Moral Psychology Today: Essays on Values, Rational Choice, and the Will (Philosophical Studies Series - Volume 110)

MORAL PSYCHOLOGY TODAY

Philosophical Studies Series VOLUME 110

Founded by Wilfrid S. Sellars and Keith Lehrer Editor Keith Lehrer, University of Arizona, Tucson Associate Editor Stewart Cohen, Arizona State University, Tempe Board of Consulting Editors Lynne Rudder Baker, University of Massachusetts at Amherst Radu Bogdan, Tulane University, New Orleans Marian David, University of Notre Dame John M. Fischer, University of California at Riverside Allan Gibbard, University of Michigan Denise Meyerson, Macquarie University Franc¸ois Recanati, Institut Jean-Nicod, EHESS, Paris Mark Sainsbury, University of Texas at Austin Stuart Silvers, Clemson University Barry Smith, State University of New York at Buffalo Nicholas D. Smith, Lewis & Clark College Linda Zagzebski, University of Okalahoma

The titles published in this series are listed at the end of this volume.

MORAL PSYCHOLOGY TODAY Essays on Values, Rational Choice, and the Will

Edited by

DAVID K. CHAN University of Wisconsin, Stevens Point, WI, USA

David K. Chan University of Wisconsin – Stevens Point Department of Philosophy 2100 Main Street Stevens Point WI 54481 USA

ISBN: 978-1-4020-6871-3

e-ISBN: 978-1-4020-6872-0

Library of Congress Control Number: 2007939559 c 2008 Springer Science+Business Media B.V.  No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work.

Printed on acid-free paper 9 8 7 6 5 4 3 2 1 springer.com

Acknowledgements

All of the papers in this volume were presented at the Conference on Values, Rational Choice, and the Will held at the University of Wisconsin – Stevens Point in April 2004. A total of close to a hundred original papers were presented over the course of three days. The conference and the publication of the selection of papers in this volume were made possible by the kind support of the Interim Chancellor Virginia Helm, College of Letters and Science Dean Justus Paul, and the Department of Philosophy under the leadership of Professor Donald Fadner. Department secretary Carolee Cote helped in innumerable ways before and during the conference. Professor J David Velleman gave the opening keynote paper ‘The Centered Self’ at the conference, which is now published in his Self to Self (Cambridge University Press, 2005) and therefore not included here. Floor Oosting and the editorial staff of Springer have been very helpful and professional in guiding me through the process of getting this volume reviewed and out in print. Laurie Winship is responsible for compiling the index for the book.

v

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David K. Chan

1

Part I Meta-ethics of Values 1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism . . . . . . . . . . 17 Drew Khlentzos 2 Buck-passing Personal Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Toni Rønnow-Rasmussen Part II Reason and Choice 3 Volitions, Comparative Value Judgments, and Choice . . . . . . . . . . . . . . 55 Joseph Boyle 4 The Insignificance of Choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 J.S. Biehl Part III Desire and Intention 5 The Indeterminacy of Desire and Practical Reason . . . . . . . . . . . . . . . . . 95 Patrick Fleming 6 The Myth of Objectively Alien Desires . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 James Stacey Taylor 7 On the Intelligibility of Bad Acts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Michael Stocker vii

viii

Contents

8 After Anscombe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 David K. Chan Part IV Practical Reason 9 The Limits of Teleology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Crystal Thorpe 10 Motivation to the Means . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Stephen Finlay 11 Thresholds, Vagueness and the Psychology of Small Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Erik Schmidt Part V Freedom and Moral Agency 12 Taking Liberty with Humean Necessity: Compatibilism and Contingency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Troy L. Booher 13 Rational Choice and Evolutionary Fit . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 Malcolm Murray Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

Contributors

J. S. Biehl earned his Ph.D. in philosophy from the Graduate School of the City University of New York. He has taught at Baruch College and at University College, Cork, Ireland. His current work is concerned with the significance of normative nihilism for philosophical activity. Troy L. Booher is an appellate attorney in Salt Lake City at the law firm of Snell & Wilmer, L.L.P.; Adjunct Professor at the S.J. Quinney College of Law (University of Utah); Adjunct Professor in the Department of Political Science at the University of Utah; and Adjunct Instructor in the Department of Philosophy at the University of Utah. His recent publications include Drugs and Justice: Seeking a Consistent, Coherent, and Comprehensive View (Battin et al (co-authors), Oxford University Press, 2007), and ‘Putting Meaning in Its Place: Originalism and Philosophy of Language,’ Journal of Law and Philosophy 25 (2006). Joseph Boyle is Professor of Philosophy, The University of Toronto; Fellow and Former Principal (1991–2002) at St. Michael’s College; and Member of the Joint Centre for Bioethics. Boyle received his BA from LaSalle University in Philadelphia, and his Ph.D. in Philosophy from Georgetown. He does research in the area of moral philosophy, particularly in the Roman Catholic moral tradition, and has collaborated with Germain Grisez and John Finnis in developing and applying a distinctive version of natural law theory. David K. Chan is Associate Professor of Philosophy at the University of Wisconsin – Stevens Point. He studied philosophy at the University of Melbourne in Australia and at Stanford University (Ph.D.). His research interests are in moral psychology, virtue ethics, applied ethics and Greek philosophy, and he is working on Action Reconceptualized, a book on human agency. Stephen Finlay is Assistant Professor of Philosophy at the University of Southern California. His research interests are in meta-ethics, ethics, and moral psychology.

ix

x

Contributors

Patrick Fleming received his PhD from the University of Connecticut in 2006. He is currently a visiting assistant professor at James Madison University. His work focuses on meta-ethics, practical reason and philosophy of action. Drew Khlentzos teaches philosophy in the School of Social Science at the University of New England in Australia. He is the author of four books, most recently Naturalistic Realism and the Antirealist Challenge (MIT Press, 2004). He has interests in logic, philosophy of logic, philosophy of language, cognitive science and philosophy of mind, and is a member of the Language and Cognition Research Center. Malcolm Murray is Associate Professor of Philosophy at the University of Prince Edward Island, Canada. His publications include The Moral Wager: Evolution and Contract (Springer, 2007), Liberty, Games, and Contracts (ed, Ashgate, 2007), Critical Reflection with Nebojsa Kujundzic (McGill-Queen’s, 2005), and articles on contractarianism in Canadian, American, Australian and European philosophy journals. Toni Rønnow-Rasmussen teaches in the Department of Philosophy at Lund University in Sweden. He is the author of a book about R. M Hare’s moral philosophy, co-editor of 4 books, and has published articles notably in value theory and metaethics. He is also an Associate Editor for the journal Ethical Theory and Moral Practice. Erik Schmidt is Assistant Professor of Philosophy at Gonzaga University. He completed his Ph.D. at Syracuse University in 2003. Erik has written on the topics of moral psychology, rationality, and the philosophy of art. He is currently writing a book entitled Beyond Comparison, which explores the limits of economic models of valuation and choice. Michael Stocker is Irwin and Marjorie Guttag Professor of Philosophy at Syracuse University. He teaches and writes on ethics and moral psychology. He is the author of Plural and Conflicting Values (Oxford University Press, 1990) and Valuing Emotions (Cambridge University Press, 1996). James Stacey Taylor is Assistant Professor of Philosophy at The College of New Jersey. He has special interests in medical ethics, ethical theory and action theory. He is the author of Stakes and Kidneys: Why markets in human body parts are morally imperative (Ashgate, 2005), and editor of Personal Autonomy: New essays (Cambridge University Press, 2005). Crystal Thorpe has a Ph.D. from Stanford University, and has taught at the University of Florida. Her research interests include practical reasoning, moral theory and bioethics.

Introduction Moral Psychology Today David K. Chan

Philosophers working in the field of moral psychology concern themselves with issues and questions that straddle across more than one sub-discipline in philosophy. The answers that they propose and the theories that they propound have implications for moral philosophy, the philosophy of mind, epistemology and metaphysics. Philosophers with an interest in the ‘will’ or the motivational psychology of human action divide between those investigating psychological and behavioral concepts for their own sake, and those working on action theory for the sake of answering larger questions, especially ethical ones. The essays in this volume, all of which were presented at a conference on Values, Rational Choice, and the Will, belong largely in the latter category of research. The essays here contribute to debates and discussions on a wide range of topics that are of current interest in moral psychology. These essays demonstrate both the breadth of the field of moral psychology and the importance of the questions investigated. Artificial boundaries between sub-disciplines of philosophy are crossed in putting these essays together in the same anthology. There are essays here in meta-ethics and in normative ethics. There are essays in ethical theory and in applied ethics. There are essays that draw on discoveries in empirical psychology. There are essays that draw on the history of philosophy and those that analyze concepts and arguments. There has been a tendency in some circles in analytic philosophy to examine concepts and arguments apart from the historical context of these ideas. But not only does the philosophy of action have a long and illustrious history that began with Aristotle, with subsequent contributions along the way from Aquinas, Hume, and Kant, but recent philosophers such as Anscombe have borrowed and developed ideas from the history of philosophy. Some of the essays here reflect a renewed interest in the work of Anscombe among philosophers of action. Moral psychology today is progressing in several directions at once. Readers of this volume will be exploring a domain of philosophy that is both unified and D.K. Chan Department of Philosophy, University of Wisconsin – Stevens Point, Stevens Point, WI 54481, USA, Tel.: 1(715) 346-3735, Fax: 1(715) 346-4215 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

1

2

D.K. Chan

diverse. I shall introduce the papers here by articulating what moral psychology is, and describing some of its many directions of research. I will then highlight the new ideas presented in this anthology and discuss how the essays here contribute to questions and issues that engage philosophers in moral psychology today.

1 From Philosophy of Action to Moral Philosophy At the heart of moral psychology is the philosophy of action, narrowly construed, as a branch of the philosophy of mind having to do with ostensible behavior. Philosophers of action have focused on two central questions (see p. 1 in Mele 1997). (1) What are actions? That is, how do we distinguish human actions from other things that humans do, such as digesting their food? (2) How are actions to be explained? That is, can human actions be explained in the same way that we explain other events that occur in the natural world, such as planetary motion, earthquakes, or the behavior of insects? Within the framework of these questions lie a number of further sub-topics. Under (1), questions arise as to what intentional actions are, and how these are to be distinguished from other kinds of action. Under (2), questions arise as to the status of explanation by reasons, and how such explanations are related to causal explanations. Answers to these questions are inter-related. The most popular approach in recent philosophy of action takes the form of a causal theory of action that accounts for the nature of action in terms of mental antecedents such as beliefs, desires and intentions. The theory also explains action in terms of some of the very same causes of action. Philosophers of action do conceptual work defining the psychological items necessary for human agency, and articulating the relation between these concepts and the various categories of action, such as trying, intentional action, and unintentional action. They give accounts of the difference between rational and irrational action, and explain how each of these is possible. They explore borderline and problematic examples of action, such as weakness of will, the role played by luck in cases when success is unexpected or in cases known as causal deviance, and the implication of beliefs regarding the outcome and effects of action. They examine whether actions are caused in the same way that non-action events are caused, and whether the story of how an action is caused is necessarily the story of how the action is rationally justified. The work of philosophers of action draws on but is distinct from empirical studies in psychology. Without the distinction, doubts could be raised about whether moral psychology is possible.1 Normative ethics is concerned with ideal standards of feeling, thinking and conduct, whereas empirical psychology is concerned with facts about the nature of cognition, reasoning, and behavior. The conceptual work of philosophers suggests ways to think about the agency of human beings. Clearly, such analyses have to be sensitive to the facts about the human psyche, at the micro- and macro-levels. But what the facts are partly depends on the concepts

Introduction

3

used to describe and interpret them, and the facts underdetermine the theories of philosophers. For instance, does a person desire to do everything that he intentionally does? The answer depends on our concept of desire. Or how many actions does an agent perform when he moves a lever to activate a pump? The answer depends on how we individuate and count actions. Philosophers of action contribute to the broader issues of moral psychology when they turn their attention to the implications of their theories and concepts for moral philosophy. A number of questions of interest to action theorists are relevant to ethics. What kind of freedom an agent has regarding whether to act, and what kind of actions an agent can be held responsible and accountable for? What do the agent’s actions tell us about her moral character, and what do they contribute to the formation of character? How are moral choices made, and are moral choices always the rational ones to make? Is moral reasoning a form of instrumental reasoning, and is the goal of moral action always something good? What are values, and how are they transmitted or reflected in the agent’s choices? How are the causal antecedents of action such as desire, intention and choice relevant for moral evaluation? How do we evaluate an action motivated by a desire that the agent prefers not to have? How does an agent constitute her practical identity, and what sort of self-awareness is required for agency?

2 Current Issues in Moral Psychology Moral psychology today is a rich and exciting field of research, providing a variety of answers to the questions listed above. The authors in this volume make contributions to ongoing debates by re-working earlier views, challenging popular views, and presenting their own alternatives. In the following chapters, essays are organized around the concepts and issues in moral psychology that they focus on. The first essays (Khlentzos, Rønnow-Rasmussen) are concerned with what values are. In moral discourse, it is assumed that there can be meaningful discussion of values. Are they real and objective? If values are real, should we think of them as natural or invented? If there are objective values, are personal values objective? Since thinking that something is of value is one thing, and acting to bring about what is of value is another, the concept of choice provides a crucial link between thought and action. A moral agent is one whose actions are chosen for the sake of what she values. The papers in this volume on choice are concerned with whether and how an agent can choose between competing values (Boyle), and whether a person’s choice is sufficient for the moral evaluation of the agent (Biehl). Concepts of desire and intention are introduced in causal theories of action as mental antecedents with distinctive roles in human agency. As a motivational state that humans share with many animals, desire seems a bit mundane and the concept of desire has not been given enough attention in the philosophy of mind. But there are problems with the concept when applied in ethics. In the section on desire, the papers deal with the lack of determinacy for the purposes of practical reasoning

4

D.K. Chan

(Fleming); the difference between desires that agents own up to and ‘alien’ ones (Taylor); and whether desires for what is bad are possible and intelligible as reasons for acting (Stocker). Philosophers of action working on theories of intention in recent years have been successful in showing intention’s role in practical reasoning and action. To what extent is this concept of intention suited for use in ethical theory? The question to be answered is whether the concept can be used to distinguish between right and wrong actions (Chan). Desire and intention both have roles in practical reasoning, and intention is itself an output of practical reasoning. Practical reasoning is today an important topic in ethics,2 and there is a debate between Humeans who see moral motivation as desire-based and others who think that practical reasoning can provide reasons for acting independent of desire. Another debate concerns the possibility of incommensurable choices: whether every alternative is comparable for the purpose of practical reasoning, or whether there are moral conflicts irresolvable by reason.3 Contributors in this volume challenge the assumption that what drives practical reasoning are goals that are set by either a desire or an intention (Thorpe); defend a Humean ‘motivation-by-desire principle’ against rationalist objections (Finlay); and argue that incomparability is a misdiagnosis of why human choices do not conform to requirements of rational choice theory (Schmidt). The relation between moral responsibility and freedom of the will has been the focus of a lot of work in moral psychology. In this volume, Hume’s compatibilist view of free action is examined and defended (Booher). The view of rational choices as self-interested ones poses a serious challenge to moral theorists who need to show that moral action is justified by reason. As an alternative to the contractarian response to this problem, an evolutionary model of moral choice is presented here (Murray).

3 The Essays in This Volume The essays here present new answers to many of the questions arising in contemporary moral psychology. As a guide for readers, the contributions of each author are briefly described below, and placed in the context of the issues that they each address.

3.1 The Nature of Values Moral actions reflect moral judgments. In other words, moral agents do what they judge to be right. Moral judgments are evaluative judgments, that is, they convey the agent’s view concerning the value of doing a thing rather than another. The appreciation of value has normative or action-guiding force on the agent. Meta-ethicists have puzzled about the nature of values. Are these special kinds of facts, and could the reality of values be reconciled with a naturalistic metaphysics? Mackie (1977)

Introduction

5

has argued that moral norms and values are too ‘queer’ to be part of the furniture of the world. Drew Khlentzos in ‘Moral Realism, Meta-Ethical Pyrrhonism and Naturalism’ points out that moral naturalists have to deal also with an epistemological problem: If there is a natural property of objective impermissibility, and if folk moral judgments can be correctly formed, then ‘the folk must unwittingly detect and respond to the relevant natural moral properties.’ The question of how they do this is called the Hard Problem. Khlentzos examines Nagel’s form of moral realism (Nagel 1997) that is grounded in the objectivity of moral reasoning: The practice of first-order moral evaluations would be incoherent without believing in objective norms and values, and we should prioritize these over second-order moral reflections about how things really are with these values. Taken as an inference to the best explanation (of our inability to think of first-order moral evaluations as anything but objective), Nagel’s argument is not fallacious. But Khlentzos suggests that the best explanation is not moral realism but what he calls ‘Meta-Ethical Pyrrhonism’, the view that ‘neither the available arguments for moral realism nor the arguments against are in the end rationally compelling.’ He argues however that unlike a religious agnostic, the Meta-Ethical Pyrrhonist should judge and act in accord with folk morality, for he shows that ‘it will be irrational to treat moral norms in the context of decision-making as anything other than fully objective.’ Toni Rønnow-Rasmussen in ‘Buck-passing Personal Values’ makes use of the so-called ‘Buck-passing’ account of value4 to provide an analysis of personal values, the kind of value found in a person’s cherished objects for instance. As applied to values in general, the account that he calls the Fitting-attitude analysis holds that ‘being valuable is being a fitting object of a pro-attitude.’ The analysis reduces the evaluative to deontic claims about attitudes towards the valuable objects that we have reason to take up towards the objects. Rønnow-Rasmussen rejects the view that ‘personal values are a special subgroup of extrinsic values.’ What needs to be introduced in analyzing personal values is a very particular range of attitudes, namely those that have the form ‘x favors O for y’s sake.’ Instances of such attitudes include both favoring O for its contribution to y’s welfare, as well as attitudes of respect, admiration, honor, and so forth, that do not necessarily positively impact y’s welfare. The proposed Fitting-attitude analysis of personal value defended in the essay is stated as follows: O has personal value for y, if and only if there is reason to favor O for y’s sake. There are some problematic examples addressed in his essay, and Rønnow-Rasmussen admits that there are general problems for the buck-passing account as a whole that he does not address. The Fitting-attitude analysis of values does have the advantage of neutrality on the issue of value objectivism.

3.2 Choice If human thought led directly to action, there would be no need for philosophers to discuss the will. However, the relation between thinking and acting is highly mysterious. Much of human thought comprises of beliefs acquired through perception or reasoning. How does thinking cause one to act? There are other states of mind

6

D.K. Chan

besides belief. In particular, desires are concerned not with the actual state of affairs in the world, but with a possible state of affairs that the agent favors. But not every desire is followed by action to bring about that state of affairs. It seems that an agent must make a judgment as to which desire is to be satisfied, and when it is to be satisfied. The former requires an evaluation of what is desired, and the latter requires a belief that the opportune time to act has arrived, given the condition of the agent and the surrounding world. In ‘Volitions, Comparative Value Judgments, and Choice,’ Joseph Boyle examines what is needed to link value judgments with action, and finds a place for human freedom in his theory of motivation. Boyle distinguishes between the motivations behind the cognitively based behavior of non-human animals, and the volitions characteristic of human action that are ‘responsive to specifically human cognition, namely, to knowledge by way of propositions, including their combination in reasoning.’ In all voluntary human actions, the agent’s judging that it would be good to do x is connected to the agent’s doing x by the agent’s wanting to do x, and the latter involves ‘an interest in a good, an intention of specific goals instantiating the good, and the rational adoption of means to those goals.’ Boyle discusses cases where the connection is blocked by a conflict of desires where ‘a further initiative of the agent to overcome the conflict’ is needed. Human choice is not simply adopting a means to an end, but ‘the kind of volition called for to overcome conflicts of motivation that block action.’ In comparing the options for action, the agent seeks to uncover common measures for ordering the options. Free choice involves the selection of one from a set of options that is judged incommensurable in value: it is ‘a choice without sufficient causal conditions prior to the agent’s choice itself,’ where the rational desirability of the rejected option is not silenced or undercut by the endorsement of the other option. Free choice is not a mysterious twitch or accident, but ‘a rational and appropriate, though undetermined, response to the judgment that one proposed action is incommensurable in desirability in relation to other actions the agent has reason to do.’ Choice has always interested moral philosophers. Wrongdoers may be excused for what they do if they did not choose to do it. It is only appropriate to judge a person by his action if that action is one that is chosen by him. Joseph Biehl in ‘The Insignificance of Choice’ questions such assumptions about choice because he doubts ‘our notion of choice is capable of bearing the burden that many want our moral discourse to impose.’ He distinguishes between our practices of holding people accountable as objects of moral criticism with the purpose of altering their behavior, and the sense in which moral claims and judgments are normative where there is a standard of correctness by which behavior is measured. The ability to choose a different course of action may have a significant role in the former practices, but Biehl thinks that this is not so in the latter sense of morality. Moral standards, like standards of rationality, are inescapable and not conditional on the aims of the agent. The notion of choice is required to understand rationality; hence if the normativity of morality could be accounted for in terms of rationality, choice would indeed be important for morality. But although this is an attractive idea, it cannot be maintained. Biehl sees choice as ‘always a matter of selecting something for an explicit reason.’

Introduction

7

The standard of practical rationality is the standard of instrumental reasoning, and not a standard that applies to ends rather than to means. Thus, a person can be rational even in the pursuit of immoral ends. On the other hand, if there were a standard of practical rationality that is concerned with ends, it would not be one that is used to evaluate the agent’s choices.

3.3 Desire Wanting something provides an agent with a starting point for deliberation, and he acts for a reason if he acts to satisfy that desire in accordance with his practical reasoning. Models of practical reasoning assume that the agent’s desires can be clearly articulated for the purpose of his deliberation. Against this, Bernard Williams has suggested that there is an essential indeterminacy regarding what an agent has a reason to do (Williams 1981). Patrick Fleming in ‘The Indeterminacy of Desire and Practical Reason’ examines how the agent’s motivational set can be indeterminate, and sheds light on the nature of practical reason. The basis of Williams’ claim, that it is indeterminate as to what we have a reason to do, rests on two kinds of indeterminacy. Indeterminacy in procedure concerns what considerations are brought to the agent’s mind through the use of his imagination. Indeterminacy of desire concerns the unsettled content of the agent’s subjective motivational set. The rationalist response to Williams is ‘to introduce formal constraints on desires to yield an account of reasons.’ Fleming criticizes the attempt by Michael Smith to constrain desires by what our fully rational selves would be guided by in acting, namely those that can be systematically justified (Smith 1994). But seeking coherence among one’s desires may not be rational or achievable in a world of plural and conflicting values. If there are different fully rational selves that we may become, an agent needs to decide what type of person to be. Any hope of settling this by a rational process founders on the indeterminacy of desire. Either the content of the desire is unsettled, or the weight of the desire may be indeterminate. The problem, Fleming suggests, is not merely epistemological. There may be ‘no fact of the matter as to what the agent desires or most desires.’ This is not the claim that problems of indeterminacy of desire are irresoluble, but that a rational examination of what has value is not always the way to resolve them. The lesson seems to be that there are practical questions that need to be settled that cannot be settled by a narrow form of practical reasoning, and that ‘no set of formal principles will answer all the questions that are of interest in practical reason.’ In particular, the most important contribution of deliberation has to do with constitutive questions about what kind of practical identity to adopt. Motivation by desire does not seem to be sufficient for agency, for some desires are ones that a person may rather not have. Harry Frankfurt distinguishes between desires that a person identifies with, and those that are alien to him, depending on whether he reflectively endorses the desires in question (Frankfurt 1971). In contrast, critics of Frankfurt suggest that a structural analysis of identification can provide an objective basis for desires to be the agent’s own. James Stacey Taylor in ‘The Myth of Objectively Alien Desires’ attempts to defend an attitudinal analysis

8

D.K. Chan

of identification similar to Frankfurt’s by showing that the three most prominent kinds of objective criteria of identification do not succeed in distinguishing between alien and non-alien desires. On the basis of counter-examples, he shows that a desire may be irresistible, uncontrolled, or induced by other agents, and still be one that the agent identifies with. The identification is achieved by the agent’s decision ‘to act on a desire such that it becomes part of her motivational psychology.’ There is a well-known regress objection against Frankfurt’s account in its use of higher-order desires to endorse lower-order ones (Watson 1975). Taylor suggests that an attitudinal analysis of identification could avoid the objection on the basis of the idea that ‘a person must necessarily be active with respect to her decisions.’ No higher-order decisions would be necessary in order to ensure that the endorsed desire is one that the agent identifies with. The assumption that desires motivate only because they aim at something that an agent finds to be good is challenged in ‘On the Intelligibility of Bad Acts’ by Michael Stocker, whose views were first presented in an influential paper, ‘Desiring the Bad: An Essay in Moral Psychology’ (Stocker 1979). Stocker now takes on Joseph Raz, who defends the ‘classical’ view that ‘if there is no good in doing that, there is no reason to do it, and thus, it cannot be done for a reason,’ and ‘only what is desired or done for a reason is intelligible’ (Raz 1999). Stocker argues that some desires for the bad can satisfy the requirement of a desirability condition and thereby provide intelligible reasons for action. There are actions that are done not just from a mistaken conception of the good but which are entirely bad, but which are nevertheless intelligible. Stocker accepts that what makes for intelligibility in action is the answer to the question ‘What do you want that for?’ But examples such as Augustine’s delight in stealing the pears show that citing what is bad or seen as bad can be a satisfactory answer to this question. Counting himself an Aristotelian, Stocker rejects the line of argument from its being natural for a human to do what is good for him, to the claim that ‘we can understand, see as intelligible, how just on its own a human good can attract a human’ but not how just on its own a human bad can attract a human. A person’s developed (or second) nature can be good or bad, and it can be natural for those who are bad to seek what is bad. The desires and acts that attend sadism, religious fanaticism, and abusive love are intelligible as the result of warped or deformed consciences. Although we do believe many of our acts to be seeking what is good, the pressure on agents to see their actions as good may involve ethnocentrism and narcissism. A further objection to the claim that intelligible action is always sub specie boni is that ‘there are other ways of seeing oneself and the world,’ not necessarily as someone whose desires and acts are good. Here we find people given to spite, malice, and contempt: vices that Aristotle found to be both intelligible and bad.

3.4 Intention To many philosophers of action, motivation by desire is not sufficient for human agency of the sort relevant to ethics. That is, for someone to perform an action

Introduction

9

for which she can be held morally responsible, it is required that her desire led to action in a particular way. A person’s moral responsibility is absent or diminished if the desire motivates without her awareness or control in Freudian cases, or against her will in cases of weakness of will. Typically, we hold a person responsible for her intentional actions, which are actions motivated by her reasons for acting. As Anscombe puts it, intentional actions are those to which a certain sense of the question ‘Why?’ has application (see §5 in Anscombe 1957). For an agent to act for a reason is for the agent to have gone through a process of practical reasoning with a conclusion in favor of so acting. Since the agent could make a decision that she either fails to act on, or will only do so at a suitable time in the future, the practical conclusion may either be inert or be future-directed. The mental state that is formed in making a commitment to act now or in the future has a functional role different from the desires that are evaluated by practical reason. This mental state, according to a functionalist account, is an intention (Bratman 1987). In ‘After Anscombe,’ I argue that, although Bratman’s account of intention ‘has provided a conceptual tool for many directions of research in philosophy and cognitive psychology,’ it cannot do the work in ethics that moral philosophers, especially Kantians, use it for. This can be shown by considering the problems in using intention to make a moral distinction in cases of double effect. If so, Bratman’s is not the same concept of intention that Anscombe had in mind when she wrote her book. I show that Anscombe’s account of intention is a much broader concept, with deeper historical roots. This does not mean that Bratman’s concept of intention should be abandoned. Rather, we need to make a distinction between desire and intention that is implicit in Anscombe’s account. That is, the aspects of intention that make her concept suited for ethics do not belong to Bratman’s narrower concept but should be attributed to a concept that can be distinguished from intention. Evidence for what the latter concept is can be found in Aquinas’ account of intentio from which the concept of intention has borrowed its name. Intentio for Aquinas is an act of will in regard to the end, whereas Bratman’s intention is of the means that are chosen in practical reasoning. Intentio seems to be an intrinsic desire for the end. If this is right, then it is desire that is relevant for ethics, and not intention in its modern use. Recognition that desire and not intention is the relevant concept for ethics, in both Anscombe’s and Aquinas’ accounts of human agency, dovetails perfectly with their ethical theories, for it is Aristotle’s ethics that has guided both philosophers in their ethical writings. After all, Aristotelian virtue ethicists are concerned with moral evaluation of the agent’s character, which consists in a set of dispositions to seek ends that are desired for their own sake.

3.5 Practical Reason The object of desire is a state of affairs, and a person who acts to satisfy that desire is engaged in goal-directed behavior. The ‘teleological thesis’ that affirms that all intentional action is goal-directed seems to be shared by Humeans and Kantians alike, but Crystal Thorpe in ‘The Limits of Teleology’ challenges the thesis. The

10

D.K. Chan

focus of her essay is the kind of intentional action that is deliberative, ‘guided by reasoning, reflection and careful consideration.’ Instrumental reasoning is teleological, but non-instrumental reasoning that is concerned with setting the ends of action is non-teleological. When an agent adopts an end, ‘the only change she produces in the world is a change in her psychology.’ To act is to produce a change in the world outside her mind, and for this, instrumental reasoning is required. Hence, on this standard view of practical reasoning, the teleological thesis turns out to be true. Thorpe considers two alternative conceptions of practical reasoning that include other forms of non-instrumental reasoning, namely moral reasoning mediated by rules, and practical judgment. If such forms can guide reasoning that leads to intentional action, independently of instrumental reasoning, then there is reason to think that the teleological thesis is false. Thorpe argues that subsuming a case under a rule does not necessarily involve the promotion of an end. And when a person discerns through practical judgment rather than rules which moral features are relevant, and which features matter more to the case at hand, there need be no appeal to moral rules at all. Here, ‘her action is value-driven rather than goal-directed.’ Thus, the plausibility of alternative conceptions of practical reasoning provides a challenge to the teleological thesis. There is a puzzle involved in deliberation and motivation. Does an agent desire to do what he is motivated to do? If I insert coins in a vending machine, do I desire to do that, or do I just desire the soda? Stephen Finlay in ‘Motivation to the Means’ discusses the ‘Motivation-by-Desire Principle’ of orthodox Humean theory. Either there is motivational transfer from the end to the means, or ‘there is license for attributing the motivation to sources other than desire, such as rational acceptance of imperatives of reason.’ The rationalist objection is that the story of motivational transfer does not account for how rational agents are motivated by reasons, and why they ought to do what they have reason to do. Humeans who simply analyze reasons in terms of desires and beliefs need to explain why the motivation to the means is not merely a contingent psychological fact, but a rationally appropriate response by an agent who recognizes the normative requirement of doing what she has reason to do. Finlay takes Hume and Davidson to have a response to the objection in that they ‘deny the existence of any gap between desire for the end and motivation to the means.’ The motivation-by-desire principle is defended by showing that ‘an agent’s desire for some state of affairs necessarily provides that agent with motivation towards believed instrumental means to that state of affairs.’ Building on elements of Davidson’s account, Finlay constructs his ‘argument from the constitutive nature of desire’ to fully defeat the rationalist objection. Accounts of practical decision making that link an agent’s desires with her actions provide models of rational choice that are important for two purposes. First, the rationality of the decision to act helps to explain the action. Second, moral actions can be shown to be rational by showing that they are the choices that an agent would make based on desires that reflect ethical values. A problem that Erik Schmidt in ‘Thresholds, Vagueness and the Psychology of Small Improvements’ draws attention to is that ‘people frequently fail to conform to the axioms of rational decision theory.’ In his discussion of the problem of incomparability (a term he prefers to

Introduction

11

‘incommensurability’), he argues that philosophers have misdiagnosed the problem due to their acceptance of ‘an empirically inadequate approach to practical reason.’ The ‘argument from small improvements’ is thought to establish that certain alternatives are incomparable in that the relation between them is neither superiority nor inferiority nor equality. Philosophers have appealed to some kind of vagueness to account for the incomparability. But Schmidt borrows from psychological studies of choice that show people to ‘compare alternatives in terms of their subjective rather than their absolute value.’ A threshold is the minimum amount of evaluative difference required to register a difference in subjective value. Cases that fall under the argument from small improvements can be accounted for in terms of the application of a conflicting pair of thresholds. The puzzle then is not due to incomparability, but is a product of human psychology. Although the intransitivity of our decision procedures may seem irrational, it is to our advantage that human decision-making is flexible in that ‘our preferences are not simply elicited or revealed by the choices we make, they are partly constructed during the process of deliberation.’ Philosophers neglect psychological studies of choice at their peril.

3.6 Freedom In explaining human action in terms of intention, choice and desire, philosophers are appealing to causal antecedents of action. But if these mental states are neural states, action can be treated as events in a causal nexus. In order to preserve the kind of freedom required for moral responsibility, libertarians deny that causal necessity applies to free human actions. The compatibilists in the freewill debate criticize as implausible this libertarian account for requiring agents to intervene from outside of the causal nexus. Instead, compatibilists deny that causal necessity rules out human freedom. Free action is distinguished from other events by the type of cause, namely that they are willed or chosen, and not by the absence of causal necessity. But many compatibilists have been in turn criticized for failing to refute the hard determinists who stand on the opposite extreme from libertarians, and who deny that there are free actions at all. As a prime exponent of compatibilism, David Hume also seems to have exemplified this very failure to provide a positive account of free action. Troy Booher in ‘Taking Liberty with Humean Necessity: Compatibilism and Contingency’ examines the role, in Hume’s discussion of liberty and necessity, of the Humean definition of necessity as ‘the constant union of objects and the inference of the mind.’ The definition is not needed for the possibility of free action, but to show that there is only one kind of necessity that applies to both free actions and other natural events, in Hume’s response to a historical controversy where the possibility of free action is not at issue. How then should Hume be located in current debates between compatibilists and hard determinists? Booher argues that Hume should be viewed as a forerunner of the ‘reactive attitudes’ camp of compatibilism,5 and that his positive account of moral freedom is the view that ‘ultimately all moral assessments

12

D.K. Chan

concerning the fairness of holding one accountable are derived from moral sentiments.’ As a matter of contingent fact, recognition that causal necessity applies to actions considered free has had little effect on how these actions elicit our moral sentiments. But as there is no guarantee that our moral sentiments will not change over time, the disagreement between hard determinists and compatibilists cannot be resolved in the abstract.

3.7 Moral Action Values are translated into actions by the practical choices that agents make. But doing the morally right thing seems to go against the self-interest of agents when doing so requires the sacrifice of personal interests for the benefit of others or for the sake of duty. The question arises as to whether morality is rational, or there is some other (non-rational) way to justify (and motivate) moral action. Contractarians like David Gauthier defend moral rationalism by arguing for the rationality of constrained maximization, in which conditional cooperation with others in society yields the highest utility payoff for the individual (Gauthier 1988). Malcolm Murray in ‘Rational Choice and Evolutionary Fit’ argues that the rationalist model fails, and suggests an evolutionary model in its place. The evolutionists maintain that agents who follow moral strategies are more likely to pass on their genes to future generations, but this success has nothing to do with rational choice. For ‘evolutionary success of irrational strategies is the norm, not the exception, in nature.’ To defend the evolutionary model of moral choice, Murray addresses four objections. First, it seems incoherent to advocate irrational strategies. Second, moral behavior is successful on the evolutionary model only under very narrow and unlikely conditions. Third, evolutionary fit may be aligned with moral strategy, but to say that a strategy will have evolutionary success is not saying that morality, as we understand it, is justified. What is captured by the evolutionary account does not seem to be moral agency in a robust sense. Fourth, there is a gap between the descriptive content of evolutionary theory and the normativity of moral judgments. In reply to the fourth objection, Murray suggests that the idea of a ‘real morality’ that is not captured in evolutionary accounts should not be any more of an issue than the idea that ‘real consciousness’ is left out in the explanations of consciousness by cognitive science.

Notes 1 2 3 4 5

Such a concern is raised in the ‘Introduction’ to Flanagan and Rorty (1990). See, for instance, the papers in Cullity and Gaut (1997). Recent work on this topic can be found in Chang (1997); Baumann and Betzler (2004). One of the most recent examples of such an account of value is found in Scanlon (1998). This compatibilist position was introduced in Strawson (1962).

Introduction

13

References Anscombe GEM (1957) Intention. Blackwell, Oxford Baumann P, Betzler M (eds) (2004) Practical conflicts: new philosophical essays. Cambridge University Press, Cambridge Bratman M (1987) Intention, plans, and practical reason. Harvard University Press, Cambridge MA Chang R (ed) (1997) Incommensurability, incomparability, and practical reason. Harvard University Press, Cambridge MA Cullity G, Gaut B (eds) (1997) Ethics and practical reason. Oxford University Press, New York Flanagan O, Rorty AO (eds) (1990) Identity, character, and morality: essays in moral psychology. The MIT Press, Cambridge MA Frankfurt H (1971) Freedom of the will and the concept of a person. Journal of Philosophy 68:5–20 Gauthier D (1988) Morals by agreement. Oxford University Press, Oxford Mackie JL (1977) Ethics: inventing right and wrong. Penguin, Harmondsworth Mele A (ed) (1997) The philosophy of action. Oxford University Press, Oxford Nagel T (1997) The last word. Oxford University Press, New York Raz J (1999) Engaging reason: on the theory of value and action. Oxford University Press, Oxford Scanlon TM (1998) What we owe to each other. Harvard University Press, Cambridge MA Smith M (1994) The moral problem. Blackwell, Oxford Stocker M (1979) Desiring the bad: an essay in moral psychology. Journal of Philosophy 76:738–753 Strawson P (1962) Freedom and resentment. Proceedings of the British Academy 48:1–25 Watson G (1975) Free agency. Journal of Philosophy 72:205–220. Williams B (1981) Internal and external reasons. In: Moral luck. Cambridge University Press, Cambridge

Chapter 1

Moral Realism, Meta-Ethical Pyrrhonism and Naturalism Drew Khlentzos

Abstract This paper argues that naturalistic moral realism is vulnerable to a Hard Problem that has gone largely unrecognized. This problem is to explain how natural moral properties are detected by the folk. I argue that Thomas Nagel’s persuasive case for moral realism founded on the priority of first-order moral evaluations over second-order reflection is not conclusive—a certain type of moral agnosticism which I call Meta-Ethical Pyrrhonism can account for our inability to think of firstorder moral evaluations as merely subjective or relative. Although unsatisfactory as metaphysics, Meta-Ethical Pyrrhonism is arguably all that a moral naturalist is entitled to by way of a meta-ethical theory. Keywords Hard Problem · Meta-Ethical Pyrrhonism · Moral realism · Moral skepticism · Naturalism

Suppose that there are no non-natural facts—that all that exists is determined by ultimate physical laws and boundary conditions. Can any such naturalistic scheme countenance moral norms and values? Mackie (1977) notoriously thought not. Moral norms and values are simply too queer to be part of the furniture of the world, he contended. The reason was that these norms and values as we construe them are supposed to be objectively prescriptive. Yet objective prescriptivity is a ‘queer’ property so unlike any other natural property that we could never have any naturalistic ground for believing in it. Mackie’s Argument from Queerness has not convinced naturalistic moral realists. Some hold that Mackie is mistaken in thinking that objective prescriptivity, at least as Mackie himself construed it, is a sine qua non of folk moral norms and values, others that the very notion of objective prescriptivity rests upon a confusion of ontological with motivational questions. D. Khlentzos Language and Cognition Research Centre, School of Behavioural, Cognitive and Social Sciences, University of New England, Armidale, New South Wales, Australia 2351 Tel.: +61-2-67733208 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

17

18

D. Khlentzos

My primary purpose in this paper is not to offer a substantial contribution to this debate, although I shall need to say something about objective prescriptivity and its place in folk morality and in this context shall explain why Mackie’s notion cannot be lightly dismissed by naturalists. My aim is to first examine from a naturalistic point of view the phenomenon with which Mackie starts: the conflict between our ordinary moral judgments and naturalistically grounded reflections on the nature of those judgments. I shall argue that Mackie is right to find it extremely difficult to bring first-order ethical judgments into harmony with second-order reflection on those judgments. Furthermore, moral naturalists have not appreciated the complexity of the task they face—they have, I shall argue, largely ignored a foundational problem of naturalistic moral epistemology. This will involve a brief discussion of what naturalistic moral realism is and of why it is vulnerable to a difficulty I label the Hard Problem. I shall then argue that Thomas Nagel’s persuasive case for moral realism founded on the priority of first-order moral evaluations over second-order reflection is not conclusive—a certain type of moral agnosticism which I call Meta-Ethical Pyrrhonism can account for the phenomenon Nagel emphasizes: namely, our inability to think of first-order moral evaluations as merely subjective or relative. Although unsatisfactory as metaphysics, Meta-Ethical Pyrrhonism is arguably all that a moral naturalist is entitled to by way of a meta-ethical theory.

1.1 Objective Prescriptivity and the Requirements of Naturalism Suppose you hear a voice instructing you to strangle your cat. Checking to see where the voice is coming from you realize, to your dismay, that it has issued from within your own head. Naturally, you do not think to obey. For as you now realize, you have just experienced a transitory psychotic episode. Schizophrenics hear such voices and, tragically, often obey the instruction conveyed, however bizarre it may be. Why do they do so? Given their condition, sometimes because of the tone or the urgency of the instruction, other times because of the authority they believe to have issued it. The schizophrenic who strangles his cat may well do so because he believes that this had to be done. If he believes that God or Satan instructed him to do so, he has identified the source of this to-be-doneness. On the other hand, if we can get him to acknowledge that he was merely reacting to the urgency of the instruction or the insistent tone in which it was delivered, we may just succeed in convincing him that he was mistaken in believing the action of strangling his cat really was objectively prescriptive. Any naturalist who believes that it is straightforwardly and objectively wrong to incarcerate young children who have committed no crime owes us a naturalistic account of the source and nature of this objective wrongness—one founded upon natural, ultimately physical, facts. For a number of reasons, moral naturalists seem not to have fully appreciated the complexity of this task. This is partly because of two misplaced prior commitments—to the metaphysical thesis of moral supervenience

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

19

and to the belief that central folk moral judgments must be true. The combination of these two has led naturalists to misconstrue their fundamental metaphysical task so that it becomes for them a question of either showing how the totality of natural facts fix the moral facts or, if non-reductive supervenience be judged inadequate, identifying those natural properties to which the moral properties are to be reduced. However, the fundamental metaphysical task for the naturalist is not to vindicate any sort of folk theory, however venerable that theory may be. The fundamental task is to show how, in the moral case, normative moral properties are so much as possible in a purely physical world: how does a world of fields and forces give rise to anything so strange as a moral obligation or a virtue or a duty? To be sure, this is no isolated problem. The moral naturalist’s task is part of a wider naturalistic metaphysical task that I’ve dubbed the Representation Problem in earlier work (Khlentzos 2004). The Representation Problem is the problem of explaining how mental representation of any feature of a mind-independent world can be possible, ethical and normative features included. Naturalistic realists have been slow to appreciate the centrality of the Representation Problem as well.1 One very plausible naturalistic answer to our question is that a physical world cannot give rise to any such things as duties or virtues or values except by giving rise to intelligent creatures who invent them without realizing this is what they’ve done. This was, of course, Mackie’s answer. If there are no objective moral facts or properties to begin with, then there are none to be fixed by the physical facts and there are none that are reducible to those facts. Whence, the naturalist cannot start with the assumptions most moral naturalists start with. The naturalistic task with respect to folk moral theory must therefore be one of radical interpretation. Moral language is a jungle language, to be interpreted from scratch using bona fide natural properties, and not one that the theorist already knows how to speak. Along the way, the naturalist will need to pose and answer questions that might look bizarre to the folk or that might offend common sense, questions such as this: how is the folk appeal to the voice of moral conscience to be distinguished from the schizophrenic’s appeal to voices inside his head? So, consider our example once more and assume the naturalist concurs with the folk belief that incarceration of innocent children is wrong. To the extent our naturalist is a realist, s/he will contend that such incarceration would still be wrong even if we all agreed it wasn’t and that our merely judging it to be wrong does not make it so. There is a judgment-independent fact to the matter as to whether incarceration of the young is wrong. Moreover, the confident and unanimous folk belief that it is wrong could at least in principle be mistaken, according to the naturalistic moral realist. What is there to say it is not? How do the folk manage to detect and respond to the natural property of objective impermissibility inhering within acts of incarcerating the young? The question does not merely ask: how do the folk form beliefs about what is objectively impermissible? It asks how they form veridical beliefs such as the belief that incarceration of innocent children is objectively impermissible, assuming it is. And the question is not: ‘how does the naturalistic theorist detect this natural property?’ It is: ‘how do the folk detect this natural property?’

