever was an inactive Lump in absolute Repose.,,18 Newton's ascription of powers to matter blended with Leibnizian pananimism to support variously conceived systems of organic molecules or matter as a network of forces, all of which by spiritualizing matter raddled the notion of agent-minds. John Yolton's work has recently acquainted us with the reception-history of Locke's supposition that matter might think. Yolton showed the fascination with which Locke's doubts about the very intelligibility of "thinking substance" and his suggestion were taken Up.19 According to Toland: whatever be the Principle of Thinking in Animals, yet it cannot be perform'd but by the means of the Brain. We Men are Conscious of no Thoughts, while the Functions of the Brain are suspended; we find ourselves to think there, and only there;.and we observe no signs of Thought in any things that want a Brain, whereas every Creature that has one, seems to show some degree of Thinking by its Actions. 20

But the materialization of thought hinted at in Locke is broader than the suggestion that matter may think; it extends to the role of the new way of ideas-later transformed into the "ideology" of the French-and to the psychologism and powerful sensationism-sensuality even-of the Lockean text. The improvement of human reason-and incidentally of medicine (v. Locke's Epistle to the Reader)-requires an understanding of natural (imprinting and associating) as opposed to idealized reasoning processes such as the syllogism. As is well known, Locke's diagnosis and treatment of the errors of human reason is both similar to and different from that of his method- and regimenoriented predecessors, despite some similarity in their complaints, which concern the problems of stagnation in the exact and moral sciences and individual epistemological failure, problems they attributed to vagueness and non-referentiality in both learned discourse and in the vernacular. Schoollogics, artificial regimens, are rejected: and the actual processes of ideaformation, once recognized, are legitimated by the philosopher as appropriate to our situation. Error can be avoided by hewing to the paths of nature and so the artifices of philosophical discourse are attacked on two fronts. Ideas whose source can be located in experience even though metaphysicians and theologians have attacked them as meaningless or impossible (e. g. power) are legitimated. At the same time, pseudo-ideas (e. g. thinking substance) which metaphysicians exchange freely are de-legitimated. Note that we have already moved away from the Cartesian conception of method as an inner discipline. Our problems, it might be said, are material rather than formal. We do not need to change the way in which we reason, the

258 CATHERINE WILSON way in which we move stepwise from line to line in a proof. The movement of the mind occurs naturally through association-and over that we have no control. Our task is to effect a restocking of our mental contents: to ensure that the ideas in our minds are those and only those that a mind uncorrupted by teachers and words would have achieved by leaving its senses wide open to the world. Hence the primitivist element in Locke: the experiences of babies and children are valorized as uncorrupted by deviant language and metaphysical subtlety, in contrast with the Cartesian complaints about the puerility of reliance on the senses. In keeping with this vicarious pleasure in the natural and naif is the strikingly sensualistic language of the Essay: the taste of pineapples, the scarlet of the trumpet, the scent of violets. By contrast with the subtle and qualified versions to be found in Gassendi and Locke, eighteenth-century materialism can seem crude. "Crude" is indeed its usual predicate. And indeed, it was crude in a number of ways: above all, materialism served as the foundation for encouragements to libertinage, and writers like La Mettrie are frequently lewd. But equally important, philosophical materialism did not meet the usual standards of philosophical cogency in argument. As Luzac says in his response to La Mettrie's L 'homme machine: ...the physician's ignorance of logic causes him to reason that a medicine which restores a sick person to health, has sometimes this effect, that it changes a blockhead into a man of sense: Nothing more is wanting for him to conclude, that man is no more than a watch; and that it is sufficient his spring and wheels be in good order to render him reasonable. Thus it is that man is a machine 21

Athhough we do not have here a complete lack of "positive argument" as Descartes charged, medical materialists appealed both to banal and easily countered observations such as the effects of alcohol, opium, and injuries to the head on mental performance, or to questionable observations of events of questionable frequency and representativeness, combined with dubious inferences ("no more is required") and a speculative medical ontology (springs and wheels!) left over from the previous century to create a context of total implausibility. Cabanis, too, bases his case for the rapports of the physical and the moral on anecdote and generalization. A fever can turn a genius into a dunce-or for that matter a dunce into a genius. Pedagogic regimes do not seem to be relevant. The brain, as he says notoriously, must be considered as a special organ for producing thought, "as the stomach and intestines produce digestion, the liver filters bile, and the parotids, and the maxillary and sublingual glands manufacture saliva. ,,22 Yet in its detail, subtlety, and comprehensiveness, Cabanis's materialism is far from crude. His attention to the instincts of birds and kittens, the anguish of adolescence, the brief moment of brilliance of young girls just before they reach marriageable age, and the varying appearances and textures of the brain and internal organs is unsurpassed in delicacy and sensitivity. He imagines the entire body as the determinant of error and delusion. It

Between Medicina Mentis and Medical Materialism 259 is the stomach and the organs of generation that are the sources of madness, delusions, and strange imaginings and fantasies, while amorous and ascetic melancholy produce "les plus funestes explosions et laissent apres elle des traces ineffables.,,23 The optical metaphors which dominate Bacon's and Descartes's discussion of method: mental intuition, clarity and distinctness, separation, are countered here by the privileging of touch as the basic sense: the first to appear, the last to disappear, "la sensibilite meme,,24 which unifies the animal kingdom (and even reaches down into the world of plants). In the lower species, the nervous and muscular systems are not distinct.: the animal is governed by irritability and reactivity. A higher animal is characterized by the progressive differentiation of the two systems. This brings with it the higher faculties and greatly increased scope of action and reaction, but also exposes it to nlore dangers and leaves it with less secure responses. 25 The "will" and the "self' are not metaphysical entities but ideas originating in effort and resistance against the not-self, including the lirrlbs of one's own body.26 Cabanis's recommendations for a further study of the physiological bases of differences in tenlperament, character and sensibility with an eye to the medical and eugenical improvement of the human race to make it "plus sage et bon" represent in one way, a continuation of the materialistic approach to psychological phenomena which is evident in Descartes. But it is also a betrayal of the Cartesian program. The Cartesian portrayal of human reason as a faculty uniquely human, extranatural, present in everyone from conception onwards, and improvable through the application of systems of order and method cedes at the end of the eighteenth century to a different picture. In this picture, which is encouraged by the cult of sensibility and the rise of the psychological novel and its interest in mentality and sentimental excess, reason is a mode of sensibility, humans differ from animals only in the range and fineness of this faculty, but differ considerably from each other depending on their age, sex and constitution, heredity, and circumstances. The thoughts of children whose brain substance is soft and pliable are impetuous, lively, numerous, unstable and uncertain~ the ideas of women in whom the cerebral pulp is more abundant than in men are characterized by imagination, superficiality, and vivacity: "un talent facile enleve legerment la superficie des objets.,,27 Adolescents have denser bodily tissue, which accounts for their notorious moodiness, religious susceptibility, romanticism and propensity to acute illness. Education accordingly is not a matter of the discipline of the logic book, but of instruction suitable to the way of life to be followed by the creature and its natural dispositions and tendencies. We might speak here not of the fragmentation of reason but of the destruction of reason as a faculty. As Foucault noted, women and their psychology became in the Enlightenment more interesting to nlen-and interesting as cognitive subjects, albeit as cognitive subjects organized according to nature's reproductive purposesthan they were before or have been perhaps since. It is difficult to say whether

260 CATHERINE WILSON this form of exquisitely focussed attention was good for them or not. The Cartesian program seems to include women and even babies, as we have seen, in its universal we, but it is widely claimed that this universalism is antifemale. Cabanis, (who admits to drawing, not on observations from life for his knowledge of women's psychology, but on the novelist-essayist Rousseau's character of Sophie from his Emile) is a major proponent of "difference." That such empiricisnl may support sexism and racism28 is evident, but the connection runs otherwise in the other great proponent of cultural difference, J. G. Herder, who saw a supreme value in the particularity of isolated cultures and modes of life, and who devalued universal reason accordingly?9 But this is not the place to settle the question whether more good or evil has been done by empiricisnl by contrast with the probing and articulation of the significance of the universal we. 3. Man is not a machine...Medicine is not philosophy.

According to Ernst Cassirer, eighteenth-century materialism was "an isolated phenomenon of no characteristic significance."3o Cassirer appears to base this assessment on his estimation of the crudeness and dogmatism of Helvetius, La Mettrie and Holbach,and on the rejection of materialism by Voltaire, D'Alembert, and other influential philosophes. A similar minimalizing tendency is apparent in Lewis White Beck's Early German Philosophy. Early German Philosophy has one page reference to "materialism." Rather in contradiction to Cassirer, Beck considers materialism to be a dangerous French idea from which German philosophers were concerned to distance thenlselves and which they ignored. Various explanations have been and might be called into play to explain this lack of Einklang as represented by open polemical debate anlongst intellectually equal partners in Germany by contrast with France and England31 -the underlying spirituality and theological seriousness of the populace and its rulers by contrast with French levity and libertinism, different practices of censorship or successful censorship evasion, later, horror at the aftermath of the Terror-the ultimate result of the antiauthoritarian logic of the philosophes. Even the chauvinistic belief that the national-genius Lejbniz had trounced Locke and proved the immortality and indivisibility of the soul might be held responsible for the continuation of metaphysics, and school-philosophy. The rigidity of the German academic system in which the Wolffian tradition remained strong and a few textbooks continued to form the basis of the curriculum is of primary importance. However, methodological problems complicate the picture: it is difficult to distinguish causes from effects in the above "explanations." Second, the historian may argue that there is no question of philosophy failing to come to terms with medicine in Germany; rather the philosophical tradition was carried on in Germany, but disintegrated into science, popular science and belles lettres in France.

Between Medicina Mentis and Medical Materialism 261 The conclusions of Cassirer and Beck are in any case not sustainable; they appear to be based on an assessment which takes the claims of the canonical philosophers of the period-notably Kant and Hegel-that medicine and philosophy are independent as self-evident, rather than interpreting their assertions as proof of the encroachment of mediciue on philosophy. The historians find no trees but miss the forest. For although Kant devotes a derisory six lines in the section on the "Paralogisms of Pure Reason" to the refutation of thinking nlatter,32 it is evident that the system of transcendental idealism is itself intended as a refutation of materialislu, as philosophy versus medicine, and that Kant is persistently concerned to separate the empirical side of psychology and anthropology from the study of the subject of knowledge and especially of ethical action. Kant writes: "In unserem Koerper ein feines Gewebe von Nerven ist... [Darauj] wirken unsere Ideen und zwar aufverschiedene Art.,,33 But this fabric of nerves is not I, and the Critique ofPure Reason presents a system that uses skepticism about essential natures to preserve a discourse based on faculties, immaterial entities, and unexperienced future states. Indeed Beck's own single footnote to materialism turns out to refer us to his own claim that the separation of two causal orders, mechanical and teleological is "the highest point in Kant's theoretical philosophy."34 How, one might wonder if nlaterialism was of no general significance in eighteenth-century thought and specifically in eighteenth-century German thought, could this separation constitute the highest point of Kant's critical philosophy? Recalling Goethe's description in Dichtung und Wahrheit of Holbach's System ofNature as a forbidding book whose contents, Goethe found, when he had the nerve to approach them, disappointingly dull, we may venture the hypothesis that materialism c. 1780 in Germany was a subject about which there existed a kind of derivative and vague knowledge, which nevertheless elicited a critical reaction.35 But there are many strains in materialism: Kant's insistence on a sharp distinction between the faculties of cognition and sensibility apparent as early as 1771 is normally taken as directed against visionary mystics like Swedenborg. But it should perhaps be taken as directed against the emerging nature philosophy of the romantics and forms of hylozooisnl and pananimism deriving from Leibniz and others, for, as we have seen, the near erasure of this distinction is a feature of Cabanis' s system, and generally of systems built on vegetable models of organic life. If we distinguish between various forms of materialism-including the English agnostic-mystical variety of materialism represented by Priestly,36 who constantly stresses the fact that the matter behind our experiences unlike the material object we meet with in experience is not actually solid/7 the dogmatic, profane, hedonistic version of materialism espoused by La Mettrie, and finally romantic nature-philosophies, it is possible to find evidence of a reaction to all three. Kant seeks to build on ignorance and failure-our ignorance of the natures of matter or of mind, our failure to provide a Newtonian account of the growth of a blade of

262 CATHERINE WILSON grass-something positive: the certainty of human uniqueness in the creation, dignity, autonomy, and the hope of a future world. Thus the critical philosophy introduces its terms both according to the nlinimalist principle that we cannot introduce more than we know (for example, an immaterial, immortal thinking substance) in order to conlbat spiritualism, while it employs a different principle to combat materialisnl. This is the principle that we may introduce (==resort to) a term as long as it satisfies our reason, by which Kant frequently means "effects a reduction in our anxiety." "Why do we have to resort to a doctrine of the soul founded exclusively on pure principles of reason? Beyond all doubt, chiefly in order to secure our thinking self against the danger of materialism.,,38 And more explicitly, "Rational psychology exists not as doctrine, furnishing an addition to our knowledge of the self, but only as discipline. It sets impassable limits to speculative reason in this field, and this keeps us, on the one hand, from throwing ourselves into the arms of a soulless materialism, or on the other hand from losing ourselves in a spiritualism which must be quite unfounded so long as we remain in this present life.,,39 Note that it is materialism that stands there with open arms. Some interpreters of Kant may think that, on his view, the mental operations described or referred to in the Critique of Pure Reason and elsewhere may be performed by what is in fact only a body, although we can never know that this is the case, and we must think in the absence of this knowledge one way or the other that it is not. 40 This is a bold but unlikely reading. If Kant is construed as a realist who holds that states of affairs may obtain and propositions nlay be true independently of our ability to know them, he still cannot acknowledge that it is possible that our thinking and actions are performed by what is only a body subject to mechanical laws. For in that case, despite any belief, faith or hope to the contrary, there is no noumenal self, no free will (on Kant's assumptions about bodies) and, even in light of the most ferocious attachment to its ideas, no nlorality, except as all these are figments of the imagination dreamed up by the body. Now Kant does not infer from the objectively binding character of morality to the falsity of materialism and thence to the truth of spiritualism. It is thus proper to regard him as a sort of intuitionist who rejects indirect proofs. But if he is so interpreted, then even the disjunction, "Either it is the body which thinks and acts or spiritualisnl is true," is not assertible by Kant. The situation is even worse if he is construed as a verificationist, for then the proposition that it is the body which thinks and acts is unprovable and literally meaningless. It is debatable whether spiritualism would actually ground morality by providing us with justified hopes of survival and reward, but at least Kant does not appear to think that its truth is inconsistent with morality. Thus materialism and spiritualism, though both designated as dogmatic options, are regarded differently by Kant and treated nonsymmetrically. For example, reference to

Between Medicina Mentis and Medical Materialism 263 materialism's hypothetical possibility, as a dialectical weapon against dognlatic spiritualists, is not recommended although the reverse is. 41 Is it the case, one might wonder, reflecting not only on the case of Kant, but of any of his rationalist predecessors, that the official branches of the nlajor philosophers's philosophies contradict, in the name of human hope, dignity and non-animality, the experience of mortality, limitation, and errlbodiment which they address under other headings in their works? If so, it would follow that there can be few materialistic philosophies. When philosophers do not appear to grasp this fully and insist on engaging with materialism as though it had to do with them, their editors and commentators see to the purification of the discipline themselves. This suspicion is strengthened by a look at the great synthetic historians of philosophy whose concern was to portray the history of philosophy as a continuous tradition of resistance to empiricism and science, both of which tend to attract the predicate crude. For Hegel, who saw the origins of modern philosophy as stemming from the independence from religion, philosophy is nevertheless an unfolding of the progress of the self-discovery of Spirit, as the introduction to his influential Lectures on the History of Philosophy informs us: There is an old tradition that it is the faculty of thought which separates men from beasts; and to this tradition we shall adhere...We must. ..consider it best when thought does not pursue anything else, but is occupied only with itself-with what is noblest-when it has sought and found itself. The history which we have before us is the history of Thought finding itself, and it is the case with Thought that it only finds itself in producing itself ..These productions are the philosophic systems. 42

By contrast, materialist histories such as that of Friedrich Albert Lange,43 appear to tell the story of disorganized, sporadic attempts on the part of materialists to fight against prejudice and superstition, not of matter finding itself by revealing itself. The spiritistic bias in the historiography of philosophy goes well beyond the spiritistic bias in philosophy itself. However, even if it leads commentators to find no trace of materialisnl in eighteenth century philosophy, its presence can be established, and not only in the work of minor authors. If we consider each of their oeuvres as a whole, it is apparent that not one of the canonical philosophers, with the possible exception of Hobbes, who is read as a political philosopher and not a metaphysician, has a consistent story on the mind and the body. All that is consistent is the pattern of simultaneous assertion and denial, of multiple rival accounts. Seventeenthcentury metaphysics, despite its fascination with invisible entities and unexperienced states of affairs (the soul, God, immortality) is more interested in the body, more involved with the shadowy other of nlentality, than one might expect.

264 CATHERINE WILSON I return in this connection to the problenl of Kant. Kant is undoubtedly a philosopher, for he holds to philosophical terminology~ he discusses God, the soul, and the world, he discusses modes, qualities, accidents, causation, and function. Yet Kant also discusses racial and sexual difference, diseases of the brain, the brain-water and the ether, and the empirical part of psychology. As has been convincingly shown by Robert E. Butts, Kant engages willingly throughout his career with the idea of a physiological basis for defective thinking, and with the naturalistic study of human beings. 44 At the end of 1773 he writes to Marcus Herz, himself a physician: This winter I am holding for a second time a private seminar on anthropology, a subject which I am now thinking of getting officially recognized as a academic discipline. My intention is...to reveal the sources of all knowledge, those of morals, of skill, of intercourse, of the method of educating and ruling man, and consequently of everything practical. After that I seek further the possibility of the n10dification of human nature in general. Thus, that subtle enquirydoomed in my eyes forever to failure-that subtle enquiry into the manner in which bodily organs are related to thoughts is omitted... 45

Is the failure to explain the manner in which bodily organs are related to thoughts referred to a failure of will or skill? The Preface to the Anthropology of 1798 satirizes those "physiological" anthropologists who (like Descartes, Kant says) run hither and thither chasing the traces of sensations in the brain. Our inability to find and manipulate the nerves and fibres in the brain creates a separate realm, a place for a surface-level, "pragmatic" or phenomenological study of man which will result in modification and improvement, in addition to a place for an applied metaphysics. 46 What then is the relationship of Kantian reason to the differentiated modes of mental address of the Anthropology? The anthropological works suggest that, in contrast to his ronlantic contemporaries, Kant was incapable of registering particularities of mentality other than negatively. It is safe to say that, in every work of Kant's, either human reason and action are to be discussed in extra-empirical terms, or human cultural, psychological and physical diversity are assessed as departures from an idealized and dematerialized norm. The empirical example serves typically as an antitype for the philosophical concept: properly functioning critical reason is to be understood by contrast with the ravings of a Swedenborg, as the metaphysical concept of goodness is explicated by pointing to the radical evil evident in the reported behavior of New Zealand savages. 47 But even the improvement of the human social personality which the Letter to Herz promises as a benefit of the study of anthropology is not as important for Kant as the moral improvement which depends on grasping that man is not an entirely material being, the subject of physiology, psychology, or anthropology. For Kant, as well as for Descartes, mind cannot entirely be part of the world order if their aims~~1~~~~hie~~--,--- __

Between Medicina Mentis and Medical Materialism 265 To conclude, it is evident that the major philosophers of the seventeenth and eighteenth century, for all their interest in immaterialist systems and universal regimens for reforming reason in general, had pronounced medical interests, that consciousness of the body and a conviction that it was necessary to understand its workings to improve the human character is evident in their works. Descartes's investigations into the brain and the bodily accompaniments of the emotions coexist with his official dualism. Locke's experience as a physician and his concern to recover the pure experiences of childhood co-exist with his psychological and political atomism and his conception of the mind as a box of inner objects endlessly combinable and replenishable. Spinoza's programmes for emotional control and emendation of the intellect co-exist with his rejection of individuated res cogitans. Yet they remain, like Kant, within the boundaries of philosophy, as Descartes's unfortunate disciple Regius, and, after him, La Mettrie, Cabanis, and Herder conspicuously failed to do through lack of proper attention to the problems of substance, causation,-and method. Does the idea of a medicina mentis then give way to medical materialism in the course of the early modern period? The answer must be yes and no. On one hand, subject-specific conceptions of proper method and procedure (scientific, anthropological, aesthetic), develop to replace all-purpose utopian regimens for reforming reason and thinking well in general. Kant, whose belief in world peace brought about by reason is neither more nor less commendable and naive than Leibniz's, can make light of Leibniz's Universal Characteristic in this connection. The Cartesian mathesis universalis, meanwhile, diversifies into an army of formal techniques in geometry, analysis, and probability theory. At the same time, the idea that mental functioning is dependent on the organic texture and vital activity of the brain and body comes into prominence, as I have tried to show, not just in opposition to Cartesianism but through its influence. Philosophy as a discipline nevertheless succeeded in opposing these tendencies to diversification. Until recently, the theoretical fragmentation of the intellect and the practical fragmentation of the sciences which were first recognized in Kant's time seemed scarcely to affect epistemology or the philosophies of mind and action. Philosophers redefined their territory by discovering that there are always foundational and systematic questions to be asked-and also by specifying what philosophy was not going to be concerned with: organic texture and variability. One branch of Cartesianism was sacrificed, that the other might live.

NOTES 1~ P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) XLIV-1: 358: "[1]1 faut connoltre l'homrne physique, pour etudier avec fruit l'homme moral; pour apprendre a gouvemer les habitudes de l'esprit et de la volonte, par les habitudes des organes et du temperament. Et plus on avanca dans cette route d'amelioration, qui n'a

266

CATHERINE WILSON

point de tenne, plus aussi I' on sentira combien I' etude qui nous occupe est importante: de sorte qu'un des plus grands sujets d' etonnement pour nos neveux, sera sans doute d' apprendre que chez des peuples qui passoient pour eclaires, et qui l' etoit reelement a beaucoup d'egards, elle n'entra pour rien dans les systemes les plus savans et dans les etablissemens les plus vantes d' education." 2. 1. O. de La Mettrie, Man a Machine, G. C. Bussey, trans. (Chicago, Open Court, 1912) p. 88. 3. G . W. Leibniz, Elements ofJurisprndence, in Leibniz: Philosophical Papers and Letters, 2nd ed. L.E. Loemker, trans. and ed. (Dordrecht: Reidel, 1969) p. 132. 4. Descartes, Discourse on Method, in Philosophical Writings, 3 vols. 1. Cottingham, R. Stoothoff, D. Murdoch, trans. and eds. (Cambridge: Cambridge University Press, 1985-91) vol. I, p. 151 (AT VI 78). 5. R. Descartes, Principles, in Philosophical Wn'tings, Vol. I, p. 186; (AT IX-2: 1415). 6. Compare R. Descartes, .Description of the Human Body, in Philosophical Writings, Vol. I, p. 314 (AT XI: 224). 7. Noteworthy in this respect are Amelie Rorty, "Descartes on Thinking with the Body," in The Cambridge Companion to Descartes, John Cottingham, ed. (Cambridge: Cambridge University Press, 1992) pp. 371-392; Gary Hatfield, "Descartes's Physiology and its Relation to his Psychology." in the same volume, pp. 335-70. 8. Descartes, Replies to Objections II, in Philosophical Writings, vol. IT, p. 96 (AT VII: 134). 9. Descartes, Replies to Objections V, in Philosophical Writings, vol. II, p. 245 (AT VII: 354). 10. Despite Leibniz' s sometimes critical remarks on Descartes's method. On Tschimhaus and Wolff, and a host of later users of the term "medicina mentis" see H. W. Arndt's introduction to Christian Wolff. Vernunftige Gedanken von den Kraften des Menschlichen Verstands, (1713) repro (Hildesheim: DIms, 1965) pp. 10-11, p. 53. 11. See, on the distinction, Ian Hacking, "Leibniz and Descartes: Proof and Eternal Truths," in Descartes: Philosophy, Mathematics and Physics, ed. Stephen Gaukroger, (Brighton: Sussex, Harvester, 1980) pp. 169-80. 12. This is not "foundationalism" in its usual sense: see the remarks at Pn'nciples III, .47 and IV, 204 in Philosophical Writings, vol. I, pp. 257,289 (AT IX-2: 102,327). 13. For a summary of the evidence, see J Oh11 Cottingham, "Cartesian Dualism: Theology, Metaphysics, Science," in The Cambridge Companion to Descartes, pp. 236257. 14. Descartes, Treatise of Man, in Philosophical Writings, vol. I, p. 108 (AT XI: 202). 15. Descartes, Discourse on Method, in Philosophical Writings, vol. I, p. 143 (AT VI: 62). 16. This is the conclusion of Stephen Gaukroger in Descartes: An Intellectual Biography, (Cambridge: Cambridge University Press, 1995) p. 387; more controversially Gaukroger ascribes this change to Descartes's renewed reflections on substantial union. 17. See Gaukroger on Mersenne's anti-naturalism and its influence on Descartes in Descartes: An Intellectual Biography, (Cambridge: Cambridge University Press, 1995) pp. 148-50.