20

D. Khlentzos

Suppose the natural property of objective impermissibility is its tendency to minimize overall well-being if adopted as a rule for action. Positing a brute faculty of moral intuition in the folk for detecting this property is clearly unacceptable to any naturalist. The suggestion must be that the folk infer its existence somehow. Infer it from what? Their moral feelings, perhaps? The folk feel disgusted or outraged by the prospect of seeing innocent children behind bars and infer that the best explanation of their disgust and outrage is that these feelings are elicited by the gross-well-being-inhibiting feature of incarceration. But this inference looks hazardous to the extent that it may very well tell against the moral realist. For if child incarceration is objectively impermissible, then the direction of explanation, for a moral realist at least, needs ultimately to proceed in the reverse direction: we feel disgusted because it is wrong, rather than, as non-cognitivists aver, we infer that it is wrong simply because we all feel disgusted by it. Moreover, and not insignificantly, this is precisely the explanation the folk themselves give of their attitudes. Rightly or wrongly, the folk believe that they can detect moral properties directly. Furthermore, for this inferential story to have any hope of working, we’d need to be able to distinguish objectively grounded moral disgust or outrage from nonobjectively grounded moral disgust or outrage, since there are many things that disgust/outrage some that do not disgust/outrage others and there are even things that tend to disgust or outrage (almost) everyone—such as the prospect of intense suffering of a loved one in the terminal stages of cancer—that are not objectively impermissible. Yet the task of distinguishing the relevant objectively grounded moral feelings from the non-objectively grounded moral feelings simply raises anew the very problem this type of account was invoked to solve: how do the folk detect the natural property of objective impermissibility? Do things improve for the naturalist if, instead of moral sentiments, we appeal to moral judgments as evidence for the existence of a natural property of objective impermissibility? Perhaps, as suggested in Nagel (1997), it is the folk’s inability to think of incarceration of children as anything other than objectively impermissible that provides the strongest evidence that their moral judgments track objective properties. We are going to examine Nagel’s view in more detail later but we can note this now: even if Nagel’s suggestion is right, it still doesn’t solve the moral naturalist’s dilemma. Positing appropriate natural properties to identify with folk moral properties is the easy part; saying how untutored folk detect and respond to them in arriving at their moral judgments is the Hard Problem of naturalized moral epistemology. Surprisingly, naturalists seem to have paid little if any attention to the Hard Problem. Many write as if there is no obligation on a naturalist to address it. Thus, according to Susan Hurley, ‘to say that a certain act ought to be done is to say that it is favored by the theory, whichever it may be, which gives the best account of the relationships among the specific values that apply to the alternatives in question’ (see p. 11 in Hurley 1989). Yet the folk who make true moral judgments, such as that incarceration of children is wrong, know nothing of any moral theory. If the folk moral judgment and the theorist’s moral judgment are not to simply diverge in meaning when each declares

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

21

child incarceration wrong, we need some way of connecting folk evaluations to the theorist’s. Moreover, the connection between folk ‘wrong’ and the theorist’s ‘wrong’ had better be a very direct one, otherwise, to quote Bas van Fraassen, the theorist runs the risk of ‘giving in to one of the temptations that make for really bad philosophy, surreptitiously and underhandedly turning the ordinary word . . . into technical jargon while ostensibly keeping it intact’ (see p. 7 in van Fraassen 2002).2 To underscore the interpretative task facing the naturalist, imagine that our theorist, Nat, has stumbled upon a lost race of people completely cut off from civilization, somewhere in the jungles of Malawi. Remarkably, these people (the Churls as they call themselves) appear to speak English or something that sounds awfully like it. However, their moral vocabulary, if that is what it is, seems to differ from our own. Instead of describing acts as ‘right’ and ‘wrong’ the Churls classify them as ‘cornucopian’ and ‘pandoran’. Nat is excited to learn that acts described as pandoran in Churlish can be mapped onto those acts that he knows to have high overall disutility for the Churls and that acts that are cornucopian can be mapped onto acts that possess high overall utility for the Churls. Our theorist is thus satisfied with his discovery that ‘pandoran’ just means possessing high overall disutility and that ‘cornucopian’ means possessing high overall utility. What if the Churls reject this identification though? What if Nat has failed to appreciate the religious significance of the terms ‘pandoran’ and ‘cornucopian’? What if ‘pandoran’ in Churlish means cursed by the gods and ‘cornucopian’ in Churlish means blessed by the gods? Should Nat still maintain that in spite of what the Churlish folk themselves say, he has located the natural properties that are the real Churlish moral properties on the grounds that a mature Churlish folk morality freed of its current entanglement with primitive Churlish religion would independently endorse the property identities he has posited? Of course, our theorist could maintain this and could even do so with some plausibility if he could show how the Churls are able to first detect and then respond to overall utility/disutility without realizing (even denying) that this is what they are doing. Yet even this would not suffice. Detection followed by response is necessary but it is far from sufficient. The good folk of Salem were able to detect and respond to the effects of ergot poisoning which they mistook for demon possession, but no one would think of rehabilitating witch theory by suggesting that a mature folk witch theory freed of any entanglement with the primitive fears and superstitions associated with witches would identify the property of demon possession with that of ergot poisoning.3 The allegory of Nat and the Churls demonstrates what is needed in the moral case. The folk must unwittingly detect and respond to the relevant natural moral properties—what the folk call ‘morally impermissible’ is really what the theorist identifies as gross-well-being-inhibiting features, say, even though the folk have no inkling that this is what they are detecting and responding to and may roundly reject the suggestion that it is. Let us assume that overall utility is the relevant natural property. The idea then is that moral judgments are no different from water judgments in this respect. Even if we all believed that water was a living spirit, our water judgments would still in reality be H2 O judgments.4

22

D. Khlentzos

But how do the folk manage this for their moral judgments? How do they lock onto the right natural property? Perhaps there is some module in the brain that performs the relevant bulk utility calculations? We would need more than this though. We’d need a reliable utility calculator, at least for the relatively uncontentious cases such as the incarceration one. The output of the module must map onto the facts, on pain of failing to vindicate moral realism.5 So perhaps reliable utility modules have been selected for in our evolutionary history? There are a number of difficulties with this proposal, not the least being that the empirical evidence attests to a widespread divergence in basic moral evaluations that seem to be closely linked to the moral evaluator’s culture. This affords some prima facie evidence against the existence of a shared module.6 The thing to note at this stage though is that the moral naturalist needs some such suggestion to be true. Why? In the case of water and H2 O, there was no need to posit a dedicated H2 O-detecting brain module. So why is it needed here? The reason brings us back to Mackie. Moral judgments, as Mackie argued, directly engage the will. Unless the impermissibility-grounding property were to be somehow represented as such in the brain, there would be no naturalistic justification for thinking that it was that property rather than some other one that moved the folk to evaluation and action. Yet if our theorist were to discover that the property s/he had identified with the impermissibility of incarcerating children was not causally responsible in the folk for their attitudes of disapproval, expressions of disgust and outrage, political agitation and so on, s/he would have to revise the conjectured identity. For qua radical interpreter, the naturalist wishes to know what natural property it is that the folk recognize and respond to when they issue their moral judgments and evaluations. The moral case contrasts sharply with water and H2 O. The folk have no dedicated mental module for detecting the presence or absence of the chemical property being composed of H2 O. Nor do they need any. It suffices, for the purposes that water serves for individual survival and social transactions, that there be some semantic marker for the substance water acquired in the ordinary process of learning the significance of that substance. It is unnecessary that there also be some pre-coded mental formula representing the composition of that substance. For the moral case though, no physical substance is the subject of moral predication whence the substance/composition distinction lapses. Thus, if the property of being objectively impermissible is to be identified with the natural (dispositional) property of leading to a much lower ratio of pleasure to pain, and if, unbeknownst to the folk, it is that property which they actually detect when they evaluate child incarceration as impermissible and which, once detected, leads them to decry as immoral the actions of any political regime instantiating it, then they had better have some reliable means of detection of that property. Moreover, this means of detection would need to be relatively encapsulated and not subject to top-down processing, if it is to be compatible with widespread ignorance as to its true nature. That suggests a dedicated brain module. More complex hypotheses are possible of course. There is no reason why a moral naturalist should not be a pluralist about moral values and attribute a variety of different such modules suitable for detecting other natural properties

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

23

that could serve within different individuals as value properties. Alternatively, the modularity hypothesis might be challenged by some moral connectionist model. These are details that need not concern us. Mackie of course did not posit anything like a module on behalf of moral naturalists, since he wrote many years in advance of the modularity of mind thesis. What he (famously) argued was that objective values would need to somehow attract the will. Yet some naturalists who are externalist in their theory of moral motivation reject this requirement as gratuitous, whilst others believe that it is possible to embrace less extreme versions than Mackie’s on the question of objective prescriptivity and the will. In fact, as we shall see, Mackie’s requirement seems entirely justified once one appreciates the Hard Problem. But let’s review the objections to it first. In a review of Susan Hurley’s Natural Reasons, Frank Jackson writes: If Mackie is right, if it is an essential part of M (matured Folk Morality) that there be properties which of necessity attract anyone who contemplates them, no doubt nothing is (morally) right. But more moderate views about M are possible (see p. 486 in Jackson 1992).

To anyone who appreciates the Hard Problem of saying how it is that the folk first detect and then respond to those natural properties that are the moral properties, the requirement that the relevant natural properties of necessity ‘attract the will’ does not look at all gratuitous. For this is precisely the explanation the folk themselves tender when quizzed about why they acted in the way they did on certain occasions: Nat “What moved you to report those young guys to the police?” Subject “They set fire to a cat!” Nat “But what was it about that action that disturbed you?” Subject “It was blatant cruelty to an animal! What would you think?”

If our theorist has identified high overall disutility with an action’s wrongness, then it must have been a recognition of the high overall disutility of the action of setting fire to the cat that prompted the subject’s moral disgust and subsequent action. At least it needs to have been that if the subject’s moral evaluation is to be objectively true. This is just to say that it is that natural property of that type of action that repels to some degree or other any rational, fully informed moral agent who contemplates it. What ‘more moderate’ view could Jackson have had in mind? Only if one adopts the highly immoderate view that recognition of moral value has no effect per se on the will could one think otherwise. That leads into the next criticism since the above is in effect the externalist view. Far from it being built into the nature of natural moral properties that they provide the source of moral motivation, it is simply a result of socialization that we feel motivated to care about others to any degree or to act on our moral judgments and evaluations, according to externalists about motivation. Now there is something to this externalist objection, but it is opaque whether it could do anything other than exacerbate the Hard Problem if it were true. That there is a difference between being more motivated to perform action A rather than action B and ranking A ahead of B in some preference order seems a sine qua non of any credible account of akrasia. It would be a very odd theory of action though

24

D. Khlentzos

that explained normal psychological motivation as a mere artefact of socialization, taking as its model for action the pathological case of the akratic. Moreover, the empirical claim that it is purely by dint of socialization that we care about others or feel motivated to act on our judgments seems very dubious at best. There is a mounting body of psychological and neurophysiological evidence attesting to the crucial and ineradicable role emotions play in decision-making and an equally impressive body attesting to the genetic provenance of feelings of empathy towards others and dispositions toward reciprocity.7 The crucial point for our purposes though is this: suppose that externalists are right and that moral motivation is just a socially-induced phenomenon. Suppose as before that the morally impermissible is the natural property tending to reduce the ratio of overall pleasure to overall pain. Call this natural property R. The Hard Problem of naturalistic moral epistemology is to explain how R determines the content of (or otherwise enters into) folk moral judgments. Externalists reject Mackie’s suggestion that R must somehow engage the will. So we assume that it is the socially-induced internal property S that does this instead. For simplicity, assume S is just the socially sanctioned rule ‘Reject incarceration of the young and innocent!’ Now when the folk pass the moral judgment that incarceration of children is impermissible, it is no longer R but S that moves them to decry such incarceration and agitate for its removal from political regimes that institutionalize it. For the radical interpreter, what is the point then of appealing to the external natural property R in any naturalistic explanation of why the folk issue the particular moral judgments they issue when, ex hypothesi, it is S not R which motivates them to do and say what they do and say? Externalism about moral motivation simply exacerbates the Hard Problem. I think this was one of the reasons why Mackie drew a contrast between objective and subjective values. There were, he argued (Mackie 1977), two independent constraints that moral values needed to satisfy: 1. They had to systematically and reliably guide decision and action. 2. They had to be grounded, independently of the agent, in the objective natural order. According to Mackie, whilst objective values satisfied (2), they failed to satisfy (1). Subjective values, on the other hand, satisfied (1) but failed to satisfy (2). In order to ensure that objective values, such as the moral naturalist’s natural properties, could guide decision and action and thus satisfy (1), they would first have to engage or at least attract the will. Alternatively put, the folks’ action-guiding subjective values would have to have been formed and conditioned as motivated responses to those natural properties that comprise the objective moral values. At least this is so if we believe, as moral naturalists do, that the central folk moral judgments are objectively true. Even so, there is a puzzle with Mackie’s subjective values that John Burgess has recently drawn to our attention (Burgess 1998). Mackie seems to have taken subjective values more or less for granted, apparently regarding the revelation that all our putative objective values are in reality subjective as a vindication of his Error

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

25

Theory. Yet, as Burgess observes, even if Mackie is correct in holding that the source of all our moral evaluations is to be found in our own attitudes rather than the objective natural facts, the fact is that we judge actions to be good or bad, right or wrong, and people to possess various moral attributes. If this is so, we are still committing the error of regarding as valuable something that is not in itself valuable at all. So why doesn’t Mackie reject subjective values as well?

1.2 Moral Realism We need to say a little more about how moral realism is to be construed. Without attempting a precise definition, I take moral realism to be the thesis that there are mind-independent moral entities and properties—norms, values, duties, and such like. Naturalistic moral realism is then the view that the mind-independent entities and properties that are the moral ones are simply certain sorts of natural entities or properties. An ontological construal of moral realism is, in my view, preferable to any semantic construal since the latter are too weak to distinguish realist views from various antirealist pretenders. Thus, moral nihilists who are quietists about folk moral practice and deflationist about truth might agree with the claim that folk moral judgments are either true or false, on the grounds that it is part of the socially useful practice of issuing moral judgments or resolving ethical disputes that there be definite conditions under which we can assert some ethical claims and deny others. Thomas Nagel defends an attractive form of moral realism that is grounded in the objectivity of moral reasoning. At first blush, Nagel’s view looks as if it is not amenable to any type of ontological construal. For he writes: I take it for granted that the objectivity of moral reasoning does not depend on its having an external reference. There is no moral analogue of the external world—a universe of moral facts that impinge upon us causally. Even if such a supposition made sense, it would not support the objectivity of moral reasoning (see p. 101 in Nagel 1997).

This sounds as if he may mean to reject real moral entities and properties as superfluous. He also seems to want to reject naturalistic moral realism if this means, as I have assumed it does, that moral entities and properties are reducible to natural entities and properties, when he claims: ‘. . . in the end, the contest is between the credibility of substantive ethics and the credibility of an external psychological reduction of that activity’ (see p. 115 in Nagel 1997). These appearances are however misleading. Nagel sees moral reasoning as a species of practical reasoning aimed not at finding out how things are in the world but rather at settling the question of what we ought to do. That is why the supposition that moral facts impinge upon us causally is superfluous even if coherent. Crudely, we act so as to change the world; it is irrelevant to that project that our sources of motivation stem from the world acting on us first (if indeed it does so). As to moral naturalism, Nagel is perfectly happy to countenance the possibility that rule utilitarianism, for instance, might provide the best vindication of our moral

26

D. Khlentzos

reasoning. The ‘external psychological reduction’ Nagel means to oppose is not a bona fide naturalistic reduction of value to some natural property but what he terms the subjectivist/relativist ‘reduction’ of objective moral reasoning to some merely parochial or subjective foible of our human psychological makeup. In fact, the latter ‘external psychological reduction’ is not a genuine metaphysical reduction at all but part of an error theory about the source of the illusion of objective moral reasoning. Nagel’s project is then to show how ‘. . . attempts to get outside of the object language of practical reasons, good and bad, right and wrong . . . and see all such moral judgments as expressions of a contingent, non-objective perspective will eventually collapse before the force of the first-order judgments themselves’ (see p. 103 in Nagel 1997). As we have already noted, an ‘internalist’ project such as Nagel’s which seeks to vindicate our belief in objective norms and values by demonstrating that the practice of first-order moral reasoning would be incoherent in their absence does not solve the Hard Problem of naturalistic moral epistemology. Nor does it purport to. It does not show how the folk manage to target precisely those objective natural properties that are the moral properties as they make their moral evaluations. But if successful, it does comprise a crucial first step towards any such solution—at least we’d know the relevant natural properties are there in the world and that it is to those properties that our moral reasoning is somehow differentially sensitive. Henceforth, the type of naturalistic moral realism that I shall examine is Nagelian realism. The Hard Problem will have to wait. After all, there is no certainty its solution even lies within our grasp. I am going to argue that whilst Nagel is right to contend that ‘the procedures of justification and criticism we employ can [not] be thought of as . . . something we just do,’ this does not by itself vindicate moral realism. There is an interesting type of moral agnosticism that concurs with Nagel’s appraisal.

1.3 Possible Conflict Between Moral Evaluations and Moral Theory Nagel is not alone in counterposing first-order judgments to second-order judgments in order to confute skepticism. There is currently a lively dispute in philosophy of mind about the possibility of genuinely mental causation in a purely physical world.8 If mental properties are causally otiose, how is epiphenomenalism to be avoided? Philosophers such as Tyler Burge, Lynne Rudder Baker, and Robert van Gulick have appealed to the certainty we feel in our ordinary first-order ascriptions of mental causation as an antidote to the second-order philosophical scepticism engendered by naturalistic worries about how such causation is possible. Thus Burge claims: Materialist metaphysics has been given more weight than it deserves. Reflection on our explanatory practice has been given too little. The metaphysical grounds that support the worries are vastly less strong than the more ordinary grounds we already have for rejecting them. . . . I think it more natural and fruitful to begin by assuming, defeasibly perhaps but firmly, that attributions of intentional mental events are central to psychological explanation

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

27

both in ordinary life and in various parts of psychology. We may also assume that intentional mental events are often causes and that psychological explanation is often a form of causal explanation. Given these assumptions, the ‘worry’ about epiphenomenalism seems very remote. . . . None of the metaphysical considerations advanced in current discussion seem to me remotely strong enough to threaten this conclusion (see p. 97 and p. 118 in Burge 1995).9

Jaegwon Kim responds to Burge as follows: The problem of mental causation is primarily a metaphysical problem. It is the problem of showing how mental causation is possible not whether it is possible, although of course what happens with the how-question may in the end induce us to reconsider our stance on the whether-question. In raising the how-question, we are assuming “defeasibly but firmly” as Burge says, that the whether-question has already been affirmatively answered (see p. 61 in Kim 2000).

Kim’s reply is quite correct as far as it goes. However he only alludes to the underlying philosophical question that I think is really worrying Burge: what if our theoretical deliberations lead us to conclude that mental causation is impossible yet we find ourselves unable as theorists to treat our own attributions of mental causation to ourselves and others as fictions or errors or mere rationalizations? This is Nagel’s question and Burge tenders a Nagelian reply: in that event, we would have to disbelieve the theory since the first–order attributions of mental causation are much more secure epistemically than any of the various metaphysical theories about mental causation. What of the moral case though? My judgment that incarceration of innocent children is objectively impermissible seems to me to be, if anything, even more compelling than my judgment that my desire for beer caused me to go to the refrigerator to look for it. I cannot see how I could possibly rescind the moral judgment even if, to my great surprise, it turned out that it was not my desire for beer that led me to the refrigerator. Moreover, like Nagel, I cannot see how I could regard it as just another fact about me that I happen to make this considered moral judgment rather than its contrary say or none at all. The Hard Problem of moral epistemology is a metaphysical one also. It is the problem of showing how the folk detect and are differentially sensitive to those natural properties that are the moral ones. So, insofar as I am a naturalistic realist, must I follow Burge and Kim in assuming ‘defeasibly but firmly’ that the ‘whether question’ has been affirmatively answered in the moral case? No. There is a way of respecting the perceived character of objective prescriptivity implicit in folk moral judgment that is compatible with agnosticism about the reality of the moral values and properties we take to underwrite them. In order to appreciate the type of moral agnosticism that can sustain this attitude, it is useful to view moral norms in a wider setting. One important function of moral norms is to counteract limited sympathies, as Mackie put it, and if moral norms are perceived by the folk as objectively and intrinsically prescriptive, and that they are so perceived is common knowledge amongst them, then even if this perception is erroneous, these norms may function under favorable conditions as credible replacements for various forms of

28

D. Khlentzos

external enforcement within social transactions that would otherwise yield Pareto sub-optimal outcomes. The general observation is hardly new, comfortably predating Mackie. Indeed, it was clearly appreciated by Plato. Whilst the relativisation of the efficacy of norms to favorable conditions might sound a little unfamiliar in the moral context, only my proposal about the nature of those conditions is novel. Game theorists distinguish, amongst agreements and threats, those that lack credibility from those that do not. Any agreement or threat that is not enforceable is deemed ‘cheap talk’. If moral norms are to function in the way they do to mediate folk social transactions, they need to be perceived precisely as the folk do perceive them: as intrinsically and objectively prescriptive. Any person who fails to observe the moral injunction, for example, to respect the basic rights of others, is ostracized in a number of ways since, according to folk moral thought, it is an intrinsic and objective fact about human beings that they have basic rights. ‘Respect the fundamental rights of others’ is an intrinsically and objectively prescriptive folk moral norm. Suppose now that there is in reality nothing to this talk of basic rights, that the whole idea of humans as possessing such properties is a collectively beneficial anthropomorphic illusion. If this fact were to become common knowledge amongst the folk, it would undermine the role of moral norms in social life. But wouldn’t it equally do so for any theorist apprised of this discovery? The justification for fair treatment of others, including inter alia not incarcerating young children, could no longer be that humans have certain fundamental rights. Now if Kant was right, our moral duties are the dictates of reason. A consequence of his view is that certain simple forms of the Prisoner’s Dilemma (PD) can in principle be dissolved—each autonomous agent in certain simple PD situations has a non-instrumentally rational reason to cooperate with the other, and that this is so is common knowledge amongst them since it is a priori deducible from a knowledge of the conditions of their own autonomy. If moral duties are illusions though, not only are the simpler PD situations not (dis)soluble through a priori reasoning, but there is now a choice for the theorist where there was not before as to whether to abide by moral precepts at all for those interactive situations in which moral precepts normally mediate. The challenge to the reflective theorist will then be to find a reason for acting morally in those interactive situations where doing this goes against his own perceived best interests. If moral norms and values are illusions or social constructs or reflections of purely subjective preferences, then recognition of this fact must lead the theorist to reconsider their probative force. Moral norms would then stand in danger of becoming cheap talk for the theorist. Let us call this the Ring of Gyges objection since its upshot is that knowledge of the irreal status of the moral law is a ring that renders the cognoscenti invisible to the dictates of that law. I think it is the Ring of Gyges objection that is really troubling Nagel in his discussions of moral subjectivism and relativism: if moral norms and values are not fully objective and real, then we have in the end no reason to act morally. Thus he writes, ‘The opposition here is between a theory about how things are and a practice that would be impossible if this was how things are’ (see p. 106 in Nagel 1997).

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

29

Even so, there is something very unsettling about Nagel’s inference from our inability to think of our forms of reasoning as less than fully objective to their actually being fully objective. How can our inability to see the flaws in our own (practical reasoning) program possibly show that there are no such flaws? The inference looks like a simple ad ignorantiam. Still, Nagel is in good company in endorsing this inference. Frank Jackson has proffered a related response to Mackie’s moral skepticism using a priori principles based upon interpretative constraints: Any account of M (matured Folk Morality) which makes it all but impossible for any actions in the world as it is actually constituted to be right must be mistaken. It is a constraint on interpreting our moral discourse that some actions have the moral properties our discourse putatively ascribes to them (Jackson 1992).10

It is worth taking some time to explain why this particular a priori insight is flawed since, if it were right, the Hard Problem would simply dissipate. To begin, for realists, there is no direct inference from intelligibility constraints governing the interpretation of some discourse to a metaphysical conclusion concerning the nature of the world—that we perforce interpret moral discourse as quantifying over moral values in no way shows that there are moral values. At best, it shows that we folk cannot make sense of what our fellow folk are talking about unless we assume the existence of moral values. This says something about us, not the world. Of course, one must observe the rules underwriting moral attribution to engage in the practice of moral evaluation and dispute. But this is perfectly compatible with retaining a healthy skepticism about the entities and properties the discursive practice apparently quantifies over. L.E.J. Brouwer, the founder of Intuitionism, furnishes a good example of this phenomenon—Brouwer made major contributions to classical mathematics, all the while maintaining that it was based on error. If Jackson’s idea is instead that since an interpreter must engage in the very practice she seeks to interpret, she must posit in the metalanguage the very same theoretical entities those she interprets in the object language, this is simply false. This would imply that Quine and the Churchlands are ineluctably committed to the existence of propositional attitudes by their understanding of the discourse that quantifies over those attitudes, that we could not understand primitive tribes unless we believed in their gods and so on. Moreover, even if it were the case that interpreting propositional attitude discourse can only be achieved by attributing further propositional attitudes, or that interpreting moral attitude discourse could only be a matter of attributing moral attitudes, how would this provide an answer to the skeptic about either folk mentality or morality? Consider a comparable case. Some logicians have argued that the only way to justify classical inference rules in the object language is to use classical reasoning in the metalanguage.11 Clearly, this in no way shows that classical reasoning is itself justified in some absolute extrinsic sense. Indeed, opponents of classical logic such as Michael Dummett acknowledge this feature and use it as a ground of criticism—intuitionistic inference rules can be justified in a classical as well as intuitionistic metalanguage, whence the meanings of the intuitionistic logical constants

30

D. Khlentzos

can be conveyed without circularity, or so some intuitionists hold.12 Jackson seems to mistake the imposition of internal consistency requirements for the satisfaction of external correspondence relations. Still, the idea that you can’t interpret discourse D unless most of the D-terms and D-predicates refer to the things the D-users take themselves to be referring to, or more strongly, unless the things the D-terms refer to have the properties the Dpredicates refer to, is one that seems to have some currency amongst philosophers. As far as I can see, there is no reason in the world to believe it. We have no difficulty interpreting the language of mediaeval cosmologists whose beliefs about the cosmos were widely false. Even in cases where the D-terms and D-predicates clearly do refer to their intended referents, interpretation can be successfully accomplished without imposing the extraneous requirement that the D-objects actually have the D-properties. Once objects have been assigned to terms and properties to predicates, semantic interpretation can be considered completed, at least for the purposes of understanding what D-speakers are talking about in the most obvious sense of ‘talking about’. At best, one might require, a la Quine (see Chapter 2 in Quine 1960) and Davidson, that unless the D-sentences describing simple and obvious matters turned out largely true by the lights of the interpreter, interpretation would be stymied in the sense that no accompanying propositional attitudes could plausibly be attributed to the D-users owing to the breach in rationality that would also have to be imputed to them in order to explain their cognitive failure. But both Quine and Davidson are quite happy to attribute systematic error to D-speakers, provided that it is rationally explicable from the interpreter’s point of view. The latter point highlights a crucial difference between Donald Davidson’s argument against general skepticism and Jackson’s related but more specific argument against moral skepticism, both being based on putative facts about interpretation. In an effort to secure the overall truth of our most basic beliefs about the world, Davidson appeals to the artifice of an omniscient interpreter constrained to use, as human interpreters perforce use, rationality constraints to determine which attitudes and meanings may justifiably be ascribed to the agent. Since the rationality constraints dictate that our basic beliefs turn out true by the interpreter’s lights and since the interpreter in this case is omniscient, most of our basic beliefs are indeed true and the skeptic is stultified (Davidson 1984). Absent the omniscient interpreter and Davidson has no reply at all to the skeptic. But Jackson makes no such appeal. What then secures the overall truth of the folk theory of morality? Why couldn’t it be based on a systematic but seductive error just as Mackie argued?

1.4 Meta-Ethical Pyrrhonism We are finally in a position to appreciate the type of moral agnosticism that can be used to undercut Nagel’s argument for moral realism. I should first say that I do not see Nagel’s argument for the objectivity of moral reasoning as in any way

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

31

fallacious. Rather than appealing as Jackson does to truths that can supposedly be known a priori, Nagel is offering an inference to the best explanation: what possible explanation could there be for our inability to think of our first-order moral evaluations as anything other than fully objective, than that they are fully objective?13 Like all inferences to the best explanation though, this argument fails if there is a better or at least equally plausible alternative explanation. In this case, I think there is. The leading idea behind Meta-Ethical Pyrrhonism (MEP) (and the reason for the name) is that neither the available arguments for moral realism nor the arguments against are in the end rationally compelling—the strongest reason for thinking that there are real objective moral properties is that our practice of moral reasoning presupposes their existence and lapses into incoherence if there aren’t any. The strongest argument against real moral properties is that no natural property could discharge the normative role required by our moral practices: there are no objectively and intrinsically prescriptive natural properties in the world. However, according to the Pyrrhonist, there is no breaking the dialectical stalemate since neither set of judgments can be given up. Hence, like Pyrrhonism of yore, MEP holds that first-order folk moral judgments presupposing moral properties are real are exactly counterbalanced by naturalistic second-order judgments that moral properties are fictions/ constructs/illusions. The reason there can be no present resolution of the tension between first-order and second-order judgments is that whilst moral practice is indispensable to us, fictional moral properties seem unable to ‘save all the moral phenomena’: in particular no ersatz moral property can give us a reason to act morally when it goes against our best interests to do so. The upshot here is that for the theorist, interactive situations wherein moral norms can secure cooperation between folk in the teeth of mixed or conflicting interests now become potential PD situations or else situations where the theorist has no reason not to unilaterally defect, even if the other agent believes defection to be intrinsically and objectively wrong.14 MEP is thus a form of meta-ethical agnosticism that can nonetheless supply the theorist with a reason for acting morally in certain Pareto sub-optimal situations: According to the Pyrrhonist, we should suspend judgment on whether moral properties are real or not, but we should still act as if moral norms and laws were intrinsically and objectively prescriptive. MEP is not moral skepticism at least if that is taken to mean that one cannot justify belief in any normative moral claim such as ‘incarcerating children is wrong’. Neither is MEP the claim that there is no fact to the matter as to whether moral properties are real or that it is impossible to discover whether moral properties are real. MEP is instead the thesis that it is as of now undecidable whether moral properties are real or not. It is indexed to the information and theories that we currently have at our disposal (or foreseeable extensions thereof). Further information about the human brain or the evolution of cooperative behavior and reciprocity in hominids might supply essential clues that could be used to break the epistemic deadlock. Despite being a form of moral agnosticism, MEP claims that we should acquiesce in the undeniable appearances of objective moral prescriptivity and judge and act in accord with folk morality. Yet we might well wonder why we should do this.

32

D. Khlentzos

A religious agnostic has no reason to act/judge religiously, it might be argued, so why should a moral agnostic act/judge morally? The reason has to do with the very agnosticism that underwrites the Pyrrhonist’s position: in a situation of radical uncertainty between two alternatives, if on either alternative one is better off in pursuing a certain course of action, then one should pursue that course of action. The relevant course of action is to obey the central moral norms the rest of the folk obey. If these are real and objective, one will have been saved from committing acts that are morally wrong; if they are not, one should still obey them since one is instrumentally better off in acting as if they are.15 Pyrrhonists believe that there is a 50% chance that there are objectively prescriptive moral norms and a 50% chance that the appearance of objective prescriptivity for moral norms is an evolutionarily useful illusion ‘designed’ to secure social cohesion. Then to ‘cooperate’ is to judge/act in accord with the precepts of Folk Morality. To defect is not to do so. The situation can be modeled game-theoretically as in Fig. 1.1 below with the agents radically uncertain as to whether the actual choice situation they face is a Prisoner’s Dilemma (PD) or one wherein there are real moral norms. Roughly, if the second-order naturalistic arguments are correct, we are in a PD situation; if, on the other hand, folk first-order moral judgments are objectively true, we are in a moral realist (MR) situation.16 That Pyrrhonists are indeed better off by acting in accord with the dictates of folk morality can be seen from the final payoff matrix calculated for the situation of radical epistemic uncertainty in which they find themselves, as in Fig. 1.2. The cells of the matrix in Fig. 1.2 are calculated using a 50% contribution from each appropriate cell in the two matrices represented in Fig. 1.1. The solution to the game is then for each agent to cooperate by observing the norms of folk morality, joint cooperation comprising the Nash Equilibrium.17 An interesting feature of this game can be used to model the aspect of moral practice which so impressed Nagel: our inability to think of our moral judgments and decisions as anything other than fully objective. For whilst defection will reveal to an agent which situation s/he is in, defection is instrumentally irrational.

Nature

0.5

0.5 MR

PD

C

Fig. 1.1 MEP choice situation

D

C

D

C

(85, 85)

(0, 100)

(85, 85)

(0, –100)

D

(100, 0)

(10, 10)

(–100, 0)

(–190, –190)

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism Fig. 1.2 MEP payoff matrix

33 Cooperate

Defect

Cooperate

(85, 85)

(0, 0)

Defect

(0, 0)

(–90, –90)

Cooperation through observance of the relevant moral norms, on the other hand, which is the rational course of action for each agent, will not reveal to the agents which situation they are actually in. Those who cannot resist the temptation to indulge their epistemic curiosity might discover they are really in a PD situation. Nonetheless, it will be irrational to attempt to discover this fact in any situation where the assumption that moral norms exist works to the benefit of all agents. Hence, even if they are fictional constructs, moral norms can still bind the theorist so long as their fictional status remains hidden during moral transactions and cannot be subsequently demonstrated in reflection on the nature of those transactions. Thus is the Ring of Gyges objection disarmed by the Pyrrhonist. It may be thought that even if there are sound instrumental reasons for cooperating with the dictates of folk morality, there can be no basis for a moral agnostic to endorse any moral judgments. Surely this, at least, is not in the spirit of the classical pyrrhonist who sought release from the vexatious tug of conflicting judgments by refusing to tender any? The latter point is right in one sense. Meta-Ethical Pyrrhonists do differ from their classical forebears in issuing moral judgments and engaging in debate with a view to coming to some settled opinion about what ought to be done (rather than what appears as if it ought to be done)—that is, what first-order moral judgment or course of action one should endorse. They agree with their forebears, of course, in thinking that no settled second-order judgment can be reached as to the metaphysical nature of moral norms or values. The justification for coming to an opinion about first-order judgments is one that classical Pyrrhonists did not consider, although Pascal most certainly did. Suppose with the Pyrrhonist that there is a 50% chance that moral norms and values are real and a 50% chance that they are instead evolutionarily useful illusions ‘selected for the purpose’ of counteracting limited sympathies amongst social animals. Suppose next that cooperation with those norms is beneficial in either case, as Pyrrhonists aver. Then it does not matter that the ‘ought’ as it features in the judgment ‘no child ought to be incarcerated’ is ambiguous between the deontic sense and the sense of practical rationality, the judgment comes out true on either reading. So, finally, we have an answer to Nagel’s inference to the best explanation argument for moral realism. There is in fact a better explanation of our inability to think of our moral reasoning as anything other than objective: it is not in our best interests

34

D. Khlentzos

to inquire about the reality of the putative objective norms that mediate various social transactions when we are engaged in these transactions, and since subsequent second-order naturalistic reflections on our first-order moral judgments are nowhere near robust enough to offset the hold of the latter on our decision-making or the benefits that accrue to acting upon those judgments, we should cooperate with the dictates of these moral precepts in spite of our uncertainty. The effect will be the same: it will be irrational to treat moral norms in the context of decision-making as anything other than fully objective, even if it should prove theoretically impossible to subsequently think of those same norms as anything other than non-natural and thus non-objective properties from the naturalistic point of view.

1.5 Conclusion My aim has been to undercut a persuasive argument for moral realism, not to promote an alternative that I confess to finding philosophically unsatisfying. For to the extent that one wishes to know what types of things the world contains, to the extent that one is interested in metaphysics in other words, one will not rest content with any form of moral agnosticism and will seek to advance new considerations or review old ones in an effort to decide whether moral properties are real or not. However, not every philosopher shares a passion for metaphysics. One who most assuredly does not is Bas van Fraassen who professes to reject the very project of metaphysics on empiricist grounds.18 Someone sympathetic to van Fraassen’s views might endorse Meta-Ethical Pyrrhonism as a stable and defensible position about the irrelevance of moral ontology to moral theorizing on the grounds that Moral Realism poses a pointless epistemic risk, following van Fraassen’s argument for Constructive Empiricism: If I believe a theory to be true and not just empirically adequate, my risk of being shown wrong is exactly the risk that the weaker entailed belief will conflict with actual experience. Meanwhile by avowing the stronger belief, I place myself in the position of being able to answer more questions . . . having a richer fuller picture of the world . . . a wealth of opinion I can dole out to those who wonder. But since the extra opinion is not additionally vulnerable, the risk is . . . illusory and so is the wealth. It is but empty strutting and posturing, this display of courage not under fire and avowal of additional resources that cannot feel the pinch of misfortune any sooner (see p. 255 in Churchland and Hooker (eds) 1985).

‘Empirical adequacy’ in the moral case will simply amount to a vindication of the central folk moral evaluations. So the suggestion is that a realist meta-ethics is a pointless epistemic risk caused by an unwelcome intrusion of metaphysics into a practice that is definitive of our humanity. Those less hostile to metaphysics than van Fraassen might independently endorse Meta-Ethical Pyrrhonism as the only stable meta-ethics consistent with naturalism. For if it really is true, as Nagel argues, that we cannot surrender our firstorder moral judgments or trade them in for subjectivist or relativist ersatzes, if we really are forced to regard our moral evaluations as intrinsically and objectively

1 Moral Realism, Meta-Ethical Pyrrhonism and Naturalism

35

prescriptive every time we issue them, and yet we can provide no subsequent naturalistic explanation of a property so strange as objective prescriptivity, the most plausible naturalistic response to this situation may well prove to be the Pyrrhonist’s: since our moral evaluations and our naturalistic reflections cannot be brought into harmony with each other and since neither set of judgments can be given up, as naturalists we have no choice but to remain agnostic about the reality of moral properties.

Notes 1 To the extent that they have not realized that the semantic attacks of antirealists on metaphysical realism are driven by a conviction that the Representation Problem is insoluble for realists. For the argument supporting this claim, see Khlentzos (2004). 2 I have omitted from the quote the word van Fraassen is actually discussing which is the word ‘part’ but his point is equally applicable to the word ‘wrong’. 3 For a compelling defense of the hypothesis that convulsive ergotism may have been the physiological basis for the Salem witchcraft crisis of 1692, see Capporeal (1976). 4 My conjecture is that fixation on the water-H O model has led to complacency amongst naturalists that 2 the Hard Problem is no harder in the moral case than the water case. I shall explain why this model is a highly misleading one for a moral naturalist to appeal to. 5 Once more, though, no mere mapping between the two is sufficient to establish the identity of natural property with folk moral property. 6 Although only prima facie since the fact of moral disagreement may be attributable to disagreement about the non-moral facts. 7 For recent surveys of this evidence, see Bechara et al. (2000) and Dolan (2002). See Damasio (1994) for an accessible popular presentation. 8 For a representative selection of papers on this theme which includes papers from Kim, Burge, Rudder Baker and van Gulick, see Heil and Mele (eds) (1995). 9 Quoted on pp. 60–61 in Kim (2000). 10 In fact, although I think Jackson’s argument is fallacious, I do not believe Nagel’s is. Jackson’s appeal to what can be known a priori is the difference between the two. 11 Most notably Michael Dummett. See Dummett (1991). 12 Cf. Dummett (1991). 13 Qua moral naturalist, Jackson is no more entitled to assume a priori that the extensions of ‘right’ and ‘wrong’ are non-empty than our fictitious theorist Nat is entitled to assume a priori that the extensions of ‘cornucopian’ and ‘pandoran’ are non-empty. 14 Indeed, as everyone knows, if the situation is a PD situation and the other agent is most likely to cooperate out of a sense of duty or a misplaced expectation of reciprocity or because s/he believes defection to be immoral, this supplies the canny theorist not bound by moral norms with the best possible reason to defect. 15 Readers will recognize the Pascalian character of this reasoning. Pascal, of course, famously denied that agnostics have no self-interested reason to act religiously. 16 Thanks to Arcady Blinov for making me aware of this choice-situation. See Blinov (2003) for a discussion of the use of such choice-situations in modeling the strategic benefits of what he sees as certain types of cognitive irrationality. 17 See Khlentzos (2004) for further elaboration and discussion of this choice-situation. 18 See his defense of this view in van Fraassen (2002).

36

D. Khlentzos

References Bechara A, Damasio H, Damasio AR (2000) Emotion, decision-making and the orbitofrontal cortex. Cerebral Cortex 10(3):295–307 Blinov A (2003) Rationalities in conflict: compensatory logico-cognitive irrationality in interactive contexts. Yearbook of logical investigations 10:233–241, Nauka, Moscow Burge T (1995) Mind-body causation and explanatory practice. In: Heil J, Mele A (eds) Mental causation. Clarendon, Oxford Burgess JA (1998) Error theories and values. Australas J Philos 76(4):534–552 Capporeal LR (1976) Ergotism: the Satan loosed in Salem? Science 192:21–26 Churchland PM, Hooker CA (eds) (1985) Images of science: essays on realism and empiricism, with a reply by Bas van Fraassen. University of Chicago Press, Chicago Damasio AR (1994) Descartes’ error: emotion, reason and the human brain. Grosset/Putnam, New York Davidson D (1984) The method of truth in metaphysics: Essay 14. In: Inquiries into truth and interpretation. Clarendon, Oxford Dolan RJ (2002) Emotion, cognition and behavior. Science 298:1191–1194 Dummett M (1991) The logical basis of metaphysics. Duckworth, London Heil J, Mele A (eds) (1995) Mental causation. Clarendon, Oxford Hurley S (1989) Natural reasons: personality and polity. Oxford University Press, Oxford Jackson F (1992) Natural reasons: personality and polity by Susan Hurley—review essay. Australas J Philos 70(4):475–488 Khlentzos D (2004) Naturalistic realism and the antirealist challenge. Bradford Books, Cambridge MA Kim J (2000) Mind in a physical world. Bradford Books, Cambridge MA Mackie JL (1977) Ethics: inventing right and wrong. Penguin, Harmondsworth Nagel T (1997) The last word. Oxford University Press, Oxford Quine WVO (1960) Word and object. The MIT Press, Cambridge MA van Fraassen B (2002) The empirical stance. Yale University Press, New Haven

Chapter 2

Buck-passing Personal Values Toni Rønnow-Rasmussen

Abstract So-called fitting-attitude analyses or buck-passing accounts of value have lately received much attention among philosophers of value. These analyses set out from the idea that values must be understood in terms of attitudinal responses that we have reason to or that it is fitting or that we ought to have regarding the valuable object. This work examines to what extent this kind of analysis also can be applied to so-called personal values – value-for, rather than to the impersonal value period which has been the standard analysandum. The shift from impersonal to personal values can, it is argued, be taken without any major change to the pattern. It is not the normative element that needs to be changed in the analysis but rather the kind of attitude – what is required is that the attitudes all have to be so called ‘for someone’s sake’ attitudes. Keywords Buck-passing · Fitting-attitude analysis · Good-for · Personal values · Value-for According to an influential approach to value analysis, what I will refer to as the Fitting-attitude analysis (FA, for short), being valuable is being a fitting object of a pro-attitude. The key idea behind FA is to analyze axiological concepts such as ‘value’ and ‘goodness’ in terms of deontic notions such as ‘fittingness’, ‘oughtness’, or, as more modern proposals have it, in terms of reason. Prominent advocates are Franz Brentano, C. D. Broad, A. C. Ewing, and more recently, Thomas Scanlon (1998). In this work, I want to argue that the FA analysis can be brought to bear on personal values as well. All we need to do is to formulate the analysans in terms of a special kind of attitude. In Section 2.1, I outline the characteristic features of the FA analysis. Section 2.2 discusses a suggestion as to how personal values might be understood, which I believe to be inferior to the one I put forward and examine in Sections 2.3 and 2.4. Finally, in Section 2.5, I consider some objections to my proposal. T. Rønnow-Rasmussen Department of Philosophy, Lund University, Kungshuset, Lundag˚ard, 222 22 Lund, Sweden Tel.: +46-46-222 9486, Fax: +46-46-222 4424 e-mail: [email protected] D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

37

38

T. Rønnow-Rasmussen

2.1 Fitting-attitude Analysis FA analyses reduce the evaluative to deontic claims about attitudes towards the valuable objects in question that it is fitting or that one ought to or that we have reason to take up towards the objects. A noteworthy feature, according to the more interesting versions, is that it is properties other than the value property that provide reasons to respond to the thing by taking up some attitude. The reason why I ought to take a certain responsive stance is to be sought among the subvenient properties and not in what supervenes on those properties. Following Scanlon (1998), we could say that what carries the buck is not value but what value supervenes on! In other words, for object O to be valuable is for it to have the property of having a value-making base that gives us reason to hold a pro-response vis-`a-vis O. Most versions of the FA analysis, including the one defended in this work, are ‘value-bearer pluralists,’ i.e., they acknowledge that final (nonderivative) value may accrue to different kinds of objects, such as abstract states of affairs and concrete objects (like persons and things).1 Pluralists stress that different kinds of valuable objects invite different genera of pro-responses. For instance, a brave person ought to be admired, a precious artifact calls for care and protection, and a desirable state should be desired. The response in many cases may actually be more accurately described as a conglomerate of different kinds of attitudes or even attitudes-cumbehaviors. Some attitudes are especially thing-oriented: we care for, say, our children; we cherish and preserve an object of historical importance, say, Napoleon’s hat; or we protect a part of the Brazilian rainforest that is untouched by humans. Other attitudes are rather state of affairs–oriented. The most obvious example is desire, which typically takes a state of affairs (e.g., that Sweden will qualify for the World Cup in soccer) as its object rather than a (concrete) thing or person. The FA analysis has typically been employed to analyze so-called final values of objects, i.e., what is valuable ‘as an end’ or ‘in its own right’ or ‘for its own sake’. However, none of these ‘idioms’ are in fact perfect. That something is valuable as an end, for instance, appears not to be consistent with the idea that the value bearer may be a non-propositional object, such as a thing (painting, stamp, etc.). It seems right, at least, to agree with Ross that ends are objectives and never things. In this sense, the two latter expressions seem to be better. Both ‘in its own right’2 and ‘for its own sake’ imply something about whether the value is dependent or not on something other than the valuable object. Of these two expressions, I prefer ‘for its own sake’. Admittedly, the advantage of this idiom may not be immediately obvious. There is a somewhat counterintuitive ring to this notion when it occurs in expressions such as ‘x is valuable for its own sake’ (in comparison to, say, ‘x is desired for its own sake’). Notwithstanding, I prefer this latter alternative for the following reasons: It is a natural counterpart (in a sense which ‘in its own right’ is not) to two other familiar kinds of value, viz., contributory value and instrumental value, i.e., what is valuable for the sake of some whole or for something else’s sake rather than for its own sake. Finally, ‘valuable for its own sake’ is not biased with regard to the value-bearer issue; it does not rule out either abstract states or concrete things as was the case with at least the first alternative. Therefore, in what follows I will speak of final values

2 Buck-passing Personal Values

39

in terms of what is valuable for its own sake. However, my aim here concerns in a sense only indirectly final values. My focus will be on (final or non-final) personal values. The intuitive idea behind the dichotomy personal/impersonal values may be stated in the following way. Certain evaluations are made with no eye to anything but to the valuable thing itself. For instance, knowledge has been said to carry such value (period). But sometimes our evaluations are quite different; they concern some object but are in fact very much done with a certain person in mind. For instance, I value in this sense some drawings made by my daughters. These sketches hardly carry an impersonal value, but they do have value for me. What I want to discuss is to what extent some version of the FA analysis, if true, can be helpful when it comes to understanding values of this latter kind. The above proviso, ‘if true,’ must be underlined. Recently, Wlodek Rabinowicz and I (2004) have argued that despite its many advantages, FA faces some serious problems.3 However, my aim here is not to solve these problems. I have a much more modest end in sight, namely to argue that if this analysis overcomes these problems, it will help us understand the difference between impersonal and personal values. For reasons of simplicity, let ‘favor’ be a schematic place-holder for different pro-responses that are called for by different kinds of valuable objects. Applied to final values, FA says, as we have seen, that an object is finally valuable if there are reasons to favor it for its own sake. An object of instrumental value is then said to be instrumentally valuable if there are reasons to favor it for the sake of its effects. Final values understood in terms of FA suit examples of impersonal values well – at least if by ‘impersonal’ we have in mind values that are not related to any particular person or persons. After all, it is in virtue of the non-evaluative properties of the valuable object – and not some property of, say, the evaluator – that we should take an attitudinal stance, and these properties will call for such a response regardless of who we are. This is obviously the case when final value accrues to the object in virtue of its internal properties (a so-called intrinsic final value). It is also the case when we turn to so-called extrinsic final value – where the final (in contrast to instrumental) value supervenes on an object in virtue of at least some of its (noninternal) relational properties. The spirit of the analysis prompts us to regard such values as calling forth that one ought to favor the object for its own sake in virtue of its non-evaluative intrinsic as well as extrinsic properties. Just think of Beardsley’s example of a stamp that is unique or Kagan’s discussion of the pen Abraham Lincoln used to sign the Emancipation Proclamation. The uniqueness of these objects is an obvious (non-internal) relational property of a potent value-making character.4 The uniqueness gives us a reason to, say, cherish and protect the stamp and the pen.

2.2 A First Attempt at an Analysis of Personal Value It might be thought that the notion of extrinsic final values suggests an analysans. According to this analysis, which I will refer to as the De Facto Attitude analysis of Personal Value or DEFA, for short, personal values are a special subgroup

40

T. Rønnow-Rasmussen

of extrinsic values. The idea behind DEFA is basically not a good strategy for analyzing personal value. However, it might prove useful to glance for a moment at an alternative that goes wrong. DEFA is arrived at by the following path: If final values can be extrinsic, i.e., accrue (at least in part) to objects in virtue of their non-internal relational properties, then it will be expected that some object of extrinsic value will carry this kind of value in virtue of its relation to a person. Of course, being related to some person is not a sufficiently precise description to render DEFA plausible qua account of personal value. But once we have taken the step of relativizing value to persons in this way, we have on this analysis at least paved the road for personal values: what we need next is to specify the relata more carefully on to which personal value accrues. We need to do so in order to distinguish personal value from other kinds of extrinsic values that accrue to objects in virtue of being merely related to a person. Intuitively, it seems credible that personal value refers to something more than merely a value that is related to a person. We need, that is, to analytically separate from each other cases like, for instance, the following ones: Napoleon’s hat, and a drawing made by my 5-year-old daughter. It is plausible to argue that Napoleon’s hat has some extrinsic value, and it seems plausible to argue this whether or not the hat had in Napoleon’s view any (personal or impersonal) value. Even if Napoleon never endorsed any evaluative judgment whatsoever about his hat, it might still be of value due to the fact that it was Napoleon’s well-known hat. In the case of my daughter’s drawing, it is at least much more unclear that it has any impersonal value. However, for my own part I can assure you that I at least cherish the drawing. Of course this is no proof that the drawing has personal or any other kind of value. Nonetheless, a drawing by a child which is cherished by the parent seems to be a plausible example of something that carries personal value. Suppose we accept these examples. The question then is, how should we keep the cases apart? What makes the first example into one of ‘impersonal value’ and the other one into one of ‘personal value’? The role which Napoleon played in the past, and my and my daughter’s historical insignificance may account for why the hat carries impersonal value, and why the drawing does not. But these facts do not account for why the drawing and not the hat has personal value. Here we need to look for something else. A natural suggestion would be to search for something special about the two objects involved in the drawing-example (i.e., me and the drawing). On the face of it, a likely suggestion is to say that the person, a, i.e., the person for whom, ex hypothesi, the drawing is a personal value, as a matter of fact favors the object. DEFA could but need not be cast in some kind of fitting-analysis mold, in which case it would say something to the following effect: O has personal value for a if and only if we ought to favor O in virtue of the extrinsic property that O is de facto favored by a. Thus, my daughter’s drawing would have personal value only if the value accrued to it in virtue of being an object that is favored (say, cherished, esteemed, etc.) by me, a. Thus on DEFA, we need to add that the person to whom the object is of personal value also de facto has certain attitudes towards the valuable object.

2 Buck-passing Personal Values

41

But despite whatever intuitive appeal DEFA has, it still establishes at most that the value accruing to the drawing is a relative value. And that is not specific enough if it is personal values we are looking for. The inference from ‘x is a personal value’ to ‘x is a relative value’ is valid – reversing the entailment is not. Actually, there are two further related reasons why DEFA should be regarded with skepticism. First, it identifies personal values exclusively among extrinsic (final) values. But this seems at first sight to be an arbitrary restriction. An account that also admits personal value the status of intrinsic value (i.e., the kind of final value that accrues to objects in virtue of their internal properties) is more appealing. Second, it is not obvious that personal value must be identified with what is valued by persons. Convincing cases could be made for the idea that although an object is not valued (esteemed, cherished, etc.) by a particular person, it would nonetheless be correct to say that the object has personal value for him.

2.3 Fitting-attitude Analysis of Personal Value The approach that I outline next carries, I think, intuitive appeal, qua analysis of personal value. In part the attractiveness of this kind of analysis derives from its idea that personal value calls forth a very particular range of attitudes. There is something about these attitudes that makes us understand why objects related to these attitudes carry personal and not impersonal value. This is one important advantage of this analysis. Another pro-feature is that it elucidates why we are justified in our concern for objects that are valuable for us; if value just is, as FA has it, the existence of reasons for such a concern, the justification is immediately forthcoming. A further advantage is that its appeal is not conditional on our taking sides on a controversial issue such as the subjectivism/objectivism question. The analysis is basically open to both possibilities. So far we have mentioned mainly two kinds of pro-responses: (i) x favors O for its own sake. (ii) x favors O for some other object’s sake. . .

To introduce the analysis, I suggest that we should add to this list of final and instrumental attitudes, i.e., the attitudes referred to in (i) and (ii), the following entry: (iii) x favors O for y’s sake.

The idea that we favor things for our own sake is, I take it, uncontroversial (I may, say, desire to know whether or not I have a fatal illness, for my own sake, but be absolutely indifferent to whether someone else ought to have this knowledge). The real nub of the issue concerns something else, namely to what extent ‘x favors O for y’s sake’ is shorthand for (iv) ‘x favors O for its own sake, for y’s sake,’ or (v) ‘x favors O for some other object’s sake, for y’s sake’?

Suppose, to simplify matters considerably, that there are only final and instrumental attitudes. Then (iii) must be an example of either a kind of final or instrumental

42

T. Rønnow-Rasmussen

attitude, or not be an attitude at all. But is it at all possible to favor, for example, O for its own sake, for a’s sake? And if it is possible, how are these attitudes best understood? At least two features of (iv–v) complicate an evaluation of these suggestions. First, we must not forget that ‘favors’ is here a technical term. So what might be possible or impossible with regard to one kind of attitude, say, ‘desire’ may be possible or impossible with regard to, say, ‘admire, respect’. We should therefore keep a lookout for questionable generalizations on this matter. Second, that the formulations in (iv–v) are somewhat obnoxious by their very cumbersome nature might make us reluctant to accept them as coherent. But, of course, the fact that we need a cumbersome formulation to catch a complex attitude is in itself no reason to dismiss these attitudes on logical grounds. I will assume in this work that there are these kinds of complex attitudes. There is, in fact, ample evidence that we describe ourselves as having these kinds of complex attitudes. For instance, when we treat some stranger amicably for a friend’s sake – the stranger is, say, a friend of our friend – we are exhibiting this kind of attitude. Or when I go to the cinema, I often do so for my wife’s sake. Emotions too seem open to this sort of construction. For instance, on learning that someone whom we care about has had good fortune, we are glad for his or her sake. Many more examples of attitudes, attitude-cum-actions, and emotions could be given. Although I am assuming that we can have these attitudes, I am far from sure how they are best analyzed. Displaying this kind of attitude seems to involve our ability to, in a sense, harbor attitudes for someone else’s sake. This, by the way, seems to be an important part of what it means to be empathetic. To recognize that something is of personal value to x is, on this line of reasoning, to somehow echo a favoring that this x has or ought to have vis-`a-vis the object of value. One might object that taking this route will strand us with two notions of personal value; one employable by the person for whom something is a personal value, and another for those who recognize that something carries this person’s personal value. Actually I am not sure that this is really the case. The attitude that I am ready to take towards, say, the stranger who is a friend of a friend of mine – call him a – has something to do with my friend’s attitudes towards the stranger. But it is hardly his de facto favoring that I will take over. Rather, it seems to me to be the kind of attitude that it is fitting for a to take in those circumstances. If this is the case, I see no reason why a’s claim that something has personal value should not be analyzed in the same way. These speculations apart, one can easily expect – with the focus on final rather than instrumental pro-responses – the following response: Make up your mind – either you favor O for its own sake or you favor it for your own sake. You cannot have it both ways. Actually, I think we can have the cake and eat it too. What I want to suggest is that ‘sake’ is in fact ambiguous: it does not necessarily have the same content in ‘for its own sake’ as it has in ‘for a’s sake.’ Speaking about a’s sake and O’s sake might suggest that I am ascribing, at least to the former, a sake or an end, in the sense of an objective. But, as I suggested at the outset, the kind of final value which I am interested in accrues to objects that often cannot be understood as having

2 Buck-passing Personal Values

43

objectives (e.g. like my daughters’ drawings). Therefore, in the case O is carrying a final personal value, the first ‘sake’ in ‘favoring O for its own sake for a’s sake’ indicates that the value is not dependent on being conducive to something else that is valuable. Final value is perhaps best described then as an end-point value or an ultimate value, the sort of value not determined by something else being valuable. And similarly, the second ‘sake’ indicates that the favoring attitude is not instrumental; our attitude is for a, and it is final in the sense that it is not dependent on some other favoring. An attitude having the attainment of some objective as its object will only be relevant to an analysis of personal final value in case the attainment of the objective is favored for the person’s sake. That (iv) and (v) might appear problematic has perhaps to do with the following: it is often thought that the way to individuate an attitude is to determine some proposition that is conceptually related to the attitude.5 Suppose, then, that in the case of final attitudes like the one in (i), we need, in order to describe this so-called propositional element, the notion ‘for its own sake’; in such a case there just might be a problem with idioms such as the one in (iv) above. I am not saying there is a problem. But at least it would seem as if ‘to favor O for its own sake for a sake’ would risk coming out as a logical mishmash. Whether or not the above-mentioned risk is real or not, there is, as far as I can see, no compelling reason that final attitudes have to be understood in this way, and especially, there is no compelling reason that these attitudes, even if they actually had to be understood in terms of some propositional element, must have a propositional content corresponding to the above description, i.e., a description formulated in terms of ‘for its own sake’. There are more ways to understand what I will refer to as final and instrumental ‘favoring’, i.e., those examples of favor-attitudes referred to in (i) and (ii) above, than by ‘for its own sake’ propositions. For my purposes here – which is to pave the road for a pattern of value analysis applied to personal values – it suffices to say that a final ‘favoring’ of O (i.e., that we favor O for its own sake), is the kind of favoring we have when O appears to us as being valuable for its own sake. It must be noticed here that O, of course, need not be carrying a final value. The attitude is explicated in terms of an evaluative judgment to the effect that the object of the attitude is valuable for its own sake. Thus, the point centers on the idea that O appears to us as valuable for its own sake and not that the object is as a matter of fact valuable for its own sake. Hence, I shall say that person a has a final attitude vis-`a-vis O if and only if O appears to him as valuable for its own sake. Similarly, an instrumental attitude should be understood in terms of a judgment to the effect that O is valuable for some other O ∗ ’s sake (that has final value).6 This naturally prompts the reply that an analysis of final value in terms of this kind of final attitude becomes circular. Certainly, this is true, but it is not a reason to reject the analysis. Not all circles are malignant. The one facing us here belongs to the benign kind; the analysis is informative for anyone who knows what an evaluative judgment is. The circle would pose a serious threat where there is no other means to determine the judgment ‘O is valuable for its own sake’ than by resorting to an analysis of (final) value. But this is not the case. There are different ways that

44

T. Rønnow-Rasmussen

we can determine what judgment a person endorses (ranging from the obvious one of simply asking the person, to psychologically and perhaps even neurologically more sophisticated alternatives). The point is that we do have independent means to determine judgments without having to recur to analyses of the notions that these judgments contain.7 If this is true about final and instrumental favoring, I see no reason why it would not also apply to the more complex kind of favoring we have when we favor an object for some person’s sake. Here the judgment will be slightly different. Let ‘valuable x ’ refer to value-for x, i.e., what is of personal value for x. The judgment involved with this kind of attitudes would then be something of the following sort: ‘O is valuablex for its own sake.’8 In what follows, I will, unless needed, speak of ‘favoring’ without specifying the kind of favoring involved (i.e., whether or not it is final or instrumental). This simplifying measure must not be looked upon as a fraudulent attempt to sweep away the complex problems behind the expressions ‘final attitude’ and ‘instrumental attitude’. I am first in line to recognize the need to establish a more precise meaning of this pair of notions. Nonetheless, to avoid the clumsy ‘favor O for its own sake/for something else’s sake’ I will keep to the simpler ‘favor O’.

2.4 ‘For a Particular Person’s Sake’ I have indirectly suggested that there is nothing in ‘sake’ that prevents us from favoring O for its own sake, for x’s sake – I have at least failed to determine a single precise content of ‘a’s sake’ that is not combinable with ‘for its own sake.’ It is, however, at least possible to wall in a range of bona fide properties of a, in virtue of which we can reasonably favor O for a’s sake whether or not the pro-response is final or instrumental. One suggestion is that favoring O for a’s sake is just favoring O for whatever contributes in some sense to improving (or, even looser, has something to do with) a’s, say, (i) present state of mind, (ii) his entire life, or (iii) some part of his life. The alternatives mentioned above could be referred to as three possible bearers of welfare. Naturally, there are other possible bearers of this kind of value. I am not taking a stand on whether welfare is applicable to only one or a combination of these (or some other) alternatives. That the set of personal values is in fact composed by one large group of such agent-relative ‘welfare’ values seems intuitively plausible. However, although this is one likely candidate, it is hardly the only one. Personal values, I shall say, consist of two fundamental categories of values, one of which holds those values(-for) the realization of which are welfare enhancing or preserving values. The second category contains a related but nonetheless different kind of value-for. Objects carrying these other kinds of values are what we ought to favor for a’s sake without it being necessarily the case that this somehow contributes to a’s welfare, even if welfare here is understood in quite a broad enough sense. In other words, ‘sake’ here does not mean ‘person-welfare-sake’. Despite the fact that favoring O would not have any welfare-contributing properties, values belonging to this category would nonetheless exemplify something that is good for a. But it might be objected, is that possible? Can we favor something for a’s sake that is not

2 Buck-passing Personal Values

45

welfare-promoting in some sense? A moment of reflection should show that the idea is in fact quite plausible. Consider the attitude ‘respect.’ Showing a person respect is not something one does, at least not necessarily, in order to contribute to this person’s welfare. Respecting someone might even in some cases have welfare-impairing effects. Respect has rather to do with what the person is or represents, and not with what you hope to achieve for the sake of this person’s welfare. For instance, once upon a time it was not uncommon that soldiers expressed respect for their enemies. This respect didn’t prevent them from wishing that the welfare of their enemies would come to a most abrupt and definitive end. Of course, many of these attitudes are obviously old-fashioned, and some are quite ridiculous. But this fact doesn’t change the picture that we conceptually can do things for someone’s sake that do not necessarily mean doing them for the person’s welfare. ‘Doing it for a’s sake’ must, however, in some minimal sense be counted as being in favor of a – in the sense that being respected is regarded as positive, while being ignored or disrespected counts against a person.9 Respect belongs to a group of attitudes that presents the starkest counterexamples to the idea that doing something for someone’s sake is necessarily doing it to preserve or promote someone’s welfare. But there are many more examples: to admire, honor, esteem, and like. Even a response like desire, which traditionally is intimately associated with welfare, need not always be understood as welfareregarding. Certainly, to have one’s desire satisfied seems to be a step in the right direction to fare well. However, although this might be true about certain desires, it is widely agreed today that it is not true about all desires. The reason can be as simple as that the satisfaction of a desire may have very little or nothing to do with the person’s welfare. This is what Parfit’s ‘Stranger case’ showed (see pp. 493–502 in Parfit 1984). But there is another well-known reason. Sometimes, we may think that a person lives an immoral (or inauthentic or irrational, etc.) life, and we might well desire that some states of affairs should obtain that would make his life moral (or authentic, etc.) despite the fact that it would have a negative impact on his welfare. We would still be doing it for this person’s sake.10 The above pieces of reasoning will, I hope, lend support to the idea that ‘sake’ in ‘O ought to be favored for a’s sake’ is unclear, and that there is conceptual space to be filled for the kind of complex constructions as the above mentioned one. We can next try our hands at an FA-style of analysis of personal value (henceforth, FAP): (FAP): O has personal value for x, if and only if there is reason to favor O for x’s sake.

Thus, the kind of attitude that is called forth according to the FAP advocate, when some O carries a personal value for x, is that x favors O for x’s sake. This approach is superior to DEFA. It has, I venture, a strong intuitive appeal, qua analysis of personal and not merely relative value. The object O not only has a value relative to some person. The object ought to be favored in a special way; it should be favored for some person’s sake. Undoubtedly there will be skeptics who do not share my intuition. Still, I venture to say, anyone who denies that the above has at least an intuitive advantage over DEFA has a hard time appearing reasonable.

46

T. Rønnow-Rasmussen

Like any FA analysis, FAP does not assume the truth or falsity of value subjectivism or value objectivism; FAP is quite open on this matter. Nor does FAP limit personal value to extrinsic value; that we ought to favor O for a’s sake may depend on the internal as well as the external properties of O. Another advantage of FAP is that it does not necessarily make personal value conditional on some de facto favoring. An object of personal value is one that we ought to favor for a person’s sake, and this might well be the case whether or not anyone actually favors the object for this person’s sake. This point is less counterintuitive than it may appear to be. Suppose I did not know what my parents looked like. I knew them only by name. Furthermore, assume that there existed a photo of my parents, unknown to me. Such a photo could plausibly be said to carry a personal value for me whether or not I knew about it. The point is not, however, that the value of the photo would be only hypothetical, say, conditional on me cherishing the photo. It might even be objected that all that is needed in order to account for the value in a case like this is not that I de facto or hypothetically have an attitude; what is required is that I am disposed to cherish the photo.11 The suggestion that value (whether personal or impersonal) can always be related to at least the (de facto) dispositions of an existing subject, is often voiced in order to avoid certain problems attached to the idea that value relates to desires, wants, etc. Notwithstanding, the suggestion is not convincing. A ‘dispositionalist’ would have to convince us that dispositions are for particulars and not for kinds. And as far as I can see, this would be very hard to argue. If I don’t know of, say, p’s existence (the photo), my disposition will be one of favoring a kind of object, viz., objects like p. It will hardly be a disposition to favor this particular photo p. Nothing changes by adding that the attitudes are hypothetical. Whether or not the photo has personal value is, on the analysis outlined here, a question of whether we ought to favor the photo for a’s sake, and the truth of this normative statement is not dependent on the truth of a hypothetical statement. But suppose a declared, once we had drawn his attention to the photo, that he couldn’t care less. Suppose further that it was actually true – and there is no reason to deny this possibility – a is absolutely indifferent to this photo. Does it follow that the photo then has no value – personal or any other kind of value? It might of course follow if we adhere to some kind of metavalue theory, according to which, say, values are constituted by attitudes (e.g. classical subjectivism, or more sophisticated non-cognitivism). But theories apart: Does it follow that the photo has no value – if a is indifferent to it? I don’t think so! It is obviously a substantive issue.

2.5 Some Problems As a fitting-attitude analysis, FAP faces various problems. Putting these aside, I will comment on some issues that concern FAP in particular. First, FAP is formulated in terms of a reason that ranges over all (or at least many) agents. One might object to this on the following ground. In contrast to other kinds

2 Buck-passing Personal Values

47

of value, personal value is first and foremost someone’s value, and so we need an argument to the effect that it is also, in a sense, everyone’s value. Admittedly, there is something to this idea. But not more than what can be accounted for by an FAP advocate. After all, this analysis does index a subject, and does in that sense ‘personalize’ the value – even if it does so in an impersonal way. There is a related possibility to be found in the drawers of the FAP analysis. A personal value might, so to speak, have more than one side – one facing the person for whom the value is ‘personal’, say, a, and another for all the rest of us. This might show itself in two ways: the attitude that it is fitting to take regarding the object may vary depending on, say, whether it is your daughter’s drawing or not; it may also vary with regard to the strength of the reason. Some reasons appear to have a stronger normative force than other ones and it is perhaps conceivable that one and the same object is Janus-like also with regard to this matter. But to what extent personal values are in fact in this or the former sense Janus-values is something that needs to be further examined. The idea seems intellectually inspiring. Nonetheless, I haven’t managed to come up with any really convincing cases. My suspicion is therefore that if personal values do have this ‘double-sidedness’, there are not many of the kind. Second, FAP is likely to be regarded as too demanding. Consider the following example (inspired by objections made to me by Jonas Olsson and Michael Brady): Suppose winning the New York marathon has personal value to Charlie. Furthermore, assume that you want to win this race too. Is it really the case that you should then favor his victory for his sake? Isn’t that just counterintuitive? Third, we do not confine ourselves to speaking about the welfare of people we care about. We also express judgments about the welfare of people that we are indifferent to, or that we even dislike. For example, although we dislike Hitler, many people are nevertheless inclined to say that Braun was good for Hitler. FAP appears to have counterintuitive results here. Suppose what Braun does that is good for Hitler (say, promotes his welfare) is doing y. FAP says then that we ought to desire that Braun does y for Hitler’s sake but, again, most of us could care less about what was done for Hitler’s sake. But although this might be correct, it is still plausible to say that Braun had personal value for Hitler. It could therefore be argued that FAP should be rejected because of being counterintuitive. Much of what can be said about this example can be said about the first case too. Let me therefore begin by commenting on the latter problem. There are in fact two quite familiar strategies open to an FAP advocate: one explains why these examples appear counterintuitive in such a way that the analysis does not come out as requiring that we hold actual pro-responses with an eye to Hitler. On the second approach, we are forced to swallow the bitter pill; if Braun has personal value for Hitler, then we should have pro-responses vis-`a-vis Braun for Hitler’s sake. Considering the first strategy, a combination of two well-known replies will be helpful: First, we must not forget that it is quite possible that Braun had no personal value for Hitler. But notice, to say this is not to deny that Hitler, de facto, did favor Braun. Apparently he did. But again, from the fact that he loved or cherished (etc.) Braun, it does not follow that she was a carrier of personal value. It only means that

48

T. Rønnow-Rasmussen

she was the object of some of Hitler’s pro-responses. But, it might be argued, this fact does not explain our intuition that we do seem ready to say, not merely that she was an object of his, say, liking, but that she had value for him. The explanation of why we seem to be ready to hold that Braun had value for Hitler is perhaps that we are so used to the idea of not paying attention to the metavalue issues that arise when we speak about personal values, the reason being that we take the following as obvious: If someone cherishes, desires, etc. an object, it has value for this person. But although people often reason like this, it is no reason to say that it leads to a correct position. I don’t see why personal value should be exempted from the standard issues that arise when we talk about (impersonal) value. Of course, the reason we want to say that Braun carries value-for need not be that we are metaethically na¨ıve. Perhaps the judgment ‘Braun is good for Hitler’ expresses what R. M. Hare called an inverted-commas judgment. What we say is that Braun has value for Hitler, according to Hitler. In other words, the ‘good-for’ in the judgment is parasitic on the genuine evaluation of Hitler. Both replies seem plausible. It is their generality that worries me. Is it really always the case that we are either metaethically na¨ıve or expressing some kind of inverted-commas judgment when we judge that something is good-for a person that we dislike? The second strategy acknowledges that the judgment ‘Braun is good-for Hitler’ is a genuine evaluative judgment; it handles the counter intuitiveness by underlining that the reasons involved here are so-called pro-tanto reasons (i.e., reasons that persist even in the presence of outweighing counter-reasons). Thinking that Braun is good for Hitler is thinking that we pro-tanto ought to have some pro-response regarding Braun for Hitler’s sake. Admitting this is, of course, not admitting that a person is all-things-considered worthy of our pro-responses. Both approaches are in fact combinable, which a fortiori strengthens the FAP advocate in his defense of the idea that FAP just might account for people’s personal values – whether or not they are malicious. Consider next the first example, the person who recognized that winning O has personal value to Charlie. Isn’t it counterintuitive to think that we have a reason to favor O for Charlie’s sake? Before giving in to this objection, we should investigate to what extent the judgment ‘O has personal value for C’ is a genuine evaluation. I venture that much resistance to the normativity of personal value thrives on not realizing that the examples concern inverted-commas uses of ‘value-for’, or because it is assumed that ‘O has personal value for C’ is equivalent to ‘O is de facto favored by C’. Still, suppose someone claims that he recognizes that O has personal value for C, but denies that this gives him any reason to favor O for C’s sake. Pointing out that the reasons here might be ‘pro-tanto reasons’ will hopefully clear away more of the misunderstanding. But what should we say to someone who continued denying that there was any reason at all to favor O for C’s sake? We could try reminding him that ‘favoring’ is a technical term that covers a wide spectrum of attitudes. In other words, there is a variety of ‘favorings’ that might be relevant here, such as to be glad for C’s sake or to respect the medal for C’s sake (and not, say, spit on it). Many more examples could be given. Of course, some persons might insist that nothing whatsoever follows regarding our attitudinal stance to O, from O being of

2 Buck-passing Personal Values

49

personal value to C.12 But now we have arrived at a point where at least I find it hard to understand what ‘value’ in ‘value-for’ stands for in this person’s view. Acknowledgements This work has very much developed as a result of a series of conversations over the years with Wlodek Rabinowicz. David Alm, Dan Egonsson, and Michael Zimmerman’s comments have helped me a great deal. A version of this paper was presented at a workshop at SCASSS, Uppsala, and I wish to thank the workshop members, in particular David Bengtsson, John Broome, Johan Br¨annmark, Krister Bykvist, Erik Carlsson, Sven Danielsson, Margaret Gilbert, Jonas Olson, Howard Sobel, and Daniel Svensson. Thanks also to the audience members who attended my talk at Stevens Point, especially Michael Brady and Michael Stocker, and to David Chan for organizing such a splendid conference. My work on this paper was supported by a grant from The Bank of Sweden Tercentenary Foundation. A longer version of it (2007) appeared in Journal of Ethics, volume 11:4 as ‘Analysing Personal Values’.

Notes 1 See for instance, Elisabeth Anderson (1993), Christine Swanton (1995), Wlodek Rabinowicz & Toni Rønnow-Rasmussen (2000, 2003a). For monist approaches, see Noah M. Lemos (1994), Michael J. Zimmerman (2001a, 2001b), Jonas Olson (2003). 2 That something is valuable in its own right might lead one’s thoughts to the idea that the object has value in virtue of some ability or capacity of the object that has been realized. In that case, this kind of value would, it seems, exclude certain objects from the sphere of value bearers. (Would a state of affairs have value in its own right in this sense?) ‘Own right’ also seems to suggest that the value accrues to an object without the ‘help’ of a subject. This in its turn suggests that the expression lends itself more easily to an objectivist analysis of value. 3 One difficulty is the so-called Wrong Kind of Reasons problem, to wit, the problem consisting in distinguishing those reasons that bear on evaluation from those that do not. We may well have reasons to have, say, pro-attitudes that are satisfactory but nonetheless wrong from the point of view of the FA analysis. If a bully with low self-esteem tells me to appreciate him for his own sake under a threat of severe punishment, I have, it seems, a perfectly understandable reason to appreciate him. Given the threat I ought to appreciate him. Notwithstanding, it doesn’t seem right to conclude that the bully is valuable. The analytical problem of separating those reasons that are of the right kind from those that are not ought to put a damper on the enthusiasm of the FA advocates. Cf., D’Arms & Jacobsen (2000a; see also 2000b and 2003) and their ‘conflation problem’. For some recent attempts to solve the problem, see among others Ingmar Persson (2007), John Skorupski (2007), Philip Stratton-Lake (2005), Jonas Olson (2004), and J. Br¨annmark (2003). For a reply to Olsson’s suggested solution, see Rabinowicz and Rønnow-Rasmussen (2006). See also Sven Danielson and Jonas Olson (2007). 4 Kagan (1998) does not refer to these values as final ones, but that is a mere terminological issue. Beardsley sets them aside as peculiar values. For other examples of extrinsic final values, see Rabinowicz and Rønnow-Rasmussen (2000). 5 I do so with the attitude of love in Rønnow-Rasmussen (2007a). 6 A caveat: This account is much too simplified. I am not confident that it is true about all kinds of favoring. But more importantly, I don’t deny that favoring might be individuated in some other way than by a judgment. For instance, if I can feel afraid without being afraid of something, then surely this ‘judgment-condition’ is not valid. Moreover, I can even accept the possibility that being afraid of O does not have to be understood in terms of an ‘evaluative judgment’. However, the account is valid for those ‘favorings’ that are justifiable by the object O. 7 For an illuminating work on different kinds of circles, see Humberstone (1997). See also Rabinowicz and Rønnow-Rasmussen (2006).

50

T. Rønnow-Rasmussen

8 For a discussion of some related problems concerning the notion of ‘attitude’ in relation to fitting-analyses of value, see Rabinowicz & Rønnow-Rasmussen (2004). 9 It might be argued that we are not in fact at all respecting a for a’s sake. We are rather respecting him for his dignity’s sake (see for instance Darwall 2002). It is quite inoffensive to say that we respect a for his dignity’s sake, as long as this is not taken to imply that the proper object of our attitude is dignity. Surely, what we respect is persons and not properties. See p. 399 in Rabinowicz & Rønnow-Rasmussen (2003a). 10 Here is another example. A person I once knew well lived a very self-destructive life. We lost contact, and when I recently learned she had passed away very much due to living a tragic life, I immediately felt sadness. However, I also realized that my grief had a different character than what I experienced when, for instance, my father died. In the latter case, I was sad for my own sake for having lost someone I cared a lot about. In the former case, I cannot sincerely say that I cared much for this person. My sorrow was rather of the following kind: I was sad for her sake. Her disastrous life called forth a sadness, which I experienced for her sake. 11 Differentiating between desiring that x be the case, and being disposed to make x the case, is merely thought to reflect the idea that some of our attitudes are dormant and others active. 12 An alternative approach to this issue would be to ask the persisting person the following q uestion: Wouldn’t you at least accept that if you want to do everything that can be done for x’s sake, you ought to favor O? I owe this informal suggestion (personal communication) to Margaret Gilbert. Cf. also Stephen Darwall’s (2002, p. 4) recent suggestion that ‘a person’s good is constituted [. . .] by what one (perhaps she) should want insofar as one cares about her.’ (For some objections to Darwall’s account, see here Rønnow-Rasmussen 2007b). Both suggestions are worth considering. However, whereas these suggestions regard (personal) value as normatively conditional on a person’s attitudes (wants and care for . . .), my suggestion sets out from the idea that what we have reasons to do is not necessarily a matter of what attitudes we have.

References Anderson E (1993) Value in ethics and economics. Harvard University Press, Cambridge MA Beardsley MC (1965) Intrinsic value. Philosophy and Phenomenological Research 26:1–17 Br¨annmark J (2003) Passing the buck: on reasons and values. In: Rabinowicz and RønnowRasmussen (2003b) Brentano F (1969 [1889])The origin of our knowledge of right and wrong. Routledge & Kegan Paul, London D’Arms J, Jacobson D (2000a) The moralistic fallacy: on the ‘appropriate-ness’ of emotions. Philosophy and Phenomenological Research 61:65–90 D’Arms J, Jacobson D (2000b) Sentiment and value. Ethics 110:722–748 D’Arms J, Jacobson D (2003) The significance of recalcitrant emotion (or, antiquasijudgmentalism). In: Hatzimoysis A (ed) Philosophy and the emotions. Cambridge University Press, Cambridge Danielsson S, Olson, J (2007) Brentano and the buck-passers. Mind 116:511–522. Darwall S (2002) Welfare and rational care. Princeton University Press, Princeton Humberstone IL (1997) Two types of circles. Philosophy and Phenomenological Research 57:249–280 Kagan S (1998) Rethinking intrinsic value. The Journal of Ethics 2:277–297 Lemos NM (1994) Intrinsic value, concept and warrant. Cambridge University Press, Cambridge Olson J (2003) Revisiting the tropic of value: reply to Rabinowicz and Rønnow-Rasmussen. Philosophy and Phenomenological Research 67:412–422 Olson J (2004) Buck-passing and the wrong kind of reasons. The Philosophical Quarterly 54:295–300 Parfit D (1984) Reasons and persons. Clarendon Press, Oxford

2 Buck-passing Personal Values

51

Persson I (2007) Primary and secondary reasons. In: Rønnow-Rasmussen et al. (2007) Rabinowicz W, Rønnow-Rasmussen T (2000) A distinction in value: intrinsic and for its own sake. Proceedings of the Aristotelian Society 100(1):33–49. Also In: Rønnow-Rasmussen and Zimmerman (2005) Rabinowicz W, Rønnow-Rasmussen T (2003a) Tropic of value. Philosophy and Phenomenological Research 66:389–403. Also In: Rønnow-Rasmussen and Zimmerman (2005) Rabinowicz W, Rønnow-Rasmussen T (eds) (2003b) Pattern of values: essays on formal axiology and value analysis. Lund Philosophy Reports, Lund Rabinowicz W, Rønnow-Rasmussen T (2004) The strike of the demon: on fitting pro attitudes and value. Ethics 114:391–423 Rabinowicz W, Rønnow-Rasmussen T (2006) Buck-passing and the right kind of reasons. Philosophical Quarterly 56:14–120 Rønnow-Rasmussen T (2007a) Love, value and supervenience. In: Rønnow-Rasmussen et al. (2007) Rønnow-Rasmussen T (2007b) Analysing personal values. Journal of Ethics 11:405–435 Rønnow-Rasmussen T, Zimmerman MJ (eds) (2005) Recent work on intrinsic value. Springer, Berlin Heidelberg New York Rønnow-Rasmussen T, Petersson B, Josefsson J, Egonsson D (eds) (2007) Hommage a` Wlodek: philosophical papers dedicated to Wlodek Rabinowicz. (http://www.fil.lu.se/ hommageawlodek) Ross WD (1939) Foundations of ethics. Clarendon Press, Oxford Scanlon TM (1998) What we owe to each other. Harvard University Press, Cambridge MA Skorupski J (2007) Buck-passing about goodness. In: Rønnow-Rasmussen et al. (2007) Stratton-Lake P (2005) How to deal with evil demons: comment on Rabinowicz and RønnowRasmussen. Ethics 115:788–798 Swanton C (1995) Profiles of the virtues. Pacific Philosophical Quarterly 27:335–344 Zimmerman MJ (2001a) The nature of intrinsic value. Rowman & Littlefield, Lanham Zimmerman MJ (2001b) Intrinsic value and individual worth. In: Egonsson D, Josefsson J, Petersson B, Rønnow-Rasmussen T (eds) Exploring practical philosophy: from action to values. Ashgate, Aldershot

Part II

Reason and Choice

Chapter 3

Volitions, Comparative Value Judgments, and Choice Joseph Boyle

Abstract This paper seeks to clarify some of the conceptual relationships between value judgments, volitions and actions. Following suggestions articulated by Davidson, I develop an account of motivation such that if an agent judges an action good, the agent will want to perform that action, and the presence of such motivation will lead to action unless prevented. Choice is called for when one must select among actions whose motivations are prevented from being effective by the presence of contrary motivations. When such selections are responsive to reflection on the merits of the actions, they are choices. The possible outcomes of this reflection are three: that the desirabilities of the actions compared are equal, or that one is better, or that they are incommensurable. Choices responsive to these different outcomes are importantly different. Choices among incommensurably desirable options are free choices. This account does not allow weak-willed choices as defined by Davidson, but it allows that a person can correctly judge an option contrary to his or her good and still choose it. Keywords Free choice · Incommensurability · Volition · Weakness of will

The goal of this paper is to uncover some conceptual relationships between choice and value. In particular, I will show that a person’s selection of one of a set of options that is judged incommensurable in value is a distinctive form of free choice. The selection of an option shown by deliberation to be commensurable in value to competing options is free in a different and weaker sense. These claims about choice and value presuppose that choice is a kind of volition, namely, the kind of volition called for to overcome conflicts of motivation that block action. This account of choice, in turn, presupposes an account of volition, which I understand to be a form of motivation, namely, motivation responsive to propositional knowledge. J. Boyle University of St. Michael’s College, 81 St. Mary Street, Toronto, Ontario M5S 1J4, Canada, Tel.: (416) 926-2260, Fax: (416) 978-8703 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

55

56

J. Boyle

I understand motivation generally to be the name of any psychic reality required to mediate cognition and behavior in the cognitively based behavior of animals and humans.

3.1 Value Judgments, Willing, and Doing Donald Davidson formulated two principles to capture common intuitions about the connection between willing to do something and doing it; and between judging something good to do and willing to do it (see p. 23 in Davidson 2001). Davidson formulated these principles comparatively – about what an agent wanted more or judged better – because of his concern with weakness of will. He stated them as follows: P1. If an agent wants to do x more than he wants to do y, and he believes himself free to do either x or y, then he will intentionally do x if he does either x or y intentionally. P2. If an agent judges that it would be better to do x than to do y, then he wants to do x more than he wants to do y.

These principles plainly have non-comparative analogues, which should also be true if they were true. The non-comparative formulations are less complex and focus on simple cases of willing, so likely raise fewer questions than Davidson’s comparative principles. I propose to begin with the non-comparative analogues, since, before considering choice, which I will argue is a special kind of volition, it is useful to focus on the more generic idea of a volition. My formulations are as follows: NP1. If an agent wants to do x, then he will do x, unless prevented from doing x. NP2. If an agent judges that it would be good to do x, then he wants to do x.

In NP1 I specify the conditions introduced by Davidson in P1 by holding that an agent will do what he or she wills unless prevented from doing it. I understand that an agent is prevented from doing something he or she wants to do either if he or she believes that it cannot be done, or wants to do something incompatible with doing it, or if something external to the agent prevents it, or if the agent is incapable of doing it. These cases seem to me to cover what Davidson means to allow as conditions under which willing to do something does not lead to doing it. NP1 also captures, perhaps more directly than P1, the intuitions about the nature of volition which Davidson means to articulate. He cites Anscombe: ‘The primitive sign of wanting is trying to get.’ Hampshire provides a formulation with a closer fit to the intuition Davidson is interested in: ‘A wants to do X’ is equivalent to ‘other things being equal, he would do X if he could.’ Davidson proposes that ‘other things being equal’ allows the comparative interpretation of P1 (see pp. 22–23 in Davidson 2001).1 I agree, and further believe that Hampshire’s qualifier is also captured within my ‘unless prevented’ condition. NP2, like P2, affirms a tight connection between a certain kind of value judgment and willing, that is, between judging that it would be good to do something and wanting to do it. Davidson directly defends P2 only briefly. He notes that it states ‘a

3 Volitions, Comparative Value Judgments, and Choice

57

mild form of internalism,’ which he formulates as follows: ‘It says that a judgement of value must be reflected in wants (or desires or motives)’ (see p. 26 in Davidson 2001). This is equally true of my NP2. Davidson notes that it is tempting for those looking ahead at the apparent inconsistency between the conjunction of P1 and P2, and weakness of will to tinker with them. But this is problematic because ‘P1 and P2 derive their force from a very persuasive view of the nature of intentional action and practical reason’ (see p. 31 in Davidson 2001).

3.2 A Sketch of a Theory of Motivation I believe that much of the persuasiveness of the view of the nature of intentional action and practical reason to which Davidson refers rests on considerations that are not specific to intentional action and practical reason, but rather are common to any behavior, human or animal, that is responsive to cognition. In this section, I will develop this belief by seeking to vindicate NP1 and NP2, showing their place within an account of behavior based on cognition. This more general account of cognitively based behavior provides the larger context for a theory of intentional action. Consider the following common events: A dog feels a bellyache, perceives that negatively, and chews grass that causes vomiting; an animal feels lethargic, and goes to a quiet place; a small child is hungry and cries until fed. The chewing, going and crying are cases of what I will call cognitionally based behavior. An example of such behavior in a mature human might be the person’s going for water in response to being thirsty. The case I have in mind here is not an intentional action of going for a drink, but simply doing that without reflection or decision in response to the perception. Such examples suggest the following definition: a behavior is cognitively based if and only if (1) the animal perceives something – some amalgam of colors, shapes, movements and so on – internal and/or external to itself; (2) the animal perceives the relevance of the object to its welfare; and (3) the animal behaves in accordance with the perceptions in (1) and (2). Another feature of at least some of the common events listed above is the animal’s imaginative projection of possible behavior responsive to the perceptions of (1) and (2). This feature seems characteristic of forms of cognitively based behavior most helpful for understanding human agency. For cognitively based behavior which lacks this imaginative anticipation is one step closer to instinctive behavior than much of the animal behavior in which past experience and imagination seem to allow a range of responses to the perceptions in (1) and (2). In speaking of a perception of data in (1), of a perception of relevance in (2), and of an imaginative anticipation in many instances of cognitively based behavior, I am not referring to reflective, propositional knowledge. The implications of this restriction are different for each of these cognitions: the object perceived is not a state of affairs affirmed in reflective judgment; the instinctive perception of relevance

58

J. Boyle

perceives but does not recognize a relationship, namely, that between the object perceived and the animal’s interests; and the anticipation is imaginative not conceptual. Consequently, the imaginative anticipation in much cognitively based behavior is not means/ends reasoning. Of course, small children, who may be incapable of the deliberation and choice needed for morally significant action, may well be capable of engaging in cognitively based behavior in which means are adopted for the sake of goals that are sought to satisfy desire. That requires reflective separation of the desire to be satisfied and the behavior that may satisfy the desire, and insight into the (usually) causal connections between means and ends. This reasoning, which seems present in the behavior of, say, five year olds, is not evidently present in cognitively based behavior more generally. The strictly cognitive elements which partly comprise cognitively based behavior are not by themselves sufficient to explain it. My reasoning is as follows: the cognitions relevant to behavior perceive data and relevance, and anticipate possible behavior. These cognitions are relevant to the behavior, since the behavior is responsive to them. By definition cognitively based behavior must have an explanatory dependence upon them. This sort of relationship can correctly be called ‘causal,’ but the dependence relationship in question cannot be that of a movement’s dependence upon its moving or efficient cause. For perceiving data, perceiving relevance, and anticipating behavior, just as cognitions, present or represent data and possibilities. More is needed to explain cognitively based behavior, since something is needed in addition to presentation or representation to activate the animal’s behavior in response to cognition. To develop this a bit: the object of human knowing is truth, theoretical or practical. Similarly, the objects of the cognitions on which cognitively based behavior is based are information, relation to interest, and imagined possibilities. Cognitions having these objects are not perhaps strictly true or false, but they are presentations or representations. These play their distinctive explanatory role in cognitively based behavior precisely by presenting or representing data, relevance and possible behaviors. Something in addition to that explanatory role is needed to explain cognitively based behavior, namely, something that is both responsive to the cognitions and capable of activating behavior. Of course, it is possible to stipulate that the cognitions in question include whatever it takes to activate behavior. But even if justified, such a stipulation would not eliminate the need to distinguish within the psychical component of cognitively based behavior between the strictly cognitive and the causal activation of behavior in response to the cognition. Whether two events, or two moments within a single event, having experiences, beliefs and so on remains distinct from a person’s or an animal’s actively responding to those experiences or beliefs by activating the appropriate behavior. Thus, the cognitive and the causally effective response to it remain irreducible aspects or parts of a larger, complex whole, however tightly they are intertwined. Therefore, some psychic factor (1) responsive to cognition and (2) having the causal power lacking in cognition to activate behavior is needed fully to explain

3 Volitions, Comparative Value Judgments, and Choice

59

cognitively based behavior. I will call this factor ‘motivation’ and call instances of motivation ‘motivations.’ The idea of responsiveness to cognition included in my notion of motivation does not, of course, mean ‘caused by cognition’. Motivation is informed by cognition, and so consequently the presence of the informing cognition is necessary for the identity of any given motivation. Still, the informing of the motivation is not what makes it causally active, as that must come from some other source within the animal or person. Motivations are of several kinds. Motivation towards the beneficial and motivation away from the harmful are respectively attraction and repulsion. Desire readily comes to mind as the paradigm case of attractive motivation. But the motivation to the benefit of resting is quite distinct from standard cases of desire, for example, sexual desire or hunger. Moreover, motivations to struggle, which the medieval philosophers categorized as ‘irascible appetites’ – the flight or fight motivations such as anger – are irreducible to attraction or repulsion. Motivation links the three elements of cognition – data perceived, interests perceived, and imagined responsive behavior – to the animal’s behavioral capacities, that is, at least elementally, to the animal’s ability to move, imagine, think and do anything else that might be a behavior it can do without causing it by doing something else. That link is causal, by which I mean not simply explanatory but efficiently causal, since the activation of the behavior is what the cognition cannot explain by itself. There are bound to be puzzles or mysteries about how a psychic power, such as I have defined motivation, can be both responsive to cognitions and causally active in relation to the behavior it motivates. Thus, I do not pretend to have a full account of either end of the relationship that defines motivation, of either how motivations respond or how they activate. Nevertheless, if there is such a thing as cognitively based behavior, then there is what I have called motivation. And whatever the details of our understanding of how motivations respond to knowledge or of how they activate behavior, connections at least very similar to those affirmed in NP1 and NP2 must obtain if there is to be any such thing. For the responsiveness of motivation to cognition is just what NP2 articulates, at least in respect to volitions, and the behavior-activating power of motivations is what NP1 articulates, at least in respect to volitions. Of course, many people understand motivation differently than this account requires, that is, as essentially an element in the explanation of cognitively based behavior. Accounts of motivation according to which motivations are conceived and identified introspectively will likely resist the ideas that motivations will activate behavior unless prevented, and are essentially responsive to cognitions. These and other accounts likely favor a dispositional understanding of motivations according to which certain cognitions tend to generate certain motivations, which in turn only tend to activate behavior. One reason why some people conceive of motivations as tendencies rather than as active psychic causes is the common experience of the ineffectiveness of motivation in activating motivated behavior. But NP1 includes a condition, the ‘unless

60

J. Boyle

prevented’ condition, and the factors preventing a motivation from being effective frequently obtain. So, the view of motivation expressed in NP1 is consistent with common experience. Moreover, the conception of motivation expressed in NP1 accounts for an aspect of our experience of agency and its analogues which a dispositional conception of motivations leaves mysterious, namely, the fact that motivation does activate behavior. The account of motivation expressed in NP1 plainly accounts for this fact. But dispositions do not of themselves activate behavior. They only tend to do so and some other factor is needed to explain how they do so when they do so. Consequently, if a disposition is what responds psychically to cognition, the practical effect of the cognitions remains obscure. In respect to NP2, the connection between judgment and volition may seem quite contingent because there are value judgments of several kinds that do not issue in motivations, including some in which, given prior motivation, there is motivational effect. For example, what is useful to reach a goal is not motivating unless one values the goal, but such instrumentalities may sometimes reasonably be called ‘good’ even by those who do not value the goal to which they contribute. Similarly, a striking or especially perfect instance of a kind may be evaluated highly, without that judgment implying anything practical. Again, a paradigm case of excellence in some area may be judged good even by those who have no interest in that kind of excellence. However, the abundance of kinds of evaluative judgments does not undercut NP2. For that principle formulates the connection between a specific form of evaluation and motivation. The cognitions in play in cognitively based behavior are first-personal and practical. In them, data perceived are also perceived as relevant to the animal’s welfare, and the animal imagines something that can be done in response to those bits of information. However unreflective this process may be, the practicality of the cognitions is clear, and they cannot be practical without motivations responding to them. This consideration plainly applies to the more reflective cognition and consequent volition referred to in NP2. For the judgments of value here are obviously judgments of desirability, and these could not be practical without the ‘mild form of internalism’ my account of cognitively based behavior implies. As NP1 and NP2 suggest, human action involves a different sort of response to cognition than we have been discussing, or, perhaps more precisely, a response to different forms of cognition. The motivations characteristic of human action are those responsive to specifically human cognition, namely, to knowledge by way of propositions, including their combination in reasoning. A motivation responsive to propositional knowledge is a volition. Willing was helpfully defined by medieval philosophers as ‘rational appetite’. In the language that I am using, a volition is any motivation responsive to knowledge that is propositional. This old definition of willing assumes that intentional action is either analogous to or a species of cognitively based behavior. Volition is what causally links the cognitions on which human action is based to the person’s bodily and mental capacities to act. As in the case of the strictly sensory cognitions described above, those leading to human action are not of themselves sufficient to explain the person’s action. They

3 Volitions, Comparative Value Judgments, and Choice

61

provide part of its explanation, in particular the agent’s reasons for doing it, but just as cognitions aiming at the truth, they lack moving causal power. Knowledge through propositions is distinct from sensory experience, although information provided by sensory experience is central to what people know propositionally. The distinctiveness of propositional knowledge has several aspects relevant here: (a) propositions are distinguishable from the states of affairs we know through them, and that distinction allows the reflective determination of truth or falsity; (b) propositions can be combined by reasoning so as to represent causal and other relationships between states of affairs; and (c) there is a functional differentiation of elements within a proposition. Formulable in many ways, the simple idea here is that a truth bearer must have both a referring subject and a predicate through which the knower does more than simply refer to an individual thing but renders intelligible what is referred to. The knower predicates of the object referred to an intelligible (and so general) feature formulating a complex idea that can correspond to a state of affairs. These characteristics of propositional knowledge significantly affect the nature of the specific cognitions on which intentional actions are based. The object perceived in cognitionally based behavior here becomes a conceptualized state of affairs; the object of the interest whose perception makes the perceived information relevant becomes the good-making or bad-making features of a propositionally formulated assessment of the information received; and the anticipated response is conceptualized, and reflectively and rationally related to the prior cognitions as providing a means to the end they suggest. Reasoning and the insight it provides thus become constitutive of the cognitive basis of intentional action. For example, I feel pains in my hip; I recognize that they are not transient but a threat to my mobility; I see that for the sake of preserving this aspect of my health, it is worth undertaking a regimen of prescribed exercise that includes avoiding certain postures and movements. The motivations responsive to these rational cognitions are distinguished in accord with the complexity required by these cognitions. The response to the evaluation indicating the existence of some good-making property, such as being healthy, is a motivation to realize or safeguard instances of that property in a person’s life. It is a response to the recognition that goals instantiating it are worth promoting. This volition might be called having an interest in that good. The recognition that instances of that property are threatened or can be fostered provides a goal or end which an agent can intend if there is anything he or she can do to realize it, and the performance understood as capable of fostering that goal is a means the person adopts. Minimally, therefore, the willing to do something that is an essential component of a voluntary human action – the wanting in NP1 and NP2 – must include an interest in a good, an intention of specific goals instantiating the good, and the rational adoption of means to those goals. These three interconnected motivations, ordinarily experienced as aspects of a single motive, are necessary for the production of what we regard as an intentional action, and only through them can the cognitions on which intentional actions are based have effect in action.