Between Medicina Mentis and Medical Materialism

267

18. John Toland, Letters to Serena, Facs. (Stuttgart-Bad cannstatt, Frommann, 1964)p.159. 19. John Yolton, Thinking Matter: Materialism in Eighteenth-Century Britain, (Minneapolis, University of Minnesota, 1983). 20. Toland, Letters, p. 139. 21. Elie Luzac, Man More than a Machine, (London, 1750) (orig. L 'homme plus que machine, London, 1748) pp. 1-2. 22. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and 1. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 195. 23. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 176. 24. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 226 25. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 533 26. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 546. Cabanis is here summarizing Destutt de Tracy. 27. P.-J.-G. Cabanis, Rapports du physique et du moral, Oeuvres philosophiques, C. Lehec and J. Cazaneuve, eds. (Paris, Presses Universitaires, 1956) vol. XLIV-I, p. 297. 28. Harry Bracken, "Essence, Accidents and Race," Hermathena 16 (1973) pp. 8196. 29. J. G. Herder, Ideen zur Philosophie der Geschichte der Menschheit, (Dannstdt, Melzer, 1966) pp. 513-55. 30. Ernst Cassirer, Philosophy ofthe Enlightement, (Boston, Beacon, 1951) p. 55. 31. According to Eduard Zeller, "Auch Deutschland war von dieser Bewegung [Materialism, sensualism] bertihrt worden." However, "...der Materialismus eines Lamettrie und seiner Nachfolger wurde hier von den Philosophen einstimmig zurtickgewiesen, so manchen Anhanger er auch in der franzosisch gebildeten Welt zahlte." Geschichte der deutschen Philosophie seit Leibniz, (Muenchen: Oldenbourg, 1875) pp. 249-50. 32. 1. Kant, Critique of Pure Reason, trans. by N. K. Smith, (London: Macmillan Publishers, 1933) B 419-20. 33. Otto Schlapp, Kants Lehre von Genie und die Entstehung der Kritik der Urteilskraft (Gottingen, Vandenhoeck & Ruprecht 1901) p. 194. 34. Lewis White Beck, Early German Philosophy, (Cambridge, Massachusetts: Harvard University Press, 1969) p.487. 35. Manfred Kuehn has argued that materialism was present in Gennan thought, imported from Scottish physiology and common-sense philosophy. Johann Christian Lossius, whom Kuehn describes as "the most radical nlaterialist philosopher of the Enlightenment," in Gennany compared men to clocks and speculated about the role of fibres and nerve juices in cognition and sensation in his Physiche Ursachen des Wahren of 1774. See his Scottish Common-Sense Philosophy in Germany 1768-1800, (Toronto, McGill-Queens University Press, 1987) p. 86. The anthropological and sensationalistic tum represented by Lossius, Ernst Platner, Dietrich Tiedemann, Nikolaus Tetens, et al. was not consistently materialistic in the sense of the Frenchphilosophes. 36. Kuehn notes the importance of Priestly in his book, but does not provide specific infonnation about Priestly's reception in Gennany.

268

CATHERINE WILSON

37. When J. C. Pritchard in 1829, states that we know nothing at all of the essential nature of either mind or nlatter, "...the words mean nothing at all, but the unknown substance to which another set of known properties belongs. Algebraical symbols, such as x and y would be just as expressive as these words," he is echoing a Kantian theme. 1. C. Pritchard, A Review ofthe Doctrine ofa Vital Principle, (London, 1829) p. 39 38. 1. Kant, Critique of Pure Reason, trans. by N. K. Smith, (London: Macmillan Publishers, 1933) A 383. 39. 1. Kant, Critique of Pure Reason, trans. by N. K. Smith, (London: l\.1acmillan Publishers, 1933) B 421 40. See note 46 below. 41. As one is entitled to use speculative spiritualism as a skeptical weapon against materialism: see I. Kant, Critique of Pure Reason, trans. by N. K. Smith, (London: Macmillan Publishers, 1933) B 808-9. 42. G.W.F. Hegel, Introduction, Lectures on the History of Philosophy, 3 vals., E. S. Haldane, trans. (London: Routledge and Kegan Paul, 1892) I: pp. 4-5. 43. F. A. Lange, The History of Materialism, (1879-81) (repr. New York: Arno Press, 1974). 44. R. E. Butts, Kant and the Double Government Methodology, (Dordrecht, North Holland: Reidel, 1987) esp. ch. IX pp. 283ff. Kant acted as an anthropologist early and often. His pre-critical writings include the trite generalizing of the "Observations on the Sublinle and the Beautiful," (1763); the Dreams ofa Spirit Seer, (1765) which is about mental illness and its relation to spiritualism, and the satirical "Versuch tiber die Krankheiten des Kopfes" in which Kant calls for a cooperation between philosopher and doctor and concludes that cathartic medicines in high doses may purify and cleanse mental aberrations. Interesting is his treatment of the anatomist Sommerring's book on the brain ventricles and cerebro-spinal fluid, On the Organ of the Soul; see Alexander Ruger, "Brain-Water, the Ether and the Arts of Constructing Systems," Kant-Studien 86 (1995) pp. 26-40, p. 27. 45. Kant, Selected Pre-Critical Writings, G. B. Kerferd and D. E. Walford, eds. (Manchester: Manchester University Press, 1968) p. 120. 46. Lome Falkenstein argues that Kant's theory of spatial representation implies that "... cognition insofar as it involves the representation of sensory qualities in sensory arrays, simply cannot be a purely psychic process. It must be a physical process, perfonned by a body." ("Kant's Account of Sensation," Canadian Journal of Philosophy 20 (1990) pp. 63-88, p. 85.) "Cognition" is not used here in a sense equivalent to intellectual activity, which for Kant is discontinuous with sensory perception. Even so, Falkenstein notes that Kant's "way of ideas" is ultimately subject to this interference. 47. Kant, Religion within the Grounds of Reason Alone, Theodore M. Green and Hoyt H. Hudson, trans. (New York: Harper, 1960) p. 28.

Descartes's Rules and the Workings of the Mind ERIC PALMER

1. Introduction The Rules for the Direction of the Mind is generally taken to be an odd piece among Descartes's writings. Its difference from the later philosophical works is evident at first reading: evident, for example, in the strong focus upon imagination and the sensory system in perception~ in the discussion of the "simple natures" of the mind~ in the near absence of mention of God in its presentation; and in the absence of the classic Cartesian concern for metaphysics early in the treatment of a method for acquiring knowledge. One nnght discount the material as mere juvenalia, or plunder it for traces of a development towards Descartes's later positions; but to do so, I suggest, would be to miss or misread an interesting and singular work of early modern logic and of Descartes's early career. Descartes himself suggests that the Rules are merely intended as private notes written in anticipation of decrepitude, for personal rehearsal, "...so that when old age dims my memory I can readily recall [my universal mathematics] ...by consulting this book... "l The first-person form and apparent candidness of this pronouncement brings to mind the rhetoric of Descartes's published writing in the Meditations and Discourse, however, and it seems unlikely that Descartes would have inserted in the middle of a work the key he thought might be advantageous to his own recollection of its purpose, when old age has dimmed his memory! It is at least as reasonable to expect that Descartes's pronouncement is a stylistic touch, and that the work was left as it was-unfinished, in disarray, and probably untitled2-not because it was at no point intended for wider circulation, but because it was superseded at the time he apparently left off its composition for the last time, late in the 1620's. I briefly consider why Descartes stopped work on the Rules towards the end of my paper. My main concern is to accurately characterize the project represented in the Rules, especially in its relation to early-modern logic. The Rules certainly exhibits features of an art of reasoning, as the first rule and the title we affix to it suggest. It was used by the authors of the Port Royal Logic to improve their work,3 and clearly in service of such an art, Descartes writes: Within ourselves we are aware that, while it is the intellect alone that is capable of knowledge, it can be helped or hindered by three other faculties, viz. imagination, sense-perception, and memory. We must therefore look at these faculties in turn, to see in what respect each of them could be a hindrance, so that we may be on our guard,

270

ERIC PALMER and in what respect an asset, so that we may make full use of their resources. (AT X: 398-9) But the above rationale for including treatment of the faculties of imagination and memory-standard fare for logic texts 4-masks an important difference in Descartes's discussion. Descartes goes on to develop much more sophisticated accounts of mind and of brain than one might expect in a practical art of thinking: neither is represented, for example, in the Port-Royal Logic, and only cursory mention of sensation and the brain are provided in Gassendi's Institution of Logic. Descartes's Rules for the Direction of the Mind, then, is far more than the ordinary guidebook for practical logic and problem solving that the title suggests. What is the purpose of such discussion? Like Gassendi, Descartes focuses with great care upon a theory of ideas and the "necessary connections" among them; 5 but unlike Gassendi and others, Descartes includes the accounts of sensation and brain physiology to support his method for finding correct solutions to problems. I maintain that Descartes's approach exhibits a turn from one that exclusively treats of method towards one that we would say also includes epistemology, despite that neither Descartes nor his contemporaries use the latter term. In the first part of the Rules, Descartes supports some of his maxims for problem solving by providing an argument that connects his theory of ideas and its accompanying theory of knowledge to what we might call a cognitive science: a largely empirical theory of the mind's and of the brain's workings. In the second part of the work, these accounts inform specific practical guides for successful visual representation of mathematical and logical problems that should also be read as a further philosophical treatment of the role of the imagination in certifying some knowledge, including logic. Thus, logic is viewed as a process which involves the corporeal natures conjoined in the imagination. The product of Descartes's effort, consequently, is a problemsolving manual with a decided turn towards epistemology-and naturalistic epistemology, at that-which provides a theoretical foundation for his advice for improving problem-solving skills. 2. The Process of Composition of the Rules My first task is to briefly introduce the Rules, explain its complex structure, and point out the reasons why it should not be treated simply as a manual of method for the direction of the mind, as the title we affix to it suggests. Descartes envisioned the Rules as consisting of thirty-six guidelines divided evenly among three topics, with each, presumably, accompanied by a commentary. Such a plan is detailed in the twelfth rule. (AT X: 428f.) The first twelve rules were intended to consider the most general rules of method that lead to certain knowledge. They were to present a method for inquiry, "to direct the mind with a view to forming true and sound judgments about whatever comes

Descartes's Rules and the Workings afthe Mind 271 before it." (Rule One, AT X: 359) The second and third divisions were to concern applications of the material of the first section in something more representative of an ordered method for inquiry. The unfinished second division was intended to treat of problems that "can be understood perfectly, even though we do not know the solutions to them," and would focus particularly upon mathematics and geometry. (AT X: 429) The third division, apparently never commenced, was to treat problems "not perfectly understood," especially the "mixed" mathematics, and ,vas to present the methods necessary for reducing those problems to perfectly understood problems. The reduction was to be carried out by discerning the conditions required for defining the problem and, then, for determining "mutual dependence" (AT X: 429) between what is known and what is sought, among words and things, causes and effects, and parts and wholes, as is done in the investigations of riddles, magnetism, and plumbing systems that Descartes sketches briefly in the thirteenth rule. (AT X: 431-7) Such was the intended structure according to the twelfth rule, but the structure of the actual product is quite different. Descartes appears not to have finished the work: his notebooks provided drafts of only the first twenty-one rules and the first eighteen comnlentaries, and many of the latter survive in multiple drafts and show obvious gaps that suggest a need for revision. The second complication to the structure of the Rules arises as a consequence of their incomplete condition. The Rules was composed over a stretch of time perhaps as great as ten years,6 and Descartes's conception of the project changes significantly during its composition. The order of presentation of the rules also does not correlate with their order of composition: the first rule appears not to be of the earliest date, and sonle of the re-drafting within particular rules appears to be a re-thinking of earlier material. Given the unfinished and uneven character of the Rules, then, the level of theoretical unity I attribute to portions of the work might justly be called into doubt. Should the work be treated as a treatise, or as a scrapbook? Jean-Paul Weber's careful accounting of the history of the composition of the Rules might suggest the scrapbook, for the grand plan of a "universal mathematics" for solving problems in all sciences that Descartes discusses in the fragments of earliest composition (AT X: 374, line l6f.) appears to have been dropped for the bulk of the work and in his later writing. 7 Other passages exhibit more staying power: in the next stage of writing, displayed in many of the earlier nLles, Descartes concerns himself with maxims of a sort that punctuate his published writings. These include the methodological nlaxims we might expect in a set of rules for the direction of the mind, such as, "we ought to investigate what we can clearly and evidently intuit or deduce with certainty, and not what other people have thought," and "haphazard studies and obscure reflections blur the natural light and blind our intelligence." (AT X: 366, 371) The rules of

272

ERIC PALMER

method familiar from the Discourse (AT VI: 18-19) also have their close ancestors here. 8 My concern is particularly with another group of passages, however: sections of the commentaries on the eighth and twelfth to sixteenth rules that appear to form a natural unit. In those passages, new and sophisticated epistemological turns lead the work away from the narrow discussion of method that the fanliliar title of the work suggests. The first twelve rules, though flagged as a propadeutic to method in the first, fourth, and twelfth rules, also include discussions of physiology relevant to sensation, menl0ry, and imagination, and a theory of the contents of the mind: of the atoms, or "sinlple natures" of cognition. The rules of the second division, from the thirteenth forward, consider the imagination as a corporeal tablet and the simple natures in combinations upon the tablet. They also develop a theory of the role each of these items plays in establishing certain knowledge. How do these topics fit as rules for the direction of the mind? They are present in service of such rules: in later stages of composition, Descartes has made an important shift away from maxims, developing accounts of the workings of the nlind, the brain, and the sense-organs to theoretically ground his maxims in aid of clear reasoning within a developing epistemology, elaborated in its fullest development in the fourteenth rule. Because I focus upon accounts of the imagination and of the simple natures of thought that Descartes himself explicitly intertwines, then, a case can be made for the unity of these portions of the Rules. Weber's historical divisions also do not cut deeply here, for he suggests that the passages I consider are closely connected. 9 The task ahead, then, is to reconstruct Descartes's inlplicit naturalistic epistemology, constructed as a theoretical aid to the direction of the mind, from clues present in the eighth to the sixteenth rules. Though aspects of the account will survive in Descartes's later philosophy, the roles that the imagination and simple natures play will be supplanted, replaced by Descartes's more familiar technique for the direction of the mind, the metaphysical method of doubt. 3. Epistemology in the First Division of the Rules

Like other early modern thinkers, Descartes focuses upon a method for inquiry, rather than epistemology. However, the former flows recognizably into the latter in Descartes's writings as he asks and answers a question of a suitable to launch an epistemological inquky: "What is human knowledge what is its scope? ..the question ought to relate either to us, who have the capacity for knowledge, or to the actual things it is possible to know." (AT X: 398; c. f., AT X: 411) To 'answer the question, Descartes considers how knowledge relates "to us" through a study of the nlind and a study of the brain and sense-organs, the physical apparatus that also pertains to thinking. That area is considered below,

Descartes's Rules and the Workings ofthe Mind 273 but we consider first "the things it is possible to know," which are things "in so far as they are within the reach of the intellect;" that is, as the contents of experience. This consideration ushers in Descartes's theory of simple natures, which provides one of the grounds of his theory of knowledge.

3. i. Simple Natures The simple natures are the sole contents of the mind: in any act of perception or conlprehension, only simple natures are perceived. (AT X: 419) But one's perception of natures and the potential for these natures to provide knowledge are not always simply associated, for one can be ignorant of connections among the natures one perceives. Descartes's discussion of an example indicates his understanding of the gap between perception and knowledge: Indeed, it is often easier to attend at once to several mutually conjoined natures than to separate one of them from the others. For example, I can have knowledge of a triangle even though it has never occurred to me that this knowledge involves knowledge also of the angle, the line, the number three, shape, extension, etc., ... Perhaps there are many additional natures contained in the triangle which escape our notice, such as the size of the angles being equal to two right angles...(AT X: 422)

In this example and others, the familiar visual metaphors of clarity, distinctness, luminosity and recognition guide Descartes's account of knowledge to an even greater extent than they do in his Meditations. Descartes suggests here that knowledge consists in discerning the simple natures, and in subsequently sorting out the necessary connections among them. The simples themselves are discovered by intuition, which occurs "spontaneously" to the well-prepared mind. (AT X: 428) Such intuition can never present error, for simple natures have the quality of being "clearly and distinctly" perceived whenever they are perceived at all, and are necessarily true where truth is applicable to them: they themselves "never contain any falsity." (AT X: 418, 420) Ignorance, then, is the result of an ignorance regarding connection among simples. Such a view is evident in subsequent advice that Descartes gives for the direction of the mind, like the following: whenever we deduce something unknown from something already known, it does not follow that we are discovering some new kind of entity, but merely that we are extending our entire knowledge of the topic in question to the point where we perceive that the thing we are looking for participates in this way or that way in the nature of the things given in the statement of the problem. (AT X: 438)

Descartes appears to suggest here and in the previous passage that in many, and perhaps all cases of error, the mind has an obscured view of its own contents. This presents a shortcoming in the theory of simple natures as epistemology:

274 ERIC PALMER Descartes relies primarily on visual metaphors to explicate ignorance and understanding, and he does not employ the theory of simple natures to improve upon that account. Connections among simples are discovered through "deductions," and are grounded upon the intuition of "common notions" that are themselves also sinlples. (AT X: 424, 419) Other processes, such as inspiration through the light of divine grace and conjecture, are not the topic of method, which is restricted to preparing the mind for intuitions, and for discerning the combination of simples through deduction of necessary connections among them. (AT X: 424-5)10

3. ii. Physiology and the Practical Role of the Imagination Descartes's discussion is augmented by a theory of the physiology of perception, principally grounded in the assumption that our senses detect only geometric and numerical properties of objects outside of the body. (AT X: 41213) Beyond that notable revision to the account of intelligible species, Descartes presents a roughly Aristotelian account of sensation and recollection. 11 He suggests that "sense perception occurs in the same way in which wax takes on an impression from a seal," with all the ideas from the five senses meeting in the brain, in the "phantasy...a genuine part of the body...large enough to allow different parts of it to take on many different figures and, generally, to retain them for some time; in which case it is to be identified with what we call 'memory'." (AT X: 414) Descartes continues with a discussion of the relation of the faculties to the body, with particular regard for the processes of recall of menlory and of invention in the imagination. The transition to a discussion of faculties serves to connect Descartes's treatment of physiology with his comments on method, the central concern of the Rules. Descartes's efforts are cashed out in a recommendation concerning the direction of the mind: "If. ..the intellect proposes to examine something which can be referred to the body, the idea of that thing must be formed as distinctly as possible in the inlagination." Descartes suggests that an "abbreviated representation" of the thing is . to be formed in the imagination, so as to facilitate memory; and in later rules he instructs the reader on the art of forming such representations. (AT X: 416-7; AT X: 450-69) In these passages, Descartes connects an account of the workings of the mind to his discussion of method. 3. iii. Cognitive Science To this point, I have drawn a rough sketch of th~ epistemology that can be gleaned fronl the Rules, in Descartes's treatment of perception and the contents of the mind in their relation to method. Before further examining Descartes's account of the relation of method to the workings of the mind, I should pause to

Descartes's Rules and the Workings ofthe Mind 275 note the fundamental divergence between this approach to method and that which is presented in Descartes's later work. In the Rules, the method for inquiry is not begun from radical doubt, leading on to knowledge of one's own existence, then of an understanding of the character of certainty, and then of a proof of God's existence, as is familiar in Descartes's later work. Important anticipations of the more familiar method and its familiar conclusions may be found if sought~ for example, in the following: If someone sets himself the problem of investigating every truth for the knowledge of which human reason is adequate-and this, I think, is something everyone who earnestly strives after good sense should do once in his life-he will indeed discover by means of the Rules we have proposed that nothing can be known prior to the intellect, since knowledge of everything else depends on the intellect, and not vice versa.

The passage does not anticipate Descartes's method of doubt, but it does point to the priority of knowledge of oneself. However, Descartes continues his thought in a direction quite uncharacteristic of his later philosophy: Once he has surveyed everything that follows immediately upon knowledge of the pure intellect, among what remains he will enumerate whatever instruments of knowledge we possess in addition to the intellect; and there are only two of these, namely imagination and sense-perception. (AT X: 395)

This passage, like the question that began this section, "What is knowledge and what is its scope?" instead leads Descartes toward a very different sort of inquiry than that represented in the method of doubt and subsequent metaphysical inquiry. The study of the "pure intellect" leads Descartes to a theory of simple natures, rather than the cogito~ and then on to the "instruments of knowledge," the faculties, rather than a proof of God's existence. The method of the Rules clearly represents an approach founded in a theoretical account of cognition-a cognitive science-rather than the more familiar Cartesian metaphysical method of doubt. The first part of the Rules, then, contains material that we would today include under the headings of epistemology and cognitive science, in addition to a discussion of rules for the direction of the mind in inquiry. This additional material is intended to support such straightforward rules by providing them a place in a system of thought, presenting systematic connections with these other fields~ and this strategy again suggests the distance that Descartes maintains from his later rationalist and more strictly foundationalist approach to inquiry. Descartes makes a great deal in the second division of the Rules of his suggestion that a study of the other Aristotelian faculties and their corresponding organs, the sense-organs and brain, can be useful in an approach that bears fruit for the direction of the mind.

276 ERIC PALMER 4. Epistemology in the Fourteenth Rule: Visualization and Validation Although Descartes treats imaginative visualization as an aid to the direction of the mind whenever "the intellect proposes to examine something which can be referred to the body," there is much more underlying Descartes's claim than a practical maxim. (AT X: 416) Visualization is not merely an aid: it provides another key feature of Descartes's implicit epistemology as well. In the development of the fourteenth rule, Descartes provides a theoretical basis for the clarity that he maintains will result in many areas of problem solving as a consequence of his approach through visualization. (AT X: 438) That theoretical basis lies in the view that geometrical and mathenlatical truth, and other truths related to mathematics by common natures, are also validated, or certified as true, by the process of visualization in the imagination, or by drawing diagranls, which is the topic of the fifteenth rule. The imagination can be used to validate such tnlths because the imagination harbours "the real extension of a body considered in abstraction...," i.e., the imagination, as a structure of the body, itself employs the very corporeal natures that are the province of mathenlatical truth. (AT X: 441) In preparation for the argument in support of his more ambitious thesis concerning visualization, Descartes writes: it will be to the reader's advantage...to think of all knowledge whatever-save knowledge obtained through simple and pure intuition of a single, solitary thing-as resulting from a comparison between two or more things. (AT X: 440) The "single, solitary thing" mentioned is presently identified as a simple nature. Descartes has previously indicated that the intuition of natures for the prepared mind is beyond the scope of nlethod, and so he concludes at this point with the suggestion that "the chief part of human endeavour is simply to reduce these proportions to the point where an equality between what we are seeking and what we already know is clearly visible." Because proportions are involved, such comparison is invariably a matter of comparison of magnitudes. (AT X: 440, 441) The topics of proportion and magnitude link the simple natures into Descartes's full explanation of the role of visualization, and the way is prepared for a theoretical justification of visualization. Descartes explains: it will be very useful if we transfer what we understand to hold for magnitudes in general to that species of magnitude which is most readily and distinctly depicted in our imagination. But it follows from what we said in Rule Twelve that this species is the real extension of a body considered in abstraction from everything else about it save its having a shape. In that Rule we conceived of the imagination, along with the ideas existing in it, as being nothing but a real body with a real extension and shape. (AT X: 441)

Descartes's Rules and the Workings ofthe Mind 277 The above passage provides the guarantee that what applies to any magnitude will also hold of visualized magnitude. Since "nothing can be ascribed to nlagnitudes in general which cannot also be ascribed to any species of magnitude," the particularly useful species provided in the imagination may be called upon to validate conclusions concerning other species and magnitude in general. (AT X: 440) The next several pages of Descartes's argument contain an attempt to show that the intellect, if unaided by the imagination, may go astray in solving a wide variety of problems that pertain to material natures. A clear conception of extension invariably invokes actual corporeal things, and may require the use of the imagination: although someone may convince himself that it is not selfcontradictory for extension per se to exist all on its own even if everything extended in the universe were annihilated, he would not be employing a corporeal idea in conceiving this, but merely an incorrect judgement of the intellect alone. He will admit this himself if he carefully reflects on the image of extension which he tries to form in his imagination. He will realize that he does not perceive it in isolation from every subject, and that his imagination of it is quite different from his judgement about it. (AT X: 442-3)

Since the imagination employs a species of extension in its application, the intellect should not go astray while using imagination in relation to corporeality.12 The importance of actually em.ploying the inlagination and material natures in attempts to demonstrate and validate nlathematical and geometrical truth is the focus of detailed treatment in a nUlnber of drafts in the middle portion of the explication of the fourteenth rule. (AT X: 442-9) In each case, as in the passage above, Descartes takes great pains to distinguish the common or the over-subtle understanding "obscured by many vague and illconceived principles" from a clear conception of problems relating to corporeal nature. (AT X: 442) Descartes suggests that the misconceptions that pure understanding is prone to are the result of an incomplete "modal" grasp of the subject matter. Geometrical treatments of extension, surface, line, and point are incomplete concepts if understood apart from body: "a line, whose flowing motion [one] conceives as creating a surface, is a real body, whereas that which lacks breadth is simply a mode of body." (AT X: 446) Having justified the role of imaginative visualization in validating inquiry into material natures, Descartes closes the discussion of the fourteenth rule by returning to practical method, and a detailed treatment of strategies for visualization. Descartes finds three considerations to be of relevance, "viz. dimension, unity, and shape." (AT X: 447) "Dimension" refers to any nleasurable parameter, and unity "is the common nature which, we said above, all the things which we are comparing must participate in equally." (AT X: 447,449) The shape of a visualized figure depends upon its subject matter: open figures (points and connected lines) represent sets; closed geometrical figures illustrate

278 ERIC PALMER magnitudes. (AT X: 451) Descartes's mention of the "common nature" of unity that serves to make dimensions commensurate indicates once again that the account of the contents of the mind of the first division of the Rules plays a role in the later development of rules concerning the imagination.