62

J. Boyle

Having an interest in a good, intending a goal, and adopting a means are, of course, not the only kinds of volition, nor are they the only elements in many intentional actions. These are the volitions necessary for a human action. Two other kinds of volition are worth noting. One of these is a form of wishing, that is, the conditional readiness of a person to do an action which was prevented. One whose intention is thwarted by some incapacity or some unchangeable external circumstance may still wish for the goal or to do what will achieve it. Similarly, one can wish to do an action prevented by the agent’s conflicting motivations. Wishes in response to such blocks to action can be practically important at some point. For example, the immoralist who wills ‘I would steal that, but I’d get caught,’ may well be ready to steal in different circumstances; and the continent but not yet virtuous person who wills ‘if only that were not immoral, I would do it,’ may court this immorality in the future. These acts of willing, though not immediately effective because of the preventative factors, can activate future action when those factors are not functioning to prevent the action. Another important kind of willing that does not necessarily exist in every intentional action is choice, which I believe is distinct from adopting a means to an end. I will consider choice in detail in the remainder of this paper.

3.3 Resolving Conflicts of Desires by Deliberation and Choice As already noted, the success of motivation in activating behavior can be prevented in several ways. Some of the factors preventing the effective outcome of motivation are external to motivation and independent of the knowledge to which it responds: behavioral capacities sometimes break down, and conditions in the world outside the animal block its behaving. Cognitive and motivational factors can also prevent a motivation from being effective. In particular, the behavioral effect of any motivation can be blocked by the presence of other motivations. This occurs whenever there are motivations towards activating incompatible behaviors at the same time, where the strength of each of the motives is sufficient to block the efficacy of the other in activating behavior. Motivational blocks of this kind can be resolved or overcome in several ways. Some resolutions of motivational blocks work independently of the animal’s cognitive and motivational capacities, as when external force is applied or when the physiological systems break down. For example, hesitation between fight or flight is sometimes settled when an enemy destroys an animal’s capacity to inflict harm, or, in the opposite case, destroys the capacity to run away. Perhaps there also exist within the nervous system some non-psychic overrides of motivations that cause behavior independently of cognition and motivation in at least some such situations of hesitancy to act. But sometimes the block to behavior can be overcome within the working of the animal’s psychic capacities for cognitively based behavior. Sometimes that occurs when the relative strength of the motivation for one behavior changes, perhaps

3 Volitions, Comparative Value Judgments, and Choice

63

because of new information or a different perception of its relevance, in such a way that it or the other or some new motivation prevails and activates behavior. In the case of human motivation, we are sometimes aware that action or behavior was blocked by a conflict of desires and that, without any further motivation or decision on our part, the issue is settled and one desire has prevailed. Perhaps this form of emerging dominance by one of the conflicting motivations is always sufficient to explain how motivational blocks are overcome in animals and small children. But in mature humans, and possibly also in some animals and children, the dominance of one of the conflicting motivations is sometimes established by a further initiative of the agent to overcome the conflict. This initiative is the agent’s picking or selecting one of the behaviors. An adult person’s reflective choices are perhaps the clearest examples of this sort of initiative, but other less reflective or deliberate selections are part of human life, and may also be part of the lives of children and animals. As in the cases where the dominance of one motivation emerges from the factors informing the set of motivations, the cases in which the dominance is established by the agent’s selection are also informed by the factors revealed in the set of cognitions informing the conflicting motivations and in other relevant cognitions. Otherwise the selecting would not be a component in cognitionally based behavior. The difference between the two cases is that selecting is also informed by cognitions taking into account the conflict of motivations, and consequently, selecting motivates precisely by dealing directly with the conflict. Selection is a form of motivation because selecting links cognition to behavior. By selecting one of the motivated actions, the person or animal allows the activating of the behavior by the motivations towards it. So selecting is a motivation that presupposes prior motivations and functions by allowing one of them to be effective in activating behavior. These now dominant motivations are responsive to the cognitions informing them. The cognitions informing the motivations to the behavior not selected cease to be practical as far as that unique behavior is concerned. In short, selection is a functionally distinctive motivation that responds to a cognition distinct from those informing the motivations whose conflict blocks action. That cognition is reflective in at least the minimal sense that it addresses the blocking of action by conflicting motivations. It also seems to require some form of comparison of the behaviors whose motivations are in conflict. That is easiest to understand as a rational comparison, but it may be that, short of a rational comparison, there is a further imaginative consideration of the possible behaviors and their relevance to which the new motivation of selecting responds. Thus, perhaps one of the behaviors is seen to fit better with an imagined future, and that cognition informs the selection that allows the motivations for this behavior to activate it. In short, motivations sometimes clash, and do so in a way that blocks the behavior motivated by each of the conflicting motivations. Sometimes those blocks to behavior are overcome and the animal gets on with one of the motivated behaviors. Sometimes such blocks are overcome within the animal’s motivation. In some cases one desire gets stronger. But sometimes a human being or other animal itself picks

64

J. Boyle

one behavior over the other, by a form of motivation not reducible to those in conflict and responsive to interests and information not reducible to those initially generating the block to behavior. Whether or not this idea that an agent can actively overcome conflicting motivations by its own initiative has application to animals, it is central to the idea of human choice. People face conflicts of motivation sufficient to prevent their doing what is motivated. Sometimes those conflicts of motivation are overcome when one desire dominates due to changed information. But sometimes a person settles the matter by actively picking one action rather than another, and sometimes this selection is based upon rational reflection – deliberation – on the motivated actions. The selection of a possible action on the basis of deliberation about its merits is choice. Thus, not every selection or picking of one thing over another by a human agent is a choice. Only those selections that are responsive to deliberation are choices. For example, picking an item from a menu could be, but is not necessarily, a choice, since that could be done on a whim without deliberation. Choice, therefore, is a volition since it motivates action by removing a factor preventing it, precisely in response to the rational considerations revealed in deliberation about the options. This conception of choice differs in some ways from that in the Aristotelian analysis to which my overall account is indebted. It is true that choice is always of a means – that is, of something we can do for the sake of causing a valued goal or instantiating a good-making property. The activating of the performance constituting the means is precisely the point of choosing. But choice, as I am defining it, is not simply adopting a means to an end, which is essential in any human action and need involve no conflicting motivations. Practical reflection upon one or several means to a valued end may give rise to conflicting motivations because this reflection can indicate the difficulty of the means, or because of other anticipated effects adopting a means may have in relation to our other interests. But choice is not adequately defined as adopting a means to an end, that is, as picking a performance one believes will contribute to a valued goal. For on this conception of choice, the deliberation often used to resolve conflicts of motivation, and consequently the specific forms of selection responsive to that deliberation, are left unexplained or forced into a distorting means/ends notion of practical thinking. Moreover, a person overcomes a motivational block in choosing by his or her practical endorsement of one of a set of competing options. That endorsement is plainly not limited to the performance constituting the chosen means: the valued goal to which it contributes is also endorsed as is the underlying good in which the person is interested. So choice is, just as importantly, of the goods and goals as it is of the means. Indeed, it seems that the endorsement of the claims of a good or goods as instantiated in one option is the specific contribution of choice to the motivational mix. That endorsement of the good is what co-opts or silences or renders ineffective the claims of the goods involved in other options, and that allows the formerly blocked motivation to activate the behavior.

3 Volitions, Comparative Value Judgments, and Choice

65

3.4 Responding to the Results of Deliberation Deliberation certainly involves comparisons of several kinds among options for action and the person, in making such comparisons, will observe or try to discover various kinds of commensurability. For example, options can be morally comparable when measured by the same set of moral principles. Similarly, a definite goal to which a person is committed can provide a measure for ordering that person’s options. Even something more complex, such as a person’s established preferences or commitments, could provide a measure for assessing new options. The deliberative undertaking of making comparisons among options includes trying to uncover these and other kinds of commensurability, but the interest of a deliberating person in discovering common measures to assess options does not guarantee that they exist. The result of the effort may be the judgment that the options are not commensurable in the respect considered: for example, one’s settled commitments might provide no measure that is applicable to some new project. One of the respects in which people are necessarily concerned to discover bases for commensurability among options is in respect to their overall desirability or intelligible appeal – in respect, that is, to what would gives us reason to do any one of them. Since options comprise what we have reason to do, commensurating them in respect to just that, namely, the desirability that gives us reason to do them, is plainly relevant to deliberation, since knowing this would allow us to determine which option is most desirable – what there is the most reason to do. If deliberation does uncover a principle of commensurability of this kind, it is possible that a person deliberating about two options will judge that one of them is more desirable than the other. Alternatively, it is possible that the person will judge them equal in desirability. If a person deliberating judges that the options are neither equal in desirability nor better nor worse than others in desirability, the person judges the options incommensurable in desirability. In this case, they will be judged desirable in ways lacking a common measure.2 These three forms of comparative value judgment seem to me to capture the essential possibilities. Other seemingly distinct value judgments are reducible to these. Cases in which the deliberating person is unable to identify a principle of commensuration, but is unsure whether further reflection would turn up such a principle, are reducible to those in which the agent judges the options incommensurable if choice must be made in the face of that uncertainty. The person chooses and, having no basis for commensurating the options, chooses among options for him or her incommensurable in desirability. In similar cases in which deliberation continues so as to resolve the uncertainty about the presence of a principle for commensurating the desirability of the options, the further reflection may turn up such a principle but it may not. So the prospect that one may find oneself facing options which one cannot commensurate is irreducible. In short, a person’s failure to discover a principle of commensuration in deliberating, while continuing to hope that such a principle might turn up by continued reflection, does not introduce a further fundamental category of comparative value judgment.

66

J. Boyle

Moreover, a more general consideration of the sources of uncertainty and mistakes in deliberation and in the judgments which complete it confirms that the threefold division of comparative value judgments is comprehensive. These uncertainties do not arise because of the unavailability to the deliberating person of the grounds for the desirability of the various options he or she wants to pursue. Although these can be difficult to articulate, they are present as the goods the prospect of whose instantiation gives point to the action. If these goods are such that they share some common feature that makes them both desirable, then that should also be available to the deliberating person: for the desirability making feature the goods share is what the person is fundamentally after in acting for either option. Thus, it is in principle possible to be confident that the good making features which render a goal desirable are or are not similar enough to those of others to allow for ranking by a common measure. Some uncertainties arise in deliberation because the adopted means to the goals for which people act are at best only likely to contribute to those goals, which in turn are only more or less likely to instantiate the good making features for which they are valued. This uncertainty does not affect the person’s awareness of the goods which make the options for action eligible, nor does it affect the person’s ability to understand whether there is some underlying, common principle of desirability. Of course, comparing an action likely to bring about a valued goal with one less likely to bring about a distinct valued goal might introduce incommensurabilities even if the goods underlying the distinct goals were the same. Here again, the uncertainties do not suggest the need for an amplified scheme of comparative value judgments. Another form of uncertainty arises from a consideration of the side-effects of actions, which is surely an essential part of deliberation. Although the uncertain risks of bad side-effects and probabilities of bonus outcomes complicate many deliberations, they do not fundamentally alter a person’s access to the good or bad making features which give side-effects their deliberative significance. Since sideeffects can have impact on anything a person values, the full array of a person’s conception of the good can be brought into play. The introduction of this full sense of what a person thinks good, together with issues raised by probabilities of the outcomes, suggests that judgments of incommensurability may be common. Still, that judgment of incommensurability is but one of the possible outcomes of deliberation, and, in spite of the complexities, a person might reasonably come to think that options are equal in value or that one is more or less valuable than the other. Finally, uncertainty is introduced into deliberation by the vagueness of value predicates, which leads to comparative value judgments that are rough at best. Thus, we sometimes think that options are roughly equal in value, or that the differences in what we value about them are so small that we are reasonable in treating them as equal. Similarly, we sometimes face options that seem similar in value but lack the resources to articulate the basis for a ranking. So we are stuck with treating them as incommensurable. As these examples suggest, the uncertainties raised by vagueness do not require further categories of comparative value judgment. Since there are three types of comparative value judgment that can terminate the effort of a person deliberating to ascertain which of a set of conflicting option

3 Volitions, Comparative Value Judgments, and Choice

67

is most desirable, there will be three distinct forms of choice, namely, motivations responsive to judgments that the options are equal in desirability, those responsive to judgments that one option is more desirable than the others, and those responsive to judgments that the options are incommensurable in desirability. Let us consider these three responses to the differing value judgments. I propose to do this without extensive examples, since it is the general character of the different volitional responses to judgments of these general forms that I wish to explore, rather then whether or how frequently the judgments are correctly made. The first form of choice is that responsive to the judgment that the options for action among which one must select to overcome the block to action are equal in value. An example might be the hesitation one might feel in opting for one of several possible and complex investment opportunities, and the overcoming of this hesitation by flipping a coin after an investigation showing that the risks and possible benefits of both prospects were identical, i.e., that they were equally desirable. If the selection in such cases is to be a volition responsive to the judgment of equal desirability, it will be responsive to the fact that in respect to rational desirability the options are identical. What distinguishes them and blocks action is revealed in deliberation to be something unconnected to their being worth doing. The choice of one option over the others, therefore, is not on the basis of relevant intelligible differences among them as, in just that respect, the choice is non-rational. But non-rational selection among options that are intelligibly identical in terms of what makes them desirable is appropriately responsive to the judgment made. That is appropriate because it removes the block to action. Consequently, if choice is selection responsive to deliberation, selections of this kind are properly understood as choices even though they do not select in response to something intelligibly distinctive in the option selected. Removing the block to action is rational since any of the options is desirable and, since the diversity of motivation that prevents action is revealed in deliberation to lack relevant intelligible grounds, picking one of them on a non-rational basis is the rational thing to do. Since there is no intelligible basis for preferring one to another of the motivated options, the judgment of equality of value effectively silences the appeal of the option rejected. Hankering for a rejected option equally good to the one chosen may be possible emotionally but there is no benefit promised by it that is not just as well received from the option chosen. Choices of this kind are free in the sense that any motivation is free. Their freedom is simply the fact that it is the person’s own uncoerced judgment of equality of value that determines the person to get on with picking one of the options identical in rational appeal. That judgment does not settle which option to select and that aspect of the selection must be made on the basis of some non-rational factor. But that too, though not responsive to intelligible features distinctive of any option, can be free insofar as external factors do not coerce the person or prevent the selection from being appropriately, although non-rationally, made. But choices of this kind are not free in the sense that a choice would be free if it lacked sufficient causal conditions outside the agent’s choosing itself. The non-rational selection, which is the choice as a response to the judgment of equality of value, can be sufficiently

68

J. Boyle

determined by anything and serve its purpose; flipping a coin to settle such a matter can be determined by the fact that the person can think of only this way to settle it. Choosing to flip a coin is, of course, another action which may be a free choice. The second form of choice responds to the judgment that one of the options is simply better than the others in terms of what makes them rationally desirable. In these cases, if the selection is a volition responsive to the judgment of desirability, it will be responsive to the fact that one option is rationally more desirable. Selecting the better option is responsive to this judgment and selecting any other option would not be a response to that judgment, as Davidson’s P1 makes clear. This does not mean that other ways of overcoming the motivational block that called for the deliberation are impossible, but rather that they would not be volitions responsive to the judgment of the greater good. For example, the desire for the less desirable option might become so strong that deliberation and choice were literally overwhelmed. Similarly, a person could become distracted from attending to the judgment of the greater good, which might happen if the person became so focused on the attractions of one option that the consideration of the other faded and the person no longer faced a motivational conflict. However, a choice to focus exclusively on one of the options is plainly a different matter since choices presuppose conflicting motivations, in this case motives for and against taking such a focus. But if the comparative judgment is made, all the reasons favor doing what is more desirable and no reason emerges for considering anything not captured in that judgment. It is certainly possible to choose to terminate deliberation prior to coming to a comparative value judgment about the motivations in conflict. Many choices can be put off in this way or settled by default. But choices about deliberation itself are responsive to comparative value judgments of the kinds now under discussion. In cases where one judges an option simply better, therefore, rational selection is of the better option. The connection is necessary since a selection of a worse option could not be responsive to deliberation and so would not be a choice. For the options have been commensurated by a common standard of rational desirability and one of them determined to be better. Consequently, any reason for choosing an option judged less valuable would have been considered in the overall comparison of desirability, which resulted in the judgment that the features making the less valuable option desirable were more fully instantiated in the better option. It follows that, as in choices responsive to judgments of equal good, choices of the greater good effectively silence the motivations for the lesser good. For the judgment of greater good is the judgment that whatever there is of desirability in the lesser good will be found and literally fully compensated in the greater good. Any hankering for the option judged less good is strictly emotional and without any rational ground. The mechanism for breaking the block to action caused by conflicting desires is clear in these cases: the judgment that one option is simply better in rational appeal undercuts the motivational basis for the other options because that judgment reveals that the better option promises more of what interests the person than the worse option. Therefore, it appears that the judgment of greater desirability itself removes

3 Volitions, Comparative Value Judgments, and Choice

69

the motivational block by subverting the motivation towards the lesser good without a further motivation. The willing needed to connect that judgment to action is simply the willing of the option that was blocked by the existence of other motivations, now revealed as promising nothing intelligibly desirable not contained in the more desirable option. So it may be stretching things to speak strictly here of selecting or choosing the better option. Still, action was blocked and that block was overcome by the agent’s own deliberation, and the motivation that had been blocked was allowed to become effective. Dynamically, that move to action can be construed as a selective process and surely one that is responsive to judgment. Effective volitions responsive to judgments of this kind are free in the way all volitions as such are free: a volition is the person’s own activity; it is not coerced by external factors; and it is properly determined by the reasoning that informs it. But plainly, this volition is not free in the stronger sense in which a choice would be free just in case it did not have sufficient causes outside the agent’s own choosing. I turn now to the third of the comparative judgments of value distinguished above, namely, the judgment that the options among which a person must select are incommensurable in terms of their rational desirability. In such cases, the selection among them is responsive to that judgment if it allows either of the conflicting motivations to issue in action. Each option contains a basis for rational desirability. Endorsing one of them is thus a reasonable response to the judgment of incommensurable value. Indeed, it is the only volitional response possible in the situation. Of course, there may be other kinds of rational consideration relevant to such choices, for example, the immorality of some options, but these cannot alter the fact that one has reason for the morally good choices that are incommensurable in rational appeal with the reasons supporting the immoral options. The endorsing of one option in response to the judgment of incommensurable value is usefully contrasted with the selection of one of equally valuable options. In the latter, the options are intelligibly identical in those respects relevant to their selection, which must therefore be based on a non-rational consideration. But in the former, the intelligible grounds for desiring the various options are irreducibly different. In choosing one over others, a person responds to and endorses its rational desirability by choosing to make those reasons the operative reasons in the action that had been blocked. The judgment that options are incommensurable in the grounds of their desirability does not silence or co-opt the motivations for the options not selected in the way judgments of the greater good or of one of equally good options silence their motivations. In these judgments, the appeal of the less good option or of the equally good option to be set aside is undercut because the other option is seen as also possessing that appeal or more of it. But the endorsement that is the choice of one of a set of incommensurably good options does not respond to any such rational undercutting of the motives for the unchosen option. The appeal of the unchosen option remains, even though the choice establishes another option as the action to be performed with its reasons now established as the person’s own. Unless the cognitions upon which the motivations for the unchosen option are rejected as false or

70

J. Boyle

incorrect, the motivation towards the unchosen action will remain as wishing for it. This suggests that in morally significant choices among incommensurable options, the endorsement of a morally good set of reasons will remain unstable unless the grounds for the motivations for immoral alternatives are addressed and rejected. The endorsing of one option is a volition that responds to the value of that option but not to the value in others, after considering precisely the comparison of their values. Given the character of the comparative value judgment to which it responds, that volition could not be determined by the motivations in conflict since they are revealed to motivate towards incommensurably valuable options. That comparative value judgment would equally inform the endorsement of the other of the incommensurable options. Thus, the choice is free in a way the other kinds of choice are not: it is a choice without sufficient causal conditions prior to the agent’s choice itself, that is, it is incompatibilist free will. If a person faces incommensurably valuable options and if there is a volitional response to the person’s judgment that the options are incommensurably valuable, then there will be free choices in this strong sense. Of course, this connection I am affirming between free choice and incommensurably valuable options neither proves nor disproves the existence of free choice or of incommensurable options. Independent argumentation is needed for those claims. But this connection does help clarify some of what philosophers have found so mysterious about free choice. Endorsing an option that is incommensurable in desirability in comparison to other options cannot, of course, be completely rationalized in terms of the reasons available. There are too many of them, irreducibly pointing towards different possible actions. But freely endorsing one of a set of incommensurable options is fully responsive to the judgment that the options are desirable but incommensurably so: in that situation, an agent cannot do more than endorse some instantiations of some good or goods as those that will be dominant for that agent. In effect, this carves out from the spectrum of goods and their instantiations those that shall define the agent’s living. This is not the mysterious twitch or accident alleged by some opponents of free choice but a rational and appropriate, though undetermined, response to the judgment that one proposed action is incommensurable in desirability in relation to other actions the agent has reason to do.

3.5 Compatibility of Weakness of Will with the Account of Volition The account of volition developed here surely raises the question of how weakness of will is possible on its terms. My application of the conjunction of NP1 and NP2 to motivational conflicts in which one option is judged better than another seems to imply the impossibility of weakness of will, at least as defined by Davidson. For when one judges that one option is better, the appeal of the worse falls away as a motivation blocking action. Seeing that an option is more desirable appears sufficient to silence the claims of the less desirable option, and that guarantees that the motivation towards the more desirable prevails. So, there is no possible choice and no rational motivation that is contrary to this judgment.

3 Volitions, Comparative Value Judgments, and Choice

71

Contrary to Davidson, I do not deny that implication. Davidson avoids the logical inconsistency between the conjunction of P1 and P2 and weakness of will by treating the comparative value judgment in place when a weak-willed choice is made as conditionalized, that is, as a judgment that is prima facie (or probable) in relation to all the relevant considerations. Since the comparative value judgment that would be inconsistent with a weak-willed choice is not conditionalized but categorical, as P2 indicates, Davidson avoids the contradiction (see p. 39 in Davidson 2001). But this seems to move the puzzle about weakness of will from the domain of logical contradiction into the realm of mystery. For, as Davidson notes (see p. 39 in Davidson 2001), the conditionalized or prima facie comparative judgment of value must be transformed into or replaced by a non-conditional judgment in order to function within human agency as the basis for an effective volition. The weak-willed person presumably has before him the conditionalized value judgment but, by hypothesis, that is not the value judgment to which his choice responds. So this seems to allow that one can judge that on all the relevant considerations a is better than b, and then act on some contrary judgment. Necessarily that is acting without reason. Davidson seems to allow this in the closing paragraph of his paper (see p. 41 in Davidson 2001). Deliberative judgments are unlike predictions in which the predictor knows that things might turn out differently than he or she expected on the basis of all the available evidence. In deliberating and choosing, one does not wait for a prediction to be verified but acts on the basis of the deliberation. All of the evidence that one has about what it is desirable to do supports the judgment that one option is best, yet one is somehow able to believe something contrary and to do this without grounds, which all support the conditionalized judgment and which comprise all that could be relevant to the judgment. This account renders akratic choice too mysterious for it to remain a part of people’s lives over which they have some control, and for which they bear some responsibility. Akratic choice is problematic because it is choice, not non-rational selection or the overwhelming of agency by strong desire or some other form of affliction that subverts human agency. In general terms, my account has the following form: It is surely possible to judge that one option is more desirable than another in several different ways. On my account, practical judgment and volition exclude the possibility of choices against the option judged better in only one of these ways, namely, when the choice responds to the judgment that one option is simply better in terms of whatever counts towards its desirability than its alternative does. But other comparative judgments of desirability can approximate these full judgments of comparative desirability while falling short of their full assessment of good-making features and their instantiations. These approximate but incomplete judgments of comparative desirability can be true, and they do not prevent the emergence of contrary judgments and motivations. This general strategy can be developed as follows: A definite but encompassing set of goals can become established within a person’s life in virtue of such things as vocational commitments, friendships, hobbies, and other interests. These commitments are not simply to goods as areas of rational interest, but to general goals that can encompass much or all of a person’s life. Such goals can provide a standard for what that person wants his or her life to be about. For such a person,

72

J. Boyle

an action will be regarded as desirable, at least in part, because of its fit with those goals, and at least rough judgments of fit are possible for many options a person faces. Things that do not fit will for that reason be more or less undesirable. Thus, for a chaste person, unchaste acts are always less desirable than chaste acts. And mutatis mutandis, the same holds for other virtues and for established character traits. But when a normally chaste person is seriously tempted to unchastity, the stable framework of valued goals that ranks options by their relationship to chastity is called into question. When that happens, the special (though ex hypothesi immoral) attractions of unchastity make an appeal that can compete with those of living a chaste life. That competition is possible because unchastity has its own attractions, incommensurable with those of chastity. But those attractions do not require the tempted person to repudiate the judgment that chastity is better. The person knows well that when the transient emotional conditions of the temptation recede, the judgment will be the same. So contrary to Davidson, I think that the emotional condition that allows an established valuing of some goals to be called into question by the appeal of what challenges those values is essential to weakness of will. That emotional condition is what allows the person to maintain the judgment that the desirability of a life in accord with established ends is better than anything that would challenge it, but it also introduces some element of desirability incommensurate with that established standard of commensurability. In short, I reject as inconsistent with basic truths about volition and practical judgment the stark version of weakness of will articulated by Davidson. Davidson preserves the logical consistency of this strong conception of weakness of will with our beliefs about action only by rendering it a ‘surd’. There are true judgments of the form ‘this option is more desirable than that one’ that allow for a choice contrary to those judgments, but they cannot be judgments that one option is strictly more desirable than another – that is, more desirable in any respect that would count for the desirability of either. Acknowledgements Parts of earlier versions of this paper were presented to philosophy department colloquia at the University of South Carolina and Franciscan University of Steubenville. I am grateful to the participants. I also thank David Chan and the participants in ‘Values, Rational Choice and The Will, A Special Conference on Values,’ at the University of Wisconsin at Stevens Point, April 2, 2004. I am especially indebted to Michael Stocker for his useful comments. I also thank Germain Grisez, Elmar Kremer, Thomas D. Sullivan and Chris Tollefsen who provided many suggestions and criticisms.

Notes 1

Davidson cites p. 67 in Anscombe (1957) and p. 36 in Hampshire (1965). This assumes that the idea of incommensurably desirable options is coherent. I will not defend this assumption here. For some discussion of this see Boyle (2002). I adapt some arguments from that article in this section. 2

3 Volitions, Comparative Value Judgments, and Choice

73

References Anscombe GEM (1957) Intention. Blackwell, Oxford Boyle J (2002) Free choice, incomparably valuable options, and incommensurable categories of good. The American Journal of Jurisprudence 47:123–141 Davidson D (2001) Essays on Actions and Events, 2nd edn. Oxford University Press, Oxford Hampshire S (1965) Freedom of the Individual. Princeton University Press, Princeton

Chapter 4

The Insignificance of Choice J.S. Biehl

Abstract For some time, philosophers have sought a more satisfactory understanding of the mysteries of morality through a close analysis of its assumed kinship with practical rationality, via the psychological capacity of choice. It is the view in the present paper that no such understanding is possible by these means. The significance of morality has nothing to do with choice. Keywords Choice · Morality · Normativity · Rationality · Responsibility

4.1 Philosophers, and so too the common run of men, have long found it obvious to think of the notion of choice as bearing significantly on morality, particularly with reference to the proper extension of the moral domain and the appropriateness of the moral appraisal of agents in particular circumstances. Such significance is taken to stem from the twofold way we think of choice. There is, on the one hand, the capacity to make a choice; to be able to select among alternatives seems to be a prerequisite for being a denizen of the moral domain. Only those with the capacity to choose count as moral agents, and only such agents are the proper objects of moral address. On the other hand, the possibility of real alternatives in a given situation—the actual having of a choice—seems to play a particularly significant role in the justification of morally addressing an agent in that situation. That is, we think it inappropriate to morally evaluate an agent for having done what she had no choice but to do. And so choice is deemed crucial for the very viability of our moral discourse (and practice). Obvious as all of this may be, I find myself unmoved. I rather doubt that whatever sense we can make of our notion of choice is capable of

J.S. Biehl e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

75

76

J.S. Biehl

bearing the burden that many want our moral discourse to impose. And so I believe that we shall have to look elsewhere to make sense of that discourse, if we choose to do so at all.

4.2 I shall begin by laying out in slightly more detail some ways we have come to think of choice as bearing on certain fundamental aspects of our moral practice. I have noted that we think of choice in two different, though related, senses: in terms of a psychological capacity for making choices and in terms of the circumstantial opportunity to exercise that capacity, what I referred to above as the ‘having’ of a choice. To say which of these senses is basic and which is derivative would not be to the point; both jointly contribute to fostering an atmosphere where moralized impulses are given free reign. Both, that is, play an integral role in making choice a central element of our moral discourse. A moral judgment commonly held by philosophers and non-philosophers alike has it that moral appraisal is fitting, or appropriate, only if the object of appraisal is (or can be attributed to something that is) responsible for being as it is. To be ‘responsible’ in this sense is to be something more than merely causally essential— what I shall refer to as ‘accountable’—for being a certain way.1 It is to be such that one has it in one’s power to be different than one is if one were so disposed. I’ll illustrate this difference with an example. The stereo in my living room, though quite able to play ‘standard issue’ Compact Discs direct from the manufacturer, is nonetheless unable to play CDs I have made myself; it cannot read the information that has been ‘burned’ onto them. The newer stereo in my study, however, can. Clearly, there is something about the technology of the older machine that is accountable for its inability to play these ‘homemade’ discs. In speaking of accountability here, all that is meant is that there is something about the machine that is causally essential to the result in question. But though in this instance at least, it would seem wholly appropriate to attribute accountability to the machine for the sound of silence in my living room, it would be quite inappropriate to attribute responsibility to it: to claim, that is, that the machine did not do what it could have done if it had tried and is for that reason subject to moral criticism. In more colloquial terms, it is perfectly acceptable to say that stereo in my living room is no good for playing burned CDs, but completely unacceptable to say that it is wrong for not playing them and vow to use only the stereo in my study until it changes its tune and becomes more accepting of discs of humbler origin. Stereos, even those out of step with the fashion of the times, are beyond reproach. Of course they are beyond repute as well; they simply are not proper objects of moral address. To be the proper object of such address is precisely to be a thing that is, both generally and in the present instance, worthy of praise or blame, commendation or condemnation. It is, to put it succinctly, to be subject to criticism precisely (and only) because one is responsible for being what one is. Being responsible, as was

4 The Insignificance of Choice

77

said, is a matter of having the ability (what I referred to above as ‘power’) to alter oneself or one’s circumstances—to respond to them—when one is disposed to do so.2 That stereos lack this ability and hence are not taken to be responsible for what they are and how they perform is not in question. What is in question is just what this responsive ability involves in the things that have it. Perhaps a natural thing to say in response to the second question is suggested by the tenor of the discussion so far: animate things have this ability and inanimate ones do not. But though this seems perfectly plausible given that the ability to alter one’s state or circumstances if one is so disposed would seem to apply to single-cell organisms, vegetation, and animals no less than humans, it sorely misses a basic point of talking about responsibility in the first place. The concept of responsibility is a part of our conceptual economy precisely to facilitate moral address; to justify the act of criticism. But what is it to justify moral criticism? One thing it might mean is to make such criticism worthwhile. Indeed, this is an important part of what was meant by saying that the ‘proper object’ of moral address is something that is ‘worthy’ of praise or blame. Hence withholding the attribution of responsibility to stereos can be seen to follow from the judgment that criticizing it won’t get us anywhere. Blaming the machine for its failure to play the music we want to hear will have no influence on its future performance. But the same point applies to most living things as well: morally dressing-down the bumblebee is not going to keep it from stinging you. The questions concluding the previous paragraph, therefore, may well be recast as follows: what things are worthwhile to praise and blame, and what about them makes it so? Answers to these questions are immediately suggested by the pointlessness of criticizing inanimate and (so many) animate things alike: it is worthwhile to praise or blame something only if that praise or blame will have an effect on it.3 The thought that naturally follows is that being the proper object of moral address is really a matter of being responsive to such address. So what can respond to moral criticism and why? Here would seem to be where the concept of responsibility earns its keep: being responsive in the sense in question is a matter of one’s disposition to alter oneself (or one’s circumstances), being itself capable of alteration because of moral criticism that has been directed at oneself. Given this, it is not surprising that bacteria and bumblebees are not properly subject to criticism. But our experience shows that some living things are, as any dog-owner and parent can attest. Given certain preconditions, such as effective conditioning and emotional attachment, moral criticism can be very effective in getting your dog or child to stop (or continue) behaving in a particular way, at least for the present. Indeed, given sufficient comprehension of the attitudes expressed in moral criticism, dogs and children have been known to become disposed to elicit praise and preempt blame. It may take considerable patience and hard work, but subjecting dogs and children to moral critique can prove quite worthwhile. It remains to be answered however just what it is about dogs and children (and any other living thing capable of responding to moral criticism) that enables such responsiveness. What was just said about comprehension, however, goes some way towards addressing this issue. As moral criticism essentially involves the expression

78

J.S. Biehl

of complex patterns of attitudes, usually expressed by language or some other communicative means, responsiveness to such criticism requires a similar (though perhaps not necessarily equally) complex capacity to recognize the attitudes being communicated and, in turn, to have that recognition itself influence, and be expressed by, subsequent behavior. What is required is a sufficiently sophisticated attitudinal aptitude that is combinable with the practical ability to effectively coordinate engagement with the environment. In simpler terms, what is required are capacities constitutive of choice. Aristotle defined choice as ‘deliberate desire of things in our own power,’ (see 1113a 11–12 in Aristotle 1980) and I shall make no attempt to better him. His characterization appears thoroughly useful: to choose is to select something that one could have (or do) after consideration of the alternatives. It hardly makes sense (and serves no obvious point), either to him or us, to say that one chooses what one cannot have or what is not within one’s ‘power,’ for what is not within one’s power is thereby not among the alternatives from which one can choose. Nor are we anymore able (or willing) to see our way to saying that someone chooses something when she has not considered the alternatives and selected on the basis of that consideration. Moving on one’s desire for x without regard for the fact that one could have y or z instead, and why x is to be preferred, hardly seems to be a case of choosing x. We might, perhaps, (if we so chose) attempt to extend our usage of this notion to encompass such examples but we would succeed only in rendering it discursively useless. But let’s return to the present point: how the capacities involved in choice are what make for responsiveness to moral criticism. The point, I suspect, is clear: to be disposed to alter oneself or one’s situation in response to moral criticism requires that one be motivated (desire) to be or behave in ways that one comprehends to be desired by the critic (on consideration of the alternatives). In turn, the effectiveness of the disposition to alter oneself, and thereby of the responsiveness to moral criticism, depends crucially on the ability to realize the alternative deemed appropriate, i.e., the ability to engage with an environment (including oneself) that actually presents the desired result as an alternative (something actually within one’s power).

4.3 Two comments are worth making at this stage. First, the significance of the capacity to make choices for moral address is not dependent on the actual exercise of that capacity. As it happens, much of what we seek to alter (or maintain, for that matter), in ourselves or others, is not the result of choice. Indeed, part of the point of the moral criticism of agents (creatures capable of choice) is to succeed, perhaps for the first time, in bringing the subject of criticism (that for which we criticize the object of address) under the auspice of choice, believing thereby to make the desired result more likely to obtain and easier to predict. Moreover, this contrast between what is chosen and what isn’t (but could be) is itself quite significant to our moral discourse,

4 The Insignificance of Choice

79

particularly with respect to the typology of blameworthy agents. For instance, the akratic, at least in Aristotle’s view, is distinguished in her wrongdoing by acting objectionably against choice, rather than from it, as the wicked do (see 1151a 6–7 in Aristotle 1980). Whether or not one shares this view of the akratic, it is clear that the fact that an agent has not acted as a result of choice—not acted intentionally, as it might be put—may well alter the nature of our criticism, but need not preclude it. In any event, an agent is open to criticism for what she does, not because it was chosen, but because it could have been. Second, specifying the significance of choice reminds us of the naturalness of the thought that it is not merely the psychological capacity to make a choice that renders moral address worthwhile, but also the having of a choice. It is pointless, as we have seen, to engage in moral criticism with something for which the desired state of affairs is not legitimately an alternative, whether that impossibility is a result of some psychological/physical incapacity of the agent or of an unaccommodating environment. In either case, an effective response is not an option. But natural though this thought may be, one might nonetheless be led to question it. Imagine, if you will, that my daughter has joined a cult and, if such a state of affairs is not itself sufficiently odious, imagine further that members of the cult invariably perform acts that lead to their own or others’ death, often both. What I might do in such a circumstance is criticize my daughter for deciding to associate with such objectionable people, in the hopes of moving her to decide to quit. And, in keeping with the discussion so far, if I were to subsequently come to believe that she had been brainwashed by the cult leader so that quitting was not actually a path she could choose, I might desist with the criticism due to its pointlessness and instead seek other means, if available, to dissociate her from the organization. But now imagine that my daughter admits to me that though being brainwashed is quite effective in preventing her from quitting, she would not quit the cult even if she could choose to do so, as she is now deeply in love with the cult’s leader and intends to remain faithfully by her side, come what may. Chances are I might be moved once again to engage in some criticism of my daughter, whether it has influence on her or not.4 But why? What is the point of criticizing an agent in circumstances where she is not responsible, unable as she is to performatively respond to the criticism directed at her? The answer would surely be ‘none’ if the only point to moral criticism were to alter the agent (or her behavior) in some way. But that isn’t its only point. As anyone who has ever engaged in moral criticism of someone whom that criticism cannot reach can attest, such criticism, though without hope of generating the desired—indeed, any—response, can still be immensely satisfying. Expressing the attitudes expressible by means of praise and blame serves needs many of us so clearly have (Strawson 1962). Expressing an attitude can be quite satisfying. Such satisfaction is not to be confused, however, with the satisfaction one might experience in having the attitude. Indeed, the satisfaction that one might feel in expressing an attitude might well be the only satisfaction one gets with respect to it. This is perhaps most easily seen with regard to what might be called ‘negative’ attitudes, such as anger, frustration,

80

J.S. Biehl

disappointment, fear, and guilt, for which expressing, rather than ‘sitting on,’ or suppressing them, is often the necessary first step in freeing oneself of them and the discomfort they involve. This, I think it fair to say, is why we often speak of expressing such emotions in terms of ‘discharging’ them. In any event, it simply is the case that it often feels good to express one’s frustrations, fears, and feelings of guilt, particularly to another, while it often feels bad to be frustrated, frightened, or guilty. Moral criticism, particularly of the condemnatory kind, is the expression of negative attitudes such as anger, disappointment, resentment, repudiation, and remorse.5 Expressing these attitudes often feels better than experiencing them without expressing them. Hence we can see why someone might criticize, say, the policy of a politician, the fees charged by a professional, or the comportment of a parent, though such criticism may well prove ineffectual. But the expressive function of criticism is not always so obviously self-serving. Self-directed criticism, for instance, is often employed not for the sake of changing oneself but to make oneself feel worse. When confronted with (what one takes to be) an irremediable desire or attitude, particularly one that provides illicit pleasure, one may become convinced that the only saving grace available is that one may still castigate oneself and thereby suffer.6 Evincing a more deeply satisfying phenomenal experience is not all that the expression of moral criticism achieves. Moral criticism also serves to solidify one’s ‘practical identity,’ one’s sense of self.7 Through such criticism, one can confirm to oneself and others what one stands for, what one will accept and what one forbids which, often more incisively than anything else, tells the world who one is. For one who needs to show others, or remind oneself, what one is about, moral criticism can be invaluable. Engaging in moral criticism, then, serves purposes beyond prodding the object of that criticism to choose to be in some sense different, and hence is not necessarily undermined in situations where the criticized agent ‘could not do otherwise.’ Nonetheless, there is still a significant connection between such criticism and our thoughts about choice. When my daughter tells me that she would remain in the cult even if she weren’t brainwashed, I am being told that if she could choose she would choose in a way I find objectionable. It is this belief, that an agent would choose in a certain way were it open to her to do so, that gives criticism a form of life that it would not otherwise have. It is natural, or more likely, a naturalized practice, for the attitudes expressed by moral criticism to be elicited in response to the attitudes discerned in the agent so addressed.8 This fits nicely with the idea, discussed above, that engaging in criticism is worthwhile only to the extent that the object of criticism can be understood as being (potentially, at least) responsive to such criticism. For this suggests that the interaction, of which moral criticism is the backbone, is itself a matter of the inter-responsiveness of attitudes. Hence the oddity of engaging in moral criticism of stereos or bumblebees is not only that there is no possibility of the desired attitudinal response but that the attitudes expressed by that criticism are not themselves in response to attitudes to which one objects.9 But the naturalness of the conception of moral discourse being rooted in attitudinal interaction also helps to explain our intuitions in the case of my brainwashed daughter.

4 The Insignificance of Choice

81

When I find that my daughter has been brainwashed and so cannot respond to my criticism by choosing to quit the cult, I, rather naturally, find that my anger (or disappointment, resentment, or what have you) either dissipates or directs itself elsewhere, namely towards the attitudes of those persons accountable—and presumably responsible—for brainwashing her in the first place. The attitude initially found objectionable—an attitude of my daughter—is, on further review, believed to be either non-existent or beyond her efforts to alter. But when I come to believe that the attitude in question does exist and, further, that if it were in her power to alter she wouldn’t, my reactive attitude is refocused on what initially incited it. My repudiation is, once again, of my daughter’s attitude, hence any criticism I engage in, even if only for the sake of my own benefit, will naturally be directed at her. So here too, where the agent does not properly have a choice, the capacity to make a choice is significant for rendering moral criticism fitting. To be sure, it is fitting not in correctly attributing responsibility in the standard, indicative sense but rather in a more exceptional, subjunctive sense. This is why we do not find it objectionable when someone morally criticizes an agent that could not have done otherwise if she would not have done otherwise anyway. Indeed, it appears a perfectly natural thing to do.

4.4 There is an objection to the foregoing that is of a fundamental nature. I announced at the commencement of Section 4.2 that I would elaborate on some ways we have come to think of choice as significant for our moral discourse, particularly our engagement in moral criticism, the appraisal of agents by means of praise and blame. The objection is that, even if one were to grant that I have shown how choice plays a significant role in the practices I have discussed, I have not shown in any way that choice plays a significant role for any moral practice. This, the objection continues, is because the practices I have discussed are not in any way moral. Taking the measure of this objection will be the focus of the remainder of this essay. To preview, however, what I shall say about this objection amounts to an elaboration of two claims. The first, on which I will spend little time, is the ready concession that there is an operative sense of ‘moral’ which is not applicable to the practices I have been discussing. The second, which is my primary concern, is that choice is not significant to practices which are moral in the operative sense, insofar as we can understand that sense at all.

4.5 To say that a practice, or a claim, or a judgment is moral, in the sense that will now concern me, is to say that it is normative. But the ease with which one can state ‘morality is essentially normative’ belies the tremendous difficulty awaiting anyone

82

J.S. Biehl

attempting to state just what this claim amounts to. The difficulty is so great in fact that some philosophers (and many non-philosophers) have been led to think that it doesn’t amount to anything at all. Yet while no attempt at a fully fleshed-out picture of what the normativity of morality involves has managed to rise to the level of a ‘received view,’ even an unsatisfying one, there seems sufficient agreement concerning fundamental bits of the skeleton. Hence characterization of the normative typically begins by emphasizing that it concerns what ought to be (or ‘should,’ or ‘supposed’ to be) rather than merely what ‘is.’10 This seems to be the most basic normative notion and yet the hardest to hang any flesh on, as it is also the most elusive. It is not obvious just what it is to say what ought to be, as distinct from what is, or what you want to be, or what you believe to be, or what you believe could be, or will be, or what needs to be if things are to be the way you want them to be.11 But helping to give it more substance is a second element, the idea of a ‘standard of correctness’ to which, for instance, objects of moral criticism are measured.12 Indeed, this idea, without introducing any real content, combines with the enigmatic ‘ought’ to enable the construction of the most basic moral formulae. From the perspective of morality qua normative, moral criticism just is the expression of a judgment of how an object ‘measures up’ to the standard. Praise is the expression of the judgment that an object satisfies the standard, and hence is as it ought to be, while blame expresses (the judgment of) an object’s failure to satisfy the standard, and hence to be not (entirely, at least) as it ought to be. Familiarity with these two elements of the normative conception of morality alone would have sufficed to make clear that no such idea was being discussed when I spoke earlier of ‘moral criticism,’ ‘moral address,’ ‘praise,’ ‘blame,’ ‘responsibility’ and the like. When I claimed that moral criticism was appropriate or justified only if the object of criticism was capable of choice—i.e., an agent that could alter itself or its behavior in response to that criticism—the ‘justification’ at issue was entirely pragmatic. What I was saying was that it was not worth the effort, because one’s aims would not be realized, to criticize something or someone that was incapable of responding to that criticism in the way desired. There was no implication that the object of criticism ought—whatever that actually means—to respond to such criticism at all, let alone in the way implied by that criticism. Nor was it implied that there was some standard, beyond the concerns of the critic, that the object was being measured against. Indeed, the entire discussion could have proceeded without any reference to ‘moral’ criticism at all: any kind of address of the object is worthwhile provided the object is an agent capable of responding to it in the way desired. Addressing the agent in terms of what is ‘right,’ ‘wrong,’ ‘good,’ or ‘bad’ matters not to the issue of responsibility as I have been discussing it.13 Likewise, the more ‘expressive’ dimension of criticism I touched on has, at bottom, nothing to do with morality in the normative sense. To use ‘moral’ criticism to express one’s anger, disappointment, or repudiation (or happiness, pleasure, or identification) implies nothing about what ‘ought’ or ‘ought not’ to be. Neither does improving one’s phenomenal state nor solidifying one’s practical identity need to be understood in terms of the critic’s attitude about what ought to be the case. Nor does the ‘naturalness’ of reactive attitudes in light of the attitudes of the objects of

4 The Insignificance of Choice

83

criticism suggest that the attitudes, either of the critic or the criticized, ought to be as they are or different. Again, no standard beyond that of the critic’s own temper need be (or is) involved at all. My discussion of the significance of choice has not been a discussion of the significance of choice to morality.

4.6 Morality, insofar as it is normative, is not to be conflated with some pragmatic or expressive practice one may or may not engage in.14 Morality is something different but as to what I can offer no answer. (I doubt an answer is possible though I will not argue that here.) My present concern is with what I take to be the constitutive inability of choice to help in the formation of an answer. A steady diet of choice will not put meat on the bones of morality. Before stating why choice is unhelpful for making sense of morality, it would be worthwhile to consider why one might think it is. This thought arises from two other pieces of the normative skeleton, the first of which is expressed, perhaps confusedly, in the slogan ‘ought implies can’. Without knowing to what the term ‘ought’ refers or means, it is rather hard to suggest any claim about what it implies. Nevertheless, our earlier discussion of the pragmatics of critical address can shed useful light on the matter. It was said that addressing something about its performance in a critical mode— the expressive function of that criticism aside—was only worthwhile if it could respond to that critique. The ability to respond to the critique was, in turn, analyzed in terms of the ability to alter oneself or one’s situation in a manner sensitive to both the critique and the possibilities for such change. That ability was identified with choice, the ability to select from among alternatives in light of their comparative recognition. The pragmatic point is that the utility of such subtle attempts at behavior modification presupposes ‘can.’15 Non-agents—creatures incapable of choice—can’t be modified by criticism. Considerations not too dissimilar might be applicable in the case of morality. What ought to be, ex hypothesi, is over and above what is. Hence moral criticism often involves a ‘difference claim’—and necessarily so in the case of blame—with respect to reality. That is, to criticize something for failing to measure up to the standard of correctness necessarily involves the claim that how it is is not how it ought (entirely) to be. In this, as in other respects, normative claims are not unlike claims about what I, you, or someone else might want to be, for what anyone wants to be is never merely a matter of what is and might not be at all.16 Now if the point of engaging in moral criticism as a mode of discourse isn’t merely to convey the information of the object’s current normative status but to possibly change it, then the pragmatic principle would clearly apply. If you want to change the normative status of something through moral criticism, then that thing had better be capable of choice (see pp. 119 ff. in Williams 1981).

84

J.S. Biehl

The second aspect of morality’s normativity that leads to thoughts of choice concerns the special status afforded to the standard of correctness. Unlike almost all other standards, that of morality is such that one is required to be subject to it; the appropriateness of measuring one by it is not a matter of anyone’s discretion. The regulative ideal that the standard embodies is not one that it is possible to avoid. Yet the application of most (all?) non-moral standards, like the aptness of most non-moral criticism, is a pragmatic affair. Consider, for instance, academic standards that measure the appropriateness of awarding a degree. Such standards enable us to rank students as better or worse, distinguishing who ‘measures up’ and who does not. But though these standards specify what is required of the students subject to them, the students themselves are not, in the most fundamental sense at least, required to subject themselves to the standard. Whether they do or not is for them to decide. This point is easy to underestimate, as it may appear to be merely the familiar thought that many demands on behavior are conditional on the aims of agent. Provided a certain aim, there are certain things that one simply has to do in order to achieve it, and success or failure in doing them is therefore a measure the agent cannot avoid. But the agent who does not do them cannot be said to fail to do them if she has abandoned (or never had) the aim in question. But the pragmatic point with respect to standards goes deeper than this. An agent might well have aims distinct from, or superceding the aim that subjects him to a standard, pursuit of which lead him to behave in ways designed to be measured by the standard in a particular way, even as failing to meet it. Indeed, even an agent’s only aim could be such that it subjects him to a standard the failing of which constitutes satisfying the aim. Consider an intrepid Romeo lusting after some wholesome co-ed. Believing that she will not welcome his advances under most circumstances, he comes up with a plan. He decides to enroll in a logic class that he is more than capable of passing and then proceeds to perform at a level such that he fails the midterm. This is precisely as he wants it to be since he is now required by the instructor to arrange sessions with the tenderly sympathetic tutor for the class (a veritable philosophical Florence Nightingale), none other than his wholesome prey. Most standards have, to greater and lesser degrees, peripheral applicability to those subject to them. This is why it sounds so out of place to say of someone who has measured poorly with respect to one that she ought to have performed better, where this is not meant in the eliminable—and therefore non-normative—sense of expectation.17 But there is one putatively non-moral standard that does not seem peripheral in this way. It, as morality is said to do, appears to impose a standard that all who could be measured by it necessarily are. This is the standard of rationality and measuring up to it is also frequently claimed to be an inescapable requirement. And what this standard measures, in its practical mode anyway, is just what morality measures. Practical rationality, like morality, is concerned with action motivated in light of reasons, which is to say in light of reflection upon one’s circumstances, the possible changes to them one could induce, and the relative preference, or evaluation one

4 The Insignificance of Choice

85

gives to their obtainment. That is to say it is concerned with choice.18 Understanding (practical) rationality requires an understanding of choice, so it stands to reason that the other inescapable standard might find choice important too. In being central to rationality, choice holds out the promise of not simply giving moral criticism a pragmatic utility it would not otherwise have, but of bringing the authority of the rational standard to the aid of the rather obscure authority of morality. Perhaps the very idea of what (morally) ought to be can be illuminated by the idea of what rationally ought to be. Indeed, maybe measuring up to the rational standard just is what the normativity of morality is all about. One can see why this might be an attractive idea; it would certainly appear to simplify matters. For if the moral standard and the rational standard were one and the same, then being moral would be a matter of being rational and being (or doing) what ought not to be—immorality—would be a violation of rationality, i.e., irrational. Moreover, measurement by the moral standard would be unavoidable in a way that has intuitive coherence: insofar as one is an agent one is either rational or irrational; being arational is not an option. Whereas the student can avoid failing by the measurement of an academic standard by dropping out, the agent must always find himself registered for morality. Moral assessment, because it involves rational assessment, is inescapable; here no dropouts are allowed. This idea, no doubt attractive to some, is difficult to maintain. Accounting for the normativity of morality in terms of rationality holds promise only insofar as we can hope to give an informative account of the normativity of the latter, and this we cannot do.

4.7 A choice is the determinate motivation—the intention—that usually follows in the wake of reasoning dependent on what the agent believes about her alternatives, about the possible consequences of any of them obtaining, about the purpose(s) of the choice situation, and of the relationships between all three.19 Such reasoning tends to issue in a belief that selecting this alternative will (probably) satisfy the purpose of the choice situation.20 Provided the motivation to satisfy that purpose is sufficiently strong, it will become determinate—the intention will be formed.21 That is, a choice has been made. Choosing something is always a matter of selecting something for an explicit reason. It is to decide; to be motivated in light of a reflective awareness of the possibilities. Aristotle, we will recall, defined it as ‘deliberate desire for things in our own power’ (see 1113a 11–12 in Aristotle 1980), what we desire as a result of deliberation. Choice, in other words, is reasoned desire. It is here that we can find the content behind his famous claim that ‘choice relates to the means’ (see 1111b 27 in Aristotle 1980). Indeed, his claim amounts to a conceptual truth: to choose something is always to choose it for the sake of something, as a means to some end, the limiting case being that of choosing something for its own sake, where the means and the end are one.

86

J.S. Biehl

Some view this as a welcome fact about choice, for it suggests that the standard by which it is assessed—the standard of practical rationality—is precisely what they have claimed it to be, namely the instrumental standard. Reasoning about what to (choose to) do is reasoning about means to ends. The purpose of a choice situation is the end and the selected alternative is believed to be the (best) means to satisfy it. And this in turn helps explain why those who believe instrumentalism to be the (one and only) standard find other proposed standards of practical rationality so uncompelling. Imagine I have a few desires I would like to have satisfied. One is to have another drink tonight and another is to wake up tomorrow refreshed and ready to philosophize, and a third is to keep my promise to myself to stick to my diet and not overload on alcohol (too many carbohydrates). As it happens, I have already gotten a taste for the bourbon and decide, after careful consideration of the alternatives, to have one more for the road (so to speak, of course). The inherent instrumentality of choice makes my choice perfectly rational. Insofar as I chose to have another drink, I did so for a reason, believing it to be the alternative open to me that best satisfied my purpose(s) for entering the choice situation. Whatever that purpose was, I came to believe that having another drink was the best way to achieve it. My choice, given my purpose and my beliefs, was, by the instrumental standard, rational. Now imagine that someone, upon hearing of this tale, tells me that I have chosen irrationally. Assuming it is agreed that taking what I believe to be the means to my end is rational, an immediate response I might have to this person is that they are simply confused.22 But if the person were to persist, claiming to appeal to a rational standard that applies to ends rather than to means then my likely response will be that she is equivocating. Clearly, the standard that applies to ends cannot be the same as that which applies to means if one is allowed to speak legitimately of choosing what one believes to be appropriate—i.e., rational—means to irrational ends. Equivocation in the defense of reason is no vice, it might be said, but it nevertheless forces us to face the question of just what could possibly be meant by speaking of a rational standard for ends that involves not even a surreptitious appeal to some further end in light of which the former can be seen as means. We can, after all, always ask the question why when we are told that it is rationally necessary that our chosen ends be such that they are, say, optimal—their realization coheres with the (possible) satisfaction of the majority of our present desires. What emboldens the instrumentalist about practical reason is that dissenters from his position seem faced with the dilemma of either treating this question as legitimate or not. Conferring legitimacy buys into the requisite sort of answer, which is to claim that the end in question, coherence say, or consistency, is the necessary (or best) means to some end we are assumed to have, such as to live as ‘unified’ a life as possible, or to increase the likelihood of realizing as many of the aims we in fact have. Such an answer obviously fails to escape from the instrumentalist orbit. If, on the other hand, they reject the question, preferring instead to claim that the pursuit of such an end is simply ‘constitutive’ of creatures like us (i.e., rational creatures), then they have buried the notion of a rational standard in obscurity.23 Indeed, they have actually

4 The Insignificance of Choice

87

changed the subject, from a discussion of the appropriate means by which to assess human thought and behavior to an analysis of human nature. This is not ethics or a ‘theory of rationality’. This is metaphysics.

4.8 However pleased the instrumentalist about practical reason may be with this result, it must be unwelcome to those who seek to dispel the obscurity of the normativity of morality by its assimilation to the normativity of rationality. If pursuit of the true rational standard leads one into metaphysical speculation then it must be obscure indeed. And barely comprehensible. Imagine, for instance, being told that it is constitutive of human nature to have the goal of being motivationally consistent; to develop as coherent and unified a motivational set as possible.24 On such a story, (human? rational?) agents measure up to the rational standard—flourish, as it might be said—if their natural propensity is actualized. Numerous questions immediately arise, however. The obvious failure of so many agents to flourish in this way beggars the question of just how it was determined that this in fact is the standard.25 Is this a feature that some very ‘successful’ agents have exhibited? Successful in what way? Getting what they want? And why is it that so many of us aren’t flourishing in this way? If we could wouldn’t we, given that this is our ‘nature’?26 Maybe we don’t flourish (in this sense) because we can’t.27 Whatever we may think about such an approach to a standard of rationality, it is clear that it no longer has any significant connection with the capacity to choose. If the (practical) rational standard is in fact concerned with ends, then it really isn’t measuring the choices an agent makes so much as measuring the agent herself. As a metaphysical account, about the ends that are ‘basic’ to human beings as such, the failure to have (any of) them suggests a problem of the most profound sort—such unfortunate souls were botched by Nature—not something that could be remedied by getting the agent to choose to procure them.28 Such a standard, then, is without pragmatic utility; criticizing an agent for not having them can do nothing to move her to obtain them. And though we might readily admit the metaphysical standard could still retain its ‘normative authority’, we are now left grasping for some content to the claim that an agent that does not have a requisite end nevertheless ought to have it.

4.9 I have sought to show that attempting to avoid the conclusion that the practical rational standard is instrumental necessarily leads one into the dimly lit terrain of ‘the nature of things,’ where talk of how things ‘ought to be’ is not meant in terms of expectation yet is nonetheless intelligible.29 But this might be thought to be looking at things the wrong way around. Agents without the goal picked out by the

88

J.S. Biehl

metaphysical-rational standard are beyond the pale. Rational address—and moral address if it is properly understood in these terms—is meant only for those who have the end. So constrained, we can make sense of such address because it is quite intelligible to talk about someone choosing means insufficient to some end (including an end [or ends] that all human beings by their nature have). Here is ‘failure’ for which the term ‘irrational’ intuitively fits. And here such address—criticism—has a point beyond merely conveying information: perhaps the agent can be moved to take care that in the future she chooses wisely, rather than poorly. If we take this line of thinking seriously, and here I shall offer no reason not to, we will do well to notice the following. If the agent has an end but is not realizing it then that can only be due to her having false beliefs about what would be the appropriate means to it.30 And this is not a failure of choice, per se, but of the fixation of belief. This is a failure to measure up to the theoretical standard, not the practical. Indeed, by the standards of instrumental rationality, failing to achieve the end in such circumstances is precisely what would be expected. Only the very lucky or the instrumentally irrational could possibly take the appropriate means to their end while falsely believing what those means are. In fact, only the very lucky could; instrumental irrationality is an incoherent notion. As the capacity to select the alternative one believes to best satisfy the purpose of the choice situation, choice is a mechanical psychological process that can’t ‘go wrong’.31 In any given situation, agents either choose or they don’t.32 Presumably the instrumentalist about practical reason will not be pleased about this result because it shows that the instrumental ‘standard,’ if it measures anything at all, isn’t measuring choice. Or, to put it somewhat differently, choice is such that it can’t help but measure up to it. An inviolable standard, however, is no standard at all. If the practical standard of rationality concerns ends, then its normativity—if it’s normative—is quite obscure. But if the practical standard concerns means—that is, if it’s instrumental—then it isn’t normative.33 This was Hume’s view. As his discussion of what counts as ‘contrary to reason’ clearly reveals, he thinks nothing does, at least not anything with practical import such as choice. As he puts it, ‘a passion must be accompany’d with some false judgment, in order to its being unreasonable; and even then ‘tis not the passion, properly speaking, which is unreasonable, but the judgment’ (see II.iii.3.416 in Hume 1978). For Hume, if anything counts as normative in the rational sphere it concerns belief, not choice. And it is this view, coupled of course with his conviction that morality is inherently practical, that leads him to deny that one does well to understand morality by assimilating it to rationality. Hume’s normative nihilism with respect to practical reason has at times been dismissed rather brusquely. It is often attributed to a question-begging construal of rationality as concerned with truth and falsehood.34 Given this, of course it would seem plausible that reason (understood as a normative domain) is not concerned with the practical. But on the account presented here, where the non-normativity of practical reasoning is derived through the analysis of choice, limiting normative reasoning to the fixation of belief—assuming that that is normative—is actually the conclusion.

4 The Insignificance of Choice

89

4.10 The thesis of this essay is that the obscurity of morality qua normative is not to be dispelled by attempting to see it in terms of some inescapable standard of correctness that applies to choice. Choice is a functional capacity that is constituted by the selection of an alternative the agent believes to (best) satisfy the purpose of a choice situation. As such, choice is, in these unfortunate terms, ‘by its nature correct,’ since not to select what one believes to be the best alternative is simply not to choose. If practical rationality concerns reasoning that issues in choice, then such reasoning is not normative.35 This is not to claim that every agent’s belief as to what best satisfies the purpose of the choice situation is correct, i.e., true. No doubt many choices fail to deliver all that is hoped of them. But this ‘failure’, such as it is, cannot be understood as a failure of choice but rather of belief. If we are to allow that a mistake can be made, it has to be in the fixation of the agent’s beliefs concerning the choice situation. If one maintains the hope of explaining morality by appeal to rationality, then that rationality must be theoretical (and, of course, normative). What (morally) ought to be, on this picture, would be that agents have certain true beliefs. Given the nature of choice, the rest would take care of itself. Interestingly enough, this seems to be the very picture of morality that we find championed by the Socrates of Protagoras, compressed in his infamous claim that ‘no one goes willingly toward the bad or what he believes to be bad’ (see 358c–d in Plato 1992). Given that everyone does what they believe to be best, moral criticism—and moral education—is a matter of getting people to believe what is true. If what you believe is best is best, then you will (attempt to) do what’s best: in this sense the Good and the True are one. If you find that idea rather hard to swallow, then you just might be a normative nihilist tout court and then your task is to show that a normative standard for belief is no more coherent than that for choice.36 If, on the other hand, you still hold hope for an understanding of morality, then it is not theories of practical reason to which you must turn but rather epistemology.37

Notes 1 My use of the nomenclature is largely arbitrary. The notion I am intending to capture by ‘responsible’ many would claim could be, or indeed is, captured by ‘accountable’. Nothing of consequence turns on such terminological choices, so long as consistency is maintained. Two considerations that influence my usage are, first, the preponderance of ‘responsible’ and ‘responsibility’ in the philosophical literature and, second, its rhetorical suitability. 2 A vexing question is whether the responsibility in question extends to the disposition to alter oneself or not. That is, in order for something to be responsible for itself, must its ability to alter itself when it is so disposed extend to the disposition to use that ability itself? The question appears to suggest a dilemma that threatens the cogency of the concept of responsibility: answering ‘yes’ seems to invite an infinite regress of abilities and substrata that employ them while answering ‘no’ seems to blur any real distinction between responsibility and accountability. This issue is the specter haunting every discussion

90

J.S. Biehl

of the nature and extent of responsibility, the current one being no exception. Though I do not deny this problem, I do intend to ignore it henceforth. 3 For more detailed discussion of (some variations of) this idea, see Schlick (1939); Smart (1961); Dennett (1984); and Smiley (1992). 4 Discussion of this type of example stems from Frankfurt (1969). 5 The points being made here are not peculiar to negative attitudes. Expressing ‘positive’ attitudes such as love, esteem, pride, happiness and the like often adds to the pleasure associated with the experience of those attitudes. Commendatory criticism, such as the expression of admiration and gratitude, is in this regard not fundamentally different from its condemnatory counterpart. 6 For a rather forceful presentation of this idea, see III, 20, 101–3 in Nietzsche (1998). 7 This notion of ‘practical identity’ is developed at length in pp. 101 ff in Korsgaard (1996). 8 See Strawson (1962) for the seminal discussion of what he calls ‘reactive attitudes’. 9 Indeed, in the case of stereos and bumblebees, the attitudes expressed by criticism are not responsive to attitudes at all. 10 Not ‘merely’ because, assuming the coherence of normative description, the way things ought to be in any given situation may also be the way things are. That is, normative and non-normative descriptions are not mutually exclusive. 11 This list could be extended further. 12 Railton (2000), especially in pp. 1–5, provides perhaps the most helpful account of this idea. 13 Except to the degree that such terms can render criticism more useful than it would otherwise be. 14 I am inclined to believe that the pragmatic and expressive practices I have discussed are necessary for certain creatures—namely, humans—to engage in. But this is not to believe them normative. 15 I call such attempts ‘subtle’ simply to distinguish them from cruder forms of modification, such as force. 16 Given our linguistic habit to convey obliquely to others what we want (or what we believe others want, or what we believe will satisfy what others want, etc.) by ‘ought,’ it is not difficult to imagine oneself believing that what ought to be and what one wants to be are one and the same. 17 The standards employed in sports illustrate this point nicely. We have, for instance, standards that determine who is the most outstanding player for the year, but there is no suggestion that anyone ought to be such, where that is understood normatively. 18 As I understand it here, a standard of practical reason measures choice, whereas, say, a standard of theoretical reason measures belief. If there is a rational standard that measures something else, say emotions, or desires, then that would be distinct from these two. Unless explicitly stated otherwise, talk of the ‘rational standard’ is to be understood as the practical standard. 19 This description is, on certain occasions anyway, incomplete. Perhaps the most important element it leaves out is the relationship between the believed purpose of the choice situation and other purposes the agent might have. Depending on what the agent believes about her alternatives and the possible consequences of their obtainment, these other purposes may or may not become relevant. Indeed, what the agent comes to believe about the alternatives (etc.) may enable one of those other purposes to become the purpose of the choice situation. This does not alter the characterization of choice in any fundamental way, however, and I will take it as read in what follows. 20 To be more precise, a belief that this alternative will (probably) satisfy sufficiently or best the purpose of the choice situation. I will frequently simply say ‘best’. 21 Hence the qualifier ‘usually’ in the first sentence of this paragraph. Much, perhaps most, of the time the motivation to satisfy the purpose of the choice situation is sufficiently strong, which often serves to explain why one faces the choice situation in the first place. But this needn’t always be the case, and nothing precludes the motivation to satisfy that purpose waning in light of the belief as to what alternative will best satisfy it. I take this to be a basic difference between my view of things and that of James Dreier. In effect, I am responding affirmatively to his rather rhetorical comment: ‘Quite possibly the distinction between desiring and willing is important here’ (see p. 90 at note 13 in Dreier 1997). Indeed, I believe it is. This basic difference, as will be seen, goes a long way. 22 I am, at present, purposely putting aside the possibility that the charge of irrationality in this context is meant to indicate that what you believed to be the best means to your end was in fact not. I will return to this later.

4 The Insignificance of Choice

91

23

Alternatively, we can imagine someone saying that the pursuit of such an end is just a fact, presumably a non-natural one, not so much about creatures but about reality. Such an answer would, I take it, be no less obscure than one grounded in the idea of some creature’s nature. It may, in fact, be more obscure, if that is possible. 24 This would seem rather like the view proposed in Smith (1994). 25 This is not peculiar to the example used; this question can be raised with respect to any proposed account of our (ideally) rational nature. 26 Given the undeniably formal nature of an end such as this, it would seem rather easy to satisfy it: pick a motivation, any motivation, and eradicate any and every motivation that is inconsistent with it. The result will be a consistent set. 27 Or at least not without intense spiritual discipline: the suggestion under discussion sounds rather like a Western philosopher’s take on Buddhism. 28 Imagine if one could so choose: the inherent instrumentality of choice entails that such a choice—that of basic ends—could only be made for the sake of something else. And so these ends would be means (they could not be chosen for their own sake, i.e., as ends in themselves, for that would imply that the chooser already had said ends: the motivation towards them would be ‘original’ rather than derived). 29 Is the statement ‘an anencephaletic ought to have had a brain’ normative? If not, why think the statement ‘an agent without the goal of coherence (or what have you) ought to’ is? 30 Actually, it might be due to the (motivational) predominance of other ends the agent has, but what would make this situation ‘irrational’—or more to the point, what the content of such a claim would be—I cannot say. 31 An agent that chooses not to select the alternative that best satisfies the choice situation has done so for a reason and has thereby abandoned the original purpose of the choice situation for another. 32 Consider the discussion of choice offered by Boyle (2008). Boyle offers three ‘distinct forms of choice’: motivations responsive to judgments of equal desirability of options, judgments that one option is more desirable than the others, and judgments that the options are incommensurable in desirability. The first and third forms of choice make incorrect choice incoherent. The second appears to do so as well, unless some sense can be made of the idea of reasoned desire for the less desirable option. I doubt very much that it can. 33 For discussion of the inviolability of the instrumental standard entailing its non-normativity, see Korsgaard (1997) and Finlay (2008). This, of course, is the result of distinguishing between choice and mere desire, a distinction that some (James Dreier, perhaps) are unwilling to draw. See note 21 above. 34 See, for instance, the introduction at pp. 6–7 in Cullity G, Gaut B (1997); Broome (2001). 35 As I have tried to show, if practical rationality concerns ends it is not clear what the claim that it is normative even amounts to. 36 In whose number I would include Protagoras, Hume, Nietzsche, James, Rorty, and the present author. 37 My thanks to the following for helpful discussion and comments: Bernard Baumrin, Joseph Boyle, David Chan, Stephen Ellis, Stephen Finlay, Martin Harvey, Daniel Kaufman, Mark Sheehan, and David Shein.

References Aristotle (1980) Nicomachean ethics. Ross WD (trans), Ackrill JL, Urmson JO (revs). Oxford University Press, Oxford Boyle J (2008) Volitions, comparative value judgments, and choice. In: this volume Broome J (2001) Normative practical reasoning. Proceedings of the Aristotelian Society Supplementary Volume 75:175–93 Cullity G, Gaut B (eds) (1997)Ethics and practical reason. Oxford University Press, Oxford Dennett D (1984)Elbow room: the varieties of free will worth wanting. The MIT Press, Cambridge MA

92

J.S. Biehl

Dreier J (1997) Humean doubts about the practical justification of morality. In: Cullity G, Gaut B (eds) Finlay S (2008) Against all reason? Skepticism about the instrumental norm. In: Pigden C (ed) Hume, motivation, and virtue. MacMillan, New York Frankfurt H (1969) Alternate possibilities and moral responsibility. Journal of Philosophy 66: 829–39 Hume D (1978) A treatise of human nature. Selby-Bigge LA (ed) Nidditch PH (rev), 2nd ed. Clarendon, Oxford Korsgaard C (1996) The sources of normativity. Cambridge University Press, Cambridge Korsgaard C (1997) The normativity of instrumental reason. In: Cullity G, Gaut B (eds) Nietzsche F (1998) On the genealogy of morality. Clark M, Swensen AJ (trans). Hackett, Indianapolis Plato (1992) Protagoras. Lombardo S, Bell K (trans). Hackett, Indianapolis Railton P (2000) Normative force and normative freedom: Hume and Kant, but not Hume Versus Kant. In: Dancy J (ed) Normativity. Blackwell, Oxford Schlick M (1939) When is a man responsible? In: Rynin D (trans) Problems of Ethics. Prentice Hall, New York Smart JJC (1961) Free will, praise and blame. Mind 70(271):291–306 Smiley M (1992) Moral responsibility and the boundaries of community. University of Chicago Press, Chicago Smith M (1994) The moral problem. Blackwell, Oxford Strawson P (1962) Freedom and resentment. Proceedings of the British Academy 48:1–25 (Reprinted In: Watson G (ed) (2003) Free will, 2nd edn. Oxford University Press, Oxford) Williams B (1981) Ought and moral obligation. In: Moral luck. Cambridge University Press, Cambridge

Part III

Desire and Intention

Chapter 5

The Indeterminacy of Desire and Practical Reason Patrick Fleming

Abstract Bernard Williams’ argument for internalism about reasons for action has received a great deal of attention. However, another element of Williams’ view about reason has not been much discussed. Williams claims that it is often indeterminate what reasons an agent has. This paper defends that view by arguing that one’s desires are often indeterminate. Keywords Desire · Indeterminacy · Internalism · Practical reason

5.1 Introduction Desires have some important relation to reasons for action. The nature of this relation has been the subject of a great deal of debate. For instance, Bernard Williams has spawned a great deal of literature with his claim that all reasons for action are internal reasons.1 That is, all reasons must be linked to the agent’s subjective motivational state by a sound deliberative route. In this paper, I wish to draw attention to a much less discussed aspect of Williams’ papers on internalism. Williams believes that there is an essential indeterminacy regarding what an agent has a reason to do. This paper explains indeterminacy in practical reason and gives a qualified defense of it I argue that indeterminacy has two sources, according to Williams. One source is that deliberation is guided by imagination, not by rules. The second is that an agent’s motivational set can be indeterminate. I do not attempt to evaluate or defend the first sort of indeterminacy. Rather, I argue that even if we reject this sort of indeterminacy, we are still left with the indeterminacy of desire. The indeterminacy of desire sheds light on some little discussed problems in practical reason. It points to the fact that agents have to solve constitutive questions about their subjective motivational set (or their S in Williams’ shorthand). Sometimes agents must shape P. Fleming 218 Sheldon Hall, Department of Philosophy and Religion, MSC 7504, James Madison University, Harrisonburg, VA 22807, USA, Fax: 1(540) 568-8072 e-mail: [email protected] D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

95

96

P. Fleming

their motivations by deciding what kind of person to be. This paper does not attempt to show how this can be done in a rational manner. I am content to draw attention to the issue because the indeterminacy of desire creates serious problems for some prominent accounts of practical reason. In particular, Michael Smith and Christine Korsgaard introduce formal constraints on the permissibility of desire to guide practical reason. I think that in many cases, individuals have motivational content that is too unsettled for these constraints to do the work that is needed. I will argue for this largely by example. Of course, the case will not be conclusive, but indeterminacy offers a straightforward explanation of some central features of practical thought.

5.2 Williams on Indeterminacy Williams’ basic position on internalism is that, for a consideration to be a reason, it must be possible for an agent to be motivated to act on that consideration by undergoing a process of rational deliberation. If a rational agent could not be motivated to act on a consideration, then that consideration is not a reason because it could not explain the agent’s action. It is not clear what this rules out and I will not try to settle that issue here.2 However, internalism, by itself, does not pose a threat to rationalist accounts of morality. As Korsgaard has shown, the internalist requirement on reasons does not refute ethical theories as it merely makes a ‘psychological demand on them’ (see p. 23 in Korsgaard 1986). Whether internalism results in a position that is skeptical about the force of reason in morality will depend on how one fills out what constitutes a ‘sound deliberative route’. Practical rationality might constrain deliberation in such a way that makes internalism compatible with a Kantian picture of morality. For instance, Korsgaard and Smith insist that practical reason requires an element of universalizability that insures moral requirements are binding on all agents. Williams is skeptical of such positions. We should note though that in his papers on internalism, Williams does not present an argument against such positions. He just doubts that they can be made out clearly. Williams says, ‘Someone who claims the constraints of morality are themselves built into the notion of what it is to be a rational deliberator cannot get that conclusion for nothing’ (see p. 37 in Williams 1995).3 Korsgaard and Smith do not attempt to get the conclusion for nothing, but I will not assess their attempts here. Williams’ account of a sound deliberative route is intentionally vague and imprecise. He wants to allow for any number of possibilities in deliberation. This is because Williams thinks it is impossible to give a full and detailed account of what practical reason consists in. In fact, it would be a mistake to interpret Williams as offering a procedural account of practical reason if that means there is some single procedure that constrains deliberation. Williams does not believe any such procedure exists. He argues that ‘there is an essential indeterminacy in what can be counted a rational deliberative process’ (see p. 110 in Williams 1981). He offers a partial list of the possibilities for deliberation that includes ‘thinking how

5 The Indeterminacy of Desire and Practical Reason

97

the satisfaction of elements in S can be combined, e.g., by time ordering; where there is some irresoluble conflict among the elements of S . . . finding constitutive solutions, such as deciding what would make for an entertaining evening, granted that one wants entertainment’ (see p. 104 in Williams 1981). But this is not intended as an exhaustive list. Williams wants his account to be vague because he believes it is simply indeterminate what we have a reason to do. On the connection between blame and having a reason, Williams says, ‘In both the cases . . . the vagueness or indeterminacy that follows from the internalist account matches, as it seems to me, a vagueness or indeterminacy that is a genuine feature of our practice and experience’ (see p. 43 in Williams 1995). Williams would argue that a theory that gave exacting standards for practical reason would fail to capture something important about the nature of deliberation. It is open to a critic of Williams to argue that there is no such indeterminacy, but let us first set out Williams’ explanation of it. Williams argues that the indeterminacy has two sources. The first is that rational deliberation depends upon what considerations are brought to the agent’s mind. For instance, Williams thinks that imagination plays a crucial role in rational deliberation, but imagination is a creative and explorative process the outcome of which cannot be known beforehand. Williams says, ‘It is impossible that it should be fully determinate what imagination might contribute to deliberation’ (see p. 38 in Williams 1995). Call this indeterminacy in procedure. Before examining the details of this claim, let me note the second source, indeterminacy of desire, which will receive greater attention later in this paper. Desire is a source of indeterminacy according to Williams because ‘it may be indeterminate what the condition of agent’s S relevantly is’ (see p. 38 in Williams 1995). Our motivational make-up may not be completely settled. What we have a reason to do will vary depending on how these different indeterminacies get filled out. Indeterminacy in procedure concerns what might rationally influence an agent during deliberation. At one point, Williams describes some of the possibilities of rational deliberation as ‘invention of alternatives . . . breaking out of a dilemma . . . perception of unexpected similarities’ (see p. 38 in Williams 1995). These are all cases where one cannot simply be guided by a principle. Williams seems to understand deliberation as a creative process, not a procedure in which an agent transforms desires by applying principles to initial motivations. Rather an agent casts about in his imagination testing out possible solutions to the problems as they present themselves. On such an account, it would seem that agents with exactly similar motivations might hit on equally good, but incompatible solutions to the same problem. This might be because an interesting analogy occurs to one, while the other is moved by a compelling work of art. Williams seems to leave all these as open possibilities. There is any number of options and the agent just has to carry on with his deliberation until he sees where it leads him. The only requirement seems to be that the agent keeps thinking until he finds an answer that is acceptable to him. Deliberation also has a social aspect. How we see our world and how we see ourselves will be affected by how others perceive us. What considerations are brought to mind will depend upon social influences. Agents who are exposed to different

98

P. Fleming

viewpoints might find deliberation leads them in different directions. Furthermore, advice might make us rethink the situation that faces us. For instance, Williams claims that in some cases, blame consists in ‘a proleptic invocation of a reason to do or not do a certain thing, which applies in virtue of a disposition to have the respect of other people’ (see p. 41 in Williams 1995). When those we admire blame us, it might lead us to change our S in order to be admired by them. The moral responses of others might change our evaluations. Given that it is not fixed whom we will interact with and how we will respond to them, this yields another variety of indeterminacy in procedure. Of course, Kantian internalists and a whole host of reason externalists will deny that the principles of rationality are so imprecise. To adjudicate this dispute, we would have to look at particular proposals for practical reason. A rationalist would need to show her account withstood critical attack in order to refute Williams’ skepticism. Depending on the particulars, this may not be a threat to internalism.4 However, it would be a threat to the indeterminacy of the deliberative principles that Williams takes to support internalism. I will put this issue to one side. Instead, I will focus on a different problem in practical reason. Most rationalist accounts attempt to introduce formal constraints on desires to yield an account of reasons. I think such accounts run afoul of the second sort of indeterminacy that Williams points to in his discussions of internalism. Agents’ desires can be indeterminate and, as such, do not have the proper content to be manipulated by formal principles. To introduce this problem, I will need to briefly discuss Michael Smith’s account of reasons in his book The Moral Problem.

5.3 Smith on ‘Our Fully Rational Selves’ Smith accepts the internalist requirement on reasons, but takes issue with Williams’ account of deliberation. As we have seen, imagination plays a key role in Williams’ account. Smith grants that imagination might play a role, but claims ‘by far the most important way in which we create new and destroy old underived desires when we deliberate is by trying to find out whether our desires are systematically justifiable’ (see pp. 158–9 in Smith 1994).5 The claim seems to be that a primary task of practical reason is to see whether particular desires are themselves desirable. We can accomplish this by employing a method of systematic justification. This involves constructing what our fully rational selves would tell us to do in the actual circumstances we are in. For instance, our fully rational selves will not think we should be guided by fleeting desires that are the product of fear, frustration, or anger. More importantly, when deliberating about a particular desire we should try to ‘integrate the object of that desire into a more coherent and unified desiderative profile and evaluative outlook’ (see p. 159 in Smith 1994). This process of striving for coherence and unity will leave us with desires that are likely to be rationally superior to those we start with in deliberation.

5 The Indeterminacy of Desire and Practical Reason

99

It is rational to acquire new desires on this model if by adding them to our desire set, it becomes more systematically justifiable. A new desire makes the set more systematically justifiable if it justifies and explains some other desire. In this way, it would make our subjective motivational set more coherent and unified. For example, take an agent who desires that her children get a good education. Imagine this desire does not stem from any other desire the agent has; it is unconnected to other elements in her web of desires. It would have more rational support if it were connected to other desires the agent has, like her desire that her children live well. Furthermore, the desire would have greater rational support if it fit with a more general desire. The agent might see that her specific desire would cohere better if she adopted a general desire that all children get a good education. Smith claims that in adopting a more general desire, an agent might realize that her new imagined desire set is better ‘for we may properly regard the unity of a set of desires as a virtue, a virtue that in turn makes for the rationality of the set as a whole’ (see p. 159 in Smith 1994). If the belief that the new set of imagined desires is rationally superior to her present set leads her to change her desires, she has done so through a rational process. By an analogous process, one can come to give up old desires. If desires must be systematically justifiable and coherence is a key to this process, then some desires may be rationally given up when it is believed they do not cohere well with others. Imagine our above agent has a desire that none of her tax money is spent on education outside her child’s school district. Upon reflection, she sees that this desire does not cohere well with other desires. When such a desire has no connection with the rest of her desires, it might seem ad hoc and not justifiable. Or the desire might be in conflict with some other desire she has. If she changes her desires for either of these reasons, it is a change by a rational process. There is much to admire in Smith’s account. Nevertheless, there are problems. An agent might desire to have sexual relations only with members of the opposite sex. This seems to be just a brute desire that needs no justification. I suppose an agent might adopt the more general desire that all agents only have sexual relations with the opposite sex in order to explain and justify his original desire. Some people do seem to reason this way, but I think it is misguided. Problems along these lines echo Kant’s problem regarding how we formulate the maxims for the test of universalizability. The moral rationalist needs to specify how to get the content of the proposition for her particular test of rationality. Otherwise, the formal constraints do no good. More importantly, it is unclear when coherence is a virtue and when it is a vice. Sometimes an evaluative outlook that is perfectly coherent and unified signifies an insufficient grasp of what life has to offer rather than a model of rationality. As David Wiggins remarks concerning the considerations of practical reason, ‘It is of the essence of these concerns to make competing and inconsistent claims. This is a mark not of human irrationality but of human rationality in the face of the plurality of human goods’ (see p. 145 in Wiggins 1978). It seems unavoidable that a complete absence of tension would be irrational in some cases if one accepts that this is a world of plural and conflicting values. I do not have a knockdown argument for the existence of plural and conflicting values, but it does ring true. It does not seem

100

P. Fleming

that there is anything irrational about being drawn to both work and play at the same time. Being solely concerned with work would make one a bit of a bore and probably result in a less enjoyable life. So it seems in some cases the existence of tension in an agent’s motivational makeup signifies not a lack of rationality, but an appropriate sensitivity to what life has to offer. Smith is only offering a summary style analysis of ‘fully rational’, so perhaps other platitudes about practical reason will inform us in what way our desire sets ought to be coherent and unified. I do not wish to try to settle the question whether those principles are rationally binding in an informative way. What I think is important is that these principles, even if they are right, do not do anywhere near all the work needed in practical reason. On Smith’s account, our reasons are what our fully rational selves would advise us to do. I will argue that, because of the indeterminacy of desire, there are different fully rational selves we might become. This will be the case when agents are undecided as to what to pursue in life. In such circumstances, agents will be asking, if only implicitly, what type of person should I be? Since we are asking what makes an agent ideal, this cannot be answered by asking what our fully rational selves would advise us to do. This is just the same question over again. We need to face the fact that we do not as yet know what our fully rational selves are like.

5.4 The Indeterminacy of Desire Before arguing for the indeterminacy of desire, let me to say a few words about the nature of desire. I use the term ’desire’ to cover all pro-attitudes. In some contexts, this would be unwise because there are important differences between ‘hoping that X’ and ‘desiring that X’. However, I am not here giving a fine-grained analysis of all mental states. Rather I aim to establish one possible feature of the set of intentional states that must accompany belief to produce action. I do not think this begs any substantive question in philosophical psychology. One could accept this claim and still be an externalist about moral motivation or reasons.6 If we can adopt desires simply in virtue of an evaluative belief, then it is even compatible with rejecting a Humean theory of motivation.7 All I am maintaining is that in deliberation, an agent must have some motivational states to reason from to get to others. On this account, desires are intentional states with a world to mind direction of fit. This is not to offer direction of fit as a definition or a complete explanation of desire. Rather, it is one central aspect of desire that any theory of the mind must account for. Smith has argued persuasively for why we should reject a phenomenological account of desire and instead adopt a direction of fit account (see pp. 92–129 in Smith 1994).8 If we claim that desires have an essential phenomenological component, then the epistemology of desire becomes analogous to the epistemology of sensation. Sensations have an essential phenomenological aspect, which seems to make it impossible to be mistaken about the content of sensation. An agent is sensing red only if he believes he is having what we would call a red-like experience. We

5 The Indeterminacy of Desire and Practical Reason

101

are infallible in the epistemology of our own sensation. Such infallibility is wildly implausible in the case of desire. Agents are often mistaken about what they desire. What we care about or what we most want can be deeply mysterious to us. In place of a phenomenological account, we need a more complicated account of desire that includes direction of fit. Desires have world to mind direction of fit, while beliefs have mind to world direction of fit. Beliefs represent the world. Desires offer conceptions of how we might like the world to be. A desire dictates that some aspect of the world should be made to fit what the mind conceives. A mental state that is a belief is supposed to fit the world. As such, beliefs and desires have very different relationships with other mental states. Suppose that you desire P. Now even if you believe not P is the case, that by itself will not necessarily diminish your desire for P. Now suppose you believe P. Clearly, you cannot maintain that belief if you come to believe not P. As such, desires and beliefs function differently in our mental life. This is not an entirely satisfactory account of the difference between beliefs and desires, because we have used beliefs to explain desires. No matter, all we want in a theory of practical reason is some core account of desire. I take it that any theory of the mind would need to make sense of this feature of desire, which suggests we have a workable and somewhat uncontroversial account of desire. Desire as direction of fit is best understood as a disposition. Other things being equal, a desire to excel at philosophy disposes one to carefully study the work of great philosophers. This is far from an illuminating account of that desire and much more would need to be said about the functional role desires play. The epistemology of desire now becomes an epistemology of dispositions, which will be a process of assessing the truth of counterfactuals. In claiming that desires are indeterminate, I mean that there is no fact of the matter as to what the agent desires or most desires. An agent’s motivational makeup may have elements whose content is unsettled or lacks determinate weight. Shortly, I will turn to some examples to help illustrate the indeterminacy of desire. Now though, I would like to point out the indeterminacy of desire is entailed by the conjunction of our dispositional analysis of desire and one very plausible assumption about counterfactuals. Having a desire is being disposed to make the world fit that desire in appropriate circumstances. Specifying the exact functional role of a desire will be extremely difficult and depend on the truth of various counterfactuals, but we can imagine how such a process can work out. Suppose I say ‘Sally desires Sue’s companionship’. Now if Sally flees the room every time Sue enters it, it seems likely that my claim is false.9 Smith thinks this account also sheds light on how desires can have propositional content because ‘the propositional content of a desire may then simply be determined by its functional role’ (see p. 119 in Smith 1994). So, it seems that whether one has a desire and the particular content of that desire depends upon the way one behaves and the truth of some counterfactuals. Our plausible assumption about counterfactuals that entails indeterminacy is that some counterfactuals are neither true nor false. Making this assumption commits us to denying conditional excluded middle. Fortunately, David Lewis has already provided a way for us to understand this. He has argued that the truth of a counterfactual depends upon the truth of that statement in the nearest possible world

102

P. Fleming

where the antecedent is true (Lewis 1973).10 For some counterfactuals, there may be a tie for the nearest possible world. If these worlds give conflicting answers, then the counterfactual is neither true nor false. To borrow an example from Lewis, the counterfactual ‘if Bizet and Verdi were compatriots, Bizet would be Italian’ is neither true nor false. There are worlds equally similar to this, one in which Bizet and Verdi were Italian compatriots and ones in which Bizet and Verdi were non-Italian compatriots. In the case of practical choices when an agent wonders do I prefer A to B, it may simply be indeterminate what the answer is. Borrowing from Smith’s model, we might say there is more than one possible world that contains a fully rational counterpart of ourselves. Sometimes these fully rational counterparts might disagree on what is to be done. As such, there are constitutive elements of ourselves that need to be settled to decide which of these idealized individuals we should become.