4. i. The Case of Logic Descartes has mentioned that the intellect can be aided by the imaginative faculty, and that abbreviated representations of problems under investigation may be formed in the imagination. The universal applicability of such a method might be doubted, however: for why should visualization play a role in understanding mathematics and logical reasoning, for example? To support the point that visualization is an aid to all problem solving involving corporeality, and specifically to logical reasoning, Descartes refers us back to the theory of the simple natures, and to his account of knowledge as discerning connections among them. By "logic," a word Descartes does not use in the Rules, I intend to refer to Descartes's method of intuition and deduction discussed above, as distinguished from empirical investigation and syllogistic reasoning. Descartes distinguishes his method from syllogism by claiming that there is no need for a method for the latter: "when the operation is straightforward and simple we have no need of a technique to help us intuit the truth which the comparison yields; all we need is the light of nature." (AT X: 440) Like syllogistic, and unlike empirical study, the goal of deduction is to find "conjunction" among apparently unconnected terms, or a "comparison" or common participation among the entities that those terms represent. (AT X: 425, 440, 438) In his discussion of the magnet, for example, Descartes makes the distinction of method quite clear: .. .if the magnet contains some kind of entity the like of which our intellect has never before perceived, it is pointless to hope that we shall ever get to know it simply by reasoning~ in order to do that, we should need to be endowed with some new sense, or with a divine mind. But if we perceive very distinctly that combination of familiar entities or natures which produces the same effects which appear in the magnet, then we shall credit ourselyes with having achieved whatever it is possible for the human mind to attain in this matter. (AT X: 438)

In the twelfth rule, Descartes expands on what this passage serves only to suggest: that deduction of connections among simple natures should follow upon observation or research. The investigator "carefully gathers together all the available observations concerning the stone in question, then he tries to deduce from this what sort of mixture of simple natures is necessary for producing all the effects which the magnet is found to have." (AT X: 427) Beginning with empirical investigation, experimental study ultimately arrives at logic.

Descartes's Rules and the Workings ofthe Mind 279 Descartes's reference to natures in the above passage and others links his discussion concerning the contents of the mind to his thesis regarding imaginative visualization. Since unity and equality are simple natures common to both logic and measure, (AT X: 449) visualization of logical problems in terms of measurement of items present in the phantasy should be possible~ and since terms referring to corporeal natures ultimately require expression in some species of extended substance to ensure that the intellect considering them does not go astray, corporeal natures might even be necessary to allow for certainty in some logical applications. Among the diagrams presented to serve as models for appropriate visualization at the end of the discussion of the fourteenth rule is one that represents an aid to logical manipulation: a tree figure expressing the logical features of the relation of heredity. (AT X: 450) The relation between the phantasy and deductive manipulation expressed in the diagram provides the link between logic and the workings of the mind. In portions of the Rules of early composition, the only treatments of the imagination with reference to logic are disparaging. Descartes writes that "there is nothing more futile than devoting our energies to those superficial proofs which are discovered more through chance than method and which have more to do with our eyes and imagination than our intellect~ for the outcome of this is that, in a way, we get out of the habit of using our reason." (AT X: 375, c.f. AT X: 368) The universal mathematics of the first stage of writing is abandoned, however, and the maxims that followed it also give way to a return to a systematic treatment of a new kind, focused around an implicit epistemology and a study of the imagination. Descartes's shift is profound, as evidenced by his fourteenth rule: "The problem should be re-expressed in terms of the real extension of bodies and should be pictured in our imagination entirely by means of bare figures. Thus it will be perceived much more distinctly by our intellect." (X: 438) 5. Why the Plan of the Rules Was Dropped

The difficulty one faces in piecing together the evolving epistemology of the Rules is largely a consequence of its unfinished state. To explain why the project of the Rules was abandoned by Descartes, and to further illustrate the distinctions between the epistemologies of the Rules and Descartes's later work, we need only note one decisive turn in his thinking. 13 Both contemporary correspondence and the autobiographical notes of the Discourse suggest that an important change occurred in Descartes's thinking early in 1630, just around the time that Descartes appears to have left off the Rules for the last time. Descartes reports the shift in a letter to Mersenne, in which he suggests that God· has control of the creation and maintenance of a variety of important tDlths: 14

280

ERIC PAL:MER The mathematical truths which you call eternal have been laid down by God and depend on him entirely no less than the rest of his creatures. Indeed to say that these truths are independent of God is to talk of him as if he were Jupiter or Saturn and to subject him to the Styx and the Fates. (15 Apr. 1630, AT I: 145)

With this new twist, the epistenl0logy of the Rules is found to be much less solid, since the certainty of most knowledge, including mathematical and logical knowledge, is now tied to God's condition. In later works, Descartes will hold that knowledge of things other than the immediate contents of perception, and so certainly all knowledge that calls memory into play, ultimately rests on a proof of God's goodness and constancy. This shift is reflected in the Discourse as well, in which Descartes reports a central role for God in his epistemology only after "meditations...perhaps too metaphysical and uncommon for everyone's taste" that he commenced upon his return to Holland in 1629. (AT VI: 31ff.) Descartes does not create himself entirely anew at this mOlllent: important aspects of the framework of the Rules certainly do survive in his later works. Clear and distinct perception, and illumination by the natural light-what we might call "phenomenological" aspects of Descartes's earlier account-retain important and similar roles in Descartes's later account of cognition. The theory has been changed, however, and gep.erally, it appears that cognitive science exits as metaphysics enters Descartes's study of the mind. Descartes's implicit naturalistic epistemology is overturned with the introduction of a new role for God, and henceforth, his discussion of the brain and sense-organs will remain largely separated from his method, and from discussions that could be construed as pertaining primarily to the direction of the mind. 15 NOTES 1. Quotations in the text refer to the volume nUluber and pagination of R. Descartes, Oeuvres de Descartes, Charles Adam and Paul Tannery, eds. (Revised edition, Paris: Vrin/C. N. R. S., 1964-76). Abbreviated form appears in the text as "AT" followed by volume number, page number(s). The translation is that of Cottingham et. al., in R. Descartes, The Philosophical Writings of Descartes, 1. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny, translators & eds. (Cambridge, UK: Cambridge University Press, 1984-91) 3 vols. 2. Whatever title Descartes may have affixed to this work has not survived; the family of titles we find are extrapolations from other sources, perhaps taking their cues from the first rule. On the history of the title, which begins with Chanut's inventory of Descartes's writings, see Giovanni Crapulli in Descartes, Regulae ad Directionem Ingenii, G. Crapulli, ed., notes and appendices (The Hague: Martinus Nijhoff, 1965) pp. xxii-xxiii. 3. See the extract from the Port-Royal Logic reproduced at (AT X: 471-2). For an account of the purposes of the Port-Royal Logic, see the "Introduction" to A. Arnauld et

Descartes's Rules and the Workings ofthe Mind

281

al., The Art of Thinking: Port-Royal Logic, trans. by 1. Dickoff and P. James (Indianapolis: Bobbs Merrill, 1964). 4. C.f. Port-Royal Logic Part I, p. 32; P. Gassendi Institution ofLogic, 1658, trans. by H. Jones (Assen: Van Gorcum, 1981) Preface and Part 1. 5. R. Descartes, (AT X: 425); c.f., P. Gassendi, Institution of Logic, (1658) Part I, Canon V. 6. 1. -Po Weber, La Constitution du Texte des Regulae (Paris: Societe d'Edition d'Enseignement Superieur, 1964). 7. See 1. -Po Weber, La Constitution du Texte des Regulae (Paris: Societe d'Edition d'Enseignement Superieur, 1964) pp. 4ff A variant of Weber's argument, particularly focusing on the earlier stages of the composition of the Rules and the extent to which the ideal of the mathesis universalis is preserved in the later stages, is developed in 1. A. Schuster, "Descartes's Mathesis Universalis, 1619-28," Descartes: Philosophy, Mathematics, and Physics, ed., by S. Gaukroger (New Jersey: Barnes & Noble, 1980). See also Shea, The Magic of Numbers and Motion, (Canton, Mass.: Science History Publications, 1992). 8. E. g., "we should never assume to be true anything which is false," (AT X: 372); "...reduce complicated and obscure propositions step by step to simpler ones, and then, starting with the intuition of the simplest ones of all, try to ascend through the same steps to a knowledge of all the rest." (AT X: 379; c.f. AT X: 378-9) For a detailed accounting of the relation of the Rules to the four rules of the Discourse, see Gibson, "The Regulae of Descartes," Mind 7 (1898) pp. 332-63. 9. Weber dates these passages to 1623 and later. See 1. -Po Weber, La Constitution du Texte des Regulae (Paris: Societe d'Edition d'Enseignement Superieur, 1964) pp. 234ff., and 1. A. Schuster, "Descartes's Mathesis Universalis, 1619-28," Descartes: Philosophy, Mathematics, and Physics, ed., by S. Gaukroger (New Jersey: Barnes & Noble, 1980) pp. 55ff 10. For further details concerning the role of the simple natures in Descartes's philosophy, see Palmer (forthcoming) "The Limits of Cartesian Doubt," Studies in Early Modem Philosophy. See also 1.-L. Marion, Questions Cartesiennes, (Paris: Presses U'niversitaires de France, 1991). One chapter of Marion's is translated into English by John Cottingham and reprinted in The Cambridge Companion to Descartes, 1. Cottingham, ed. (Cambridge: Cambridge University Press, 1992). 11. A detailed comparison of Descartes's theory of sensation and its relation to contemporary Aristotelian accounts can be found in Wolff-Devine, Descartes on Seeing: Epistemology and Visual Perception, (Carbondale: Journal of the History of Philosophy Monograph series, 1993). 12. For a discussion of the role of the imagination as the validator of mathematics in the Rules, see 1. A. Schuster, "Descartes's Mathesis Universalis, 1619-28," Descartes: Philosophy, Mathematics, and Physics, ed., by S. Gaukroger (New Jersey: Barnes & Noble, 1980). Descartes completely reverses the argument presented in these passages in his later writings, as the chiliagon discussion in the Meditation Six suggests, and as he unequivocally states in the Fifth Replies: "although geometrical figures are wholly corporeal, this does not entail that the ideas by means of which we understand them should be thought of as corporeal (unless they fall under the imagination)." (AT VTI: 385)

282

ERIC PALMER

13. Reasons based in the development of Descartes's mathematics may be found in Gaukroger, "Descartes's Early Doctrine of Clear and Distinct Ideas," Journal of the History ofIdeas 53 (1992) pp. 585-602. 14. E. Brehier, "The creation of the eternal truths in Descartes's system," Descartes: A Collection of Critical Essays, ed. and trans. by W. Doney (New York, Anchor, 1967). 15. Concerning the carryover from the Rules to later epistemology, see Palmer (forthcoming). Wann thanks to the members of the Early Modern Logic Project and conferees for their hospitality and help, to Jim Sheridan and to the National Endowment for the Humanities, Jonathan Bennett and the members of the Syracuse l--l E. H. seminar for the Summer of 1995, at which a redrafting of this paper was completed.

Causal Inference, Associationism, and Skepticism in Part III of Book I of Burne's Treatise LOUIS E. LOEB

There is no ·doubt that Hume discovered the materials used to formulate the problem of induction familiar to twentieth-century philosophy. These materials are found at Treatise I.iii.6. Hume did not, however, conceive of his argument in Liii.6 as having any skeptical weight or force. In the Treatise, Hume's purpose in formulating the argument is to show that causal inference is the product of association.! Or so I will argue? In doing so, I will refer to Liii.6 as formulating "the problem of induction," using the phrase neutrally to refer to the arguments of that section, without implying that Burne intends them skeptically. I begin with a summary of some central features of Liii.6. Sixteen paragraphs comprise this section of the Treatise. At paragraphs one through three, Hume argues that the inference from an observed cause to an unobserved effect, or from an observed effect to an unobserved cause, requires experience of a constant conjunction. After summarizing this result at the beginning of paragraph four, Hume announces: "the next question is, Whether experience produces the idea by means of the understanding or imagination; whether we are determin'd by reason to make the transition, or by a certain association and relation of perceptions.,,3 Hunle contrasts the understanding or reason, on the one hand, and the imagination or an association and relation of perceptions, on the other. Hume proceeds to answer this "next question." Paragraphs four through seven, at pages 88 through 90, contain the heart of Hume's statement of the problem of induction. In paragraph four, Hume argues that to produce an idea of an unobserved object, reason would require a general principle about the uniformity of nature. (T89)4 At paragraph five, Hume argues that demonstrative argument cannot establish such a principle. (T89) At paragraphs six and seven, he argues that probability cannot establish such a principle without begging the question. (T89-90)5 The heart of the argument is now complete. At paragraphs eight through eleven, Hume considers, and rejects, a last stand in behalf of the view that such a principle can be established by demonstrative argument. (T9091)6 Hume's question is whether the idea of the unobserved effect, or of the unobserved cause, is produced by the understanding or reason, or by the imagination or association. He divides the possibility that reason produces the idea into two cases, and rejects both; neither demonstrative nor probable argument produces the idea of an unobserved object. In light of the question he has formulated, we should expect HUTIle to conclude that the idea is produced

284 LOUIS E. LOEB by association. So he does, at paragraph twelve, immediately following the last stand in defense of a role for demonstrative argument: Reason can never shew us the connexion of one object with another, tho' aided by experience, and the observation of their constant conjunction in all past instances. When the mind, therefore, passes from the idea or impression of one object to the idea or belief of another, it is not detennin'd by reason, but by certain principles, which associate together the ideas of these objects, and unite them in the imagination. Had ideas no more union in the fancy than objects seem to have to the understanding, we cou'd never draw any inference from causes to effects, nor repose belief in any matter of fact. The inference, therefore, depends solely on the union of ideas. (T92)

Hume's conclusion is that the idea of an unobserved cause or effect is produced by an association of ideas, by the imagination. This conclusion says nothing about the justificatory status of the inference. In the proceeding quotation, I omitted the first two sentences of paragraph twelve: We have already taken notice of certain relations, which make us pass from one obj ect to another, even tho' there be no reason to detennine us to that transition; and this we may establish for a general rule, that wherever the mind constantly and unifonnly makes a transition without any reason, it is influenc'd by these relations. Now this is exactly the present case. (T92)

Causal inferences are made for "no reason" and "without any reason." Does this not license the inference that causal inference is unreasonable?7 To see that it does not, we need to keep our focus on the terminology Hume introduces in paragraph four. He asks whether the idea of an unobserved object is produced by reason or by association. As we have seen, paragraph twelve provides the answer: the idea is produced by association rather than reason. Taken in context, Hunle's statements that causal inferences are made for "no reason" and "without any reason" mean that they are not due to a faculty of reason that is non-associative. 8 (I say more about this passage below.) Indeed, Hume writes in paragraph seven of Ijii.6 that causal inference is "just": "The only connexion or relation of objects, which can lead us beyond the immediate impressions of our memory and senses, is that of cause and effect; and that because 'tis the only one, on which we can found a just inference from one object to another." (T89)9 Hume has argued in the proceeding paragraph that no demonstrative argument to establish the uniformity of nature is available to reason. Hume' s clainl that causal inference is "just" occurs in the second of the two paragraphs in which he rejects the possibility that reason establishes the uniformity of nature by a probable argument. These paragraphs complete the heart of his statement of the problem of induction. If Hume conceives the problem of induction as having a skeptical outcome, it is difficult to understand why he would characterize causal inference as "just" with the

Causal Inference, Associationism, and Skepticism 285 argument well underway, indeed, in the fourth and final paragraph of the heart of the argument. James Beattie's reaction to Hume's position on causal inference is instructive. In his 1770 An Essay on the Nature and Immutability of Truth, Beattie famously treats Hume as an arch-skeptic. Beattie indeed accuses Hume of skepticism,1O and attacks Hume across a broad front: Hume's distinction between impressions and ideas, (Truth, 242-48) and his account of the nature of belief, with reference to vivacity; (93-99) Hume's view that the vulgar fail to distinguish perceptions and objects, (249-254) and his account of the external world; (53, 133,217,25") the bundle theory of personal identity; (82, 217, 25559) Hume's treatment of miracles (134-38) and belief in God; (115-120) and Hume's account of virtue and moral distinctions (407-419). On topics related to the metaphysics of causation, Beattie attacks Hume's argument that it is possible for something to begin to exist without a cause, (100-115) and also Hume's view that there is no power, energy, or efficacy in causation. (254-55, 294-308) Beattie takes a markedly different tack in his section devoted to "Probable or Experimental Reasoning." He writes: "In all our reasonings from the cause to the effect, we proceed on a...presumption of continuance" in "the course of nature." (Truth, 121) He agrees with Hume that inferences about unobserved causes and effects do not result from "abstract intuition," (12l) nor from "any proof or demonstration whatsoever;" (120) they are "by no means a logical deduction of a conclusion from premises." (122) Indeed, according to Beattie, "if reasoning be at all employed, it is only in order to give us a clear view of our past experience...When this view is obtained, reason is no longer necessary." (121-22; cf. 125) The inference to the unobserved object results, Beattie says, "immediately, and without the intervention of further argument." (122) The idea that the inference proceeds "immediately," (T93, 97n.; Tl02, 103) "without any new operation of the reason or imagination," (Tl02) is basic Humean doctrine. Beattie attributes these inferences to "an irresistible and instinctive impulse." (Truth, 122) He exhibits no temptation to see Hume's position on causal inference, or his own version of it, as having skeptical implications. To the contrary, conclusions about unobserved objects have "moral certainty;" (121) we believe them "with full assurance, or at least without doubt." (121) By Beattie's lights, Hume's view that causal inference is not founded in reason poses no skeptical problem; Beattie deems an instinctual foundation for causal inference quite adequate. Why should Hume's view that causal inference is due to association undermine belief in unobserved objects? A negative assessment of the justificatory status of beliefs about unobserved causes and effects does not follow, unless we attribute to Hume the view that only a non-associative faculty could justify causal inference. In contending that causal inference is an associative process, Hume is providing a naturalistic account of the

286 LOUIS E. LOEB psychological mechanism involved in causal inference. In principle, any suitably "naturalistic" theory of justification is a candidate for being pressed into service to support Hume's taking causal inference to be justified. Hume must have such a theory in view. The position of Norman Kemp Smith, first offered ninety years ago, attributes to Hume an instance of a view of this kind. According to Kemp Smith, Hume holds that psychologically irresistible beliefs are justified, or at least not to be condemned as unjustified. 11 On this interpretation of Hume, causal inference could be justified, insofar as it results in irresistible belief. On the Kemp Smith interpretation, however, the burden of Hume's epistemology is entirely negative: since the natural beliefs are irresistible or unavoidable, to say that we ought not hold them is pointless, or even false (if ought implies can). Insofar as this is Hume's point, Hume would not have provided any positive account of why we ought to accept such beliefs. For this reason, the Kemp Smith interpretation does not provide a satisfying account of how Hume could have taken causal inference to be justified. 12 In the past few years, a number of interpretations of Hume have emerged that are alternatives to Kemp Smith, yet also naturalistic. Annette Baier has suggested that Hume takes beliefs to be justified insofar as they result from mechanisms that survive critical reflection, or self-scrutiny. 13 Frederick Schmitt has suggested that Hume holds that beliefs are justified provided they result from belief-fornling mechanisms that are reliable. 14 My own view is that Hume holds that beliefs are justified provided they result from mechanisms that produce sufficiently stable or settled beliefs. I5 The availability of these alternative accounts of a naturalistic theory of justification in Hume reminds us that Hume's providing an associationist account of causal inference is in principle compatible with his taking causal inference to bejustified. 16 Hume's reference to causal inference as "just" at paragraph seven of Ijii.6 is neither a slip, nor a courtesy extended in passing to the view under attack. The claim that causal inference is justified is reiterated in Liii.I3 and Liv.4. Indeed, Hume's characterization of causal inference at page 89 coheres with his treatment of this topic throughout part iii and into part iv of Book I. The evidence will show that Hume's epistemological project has a constructive bent over large stretches of the Treatise, certainly to include the whole of Part iii of Book 1. 17 I review this evidence in some detail. The first place to look is to the historical background for Hume' s discussion in I.iii.6. I have in mind features of Locke's discussion of sensitive knowledge at Book IV, Chapter xi of An Essay concerning Human Understanding. Locke imposes a limitation on this third degree of knowledge: In fine, then, when our Senses do actually convey into our Understandings any Idea, we cannot but be satisfied, that there doth something at that time really exist without liS, which doth affect our Senses...But this Knowledge extends as far as the present Testimony of our Senses, employ'd about particular Objects, that do then affect

Causal Inference, Associationism, and Skepticism

287

them, and no farther. For if I saw such a Collection of simple Ideas, as is wont to be called Man, existing together one minute since, and am now alone, I cannot be certain, that the same Man exists now...And therefore, though it be highly probable, that Millions of Men do now exist, yet, whilst I am alone writing this, I have not that Certainty of it which, we strictly call Knowledge... 18

According to IV.xi.9, sensitive knowledge extends only to the "present testimony of our senses...and no farther." Two sections following, in IV.xLII, Locke liberalizes his account, to allovv sensitive knowledge mediated by memory: As when our Senses are actually employ'd about any Object, we do know that it does exist; so by our Memory we may be assured, that heretofore Things, that affected our Senses, have existed. And thus we have knowledge of the past Existence of several Things, whereof our Senses having informed us, our Memories still retain the Ideas; and of this we are past all doubt, so long as we remember well. But this Knowledge also reaches no farther than our Senses have formerly assured us. (Essay IV.xi.II, some emphases added and deleted)

According to Locke at IV.xi.II, if I perceive an object now I know that it exists now; if I remember perceiving it at noon yesterday I know that it existed at noon yesterday. Strictly speaking, however, I do not know that it existed prior to noon yesterday, that it existed in the interval from noon yesterday until now, or that it will exist at any time in the future. 19 Locke's limitations on sensitive knowledge renlain severe. I suggest that Hume's familiarity with this disappointing result led him to write, at section 21 of An Enquiry Concerning Human Understanding: It may...be a subject worthy of curiosity, to enquire what is the nature of that evidence which assures us of any real existence and matter of fact, beyond the present testimony of our senses, or the records of our memory. This part of philosophy, it is observable, has been little cultivated, either by the ancients or moderns. (E2I)

The problem is to determine how assurance or belief extends beyond the point where Locke says it leaves off, beyond memory of what has been perceived and "the present Testinl0ny of our Senses," a phrase that appears both at IV~ix.9 of Locke's Essay and at section 21 of Hume's First Enquiry. Hume states his answer in the first sentence of the following section: "All our reasonings concerning nlatter of fact seem to be founded on the relation of Cause and Effect. By means of that relation alone we can go beyond the evidence of our memory and senses." (E22) Hume is proposing a "causal theory" of assurance of those matters of fact that do not result from memory or present perception alone. One can infer an object's existence at times it has not been observed, as the cause (or the effect) of its existence at times it has been observed. 20 Locke set the problem that led Hume to formulate the causal theory of assurance of matters of fact that are not based on memory or present perception alone. 21

288 LOUIS E. LOEB Hume's causal theory offers a potential solution. 22 I suggest Hume read Essay IV.xi.9-11, saw the disconcerting gaps in Locke's discussion of sensitive knowledge, and proceeded to inquire how they might be repaired. 23 Though the relationship of Hume's problem to Locke's discussion of sensitive knowledge is especially evident in the First Enquiry, Hume was working with the same problem in the Treatise. Hume writes at Treatise, page 196: "My memory, indeed, informs me of the existence of many objects; but then this information extends not beyond their past existence, nor do either my senses or memory give any testimony to the continuance of their being." I-Iume "introduces the causal theory of assurance much earlier, in I.iii.2: 'Tis only causation, which produces such a connexion, as to give us assurance from the existence or action of one object, that 'twas follow'd or preceded by any other existence or action...(T73-74) [O]f those three relations, which depend not upon the mere ideas, the only one, that can be trac'd beyond our senses, and infonns us of existences and objects, which we do not see or feel, is causation. (T74)

Subsequent statements in I.iii.4 and I.iii.5 make it clear that, as in the First Enquiry, the relation of causation extends assurance beyond present observation and the memory of past observation: [T]he mind in its reasonings from causes or effects carries its view beyond those objects, which it sees or remembers...(T82) All our arguments concerning causes and effects consist both of an impression of the memory or senses, and of the idea of that existence, which produces the object of the impression, or is produc'd by it. (T84)