5.5 An Argument from the Phenomenology of Deliberation So far, I have offered a model for how desires might be indeterminate. This model depends upon some plausible, but by no means certain, accounts of desire and counterfactuals. I could offer a further defense of those positions, but looking at our practical experience makes a stronger case for the indeterminacy of desire. I offer the indeterminacy of desire as the best explanation of some very hard cases of deliberation. Imagine a man has been cheating on his wife, but has resolved not to see his lover anymore.11 However, after some time he gives in to his passions and goes to see her. Is this action rational or not? I am inclined to think that we have not been given sufficient information to answer this question. Whether it is rational seems to depend in part on whom he ultimately ends up with. If his passion for his new lover fades quickly, it may be that he should have seen that this was just an infatuation. However, if he leaves his wife and lives happily ever after, then perhaps he was acting reasonably. Now, this does not mean that we have to know the future in order to have a reason for action. One can certainly act with reason but still have things turn out poorly because of the unpredictable element in life. However, I do not think that the agent’s deliberation is simply a cost-benefit analysis of likely consequences. It is not as if when agents make such decisions they notice some factor they forget to add to their list of pros and cons. Instead, the agent must settle on whom it is he truly loves. In many cases, this will not be a process of discovery by introspection. While the agent deliberates, there may be no psychological state that determines whether this action was rational or not. The desires of the agent may very well be indeterminate at the moment at hand. That is likely what makes such decisions so hard and troubling. One might protest that cases like this, while familiar, are better explained epistemologically.12 In such cases, it may be that the man cannot for some reason see that his wife is unhappy and that this will eventually make him profoundly unhappy.

5 The Indeterminacy of Desire and Practical Reason

103

Years from now, he might be able to look back and see this. I will gladly grant that these sorts of cases exist, but I see no reason to think all cases must fit this model. A man may be torn between two loves and know that he must choose one if he is to have either. To insist there is always a fact of the matter about whom a person truly loves or ought to end up with seems overly romantic. In fact, retrospective judgements about these cases might be rationalizations so that we can better make sense of our lives. There is likely distinct psychological pressure to makes sense of these choices later in life. So, such judgements might not be informed hindsight but rather an inaccurate and self-serving recollection. Others critics might deny that such cases are possible. One might claim that if what I say is true, then we should expect to find individuals paralyzed as a result of deliberation. If desires were indeterminate in the way I suggest, then wouldn’t we all be like Buridan’s ass that starved to death because he had no reason to choose between two equally attractive and available stacks of hay? However, the objection continues, this is not what we find. Individuals do decide to either stay with their lover or not. If there is no fact of the matter as to what they prefer, then they would never choose. As a response, let me first say that I think there are cases where agents are paralyzed in deliberation. Even after extensive weighing of options, many of us have probably thought ‘I simply do not know what to do’. Yet we eventually decided to do something and I certainly could not disagree with that. All I am arguing is that we do not uncover some desire that we have for some reason forgotten to factor into our earlier calculation. Instead, agents faced with this problem engage in a different kind of reasoning.13 One hypothesis is that when we act rationally in such dilemmas, we resolve to be a certain kind of agent. Most of us are probably familiar with agents who do this either wisely or foolishly. This suggests to me that there is some standard to be met.

5.6 Two Kinds, Two Levels, and Two More Examples Here is a brief catalogue of some different kinds of indeterminacy. There can be two kinds of indeterminacy: one stemming from the content of the desire, the other from the weight attached to the desire. It will be helpful to separate these in theory even if we cannot in practice. In some cases, one may have a desire without that desire having entirely specific propositional content. Desire might be somewhat fuzzy as it were, requiring that we specify what we want. The other sort of indeterminacy stems from the relative weight of a desire. For any number of options, it may be indeterminate what one would be willing to forgo for that desire. At one end of the scale, this will concern whether the agent has any desire whatsoever for the object, while at the other end, there are desires one is willing to sacrifice anything for. Indeterminacy might also arise at two different levels. Indeterminacy might simply be the function of one desire or it might arise at a step above, from the interaction between more than one desire. Again these may be very hard to isolate in real life, but I think it is worthwhile to keep these in mind in a philosophical discussion.

104

P. Fleming

I will illustrate these kinds of indeterminacy with examples that I hope are familiar enough to serve as more data to be explained by indeterminacy. I take it that these examples and other analogous cases are common place elements of practical life. Consider the case of the indeterminacy of a single desire first. When I first began graduate school, I had a desire to be a philosopher but that desire was indeterminate. Now this might plausibly be subsumed under the heading of means-ends reasoning, where I see if my desire to become a philosopher outweighs my aversion to trying to clearly state the arguments for and against the Myth of the Given. However, I think it a little odd to conceive the desire to think clearly as somehow distinct from the desire to be a philosopher. Rather, such capacities are partly constitutive of being a philosopher. More appropriately, the end of being a philosopher includes a set of different abilities and part of my graduate education was coming to see whether I had the necessary desires to develop some essential subset of those abilities. Does that mean when someone asked me in my first year of graduate school whether I desired to be a philosopher I should have responded ‘I am still trying to figure that out’? Well, that response would not have been out of place, but neither is a straightforward ‘Yes’. After all, I would not have given up a good job and moved half way across the country for something I wasn’t sure I wanted. I did have the desire but it was indeterminate. On some occasions, an agent can have a desire without that desire having entirely specific content. Sometimes we have distinct cravings, but we also have desires that are broad and diffuse. One might protest that indeterminacy will be resolved if we focus on the world and what has value. I agree that indeterminacy can be settled, but not always by simply examining what has value. That is because we are faced with a world of plural and conflicting values, which can be a source of indeterminacy that arises at the level of a desire set as a whole. Imagine an undergraduate student with amazing philosophical ability. Say we are fairly confident, although there are no guarantees, that with the right education, he could produce work on par with that of David Lewis. However, this young man is also the star of the basketball team and has a lucrative future in professional sports. In all likelihood, if he does not continue his education now, he will never reach his intellectual potential. What does the young man have most reason to do? I would argue that both of these forms of life are worthwhile undertakings. In assessing his options, he will want to factor in his present desires, which certainly are not irrelevant. But he also needs to reason about what kinds of desire he might wish to cultivate. His deliberation should involve trying to understand what type of life will be the most fulfilling, where this includes coming to understand what he means by the word ‘fulfilling’. The young man has to decide what type of individual he wants to be. This might seem like a unique case that we shouldn’t place too much emphasis on. However, I claim only that the level of ability is unique. Most of us possess a set of skills that could be utilized in very different kinds of life. So I think the problem is perfectly general. There are also parallel problems in tragic situations were one has to choose between two evil or equally undesirable options. Related problems arise for issues more limited in scope. Perhaps we have not settled what it means to have a desirable relationship or an enjoyable retirement.

5 The Indeterminacy of Desire and Practical Reason

105

5.7 The Implications of Indeterminacy What exactly does this show? It at least shows that the question of the rational justification of morality is not the only important question in practical reason. Practical reason concerns far more than the constraints of morality. For most of us, wanton cruelty or absolute moral indifference is not on the deliberative map. Yet we are still faced with practical questions that are not settled by instrumental or prudential reasoning. Moral reasons, instrumental reasons, and prudential reasons do not exhaust the reasons we have for action. We also seek reasons for adopting some particular projects and conception of the good life. The indeterminacy of desire shows that no set of formal principles will answer all the questions that are of interest in practical reason. The content of our motivational states at the moment of deliberation may be too impoverished for formal principles to do the all work that is needed. One might think that Korsgaard has the tools to get out of this problem. Korsgaard argues that the structure of self-consciousness forces us to act and deliberate under a self-conception she calls practical identity. She says this self-conception is ‘a description under which you value yourself, a description under which you find your life to be worth living and your actions to be worth undertaking’ (see p. 101 in Korsgaard 1996). Practical identities will include things like being someone’s child, a philosopher, a member of a religious group, and many other things. One’s reasons for action will, in part, be determined by your practical identity. That is because a practical identity partly determines the content of one’s deliberation. It structures the concerns and desires that present themselves to the agent. As such, practical identity brings content to the formal principles Korsgaard employs in practical reason. Therefore, it might be a way to answer the questions I have been raising. Korsgaard defends a broadly Kantian position. When agents deliberate, the reflective distance created by self-consciousness forces them to examine whether their impulses are reasons for action. While deliberating, the agent must side with some particular consideration that presents itself as a motivation. The agent must decide if the maxim proposed by some aspect of a practical identity can be willed in accord with the Kantian imperatives. If they can, then these maxims can be laws the agent acts on. They can be willed in the universal sense. If they cannot fit the Kantian imperatives, then we have an obligation not to perform them. These formal principles separate maxims into the permissible and the impermissible. It appears one’s practical identity determines what one ought to do in the sphere of the permissible. Again, let us put aside the question regarding whether the formal principles Korsgaard puts forth are established by her argument. Instead, I wish to argue that practical identities cannot do all the work required to solve the problem I am raising. It may be that practical identities are indeterminate in just the way desires are. In that case, we have made no progress at all. If they are not indeterminate, it seems to me that they are conceptually downstream from the problems I am addressing. To see this, consider how we should understand Korsgaard’s practical identities.14 Are they simply given or are they the product of our conscious choices? Korsgaard writes that, ‘Our contingent practical identities are, to some extent, given to us – by our cultures, by our societies and their role structures, by the accidents of birth, and

106

P. Fleming

by our natural abilities – but it is also clear that we enter into their constitution’ (see p. 239 in Korsgaard 1996). Korsgaard has to say this given her insistence that the structure of human consciousness can create reflective distance from any particular concern. So it would seem we have to stand back from our practical identities and decide to adopt or create them. Here the problem under discussion crops up. Most of us are faced with choices between different practical identities that fall within the bounds of moral permissibility. But that does not mean that nothing important hinges on the decision. Choosing a career or a mate is a decision that is important and perplexing. An account of practical reason should have something to say about it. One cannot sidestep this problem with an appeal to practical identity. Often, agents will need to decide what practical identity to adopt. They may be wondering whether to be a monk or a banker or an artist. To reply by saying they should be guided by the obligations that are constitutive of their practical identity is unhelpful. It assumes the question is answered; it does not answer it. Our motivational makeup does not always provide sufficient data to determine what we have reason to do. It is not clear to me how common this is. I imagine it is fairly infrequent, but I do not think that diminishes its importance. Sometimes the content of one’s desires is very precise, such as when something sharp and painful is digging into one’s foot. Naturally, one removes the cause of the pain. However, the action that results hardly seems the product of deliberation. Indeed, I think the more specific our desire set, the less need for deliberation. Harry Frankfurt reminds us that deliberation is in an important sense de-liberation (see p. 175 in Frankfurt 1988). On my preferred account, it is an activity where some measure of freedom is lost because some alternatives become ruled out. It is an attempt to rationally constrain our choice. It is most difficult, where we do not know what considerations properly constrain our judgements. Which suggests to me that the cases I have discussed is where deliberation is most important. A more adequate conception of practical reason would address the constitutive question about what kind of agent to be.15

Notes 1 Williams first makes the argument in ‘Internal and External Reasons’ (Williams 1981). He further discusses it in ‘Internal Reasons and the Obscurity of Blame’ (Williams 1995). His final contribution on the subject appears to be ‘Some Further Notes on Internal and External Reasons’ (Williams 2001). 2 T.M. Scanlon does an admirable job clarifying this issue in his appendix on Williams (Scanlon 1998). 3 Also see Williams’ discussion of ‘broadly Kantian’ approaches (see pp. 93–94 in Williams 2001). 4 If Korsgaard or Smith’s position were to withstand skeptical attack, then that would yield a variety of internalism. 5 By ‘underived’ desires, Smith means desires that are not derived from other desires. 6 For instance, it seems compatible with Zangwill (2003) and Gibbard (2003). 7 See Wallace (1990) for an account of how this position might be worked out. 8 Smith sometimes seems to take direction of fit as an exhaustive account of desire. As I stated earlier, I take it only to be one element of the nature of desire. 9 Of course, given the complexity of human psychology, this isn’t always definitive evidence. One can imagine other beliefs and desires that might still make my ascription plausible.

5 The Indeterminacy of Desire and Practical Reason

107

10

See pp. 79–83 for discussion of the denial of conditional excluded middle. A similar example is in pp. 44–46 (Williams 1993). Williams uses the example in an attempt to show that akrasia is an ethical concept, not a psychological one. 12 I thank Paul Bloomfield for helping me see the force of this point. 13 Michael Bratman discusses the importance of Buridan cases to practical reason (see pp. 22–23 in Bratman 1987). 14 Thomas Nagel raises a similar point in his reply to Korsgaard (see p. 204 in Korsgaard 1996). Also note that when I discuss practical identity, I mean it in the sense of a rich self-conception. Korsgaard goes on to argue that we all have the practical identity of being human or a citizen in the kingdom of ends. I take that to be too devoid of content to handle the problems under discussion. 15 I would like to thank David K. Chan, Joel Kupperman, and John Troyer for written comments on earlier drafts of this paper. I would also like to thank an audience at the University of Connecticut and the participants at the conference on Values, Rational Choice, and the Will for helpful suggestions on this topic. 11

References Bratman M (1987) Intention, plans, and practical reason. Harvard University Press, Cambridge MA Frankfurt H (1988) The importance of what we care about. Cambridge University Press, New York Gibbard A (2003) Reasons and thin and thick. Journal of Philosophy 100(6):288–304 Korsgaard C (1986) Skepticism about practical reason. Journal of Philosophy 83:5–25 Korsgaard C (1996) The sources of normativity. Cambridge University Press, Cambridge Lewis D (1973) Counterfactuals. Blackwell, Oxford Scanlon TM (1998) What we owe to each other. Harvard University Press, Cambridge MA Smith M (1994) The moral problem. Blackwell, Oxford Wallace RJ (1990) How to argue about practical reason. Mind 99:355–385 Wiggins D (1978) Deliberation and practical reason. In: Raz J (ed) Practical Reasoning. Oxford University Press, Oxford Williams B (1981) Moral luck. Cambridge University Press, Cambridge Williams B (1993) Shame and necessity. University of California Press, Berkeley Williams B (1995) Making sense of humanity. Cambridge University Press, Cambridge Williams B (2001) Some further notes on internal and external reasons. In: Millgram E (ed) Varieties of practical reasoning. MIT Press, Cambridge MA Zangwill N (2003) Moral motivational externalism. American Philosophical Quarterly 40(2): 143–154

Chapter 6

The Myth of Objectively Alien Desires James Stacey Taylor

Abstract Since Harry Frankfurt’s seminal 1971 paper ‘Freedom of the Will and the Concept of a Person,’ philosophers of mind and action have been concerned to distinguish those desires that a person identifies with from those that are alien to her. Frankfurt’s original attempt to distinguish desires that a person identified with from those that were alien to her was a purely attitudinal account of identification. This first attempt by Frankfurt has been heavily criticized, with the most pressing criticism being the Regress-cum-Incompleteness objection. Owing to the force of this criticism, those concerned with providing an account of identification turned to developing structurally-based accounts of identification that focus on the structural relationships that such desires hold to other elements of her mental economy. In this paper, I will argue that this move is mistaken. However, my aim in this paper is not merely the negative one of seeking to show that structural analyses of identification are likely to fail. If I am right that none of the most prominent objective criteria succeeds in distinguishing those desires that a person identifies with from those that she is alienated from, then an adequate analysis of identification must be attitudinally based. I sketch what one such account will look like at the close of this paper. Keywords Alien · Autonomy · Decision · Desire · Identification Since Harry Frankfurt’s seminal 1971 paper ‘Freedom of the Will and the Concept of a Person,’ philosophers of mind and action have been concerned to distinguish those desires that a person identifies with (that she is active with respect to, that are internal to her), from those that are alien to her (that she is passive with respect to, that are external to her) (Frankfurt 1988a).1 Frankfurt’s original attempt to distinguish desires that a person identified with from those that were alien to her was a purely attitudinal account of identification. That is, Frankfurt took the agent’s attitude towards those of her desires that were in question to be the sole criterion J.S. Taylor Department of Philosophy & Religion, Bliss Hall 103, The College of New Jersey, P.O. Box 7718, 2000 Pennington Rd., Ewing, NJ 08628, USA, Fax: 1(609)637-5167 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

109

110

J.S. Taylor

by which these two classes of desires could be differentiated from each other. For Frankfurt, a person identified with his effective first-order desires (i.e., those desires that moved him to act) if he volitionally endorsed them; that is, if he both wanted to have the desire in question, and also wanted it to move him to act (see p. 15 in Frankfurt 1988a). This first attempt by Frankfurt to distinguish between those desires that a person identified with and those that were alien to him has been heavily criticized.2 The most pressing of the criticisms that have been leveled against this original attempt by Frankfurt to distinguish between these two classes of desires has been the Regress-cum-Incompleteness objection. The proponents of this objection (which was first developed in Watson 1975) question how it is that a person identifies with his second-order desires (i.e., his desires for those of his first-order desires that he is said to identify with on this view). If the answer to this question is that the person identifies with these desires as he volitionally endorses them (i.e., he both wants to have them at a third-order level of desires, and wants them to move him to act), then the question is simply pushed back to the next level (how is it that the person in question identifies with his third-order desires?) and a regress ensues. If, however, the answer to this question is that the person identifies with his second-order desires by some other means, then Frankfurt’s original account of identification is incomplete. It is now widely accepted that the Regress-cum-Incompleteness objection shows that no purely attitudinal analysis of identification can work.3 It seems that it will always be possible to ask why it is that the person concerned identifies with the endorsing attitude that such an analysis will use to distinguish the desires that he identifies with from those that he is alienated from. Once this question is posed, it seems that it must be admitted that such an analysis is either subject to a problematic regress, or else is incomplete. Since this is so, persons who are concerned with providing an account of how to distinguish those desires that a person identifies with from those that he is alienated from have turned to developing structurallybased accounts of what it is for a person to identify with her desires. That is, they have turned to developing accounts of identification that focus not on the person’s attitudes towards those of her desires that are in question, but on the structural relationships that such desires hold to other elements of her mental economy. Such structurally-based analyses of identification rest on the view that the class of desires that a person identifies with can be distinguished from the class of desires that she is alienated from without appealing solely to the person’s own attitudes towards the desires in question. Structurally-based analyses of identification, then, are based on the assumption that the desires that a person is alienated from can be distinguished from those that she identifies with by appeal to objective (i.e., non-attitudinal, nonsubjective) criteria. (Note that this assumption is not that no attitudinal component could feature in a correct account of identification. Rather, it is that no attitudinal component could alone provide sufficient conditions for a person to identify with his effective first-order desire.) Given the difficulties that are faced by attitudinal analyses of identification, this seems to be a reasonable assumption to make. However, as I will attempt to show in this paper, this assumption is mistaken. This is so, I will argue, because none of the standard ways of objectively distinguishing

6 The Myth of Objectively Alien Desires

111

between those of a person’s effective first-order desires that she is identified with and those of her effective first-order desires that she is alienated from achieve this. That is, it is not possible, as the proponents of structural analyses of identification assume, to use objective criteria to distinguish the class of desires that a person identifies with from the class of desires that she is alienated from. My strategy in this paper will be a simple one. To show that the class of desires that a person is alienated from cannot be distinguished from the class of desires that she identifies with, I will outline the three most prominent objective criteria that are given as hallmarks of alienation and show how none of them succeed in distinguishing these two classes of desires. However, my aim in this paper is not merely the negative one of seeking to show that structural analyses of identification are likely to fail. If I am right to claim that none of the most prominent objective criteria succeeds in distinguishing those desires that a person identifies with from those that she is alienated from, an adequate analysis of identification will be one that must be attitudinally based. I sketch what one such account will look like at the close of this paper.

6.1 Some Background Clarifications Before moving to argue that structural, objective criteria cannot readily be used to identify which of a person’s desires are alien to her, some terminological clarification is needed. An attitudinal analysis of identification is an analysis on which it is sufficient for a person to identify with an effective first-order desire that she adopt a particular attitude towards it, or else that she fail to adopt a particular attitude towards it. Frankfurt’s original hierarchical analysis of identification is clearly an attitudinal analysis of this sort, for the question of whether a person identifies with her effective first-order desires is answered solely by appeal to her attitudes towards them. Frankfurt’s more recent analysis of identification as satisfaction is also such an attitudinal analysis (see pp. 105–106 in Frankfurt 1999b). On this analysis a person identifies with her effective first-order desire if she reflectively endorses it (as on his original hierarchical analysis) and is also satisfied with this endorsement. For Frankfurt, a person’s being satisfied with his effective first-order desires ‘does not require that . . . [he] . . . have any particular belief about it, or any particular feeling or attitude or intention . . . There is nothing that he needs to think, or adopt, or accept; it is not necessary for him to do anything at all’. Instead, his being satisfied simply consists in his ‘having no interest in making changes’ in his motivational economy (see pp. 104–105 in Frankfurt 1999b). By contrast with such attitudinal analyses, a structural analysis of identification holds that a person’s holding a certain attitude (or lack thereof) towards his effective first-order desires is not sufficient for him to identify with it. Instead, on such an account, a person’s effective first-order desire must meet certain objective criteria for her to identify with it. An example of such a structural approach is Laura Waddell Ekstrom’s coherentist analysis of what it is for a person to identify with her preferences. For Ekstrom, a preference ‘is a very particular sort of desire: it is one (i) for a certain first-order

112

J.S. Taylor

desire to be effective in action, when or if one acts, and (ii) that is formed in the search for what is good’ (see p. 603 in Ekstrom 1993).4 Ekstrom distinguishes between a person’s ‘self’ and her ‘true or most central self’. For Ekstrom, a person’s ‘self’ consists of her character together with the power for ‘fashioning and refashioning’ that character, where a person S’s character at time t is constituted by ‘the set of propositions that S accepts at t and the preferences of S at t’ (see p. 606 in Ekstrom 1993). A person’s ‘true or most central self,’ however, consists of that subset of these acceptances and preferences that actually cohere together. With this distinction in place, Ekstrom argues that a person identifies with her preferences when ‘they cohere with . . . [her] . . . other preferences and acceptances,’ and so can be recognized as members of her true self.5 Ekstrom’s structural analysis of identification is thus a purely objective one. However, a structural analysis of identification need not preclude the possibility that for a person to identify with her effective first-order desires (or her preferences), she must adopt (or fail to adopt) a particular attitude towards them. All that such an analysis of identification requires for it to be a structural analysis is that the adoption of (or failure to adopt) such an attitude is not sufficient for the person concerned to identify with her effective first-order desires. Michael Bratman, for example, has developed a structural analysis of identification on which a person’s effective first-order desire must be compatible with her standing decisions or policies concerning what to treat as reason-giving (Bratman 1996).6 For Bratman, what counts as a person’s standing decisions, intentions and policies is given by his broadly Lockean account of personal identity, on which an agent helps ‘ensure appropriate psychological continuities and connections [to retain her identity over time] by sticking with and executing . . . [her] . . . prior plans and policies, and by monitoring and regulating . . . [her] . . . motivational structures in favor, say, of . . . [her] . . . continued commitment to philosophy’ (see p. 43 in Bratman 2000). Given these requirements, Bratman’s account of identification is clearly a structural account of identification as this description is used here. However, this does not preclude him from introducing an attitudinal component into his account, for he argues that a person will only identify with her effective first-order desire if decides to treat it as being reason giving (in the sense of being end setting) in the relevant circumstances (see p. 9 in Bratman 1996).7 With these terminological clarifications in place, how analyses of identification are held to recognize alien desires should also be clarified prior to arguing that structural, objective, criteria cannot readily be used to identify which of a person’s desires are alien to her. The first and simplest point to make here is that if an analysis of identification commits its proponents to claiming that a (clearly) alien effective firstorder desire is one that its possessor identifies with, this will serve as a counterexample to the analysis in question. Recognizing this, many proponents of structural, objective analysis of identification have attempted to provide counterexamples along these lines to the attitudinal analyses of identification. Such counterexamples have been implicitly based on the view that unless certain structural, objective criteria are met by a person’s effective first-order desire, then that desire will not be one that the person in question identifies with. As I will argue below, however, such

6 The Myth of Objectively Alien Desires

113

counterexamples fail, for it is not possible to use such objective, structural criteria to identify those effective first-order desires that a person is alienated from. The second point to make in this context is that the ability to identify those effective first-order desires that a person is alienated from is a necessary but not a sufficient condition for an account of identification. This is because an account of alienation (i.e., an account that can show which of her effective first-order desires a person is alienated from) is not also necessarily an account of identification. (Although, it is true that an account of alienation is necessary for one to have an account of identification.) This is so for two reasons. First, one needs to provide a positive account of why a person is held to identify with those of her effective first-order desires that she is not alienated from. If this is not done, then the proffered account of identification will not advance one’s understanding of what it is for a person to identify with her effective first-order desires over and above the truism that she is not alienated from them. Second, it cannot simply be assumed that an effective first-order desire that a person is not alienated from is one that she identifies with, such that if she acts on it she will be acting of her own free will. It is possible that the taxonomy of desires is richer than that which this assumption takes to be true. That is, rather than having merely a binary taxonomy of desire, such that a person either identifies with an effective first-order desire (i.e., acts of her own free will) or else is alienated from it (i.e., does not act of her own free will), it is possible that there is a tripartite taxonomy of desire. Thus, it is possible that a person could be alienated from a desire (i.e., if she acts on it she does not act of her own free will), she identifies with it qua person (i.e., if she acts on it she acts of her own free will), or she identifies with it in some other way.8 She could, for example, identify with it qua agent, where if she acts on it she does not act of her own free will, as small children and animals do not act of their own free will but are not necessarily alienated from their effective first-order desires. For these two reasons then, even if the structural accounts of desire alienation that I will outline below are correct, they will not suffice as analyses of identification without theoretical supplement. Finally, my strategy in this paper must also be clarified. In the next section, I will outline the three most prominent ways in which certain proponents of structural analyses of identification have sought objectively to distinguish those effective firstorder desires that a person identifies with from those that she is alienated from. I will argue that since none of these ways of identifying alien desires works for all cases, we have reason to doubt that the structural analyses of identification will succeed where their attitudinal rivals fail. A proponent of a structural analysis of identification might, however, be suspicious that I am implicitly relying on the attitudes of the agents in my examples to show that the desires in question are not ones that they are alienated from. Of course, this might not be a problem for my strategy in this paper, since were the proponent of the structural analysis of alienation in question to insist that I ignore such attitudes in my responses to him, he would simply be begging the question in favor of his own view. Yet given that I have explicitly acknowledged the force of the Regress-cum-Incompleteness objection against attitudinal analyses of identification, implicitly drawing on the agent’s attitude towards the desires that are in question to respond to the criticisms of the proponents of the objective,

114

J.S. Taylor

structural analyses of alienation do seem illegitimate. Despite appearances, however, this is not so. Once I have argued that the structural, objective approach to analyzing identification has unrecognized problems (i.e., its attempts to identify paradigmatically alien desires using objective criteria are not as successful as its proponents believe), I will suggest an alternative attitudinal approach to analyzing identification that avoids the Regress-cum-Incompleteness problem that beset its attitudinal ancestors. And if this approach can avoid the Regress-cum-Incompleteness problem (and so could be a viable attitudinal analysis of identification), then implicitly drawing on attitudinal intuitions to object to structural, objective accounts of alienation will not be illegitimate at all.

6.2 Candidates for Objectively External Desires With the above clarifications in place, I will outline in this section the most prominent ways in which the proponents of structural accounts of identification have sought objectively to distinguish those desires that a person identifies with from those that she is alienated from. For each of these ways in which persons have sought to distinguish these two classes of desires, I will show that, although it is initially plausible, it cannot be used definitively to distinguish them—and thus cannot be used as the basis for an objective, structural analysis of identification.

6.2.1 Irresistible Desires It is natural to assume that if a desire is strongly irresistible (what Ishtiyaque Haji has termed a ‘defeating desire’) such that whenever the person who possesses it experiences its prompting, he will effectively be moved to perform (or attempt to perform) the act that would satisfy it, then this desire would be one that the person who is moved to act by it is alienated from it even if he endorsed his so acting (see p. 209 in Haji 1998). Mark Leon, for example, has recently argued against Frankfurt’s view that a drug addict who volitionally endorsed his first-order desire for the drug to which he was addicted would identify with this desire, and so would act freely and of his own free will when he took it (Leon 2001).9 This is because, Leon argues, if the addict’s desire for the drug that he was addicted to was truly irresistible, when he was beset by it he would be passive with respect to its operation through him. He would be, claims Leon, merely a ‘passive bystander’ to this desire, since nothing that he could do would alter how it worked. Since this is so, Leon concludes, the addict would be alienated from this desire, since he would not be active with respect to it in the way that is required. Yet despite the plausibility of the view that strong irresistibility is an objective hallmark of alien desires, it should be rejected. To see this, consider a case in which the willing addict’s desire for the drug to which he was addicted had started out as a resistible desire; one that he was active with respect to, one that he was ‘in

6 The Myth of Objectively Alien Desires

115

charge’ of. As time goes by, this desire became strongly irresistible. However, the new irresistibility of this desire has no effect on the willing addict’s behavior. His drug-taking continues to follow the same patterns that it did before, and he continues to satisfy this desire for the same reasons that he did before. Since the resistibility (or otherwise) of this desire is irrelevant to the willing addict’s view of the desire, his actions, and his reasons for action, there seems no reason to claim that the mere fact that this desire is irresistible precludes the willing addict’s identification with it. That is to say, since the change in status that this desire underwent (i.e., from being one that was resistible to one that was strongly irresistible) has no effect at all on its possessor, there seems no reason to deny that the addict continues to identify with it.

6.2.2 Uncontrolled Desires Just as it is intuitive, but mistaken, to hold that a person does not identify with his strongly irresistible effective first-order desires, so too is it intuitive to think that a person would not identify with his effective first-order desires, that they are not attributable to him as a person, if he had no control over the process by which he came to possess them. This intuition lies behind many manipulation-based counterexamples that are offered against attitudinally-based analyses.10 The proponents of such examples posit a person who desires to possess a certain desire, such as, for example, a Walter Mitty-type person who desires to be a more derring-do individual. This person is then, without his knowledge, subjected to the inculcation of the desired desires into his motivational economy by someone such as a benevolent hypnotist, or a kindly neurosurgeon, and begins to act on these desires. Since his possession of the desires in question were not under his control, since both they and their stimuli were inculcated into him by another, the proponents of such examples hold that when he acts on them, he is acting on desires that are alien to him even though he might endorse his so acting. But just as the intuition that irresistibility precludes identification was mistaken, so too is this intuition. To see this, consider the case of Lord Fawn from Trollope’s novel The Eustace Diamonds, who desired to ask his estate steward, Gowran, if he had seen his (Fawn’s) fianc´ee in the company of another man (see p.183 in Frankfurt 1988d). Despite Fawn’s desire to learn about his fianc´ee’s activities from Gowran, he found himself unable to discuss the matter with him for ‘every feeling of his nature revolted’ against his doing so. Yet although Fawn’s aversion to discussing his personal affairs with his servant, and hence his withdrawal from the discussion, were not under his voluntary control he still identified with his effective first-order desire to withdraw. Owing to Fawn’s nature, to continue with this discussion was, in Frankfurt’s terms, ‘unthinkable’ for him. Fawn cared so much about the social order of which he was a part that although he was not logically or causally necessitated to withdraw from the conversation, he could not will himself to overcome his will to withdraw. To do so would be, in a sense, to betray himself. It would be to go against something that he cared deeply about, where this caring informed his choices

116

J.S. Taylor

over which actions to perform. Thus, although Fawn’s effective first-order desire to halt the conversation with his steward in its tracks was not a desire over which he had volitional control, it was nonetheless a desire that was immediately and fully attributable to him, and so it was a desire that he fully identified with. That a person cannot control the process by which he came to have his effective first-order desires does not then preclude him from identifying with them. Moreover, the example of Lord Fawn also shows that a person is not precluded from identifying with his effective first-order desires if they originate from a motivational set that he came to possess through a process over which he had no control. Thus, even if Fawn came to care about the social order of England and his place within it as a result of a socialization process over which he had no control, the fact that he has come to care in this case is a constitutive element of who he now is. That he had no control over his coming to have this volitional structure thus does not preclude any desires that flow from it from being attributable to him, and thus being a desire that he identifies with.

6.2.3 Desires Induced by Other Agents Given that neither the irresistibility of an effective first-order desire, nor its stemming from a motivational state over which the acting agent had no control, are sufficient to preclude the acting agent from identifying with it, perhaps the explanation for the plausibility of manipulation-based objections (such as that outlined above) to attitudinal analyses of identification lies in their focus on the manipulations of other agents. That is, perhaps it is the case that a person is objectively alienated from a desire if it is induced into her motivational set by another agent. Prior to assessing the plausibility of this claim, two distinct ways in which one agent can induce desires into another must be distinguished: Invasive inductions, and noninvasive inductions. A desire that is invasively induced into a person’s motivational set is one that is directly implanted into it by another agent. Desires that are implanted through hypnosis (as in the example above), demonic or neurosurgical manipulation, and brainwashing are of this sort. A desire that is noninvasively induced into a person’s motivational set is one that he possesses directly through the operation of his reason. It is uncontroversial that a person will typically identify with desires that he has been noninvasively led to have by another agent. If Bill tells Ben about a new restaurant in town that serves wonderful sushi, knowing that Ben would be interested in trying it out, Ben would (if everything about his situation is normal) identify with the desire that Bill led him to have. Given this, the proponents of manipulation-based counterexamples to attitudinal analyses of identification focus on the invasive inculcation of desires, especially desires that do not cohere with those that the victim of such inculcation already possesses. Consider, for example, an objection that John Christman levels against Frankfurt’s analysis of ‘decisive’ identification, which is an attitudinal analysis of this sort:

6 The Myth of Objectively Alien Desires

117

Imagine someone who has been secretly hypnotized to want strawberries, where the hypnotist includes a directive to ignore any information concerning the hypnosis itself. Such a person would ‘decisively’ endorse the desire for strawberries and no new information she might gather would dissuade her of her preference. But such a person is surely not autonomous relative to this desire for strawberries (see p. 8 in Christman 1991).

First, a distinction needs to be made between a person identifying with his inculcated first-order desire for strawberries, and his identifying with his effective first-order desire for strawberries that he possesses as a result of decisively endorsing this desire and then being moved to act on it. It is uncontroversial that the victim of hypnosis in Christman’s example would not identify with the desire for strawberries at the moment of its inculcation. (Christman clearly states that he would not, claiming that ‘. . . such a person is surely not autonomous relative to this desire for strawberries,’ while this person would not yet have decisively identified with this desire, and so would fail to meet Frankfurt’s conditions for identification.) (See p. 9 in Christman 1991.) The question, then, is whether this person should be said to identify with his effective first-order desire for strawberries that he forms as a result of the hypnotic inculcation of a desire for strawberries into him. On Frankfurt’s view, he would, provided that he has decisively identified himself with this desire. Such decisive identification will occur after he has reflected ‘on his own desires either because they conflict with each other or because a more general lack of confidence moves him to consider whether to be satisfied with his motives as they are’ (see p. 169 in Frankfurt 1988c). A person who so reflects will ‘without arbitrariness terminate a potentially endless sequence of evaluations when he finds that there is no disturbing conflict, either between results already obtained or between a result already obtained and one he might reasonably expect to obtain if the sequence were to continue’ (see p. 169 in Frankfurt 1988c). On Frankfurt’s view, then, a person will decisively identify with a desire after he has canvassed his other desires and decided that the desire in question does indeed fit with those that he already identifies as his. Turning now to Christman’s example of the person who decisively identifies with a desire for strawberries after this desire has been hypnotically inculcated into him, on Frankfurt’s view, this person would only decisively identify with this desire after evaluating it in the light of his other desires and deciding that it fits into his motivational set. Since through this process of evaluation and acceptance this person would become active, rather than passive, with respect to this desire, and would make it internal to his motivational psychology, it seems that he should be said to identify with it, its hypnotic origins notwithstanding. Thus, the mere fact that a person comes to possess a desire after another has invasively inculcated it is no bar to his identifying with it.

6.3 Jointly Sufficient Objective Conditions for Alienation? That a person’s effective first-order desire is irresistible, owes its origins to a process beyond his control, or has been invasively inculcated within him by another, need not preclude him from identifying with it. But this need not deter the opponents

118

J.S. Taylor

of attitudinal analyses of identification who seek to establish objective, structural means of distinguishing those desires that a person identified with from those that she is alienated from. An opponent of such attitudinal analyses of identification could claim that even if a person will not necessarily fail to identify with one of his effective first-order desires if it exhibits one of these properties, if it exhibits all of them he will necessarily be alienated from it. This, she might claim, is why manipulation-based counterexamples to attitudinal analyses of identification are so intuitively compelling—they have all three of these conditions built into them. Unfortunately for the opponent of attitudinal analyses of identification, things are not as clear-cut as this. First, such manipulation-based counterexamples trade on the intuition that the person subjected to such manipulation (such as the Walter Mitty-type mentioned above) is moved by an effective first-order desire that operates entirely independently of his motivational psychology. If this is true, this certainly supports the view that he is alienated from it, no matter what attitude he might adopt towards it. For a person to identify with an effective first-order desire that moves him to act, however, he must act. But a person whose bodily movements are utterly independent of his motivational psychology does not perform an action. Instead, he is merely subject to his body moving in certain ways. He experiences merely a bodily spasm. In the clear case then, where a person is hypnotized into making certain bodily movements, and where these have no connection at all to his motivational psychology and so come as a surprise to him, he is certainly alienated from both his movements and the motivational forces that act within him to cause them. But in such a case, he does not act at all—and so fails to satisfy the basic conditions required by attitudinal analyses of identification for him to identify with these effective motivations. This is an extreme case of the type of manipulation-based example that rests on the view that the three conditions outlined above are jointly sufficient for a person to be alienated from a desire that satisfies them. A less extreme case would be one in which a person accepted the desire that he was hypnotized into having, and engaged with it as he would with any other desire that he decided to satisfy. If this person engaged with this desire in the way that the proponents of attitudinal analyses consider to be sufficient for him to identify with it (and such engagement would vary according to the requirements of the analysis in question), then it seems that he would act rather than be moved by a bodily spasm. If this is so, then the fact that he was manipulated into having this desire—and so, claim the opponents of such attitudinal analyses, it is clearly alien to him—would form the basis of a counterexample to the attitudinal analysis in question. Yet again, things are not as clear-cut as they might at first seem. Frankfurt, for example, denies that a person will necessarily fail to identify with an effective firstorder desire that has been hypnotically inculcated into him. He writes: A passion is especially likely to be external when it is artificially induced by such means as hypnosis . . . In cases of this sort, the passion generally does not arise as a response to a perceived experience. It may well present itself to the person in whose history it is contrived to occur, accordingly, as discontinuous with his understanding of his situation and with his conception of himself. Even so, the person often appears by a kind of instinct

6 The Myth of Objectively Alien Desires

119

to circumvent these discontinuities with rationalizations: he instantaneously provides the passion with meaning, or somehow construes it as having a central place in his experience. Then, despite its origin, the passion becomes attached to a moving principle within the person; and the person is no more a passive bystander with respect to it than if it had arisen in more integral response to his perceptions (see p. 62 in Frankfurt 1988b).

Frankfurt’s claim that a person might identify with an effective first-order desire that he only possesses as a result of hypnosis might at first appear mysterious. Such mystery can be dispelled once it is recalled that an analysis of identification is an analysis of what it is for a person’s effective first-order desire to be her own, one that is legitimately attributable to her as a person such that when she acts on it, she acts of her own free will. Provided that a person decides to act on a desire such that it becomes part of her motivational psychology, it will be a desire that can be attributed to her. It will become one that is internal to her, rather than external to her; it will become one that she is active with respect to, rather than passive with respect to. For example, the Walter Mitty-type person outlined above might, after his unwitting course of hypnosis, suddenly find himself beset by the (hypnotically inculcated) effective first-order desire to save a cat from a tree when he sees one stranded. Unaware of the origins of this desire, but believing that satisfying it will make him more heroic, he decides to act on it after deciding that, given his other motivations, he should, on balance, do so. Despite the origins of this desire, then, this person’s unmanipulated response to the appearance of this desire in his motivational set renders him active with respect to it and so makes it legitimately attributable to him as a person. To make this claim more palatable to the proponents of manipulation-based counterexamples, one could imagine a person out shopping who suddenly desires to buy some daffodil bulbs despite being entirely indifferent towards horticulture in general and daffodils in particular. After being puzzled by the onset of this desire, he thinks to himself ‘Why not?’ (or else he rationalizes why he is beset by it) and buys some bulbs. Given his response to the onset of this desire, there is nothing odd about attributing it to him after he has decided to act on it, irrespective of its origins. His response to this desire renders him active with respect to its operation through his incorporating it into his motivational psychology in such a way that when he acts on it, he does so of his own free will. Why a person can be held to identify with a desire that he has been manipulated into possessing is thus unmysterious, provided that he responds to this desire in such a way as to make it his own.

6.4 Conclusion At the start of this paper, I noted that the Regress-cum-Incompleteness objection appeared fatal to attitudinal analyses of identification and that this has led philosophers to develop alternative, structural analyses of what it is for a person to identify with her effective first-order desires. Such structural analyses do not require a person to endorse her effective first-order desires for her to identify with them and so are immune to the Regress-cum-Incompleteness objection. The development of a structural analysis of identification, however, presupposes that one can distinguish

120

J.S. Taylor

between those effective first-order desires that are alien to a person and those that she identifies with, purely on the basis of the relationships that the desires in question have with the rest of her motivational economy. But as I have sought to show in this paper, it is not clear that such a structural analysis of identification is possible. This is because, I have argued, none of the three most common ways in which persons have sought structurally to characterize alien desires can readily distinguish between the effective first-order desires that a person is alienated from, and those that she identifies with.11 This conclusion might appear to leave analyses of identification at an impasse. On the one hand, attitudinal analyses of identification face the apparently fatal objection of the Regress-cum-Incompleteness problem. On the other hand, structural analyses of identification face the problem that there is no obvious way to distinguish alien desires from desires that their possessors identify with on structural grounds. The way out of this apparent impasse has, however, been suggested in the discussion above. I noted in that discussion that in Frankfurt’s view, a person could identify even with a desire that he had been hypnotized into having, provided that he rendered it his own through deciding to identify with it. It is part of the nature of a decision that the person who makes it is necessarily active with respect to it. One must make a decision; one cannot simply have a decision come upon one.12 Since this is so, the above impasse could be avoided by returning to Frankfurt’s analysis of decisive identification, on which a person identifies with her effective firstorder desires if she decides to endorse them (see pp. 164–176 in Frankfurt 1988c). Although this is an attitudinal analysis of identification, it would not be subject to the Regress-cum-Incompleteness problem, since the fact that a person must necessarily be active with respect to her decisions means that she need not make a higher-order decision to endorse them to ensure that they are hers in the relevant sense. Of course, for such an analysis to be complete, the nature of decisions must be spelled out. Moreover, one might object that a person could be manipulated into making a decision (as Iago manipulated Othello) and thus suffer a diminution in her autonomy both with respect to it, and with respect to those of her actions that flow from it.13 Since this is so, the relationship between a person’s identifying with her effective first-order desire and her being autonomous with respect to it must be spelled out, such that a person might both identify with her effective first-order desire and yet still suffer from compromised autonomy with respect to it.14 But these are matters for another paper. All that I have (modestly) sought to establish here is that such a paper is worth writing, for reports of the death of attitudinally-based approaches to analyzing what it is for a person to identify with her desires have been greatly exaggerated.15

Notes 1 The above metaphors (of ‘active’ and ‘passive,’ ‘internal,’ and ‘external’) for what it is for a person to identify (or to fail to identify) with her desires are from pp. 58, 59, 61 in Frankfurt (1988b). See also

6 The Myth of Objectively Alien Desires

121

pp. 132–133 in Frankfurt (1999a). For an excellent discussion of the relationship that the concepts of ‘activity’ and ‘internality’ hold to that of identification, see Moran (2002). 2 These criticisms are outlined in Taylor JS (2005). 3 Frankfurt acknowledges this (see p. 104 in Frankfurt 1999b). 4 Ekstrom’s concept of a preference is like Frankfurt’s concept of a second-order volition, except that Frankfurt allowed that a person might form a second-order volition for any reason at all, whereas for Ekstrom, a person forms a preference for a first-order desire because she finds a certain first-order desire to be good. 5 For Ekstrom, the definition of coherence is: a preference for desire d coheres with the character system of S at t if and only if, for any competing preference for desire g, it is either (i) more valuable for S to prefer that d than to prefer that g on the basis of the character system of S at t, or (ii) as valuable for S to prefer that the conjunction of g and a neutralizing desire n as it is for S to prefer that g alone on the basis of the character system of S at t. The relevant definition of competition is: the preference for desire g competes with the preference for desire d for S on system Y at t if and only if it is less valuable for S to prefer that d on the assumption that g is good than on the assumption that g is bad on the basis of the system Y at t. Given these definitions, personal authorization of a preference is defined as: S is personally authorized at t in preferring that d be effective in action if and only if the preference for d coheres with the character system of S at t. See p. 610, note 26 in Ekstrom (1993). 6 Bratman’s account of identification is outlined here to show that a structural, objective account of identification could include an attitudinal component. Its inclusion should thus not be taken to imply that it necessarily falls prey to the same difficulties of correctly identifying alien desires that the objective, structural analyses of identification that I consider below fall prey to. However, insofar as Bratman’s analysis of identification is based on an analysis of personal identity, it is vulnerable to objections of a different sort for in addition to his analysis of identification being subject to criticism, so too is his analysis of personal identity on which it is based. 7 Bratman recognizes that this account of what it is to treat a desire as reason-giving is only partially complete, for certain desires (such as a desire to favor a side-constraint on action) might be reasongiving without being end-setting. However, he limits his attention to those reason-giving desires that are end-setting. 8 I discuss the possibility of such a tripartite taxonomy of desire in relation to Frankfurt’s work on identification in Taylor JS (2001). 9 Frankfurt’s discussion of the willing and the unwilling addict is contained in pp. 17–19 and pp. 24–25 in Frankfurt (1988a). 10 Such counterexamples to attitudinally-based analyses of identification are outlined in pp. 4–10 in Taylor (2005). 11 Here, I follow Frankfurt in focusing on a person’s effective first-order desires as the primary objects of identification or alienation. However, this argument could also apply mutatis mutandis to the question of whether a person identifies with her first-order desires, whether effective or not in any given situation, simpliciter. 12 In instances where persons write of having decisions come upon persons in this way, it seems that they are writing not of decisions proper but merely of coming to the realization that they believe or feel a certain way. This seems to be the case in David Velleman’s famous passage in which he discusses a person having a conversation with a friend which becomes increasingly acrimonious leading them to part in anger, and who later came to believe that he had subconsciously decided to break off the friendship. Since the person concerned did not actively decide to break off this friendship, and it is genuinely mysterious as to how a decision could sneak up on a person in this way, unbidden, it seems that Velleman would do better to describe the mental progression that the person in question experienced as a realization that he felt a certain way about his friend, rather than a recognition that he had previously and unknowingly made a certain decision. See pp. 464–465 in Velleman (1992). 13 For a discussion of this point, see Taylor (2004). 14 It is widely believed that this is not possible, for the phrases ‘autonomous with respect to her effective first-order desire’ and ‘identifies with her effective first-order desire’ are taken to be synonymous (see p. 4 in Taylor 2005). 15 An early version of this paper was presented at a conference on Values, Rational Choice, and the Will that was held at the University of Wisconsin—Stevens Point in April 2004. I thank my audience on that

122

J.S. Taylor

occasion for their helpful comments. I am especially grateful to David Chan, both for encouraging me to develop this paper and for his excellent and insightful comments on an earlier draft, all of which helped me to avoid much philosophical opacity and many philosophical infelicities.

References Bratman ME (1996) Identification, decision and treating as a reason. Philosophical Topics 24:1–18 Bratman ME (2000) Reflection, planning, and temporally extended agency. Philosophical Review 109:35–61 Christman J (1991) Autonomy and personal history. Canadian Journal of Philosophy 21:1–24 Ekstrom LW (1993) A coherence theory of autonomy. Philosophy and Phenomenological Research 53:599–616 Frankfurt H (1988a) Freedom of the will and the concept of a person. In: Frankfurt H (ed) The importance of what we care about. Cambridge University Press, Cambridge Frankfurt H (1988b) Identification and externality. In: Frankfurt H (ed) The importance of what we care about. Cambridge University Press, Cambridge Frankfurt H (1988c) Identification and wholeheartedness. In: Frankfurt H (ed) The importance of what we care about. Cambridge University Press, Cambridge Frankfurt, H (1988d) Rationality and the unthinkable. In: Frankfurt H (ed) The importance of what we care about. Cambridge University Press, Cambridge Frankfurt H (1999a) Autonomy, necessity and love. In: Frankfurt H (ed) Necessity, volition and love. Cambridge University Press, Cambridge Frankfurt H (1999b) The faintest passion. In: Frankfurt H (ed) Necessity, volition and love. Cambridge University Press, Cambridge Haji I (1998) Moral appraisability: puzzles, proposals and perplexities. Oxford University Press, New York Leon M (2001) The willing addict: actor or (helpless) bystander? Philosophia 28:437–443 Moran R (2002) Frankfurt on identification. In: Buss S, Overton L (eds) The contours of agency: essays on themes from Harry Frankfurt. MIT Press, Cambridge MA Taylor JS (2001) Book review: Harry G. Frankfurt, Necessity, volition and love. Philosophical Quarterly 51:114–116 Taylor JS (2004) Autonomy and informed consent: a much misunderstood relationship. Journal of Value Inquiry 38:383–391 Taylor JS (2005) Introduction. In: Taylor JS (ed) Personal autonomy: new essays on personal autonomy and its role in contemporary moral philosophy. Cambridge University Press, Cambridge Velleman JD (1992) What happens when someone acts? Mind 101:461–481 Watson G (1975) Free agency. Journal of Philosophy 72:205–220

Chapter 7

On the Intelligibility of Bad Acts Michael Stocker

Abstract Do desires provide reasons and intelligibility or are desires intelligible only if had for, based on, reasons? To be intelligible, must a desire or act be had or done for a reason? Many classical ethicists, joined by Joseph Raz and many others, argue that reasons are necessary for intelligibility, that intelligibility requires desiring and acting under the aspect of the good, sub specie boni. I agree with Raz, Anscombe, and others that not all desires make for intelligibility, only those with desirability characterizations that answer Anscombe’s question, ‘What’s the good of that?’ But I argue, against Raz, that such characterizations, as well as intelligible desires and acts, need not be sub specie boni. I use Foot’s neo-Aristotelian teleology to argue that Raz may be right about people who have good second natures, good desires and good acts. But he is wrong about people whose second nature is not so good. Their desires and acts may be deformations of the desires and acts of good people. This suggests a position somewhat like Raz’s: the intelligibility of bad desires and acts may be ‘borrowed’ from the good desires and acts that they are deformations of. Keywords Desirability characterization · Intelligibility · Raz · Reasons · Sub specie boni

7.1 The Rational and the Classical In his ‘Incommensurability and Agency’ (1997) and also his Engaging Reason: On the Theory of Value and Action (1999), Joseph Raz presents and defends what he calls the classical as opposed to the rationalist conception of human agency. Putting the classical view, he argues that reasons have to do with the good, that only M. Stocker e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

123

124

M. Stocker

the good can provide reasons: We desire and act under the aspect of the good, sub specie boni.1 He likens the classical view to the views held by Plato and Aristotle, crediting Elizabeth Anscombe’s Intention (1957) for its modern formulation and interest (see p. 22, fn. 1 in Raz 1999). In §38 of that work, she writes, ‘the question “What do you want that for?” arises – until at last we reach the desirability characterization, about which “What do you want that for?” does not arise, or if it is asked has not the same point . . .’ An alternate, general form of her question is ‘What’s the good of that?’ According to the classical view, if there is no good in doing that, there is no reason to do it, and thus, it cannot be done for a reason. On Raz’s view, this is also a requirement for wanting: ‘We cannot want what we see no reason to want any more than we can believe what we think is untrue or contrary to the evidence’ (see p. 116 in Raz 1997). Raz puts this in terms of intelligibility: only what is desired or done for a reason is intelligible. (He relies, so far as I can see, on an intuitive understanding of ‘intelligible’, as I will, too.) This great set of issues in ethics and theories of agency finds Aristotelians and Kantians on one side and Humeans on the other. Rationalists hold that desires make for intelligible action – acts are intelligible where they are done for, to satisfy, a desire. However, some rationalists such as Hobbes understand goodness in terms of desire. In Leviathan, Ch. VI, he writes, ‘whatsoever is the object of any man’s appetite or desire, that is it which he for his part calleth good’. A Hobbes-like, rationalist view can thus be taken as entailing that the action and desire do have to do with goodness. (Throughout, I use ‘act’ and ‘action’ interchangeably.) These rationalists differ from classical theorists in that they, the rationalists, hold that desires are (or can be) constitutive of and prior to intelligibility-making reasons and, for that matter, constitutive of and prior to goodness. To be the target of Raz, they need not hold that every desire makes for intelligibility. They need hold only that at least some desires that are not, themselves, sub specie boni can make for intelligibility. Raz holds that to provide a reason for action, a desire must satisfy a desirability characterization – to use Anscombe’s term – and not all desires do. I will argue that some desires for the bad – desires that are sub specie mali – can provide intelligible reasons for action. I will also argue – perhaps courting confusion or a charge of a contradiction in terms – that some desires for the bad can provide desirability characterizations. I count myself as a follower of Aristotle, at least on many points, and as someone who has been strongly influenced by Anscombe. I am inclined to agree with Raz about the importance of desirability characterizations. I will not argue for our common position here. But I disagree with him, both in his interpretation of Aristotle and Anscombe and also about what intelligibility requires. In ‘Desiring the Bad’ (1979), I did take Aristotle to hold the position Raz attributes to him. I no longer agree with this interpretation of Aristotle. It is false to his discussion of contempt, spite, and malice in the Rhetoric. And, as will be discussed below, it is false to his account of non-ideal, especially deformed, second natures.

7 On the Intelligibility of Bad Acts

125

7.2 How Must Goodness Figure When Raz holds that intelligible action is sub specie boni he is not holding what we might call the explicitly evaluative view. This view, which Raz rejects, has it that in order for my desiring X or doing X to be intelligible, I must believe X to be good. He writes that he does not attribute a kind of high order reflectiveness to people: for example that they not only think of their actions as pleasurable or thrilling or beneficial to X or Y, etc. but also think of them as good in virtue of possessing those properties. All I ever claimed is that people act for considerations which we classify as a belief in the possession of a good making property.2

Raz holds that my desiring or doing X is intelligible if X has properties that are good or good making, whether or not I believe or even note this, and I do X on account of those properties. For example, he holds that the pleasure I anticipate from swimming, on account of which I go swimming, makes intelligible my desire and my going to swim. I need not believe the pleasure to be good. Several assumptions that I will make should be noted. First, I will take Raz to be holding that being sub specie boni is necessary for intelligibility and not that it is also sufficient for it. Holding that it is not sufficient allows what I think is right, but am unsure Raz would: that it can be unintelligible for a person to be attracted to doing what is good or what is believed good. I think it can be unintelligible for a particular person by being contrary to that person’s character or second nature. (Second natures are discussed below.) So, we might find giving generously to Oxfam unintelligible for a convinced and hardened xenophobe. Second, I will limit my examination to cases where the person is aware of what is good or good making. This is what Raz refers to by ‘property’, in that quote. What the person need not be aware of is that this is good or good making. This is the content of the ‘high order reflectiveness’. Third, although I cannot find where he says or shows that he does allow that explicitly evaluative beliefs can make for intelligibility, this must be allowed. Otherwise we would have to find unintelligible someone’s desire and act expressed by ‘I want to get you something good’, ‘I want to do what is best’, ‘I want to do what is right’, and so on. These can be particular desires and about particular acts. Such explicitly evaluative concerns can also be more general, perhaps even giving thematic considerations that inform major portions of a person’s life. I am thinking here of people who do not want or do anything unless they believe it good; or more strongly, who want things in virtue of believing them good, perhaps even in proportion to how good they believe them to be. We might find parts or aspects of such lives unattractive, but they certainly can be intelligible.3 Further, if we do not allow that intelligibility can be secured by explicitly evaluative beliefs, we would presumably have to find unintelligible Augustine’s desire to steal the pears which, he said, he ‘loved only for the theft’s sake’ and ‘Doing this pleased us all the more because it was forbidden. . . . I was being gratuitously

126

M. Stocker

wanton, having no inducement to evil but the evil itself’ (see Book II, Ch. viii, Sect. 16 in Augustine 1955). To be sure, those who hold that desires and acts are intelligible only if they are sub specie boni will, on account of that requirement, find Augustine’s desire and act unintelligible. I find them intelligible, all too intelligible. But all I want to say on this now is that if they are unintelligible, this is not due to their being explicitly evaluative.

7.3 Arpaly on De Dicto and De Re Moral Beliefs For these reasons, then, I will examine Raz’s claim about intelligibility in terms of non-exclusive disjunctive requirements: desire and action are intelligible only if they concern what is believed good or what is good, whether or not the agent believes or even notes this. The central claims of my work do not depend on our understanding the difference between these two disjuncts, but some comments on them may be useful, if only to show how it can be difficult to distinguish between them and also to show also why it may not be important to do so. A somewhat similar use of these disjuncts – put in terms of de re and de dicto moral beliefs – is argued for by Nomy Arpaly in Unprincipled Virtue (2003).4 She is concerned to argue that virtue – say, a morally good action or a morally good person – does not require de dicto moral beliefs, but only doing what is de re good. I think it reasonable to assume that virtue requires intelligibility and something thing(s) else in addition – virtuousness requires intelligibility but intelligibility is not sufficient for virtuousness. Nonetheless, a brief discussion of some of her claims about virtuousness and de dicto and de re moral claims should advance our understanding of some of Raz’s claims about intelligibility. Writing about Huckleberry Finn’s letting Jim go, she writes: That Huckleberry begins to perceive Jim as a fellow human being becomes clear when Huckleberry finds himself, to his surprise, apologizing to Jim – an action unthinkable in a society that treats black men as something less than human. As mentioned above, Huckleberry is not capable of bringing to consciousness his nonconscious awareness and making an inference along the lines of ‘Jim acts in all ways like a human being, therefore there is no reason to treat him as inferior, and thus what all the adults in my life think about blacks is wrong.’ He is not a very clear abstract thinker. . . . his reluctance [to turn Jim in] is to a large extent the result of the fact that he has come to see Jim as a person, even if his conscious mind has not yet come to reflective awareness of this perceptual shift. To the extent that Huckleberry is reluctant to turn Jim in because of Jim’s personhood, he is acting for morally significant reasons. This is so even though he does not know or believe that these are the right reasons. . . . He is also unaware, or only dimly aware, of the fact that he is acting for these reasons in the first place. But he is acting for moral reasons all the same, in the de re sense of the expression ‘moral reasons’. (See p. 77 in Arpaly 2003, italics in original.)

Arpaly holds that Huckleberry does not know he has moral beliefs and, in fact, does not have moral beliefs. She has him perceiving Jim as a fellow human being

7 On the Intelligibility of Bad Acts

127

and as concerned for his personhood. She describes him as apologizing to Jim. On the basis of these and similar facts, she holds that Huckleberry or his act is virtuous, morally good. I find some of her claims difficult to accept. To show why, I will start by noting that there are any number of scenarios in which Huckleberry decides to release Jim, but he is not thereby shown morally good. For example, he thinks it would be fun or amusing to release Jim, or he wants to see what will happen if he does it, or he does it because he wants to do something transgressive. But, of course, doing it for those reasons is not to do it as described by Arpaly: on account of Jim’s being a fellow human being, in light of harm to his personhood. These, she says, are de re moral beliefs; they are not de dicto moral beliefs; they show Huckleberry, or his desire and act, morally good. My difficulty can be brought out by focusing briefly on Huckleberry’s apologizing to Jim. I can be said to apologize if I utter the appropriate words, e.g., ‘I’m sorry’. So too, I can apologize as part of a joke or to see what will happen next or to be transgressive. But such apologies would not show me or my apologizing to be morally good. That requires that I mean the apology as an apology. I think that this requires my thinking or realizing, in some way or another, that I have harmed or wronged or failed the person I am apologizing to (or someone else suitably related to that person). This seems at least close to being a de dicto moral belief. I think much the same holds for some others of Arpaly’s claims. She says that Huckleberry is sensitive to what is morally relevant in that he sees Jim as a person and he thinks that as a person or in his personhood, Jim would be harmed by being returned to his owner and that it is in light of these considerations that he lets Jim go. They are his reasons for doing so. I think each of these emphasized terms – ‘as a person’, ‘in his personhood’, ‘harmed’, and perhaps even ‘reason’ – is explicitly evaluative. I do not see how they differ from the explicitly evaluative, de dicto, ‘Huckleberry thought it would be bad or wrong to return Jim to his owner’. Another way to put my worry starts by accepting Arpaly’s description of Huckleberry, about what he takes into consideration (e.g., Jim’s personhood), and how and when he takes it into consideration (e.g., as his reason to let him go now) – that is, those conditions which show him and his act virtuous. Having accepted this, I would ask: what in addition would have to be said to say that he also had the relevant de dicto moral belief(s) – e.g., that he believes that letting Jim go is morally required or is good? And what would be different if he did have the ‘absent’ de dicto moral belief(s), not just the de re moral belief(s)? My point in going through this is not just to ask whether we can agree with the ways Arpaly distinguishes – more or less joined by Raz – between de dicto and de re moral beliefs and claims. It is also – and here more to the point – to ask why and how it is important or even useful to distinguish them. I do agree with Arpaly that Huckleberry, as portrayed, does not use or have finegrained moral thoughts or fine-grained moral reasoning. (Similarly for Raz.) I agree also that he is not reflective about his moral reasoning; and that he does not conclude what to do, at least not from articulated principles; and that he does not use, think

128

M. Stocker

in terms of, such out-and-out moral terms as ‘this is required’ or ‘this is good’. But I do not see how any of this has to do with distinctions between de dicto and de re moral beliefs or with the importance of these distinctions. As for fine-grained thinking, having a de dicto belief such as ‘this is good’ or ‘this is the right thing to do’ is entirely consistent with – and indeed is characteristic of – moral reasoning that is an opposite of fine-grained: e.g., that is crude, narrow, prejudiced, stereotyping, one-eyed, monomaniacal, and the like (in some combination or other). Much the same holds for being reflective and what is articulated. These can involve de re or de dicto moral thoughts or both sorts of moral thoughts. To start on this, we might think of many crafts – e.g., painting, philosophy, and psychoanalysis.5 I do not mean that these and other crafts do not involve principles, nor that there is no need to learn and articulate any principles. I mean, rather, that in at least many cases, doing well at these crafts does not require having the principles in mind or rehearsing them – except perhaps on the final exam. In the practice of the craft, repetition may well give only what is ‘mechanical’, or unoriginal and derivative, or academic (as in ‘academic art’). As well – in these crafts and elsewhere – there are processes of acquiring beliefs and learning principles and then giving them up, no longer using them, transcending them. Sometimes the relation here is that what is done after giving them up is informed by what one has learned and given up.6 Sometimes it is that acquiring and using the beliefs and principles have ‘strengthened’ one’s mental muscles, much as philosophy is often said to be good training for law school. And so on. This is to speak about disciplines. Different people, at different stages of their learning and lives, may vary. Some people learn best by learning the principles; some, by applying them; and some, by transcending them. A student who is good at rehearsing an argument may go on to original work. A professor who has been original may fall back on repetition. It can sometimes be useful to call principles to mind, to reflect on what one is doing, to ‘define one’s terms’, and the like. But sometimes, doing this changes what one is doing, perhaps ruining it. Some practitioners are good, while others are not, at simultaneously doing and attending to what one is doing – e.g., in the ‘evenly-suspended attention’ of psychoanalytic practice (see p. 111 in Freud 1953). Some expert practitioners are able, and some are unable, to articulate how they learned their craft or which of its leading principles they rely on or can even call to mind. Some expert teachers or students of the principles of a craft are, and some are not, good at the craft. Here we might think of fluent and elegant speakers: some are, and some are not, able to articulate ‘rules’ of grammar or style. Seeings: These make up what Arpaly says are Huckleberry’s de re moral thoughts and which show him to be morally good. These can be ‘immediate’ and nonreflective. But seeings can also be reflective, reached slowly, with considerable articulated attention to de re considerations. (Some connoisseurs’ discussions and evaluations are wonderfully or tediously detailed, lengthy, well thought out, and articulate.)

7 On the Intelligibility of Bad Acts

129

7.4 Natural Goodness Raz is concerned with intelligibility, not moral goodness. His requirements for intelligibility are put in terms of de re moral or in terms of a disjunction of de re or de dicto. For reasons similar to those advanced in the discussion of Arpaly, I think that so far as intelligibility is concerned there is very little, if any, useful difference between de dicto and de re beliefs. This will come out in my discussion of Raz’s claim that to be intelligible, acts and desires must be sub specie boni. There are many considerations that help show that what is not good, even what is bad, can be intelligible. These include the following: These desires and acts can play the role of a standard of what will be attempted. They can figure in what shows whether the person or act succeeded or failed. They can serve in or as a prelude to the deliberations typical of practical reasoning. They can be pursued with tact, intelligence, nuance, critical assessment of alternative means, and so on. As R. Jay Wallace comments, People sometimes exhibit great intelligence and skill in executing plans that they view as dubious or questionable – think, for instance, of the extraordinary talent many of us display at procrastinating when it comes to tasks that we regard as worthy but difficult (see p. 1 in Wallace 2001).7

In what follows, I want to show in another way that desires and acts that are sub specie mali can make for intelligibility. To do this, I will start with a line of reasoning – that I will call Aristotelian teleology – that may at first seem to support Raz’s claim that intelligibility requires acting and desiring sub specie boni. (I am not concerned to defend Aristotelian teleology. I use it because of what it helps us see about Raz’s claim. Other very different views could also do this.) Its main claims are that in an important sense of ‘natural’ and of ‘natural necessity’, it is natural for a living being of a certain sort – e.g., human or wolf or orchid – to do what is good for it. Indeed, we can identify and understand the workings of what is natural for it only by understanding what it is for it to live well, where ‘live well’ is understood in terms of its species: i.e., to be a good human, wolf, or orchid. (Aristotelian teleology is important for us by allowing for disparate development and by giving a standard of goodness – in ways shown and discussed below. What it allows and gives us are to be taken quite generally. Its particular details, including its being a naturalism and its being a teleology, are not important for our concerns.) Three passages from Philippa Foot’s Natural Goodness8 are useful in putting this Aristotelian teleological theme: First, We start from the fact that it is the particular life form of a species of plant or animal that determines how an individual plant or animal should be . . . And all the truths about what this or that characteristic does, what its purpose or point is, and in suitable cases its function, must be related to this life cycle. The way an individual should be is determined by what is needed for development, self-maintenance, and reproduction . . . (pp. 32–3, italics in original).

130

M. Stocker

Second, Goodness in plants and animals nests in an interlocking set of general concepts such as species, life, death, reproduction, and nourishment . . . (p. 36, italics in original).

And third, In spite of the diversity of human goods – the elements that can make up good human lives – it is therefore possible that the concept of a good human life plays the same part in determining goodness of human characteristics and operations that the concept of flourishing plays in the determination of goodness in plants and animals (p. 44).

Of central importance to us is the conjoint claim that we understand what is good for a being by understanding its species being and understanding also that it is natural for any being to seek its species good. When the being does not seek its species good, that can be understood only in terms of its being in unnatural circumstances or its being defective – not defective in general, if that even has a meaning, but defective in ways specific to, and understandable in terms of, that sort of being. (There are specifically human ways for humans to be defective, e.g., have certain neuroses.) Using this Aristotelian teleology, we can advance a claim similar to Raz’s, if it is not Raz’s. The claim is, first, that all we need do to explain a human seeking a human good is to note that it is a human doing the seeking; and second, to explain a human seeking what is a human bad, we need to cite both that it is a human and that it or the circumstances are humanly defective. This has it that desiring and doing what is good is or can be intelligible – in light of how Aristotelian teleology holds that humans are to be understood. It does not show that in order to be intelligible, desires and acts must be sub specie boni. What might do this is showing that desiring or doing what is bad is intelligible only if the desire or act is really desiring and trying to do what is good. Citing mistakes is a time-honored way – found in Socratic dialogues and, we should note, in Raz (see p. 28 in Raz 1999) – to try to do this: I want to drink this glass of gasoline, which would be bad to do, because I mistakenly think it is gin and gin, we may stipulate, is naturally good. (For reasons given below, e.g., when considering ‘substitutes’, there are many bad desires or acts that cannot be accounted for in terms of mistakes.) I find Aristotelian teleology powerful and attractive and it is certainly useful for examining Raz. As put so far, it does seem to give him support. But as put so far, it is inadequate as a complete account of human desire, activity, and intelligibility. In particular, it is not right about a developed nature – about the second nature – of many, if not most, people. For this reason, it is inadequate for Raz. It may be that we get the fullest, best understanding – perhaps the only adequate understanding – of humans in terms of good humans, living good lives. This might seem to tell us how humans are when born, how they develop, how circumstances bear on this development, and what they are like when mature. But what it tells us is what healthy human infants are like, what good forms of development are like, what circumstances are good for this development, and what good outcomes of good development of healthy infants are like.

7 On the Intelligibility of Bad Acts

131

It is part of this account that humans develop in complex ways. Put very schematically, human development is to be seen as involving multi-potential stages and natural transitions from one stage to another, resulting in (teleologically: moving toward and achieving) full and fully good adulthood. We can say that such adulthood is natural, a natural outcome of what is also natural as it develops naturally in natural (good enough) circumstances. But when we look at the resultant adults, it would be better to say that they have developed a second nature, a developed nature that naturally differs from the earlier stages it developed from. It is a commonplace that not all humans develop into fully good adults. Because of defective early stages, or defective development, or circumstances that make good development hard if not impossible, many adults have second natures which are less than wholly good. On the assumption of good beginnings, good development, good circumstances, it is natural that a fully good adult will desire what (perhaps only what) is good for it. Indeed, this is or is close to being a truism that comes very close to being a tautology: what is good for a human is shown by what a good human naturally wants. Intelligibility and goodness go hand in hand. But once we identify and understand those earlier stages, how they can be less than wholly good, and how development can go well or poorly, and thus how a person’s second nature can be good or bad, we can understand, find intelligible, many bad acts. That is, bad from a species perspective – the perspective, now in play, for both good and bad. This will be to understand at least some bad acts as second natural – not for all humans and not for good humans, but for a human who has developed in a relevant defective way. For such people, it can be natural to seek what, in the relevant way, is bad. This allows, but does not require, that we can understand how a bad second nature works – how it provides goals and how its reasoning proceeds – by understanding it in terms of a good second nature. So taken, we would understand how, say, greed works by understanding how liberality works – in particular, how a deformed sort of liberality works. This also allows, but does not require, that we can understand how a bad second nature works only if it is not too different from a good second nature. A second nature which is too bad or too different from a good one – in terms of what it seeks or how it thinks – may force us to see the person as a beast or otherwise inhuman. As so far put, Aristotelian teleology does not support Raz’s claim that intelligible desires and acts are sub specie boni; nor, for reasons discussed above, does it show that to be intelligible, desires and acts must be sub specie boni. At most, it supports his claim when it is restricted to good people and their desires and acts. But for people who are not so good, whose second nature is less than wholly good, it also shows, to the contrary, how intelligibility and human goodness come apart. I said earlier that a second nature which is too bad or too different from a good one – in terms of what it seeks or how it thinks – may force us to see the person as a beast or otherwise inhuman. We might, by the same token, find unintelligible many of their desires and acts (if they can even be said to be desires and acts). But this still leaves a very large middle ground of people who are neither wholly good

132

M. Stocker

nor beasts. Much of what is natural for them – in accord with their less than wholly good second nature – is not good and may well be bad, but, nonetheless, intelligible.

7.5 Nurture as Part of Nature The Aristotelian theme just presented is concerned with nature and natural goodness. It is vital for understanding humans to understand that nurture is natural. People develop in the ways they do because of nurture, socialization, and the like (nurture for short). One can become a good person, as set out by Aristotelian teleology, only through good enough nurture. It is extremely unlikely that human infants or children can survive into adulthood without nurture; and it is extremely likely that if, against all odds they do survive, they will do so more as beasts than people. Many forms of not so good people, of bad, defective people – again, as understood by Aristotelian teleology – are as they are due to nurture. The nurture may, itself, be bad, helping to turn good infants and children into less than good, perhaps even bad, adults. Nurture that is naturally good – good for the typical infant and child – may not be good for an atypical infant or child, and may (help) turn these into less than good, perhaps even bad, adults. So too for ordinary, naturally good nurtures in bad circumstances. (To simplify matters, I will focus on not so good, even bad, nurtures for good infants and children.) A few more words on nurture may be of use. To this end, I want to sketch a line of thought about nurture that can suggest strong connections between goodness and intelligibility. This will, however, not be to support Raz and his classical view, but rather to support a rationalist view, somewhat like Hobbes’, discussed at the outset. I will mention only a very few of the multitude of considerations favoring this line of thought. These will be put in terms of societies and in terms of individuals. Almost every society and group thinks well of itself, thinking that it and its ways are good, perhaps even supremely good, certainly better than others. This is a standard component of ethnocentrism – found in most every society. Turning now to individuals, we find narcissism as commonly thought of: as people who, quite generally, think well (perhaps very well) of themselves, thinking that they are good people and that what they want and do is good.9 We could here examine the great power of ‘the narcissism of minor differences’, to borrow from Freud’s ‘Civilization and its Discontents’. So too, of course, for relations between societies. Other considerations are provided by the importance many place on pride, not only in one’s country or religion or . . ., but also in oneself, and of course the importance of self-esteem. So too, we could look to some of the most powerful themes grounding successful advertising, evocations of beliefs and feelings of entitlement: that we deserve or owe it to ourselves to get and have. I mean get and have quite generally, not just whatever is being advertised. We could, in addition, call as evidence the danger and power of flattery. As well, there are the attractions and near impossibility of avoiding rationalizations.

7 On the Intelligibility of Bad Acts

133

For these and other reasons, at least many people and many societies find it unbearable not to think of themselves as good. Putting matters in terms of the unbearability of thinking poorly of ourselves, the pressing need to think well of ourselves, might seem to support Raz’s linkage of intelligibility and goodness. But it does not do this. His view is part and parcel of his classical view. These implications of ethnocentrism and narcissism turn his classical and anti-rationalist position on its head. His classical view focuses on people whose desires are conditioned, even informed, by their values. But this position has it that wants condition and inform values. We are driven by our powers of rationalization and our need to think well of ourselves, to believe that we are good, and the unbearability of not being able to think this. Consider, now, people who, for these or other reasons, have turned out to be not so good. Speaking at the level of society, perhaps they have absorbed, and now accept and further, vicious social and political practices – as might be shown by volunteering for service in the Nazi military or by ordering the whipping of a recalcitrant slave in antebellum USA. At the level of individuals, perhaps they have been raised over-indulgently, which helps explain why they now lack impulse control and why their desires know few limits – which, in turn, helps explain, for example, why they think that they are entitled to seek and do pretty much whatever they want. These are bad desires and bad acts, but they are intelligible. It might be objected that in these two examples, the desires and acts are in part good and thus their being intelligible does not show Raz mistaken: The first includes the good of being true or loyal to one’s society and supporting one’s political ideals; the second involves pleasure or autonomy. In the following section, I will suggest that these are not goods in the relevant sense.

7.6 Presumptively Good; Good, Other Things Being Equal I have so far presented some lines of thought that at least initially seem to give some support to Raz’s claim. I want now to consider another line of thought that may give some support to what I see as a weakened version of Raz. The claim is that what makes for intelligibility is what is or is believed conditionally or presumptively good. I will start with a comment by Anscombe from Intention, §38: Hobbes [Leviathan, Part 1, Chap VI.] believed, perhaps wrongly, that there could be no such thing as pleasure in mere cruelty, simply in another’s suffering; but he was not so wrong as we are likely to think. He was wrong in suggesting that cruelty had to have an end, but it does have to have a point. To depict this pleasure, people evoke notions of power, or perhaps of getting one’s own back on the world, or perhaps of sexual excitement. No one needs to surround the pleasures of food and drink with such explanations.

I join Augustine in holding that it is possible to take pleasure in others’ pain. He holds that evil acts ‘may even be done for the mere pleasure in another man’s pain as the spectators of gladiatorial shows or the people who deride and mock at others’

134

M. Stocker

(see Book III, Ch. viii, Sect. 16 in Augustine 1955, my emphases). But this is not the important issue here. What is, is intelligibility and goodness. It may seem that Anscombe is here supporting Raz’s claim that intelligibility requires goodness. But this is not what she says. She writes about desirability characterizations to show how, even on Hobbes’ view, taking pleasure in others’ suffering can be intelligible. The desirability characterizations she cites here are ‘notions of power, or perhaps of getting one’s own back on the world, or perhaps of sexual excitement’, and perhaps pleasure. This will help Raz only if power, revenge, or sexual excitement – or, now agreeing with Augustine, mere pleasure – are good: that is, only if, where they are the point of acting, acting is thus sub specie boni. I think it clear enough that under certain conditions, power, getting one’s own back, sexual excitement, and pleasure can be good. But I see no reason to think that in the circumstances Hobbes is concerned with they are good. Nor for that matter, do I think that Anscombe thought that in these circumstances they are good. As I just said, I read her as offering them as providing desirability characterizations, as intelligible answers to ‘What’s the good of that?’ and ‘What do you want that for?’ I take this to show that she does not hold that desirability characterizations must, themselves, be sub specie boni. Let us start with sadistic pleasure. My claim is that such pleasures are bad, but that even so, they can make for intelligibility. I am here talking about pleasure-inothers’-suffering, which is what Augustine’s people want. Although, it, too, is bad enough, I am not talking about the pleasure of those who simply want pleasure, any pleasure, and who do not care whether the pleasure comes from watching others suffer or from, say, swimming. My claim requires that at least some pleasures are to be individuated and differentiated qua pleasure by what they are in – as Aristotelians would say, by the activities they are or complete. Such pleasures are to be seen as indissoluble complexes, not pleasures and something else. There is no separable pleasure here. Some bads do contain good aspects or parts that are really why someone desires and seeks that bad thing: I might know that all in all, a concert will be bad, but for all that I might want to go to it in order to hear what I know will be the one marvelous piece. But pleasure-in-others’-suffering is not like this. The pleasure and others’ suffering form an indissoluble whole. Following the dictates of one’s conscience can be good. But it can also be atrocious. I think this may be Anscombe’s view, too, at least in ‘Modern Moral Philosophy’, where she writes, ‘Butler exalts conscience, but appears ignorant that a man’s conscience may tell him to do the vilest things’ (see p. 2 in Anscombe 1958). Consider here people who think, perhaps accurately, that their consciences dictate killing nonbelievers. The attendant desires and acts are bad but intelligible. Especially if we think of such consciences as being warped or deformed, there is little reason to hold that there is something good – viz., following one’s conscience – amidst much that is bad, and that it is desiring and seeking that good element that makes for intelligibility. Some people may agree that acting on such a conscience is bad and that its ‘source’, religious fanaticism, is also bad; while also holding that religion and

7 On the Intelligibility of Bad Acts

135

religious belief are good. Rather than take this up, I would reply that the people I have in mind do not want to be religious or to follow religious teachings. They are not like Abraham (who, in any case, was bad enough). They do not see themselves as disciples of the devil, who glory in doing what they hope is wonderfully evil. It is ‘only’ that their consciences are warped, deformed. To use Anscombe from ‘Modern Moral Philosophy’ again, they are corrupt. Love can be good, as is obvious. Indeed, it can be a great good. But consider abusive acts, acts of violent possessiveness, done from and characteristic of deformed senses of love, e.g., battering and killing one’s wife rather than letting her get a divorce. These desires and acts are bad but all too intelligible. (More on this below.) Much the same holds for revenge. Getting one’s own back can be sweet – as is said in the Iliad (XVIII, 109) about wrath. ‘Sweeter it is by far than the honeycomb dripping with sweetness’. It may sometimes be good – e.g., if justice requires or allows it and it is done justly. But it can often be bad – e.g., simply a manifestation of unjustified, illegitimate hate. Nonetheless, desiring and acting on it can be intelligible. I mention these because they are often enough offered as making for intelligibility. I have in mind such claims as ‘I did that because my religion required it’, ‘I did that because my conscience required it of me’, ‘I did that because I thought it good’, ‘I did that because it was pleasant’, ‘I did that because I love her’, ‘I did that because I hate him’. As I see matters, these desires and acts can be bad and also intelligible. At various places in this section, I said that these bads that make for intelligibility can be seen as conditionally good, as good under certain circumstance. I think it can be misleading or at least incomplete to leave it at this. Many of these conditional goods are also conditionally bad – good in certain circumstances and bad in others. If desiring and seeking them is intelligible, why give pride of place to goodness, albeit conditional goodness? Why say that the intelligibility has to do with the conditional goodness rather than the conditional badness, or sometimes one and sometimes the other? Some goods are conditionally good, and some bads are conditionally bad. This may be the right way to think of some of these. But I think it better to think of some of them, not just as conditionally good, but as presumptively good, as good, other things being equal. Perhaps there are also some bads that are presumptively bad. Hate is a good candidate here. I will not try to justify my thinking of them this way. I will simply make three suggestions: First, Aristotelian teleology and its claims about deformed and defective second natures can be used to explain why they should be thought of as presumptively good (or, on the contrary, as presumptively bad). On that view, what is natural, including natural wants, may be thought of as presumptively good. To be sure, these are rebuttable presumptions and, in the cases just considered, the presumptions are rebutted. (I am unsure whether Aristotelian teleology allows for conditional goods that are not also presumptive goods. What I say below will reflect this uncertainty.) Second, thinking of them this way may seem to give some support to a Raz-like position, suggesting that intelligibility depends on the default, natural understanding of goodness – not on goodness, but on conditional goodness or presumptive

136

M. Stocker

goodness. So it could be held that what is bad can make for intelligibility provided that it is conditionally or presumptively good – despite the fact that the condition is not satisfied or the presumption has been rebutted. To see this as supporting a Raz-like position comes close to holding that to show that we do not desire and act sub specie boni we must show that what we desire and seek is in no way good, not even conditionally or presumptively. This has certain attractions. To the extent that we are concerned with what is in no way good, we would not have to distinguish between what is at once bad and also presumptively or conditionally good, or what is at once overall bad and in part good. But, going this way would, I think, deprive the claim about sub specie boni of a good deal of its interest. It would come close to – too close to – holding that it is unintelligible to desire or seek only what is wholly bad. Third, it is unclear to me how, in general, we are to distinguish between bads that are in part good (contain good aspects), bads that are conditionally bad, bads that are presumptively bad, and bads that are indissoluble bad complexes; so too, of course, for goods. I think sadistic pleasures are clearly enough in the last category. Abusively possessive love and its desires and acts can, I think, also be in that category. But depending on how we fill out the case, it can be in another one. What the person wants may just be love – even innocent romantic love. He may only be so ‘misinformed’ or ‘misguided’ that he thinks the way to get and keep such love involves physical force and abuse. (After all, he might ‘reason’, that’s how he gets and keeps buddies.) But perhaps what he wants is to punish, to hurt, someone for, as he thinks, betraying him and his love. The desire to punish and hurt is here integral to his desire. What about Augustine’s delight in stealing the pears? We could put this, the delight, down to thrill seeking or rebelliousness. I have trouble in seeing them as presumptively good. I do see them as conditionally good and also as conditionally bad, and for that matter, bad in these circumstances.

7.7 Raz-like Positions I agree with Raz’s anti-rationalist claim that not just any and every desire makes for intelligibility. I also am inclined to agree with him that desirability characteristics are needed for intelligibility. However, I disagree with him that these characteristics can be given only by what is good or by what is believed good. This last claim about disagreement must be divided at least in two. To the extent that Aristotelian teleology is right, what good people desire and do will be good or good making. Further, these people can be reflective about the values of their lives, desires, and acts. They may, further, be sensitive to these values as such, as values. If they believe of something – have the explicitly evaluative (de dicto) belief – that it is not good, they will not want to do it or may even want not to do it. Only what is good or believed good will provide desirability characteristics for them. Their desires and acts will be intelligible. After all they are the standards

7 On the Intelligibility of Bad Acts

137

of human life, beings in terms of which humans are understandable. In these ways, and in regard to these people, only what is or is believed good will make for intelligibility. Raz’s view about intelligibility and goodness thus seems right about good people. Matters are different with people who are not so good. What they desire and do may well not be good or good making. They may, rather, be ‘substitutes’ or ‘replacements’ for what good people desire and do. In some cases, these substitutes and what they are substitutes for can be understood in terms of how and what these different people appreciate – desire and act in – the same circumstances. A medical example may be of use here: after running up a flight of stairs, people who have let themselves go soft may have a racing heart and be near collapse but people who get regular exercise may have only a moderately raised pulse rate and be only slightly winded. Here we have the same systems and organs and the same stresses, but different outcomes. Borrowing from Aristotle, we may say, similarly, that men who are not so good, who are, say, given to cowardice, may be in conflict about staying and fighting in battle. They may even desert their posts and their fellows. Not so for good men. These latter are not insensitive, they see and feel the dangers. But they will not be moved – or not moved very strongly – to run away. They are not conflicted about this. They are not enkratics. Both sets of desires and acts are intelligible. But, so Aristotle holds, only the desires and acts of good men are good. Other substitutions need not be nearly so simply and transparently related to what they are substitutes for. These more complex or more opaque substitutes are the stuff of psychoanalysis as well as many other sorts of interpretation. Drawing just from psychoanalysis, they can be, for example, symbolic, in any number of ways, of what they are substitutes for; they may be neurotic symptoms, compromise formations, reaction formations, perversions, personality disorders; and so on. In this, they contrast with the desires, thoughts, and acts of good people who live their lives easily and without experiencing significant problems. On the view just presented, these intelligible bad desires and acts are substitutes for good desires and good acts, e.g., what Aristotelian teleology holds to be natural. This could allow for a Raz-like claim: desiring and doing what is bad can be intelligible if that is a substitute for what is good. This could be seen as a natural extension of Raz’s use of mistakes: that desiring and doing what is bad can be intelligible – if a mistake accounts for this, e.g., thinking the gasoline to be gin. It remains to be seen whether all intelligible desires and acts are such substitutes or such mistakes. In any case, I think the claim using substitutes is only Raz-like (some might say, Raz light). I also think that there is no real point in arguing whether or to what extent that claim is weaker than Raz’s. It is sufficient if the differences are seen and appreciated. I want now to return to some earlier issues. I suggested that good people may be reflective and self-aware about their values. They may think of what they desire and do, and also of themselves, as being good, and they may care about this. But there are other ways of seeing oneself and the world. Here we might consider not so good people who see themselves as out for number one; or who see

138

M. Stocker

themselves as tough guys, as ruthless people; or who see the world in terms of constant struggle, with people pitted against each other as in the Social Darwinism of the captain in Jack London’s Sea Wolf; or who see justice as imposed on us by the danger of retaliation, and who would jump at the chance to have Gyges’ ring; or who see others more or less as objects to be used and disposed of or to be played with; or who, for narcissistic reasons, require being the center of love and attention; or who believe that striking back is the natural response to being struck; or people given to spite, malice, and contempt. These people, in larger or smaller areas, are not concerned to moralize or rationalize their desires and actions. For them, ‘I could get away with it’ or ‘I thus gained an advantage over him’ or ‘He harmed me, so I got him in return’ or ‘I was bored’ or ‘I wasn’t getting enough attention’ all make for intelligibility. All of these can be intelligible even though they are not good, but are bad. As I read him, Aristotle, one of Raz’s classical ethicists, holds in Rhetoric, II, 2 that spite, malice, and contempt are both intelligible and bad. However, as said several sections ago, even these people and peoples can think of themselves as good – on account, perhaps, of ethnocentrism, or narcissism, the unbearability of thinking poorly of oneself, and people’s remarkable powers of rationalization. So, even bad people, such as those just sketched, can think of themselves and their desires and acts as being good. The captain in Sea Wolf prided himself on his hard-headed realism, on his Social Darwinism. So, these societies and people may well think of themselves, their desires, and their acts as good. It is unclear to me whether their explicit evaluative beliefs about goodness – their beliefs that they, their desires, and their acts are good – are the same beliefs good people have using the same words. It is unclear to me whether, despite their words, these not so good people do think that they, their desires, and their acts are good. It may be that their understandings of goodness are so mistaken that they do not believe that they are good: in thinking or saying such words as ‘I am good’, they mis-speak. These issues should be taken up elsewhere, not here. They are the proper concerns of discussions of meaning, reference, the philosophy of language. I think it is sufficient for our purposes to say again what I said earlier: The words and beliefs of these not so good people may seem to support Raz’s claim about intelligibility and goodness. However, I think it better to take them as helping sustain my general claim that intelligibility and goodness do go together in various ways in regard to good people, good desires, and good acts; but they can come apart, especially in regard to people, desires, and acts that are bad or simply not so good.10

Notes 1

For another noteworthy presentation and defense of a somewhat similar view – focusing on Aquinas’s claim that action is sub specie boni – where the good is understood in terms of pleasure, usefulness, and fittingness – see Vogler (2002).

7 On the Intelligibility of Bad Acts

139

2 See p. 7, fn. 7 in his reply to Stocker (2004), both presented at the ‘Conference on the Moral Philosophy of Joseph Raz’ (March 22–24, 2002, Columbia University). 3 On their unattractiveness, see, e.g., Stocker (1976) and Stocker (1981). I think that Michael Smith, in p. 186 in Smith (1994), is too harsh when he says that to have de dicto concerns with rightness (and presumably goodness, too) is to fetishize rightness (goodness). Some such desires are neither unintelligible nor unattractive, nor fetishistic. We might think here of a trial judge or a jury trying to reach a verdict that is fair and just. And, going outside the moral realm, we might consider wanting to get the right answer when adding a column of numbers. 4 Arpaly’s understandings of ‘de re’ and ‘de dicto’ may be somewhat unusual, but the discussion below should make them clear enough. For some general considerations about and difficulties with various de re/de dicto distinctions, see Plantinga (1969) and also McKay and Nelson (2006). 5 My thanks are owed to Robert Daly, M.D. for discussion and for referring me to Schon (1983). 6 See, e.g., Iris Murdoch and Simone Weil, on attention and ‘a slow shift of attachments’ (p. 25 in Murdoch 1993). 7 This is directed against a claim made by Christine Korsgaard that, in effect anyway, is very similar to Raz’s present claim. As Wallace adds, this is ‘. . . a moralizing tendency . . .[to hold] that instrumental requirements come on the scene only in relation to ends that have themselves been endorsed in some way by the agent, as ends that it would be good or desirable to achieve.’ 8 Foot (2001). See also two works that Foot cites, Hursthouse (1999) and Thompson (1995). 9 Many psychoanalytic understandings of narcissism see it as involving, or even devolving from, deep insecurities, fears or worries of unworthiness, failure, and being bad. Typically, it is also held that assurances of goodness from other people are needed to keep these other thoughts at bay. My point about narcissism is, however, well served by these psychoanalytic understandings, provided that, as they typically do, they hold that narcissists also have (often insistent) beliefs that they are good. To this extent, narcissism is understood as involving serious conflicts. My thanks are owed to the psychoanalyst and anthropologist Elizabeth Hegeman for discussion of narcissism and ethnocentrism. 10 This is a revised version of my paper, ‘On the Intelligibility of Bad Acts’, delivered at the ‘Values, Rational Choice, and the Will’ Conference held at the University of Wisconsin-Stevens Point, April, 2004. It is a development of my paper, ‘Raz on the Intelligibility of Bad Acts’ delivered to the ‘Conference on the Moral Philosophy of Joseph Raz’ (March 22–24, 2002, Columbia University) and published in Wallace RJ et al (eds) (2004) Reason and Value: Themes From the Moral Philosophy of Joseph Raz, Oxford University Press, Oxford. My thanks are owed to discussants at these conferences, especially David Chan, Joseph Raz, and Michael Smith, and to Jonathan Adler, Stephanie Beardman, Jeffrey Blustein, Ruth Chang, Robert Daly, Christopher Gowans, Elizabeth Hegeman, and Erik Schmidt.