In these passages, and elsewhere in the Treatise, in sections as late as Liv.2, (T87, 89, 107-8, 193, 198, 212) Hume claims that all assurance of matters of fact that is not based upon present observation or upon lllemory of what has been observed is based on causal inference, inference about objects that are causally related. 24 This is the same causal theory of assurance as in the First Enquiry, and it responds to the same problem in Locke. It is significant that Hume's initial response to Locke's problem is constructive. Hume offers the causal theory of assurance as an account of what does assure us of the existence of objects that we have not perceived. Considering what I-Iume does not say throws the constructive character of his response into better relief. Hume does not challenge Locke's claim that the senses and memory do supply knowledge, however fragmentary, of the existence of external objects. Hume does not, for example, challenge Locke's claim that "the present Testimony of our Senses" generates knowledge of the present existence of external objects. Though such an attack would not have been difficult to mount,25 challenging Locke's confidence in sense-perception and memory is not what Hume is about in Part iii of Book 1. Hume' s interest in

Causal Inference, Associationism, and Skepticism 289 Part iii is in ascertaining what does assure us of the existence of objects that we have not perceived. As we have seen, Hume introduces the causal theory of assurance of matters of fact as early as I.iii.2. In I.iii.4, "Of the component parts of our reasonings concerning cause and effect," Hume identifies a role for perception and memory, which are further discussed in I.iii.5, "Of the impressions of the senses and memory." The problem of induction arises when Hume asks in I.iii.6 how we make the transition to the idea of the unobserved cause or effect. In Part iii, Hume's interest is in how we advance beyond present perception and memory of what has been perceived. I.iii.6 provides the answer-assurance about matters of fact that have not been observed is derived from association rather than reason. If Hume intends the problem of induction to have skeptical force, we should expect to have seen the last of the claim at page 89 that causal inference is justified. This is not what we find. At Treatise, Ijii.9, pages 107-8, there is an extended restatement of the theory that all belief in matters of fact is based on perception, memory, and causal inference. Hume describes two systems of beliefs or "realities": Of these impressions or ideas of the memory we fonn a kind of system, comprehending whatever we remember to have been present. .. ; and every particular of that system join'd, to the present impressions, we are pleas' d to call a reality. But the mind stops not here. For finding, that with this system of perceptions, there is another connected by custom, or if you will, by the relation of cause or effect, it proceeds to the consideration of their ideas... [1]t fonns them into a new system, which it likewise dignifies with the title of realities. The first of these systems is the object of the memory and senses; the second of the judgment. 'Tis this latter principle which peoples the world, and brings us acquainted with such existences, as by their removal in time and place, lie beyond the reach of the senses and memory. (T107-8)

The first system of beliefs or realities is based on the senses and memory alone; these are beliefs based solely on the resources of Locke's sensitive knowledge. The second system supplenlents the first, and is based on custom, or the relation of cause and effect. These are the additional beliefs based on causal inference. Hume writes approvingly of both systems of beliefs, and attributes the second system, beliefs based on causal inference, to "the judgment." Hume seems satisfied that he has repaired the gaps in Locke's analysis of sensitive knowledge. Beyond the claim at page 89 that causal inference is "just," and the attribution of causal inference to "the judgment" at page 108, there is massive evidence that Hunle regards causal inference as reasonable or justified in Part iii, and even in portions of Part iv, of Book I. In I.iii.13, Hume discusses the effect of "a long chain of connected arguments." (T144) The individual arguments, where the "inference is drawn immediately..., without any

290 LOUIS E. LOEB intermediate cause or effect" (T144) may be "just and conclusive in each part." (T144) In I.iii.15, titled "Rules by which to judge of cause and effects," (T173) Hume writes that "it may be proper to fix some general rules, by which we may know," (TI73) when objects cause one another. He proceeds to provide eight such rules, (TI73-75) said to constitute "all the Logic I think proper to employ in my reasoning." (TI75) Hume's language-"rules" that are "proper" for ascertaining causes and effects, and which constitute a "Logic"-is hardly what we would expect had ·he concluded, fewer than ninety pages earlier, that no inference based on the relation of cause and effect is justified?6 In I.iv.4, "Of the modern philosophy," Hume considers an example of "the customary transition from causes to effects, and from effects to causes": (T225) "One who concludes somebody to be near him, when he hears an articulate voice in the dark, reasons justly and naturally." (T225) The just reasoning here is causal inference; the presence of an unobselVed person is inferred as the cause of the observed voice. The claim that causal inference is justified appears in Liii.6 at page 89, recurs at I.iii.14 at page 144, and remains in place mid-way through Part iv, at page 225. Similarly, Hume's approval of causal inference is manifest in his discussions of the two systems of realities or beliefs at pages 107-8 of I.iii.9, and of rules by which to judge causes and effects at pages 173-75 of Liii. 15. All this is inexplicable on an interpretation that takes Hume to intend a skeptical result in I.iii.6. Also inexplicable, on a skeptical interpretation of I.iii.6, is Hume's commitment to the existence of degrees of probabilistic evidence in I.iii.11-13. The epistemological force of the skeptical problem of induction is to show that there can be no justification whatsoever for believing the conclusion of any non-demonstrative argument, and hence that all non-demonstrative arguments have the same evidential force-none at all. In I.iii.II-13, however, Hume draws a number of epistemic distinctions or discriminations within nondemonstrative inference. Some preliminaries will help to explain the import of these sections. I.iii.ll, 12, and 13, treat "Of the probability of chances," "Of the probability of causes," and "Of unphilosophical probability," respectively. In order to see what Hume has in mind by "the probability of causes," we need to consider his contrast between "inlperfect" (TI31) and "perfect" (cf. T134, 135) experience?? Perfect experience consists in the frequent observation of a constant conjunction. Imperfect experience is experience that is not perfect. There are two cases: first, probability based on a constant conjunction which is infrequently observed; (TI30-31) and second, probability based on "contrary causes," a conjunction that is frequently observed but not constant. (T131-141) We arrive at beliefs about unobserved causes and effects on the basis of imperfect experience, of the two kinds described, as well as perfect experience. What is Hume's epistemic assessment of the various inferenc_~~_~~~~~~!!~!!7 -- --------- -

Causa/Inference, Associationism, and Skepticism 291 At Liii.ll, Hume divides non-demonstrative or probable reasoning into "proofs" and "probabilities." Proofs are "arguments from causation [that] exceed probability, and may be receiv'd as a superior kind of evidence." (TI24) Hume states that the distinction between proofs and probabilities selVes not only "to preselVe the common signification of words," but also to "mark the several degrees of evidence." (TI24) Hume shares the "common" (T124) view that some arguments from causation-those that are based on a perfect experience-constitute a superior degree of evidence. He designates these arguments "proofs," in distinction to "probabilities." (Rume's examples, at T124, are that "the sun will rise to-morrow, or that all men nllist dye,,,28) Hume proceeds to examine probability, non-demonstrative arguments that fall short of proofs, at Liii.II-12?9 I.iii.l1 deals with the probability of chances. Liii.12 deals with the probability of causes, that is, causal inferences based on imperfect experience, whether infrequent obselVation or contrary causes (obselVation of conjunctions that are not constant). In the final paragraph of Liii.12, Hume discusses "these two species of probability, which are deriv'd from an imperfect experience and from contrary causes," (T142) together with a third species of probability, that arises from analogy. In the first paragraph of the next section, I.iii.13, Hume writes: "All these kinds of probability are receiv'd by philosophers and allow'd to be reasonable foundations of belief and opinion." (T143) Even causal inference based on infrequent experience or contrary causes is justified; causal inference based on a frequent experience of a constant conjunction must be at least as justified as that based on imperfect experience. It is worth noting that Beattie follows Hume in allowing that there are "degrees of probability," (Truth, 125) "degree[s] of conviction in reasoning from causes to effects," (121) with the "inferior degrees" (121) based on "experience of the past [that] has not been uniform nor extensive." (122) Hume's position is confirmed in the penultimate paragraph of Liii.13. (T153-54) Hume is explaining the notion of probability, or judgment that falls short of demonstrative knowledge. He distinguishes kinds of belief with respect to "degree of evidence." Memory provides the highest degree of evidence. The relation of cause and effect, based on a perfect experience, constitutes the strongest degree of evidence, apart from that provided by memory. Causal inference based on imperfect experience or contrary causes, together with inference based on analogy or resemblance, fall next in line. For all Hume claims in Part iii of Book I, causal inference based on perfect experience, probability based on imperfect experience or contrary causes, and analogy, provide reasonable or justified belief. 30 In sum, far from claiming that there is a skeptical problem of induction that obliterates distinctions between good and bad non-demonstrative arguments, Hume is committed in Part iii to a number of such distinctions. 31 By documenting Hume's endorsements of causal inference, and his commitment to degrees of probabilistic evidence, I have tried to undermine the

292 LOUIS E. LOEB conviction that Hume intends the problem of induction to have the conclusion that causal inference is not justified. To find the intended conclusion, we must look elsewhere, to the claim that causal inference is due to association. One can hardly overestimate the inlportance to Hume of his associationism. The subtitle of the Treatise is "An Attempt to introduce the experimental Method of Reasoning into Moral Subjects." (Txi) Hume writes in the "Introduction" that he is proposing a "science of man," (Txx) a "science of human nature," (Txxixxii) and at I.iv.4 that the imagination must be "guided by some universal principles which render it, in some measure, uniform with itself in all times and places." (TIO) These principles of connexion, union, or association constitute "a kind of Attraction, which in the mental world will be found to have as extraordinary effects as in the natural." (TI2-13; cf. T289) The idea that Hunle sees himself as the Newton of the moral sciences is familiar enough, but it needs to be taken seriously. In the final section of Part i of Book I, (T20-24) Hume is already displaying what he takes to be the resources of associationism in providing an associationist underpining for a Berkelian position on abstract ideas. Associationist mechanisms are implicated in much of the large-scale structure of the Treatise. Book I, Part iii is principally devoted to the association of ideas by the relation of cause and effect; Book I, Part iv, to the association of ideas by the propensity, itselfbased on association by resemblance, to nlistake related for identical objects32 ; Book II, to the double association of ideas and impressions; and Book III, to sympathy, which converts a lively idea into an impression. Hume writes in the "Introduction" that his account of "the principles of human nature" is "built on a foundation almost entirely new;" (Txx) that foundation is an associationist psychology. It is in the final paragraph of An Abstract of a Treatise of Human Nature that Hume writes: "[I]f any thing can intitle the author to so glorious a name as that of an inventor, 'tis the use he makes of the principle of the association of ideas, which enters into most of his philosophy." (A31) Hume's main project is to confirnl associationisnl as a general theory of the operation of the human mind. 33 Hume's associationist ambitions are central to his conception of the problem of induction. In Part i of Book I, Hume lays the groundwork for his contrast at Liii.6 between reason and association. At I.i.3, the imagination is identified with the faculty that forms ideas fainter than perceptions of memory. (T8-9) At I.i.5, "Of the connexion or association of ideas," Hume introduces a nUlIlber of associative principles to explain how perceptions fainter than those of nlemory are formed. 34 In doing so, he treats the association of ideas by resemblance, contiguity, and cause and effect as the principles of the imagination. (cf. TIO13) In Part i of Book I, the imagination is explicitly an associative faculty. This sets up the contrast between the understanding or reason, and the imagination, that is operative at I.iii.6, and leads to the conclusion that causal inference is .due to association. 35

Causal Inference, Associationism, and Skepticism 293 The contrast between reason and association, much as the apparatus of the theory of ideas in Part I-would have been familiar to readers of Locke. In "Of the Association of Ideas," Locke contrasts ideas that "have a natural Correspondence and Connexion with one another" (Essay II.xxxiii.5) with ideas that are "not ally'd by nature," (II.xxxiii.6) but by association. The connection of ideas that is "Natural, depend upon our original Constitution, and are born with us." (II.xxxiii.7) It is "the Office and Excellency of our Reason to trace" (II.xxxiii. 5) the ideas that have a natural connection. The association of ideas is due to "Chance or Custom" (II.xxxiii.5) or to an "accidental Connexion" (II.xxxiii.7), resulting in an "undue" (II.xxxiii.8) or "wrong" (II.xxxiii.9) connection of ideas. Locke introduces association to account for opinions and reasoning that are "Extravagant," (II.xxxiii.1) in "opposition to Reason," (II.xxxiii.4~ cf. II.xxxiii. 3) and characterized by "Unreasonableness." (II.xxxiii.1, 3) In section 41 Of the Conduct of the Understanding, Locke seek to "inquire into the remedies that ought to be applied to" association. 36 As Laird points out, Locke's negative appraisal of beliefs based on association finds its way into Isaac Watts's Logic, published in 1724. 37 In light of Locke's texts, it is a familiar observation that Locke confines the province of associationism to belief that is unjustified. 38 Considered in relation to Locke, Hume seeks to expand the territory in which associationism is recognized as holding sway. In particular, Hume seeks to extend the domain of associationist mechanisms to justified as well as to unjustified belief. As we have seen, Hume both attributes causal inference to association at l.iii.6, and regards causal inference as justified throughout Part iii, and well into Part iv, of Book 1. Whereas Locke sees the role for associationism as linlited to the explanation of ways of thinking that require "remedies," Hume holds that many beliefs that arise associationistical1y are epistemically respectable. This is the crucial difference between Hume's associationism and that of Locke. 39 Hume has this in view in the Abstract when he lays claim to being an "inventor" with reference to "the use he makes of the principle of the association of ideas." Hunle's view of his own contribution seems correct in comparison with Locke's use of associationism. Hobbes, on the other hand, clearly contemplated association as having a role in causal inferences that are perfectly justified. This emerges in his discussion of prudence or foresight in Leviathan: "he that foresees what will become of a criminal re-cons what he has seen follow on the like crime before, having this order of thoughts: the crime, the officer, the prison, the judge, and the gallows. ,,40 Elsewhere, the same example is invoked to illustrate the principle that "after a man hath been accustomed to see like antecedents followed by like consequents, wheresoever he seeth the like come to pass to any thing he had seen before, he looks there should follow it the same that followed then.,,41 In his use of associationism, Hume is returning to a path set by Hobbes, a path from which Locke strayed. 42

294 LOUIS E. LOEB Hume's interest in applying associationism to justified as well as unjustified belief is reflected in shifts in terminology that occur in Part iii of Book 1. As we have seen, Hume introduces the Lockian contrast between reason and association, in paragraph four of IjiL6. In paragraph twelve, causal inference is made for "no reason" and "without any reason," in the Lockian sense that it is not due to a non-associative faculty. Subsequent to this passage at T92, Hume proceeds to coopt the terminology of "the understanding" and "reason.,,43 Just one page later, in the fifteenth and penultimate paragraph of Ijii.6, Hume is writing of "reasonings" (T93) based on the relation of cause and effect. Similarly, Hume writes in the final paragraph of Ijii.6 "that we are able to reason upon" (T94, emphasis added) the relation of causation. All this is within Ijii.6, after the formulation of the problem of induction. Similarly, Hume writes in Liii.7 of "reasonings from causes or effects" (T94) and "reasonings from causation." (T95) In a footnote in the same section, he writes: "We infer a cause immediately from its effect~ and this inference is not only a true species of reasoning, but the strongest of all others." (T97)44 Later in the footnote, Hume writes that causal inference is one of a number of "acts of the understanding." (T97) Whereas for Locke reason is the faculty that distinguishes humans from other animals, (Essay IV.xvii.!) Hume can write "Of the reason of animals" at I.iiLxvi because animals engage in causal inference utilizing associationist mechanisms. Hume's willingness to see causal inference as a species of "reason" is also prominent in Part iv. In Liv.2, "Of skepticism with regard to the senses," Hume asks whether the belief in body is produced by the senses, reason, or the imagination. (T188) Hume rules out reason, on the ground that the belief in body cannot result from an argument based on the relation of cause and effect. (T193)45 Similarly, in the last paragraph of I.iv.4, "reason" is explicitly identified with "those conclusions we form from cause and effect." (T231) If the understanding or reason is non-associative, and if causal inference is associative, how can causal inference be a species of reasoning or an act of the understanding? There is no inconsistency. If, following Locke, "the understanding" or "reason" is a non-associative faculty, ordinary causal inference is not due to reason. Our response, Hume thinks, should be to adopt his alternative conception of the understanding or reason. Rather than regard it as non-associative, we should regard it as associative, and hence an operation of the imagination. One of Hume's objectives is to reduce prima 'facie distinct cognitive faculties to a common associationist foundation. Any presumed nonassociative faculty is not the source of inference about causation. In sum, the conception of the understanding or reason as a non-associative faculty that explicitly governs Hume's statement of the problem of induction in Ijii.6 is replaced by a conception of reasoning that embraces an associative process, causal inference itself. There is no hint of skepticism here. Such terms as "reason" and the "understanding" look like honorifics.

Causal Inference, Associationism, and Skepticism 295 This not to say that any associative process is an instance of reason, construed associationistically. It is at pages 107-8 of l.iii.9 that Hume writes approvingly of the two systems of beliefs or "realities," the first based on the senses and memory alone, the second based on custom or the relation of cause and effect. He attributes beliefs based on causal inference to "the judgment." (TI08) He contrasts the beliefs that comprise the two systems of realties with those that "are merely the offspring of the imagination," or "mere fictions of the imagination." (TI08) Hume seeks to draw a distinction within the imagination between two classes of belief: those, such as beliefs based on the relation of cause and effect, which deserve epistemic approval, and those that are to be condemned as mere offsprings of the imagination. These discriminations call for a distinction drawn within the imagination itself. Hume writes in a footnote in I.iii.9: [A]s our assent to all probable reasoning is founded on the vivacity of ideas, it resembles many of those whimsies and prejudices which are rej ected under the opprobrious character of being the offspring of the imagination. By this expression it appears that the word, imagination, is commonly us'd in two different senses...When I oppose the imagination to the memory, I mean the faculty, by which we form our fainter ideas. When I oppose it to reason, I mean the same faculty, excluding only our demonstrative and probable reasonings. (Tl17-18n.)

Hume identifies the imagination in a wide sense with the faculty by which we form ideas fainter than perceptions of memory. This is the faculty of association-as introduced in Part i of Book I-,based on resemblance, contiguity, and cause and effect. It is the faculty with which Hume contrasts the understanding or reason for the purposes of his formulation of the problem of induction. In the footnote at I.iii.9, Hume identifies "demonstrative and probable reasonings" with "reason." In a footnote at II.ii.7, he identifies "demonstrative and probable reasonings" with the "understanding." (T37In.) In both notes, he identifies the imagination in the narrow sense with the inlagination in the wide sense, exclusive of reason or the understanding. The inlagination in the wide sense, therefore, is comprised of "the understanding" or "reason"-that is, demonstrative and probable reasonings-,and the imagination in the narrow sense. Since the imagination in the wide sense is an associative faculty, "the understanding" or "reason," as now defined, is itself an associative faculty, a faculty within the imagination in the wide sense. 46 The footnote at I.iii.9 draws a distinction between two components of the faculty of association-reason or the understanding, and the inlagination in the narrow sense. The discussions of the two systems of realities and the two senses of "imagination" are located ten pages apart within a single section of the Treatise. When Hume writes at page 108 that beliefs not based on perception,memory, or the relation of cause and effect "are merely the offspring of the imagination," he has in mind the imagination in the narrow sense-

296 LOUIS E. LOEB otherwise there would be no room for a contrast with "the judgment." Beliefs based on causal inference, the ones which Hume approves and attributes to "the judgment," are due to "the understanding" or "reason," as characterized at this stage of the Treatise. The terminological preferences that are implicit beginning at T93, when Hume writes of "reasonings" from the relation of cause and effect, are now explicit. For Locke, reasonable belief is non-associative, and unreasonable belief associative. For Hume, the distinction between reasonable and unreasonable belief is drawn within the faculty of association. The footnote in I.iii.9 makes official Hume's attempt to coopt such terms as "understanding" and "reason" within his associationist psychology. From this point on, Hume's official terminological commitments are relatively stable. Indeed, the Part iii distinction between two kinds of associative faculty-reason or the understanding and the imagination in the narrow sense-persists into Part iv. In the final paragraph of I.iv.iii, "Of the ancient philosophy," Runle writes that the Peripatetics are "guided by every trivial propensity of the imagination." l.iv.4 opens with an objection: But here it may be objected, that the imagination, according to my own concession, being the ultimate judge of all systems of philosophy, I am unjust in blaming the antient philosophers for makeing use of that faculty, and allowing themselves to be entirely guided by it in their reasonings. In order to justify myself, I must distinguish in the imagination betwixt the principles which are permanent, irresistable, and universat such as the customary transition from causes to effects, and from effects to causes: And the principles, which are changeable, weak, and irregular; such as those I have just now taken notice of. (T225)

It is in the continuation of this paragraph that Hume maintains that one who hears an articulate voice in the dark "reasons justly" (T225) in inferring, by the customary transition from cause to effect, that someone is present. Once again, Hume seeks to draw a distinction within the inlagination between beliefs that are justified and those that are not. The distinction between principles that are "permanent, irresistable, and universal" and those that are "changeable, weak, and irregular~' (T225) is drawn within "the imagination." (T225) I take it that the permanent, irresistible, and universal principles constitute the understanding or reason, and that the changeable, weak, and irregular principles constitute the imagination in the narrow sense. At T267, for example, "the understanding" is identified with "the general and more establish'd properties of the imagination." (T267; cf. T149, 150) My approach to the interpretation of I.iii.6 provides an answer to an interesting puzzle. We have seen that, in Part iii, Hume's response to Locke's account of sensitive knowledge is not to call into question beliefs based on the senses or memory. The puzzle arises for those who think that Runle intended the problem of induction skeptically. If so, how could he have overlooked the analogous skeptical problem about memory? To see the force of the puzzle, it

Causal Inference, Associationism, and Skepticism 297 is important to recognize that Hume nowhere expresses systematic or global doubts about memory. Hume repeatedly conveys a general confidence in memory: 'Tis evident, that the memory preserves the original form, in which its objects were presented, and that where-ever we depart from it in recollecting any thing, it proceeds from some defect or imperfection in that faculty. (T9) [The associative relations] are therefore the principles of union or cohesion among our simple ideas, and in the imagination supply the place of that inseparable connexion, by which they are united in our memory. (TI2) 'Tis impossible for us to carry on our inferences in infinitum; and the only thing, that can stop them, is an impression of the memory or senses, beyond which there is no room for doubt or enquiry. (T83) [1]t be a peculiar property of the memory to preserve the original order and position of its ideas, while the imagination transposes and changes them, as it pleases. (T85)

At the same time, the skeptical problem about memory is obvious. Beliefs based on memory would seem to require a general principle to the effect that nlemory beliefs are usually preceded by the states of affairs they report. There is no demonstrative argument to establish such a principle. 47 Any probable argument to establish such a principle would rely on memory, and hence beg the question. How could Hume have overlooked this? My answer is as follows. Hume's problenl of induction was not intended to generate skepticism about causal inference. Hume's objective was to show that causal inference is due to association. HUTIle saw no need for an analogous argument in the case of memory~ he took the picture of memory as an associative process to be unproblematic. Hume does not pursue this idea in any detail, but it is implicit in his view that decay in the vivacity of a memory belief explains the diminishing degree of assurance or confidence in a memory belief over time. (cf. T85, 144, 154, 426) Here Hume could rely on predecessors. Hobbes, writing in Leviathan ninety years before the Treatise, also appeals to the image of memory as "fading."48 Indeed, Hobbes had defined "imagination" as "decaying sense." (L I,ii)49 He maintained "that imagination and memory are but one thing, which for diverse considerations have diverse names." (L I,ii)50 Though he regarded memory as reasonable, Locke had advanced the view that ideas in memory "fade," are subject to a "constant decay," and "are laid in fading Colours." (Essay II.x.4-5) Degradation in vivacity is a characteristic of associationist transitions, as Hume conceives them. In association founded on the relation of causation, for example, vivacity is lost in the transition from a perception to the associated idea. (cf. T98, 144) Though Hume often characterizes the imagination as the faculty that operates on perceptions fainter than those of memory, he writes at Liv.7 that "The memory, senses, and understanding are, therefore, all of them

298 LOUIS E. LOEB founded on the imagination, or the vivacity of our ideas." (T266) The central idea for Hume is that all belief is to be explained with reference to vivacity, which is transmitted associationistically. Memory lends itself so naturally to the associationist framework, that Hume need not argue for an associationist picture of that faculty; it was sufficient to advance it. It would have seemed pointless to Hume to use an argument analogous to the problem of induction to show that memory, like causal inference, must be associative. This once again illustrates the centrality of Hume's associationism to a proper understanding of the Treatise. In closing, let me return to one of my central claims. Some care is needed in formulating the thesis that Hume regards causal inference as justified, and that he is committed to degrees of probabilistic evidence. We have seen that such commitments persist as late as page 225 in I.ivA. The view that causal inference is justified, however, proves provisional. In I.iv.7, the "Conclusion of this book," Hume takes the constructive results of Part iii to unravel. 51 The collapse of his constructive position occurs with Hume's announcement that he is ready to "reject all belief and reasoning, and can look upon no opinion even as more probable or likely than another." (T268-69) The problem of induction, understood skeptically, obliterates distinctions between good and bad nondeductive arguments; all inductive arguments are equally good, in that they are all worthless. Though there is nothing of this in Part iii, Hume does get to a point in the Treatise where he is ready to accept such a conclusion. Hume reaches this point, however, one hundred seventy-five pages after formulating the problem of induction. Based on his own discussion and references in I.iv.7, his negative conclusions rely on the arguments of I.iv.l, "Of skepticism with regard to reason," and I.ivA, "Of the modern philosophy." l.iii.6 lies far in the distance. In the immediate aftermath of formulating the problem of induction, in Part iii, Hume distinguishes various degrees of evidence and probability based on causal inference. It is not until l.iv.7 that a generalized skepticism asserts itself. Hume reaches a point where he is ready to regard all causal inference (indeed, all belief) as unjustified; his readiness to do so, however, is not supported by the problem of induction. Beattie understood the distinction here. Though he sees no skeptical problem in I.iv.6, Beattie derisively quotes Hume's I.iv.7 announcement of his readiness to reject all belief and reasoning, and attacks Hume's basis for it in I.iv.I. (Truth, 165-66, 218) Though we encounter a generalized skepticism in the Treatise, its location, and the grounds for it, support rather than undermine the position that Hume does not intend the problem of induction to have skeptical force. Part iii of Book I is thoroughly constructive, as reflected in its title, "Of knowledge and probability." Its constructive results, however, prove provisional, giving way in Part iv, which treats "Of the sceptical and other systems of philosophy."52 How is it possible for the constructive assessment of causal inference to collapse? As I have noted, the considerations that lead Hume to

Causal Inference, Associationism, and Skepticism 299 this result are advanced in Part iv, not Part iii. We must suppose that in Part iv Hume has located considerations which he takes to show that his favored naturalistic theory ofjustification has systematically destructive results after all. As we have seen, the literature contains a number of suggestions about the shape of a naturalistic theory of justification in Hume, depending on whether we focus on irresistibility (as with Kemp Smith), critical reflection (Baier), reliability (Schmitt), or settled belief (as in my own view). In I.iv.7, Hume must take himself to have found that the conditions of his favored theory of justification cannot be met. More specifically, he must take himself to have found either that no belief-forming mechanisms are genuinely irresistible, or survive critical reflection, or are reliable, or produce settled beliefs-as the case may be-,and concluded that no beliefs are justified relative to his theory of justification. 53 In this paper, I have not tried to adjudicate between the competing naturalistic interpretations of Hume's epistemology. I have, however, called attention to some of the data for which those interpretations will have to account: Hume's commitment to causal inference and to degrees of evidence in Part iii and portions of Part iv, and his readiness to reject both in Liv.7. 54 NOTES 1. Though I draw on evidence from other works, my conclusions are restricted to Hume's intentions in the Treatise; his position in An Enquiry Concerning Human Understanding would require separate discussion. 2. Annette Baier advances this position in A Progress ofSentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) esp. pp. 54-56 and 65-70. As she recognizes, (p. 301n.5) Tom L. Beauchamp and Alexander Rosenberg develop a similar view in Hume and the Problem of Causation (New York: Oxford University Press, 1981) ch. 2, esp. sections ill-IV. I believe I provide considerations in support of this position that go beyond these discussions. In any event, I attempt to be scrupulous in acknowledging in footnotes evidence that Baier and Beauchamp and Rosenberg have brought forward. 3. L. A. Selby-Bigge, ed., A Treatise of Human Nature, second edition, with text revised and variant readings, by P. H. Nidditch (Oxford: Oxford University Press, 1978) pp. 88-89. References to this edition are abbreviated "T" followed by page number(s). Quotations of the Treatise and other works by Hume are based on T. L. Beauchamp, D. F. Norton, and M. A. Stewart, Humetext 1.0 (Washington: Georgetown University, 1990). This is (the first phase of) an electronic edition of Hume's works. References to Hume's works other than the Treatise are to the following editions: J. M. Keynes and P. Sraffa, eds., An Abstract ofa Treatise ofHuman Nature (Hamden, Connecticut: Archon Books, 1965) abbreviated by "A" followed by page number(s); L. A. Selby-Bigge, ed., Enquiries Concerning the Human Understanding and Concerning the Principles of Morals, third edition by P. H. Nidditch (Oxford: Clarendon Press, 1975) abbreviated by "E" followed by section number(s). 4. Hume fonnulates the principle as follows: "that those instances, ofwhich we have had no experience, resemble those, ofwhich we have had experience." (T89)

300

LOUIS E. LOEB

5. As I-Iume puts it: "The same principle cannot be both the cause and effect of another." (T90) 6. The suggestion to be rej ected is as follows: "It may, perhaps, be said, that after experience of the constant conjunction of certain objects, we reason in the following manner. Such an object is always found to produce another. 'Tis impossible it cou'd have this effect, if it was not endow'd with a power of production. The power necessarily implies the effect; and therefore there is a just foundation for drawing a conclusion from the existence of one object to that of its usual attendant. The past production implies a power: The power implies a new production: And the new production is what we infer from the power and the past production." (T90) 7. D. C. Stove, for example, assumes that it does. See his Probability and Hume 's Inductive Scepticism (Oxford: Oxford University Press, 1973) p. 34. For criticism of his position, see T. L. Beauchamp and A. Rosenberg, Hume and the Problem of Causation (New York: Oxford University Press, 1981) pp. 67-75. 8. Cf A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) p. 66. 9. Cf John P. Wright, The Sceptical Realism of David Hume (Minneapolis: University of Minnesota, 1983) pp. 12-13. 10. James Beattie, An Essay on the nature and Immutability of Truth, In Opposition to Sophistry and Scepticism, sixth edition (Edinburgh: William Creech, 1778) pp. 14, 15, 23, and 30. References to this edition are abbreviated "Truth," followed by page number(s). 11. Norman Kemp Snlith, "The Naturalism of Hume (I)" and "The Naturalism of I-Iume (II)," Mind 14 (1905) esp. pp. 152, 161, and 162; and The Philosophy of David Hume (London: MacMillan, 1941/66) esp. pp. 87, 388, 485, and 486. 12. For further discussion, and additional difficulties, see my "Instability and Uneasiness in Hume's Theories of Belief and Justification," British Journal for the History ofPhilosophy 3 (1995) pp. 301-329, esp. sections 1 and 5. 13. Cf A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) pp. 15,58,282,284 (for surviving reflection) and pp. 55, 93, 97, 99, 196, 284-85 (for surviving se!fscrutiny). 14. Frederick Schmitt, Knowledge and Belief (London: Routledge, 1992) ch. ill, pp. 53-83. Schmitt also raises, though in the end rejects, another naturalistic possibility, that beliefs are justified insofar as they result from mechanisnls that are adaptive (pp. 68-72). 15. See a number of my papers: "Stability, Justification, and Hume's Propensity to Attribute Identity to Related Objects," Philosophical Studies 19 (1991) pp. 237-270; "Hume on Stability, Justification, and Unphilosophical Probability," Journal of the History of Philosophy 33 (1995) pp. 91-121; "Instability and Uneasiness in Hume's Theories of Belief and Justification," British Journal for the History of Philosophy 3 (1995) pp. 301-329; and "Sextus, Descartes, Hume, and Peirce," (unpublished manuscript). What I want to emphasize in this paper is that Hume does not deploy the problem of induction with skeptical intentions. Though I prefer my interpretation of Hume's theory of justification to those of Baier and Schmitt, I try to preserve neutrality on this question in the present context. 16. In Locke, Berkeley, Hume: Central Themes (Oxford: Clarendon Press, 1971), Jonathan Bennett writes: "Hume does not, simply by assigning causes for our causal beliefs, imply that there are no reasons for them; but by treating only of their causes-

Causal Inference, Associationism, and Skepticism

301

by his silence about reasons for them-he does suggest that these beliefs caIUlot be supported by reasons, and even that they are unreasonable." (p. 300) Bennett, writing before naturalist and reliabilist theories ofjustification gained popularity in this century, overlooks the ways in which genetic accounts of belief can mesh with a naturalistic account of justification. Each of Kemp Smith, Baier, Schmitt, and myself attribute to Hume a naturalistic account, on which the genesis of a belief, or features that derive from its genesis;, determine its justificatory status. 17. There is one possible exception when Hume claims to uncover "a new and signal contradiction in our reason" (T150) in I.iii.13. For discussion, see my "Hume on Stability, Justification, and Unphilosophical Probability;," Journal of the History of Philosophy 33 (1995) pp. 91-121, sections 1-3, esp. p. 119. 18. Peter H. Nidditch, ed., John Locke, An Essay concerning Human Understanding (Oxford: Clarendon Press, 1975) IV.xi.9 (some emphases added and deleted). References to this edition are abbreviated "Essay." 19. It is possible to interpret Locke as holding that "sensitive knowledge," even at its best;, does not qualify as "knowledge," strictly speaking. He writes that the third degree of knowledge "passes under the name of Knowledge," so that we "allow" sensitive as well as intuitive and demonstrative knowledge. (Essay IV.ii.14) Cf. Anthony Quinton, The Nature of Things (London: Routledge & Kegan Paul, 1973) pp. 131-32, and my From Descartes to Hume: Continental Metaphysics and the Development afModern Philosophy (Ithaca, New York: Cornell, 1981) pp. 55-57. See Michael Ayers, Locke, Volume I, Epistemology (Routledge: London;, 1991) esp. pp. 9394 and 156-57, for the view that Locke does hold, as he says, that sensitive knowledge "deserves the name of Knowledge." (Essay IV.xi.3) I need not take a position on this question for the purposes of the present paper. In saying that, on the account of IV.xi. 11, I do not know that the object exists at times that I have not perceived it, I mean that our beliefs about the object's existence at times that it has not been perceived do not even qualify as the third degree of knowledge, whatever the ultimate status of "sensitive knowledge" might be. 20. Hume does not use Locke's example in elaborating his theory. The reason for this, I suspect, is that Locke's example involves the persistence of an object, a man;, through time. Many philosophers have had the intuition that persistence;, or at least the persistence of an unchanging object, caIUlot be analyzed as a sequence of casually related states of an object. Hume himself had intuitions in this region. For part of the story, see my "Causation, Extrinsic Relations, and Hume's Second Thoughts about Personal Identity," Hume Studies 18 (1992) pp. 219-232, esp. pp. 226-28. As a result of these intuitions, Hume prefers examples that do not involve the persistence of an object: the presence of fire, heat, and light (E22 and T87) a letter received from a friend in France;, (E22) or the presence a watch left on the beach. (E22) It should be said that Hume's use of examples is sparse;, especially in the Treatise. 21. This is not to say that Locke saw the restrictions on sensitive knowledge as a problem. He is content to assert that it is "highly probable that Millions of Men do now exist," (Essay IV.xi.9) though this probability does not constitute sensitive knowledge. 22. One might think that Berkeley's doctrine that to be is to be a collection of ideas in the mind of a perceiver or perceivers set Hume's problem. Berkeley's problem is to give an account of what the existence of objects consists in, at times they are not perceived. Hume's problem, at this stage of his inquiry, is different: it is to see how we come to believe that objects exist at times when they are not perceived, and whether

302

LOUIS E. LOEB

those beliefs are reasonable. My point in the text is that the tenus in which Hume fonnulates the problems fall directly out of Locke's discussion at Essay IV.xi. I also remind the reader that the extent of Hume's substantive engagement with Berkeley is a subject of controversy; such engagement on a particular topic cannot simply be assumed. See Richard Popkin, "Did Hume Ever Read Berkeley," Journal ofPhilosophy 56 (1959) pp. 533-545, and "So, Hume Did Read Berkeley," Journal ofPhilosophy 61 (1964) pp. 773-78; and John Passmore, Hume's Intentions (London: Duckworth, 1968) p. 86. 23. The Locke literature rarely takes note of the limitations of Locke's sensitive knowledge to present perception and memory. For two exceptions, see A. C. Fraser, An Essay Concerning Human Understanding by John Locke, in two volumes (Oxford: Oxford University Press, 1894) p. 336n.2, and James Gibson, Locke's Theory of Knowledge and its Historical Relations (London: Cambridge, 1917) pp. 175-76. Similarly, the Hume literature, so far as I know, does not take note of the relationship between Hume's inquiry into our knowledge of matters of fact and Locke's discussion of sensitive knowledge. Fraser, in the note cited, does draw some connection between Locke's discussion and that of Hume. David Owen cites IV.xi.9, but for a somewhat different purpose, to introduce Locke's conception of probability. See his "Hume and the Lockean Background: Induction and the Uniformity Principle," Hume Studies 18 (1992) p. 191, and "Hume's doubts about probable reasoning: was Locke the target?" in M. A. Stewart and John P. Wright, eds., Hume and Hume's Connexions (Edinburgh: Edinburgh University Press, 1994) p. 149. 24. Hume's causal theory of assurance has much in common with the causal theory of knowledge due to A. Goldman. See his "A Causal Theory of Knowing," Journal of Philosophy 64 (1967) pp. 355-372. 25. It might be said that Hume does challenge Locke's position, insofar as he attacks "the philosophical system"-representative or indirect realism, in I.iv.2. Though this is the case, Hume's attack at that stage presupposes the causal theory of knowledge of matters offact-cf. T193, 198, 212. 26. Cf. T. L. Beauchamp and A.Rosenberg, Hume and the Problem of Causation (New York: Oxford University Press, 1981) pp. 52-53; and A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) p. 56. 27. Strictly speaking, at T134 and 135 Hume is describing a habit as "perfect." Such habits derive from experience that is perfect in the sense I describe. 28. "That all nlen will die" and "that the sun will rise to-morrow" are the first two of Beattie's examples of morally certain beliefs in laws of nature. (Truth, p.121) 29. In Liii.9, beliefs based on the relation of cause and effect are due to reason, the faculty of demonstrative and probable reasoning. Since causal inference is not demonstrative, causal inference must be a kind of probable reasoning. At I.iii.I1, some causal inferences-the proofs-exceed probability. This discrepancy is purely tenninological. At I.iii. 9, Hume writes of "probable reasonings" in a wide sense in contrast to demonstrative reasoning. At Liii.II, Hume draws a distinction within probable reasoning in the wide sense between proofs and probability. 30. A number of commentators recognize that in I.iii.6 Hume seeks to establish that causal inference is due to the imagination or association, while insisting that he has a skeptical aim in stating the problem of induction. In Hume 's Skepticism in the Treatise ofHuman Nature, (London: Routledge & Kegan Paul, 1985) Robert Fogelin recognizes that the problem of induction "occurs as a step leading to the conclusion that causal

Causal Inference, Associationism, and Skepticism

303

inferences (so called) are the product of the imagination," (p. 56) a "development in the text which is, if anything, more important for following the main lines of Hume's thought." (p. 53) Fogelin insists, however, that Hume's statement of the problem of induction is skeptical in intent. (cf. pp. 6, 13, 48, and 152ff.) See also Barry Stroud, Hume (London: Routledge & Kegan Paul, 1977) pp. 52-53. It is difficult to see how these commentators can accommodate the overwhelming evidence that Hume regards causal inference as justified or reasonable in Part ill and well into Part IV. 31. Cf T. L. Beauchamp and A. Rosenberg, Hume and the Problem of Causation (New York: Oxford University Press, 1981) p. 53; and A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) pp. 87-88. 32. I discuss this propensity in detail in my "Stability, Justification, and Hume's Propensity to Attribute Identity to Related Objects," Philosophical Studies 19 (1991) pp. 237-270, esp. section 2. 33. Cf my "Hume's Moral Sentiments and the Structure of the Treatise," Journal of the History ofPhilosophy 15 (1977) pp. 395-403, section 5, pp. 402-3, and my Review Essay, Annette C. Baier, A Progress of Sentiments, Reflections on Hume's Treatise, Philosophy and Phenomenological Research 54 (1994) pp. 467-474, p. 469. 34. Prior to I.iii.4, perceptions of memory are ideas. (T8, 9, 12) At Liii.4, "ideas of memory. .. are equivalent to impressions." (T82) Perceptions of memory are often impressions thereafter. (e.g., T83, 86, 88, and 89) These shifts in tenninology do not matter for purposes of this paper. 35. In D. Owen,"Hume and the Lockean Background: Induction and the Uniformity Principle," Hume Studies 18 (1992), and his "Hume's doubts about probable reasoning: was Locke the target?" in M. A. Stewart and John P. Wright, eds., Hume and Hume's Connexions (Edinburgh: Edinburgh University Press, 1994), Owen takes Hume's target to be a Lockian conception on which reason generates probable arguments by the intervention of intermediate ideas. My own sense is that Hume's target is much broader, and-as with his attack on necessary connection-more amorphous. For a useful discussion, see Edward Craig, The Mind of God and the Works ofMan (Oxford: Oxford University Press, 1987) esp. pp. 77-90. 36. John Locke, O/the Conduct of the Understanding, in The Works ofJohn Locke, Volume ill (London: Thomas Tegg, 1823) p. 276. 37. John Laird, Hume's Philosophy of Human Nature (London: Methuen, 1932) p. 40. 38. See Richard 1. Aaron, John Locke (London: Oxford University Press, 1937) p. 141; Hans Aarsleff, "Locke's Influence," in Vere Chappell, ed., The Cambridge Companion to Locke (Cambridge: Cambridge University Press, 1994) p. 269; M. Ayers, Locke, Volume I, Epistemology (Routledge: London, 1991) pp. 112, 254; George Sidney Brett, A History ofPsychology, Volume II, Medieval & Early Modern Period (London: George Allen & Unwin, 1921) pp. 262-63; 1. Gibson, Locke's Theory ofKnowledge and its Historical Relations (London: Cambridge, 1917) pp. 22, 236; J. Laird, Hume's Philosophy ofHuman Nature (London: Methuen, 1932) pp. 39-40; Howard C. Warren, A History of the Association Psychology (New York: Charles Scribner's, 1921) p. 5; and John W. Yolton, The Locke Dictionary (Oxford: Blackwell, 1993) pp. 21-22. 39. I am indebted to David Hills for bringing this point to my attention. For its mention in the literature, see R. 1. Aaron, John Locke (London: Oxford University Press, 1937) p. 141; H. Aarsleff, "Locke's Influence," in Vere Chappell, ed., The Cambridge

304

LOUIS E. LOEB

Companion to Locke (Cambridge: Cambridge University Press, 1994) p. 269; 1. Laird, Hume 's Philosophy ofHuman Nature (London: Methuen, 1932) p. 41; 1. P. Wright, The Sceptical Realism of David Hume (Minneapolis: University of Minnesota, 1983) pp. 153-54; and 1. W. Yolton, The Locke Dictionary (Oxford: Blackwell, 1993) pp. 21-22. 40. Thomas Hobbes, Leviathan, with selected variants from the Latin edition of 1668, Edwin Curley, editor (Indianapolis: Hackett, 1994) ill, 7. References to this edition are abbreviated "L" followed by book and section number. 41. Thonlas Hobbes, The Elements of Law, IV, 7, in William Molesworth, ed., English Works of Thomas Hobbes. 42. For the point that Locke represents a step backward from Hobbes in this regard, see 1. Gibson, Locke's Theory of Knowledge and its Historical Relations (London: Carrlbridge, 1917) pp. 235-36. 43. Cf. T. L. Beauchamp, and A. Rosenberg, Hume and the Problem of Causation (New York: Oxford University Press, 1981) pp. 41-42, and A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) pp. 60-61. 44. Cf. T. L. Beauchamp, and A. Rosenberg, Hume and the Problem of Causation (New York: Oxford University Press, 1981) pp. 43 and 63. 45. Hume considers two cases, depending on whether or not we distinguish perceptions from obj ects, and claims that causal inference is not in play in either case. Hurne's concern with causal inference here is often overlooked, for example, in J. Bennett, Locke, Berkeley, Hume: Central Themes (Oxford: Clarendon Press, 1971) p. 320; and R. Fogelin, Hume's Skepticism in the Treatise of Human Nature (London: Routledge & Kegan Paul, 1985) p. 69. 46. The tenninology has the consequence that demonstrative reasoning is itself associative. I do not know whether Hunle has an account of intuition and demonstration that could sustain this claim. For relevant discussions, see Alan Hausman, "It Ain't Necessarily So," Hume Studies 8 (1992) pp. 87-101 and Charles Echelbarger, "Hume on Deduction," Philosophy Research Archives 13 (1987-88) pp. 351-365. In any event, Hume circumscribes such inference very narrowly. 47. For a statement of the argument to this point, see Bertrand Russell, The Analysis afMind (London: George Allen and Unwin, 1921) pp. 159-160. 48. Cf. Elements ofLaw, III,1,7 and De Corpere, XXV, 8 in Molesworth, volume i. 49. Cf. De Corpere, XXV,7. 50. Hobbes writes: "fancy and memory differ only in this, that memory supposeth the time past, which fancy doth not." (De Corpere XXV, 8) 51. Baier stresses the contrast in tone and results in Parts iii and iv. See her A Progress of Sentiments, Reflections on Hume's Treatise (Carubridge, Massachusetts: Harvard University Press, 1991)esp. pp. 7,12,32-33, and 105-7. For some reservations about her depiction of the contrast, and her explanation for it, see my review of A. Baier, Philosophy and Phenomenological Research 54 (1994) pp. 467-474. 52. C£ A. Baier, A Progress of Sentiments, Reflections on Hume's Treatise (Cambridge, Massachusetts: Harvard University Press, 1991) p. 55. 53. For detailed development of my view, see L. Loeb, "Stability, Justification, and Hume's Propensity to Attribute Identity to Related Objects," Philosophical Studies 19 (1991) section 6, pp. 261-67; "Unphilosophical Probability," Journal of the History of Philosophy 33 (1995) section 5; and "Instability and Uneasiness in Hume's Theories of

Causa/Inference, Associationism, and Skepticism

305

Belief and Justification," British Journal for the History ofPhilosophy 3 (1995) section 2.

54. I am grateful to participants in the conference "Ramus to Kant: Logic and the Workings of the Mind," University of Westem Ontario, May, 1995, for fine collegiality and helpful comments.

Kant's Dialectic and the Logic of Illusion ROBERT E. BUTTS

A sense of universal illusion ordinarily follows the reading of metaphysics. Herbert Spencer, First Principles, ii, iii, § 46, 158 (1875)

On September 21, 1798, Immanuel Kant wrote the following in a letter to .Christian Garve: It was not the investigation of the existence of God, immortality, and

so on, but rather the antinomy of pure reason-"the world has a beginning; it has no beginning, and so on," right up to the 4th [sic]: "There' is freedom in man, versus there is no freedom, but only the necessity of nature"-that is what first aroused me from my dogmatic slumber and drove me to the critique of reason itself, in order to resolve the scandal of apparent contradiction of reason with itself.l

In thus proclaiming the centrality of the antinomies for the development of the critical philosophy, Kant invites us to hazard a fresh look at the transcendental dialectic of the Critique ofPure Reason. I propose to accept the invitation, and to offer what I hope is at least a partial new understanding of Kant's extraordinary views on dialectical logic. To begin, it is to be noted that Kant holds a very negative view of dialectical logic as it had been understood and employed by the ancient Greeks. Clearly he is referring to the Sophists, but, given his generally unfavorable opinion of the philosophy of Plato, I think he would have wanted to include him in the general condemnation of those ancients who relied on dialectical reasoning. Kant conceives what he calls "(pure) general logic" as a "canon" abstracting from all content, and as a science whose operations are governed only by the laws of non-contraction and excluded middle (and, of course, identity). The ancients, and, I suppose, many others, were tempted to believe that this formal logic could be used to produce new objective knowledge, contentful assertions generated, not by any experiences, but by dialectical logic alone. This belief illicitly converts dialectic into an "organon." The conclusions of the application of the logic produce only illusions, mere semblances of the tmth. 2 For this reason, Kant thinks dialectic must be removed from its lofty perch as the highest form of knowledge-production, and demoted to the status of a "critique of dialectical illusion." (A60-64/B84-88) Perhaps the logical indiscretions of the ancients are obvious. Of greater interest is Kant's claim to have discovered dialectical illusion in the arguments of the antinomies. Kant understood the modus operandi of dialectical logic to be apagogic: nothing is directly confirmed as tme~ the truth of a proposition follows inferentially from the falsity of its logical opposite. 3 On this conception of

308 ROBERT E. BUTTS dialectical logic, one who wants to establish that the world began in time first proves that it cannot be true that the world is eternal because the concept of an eternal world is self-contradictory. The trouble is that the logic works in both directions: one who wants to establish that the world is eternal proves that it is a contradiction to regard the world as having begun in time. What nlatters is that one reduce one's opponent's view, by deductive means, to absurdity. I will not here rehearse in detail Kant's "solution" to the first two ("mathematical") antinomies. It is enough for present purposes to remember that Kant thought that his conception of transcendental idealism solves the problem of reason's lamentable clash with itself. (A490-497fB518-525) The World, on his view, is not an object existing in itself, it is only represented to us in ongoing experience anchored in sensation. Hence both the thesis and antithesis of each of the first and second antinomies are false. To put it somewhat more accurately: the predicates "began in time," "is spatially limited," and "is made up of simple parts," refer only to temporally transient representations of objects of outer intuition. They are referentially or semantically empty when applied (through illusion) to Worlds. Reason as canonical, perhaps because of the ease with which we can employ it and the regularity with which our best hopes are logically satisfied, fails as an organon, leaving us with the scandal to which Kant refers in the letter to Garve. Why the scandal? Transcendental idealism, together with Kant's distinctions between the phenomenal and the noumenal, and between the constitutive and regulative employment of ideas, would appear to be all we need in order to avoid embarrassment in the face of the discovery that metaphysics (as dialectic) can no longer be regarded as cognitively respectable (as a source of knowledge) but continues to be essential in its nlethodological role as seeker after systenl. Reason may not produce new knowledge, but it does provide the logical structure that all good science seeks to instantiate. Perhaps Kant's considered view is this one just outlined. Certainly he struggles mightily to sustain the transcendental solution in the Appendix to the Transcendental Dialectic of the first Critique, and in the discussion of teleology in the Critique of Judgment. These parts of the critical philosophy have received much scholarly discussion. Often his interpreters largely ignore the fact that Kant did, after all, hold that metaphysics is beset by illusions, that metaphysical thinking is delusive. Like it or not, his views on transcendental illusion are crucially central to his entire theory of the philosophical status of nletaphysics. It is one of Kant's most cherished central doctrines that the faculties of the nlind be understood as having entirely separate functions. It is the business of sensibility to provide the sensory content of experience without which nothing can be known. It is the function of the understanding to seek to know about what is given in sense-contentful experience. Reason supplies the logical form for our science, without which no assurance can be obtained that_ ~~~~~Il~_ ~~

_

Kant's Dialectic and the Logic ofIllusion 309 know is related to anything else that we know. Such a neat compartmentalization is disrupted, however, by the fact that humans judge, make assertions, have beliefs, inclinations, and, sometimes, fall prey to unrepentant stubbornness. Put differently, the objectivity that the faculties of the mind are designed to preserve is often interfered with by our subjective responses to what we are thinking. It is this idea that seems to me to be uppermost in Kant's thought as he discusses the dialectical logic of illusion. When, for example, I take it as true that the world has a beginning in time, I am, although I may not be aware of it, confronted with an illusion. The "...unobserved influence of sensibility..." on the objective grounds of my judgment unites with these grounds to produce an error (of judgment) by causing these grounds to: deviate from their true function-just as a body in motion would always of itself continue in a straight line in the same direction, but if influenced by another force acting in another direction starts off into a curvilinear motion. In order to distinguish the specific action of understanding from the force which is intermixed with it, it is necessary to regard the erroneous judgment as the diagonal between two forces-forces which determine the judgment in different directions that enclose, as it were, an angle-and to resolve this composite action into the simple actions of the understanding and of sensibility. (A294-951B35l)

Kant is fond of this way of thinking, introducing what might be called "the geometry of illusion." He also employs this geometry in his lectures on logic. For example, in The Vienna Logic, he is reported to have said: When a body is affected by 2 powers, there arises a third movement, where it does not remain on the track of the one or of the other, i. e., the diagonal force. E. g., a ball that would have gone directly into the center of the target goes somewhat toward the edge of the target when it is pushed sideways by the wind, and it was pushed there neither by the power of the [gun]powder nor of the wind. 4

Similarly, the Dohna-Wundlacken Logic reports: "All our errors are turnings ofjudgments to the diagonal-erosswise (the [German] word [for this] comes perhaps from the English [word] square). (720) (The German for crosswise is in die Quere. Square=quadrat) (LL, m. 213)5 At first glance it seems that what Kant had in mind is the famous parallelogram offorces (figure 1).

A ....

---\ 8

~u Figure 1

310 ROBERT E. BUTTS A body i1' carried by an impressed force from A to B in a given time, and from A to C by a second force. Complete the parallelogram ABeD, and by both forces acting together, it will in the same time be carried in the diagonal A to D. However, in dialectical illusion, no body is in motion, and the two "motions" of the objective concept and the subjective principle do not act together to produce the motion along the diagonal. So figure 1 does not accurately represent Kant's intentions 6 The question is: what does this quasi-geometry of illusory thought really mean? Just what kind of illusion are we dealing with when we encounter metaphysical notions? Some candidates can be eliminated at once. Metaphysical illusion is not optical illusion, as when I see the straight stick appear to be bent when placed in water. Nor is it moral illusion, which occurs when our best self-interest seems to arise from duty. Nor is it logical illusion (fallacious reasoning arising from inattention to the rules of logic). (LL, The Vienna Logic, m. 832; A2961B353) Kant repeatedly refers to metaphysical illusion as transcendental illusion, (see A295-971B352-53) although, as I will try to show shortly, the use of "transcendental" here is misleading. Consider figures 2,3 and 4.

Figure 2

Figure 3

Figure 4

Figure 2 is the embodied mind as a tabula rasa. But there are no mental blank tablets in Kant's view of knowledge. The chalk of sensation writes upon a tablet possessing the intricate patterns of space and time. Figure 3 represents normal cognition, where sensibility (operating from A to B) and the understanding (operating from A to C), are construed as combining to carry the mental activity from A to D, thus forming the sense-contentful concept. If you like, figure 3 is the mentalistic equivalent of the parallelogram of physical forces. Figure 4 represents the production of metaphysical illusion, where the motions of understanding and sensibility combine to generate, not an object of possible experience, but a crooked diagonal, one that only seems to form a genuine concept, but instead yields a mere semblance of such a concept.

Kant's Dialectic and the Logic ofIllusion 311 Unfortunately, the figures do not really help us to understand just what Kant has in mind when he writes that it is the unnoticed influence of sensibility on our attempts to correctly conceptualize that generates the illusions. He repeatedly reminds us that the senses do not err. (A293fB350; LL, The Vienna Logic, m. 825; LL, Dohna-Wundlacken Logic, m. 720; The Jiische Logic, m. 53) The point is irrelevant here. For there is no proper nlanner in which I can say that I see, or touch, or smell, or hear a "world," or any other presumed metaphysical object; that is, if I anl talking about sensing as the ordinary having of sensations of things in spatio-temporal contexts. Nor, in an extended meaning of "sensing," can I be said to have distorted inlages (nletaphysical worlds) of what I actually (in the ordinary meaning of the term) sense. It is to be regretted that Kant does not show in any detail just what happens when sensibility deceives us. Instead, he refers us to the fact that we judge, thus suggesting that sometimes our judgments are tainted by our inclinations to believe this or that without really having a warrant to do so. What kind of judgments? Certainly not properly formed categorial judgments. Certainly not judgments of taste (aesthetic judgments). And surely not judgments of purpose, either as claims to probability or as guides to practice. Judgments register beliefs. Beliefs can be either true or false, and are often based on confusions or illusions. But Kant does not tell us precisely how contaminated judgments can create metaphysical illusion. In other passages, perhaps wisely, Kant abandons the talk about sensibility and judgment and takes a different line. Reason, in the form of deductive logic, seeks to systematize knowledge by accommodating the results of conceptual knowledge in ever more general, hence simpler, deductive systems. Reason's quest is parsimonious. The old metaphysical adages of parsimony (nature does nothing in vain, nature always takes the shortest path) are replaced by the rational demand to seek simplicity, to engage in an endless quest for empirical predicates that are expressively very rich, that, if you will, say a lot. These transformed principles are, Kant thinks, subjectively necessary as providers of logical connections between our otherwise logically quite lonely conceptual claims. The problem arises when we transform the transformation-when we mistakenly construe the subjective principles as objective, thus permitting our crosswise thought to regard metaphysical sentences as making objective claims about things as they are in themselves, as being about a world lying beyond the limits of categorized and schematized human experience of things occurring in space and time. This is the kind of illusion Kant calls "transcendental." (A297;B353-54) Its creation depends neither on the disruptive influence of sensibility, nor on idiosyncracies of individual judgments. 7 It is important to note that it is not that any particular (idiosyncratic) subjective response on my part conles to interfere with logic properly applied. A person's like or dislike of temporal beginnings is not at issue. A sudden muscular spasm near my broken femur is not going to help convince me that

312 ROBERT E. BUTTS the world is indeed made up of simple parts. Kant tells us that metaphysical illusion cannot be prevented. Just as the sea always appears higher at the horizon than at the shore (since we see it through higher rays of light) and just as the astronomer necessarily sees the moon as larger at its rising, so also we are ceaselessly and unavoidably entrapped by the deceptions of dialectic. Although Kant, as I have mentioned, calls such deceptive metaphysical principles (revealed in the dialectic) "transcendental," an additional distinction he marks between immanent and transcendent principles nlust be attended to. (A295-96;B352-53) Immanent principles are those "...whose application is confined solely within the limits of possible experience." The principles of the analogies are, for example, immanent principles. Principles whose application seeks to transgress the limits of possible experience, "...to seize possession of an entirely new domain which recognizes no limits of demarcation," are transcendent principles. Transcendental errors seek to apply the categories beyond the limits of experience. Such errors are corrected by a proper understanding of the distinction between the phenomenal and the noumenal, where the noumenal counts as the limit of demarcation. Dialectical errors, on the other hand, are not in this respect transcendental, but are transcendent. Swedenborg's alleged discourses with the angels are in this respect transcendent, determined to confirm the reality of a realm of beings lying outside the limits of (ordinary) human experience. Strictly speaking, such transcendent principles cannot be eliminated. Dialectic can disclose thenl, and we can then guard against the inclination to take them too seriously, but we cannot get rid of them once and for all. Enthusiasm, die Schwarmerei, is a permanent threat to the furtherance of proper philosophical analysis. 8 One nlust ponder Kant's claim that transcendent illusion is inevitable. For all persons at all times? Was there ever a time when Hume construed the principles of parsimony as truths about an experientially inaccessible nature? Do empiricists all begin as comnlitted metaphysicians and only later awaken from their dogmatic naps? Kant's astronomer has a correct theory of the sizes and motions of bodies in our solar system. The theory gives him reasons to allow for the fact that the moon looks so large as "it rises" even though· he "knows" that the determination of its actual size has nothing to do with this appearance. He will never again be deceived by the appearance of the rising moon-the theory dispels the epistemic influence of the optical illusion. The point is that Kant's critical analysis of metaphysical illusion should give us theory sufficient to dispel the epistemic effects of such illusion, even if the dreams are recurrent and inevitable. A short form of Kant's story might go like this. We begin our metaphysical speculations with a certain deserved confidence in general logic. Indeed, the logic never lets us down. It is our disposition to believe that there are certain objects with knowable characteristics whose ontological status is both atemporal and aspatial that leads us to the conviction that logic, unaided by sense

Kant's Dialectic and the Logic ofIllusion 313 experience, can produce justifiable knowledge. Since there are no such objects, so Kant's story goes, dialectic deceives us. The emphasis here, I firmly believe, should be put upon the fact that those who limit metaphysical results to consequences of an exercise in deductive logic allow themselves to be deceived by that very logic. It is all so easy, and so convincing. Logic is the ground and savior of the a priori. Its results are secure. The highest form of knowledge is metaphysica ordine geometrico demonstrata. What happens, however, if the logic, so to speak, turns upon itself, if it allows the deductive proofs of two apparently logically incompatible propositions? At this point we do not dispense with the logic, we insist that dialectic give up its pretensions to being an organon, to possessing a special content. Very well. Why not, then, simply purge dialectic of the unwanted objects, go on to reformulate metaphysical generalizations as methodological maxims (as Kant insists that we dO),9 and in other respects let the chips fall where they may? I suggest that Kant gives us at least part of the answer to this question in the important Preface to the second edition of Critique of Pure Reason. It is not often enough noted that Kant's posture in this Preface is agonistic. The critical philosopher is likened to a policeman (Bxxv)~ philosophical works are referred to as "armed." (Bxliv) What is it that is to be protected, what fought against? Kant tells us: "...speculative reason [must) be deprived of its pretensions to transcendent insight." (Bxxx) More fully, Kant catalogues the enemies: "Criticism alone can sever the root of materialism, fatalism, atheism, freethinking, fanaticism, and superstition, which can be injurious universally~ as well as of idealism and scepticism, which are dangerous chiefly to the Schools, and hardly allow of being handed on to the public." (Bxxxiv; Kant's emphasis) Of course Kant believes that he can effectuate this botanical murder without loss to the minds of members of the public. They can still believe in god and immortality. The meanings of some of the sylubols may have to change, but practical affairs will be unaffected, in fact, greatly improved. The main point is that dialectical logic reveals that not only is metaphysical speculation lacking in cognitive respectability, the belief in the existence of transcendent objects is downright dangerous if allowed to continue unchecked by philosophical criticism. It is for this reason, I believe, that the aged Kant could confidently write (in the letter to Garve) that is was the antinomies that awakened him from (at least one of) his dogmatic naps. One mild intellectual irritant remains. What are we to nlake of Kant's geometry of dialectical illusion? Is it merely a metaphor, an attempt on Kant's part to show that mathematical pictures, although often not to be taken literally, are nevertheless aids to thought? Or is there something of a nlore substantial suggestion involved? If so, what does the geometry invite us to conclude, if only hypothetically?

314 ROBERT E. BUTTS Notice that a subjective principle's interference with an act of objective cognizing (a categorized, schematized conceptualization of something happening in time and perhaps also in space) is itself an event. When the lines meet, they point to something happening in me. If this were not so, there would be no point in worrying about the deleterious effects of the free exercise of pure reason. It seems obvious that we must read Kant as having been concerned about the actual cognitive happenings in those "afflicted" with metaphysics. Metaphysics is epistemologically specious~ it is also psycho-physically traumatic. The diagonal thus represents a datable part of my personal history occurring in me somewhere. Where? I think the only place available is in my brain. Dialectical illusions, just like optical ones, are neurological events taking place in my brain. In Figure 4, the lines moving in different directions and involving incompatible forces, represent paths through nerves, and so does the line expressing the diagonal. If such a diagonalization or crosswise nerve impulse occurred only once or at a few widely separated times, there would, one supposes, be no problem. It is only when the occurrence of that impulse is frequently repeated that trouble arises. If the line representing the diagonal is viewed as frequently reinforced, as making a kind of groove in the brain, a person is afflicted with a disease (metaphysics) that, for Kant, needs a cure. The diagnostic syndrome is depicted in Figure 5.

Figure 5 Atheism and fanaticism and the other enemies of criticism involve attitudes and they also act as stimuli to behavior. It is not so much that they are false, they are threats to personal and social moral order. It is not often enough noted that much of Kant's career was devoted to attempting to prevent the formation of nerve grooves in the brain. He was much concerned with the avoidance of counter-productive reinforcement: his unwillingness to participate in daily religious exercises in the university chapel~ his admonition to his students to read less and think more; his (Protestant) objection to lavish ornamentation and monotonous music in churches; his constant warning ~~~i!1~! _~~~~r!~i!1g

beliefs~

Kant's Dialectic and the Logic ofIllusion 315 philosophy, because one cannot learn philosophy, but only to philosophize. And of course the touchstone of it all, the message of the essay on enlightenment: to be enlightened is to be free of tutelage, to be the guardian of oneself. Perhaps in view of these essentially moral considerations we should grant Kant his point about sensibility as what influences us to accept illusion. It could be that he thought not that sensibility causes metaphysical illusion, but that the presence of illusion frees sensibility to become sensuality. As Foucault notes, Erasmus ,vrote that madness is that "sweet illusion" that frees the soul from "its painful cares and returns it to the various forms of sensuality." 10 For Kant, such freedom is most definitely not a consummation devoutly to be wished. He was terrified of the potential of the passions of the body to overtake the life of the rational mind. Still, the groove in the brain corresponds to what Malebranche called the "sensible emotion of the soul," the passion experienced when animal spirits in the body congregate in one sensory receptor. l1 Foucault correctly notes (1988: 134-135) that there is geometry at work here. The rallying together of the otherwise dispersed animal spirits (or what for Kant is the creation of the groove in the brain) has a form mirrored in the felt emotions of the soul. The motions of the body and the "motions" of the soul are thus reciprocal. Perhaps it is considerations of this kind, involved as they are in the explication of the physiological basis of passion, that we should bring to bear in our attempt to understand Kant's horror of metaphysical illusion. Illusion become passionate-well-grooved-is the basic pathological condition preventing the genuine quest for certifiable knowledge. Kant may have read Fran90is Boissier de Sauvage's Nosologie methodique (Lyon 1772) and would have been much taken with that author's statement that "We call madmen those who are actually deprived of reason or who persist in some notable error~ it is this constant error of the soul manifest in its imagination, in its judgments, and in its desires, which constitutes the characteristic of this category." Kant's groove in the brain is much like the constant error here selected as the deprivation of reason. What makes madness a matter of logic has to do, no doubt, with the fact that a constant error, a well-grooved illusion, can function in the thinking and behaviour of a person within the strictest rules of logical procedure. 12 Although my evidence for this point is circumstantial, it appears at least very plausible that Kant's critique of metaphysical illusion is carried out in the context of well-received contemporary ideas in physiology and in the classification of mental aberrations. 13 Elsewhere I have tried to demonstrate the great emphasis Kant placed upon physiological and psychological matters of the sort involved here. (Butts, 1986) I regard my reading of the dialectical revelation of transcendent illusion as an additional chapter in what has turned out to be a continuing research project.

316

ROBERT E. BUTTS NOTES

1. Immanuel Kant:> Kant: Philosophical correspondence 1759-99, ed.:> trans. Amulf Zweig (Chicago: University of Chicago Press:> 1967). 2. Immanuel Kant, Immanuel Kant's Critique of Pure Reason, trans. by Norman Kemp Smith (London: Macmillan ). I follow the convention of referring to the first and second editions of Critique ofPure Reason as A and B:> and abbreviated references will appear hereafter in the text by letter and number. The present reference is at (A60641B84-88). 3. Kant had a low opinion of the use of indirect reasoning in philosophy. He thought that, strictly speaking, apagogic reasoning is only legitimately employable in mathematics. 4. For the lectures on logic I employ the translations of J. Michael Young:> Lectures on logic. The Cambridge edition of the works of Immanuel Kant (Carubridge: Cambridge University Press:> 1992):> margin numbers 824-25. The number references are to the marginal numbers of the Academy Edition of Kant's works, the texts used for translation, and not to page numbers in Young. Abbreviated references will hereafter appear in the text under LL, with title of text used for translation and marginal number(s). I think it is highly unlikely that Kant said exactly what is reported in the quoted passage. Consistently with what he says in the first Critique and elsewhere, he more probably held that the ball moves crosswise because of the conjoining of the two forces, gunpowder and wind. Some force has to cause the diagonal motion of the ball. I am grateful to Rolf George for suggesting this line of interpretation, although I do not hold him responsible for my way of developing it. Certainly Kant did not believe that dialectical illusion introduces the action of occult causes. 5. In The Blomberg Logic (LL, m. 87), we have: "All errors are, so to speak:> crooked lines, which we determine while being driven from the one side by the understanding, from the other side by sensibility." 6. In discussion of the short form of this paper Phillip Cummins suggested that the parallelogram of forces seems better to fit Kant's idea of normal cognition, with sensibility and understanding as the two "forces" which, when they act conjointly:> produce knowable objects of possible experience. 7. In the Transcendental Dialectic of the first Critique Kant teaches us that the central task of reason is to seek logically simple concepts (he calls them "conditions") from which less simple concepts may be deductively derived. The ultimate task of pure reason is to find "the unconditioned," that most simple of all concepts from which the entire aggregate of known laws of nature can be deduced, and hence unified and systematized. It is subjectively necessary for reason to seek "to find for the conditioned knowledge given through the understanding the unconditioned whereby its unity is brought to completion." (A307/B364) However, "this logical maxim can only become a principle of pure reason through our assuming that if the conditioned is given, the whole series of conditions, subordinated to one another-a series which is therefore itself unconditioned-is likewise given, that is, is contained in the object and its connection." (A307/B364) This (apparently objective) principle is, however, transcendent (see below) in Kant's sense of this term, and can have no empirical employment. To think that it can have such an employment is to be deluded-once again, by the irresistible energy and charm of d~~~!!!~JQgi~. -

Kant's Dialectic and the Logic ofIllusion

317

Note that although Kant says that the strong fonn of the principle must be assumed, and although his central doctrine appears to be that the assumption is justified in practice, with the principle employed regulatively as a methodological rule, he also at some places suggests that in this fonn the principle is objective. See A6611B689, where he says that principles of parsimony are "not merely methodological devices." But compare this strange admission with the entire discussion of maxims (methodological rules) in the Appendix to the Transcendental Dialectic, a discussion that surely contradicts the slip at A6611B689. Margaret Morrison ("Methodological rules in Kant's philosophy of science," Kant-Studien, Heft 2.,1989, pp. 155-172), has a useful, wellcrafted, discussion of the tensions generated by Kant's perhaps inadvertent slides from the subjective to the objective. 8. Those present-day prophets who at the end of day trumpet the demise of the usefulness of epistemology might do well to ponder this Kantian observation. 9. See note 6 above, and my, Kant and the double government methodology: Supersensibility and method in Kant's philosophy ofscience 2d ed. (Pallas Paperbacks. Dordrecht: D.Reidel, 1986). 10. Michel Foucault, Madness and civilization (New York: Vintage Books. A division of Random House, 1988) pp. 27-28. Abbreviated reference will appear hereafter in the text as Foucault, 1988: page number(s). 11. Nicolas Malebranche, The search after truth, trans. by Thomas M. Lennon & Paul 1. Olscamp (Columbus: Ohio State University Press, 1980) p. 349. 12 It is interesting to note that Sauvage contended that diseases of the soul require philosophical persuasion to effect cures. Error must be replaced by truth. Kant's thought holds a mirror to this therapeutic strategy: sound understanding must replace illusion. See Butts 1986, pp. 298-318, for discussion of Kant's own attempt to develop a psychiatric nosology and his views on philosophy as curative and preventative medicine. 13. There are many indications that Kant was well-versed in the contemporary literature on psychological and physiological matters. He had read Sauvage, and Cullen, and perhaps Willis. See Butts 1986, pp. 282-318, for a full discussion of this background to Kant's thought.

Bibliography ANTHONY LARIVIERE and THOMAS TEUFEL This bibliography is a survey of the variety of European texts, published both for the schools and for the public, under the headings: logic, dialectic, and method. It is in no way comprehensive. Indeed, until the library catalogues of the many Continental and British libraries, both large and small, are made widely available, we cannot be sure what exactly. will tum up. Due to many factors, including damage and destruction from wars and natural causes, the listings of many Continental libraries do not always match their holdings. We have included here bibliographical listings only for texts we have by various means ensured the present existence of, in American, British, or European libraries. This is to ensure that this bibliography can serve its intended purpose as a starting point for scholars. We would be renliss, then, if we did not mention Wilhelm Risse's invaluable work Bibliographica logica (Hildesheim: G. Olms, 1965), which has served as a starting point for our own research and which remains a valuable tool. Wherever possible we have included full bibliographical citations, including publisher/printer details~ this, however, has not always been possible. A significant number of the cited texts exist in facsimile editions~ however far too few do, and we have not found it useful to indicate which, since this process is ongoing, new titles being photo-reproduced yearly. Aconcio,.Iacopo. De methodo (1582). Dusseldorf: Stein-Verlag Janssen, 1971. Aldrich, Henry. Artis Logicae Compendium. Oxonii: E Theatro Sheldoniano, 1692. Alsted, Johannes Henricus. Clavis artis Lullianae et verae logices. Argentorati: sumptibus L. Zetzneri, 1609. Alsted, Johannes Henricus Logicae systema harmonicum. Herbornae Nassoviorum: typis G. Corvini, 1614. Alsted, Johannes Henricus. Compendium logicae harmonicae. Herbornae Nassoviroum: typis G. Corvini, 1623. Aneponymus, Gregorius. Gregoriou tou Aneponymon Synoptikon tes philosophias syntagma. Augustae Vindelicorum: Excudebat Michael Manger, 1600. Anonymous. Logicae Compendium. Glasgow: 1764. Argens, Ie Marquis d', Jean-Baptiste de Boyer. La philosophie du bons sens, ou Reflexions philosophiques sur I'incertitude des connoissances humaines, a I'usage des cavliers et du beau sexe. Londres: aux depens de la Compagnie, 1737. Arnauld, Antoine, and Pierre Nicole. La Logique, ou I'Art de penser. Paris: C. Savreux, 1662.

320 BffiLIOGRAPHY Bacon, Francis. The two bookes of the projiciencie and advancement of learning (1605). AmsterdarnlNew York: Theatrum Orbis Terrarum/Da Capo Press, 1970. Barbay, Pierre. Compendium Logicae conimbricencis e Societati Jesu. Evorae: Ex Typographica Academiae Eborensis, 1683. Barbay, Pierre. Commentarius in Aristotelis Logicam ed. 5. Parisiio: Apud Georgium & Ludovicum Josse, 1690. Bardili, Christoph Gottfried. Grundriss der ersten Logik. 1800. Bartholin, Caspar Berthelsen. Enchiridion logicum ex Aristotele et opt. ejus interpretum monumentis ita concinnatum, ut contineat praecepta acurate de c larata, controversias praecipuas omnes enucleatas, praxin brevem et perspicuam syllogismos conjiniendi, scripta resolvendi, dejiniendi, disputandi, docendi, refutandi, etc. Wittebergae: excudebat M. Henckelius, 1612. Bartholin, Caspar Berthelsen. Logica major locupletata. 1625. Bary, Rene. La fine philosophie accommodee a I 'intelligence des dames. Paris: S. Piget, 1660. Baumann, Benedict. Intellectus operationes pro exercitio... Fuldae: Typis Stahelianis, 1789. Beattie, James. The Theory ofLanguage. London: A. Strahan, T. Cadell, 1788. Bekker, Balthasar. De philosophia Cartesiana admonitio candida et sincera. Alt-Wessel: 1668. Bentham, Edward. An introduction to logick (1773). Merston: Scolar Press, 1967. Bentzel, Ignatz. Logica in compendium redacta... Herbipoli: Typis Marci Antonii Engmann, 1732. Berault, Peter. Logick, or, The Key ofsciences, and the Moral science, or, The lvay to be happy. London: Printed by Thomas Hodgkin, 1690. Bernouilli, 1. Parallelismus ratiocinii logici et algebraici. 1685. Bertius, Petrus. Logicae Peripateticae libri sex. Lugduni Batavorum: ex off. 1. Patii, 1604. Best, William. A concise system of logics, in question and answer. New York: Printed by S. Campbell, for the author, 1796. Blundeville, Thomas. The Arte ofLogike. London: 1599. 170(4°). Bossuet, Jacques-Benigne. "La logique." Oeuvres inedites de Bossuet(1828). Paris: Beauce-Rusand, 1670-78. Brerewood, Edward. Elementa Logicae. 1619. Brirthaupt, Christian. "Dissertatio de tribus logicae instauratoribus Verulamio, Ramio et Cartesio." Jena, 1712. Broedelet, 1. Disputatio philosophica qua disquirtur an philosophia Cartesiana sit Aristotelicae, prout hodies docetur, praeferenda. Utrecht: 1676.

Bibliography

321

Buffier, Claude. Les Principes du raisonnement exposes en deux logiques nouveles, avec des remarques sur les logiques qui ont eu Ie plus de reputation de notre temps. Paris: P. Witte, 1714. Buffier, Claude. Cours de sciences sur des principes nouveaux et simples, pour former Ie langage, I'esprit et Ie coeur dans I'usage ordinaire de la vie. Paris: G. Cavelier et P.-F. Giffort, 1732. Burgersdijck, Franco Petri. Institutiones logicae. Cantabrigiae: Ex officina Rogeri Daniel, 1626. Burgersdijck, Franco Petri. Monitio Logica, or, An Abstract and Translation of Burgersdicius His Logic by a Gentleman. London: 1697. Burthogge, Richard. Organum vetus et novum, or a Discourse of Reason and Truth wherein the Natural Logic Common to Mankinde is briefly and plainly described. London: 1678. 72(8°). Busch, M. Anfangsgrunde der logikalischen Algebra. 1768. Castillon, G.F. Reflexions sur la logique. 1802. Castillon, G.F. Sur un nouvel algorithme logique. 1803. Chanet, Pierre. Traite de I'esprit de I'homme et de ses fonctions. Paris: Vue 1. Camusat et P. Ie Petit, 1649. Clauberg, lohan. Defensio Cartesiana... Amsterdam: 1652. Clauberg, Johan. Johannis Claubergii initatio philosophi, adversus Jacobum Revium et Cyriacum Lentulum...pars prior exoterica, in qua Renati Cartesii dissertatio de Methodo vindicatur, scinul illustria Cartesianae logicae et philosophiae specimina exhibentur. Amstelodami: apud L. Elzivirium, 1652. Clauberg, Johan. Logica vetus et nova. Amstelodami: apud L. Elzivirium, 1654. Cochet, Jean-Baptiste. La clef des sciences et des beaux arts, ou la Logique. Paris: J. Desaint et C. Sai1lant, 1750. Coke, Zachary. The Art of Logick, or, the Entire Body of Logick in English Unfolding to the Meanest Capacity the Way to Dispute Well. London: Printed by Robert White for Geoge Calvert, 1654. Crakanthorpe, Richard. Logicae libri quinque de praedicabilibus, praedicamentis, syllogism 0 ejusque speciebus.. .Editio secunda. Landini: impensis R. Young, 1641. Crousaz, Jean-Pierre De. Systeme de reflexions que peuvent contribuer a la nettete et a I 'etendue de nos connaissances: ou Nouvel Essai de logique. Amsterdam: F. L'Honore, 1712. Crousaz, Jean-Pierre De. Summa logicae. Groningae: apud 1. Spikes, 1725. Crousaz, Jean-Pierre De. Systeme de logique abrege par son auteur. Lausanne: J. Zimmerli, 1735. Cruz, fray Alonso de la VeT. Resolutio dialectica cum textu Aristotelis. Salamanticae: Apud Ivannem Baptistam a Terranona: Expensis Vencentij & Simonis Portonarijs, 1573.

322 BffiLIOGRAPHY Darjes, 1.G. lntroductio in ortem inveniendi seu logicam theoretico-practicam. 1747. Darjes, 1.G. Weg zur Wahreit. 1776. Derodon, David. Compendium logicae. Genevae: sumpt. 1. A. et S. de Tournes, 1663. Dumfe1d, Jacob. Compendium logicae perpateticae perbreve. Rinthelii: Impressum typis Petri Lucii, Academiae typographi, 1655. Duncan, William. The Elements ofLogick in Four Books. London: 1. Dods1ey, 1748. Dup1eix, Scipion. La logique, ou, Art de discourir et raisonner.. .Edition seconde. Paris: D. Salis, 1604. Dup1eix, Scipion. Corps de philosophie, contenant la logique, la physique, la metaphysique et I 'ethique. Geneve: B. Labbe, 1623. Durant, 1. Himbert. Les Lumieres de I 'eloquence. 1625. Eliot, John. The Logick Primer. London: 1672. Farcio1ati, Jacobo.lnstitutiones logicae. Venetiis: Typiis Antonii Zatate, 1789. Farde11a, Michelangelo. Universa philosophiae systema. 1691. Gassendi, Pierre. Opera omnia. Lyon: Ant. de 1a Poterie, 1658. Gassendi, Pierre. P.G. institutio logica et philosophiae Epicuri syntagma. Londini: 1668. Genuensis, Antonius (Genovesi). La logica per gli giovanetti. Bassano: a spese Remondini, 1766. Genuensis, Antonius (Genovesi). Elementorum artis logicae-critica libri quinque. Neapo1i: apud fratus Raymondios, 1778. Geulincx, Arnout. Logica fundamentis suis a quibus hactenus collapsa fuerat restituta. Lugduni Batavorum: paud H. Verbiest, 1662. Goc1enius, Rudolph. Praxis Logica... Francofurti: Ex officina Typographica Zachariae Pa1theri, 1595. Good, Thomas. A briefEnglish tract oflogick. Oxford: Printed by L. Lichfield, 1677. Gottsched, Johann Christoph. Erste Grande der Vernunftlehre... Leipzig: Bey Bernh. Christoph Breitkopfund Sohn, 1766. Gramnlont, Scipion de. La Rationelle ou I 'art des consequences pour bien inferer et conclure. 1614. Grand, Antoine Le. lnstitutio philosophiae secundum principia D. Renati Descartes. Londini: 1672. Granger, Thonlas. Syntagma Logicum. 1620. Grebner, Thonlas. Dissertatio philosophica de scientia. Wurzburg: Typis Joannis Jacobi Chirstophori K1eyer, 1754. Hauno1d, Christoph. Logica practica in regulae digesta. Bambergae: Martini Gobhardt, 1752. Havichorst, Aloysius. Institutiones logicae. Westphalia: P.H. Perrenon, 1776.

Bibliography 323 Heereboord, Adriaan. Selectarum ex philosophia disputationem volumen primum. Lugduni Batavorum: 1650. Heereboord, Adriaan. Meletemata philosophica, maximam partem metaphysica. Lugduni Batavorum: 1654. Heereboord, Adriaan. Philosophia naturalis, moralis, rationalis. Lugduni Batavorum: 1654. Heereboord, Adriaan. Ermhneia logica, seu explicatio, tum per nota, tum per exempla synops. logicae Burgersdicianae. Lugduni Batavorum: 1657. Heimbach, Matthias. Dialectica Aristotelico-rationalis. Coloniae Ubiouem: Ex officina Henrici Rommens-kirchen, 1730. Heimmann, Jakob Fr. De Logices Aristotelinae, Ramae, Cartesiane et eccelisticae insufficienta. Halberstadt: 1657. Herbert, Edward, Baron Herbert of Cherbury. De causis errorum. n.p., n.d. Hobbes, Thomas. Computatio, sive Logica. Trans. Aloysius Martinich. New Yark: Aberis Books, 1981. Holland, G.J. Abhandlung uber die Mathematik, die allgemeine zeichenkunst, und die Vershiedenheit der Rechnungsorten. 1764. Holland, G.J. Briefe an J.H. Lambert. 1981. Horvath, Kereztelo J'anos. Institutiones logicae. Venetiis: Antonius Zatate, 1782. Hundeshagen, Johannes Christophorus. Logica tabulis succinctis inclusa. 1674. Hutcheson, Francis. Fransisci Hutcheson logicae compendium. Argentorali: Impensis Joh. Frid. Stein, Bibliopolae; Excudsbat Jonas Lorenz, Typographus, 1771. Insedoorn, Gisbert abo Cursus logicus systematicus et agonisticus. Oxonii: Typis Gulielemi Hall, Sumptibus Rob. Blagrave, 1658. Jacquier, Francois. Institutiones philosophicae. Venetiis: Apud Simonem Occhi, 1785. Jakob, Ludwig Heinrich von. Grundriss der allgemeinen Logik. Halle: In Komission bey Fraube und Bispick, 1788. Jucain, Henri. La logique: ou, I 'art de penser, degage de la servitude de la dialectique. Paris: Des Ventes, 1765. Jungius, Joachim. Logica Hamburgensis. Kames, Henry Home. Introduction to the Art ojThinking. Edinburgh: 1761. Kant, Imrnanuel. Logic (Logik). Trans. Robert S. Hartman and Wolfgang Schwaz. Indianapolis: Bobbs-Merrill, 1974. Keckerman, Bartholomaeus. Gymnasium logicum, id est de Usu et exercitatione logicae artis absolutiori et pleniori, libri tres, annis ab hinc aliquot in Academia heidelbergensi privatus praelectionibus traditio Hanoviae: apud G. Antonium, 1605. Keckermann, Bartholomaeus. Systema Logicae, tribus libris adornatum pleniore praeceptorum methodo et commentariis scriptis ad

324

BIBLIOGRAPHY praeceptorum illustrationem et collationem cum doctrina Aristotelis atque aliorum...Editio 3a. Hanoviae: apud G. Antonium, 1606. Kesler, Andreas. De Consequentia Tractatus Logicus. 1623. Kesler, Andreas. Logicae photoniane examen, Principiorum logicorum quae in Photinianorum scriptis occurant consideratio, cui praemissus est Tractatus... de illegitimo Photinianorum disputandi modo. Wittebergae: impensis haered. I. Bergeri, 1663. Lambert, Jean Henri. Abhandlung vom Criterium veritatis. Lambert, Jean Henri. Anlage zur Architectonic. 1711. Lambert, Jean Henri. Neues Organon. 2 vols. Leipzig: J. Wendler, 1764. Lambert, Jean Henri. De Universaliori calculi idea, disquistio, Nova Acta Eruditum. 1765. Lambert, Jean Henri. Logische und philosophische Abhandlungen. 1782-87. Lange, I.C. Neucleus logicae Weisanae. 1712. Lange, 1. C. Inventum novi quadrati logici universalis in trianguli quoque formam commode redacti. 1714. Launay, Gilles de. Dissertation de la philosphie en general. Paris: C. Barbin, 1668. Launay, Gilles de. La Dialectique. Paris: C. Barbin, 1673. Leibniz, Gottfried Wilhelm von. General investigations concerning the analysis of concepts and truths. Athens, Georgia: University of Georgia Press, 1968. Lentulus, Cyriacus. Nova R. Descartes sapientia faci/ori quam antebac methoda detecta. Herbornae Nassoviorum: n.p., 1651. Lever, Ralph. The Arte ofReason(1573). Merston: The Scalar Press, 1972. Libavius, Andreas. Singularium. Francofurti: 1. Saurii, impensis P. Kopfj, 1599-1601. Lipstorp, Daniel. Specima philosophiae cartesianae... Lugduni Batavorum: 1653. Lossada, Luis de. Institutiones dialecticae. Salamonticae: Fransisci Garcia Onocata & San Miguel, 1721. Maass, Johann Gebhard Ehrenreich. Grundriss der reinen Mathematik. Leipzig: 1796. Maass, lG.E. Grundriss der Logik. 1807. Maimon, S. Versuch einer neuen Logik. Leipzig: 1798. Mang, Theodor O. Institutiones logicae et metaphysicae. Parisiis: Apud Crapart; Tolosae: Apud Dupleix, 1789. Maredo, Benito de. Pharus dialectica. ex Typographia Academiae. 1720. Mariotte, Edme. Essai de logique. Paris: E. Michallet, 1678. Marsh, Narcissus. Institutiones logicae. 1681. Martinus, Cornelius. De analysi logica tractatus. 1619. Melles, Etienne de. Phi/osophia... Parisiis: Apud Edmundum Conterot et Fransiscus Le Cointe, 1678.

Bibliography 325 Monteiro, Ignacio. Ars critica rationis dirigendae: seu philosophica humane mentis institutio, logica communi usu nuncupata. Venetiis: Typis Antonii Zatta, 1777. Moulin, Pierre Du. Elements de logique. Sedan: 1621. Musschenbroek, Petrus von. Institutiones logicae. Neapoli: Expensis Ignatii Gessari, 1758. Newton, John. An Introduction to the art of Logick Composed for the Use of the English Schools, The second edition enlarged and amended by the author. London: T. Passinger, 1678. Oliva, Johann Paul. Theses logica ac physica... Ingolstadij: Excudebat Alexander Vueissenhom, 1568. Olivier, Pierre. De inventione dialectica libellus.Parisiis: apud P. Vidoveum Vernoliensem, 1540. Ortiz, Didacus. Logicae brevis explicatio. 1650. Pare, Giulio. luli) Pari) a Beriza Doctrinae Peripateticae tomi tres... Aureliae Allobrogum: Excudebat Petrus de la Rouiere, 1606. Paulo, Eustachius a Sancto. Summa philosophiae quaddripartita. Coloniae Allobrogum: Typis Phillippi Alberti, 1638. Petit, Pierre. De Nova Renati Cartesii philosophia dissertationes. Paris: 1670. Platner, Ernst. Lehrbuch der Logik und Metaphysik. Leipzig: 1m Schroicherten Verlage, 1795. P1oucquet, G. Fundamenta philosophiae speculativae. Tubingae: 1759. Ploucquet, G. Untersuchung und Abandrung der logicalischen constructionem des Herrn Professor Lambert. Tubingae: 1765. P1oucquet, G. Sammlung der Schriften welche den logischen Calcul Herrn Professor Ploucquet 's betroffen. Tubingae: 1766. Ploucquet, G. lnstitutiones philosophiae theoreticae sive de arte cogitandi. Tubingae: 1772. P1oucquet, G. Elementa philosophiae contemplativae, sive de scientia ratiocinandi. Tubingae: 1778. Ramee, Pierre (Petrus Ramus) de lao Dialectique. Paris: A. Wechel, 1555. Ramee, Pierre (petrus Ramus) de La. Dialecticae libri duo. Parisiis: apud A. Wechelunl, 1560. Regis, Pierre-Sylvain. Systeme de philosophie contenant la logique, la metaphysique, la physique, la logique et la morale. Paris: impr. de D. Thierry, aux depens d' Anisson, Posuel et Rigaud, libraires a Lyon, 1690. Regnault, Noel. Logique en forme d'entretiens ou l'art de trouver la verite. Paris: Clousier, David fils, Durand et Danl0neville, 1742. Reid, Thomas. "A Brief Account of Aristotle's Logic." The Works of Thomas Reid. Ed. William Hamilton. Edinburgh: Maclachlan and Stewart, 1880. 2: 681-713. Reimarus, Hermann Samuel. Vernunfilehre. HamburgfBohn: 1790.

326 BIBLIOGRAPHY Reusch, Johann Peter. Systema logicum antiquorum atque receptorum item propria exhibens (1734). Wolff, Christian, Freiherr von, 1679-1754, Works, 1962; Abt.3, Bd.6. New York: G. Olms, 1990. Rubio, Antonio. Logica mexicana. Lugduni: Sumptibus Antonii Pillehotte, 1620. Saint-Joseph, Dom Pierre de. Idea philosophiae rationalis. Coloniae Agrippinae: Apud Contantinium Munich, 1655. Saller, Johann. Diputatio philosophica de ente rationis logico. Ingolstadii: Ex officina typographia Davidis Santorii, 1595. Sanderson, Robert. Logicae artis compendium. Oxoniae: execud. 1. Lichfield et 1. Short, 1618. Scharfius, Johannes. Institutiones logicae. 1632. Scheibler, Christoph. Liber commentariroum topicorum... Oxoniae: Excudebat H. Hall, inlpensis 1. Godwin, 1. Adams, & E. Forest, 1653. Scheib1er, Christophorus. "Epitoma logica." Opera philosophica. Francofurti: sumptibus B.C. Wustii, typis II. Friesi, 1665. Schoock, Marten. Collegium logicum, compendiose non minus exhibens praecepta et controversias logicas quam utrorumque praxin. Groningae: typis Vae E. Agricolae, 1658. Schuster, Friedrich. Dissertatio de recentiorum quarta mentis operatione sive methodo. Bamberg: Typis Joannis Georgii Klietsch, 1750. Schweling, Johann Eberhard. Clavis scientiarum. Bremae: Typis & impensis Hermani Braueri, 1678. Secchi, Ferdinando Maria. Logicae ac metaphysicae institutiones. Augustae Taurinorum: Ex Typographia Regia, 1741. Segner, Johann Andreas von. Specimen logicae universalites demonstrate. Jena: 1740. Seton, John. Dialectica. London: 1611. Smiglecius, Martinus. Logica. 1618. Spencer, Thomas. Logicke unfolded. London: by W.H. for Nicolas Bourn, 1656. Spencer, Thomas. The art oflogic (1628). Merston: ScolarPress, 1970. Stahl, Daniel. Axiomata philosphica sub titulis A:¥ comprehenda. Cantabrigiae: Ex officia Rogeri Daniel, 1645. Stahl, Daniel. Doctrina propositionum disputationibus XII, comprehensa. Londoni: Ex officina Rogeri Danielis, 1651. Sterne, Richard. Summa logicae. Londini: Typis S. Roycroft, Impensis R. Clave!l, 1685. Stier, Johann. Praecepta doctrinae logicae, ethicae, physicae, metaphysicae, sphaericaeque, brevibus tabel/is compacta. Cantabrigiae: ex officina Rogeri Danielio, 1647. Stier, Johann. Praecepta logicae peripateticae... lena: sumtibus M. Birckneri, 1662.

Bibliography 327 Storchenau, Sigmund von. Grundsatze der Logik. Augsburg: Bey den Gebrtidern Xeith, 1775. T.G., (attributed to Thomas Gove or Thomas Gowan). Ars sciendi sive logica: nova methodo disposita, et novis praeceptis aucta. London: Typis Johannis Richardson, prostat apud Thomae Parkhurst, 1681. Tamugna, Giuseppe. Institutiones logicae et metaphisicae. Romae: Ex Typographis S. Michaelis, apud Paulum Junchi, 1778. Thomasius, Christian. Einleitung zur Vernunftlehre. Halle: gedmckt bey C. Salfelden, 1691. Timpler(us), Clemens. Logicae jystems methodicum, libris V comprehensum, in quo universus ben disserendi et sciendi modus...per praecepta et questiones... explicatur et probatur. Hanoviae: typis haeredum G. Antonii, 1612. Tittel, Gottlob August. Erlauterungen der theoretischen und praktischen Philosophie nach Herrn Feders Ordnung: Logik. Mannheim: Schwan und G6tz, 1790. Tonnies, I.H. De logicae scientia ad exemplar arithmeticae instituenda ratione. 1752. Tooke, John Horne. Epea pteroenta, or the Diversions of Purley. London: the author, 1786. Trieu, Philippe duo Manuductio ad logicam. Duaci: Typis Baltazaris Belleri, 1616. Valthuysen, Lambert Van. De initius primae philosophiae junta fundamenta Cartesii. Utrecht: 1662. Vater, 1.S.Pasigraphie und Antipasigraphie; oder uber die neueste Erjindung einer allgemeinen Schriftsprache fur aIle Volker, und von Wolkens, Leibnitzens, Wilkins's und Kalmars's pasigraphischen Ideen... Weissenfels und Leipzig: F. Severin,· 1799. Vater, Johann Severin. Versuch einer allgemeinen Sprachlehre. Halle: Renger, 1801. Vicaeus, Johann. Pronunciata philosophica... Ingolstadij: Ex offcina VVeissenhornianae, 1571. Vietoris, Matthaeus. Dialectica Laurentiana. Coloniae: Sumptibus gymnasii Laurentiani, 1770. Vincent, Le P. Jean. Cursus philosophicus, in quo totius quaestiones fere omnes... in utramque partem propugnantur. Tolosae: excudebat 1. Boude, 1658-71. Vincentio, Father Gabriel a Sancto. Logica Gabriel a Sancto Vincentio. Romae: Ex typographia Philippi Mariae Mancini, 1669. Vossius, Gerardus Joannes. Gerardi Jos. Vossii de logices et rhetoricae natura & constitutione libri II. Hagae-Comitis: Apud Adrianum Vlacq, 1658. Wagner, L.H. Grundriss der reinen und allgemeinen Logik. 1806.

328 BffiLIOGRAPHY Walker, Obadiah. Artsi rationis, maxima ex parte ad mentem nominalium, libri tres. Oxoniae: ETS, 1673. Wallis, John. Institutio Logicae... 1715. Watts, Isaac. Logick, or the Right use of reason in the inquiry after truth, with a variety ofrules to guard against error in the affairs of religion and human life as well as in the sciences... The 4th edition. London: E. Matthews, 1725. Watts, Isaac. The Improvement of the mind, or a Supplement to the art of logick containing a variety of remarks and rules for the attainment and communication of useful knowledge in religion, in the sciences and in common life... the 2nd edition. London: 1. Brackstone, 1743. Weis, Dam Ulrich. Liber de emendatione intellectus humani, in duas partes digestus, verum operationum omnium intellectus theoriam, tum earundem directionum solide ediserens. Kauffburen: typis G. Starck, prostat apud authorem, 1747. Weise, Christian. Christian Weise Curieuse Fragen aber die Logica. Leipzig: In Verlegung Joh. Grossens seI. Enben: Dmckts Christian Scholvien, 1700. Wilkins, 1. Essay toward a real character and philosophical language. 1668. Wilson, Thomas. The Rule of Reason, Conteiyning the Arte of Logic(1551). Northridge, Calif.: San Fernando Valley State College Foundation, 1972. Wolff, Christian. Vernunftige Gedanken von der Kraften des Menschlicken Verstandes und ihrem richtigen Gebrauche in Erkenntnis der Wahrheit. Wolff, Christian, Freiherr von, 1679-1754, Works, 1962; 1.Abt, Deutsche Schriften; Bd.1. Hildesheim: G. alms, 1965. Zabarella, Jacopo. Opera logica. Coloniae: sumptibus L. Zetzneri, 1597.

Index

-AA Free Enquiry, 117, 135 a posterion', 193, 198, 199, 234, 235, 237,243,248 a pn'ori, 22, 35, 165, 176, 192, 193, 195,197,198,199,200,201,202, 203,205,207,212,231,234,235, 236,237,238,239,240,241,242, 243,244,246,248,249 abstraction, i, 15-16, 24, 25, 29, 40n.ll, 56-57, 90, 92, 132, 175, 219,220,232,237,239,276. See also Ideas, abstract accidents, 80, 109, 110, 124 Agricola, R., 19, 124 Aldrich, H., 85 Alnwick, William of, 48 analogy, 15,26,30,34,47,48,49,53, 54,59,103,192,198,199,200, 202,203,205,291 analysis, see method, analytic analytic, see method, analytic Anthropology, 177, 179, 189,264 apprehension, 9, 11, 14, 88, 92, 95, 96, 98, 99, 103, 199 Aquinas, 81. Thomas, 24, 25, 32,40, 48,53,62,73,109,132,133,335 Arabic, 24,40,41 Aristotle, 8, 9, 24, 25, 40, 43, 50, 54, 58, 109, 110, 127, 130,137, 138, 142,335 Arnauld, A., 11, 12, 13, 60, 65, 66, 67, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 97, 99, 100, 101, 102,106,133,144,149,191,193, 206,280,337 Ars Cogitandi, 85, 93, 94 Art o/Thinking, 139, 141, 150 Ashworth, 1., iv, 138, 150 associationism, 292, 293, 294, 298 associationist, 103,286,292,293,294, 296,298 attribute, 75, 116, 221, 271, 285, 301

Augustine, 81., 50 Augustinian, 50, 73, 77 Aureol, P., 48, 61 Averroes, 5, 8

-BBacon, F., 85, 138 Barron, W., 84, 88, 93 Beattie, 1., 285, 291, 298 Beck, L. W., 248, 260, 261 Berkeley, G., iv, 22, 42, 96, 102, 116, 119,121,122,125,133,134,135, 302,338,339 Bochenski, I. M., 2 Boyle, R., 117, 135, 141 brain, vi, 68, 89,258, 259, 264, 265, 268,270,272,280,315 Burgersdijck, F., 110, 111, 112, 114, 121, 123, 124, 125, 133, 134 Buroker, 1. V., iv, 106 Butts, R. E., vi, 9, 317

-cCabanis, P.-1.-G., 251, 252, 267 Caesar, 124 Cajetan, 49, 54, 63 Capreolus, 1., 48, 60 Careil, Foucher de, 197 Cartesian, 11, ii, iv, v, 1, 13, 14, 17, 22, 28, 35,43, 78, 85, 97, 98, 100, 102, 103, 133, 141, 150, 156,254, 256,258,259,260,265,266,269, 275,281,335,337,338 Cassirer, E., 260, 261 Categories, 49,50,62 category, 111, 112, 123, 196, 218, 221 causal inference, vi, 153, 204, 283, 284,285,286,289,291,292,293, 294,295,297,298,299,302,303, 304 causation, 120, 133, 155,212,227, 288,291

330

INDEX

cause, 7, 33, 41n.14, 53, 61n.26, 74, 76, 103, 111, 116, 121, 126, 127, 128, 137, 144, 154, 155, 164, 169, 211,212,216,224,227,283,284, 285,287,289,290,291,292,294, 295, 296, 300, 302, 316; occasional, 74, 81n.37, 88 certainty, 30, 160, 161, 169, 171, 192, 197,199,254,262,279,280 Chatton, W., 48 Church, R. W., 66 Cicero, 5, 7, 10,19 Clarke, S., 238, 241 clear and distinct, see ideas clock, 89 Clow, J., 83, 93 cogito,192 cognitive science, 280 color, 15, 34, 91, 92, 118, 120, 160161,297 common nature, 24, 52, 55, 277 comparison (mental), iv, 16, 53, 92, 113-114, 128, 130-131, 155, 157, 160,162,164,169,175,276,278 composition (mental), 11, 15, 29, 79, 144, 181-185, 236 conceiving, 31, 58, 99, 104, 144, 148, 149,164,183,212,217,227,277 concept, ii, 43, 50, 51, 52, 55, 56, 57, 58, 100, 104, 127, 175, 180, 181, 182,183,184,186,188,193,194, 195,202,206,227,231,232,233, 234,235,236,237,238,240,241, 247,248,264,308,310 conceptualizing, 104, 133 Condillac, E., ii consciousness, v, 90, 102, 146, 175, 176,177,178,179,180,181,183, 184, 185, 187, 188, 189, 232,238, 265 contingent, v, 35, 72, 129, 153, 155, 165,166,167,172,191,192,193, 194,195,196,197,198,199,200, 201,202,203,204,205,217,220, 224 contradiction, 13, 76, 164, 168, 169, 170,191,202,211,212,226,235, 247,248,301,307,308

convention, 316 conventional sign, 101 copula, 99 corpuscular, 30, 43 Couturat, L., 192, 199, 207 Critique ofPure Reason, ii, 39, 45, 73, 103,104,106,179,185,189,190, 231,239,245,261,262,267,268, 307,308,313,316 Crousasz, l-P., 150 Cummins, P., v, 316 curriculum, ii, 18,47,83, 86, 93, 138, 141, 150 custom, 222, 295

-Dd'Ailly, P., 48 Dalbiez, R., 47 De Anima, 83 De Origine, 85, 93 de Soto, D., 49, 60, 61 deduction, v, 99, 138, 145, 153, 155, 160,165,166,197,240,245,249, 274,278,285 deductive, 137, 138, 145, 148, 149, 153,156, 157, 158, 161, 162, 163, 164,165,166,170,171,172,197, 205,279,298,308,311,313,316 Defensiones theologiae, 48, 60 definilion, 8, 11,14,17,103,129, 137,143,185,191-193,196,199, 200-204,221,254; nominal, 17, 193, 195; ostensive, 103; quidditative, 52; real, 17, 129, 133, 137,193,195 demonstration, v, 9,30,137,138,153, 154, 155, 156, 158, 159, 160, 161, 162, 163, 165, 166, 167, 168, 169, 170,171,172,174,195,199,200, 202,211,212,213,229,285,304; legal, v, 123, 124, 245 Descartes, R., i, iii, v, 1, 10, 12, 21, 22,26,27,28,29,32,34,35,36, 37,39,42,44,47,50,59,60,78, 80,85,98,99,100,101,103,104, 105, 106, 107, 122, 130, 132, 138, 156,158,252,253,254,255,256,

Index 258,266,269,270,271,272,273, 274,275,276,277,278,279,280, 281, 301, 335, 337 dialectic, 5, 6,19,51,126,131,307, 313 Dialecticae Libri, 7 Dialectique, 7,19,124,125,126,135 Didiscalicon, 18 Difference Principle, 218 Digby, W., 126 Discours de metaphysique, 193, 197 discursive, 5, 6, 9, 132, 144 disposition, 7, 129, 256, 312 disputation, 4, 6, 10 Disputationes Metaphysicae, 60 distinctness, i, 17, 180, 181, 224, 226 Dominic, of Flanders, 49, 60, 61, 63 Drununond,C., 140, 141 dualism, 73,97,99,255,256 Duchesneau, F., v, 191 Durandus, of Saint Pouryain, 48,59

-EEchelbarger, C., v, 137 Edinburgh, 86, 93, 94,139, 141, 142, 150,171,302,303,338 effect, 19, 30, 72, 73, 81, 111, 119, 123, 124, 125, 150, 155, 164, 201, 211,224,226,227,254,258,283, 284,290,292,295,297,300,302, 317 Emile, 43 eminent, 245 empirical psychology, 33, 35 empiricism, ii, 39, 85, 146, 197, 229, 252,260,263,339 enunciation, 7,159,192,193 Epicurean, 11, 14, 17, 93 Epicurus, S., 11 epistemology, i, iii, 1, 18,23, 31, 33, 35,43,48,187,197,252,265,274, 275,276,279,280,286,317,335 Erasmus, 39, 315 Erotemata Dialecticum, 7 error, vi, 14, 17, 43, 66, 67, 87, 94, 143, 146, 153, 199, 200, 258, 309, 315

331

Essay on the Human Understanding, ii, 29, 30, 43, 45, 85, 134, 138, 140, 143,156,157,172,225,226,285, 286,287,293,294,297,300,301, 302 esse obiective, 49, 61 essence, 21-24, 26, 29-37, 52, 54, 56, 57,70, 110, 115, 119, 131,212 Euclid, 233, 235 experience, 16, 22, 33, 98, 116, 129, 140, 149, 164, 191, 197, 201, 204, 214,224-229,233,236-246,248, 253,257,261-265, 283, 290-291, 300,302,308,311 extension, 15, 16, 37,42, 91, 95, 96, 120,155,196,213,214,223,225, 276,277,279 externalism, 97, 104, 105 externalist, v, 102-104 extraction, 25

-Ffaculty psychology, i, iii, v, vi, 244. See also psychology fallacy, 31, 33 figure, iv, 16, 30,47, 70, 94, 115, 128, 137,234,277,279,309,310 Fonseca, P., iv, 47, 49, 50, 51, 52, 53, 54,56,57,58,59,61,62,63 formal and objective concepts, 47, 48, 49,50,55 Foucault, M., 259, 315 Foulis, R., & A., 87 Frege, G., 97, 103, 106

-GGassendi, P., 1, 9, 10, 11, 12, 13, 14, 15,16,17,18,20,85,88,93,94, 95,258,270,281,337,338 genus, 52, 56, 58, 94, 110, 111, 128, 129, 187 geometry, 10, 13, 35, 37, 173, 196, 199,309,310,313,315 Gerhardt, C. 1., 197, 206 Ginsberg, M., 66 Glasgow, 11, 83, 86, 87, 93, 94, 95

332

INDEX

God, 15,26,28,30,34,40,42,52,53, 54, 56, 58, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 80, 81, 83, 90, 98, 116, 122, 132, 135, 144, 145, 146, 162, 165, 193, 194, 197, 199, 203, 216,248,254,269,279,280,303 Goethe, 1. W., 261 Goldsmith, 0., 125 grace, 65, 74, 77, 78,81,274 grammar, i, 51, 61, 87, 110, 249

-HHartman, R., 175 Hatfield, G., iii, vi, 21 health, v, 53,252 Henrich, D., v, 244, 245 Hervaeus Natalis, 48, 60, 61, 62, 63 Herz, M., 264 historiography, 39, 263 Hobbes, T., ii, 10, 68, 149, 151, 297, 304 humanist, iv, 1, 4, 6, 7, 9, 10, 12, 18, 19,47,86 Hume, D., 11, 12, ii, iv, v, vi, 15, 22, 102, 125, 137, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 153, 154, 155, 156, 158, 159, 160, 161, 162, 163, 164, 165, 166,167,168,169,170,171,172, 173,174,211,212,213,214,215, 216,217,218,219,220,221,222, 223,224,225,226,227,248,283, 284,285,286,287,288,289,290, 291,292,293,294,295,296,297, 298,299,300,301,302,303,304, 312,336,337,338 Hutcheson, F., iv, 83, 86, 87, 88, 89, 90,91,92,93,94,95,96,338

-1ideas, abstract, i, ii, 24-25, 56, 72, 8893, 105n2., 115, 140,171,192194, 215-224, 254, 285; agreement of, 17,84,115,131,156-157; bundle theory, 285; clear and distinct, i, 14, 26, 30, 32, 44, 72,

99, 118, 180-181, 183, 216, 221, 254,255,259,273,276; disagreement of, 17, 29, 36, 65, 66, 69, 84, 156, 157; innate, 25, 29, 32, 34, 98 identity, 159, 192 illocutionary force, 104 illusion, 55, 307, 308, 310, 311, 312, 315,316 image, 11, 14, 15, 24, 27, 68, 89, 277, 297; material image, 24 imagination, 14,24, 25, 26, 27, 28,41, 149,171,218,222,224,226,262, 269,270,272,275,276,277,279, 281,283,292,295,296,297,302 impossibility, v, 105, 161, 162, 164, 200,211,212,213,214,216,217, 220,227 impossible, 33, 71,99, 160, 167, 169, 181,194,200,204,211,214,215, 216,217,218,220,221,223,224, 227,231,232,233,234,238,257, 297,300 impression, 25, 40, 89, 95, 146, 212, 214,215,217,226,274,284,288, 297 In P. Rami Dialecticam, 125 induction, vi, 283, 284, 289, 290, 291, 292,294,296,297,298,300,303 inductive, ii, 15, 139, 145, 148, 167, 298 inference, v, vi, 9, 30, 31, 98, 99, 105, 128, 129, 130, 132, 144, 153, 155, 158, 163, 164, 165, 170, 180, 184, 185,186,187,193,194,204,212, 223,226,283,284,285,286,288, 289,291,292,293,294,295,297, 298,299,302,303,304 infinite regress, 203 Institutio Logica, 3, 13, 14, 17, 18,20, 85 intellect, i, 22, 23, 24, 25, 26, 27, 28, 30, 32, 33, 35, 36, 40, 41, 42, 43, 44,49, 98, 104, 109, 137, 138, 148, 203,265,269,274,276,277,278, 279, 336; active intellect, 24-26, 33, 40-41; patient intellect, 25, 41; pure intellection, 89, 90, 91, 92

Index intelligible species, 25, 26, 41, 110 introspection, 182, 183, 189 intuitive, 22, 30, 98, 99, 156, 157, 158, 160, 161, 164, 166, 168, 169, 170,171,173,180,211,212,225, 227,301

-JJasche, 175 Jesuit, 47, 59, 150 judgement, 84, 88, 113, 114, 144, 160, 161,277 judging, 12, 99, 103, 104, 106, 109, 148, 185 judgment, iv, 1,7,8,13,19,68,79, 99, 103, 111, 113, 129, 137, 143, 144, 145, 146, 147, 175, 182, 185, 186,230,236,291,309,311,335

-KKant, I., 1, 3, 5, 21, 22, 31, 33, 36, 39, 42,43,44,59,103,104,105,107, 175,176,177,178,179,180,182, 183,184,185,186,187,189,190, 226,231,235,236,239,240,241, 242,243,245,246,247,248,249, 250,251,252,261,262,264,268, 307, 308, 309, 310, 311, 312, 313, 314,315,316,317,335,336,337 Keckermann, B., 5, 6, 9, 12, 150 Kemp Smith, N., 106, 248, 286, 316 Kilkenny College, 126 King's College, 93 Kitcher, P., v Kneale, W. & M., 1 Kremer, E., iv, 79, 337 Kuehn, M., v, 267

-LLa Mettrie, J. O. de, 258, 260 Lambert, J. H., 231, 236, 237, 238, 243,248 language, 3,4,47,48, 50,54,57,62, 97, 99, 100, 101, 102, 103, 105, 106, 110, 114, 115, 116, 121, 122,

333

124, 125, 132, 133, 149, 151, 187, 213,219,220,221,248,255,258, 335 Latin, 4, 7, 20, 24, 42, 54, 58, 69, 80, 83,87,93,125,150,254,304 law, 96, 127, 129, 195, 197, 201, 202, 203,204,205 Leibniz, G. W., ii, v, 12, 22, 71, 80, 143,155,191,192,193,195,196, 197, 198, 199, 200, 201, 202, 203, 204,205,206,207,229,230,238, 241,251,254,255,260,261,266, 335,336,338 Locke, J., ii, 10, 22, 29, 30, 31, 36,43, 62,87,88,89,91,113,114,116, 117,119,120,122,128,131,132, 138, 141, 142, 149, 156, 158, 162, 171,172,214,215,222,232,236, 238,253,257,260,286,287,293, 294,297,301,302,303,304,336, 338 Lockean, 30, 84, 88, 91,113,115, 141,157,240,257,302,303 Loeb, L., vi, 283 logic, Aristotelian, iv, 7, 9, 10, 19, 21, 24, 26, 28, 29, 32, 33, 35, 40, 43, 47, 50, 86, 110, 117, 119, 123, 124, 125, 127, 128, 133, 134, 138, 140, 150,157,186,247,274,275,281; of ideas, i-vii, 2-3, 18, 84ff.; of invention, iv, 254; medieval, i, ii, 1, 2, 4, 6, 10, 12, 18,47,48, 61, 137, 335; Ramist, iv, 7, 11, 109, 125, 127, 128, 133; Scholastic, iv, 47, 48, 62; Scottish School, iv, 83, 86, 87, 88, 89, 90, 93, 94, 95, 96, 139, 142, 150, 248, 267, 337; traditional, i, 123, 127, 133, 138 Logica sive Ars, 140 Logicae Compendium, 11, 83, 85, 93 logical possibility, 240, 241 Logick, 151 Loudon, J., iv, 87, 88, 89, 91, 92, 94, 95

334

INDEX

-Mmachine, 173, 233 madness, 315 Magnus, A., 9 Malebranche, N., iv, 23, 65, 66, 69, 70, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 141, 315 Marsh, N., 125 materialism, 251, 253, 258, 260, 261, 262,263,265,267 materialist, 253, 263, 267 materialistic, v, 259, 263, 267 mathematical, 131, 162, 168, 186, 198, 200,203,204,205,254,255,270, 276,277,280,313 mathematics, 186, 202, 203, 205, 255, 271,276,278,279,281,282,338 mathesis universalis, 281 mechanism, 159, 256, 286 mechanistic, 253 medicina mentis, 251, 253, 265 Meditations on First Philosophy, 27, 42,79,80,81,230,247,248,254, 269 Melanchthon, P., 5, 6, 7, 8, 9, 19 menl0ry, 61, 100, 120, 146, 149, 172, 269,274,280,284,287,288,289, 292,295,297,303,304 mental act, 69, 70, 72, 113 Mersenne, Pere, 42, 49, 59 Metaphysical Disputations, 47 Metaphysics, 11, ii, 21,23, 29,35, 36, 39,42,43,47,49,52,73,83,86, 87, 91, 109, 117, 123, 124, 128, 133,140,150,246,252,269,280, 285,301,308,314,337,338 method, ii, v, vi, 13, 29, 37, 96, 122, 130, 156, 177, 184, 205, 222, 223, 236,244,253,254,255,257,259, 264,265,269,270,271,272,275, 278, 279~ 317; analytic, method of analysis, ii, 37, 84,191,194,195, 197-209, 235, 238, 239, 252, 254255, 265; synthetic, method of synthesis, ii, 99, 104, 177, 180, 191, 195,196,204,206,235,243,254, 263

methodology, 117, 156, 196, 206, 222, 336 Michael, F., i, iii, iv, 1 middle tenn, i, 18, 130 miracles, 81, 285 Molyneux, 91, 96 monads, 44 Monilio Logica, 110, 150 Mossner, M., 140

-NNadler, S., 65, 66, 79 natural language, 3, 4, 100, 122 naturalistic epistemology, 280 necessary connection, 129, 193, 227 necessity, 35, 139, 163, 169, 171, 176, 179,185,195,199,211,240,245, 307 Newton, 1., 141, 292 Newtonian, ii, 34, 261 Nicole, P., 12, 65, 78, 79, 99, 100, 101, 102, 138 nominalist, 16 normative, i, iii, vi, 35, 186, 252 vs. descriptive, i, 21, 33, 35, 39, 53, 204,252 nonnativistic, iii Nonnore, C., 48

-0objective being, 48, 56, 69 objective reality, 47, 98 objective validity, 240, 243, 245 Ockham, 3, 50 On Discursus, 88 On interpretation, 50 On the Analogy ofNames, 49 On True and False Ideas, 65, 66, 69, 79 Ong, W. 1., 125, 132, 133, 136 ontology, ii, iv, 22, 110, 111, 114, 116, 121, 124, 125, 129, 130, 133, 134, 220,230,231,258,336 Opera Omnia, 9, 11, 19, 20, 62, 93 opinion, 11,65,77,215,298,307,316 Organon, 83, 248

Index Owen, D., v, 153

-pPalmer, E., vi, 269 pananimism, 257, 261 parallelogram, 309, 310, 316 Pascal, B., 13 passion, 315 Passions ofthe Soul, 256 perception, 10, 11, 18, 26, 66, 70, 72, 84, 88, 89, 90, 92, 98, 101, 102, 115, 118, 146, 156, 160, 169,232, 253,254,273,274,287,288,289, 297,302 perspicuous, v, 212, 213, 214, 216, 219,220,221,222,223,227 phenomena, 117, 135, 183, 196,204, 205,253,259 Plato, 57 Platonism, 41 Port Royal, Port Royal logic, 11, ii, iv, 1,3,9,12,13,14,15,16,17,18, 85,87,95,128,131,133,269 possibility, ii, v, 34, 36, 95, 103, 104, 118, 154, 162, 164, 167, 170, 172, 176,186,192,194,199,201,204, 211-248,264,283,284,337 possible object, ii, 233 predicables, 7, 9, 12, 111, 125, 127 predicaments, 111 predicate, ii, iv, 56, 96, 110, 129, 153, 186,193,197,198,199,202,221, 263 predication, iv, 110, 111, 114, 116, 119,121,123,127,133,134 Priestly, 1., 267 principium probandi possibiltatem, v principle of charity, vi Pn·nciples ofPhilosophy, 42, 100, 105, 106,134,138,216,219,220,221, 235,252,266,299,312 probable, v, 86,119,123,129,131, 162,163,165,167,168,169,170, 171,174,194,283,284,291,295, 297,298,301,302,303 proposition, ii, 3, 15, 56,68, 69, 73, 99, 103, 104, 105, 111, 115, 128,

335

129, 131, 137, 144, 147, 148, 153, 154, 155, 158, 161, 162, 163, 164, 165,168,169,170,173,177,195, 198,199,202,203,211,262,307 psychologistic, i, 31, 33, 36, 38 psychology, i, iii, v, vi, 33, 35, 38, 44, 87,244,259,261,262, 336. See also faculty psychology

-Qqualities, ii, 14,30,91,112,113,116, 11 7, 118, 11 9, 121, 122, 132, 133, 204,205,214,215,217,218,220, 222,256,268 Quesnel, P., 80 Quolibeta, 48, 60

-RR. W. Church, 66 Ramus, 1, 5, 7, 116, 122, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135,251, 305 Ramus and Talon Inventory, 135 rational, 22, 23, 32, 44, 84, 137, 192, 194,196,197,198,199,200,201, 205,255,311,315 rationalism, ii, 169, 197, 263, 275 real possibility, 240, 244 Recherche de la Verite, 80, 143 reflection, 22, 23, 29, 34, 35, 38, 67, 68, 70, 90, 118, 132, 175, 179, 188, 235,299 Reflexions, 80 Regulae, 280, 281 Reid, T., iii, 21, 39, 78, 221, 226 resemblance, 161, 222 rhetoric, 19, 35, 269 Risse, W., 19 Rorty, R., 21, 43 Rubio, A., 41, 49 Rules for the Direction ofthe Mind, vi, 12, 99, 105, 135,254,269, 270, 271,275,278,279,280,281,282, 290

336

INDEX

-8Sanderson, R., 85, 123, 124, 133 Sauvage, F., 317 sceptical, vi, 117, 205, 298 Scheibler, C., 142, 150 schematize, 200 Schneider, M., 204, 209 science, 3, 4, 6, 20, 23, 33, 51, 67, 83, 117,122,130,137, 143, 144, 146, 149,185,195,196,199,206,229, 230,244,254,255,260,280,292, 307,308,317,335,336,338,339 scientia, 3,26, 31, 51, 151 Scotus, D., 47, 48, 55, 59, 62 sensation, vi, 86, 88, 90, 92,212,237, 238,239,253,267,270,274,281, 310 sensibilia, 92 sensibility, 177, 182, 189, 259, 261, 308, 310, 311, 315 sensible qualities, 204, 205 sensory, vi, 22, 24, 25, 26, 27, 28, 29, 30, 34, 37, 40, 43, 97, 102, 106, 107,177,196,253,254,268,269, 308,315 Sentences, 48 Sergeant, 1., 138 signification, 52, 101, 291 signs, 3, 10,28, 50, 51, 100, 101, 109, 116, 119, 120, 257 singular, 14, 15, 16, 196,269 skeptic, 245, 246 Smiglecki, M., 125 Socrates, 57 Soncinas, P., 48, 60, 61, 63 soul, 26, 33,41,44, 68, 71, 74, 138, 149, 178, 189, 221, 254, 256, 260, 262,315,317 space, 31, 33, 38, 97,119,310,311 Spinoza, B., 22, 39 St.J\ndrevvs,iv, 84,86, 87,88, 89, 90, 91, 93, 94, 95 Steuart Library, 141 Stove, D., 153, 165 Suarez, 41, 48,49, 59,60,61,62 substance, ii, 22, 28, 33, 43, 53, 54, 55, 56, 57, 58, 79, 97, 105, 109,

110,113,115,116,117,119,121, 122, 123, 124, 127, 128, 131, 132, 133,134,194,195,206,214,215, 217,218,221,227,245,249,259, 268,279 substantial fonns, ii, 26 Summa Logicae, 18 supposilion, 52,122,170,218,226, 257 syllogism, i, 7, 9,22, 94, 129, 131, 133,137, 138, 153, 156,254,278 middle tenn, i syllogistic, ii, 7, 130, 131, 137, 138, 145,148, 151, 153, 155, 156, 157, 278 Systema Logica, 9

-Ttabula rasa, 14 Temple, W., 125 tenn(s), i,ll, 18, 24, 39, 43, 52, 53, 54, 55, 110, 112, 113, 129, 130, 153,158,171,174,186,189,201, 240,244,245,262,266 Tetens, 1. N., 231, 238, 239, 241, 244, 248 textbook,i,3,24,50, 150, 186,235 The Development ofLogic, 1, 18 theodicy, 65, 66, 73, 77, 79, 80 theory of ideas, 2, 12, 14, 39, 85, 270, 293 Toledo, F., 43 Transcendental Deduction, v transcendental idealism, 188, 261, 308 Treatise ofHuman Nature, 65, 77, 81, 139, 141, 142, 143, 147, 149, 150, 159,170,171,220,224,256,266, 283,292,298,299,300,302,303, 304; An Abstract ofa Treatise of Human Nature, 142, 292, 293; Appendix, 147, 173n.19 Treatise ofMan, 256 Treatise ofNature and Grace, 65 triangle, 16, 37, 70, 98, 158, 216, 233, 234,273 Trinity College, 125, 135

Index truth, 2,5,6, 10, 11,32,65,71,90, 9~ 103, 105, 10~ 110, 113, 120, 127,129,130,131,132,144,168, 169,186,191,192,193,195,196, 197,198,202,203,211,217,225, 262,273,275,277,278,307 Tschirnhaus,266

-uunity of consciousness, 175, 184 universal logic, 8 universals, i, ii, 16,26, 90, 137 universities, 47,86,94, 138, 139 university, 2, 86, 87, 95, 140, 142, 314 univocal, 52, 53, 54, 55, 57, 58

-vValla, L., 7 VanGulick,R., 177, 178, 188, 189 Vanity ofDogmatizing, 149 Vives, J. L., 4

337

volition, 72, 74

-wWatts, I., 144, 151 wax, 28, 274 Wells, N., 48 Wilson, C., v, 251 Wilson, F., iv, 109 Winkler,K., 116, 117 Wittgenstein, L., 103 Wolff, J., v, 33, 229, 230, 232, 233, 234,235,236,237,240,241,244, 246,247,250

-yYolton, J., 257

-'ZrZabarella, J., 5, 8, 9