References Anscombe E (1957) Intention. Blackwell, Oxford Anscombe E (1958) Modern moral philosophy. Philosophy 33:1–19 Aristotle (1984) Rhetoric. In: Barnes J (ed) The complete works of Aristotle. Princeton University Press, Princeton Arpaly N (2003) Unprincipled virtue. Oxford University Press, Oxford Augustine (1955) Confessions. Outler AC (trans). Westminster Press, Philadelphia Foot P (2001) Natural goodness. Oxford University Press, Oxford Freud S (1953) Recommendations to physicians practicing psycho-analysis. In: Strachey J et al. (trans). The standard edition of the complete psychological works of Sigmund Freud Vol XII. Hogarth Press, London Hursthouse R (1999) On virtue ethics. Oxford University Press, Oxford McKay T, Nelson M (2006) Propositional attitude reports. In: Zalta EN (ed) The Stanford encyclopedia of philosophy, Winter Edition. http://plato.stanford.edu/archives/win2006/entries/propattitude-reports/

140

M. Stocker

Murdoch I (1993) Metaphysics as a guide to morals. Penguin Press, New York Plantinga A (1969) De re et de dicto. Nous 3:235–258 Raz J (1997) Incommensurability and agency. In: Chang R (ed) Incommensurability, incomparability, and practical reason. Harvard University Press, Cambridge MA Raz J (1999) Engaging reason: on the theory of value and action. Oxford University Press, Oxford Schon DA (1983) The reflective practitioner. Basic Books, Boston Smith M (1994) The moral problem. Blackwell, Oxford Stocker M (1976) The schizophrenia of modern ethical theories. Journal of Philosophy 76:453–466 Stocker M (1979) Desiring the bad. Journal of Philosophy 79:738–53 Stocker M (1981) Values and purposes: the limits of teleology and the ends of friendship. Journal of Philosophy 81:747–765 Stocker M (2004) Raz on the intelligibility of bad acts. In: Wallace RJ et al. (eds) Reason and value: themes from the moral philosophy of Joseph Raz. Oxford University Press, Oxford Thompson M (1995) The representation of life. In: Hursthouse R et al. (eds) Virtues and reasons. Oxford University Press, Oxford Vogler C (2002) Reasonably vicious. Harvard University Press, Cambridge MA Wallace RJ (2001) Normativity, commitment, and instrumental reason. Philosophers’ Imprint 1(3). http://www.philosophersimprint.org/001003

Chapter 8

After Anscombe David K. Chan

Abstract In this paper, I trace the concept of intention in theories of action after the publication of Anscombe’s monograph Intention. Bratman has the most viable account of intention that successfully refutes intention skeptics, but it cannot do the work in ethics that some moral philosophers use it for. This is shown by the problems in using intention to make moral distinctions in double effect cases. I argue that Anscombe’s concept of intention is broader than Bratman’s and I suggest that the aspects of intention that make her concept suited for ethics should be attributed to a concept distinct from intention, namely, an intrinsic desire for the end, corresponding with Aquinas’s account of intentio. A role for desire in ethics dovetails perfectly with the Aristotelian virtue ethics that Anscombe has been credited with reviving in her writings. Keywords Anscombe · Aquinas · Desire · Double-effect · Intention This is an essay in the history of philosophy of action after Anscombe, but the history is relatively recent. G.E.M. Anscombe’s monograph Intention (1957) was the work of an important and highly respected British philosopher. Looking back on the concepts of intention developed in action theory in the last half century, there seem not to be a specifically Ancombean account of intention.1 Anscombe was a historically informed philosopher in a period of mostly analytical works in the philosophy of action. Yet even in recent times, when philosophers considered to be moral psychologists have drawn on historical figures such as Aristotle, Hume, and Kant, insufficient attention has been given to Anscombe’s Intention. The title of my paper is meant to call to mind the work (MacIntyre 1981) of another philosopher credited with attempts to resolve contemporary ethical problems by rediscovering historical traditions in philosophy. Alasdair MacIntyre has also sometimes been credited with the recent revival of virtue ethics as a competitor

D.K. Chan Department of Philosophy, University of Wisconsin – Stevens Point, Stevens Point, WI 54481, USA, Tel.: 1(715) 346-3735, Fax: 1(715) 346-4215 e-mail: [email protected] D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

141

142

D.K. Chan

to the formerly dominant Kantian and utilitarian theories. I think that Anscombe’s influence in this revival has also not been fully recognized. In this paper, I will suggest that her philosophy of action is best understood by linking it with her moral philosophy.2 In this paper, I will relate a story about the concept of intention and its role in theories of action after Anscombe’s Intention, doing so in a way that typifies developments after Anscombe, but in broad strokes that inevitably bypass many contributions to the corpus on intention. I will show that the most viable account of intention nowadays is different in important ways from Anscombe’s concept. But I will argue that two specific features of Anscombe’s work are worth preserving. First, the more analytic work on intention today continues to lack the kind of historical awareness that was typical of Anscombe’s writings. Second, Anscombe has more consideration for ethics than those who recently articulate concepts of intention. Should we then forget the last fifty years and replace later concepts of intention with Anscombe’s? I suggest instead that the lesson to learn from a study of Anscombe’s Intention is that the concept of desire should be given prominence alongside the concept of intention to constitute a theory of action that encapsulates what is best in moral psychology, Anscombe and after.

8.1 Anscombe’s Intention In Intention, §1, Anscombe set out for us the topic of her investigations thus: ‘Very often, when a man says “I am going to do such-and-such”, we should say that this was an expression of intention. We also sometimes speak of an action as intentional, and we may also ask with what intention the thing was done.’ It is ‘implausible,’ Anscombe wrote, ‘that the word is equivocal as it occurs in these different cases.’ Clearly, she was seeking a concept of intention that includes all the aspects of its usage in ordinary language, especially the following: a) Expressions of intention are like predictions in being concerned with the future. b) An agent has a special first-person authority regarding what he intends. c) A person’s intention is connected with his reason for acting. d) A person’s intentions can usually be read off his intentional actions. These four aspects provide a number of directions that philosophical investigation of intention could take. Although according to Anscombe, intention can be ‘a purely interior thing,’ she thought the first thing to consider were the actions that intention issues in (§4). And intentional actions are distinguished as actions to which a certain sense of the question ‘Why?’ is given application, the sense in which the answer, if positive, gives a reason for acting (§5). By §16, only a quarter of the way into her monograph, Anscombe had already reached conclusions as to what this sense of the question ‘Why?’ is. So what did she do in the remainder of Intention? I will return to this question later, after I have described views of intention after Anscombe.

8 After Anscombe

143

8.2 The Intention Skeptics In 1963, Donald Davidson published ‘Actions, Reasons, and Causes,’ in which he argued that a reason for acting comprises a belief and ‘pro-attitude’ pair with appropriate content regarding that action, and that the reason for which someone acts is the cause of his bodily movement. Given that intentional action is action done for reasons, intentional agency can be accounted for without reference to a concept of intention as an internal state of mind. Davidson’s paper3 marked a turning point in philosophy of action, making the causal theory of action the centerpiece in conceptualizing human action. Since then, the difference between intentional action and things that one does without agency was thought to lie in the causal antecedents of events brought about in acting. Work in the philosophy of action became concerned with identifying and describing what these causal antecedents are and how they relate to bodily movements and events outside the body. Intention was understood in terms of its role as one of these causal antecedents. But Davidson was a reductionist concerning intention when, in the 1963 paper, he called intention ‘syncategorematic’ or a kind of shorthand to describe the relation between desire and belief on the one hand, and intentional action on the other. What about ‘pure’ intentions that are formed and never carried out? In his 1978 paper, ‘Intending,’ intention was recognized as an ‘all things considered’ judgment in favor of a course of action that, like desire, is a species of pro-attitude. So conceived, it is merely an output of practical reasoning preceding intentional action. Davidson denied that the expression of an intention refers to an ‘entity, state, disposition, or event’ (see ‘Actions, Reasons, and Causes,’ p. 8 in Davidson 1980). He saw it as an advantage of his account of acting with an intention that it required no ‘mysterious act of the will or special attitude or episode of willing’ (see ‘Intending,’ p. 87 in Davidson 1980). Did he thereby show that the intentions in Anscombe’s account from six years before do not exist? Anscombe had allowed that intention could be a purely interior thing, while Davidson denied that the expression of an intention could refer to any such thing. While Anscombe’s approach to intention went by way of an account of intentional action, Davidson argued that such an account does not require an internal state of intending. Anscombe had said that intentional actions are those for which it is appropriate to ask for reasons for acting. But Davidson suggested that reasons for acting comprise of pro-attitude and belief, and that intention and desire are both species of pro-attitude corresponding to all-out and prima facie evaluative judgments respectively. As we have noted, Anscombe did recognize two other related aspects of intention: what the intention says about the future; and the agent’s authority as to what he intends to do. There is a discussion of non-observational knowledge and ‘practical knowledge’ in Anscombe’s Intention. Although Davidson’s ‘Intending’ included a discussion of beliefs regarding what one will do that are entailed in one’s intending, he did not discuss non-observational knowledge and first-person authority, which are characteristic hallmarks of the mental. Davidson’s view of mental events is that they are identical with physical events,4 and that intentions cannot be pure acts of will or special mental episodes. Before I report on what Anscombe wrote about

144

D.K. Chan

practical knowledge, I will describe Bratman’s view of intention, one that responds to both Anscombe and Davidson.

8.3 Functionalism and Future-directed Intentions Davidson’s work coincided with the development of psychology as a science. One issue in the philosophy of psychology that he wrote about concerned the possibility of law-like explanations in psychology. The reduction of mental states to physical states was an attractive one in psychology. In philosophy of mind, reductionism was popular as a way of keeping irreducible mental states to a minimum. The more parsimonious the theory of mind was, the easier for it to win acceptance. In the 1980s, functionalist accounts of mental states began to take hold in philosophy and psychology. Mental states were demystified without being reduced away by conceptualizing them in terms of extrinsic properties and relations. Each type of mental state would have characteristic roles in relation to other mental states and to behavior. Mental states may also be subject to different norms and standards of rationality by which they may be evaluated. Michael Bratman was a student of Davidson’s who, from the standpoint of functionalism, articulated a ‘Planning Theory’ of intention by showing that intention has a distinctive role in the formation and execution of future-directed plans of action that no other mental state, in particular desire, can fulfill (Bratman 1987). Therefore, intention and desire have each its own unique role in a theory of action. According to Bratman, intention is both an input into and output from practical reasoning. Seen as an output, intention differs from desire in that the agent who intends to act is committed to acting whether at once or at the appropriate time, and can be counted upon to so act if he does not change his mind, and nothing prevents him from acting. Seen as an input, the agent is under rational constraints to do further practical reasoning to find appropriate means in a timely fashion and to screen out intentions that are inconsistent with prior intentions. After Bratman, the concept of intention is no longer easily dismissed as theoretically redundant or unnecessary in theories of action. Moreover, there is now a need for discussions of practical rationality and choice to keep apart the roles of desire and intention. Bratman showed that in practical reasoning, prior intentions form a background against which choices are made, but the reasons that are weighed in practical reasoning are desire-belief reasons. An intention to act cannot by itself constitute a reason to act. Otherwise, an irrationally formed intention can make it rational to choose actions that are the means toward the wrongly intended goal (see pp. 24–6 in Bratman 1987). Bratman may have rescued intention from redundancy but he had not in fact recovered the interior thing that Anscombe had written of in Intention. His book actually began with criticisms of both Anscombe and Davidson. The mistake common to both is the ‘methodological priority of intention in action’ (see pp. 5–6 in Bratman 1987). That is, they tried to conceptualize intention in terms of its role in intentional action, rather than its role in practical reasoning about future actions.

8 After Anscombe

145

This is in fact not entirely appropriate as a criticism of Anscombe who had picked out, as one of the aspects of intention found in ordinary language, the fact that expressions of intention, like predictions, are concerned about the future. She then noted that expressions of intention are unlike predictions in that as descriptions of future events, they are justified not by evidence but by reasons for acting (§3). But she found this idea ‘mystifying’ and turned to a more ‘fruitful line of enquiry’: a discussion of intention in action. Bratman’s method straddled what for Anscombe were the fruitless and the fruitful. He was not just concerned with how intention (as output) is justified by a reason for acting, but with how intention (as input) provides a framework for practical reasoning. He was not just concerned with the role of intention in an action here and now, but with the role of intention in planning for an action in the future. Why did Anscombe not explore this possibility? In fact, Anscombe did discuss practical reasoning in Intention, but her discussion shed light on the concept of desire, not intention (§34 onwards). As mentioned earlier, Anscombe construed her topic of intention very broadly to include aspects that succeeding philosophers did not. What Bratman did was to narrow down the concept of intention to one specific distinguishing feature: its role as input into further practical reasoning. This provided the basis for a functionalist to separate intention from desire and to counter skepticism regarding intention. But the very broadness of Anscombe’s approach is what prevented her from taking Bratman’s path between the fruitless and the fruitful. Bratman’s and other similar accounts of intention have provided an effective conceptual tool for many directions of research in philosophy and cognitive psychology. There are still skeptics about intention around but philosophers who make use of the concept of intention no longer find themselves on the defensive. That the concept has become so useful in moral psychology is reasonable evidence of its correctness.5 Is intention the final piece of the puzzle for resolving philosophical problems in action theory and in ethics? I want to suggest that the success of Bratman’s account in action theory creates a problem, and that we need to recover something from Anscombe to address the problem. As I shall explain, intention cannot do the work in ethics that some intention-theorists and ethicists would like it to do.

8.4 Intention and Ethics Bratman’s book concluded with a chapter that applied his concept of intention to make sense of an ethical principle known as the Doctrine of Double Effect (DDE). As he stated it, DDE holds that it is sometimes permissible knowingly to bring about (or allow) some bad effect in the course of achieving some good end, even though it would not have been permissible to bring about (or allow) that bad effect as one’s intended means to that good end.6 Bratman tried to show how on his account, an expected side-effect may be chosen on the basis of practical reasoning without being intended. If correct, his account of intention would lend support to an intuitive ethical doctrine. To see if this is so, we must now discuss the DDE.

146

D.K. Chan

Let’s begin with moral theories that make use of the DDE. Traditionally, the doctrine is part of Aquinas’s natural law ethics, and Aquinas himself is thought to be the first to formulate it. In absolutist moral theories, the doctrine helps deal with moral conflict by justifying a choice of the lesser of two evils in some cases. Recent discussions of the DDE have attributed a Kantian rationale for the doctrine.7 Bratman’s application of his concept of intention in support of the DDE would fit in with this agenda. But I believe that there are objections to the appropriation of the DDE by deontologists to be found in the writings of both Aquinas and Anscombe. When Aquinas formulated what was thought to be the original version of the DDE, he was concerned to justify killing in self-defense in the context of a natural law ethics that deems life as a basic human good. He differentiated between acts that are intended and acts that are praeter intentionem (beside the intention). It would be wrong to act with the intention of hindering life, but killing in self-defense is permissible if killing is beside the intention (see II–II 64, 7 in Aquinas 1981). Aquinas, however, pre-dated Kant by several centuries, and his philosophy was a development of Aristotle’s. This tells us two things: he was no deontologist; and he could not have had the same conception of intention in mind as Bratman, since intention did not exist in Aristotle’s philosophy of action. Anscombe has been scathing in her criticism of ‘abuses’ of the DDE from the Seventeenth Century on (Anscombe 1961). Modern philosophers, she wrote, were using a perverse version of the DDE in which intention was ‘an interior act of the mind which could be produced at will’ by making a little speech to yourself! In Kantian ethics, the maxim of one’s action is to be evaluated using the Categorical Imperative (C.I.) to determine whether one has a duty to do it or to avoid it. The maxim is one’s subjective principle of action, characterizing it as an action of a certain type done with a certain motive or intention. Critics of Kant have pointed out that whether one’s action is permissible depends on how one’s maxim is formulated. If one can choose one’s intention at will, one can always characterize one’s action as a type that passes the C.I. test. To use Anscombe’s example, one could drop an atomic bomb on Hiroshima without intending to kill the people living there. Such abuse of the DDE amounts to self-serving sophistry, not the application of a respectable moral doctrine. Bratman’s account may be thought to lend the concept of intention an objectivity that avoids Anscombe’s criticism. Whether or not someone intends to do something is now a matter of whether he has a state of mind that plays the defining roles that intention plays in planning and action. The pilot who bombed Hiroshima cannot in all sincerity profess not to have an intention to kill the inhabitants. So far, so good. But suppose another pilot is bombing a North Korean nuclear reactor that is being used to make and store nuclear weapons that will be detonated in the South Korean capital where millions live, and a small number of innocent people are being held at the reactor as human shields. The pilot knows all this. Is he morally justified to bomb? If the DDE were going to justify doing the lesser of two evils in this case, it would have to distinguish this pilot’s action from that of the Hiroshima pilot. Despite Bratman’s efforts, I have argued elsewhere that on his account, foreseen side-effects in cases like this are intended.8 My argument is that an agent’s knowledge of a

8 After Anscombe

147

side-effect places him under a rational constraint to plan on the basis of his bringing about the side-effect. Bratman would consider the pilot who bombs the North Korean reactor to be a strategic bomber (SB) in contrast with the terror bomber (TB) who bombed Hiroshima. He would argue that a SB’s attitude towards the killing of innocents does not put him in a state of mind that plays the roles that typify intention. But whether SB is in such a state depends on whether he is certain that in destroying his target, he will have to bring about the deaths of the innocent humans. In order to show that the moral difference between SB and TB rests on a difference of intention, the example must be one where the probability of the bad effect must be held to be the same in both cases. If killing the human shields is unavoidable in the act of bombing the North Korean reactor, and knowing this, the pilot judges it best to go ahead with the bombing, then he is settled on killing them. Bratman holds that a SB is not under any rational constraint to find means to kill the innocent and to filter out options incompatible with killing them. But allowing the possibility of such options, e.g., doing something that would cause the human shields to be relocated, changes the example into one where it is possible for the pilot to bomb without killing the innocent, contrary to the hypothesis that such killing is unavoidable. We have to assume that the human shields will be moved only if the reactor is itself relocated. Then having settled on killing them as an unavoidable part of his plan to bomb the nuclear reactor, if the pilot can best identify the reactor by looking for signs of the human shields and guide his action by targeting them, it would violate instrumental rationality for him to not do that. Which is to say that he is in the state of mind with regard to killing the human shields that Bratman would call an intention. Thus, if bombing the North Korean reactor is justified, and bombing Hiroshima is not, Bratman’s account of intention would not support a version of the DDE that explains the moral difference between the two bombings. Do we then have a good reason to give up on Bratman’s account of intention? No, because as a conceptual account of one source of intentional agency, as an account of the difference between desire and intention, and as an account of practical rationality, Bratman’s theory has given intention an important place in action theory. What we should give up is the attempt to use the concept of intention to do the work in ethics that deontologists would like it to do. Now, it may be objected that no ethical theory, especially one with the influence and strengths of Kantian ethics, can be undermined simply because one conception of intention seems unsuited for the theory. Accordingly, we should instead either change the conception of intention to one that is better suited, or give up trying to connect concepts in action theory with concepts in moral theory. The latter option would be to give up on doing moral psychology. The former option would be to tailor our psychological concepts to fit our ethics. In another of Anscombe’s papers, we find an explanation why we should reject both these options. In ‘Modern Moral Philosophy,’ she argued, ‘it is not profitable for us at present to do moral philosophy . . . until we have an adequate philosophy of psychology’ (Anscombe 1958). Anscombe castigated the ethics of moral duty and moral obligation as orphaned survivors of a ‘law conception of ethics’ that is

148

D.K. Chan

no longer relevant today. She suggested that to describe actions and character in terms of the categories of virtues and vices is a better way of evaluating them than to describe them in terms of a special ‘moral’ sense of right and wrong, one that depends on the discarded law conception. But to have an account of what a virtue is and how it relates to virtuous action requires that we first have accounts of what a human action is, and of the concepts of motive, intention and desire. Hence, we cannot even begin to do ethics until we have worked out a sound philosophy of psychology. Anscombe’s argument, which is at the same time an argument for doing virtue ethics, called for the kind of work done by Bratman in his account of intention. The result of this work is not, however, what I think Anscombe would have wanted (or expected). In ‘Modern Moral Philosophy,’ Anscombe stated that no one would defend a view of intention in which one is said to intend any foreseen consequences of one’s action. But I did exactly that a few paragraphs earlier.9 Moreover, I claimed that this view of side-effects is supported by the best concept of intention available in the philosophy of psychology, namely Bratman’s account of intention. Is it possible to keep Bratman’s account, and also the DDE that it seems to undermine? I think so. The DDE as stated earlier assumes that if there is a moral distinction between the two bombers in our examples, there must be a difference in what they each intend. But it is not possible to show that one bomber intends to kill innocent people and the other does not. So we should give up the assumption. The moral distinction that the DDE captures, and which makes it so useful in justifying sometimes doing the lesser of two evils, is one that depends not on a difference in intention regarding foreseen side-effects, but a difference in desire. The pilot who is willing to drop an atomic bomb on a city of more than a million civilians with no military targets must either desire to kill civilians, or have too weak a desire not to kill them. The pilot who bombs a nuclear reactor that is being used to produce and store weapons that will kill millions, knowing that his action will kill a small number of innocent people, is doing something compatible with a strong desire not to kill innocent people. Given the alternatives, it may be rational for his desire not to kill to be outweighed by his desire to save the lives of millions. Although my account of the DDE that was published recently may initially seem to represent a break with the traditional version, there are some historical facts that count in its favor. First, a version of DDE in terms of desire will have a rationale in an Aristotelian virtue ethics, exactly the kind of ethics that Anscombe has helped to revive. Second, there is reason to think that my formulation of the DDE is closer to that of Aquinas. I shall explain this in the next section. Third, intention for Anscombe is a very broad concept, unlike Bratman’s concept, and may well include aspects that belong to desire. The broader concept may be applied in ethics in a way that the narrower concept may be unsuited for. If so, the ethical work is done by another concept besides Bratman’s intention, and it is my contention that desire is that very concept. This points to a distinction between practical reasoning by which intention is formed, and moral reasoning by which certain ends come to be desired, which is a distinction Anscombe was herself aware of in Intention.

8 After Anscombe

149

8.5 Intentio and Intention From where in the history of philosophy does the concept of intention originate? Aquinas formulates the DDE in terms of what is intended and what is beside the intention. But the word that is intention in translation is intentio in the Latin, defined as an act of the will in regard to the end. Intentio is distinguished from electio (usually translated as choice), which is an act of the will in relation to the means. And one can intend without having yet chosen.10 When Aquinas discusses intentio as an act of will, it is one of 12 or so stages in the process of human action that he distinguishes. The will is defined as a rational appetite, meaning that it is an inclination toward something extrinsic that is apprehended as a good by the intellect (see I 78, 1c and I–II 8, 1c in Aquinas 1981). The idea of a rational appetite comes from Aristotle, who sees rational beings as having a will, in addition to the natural appetite found in all living things and the sensitive appetite found in all animals as sources of movement. Aristotle calls the rational appetite proairesis, and actions that are motivated by this faculty are reflective of the agent’s character and can be either virtuous or vicious. What Aquinas does when he introduces intentio and other acts of will is to make finer distinctions regarding the stages that lead from the agent’s dispositions to act, through his deliberation, to his practical conclusion and action. This refinement does not change the Aristotelian connection between the will and the agent’s character. Intentio is an act of will and is used by Aquinas in his version of the DDE to make the point that some wrongful acts do not reflect badly on the agent’s character. We are told by Aquinas that human actions are to be evaluated in genere moris, i.e., under the description that is morally relevant, namely the description they have as willed. In other words, an act derives its moral character from the ends that the act is ordered to by the will (see I–II 1, 3 ad 3 in Aquinas 1981 and 2, 2 ad 13 in Aquinas 2003). When an act of killing an aggressor in self-defense is beside the intention, it is not as such an act willed by the agent. To say this is to say that it is not for the agent the object of the rational appetite. The concept of intention today is of a state of mind that motivates an agent to carry out the means to achieve his end. An intention is typically formed by making a rational decision after considering alternative actions that are evaluated in terms of the contributions they make towards the agent’s goals. It is true that the intended means can provide a goal for further practical reasoning. For example, if one intends to go to the library as a means of doing research, one is required to figure out how to get to the library. But one may also intend means, e.g., to swing one’s arm, that requires no further reasoning. Such means are not intended as ends. In fact, intention is the state of mind that is needed to motivate one to do actions that one does not want as an end, but that one is committed to doing solely as means to things that one really wants.11 This concept of intention found in modern philosophy of action is quite different from Aquinas’s concept of intentio, in at least two important ways. First, intention is formed through practical reasoning in which the standard is that of instrumental rationality: the suitability of the action as a means to achieve given ends. Unlike

150

D.K. Chan

Aristotle’s proairesis and Aquinas’s intentio which is an act of will, an intention to do what is most likely to satisfy given ends has no necessary implications for the ethical evaluation of the action. An agent need not want to do what he intends to do as a means to his ends. The pilot who bombs the North Korean nuclear reactor intends to kill the innocent civilians he knows to be there, but he could very well have a strong desire not to kill them. His act of killing is praeter intentionem, and does not impugn his character even though it is something he intends to do. Second, intentio for Aquinas is an act of will, and therefore a form of rational appetite. It is a desire of the kind that has a role in means-end reasoning, and that is characteristic of rational beings. But if something is not the object of appetite at all, it cannot be the object of intentio. We saw earlier that Bratman’s concept of intention is defined by contrast with the concept of desire. Such a concept of intention would not be found among the acts of will that Aquinas distinguishes as stages in a process leading from the agent’s character to action. There is also an account of rational appetite in Anscombe’s Intention. Not only did she discuss intention, but some twenty pages of her 94-pages monograph are devoted to a discussion of ‘wanting’ examined by way of an analysis of practical deliberation. I have already mentioned that non-observational knowledge of the agent regarding what he is going to do is an aspect of intention in ordinary usage that Anscombe had noted. At the beginning of §33 of Intention, we find the statement: ‘The notion of “practical knowledge” can only be understood if we first understand “practical reasoning”.’ The question Anscombe puzzled over is how reasoning could lead to action. Her Aristotelian answer was that ‘whatever is described in the proposition that is the starting-point of the argument must be wanted in order for the reasoning to lead to any action’ (§35). Thus, she finds that an integral part of her project on intention involved an examination of what a desire is, and an account of its role in practical reasoning and action. What is Anscombe’s view of desire? Noting that someone cannot just desire something without being able to say what he wants it for, Anscombe concluded that the thing wanted must have a desirability characterization that reveals what about it that the agent conceives to be good, so that no further question of ‘What do you want that for?’ could arise. But what the agent conceives to be good may not be what is really good. Anscombe did not go further into questions of what the human good is, and whether there are moral premises that govern practical reasoning. But she remarks that if practical reasoning has anything to do with ethics, it will be about such questions rather than about the right way to choose the means to whatever ends people happen to have (§41). Anscombe here marks the distinction that I have noted between standards of instrumental rationality, and accounts of human ends used for ethical evaluation. Given my interpretation of Aquinas and the DDE, it is no longer surprising that Anscombe paid so much attention to desire in Intention. We know that Anscombe wanted her philosophy of psychology to lay the foundation for an account of virtues. Intention that is distinct from desire, such as Bratman’s concept of it, cannot do ethical work, as the norms that govern intention are those of instrumental rationality. By contrast, Aquinas’s intentio is of ethical relevance because it is a form of

8 After Anscombe

151

rational appetite. Unfortunately, Anscombe’s account of intention is too broad to be just about intentio, or just about intention. We have seen that philosophers of action writing after Anscombe did not pay enough attention to the implications of philosophy of psychology for ethics. The concept of intention found nowadays in the philosophy of action is useful for understanding what goes on in the mind when someone acts for a reason, i.e., when appropriate means are chosen and carried out to achieve given ends. But to do the work that Anscombe envisages for philosophy of psychology in ‘Modern Moral Philosophy,’ moral philosophers need to equip themselves with a proper account of desire, and not just with a theory of intention.

8.6 Conclusion: Desire and Intention In the years after Anscombe’s Intention, philosophers of action and moral psychologists have shown somewhat greater interest in the concept of intention than in the concept of desire. Part of the explanation for this is that intention was confronted with skeptics such as Davidson, and it was thought to require more conceptual examination than the relatively unproblematic and commonplace concept of desire. It is not surprising that philosophers of action have produced more work on intention than on desire. But if we accept Anscombe’s stricture that a sound philosophy of psychology is needed to do ethics, would there not be a demand, especially among virtue ethicists, for moral psychologists to turn their attention to desire? With the renaissance in virtue ethics, such a demand has indeed arisen, and I am hopeful that more interest in the concept of desire, and greater attention to the ethical implications of research on the motivational sources of human agency, will result in Anscombe’s clarion call being answered.

8.7 Postscript: The Wrong Turn After Anscombe If the importance and relevance of Anscombe’s monograph Intention for philosophy of action and ethics today is what I have claimed for it here, why have many of the ideas that I’ve mentioned here been relatively neglected and overlooked? Let me suggest some explanations. I will begin with the influence of Wittgenstein’s philosophy on Anscombe’s writing. Anscombe is, of course, famous for her translations of some of Wittgenstein’s most important works, including the Philosophical Investigations. Her monograph on Intention is replete with references to Wittgenstein and her writing is in the form of remarks and aphorisms in the style of Wittgenstein, without chapters or sections organized around topics. It is easy to overlook an important point or to miss the relation between different parts of the book. A follower of the later Wittgenstein will also be skeptical about the private aspects of the mental, preferring to describe the mental in terms of the conventions that govern our use of mental terms in ordinary language. It is not surprising that Anscombe’s approach to intention is to

152

D.K. Chan

describe the different aspects found in the ordinary usage of the term. We saw that she presented us with a very broad concept, and it is not clear that all the different aspects relate to the same concept. She does not try to do what, for instance, Bratman does in delineating a narrow and precise meaning of intention. But philosophers working towards a degree of conceptual clarity not found in ordinary speech will recognize that we do not make a clear and sharp distinction between intention and desire in ordinary language. To take an example that I have written about, it seems equally appropriate to say ‘I want to take medicine to get well’ or to say ‘I intend to take the medicine to get well, but I don’t want to.’12 Anscombe’s reliance on ordinary talk prevented her from seeing that intention itself, unlike rational appetite or desire, may not be suited to do ethical work. The kind of philosophical analysis that arose from the middle of the 20th Century has not always been a good thing in that it provided answers to philosophical problems devoid of any sensitivity to the history of ideas. Anscombe’s historically-aware writing could have worked to her disadvantage at a time when such writing was not valued. Added to that was the beginning of a shift away from Britain and towards the United States as the new center of philosophy in the English-speaking world, or what we call Anglo-American philosophy. Cambridge, where Anscombe worked, and Oxford declined in relative importance, and philosophers such as Davidson, working in the New World found it easier to have their ideas discussed and studied. Recent developments in psychology and cognitive science have also favored the views of Davidson and Bratman. I have mentioned that Davidson’s philosophy developed at the same time that psychology was rapidly shaping itself into a science, and he in fact had worked closely with experimental psychologists.13 Bratman’s account of intention has gained broad acceptance partly because his concept has been modeled in cognitive science and used by scientists at Stanford in attempts to build thinking machines that can form intentions (Bratman et al. 1988). Finally, I think philosophers of action miss what is important in Anscombe’s Intention simply because it has been misread. This is partly due to the breadth of her concept of intention, and the style of her writing. It can also be attributed to the fact that it is difficult to see the world as Anscombe saw it at the time that she wrote Intention. The ideas in the text look different for one who reads it after reading Davidson. And the ideas can look different again after one has read both Davidson and Bratman. I too am unlikely to read it exactly the way that Anscombe meant for it to be read. For my reading is based on work I have done on the DDE and on desire, which requires that desire and intention be distinguished as different concepts with different uses. I think the distinction I make is implicit in Anscombe’s work, but I do not think she saw it. Why then should I, or any philosopher today, return to Anscombe’s work? What are the advantages of re-reading Anscombe’s Intention? If we seek a proper understanding of where our concepts in moral psychology came from and how they developed, it is crucial to return to a text that is not only historically important, but also historically sensitive. It is historically important because without the publication of Anscombe’s Intention, the last fifty years of philosophy of action may not have been the same. Many important issues in the philosophy of action

8 After Anscombe

153

were framed in Anscombe’s monograph. We find, for instance, questions about the belief-conditions on intention, about side-effects of action, about the proper description of actions, about the individuation of actions, about mental causes, and about practical reasoning, to name a few. The historical sensitivity of Anscombe’s ideas provides an important counterbalance to the wealth of analytical writing in philosophy of action that we have seen during the intervening years. Most importantly, moral philosophers should read Anscombe’s work. As she pointed out, it is important for virtue ethicists, writing about reasons for action, character and practical wisdom, to have a sound philosophy of psychology before doing ethics. Just as deontologists need to work on their concept of intention, virtue ethicists need to say what desire is and work on the psychological aspects of character and the rational choice of ends that comprise human good.14

Notes 1 Apparently, the favorite references to Anscombe’s text are to the famous quote that ‘The primitive sign of wanting is trying to get’ (§36); and to her view on the individuation of action (§26), something that Davidson calls the ‘accordion effect’ whereby answers to the ‘Why?’ question generate further descriptions of the same action. But it is obvious that neither passage discusses the concept of intention itself. 2 This is basically the theme of her seminal essay ‘Modern Moral Philosophy’ (Anscombe 1958). 3 The articles by Davidson referred to in this paper have all been reprinted in his collected essays (1980). 4 ‘Mental Events’ in Davidson (1980) sets out Davidson’s version of the identity theory, known as ‘anomalous monism’. 5 Bratman has written that ‘the proof of the pudding is in the eating.’ 6 Certain conditions, such as the requirement of proportionate reason, have to be met for the DDE to apply. 7 The most prominent work is from Quinn (1993). 8 I summarize a much longer and detailed argument from Chan (2000) in the next paragraph. 9 As I will argue below that Anscombe’s concept of intention is broader than Bratman’s concept, I am not actually contradicting Anscombe in what I have said about the Strategic Bomber. 10 Aquinas (1981) discusses intentio at I–II 12. I summarize here a much more subtle discussion from an unpublished paper, ‘Aquinas, Intention, and the Doctrine of Double Effect.’ In that paper, I consider the claim in Finnis (1991), pp. 7–8, ‘there is intention only when and insofar as one settles on something as a means of achieving an end . . . so there is no intending until there is adoption of means, typically by choice.’ I show that intentio and electio are distinct acts of the will even when they have the same action as object. 11 In Chan (2004), I argue against the view that agents are motivated to bring about the means to their ends by a state of ‘extrinsic’ desire. 12 I argue in Chan (2004) that the first statement is incorrect and that if I want to take medicine, I can only want it for its own sake. 13 Some of this work is reflected in Davidson et al. (1957), and the essay ‘Hempel on Explaining Action’ (1976), reprinted in Davidson (1980). 14 I wish to thank the audience at a plenary session of the Conference on Values, Rational Choice, and the Will held at the University of Wisconsin – Stevens Point in 2004, at which a version of this paper was presented and helpful suggestions and comments were received.

154

D.K. Chan

References Anscombe GEM (1957) Intention. Blackwell, Oxford Anscombe GEM (1958) Modern moral philosophy. Philosophy 33:1–19 Anscombe GEM (1961) War and murder. In: Stein W (ed) Nuclear weapons: a Catholic response. Sheed & Ward, New York Aquinas (2003) De Malo. Regan R (trans). Oxford University Press, Oxford Aquinas (1981) Summa Theologica. 2nd rev edn. Christian Classics, Westminster MD Bratman M (1987) Intention, plans, and practical reason. Harvard University Press, Cambridge MA Bratman M, Israel D, Pollack M (1988) Plans and resource-bounded practical reasoning. Computational Intelligence 4:349–355 Chan DK (2000) Intention and responsibility in double effect cases. Ethical Theory and Moral Practice 3:405–434 Chan DK (2004) Are there extrinsic desires? Nous 38:326–350 Davidson D (1980) Essays on actions and events. Oxford University Press, New York Davidson D, Suppes P, Siegel S (1957) Decision-making: an experimental approach. Stanford University Press, Stanford Finnis J (1991) Object and intention in moral judgments according to Aquinas. The Thomist 55: 1–27 MacIntyre A (1981) After virtue. University of Notre Dame Press, Notre Dame Quinn W (1993) Actions, intentions, and consequences: the doctrine of double effect. In: Morality and action. Cambridge University Press, Cambridge

Part IV

Practical Reason

Chapter 9

The Limits of Teleology Crystal Thorpe

Abstract The teleological thesis says that an essential feature of intentional action is that it is goal-directed or teleological. To say that a person’s action is goal-directed or teleological is to say that a person performs that action in order to promote some goal, end or desire. The Humeans accept the teleological thesis, as do some rationalists. My aim is to cast doubt on this thesis. More than two decades ago, Michael Stocker argued that teleological considerations are inappropriate for understanding acting out of or from character. Here, I deepen Stocker’s original insight by isolating another class of actions, in addition to those done out of or from character, that are intentional yet not teleological. I focus on intentional deliberative action, that is, action that is guided by reasoning, reflection and careful consideration. Intentional deliberative action differs from action done out of character in that it is deliberative. I demonstrate that a species of intentional deliberative action – what I call valuedriven action – is not teleological. Keywords Goal-directed action · Hume · Intentional action · Moral reasoning · Teleological

9.1 Introduction To say that a person’s action is goal-directed or teleological is to say that a person performs that action in order to promote some goal, end or desire. An essential feature of intentional action is that it is goal-directed or teleological. Call this the teleological thesis. The Humeans accept the teleological thesis, as do some rationalists. On the Humean side, the teleological thesis serves as the starting point of an influential argument, most clearly articulated by Michael Smith, purporting to show that the Humean theory of motivation is correct (see pp. 36–61 in Smith 1987). Furthermore, C. Thorpe e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

157

158

C. Thorpe

the teleological thesis seems to undergird the belief-desire theory of action.1 On the rationalist side, Kant claims that every action has an end, and that there are objective ends that we have a duty to have.2 My aim is to cast doubt on the teleological thesis. I am in good company here. More than two decades ago, Michael Stocker argued that teleological considerations are inappropriate for understanding some forms of intentional action (Stocker 1981). He stressed what he called the archeological – acting out of or from – and contrasted it with the teleological – acting in order to.3 On Stocker’s view, acts of friendship and courage are performed out of or from character rather than in order to promote the goal of friendship or courage. The kind of intentional action that Stocker focuses on – acting out of or from – isn’t deliberative.4 The kind of intentional action that I will focus on, in contrast, is deliberative. Deliberative action is guided by reasoning, reflection and careful consideration. We deliberate about how to deal with terrorism, whether to make a career move, how to maneuver through an unmarked detour, and what to do about world hunger. My claim is that there is a kind of intentional deliberative action – what I call value-driven action – that is not teleological. To say that a person’s action is value-driven is to say that a person performs that action because she takes it to be right, good, required by duty or supported by reasons. If I am right, I will have deepened Stocker’s original insight by isolating another class of actions, in addition to those done out of or from character, that are intentional yet not teleological. My arguments, if successful, also serve as a back door argument against the Humean. Because the teleological thesis undergirds both the Humean theory of motivation and the belief-desire theory of action, casting doubt on it casts doubt on these other theories as well.

9.2 Intentional Deliberative Action Seems to be Teleological Some actions, like grimacing in the mirror or whistling a tune, are ready counterexamples to the teleological thesis. In standard cases, they are intentional yet not performed in order to further some goal. What if we narrow the class of intentional action to intentional deliberative action? Is the teleological thesis true with respect to intentional deliberative action? It seems to be. I think the reason for this is that a certain picture of practical reasoning – call it the standard view – has molded our thinking about deliberative action. The standard view says that there are two kinds of practical reasoning – instrumental reasoning and non-instrumental reasoning. Instrumental reasoning is teleological. When a person reasons instrumentally, she focuses on how to promote her goals, ends or desires. Non-instrumental reasoning is not teleological. When a person reasons non-instrumentally, she figures out which goals, ends and desires are worth having in the first place, based upon some substantive, non-instrumental standard.

9 The Limits of Teleology

159

It will be helpful to take a closer look at both instrumental and non-instrumental reasoning. Many of our goals, ends and desires are complex. When we reason instrumentally, we break these complex goals down into subgoals and figure out how to further each subgoal. Say, for instance, that my neighbor’s car won’t start, and that my goal is to give her a jump. With this goal in mind, I reason instrumentally, figuring out that it can be broken down into subgoals: the goal of getting the cars in the right position, the goal of hooking up the cables, etc. I then figure out what I have to do in order to further each subgoal. When I use instrumental reasoning to break a complex goal into subgoals, I am simply adopting new goals that are instrumentally related to my initial goal. Before deliberating, although I had the goal of giving my neighbor a jump, I didn’t have a goal of getting the cars in the right position. I acquired the latter only after reasoning instrumentally. Acquiring goals through instrumental reasoning is different from acquiring goals through non-instrumental reasoning. In the latter case, I acquire goals on the basis of some substantive, non-instrumental standard. For example, while watching my neighbor in despair, I come to see, through a non-instrumental process of reasoning and reflection, that a world in which people help their neighbors is better than a world in which they don’t. Based on this, I decide to help my neighbor. Once I adopt the goal of helping my neighbor, I reason instrumentally to figure out how to promote it. Notice that on the standard view, when we reason practically we are either figuring out which goals, ends or desires are worth having based on some substantive standard, or we are figuring out how to promote the goals, ends and desires that we already have. On the standard view, therefore, practical deliberation itself is structured around goals. So long as we think that the standard view is right, it is natural to think that intentional deliberative action is goal-directed. Look at it this way: When a person reasons non-instrumentally, she evaluates goals, ends and desires. When she acts on the basis of non-instrumental reasoning, she adopts, abandons, or retains goals, ends and desires on the basis of some substantive, non-instrumental standard. These actions are not teleological. Notice, however, that these actions are mental acts. In standard cases, when a person performs a mental act, the only change she produces in the world is a change in her psychology.5,6 In contrast, when a person reasons instrumentally, she figures out how to promote her goals, ends and desires. When she acts on the basis of instrumental reasoning, she either (1) adopts, abandons or retains some goal, end or desire, as when she adopts sub-goals; or (2) does something other than adopt, abandon or retain a goal, end or desire, like get her body in motion towards some goal or allow something to happen so as to further some goal. Actions like (1) and (2) are teleological. They are done with an eye towards promoting a goal, end or desire. Actions like (1) are mental acts, whereas actions like (2) are not. Actions like (2), unlike those like (1), include actions that produce change in the broader world outside the agent’s psychology. Included here are acts like hauling bricks, transferring money from one account to another, winking, signing a contract, allowing a cashier’s mistake to go unnoticed, cracking a nut, etc.

160

C. Thorpe

I take it that the teleological thesis is not about mental acts. Rather, it is about those things that we intentionally do that produce change in the broader world outside our psychology. If this is right, the teleological thesis says that an essential feature of intentional action that produces change in the outside world is that it is goal-directed. In this paper, I am concerned with the teleological thesis with regard to intentional deliberative action. So here, I am concerned with the truth of the following claim: An essential feature of action that is intentional, deliberative, and productive of change in the outside world is that it is goal-directed. On the standard view, this claim is true. This is because on the standard view, the kind of action at issue – action that is intentional, deliberative, and productive of change in the outside world – is always based, at least in part, on instrumental reasoning. If non-instrumental reasoning has any role to play in guiding this sort of action, its role is limited to one of setting the parameters for action by setting goals, ends and desires.

9.3 Alternative Conceptions of Practical Reasoning If the standard view is right, the teleological thesis is true, at least with regard to intentional deliberative action. Is the standard view right? I do not think it is. However, rather than argue that the standard view is mistaken, I will argue that there are other plausible conceptions of practical reasoning. On these views, intentional deliberative action isn’t necessarily teleological. So long as these views are plausible, it is likewise plausible to think that the teleological thesis is false. A conception of practical reasoning is an account of the kinds of practical reasoning that exist and how each functions. Because it is beyond question that instrumental reasoning exists, is teleological, and functions as described above, instrumental reasoning is a part of any conception of practical reasoning worth taking seriously. What differentiates one conception of practical reasoning from another, therefore, is whether it includes non-instrumental forms of reasoning, and if so, how those non-instrumental forms of reasoning are characterized.7 On the standard view there are only two kinds of practical reasoning – instrumental reasoning and substantive reasoning about goals. I will discuss two alternative conceptions of practical reasoning. On the first, which I call the moral reasoning mediated by rules view, there are four kinds of practical reasoning – instrumental reasoning, substantive reasoning about goals, moral reasoning mediated by rules, a second kind of non-instrumental reasoning, and practical judgment, yet a third kind of non-instrumental reasoning. On the second alternative conception that I will discuss, which I call the moral reasoning unmediated by rules view, there are three kinds of practical reasoning – instrumental reasoning, substantive reasoning about goals, and practical judgment. On both alternatives, when a person reasons noninstrumentally, she either figures out which goals, ends or desires to adopt, abandon or retain, or she comes to believe that she ought to perform some action because that action is right, good, required by duty or supported by reasons. In standard cases, when a person does something because she believes it to be right, good, required

9 The Limits of Teleology

161

by duty or supported by reasons, she produces change in the world. On each of the alternative conceptions I discuss, therefore, non-teleological forms of reasoning, independently of instrumental reasoning, can guide the sort of action at issue in the teleological thesis. Clearly, to make good on these claims, I must show that the reasoning that leads a person to see that she ought to do something because it is right, good, required by duty or supported by reasons is not merely a combination of substantive reasoning about goals and instrumental reasoning. I will argue that it is a kind of non-instrumental, non-teleological reasoning that cannot be accommodated by the standard view. My position, therefore, is that on these alternative conceptions of practical reasoning, a person’s action can be intentional, based on reasoning, and productive of change in the outside world, all without being based on instrumental reasoning, not even in part. In order to show that there are forms of non-instrumental reasoning that not only can guide the sort of action at issue in the teleological thesis, but can do so independently of instrumental reasoning, it will help to say a bit more about instrumental reasoning. First, it is useful to think about instrumental reasoning in terms of a means-end syllogism.8 The major premise tells us to promote some goal, end or desire. The minor premise tells us which actions promote the goal, end or desire given in the major premise. The conclusion, an inference from the major and minor premises, tells us to perform those actions that promote the goal, end or desire given in the major premise, that is, to perform those actions specified in the minor premise. We might represent this pattern of reasoning as follows: 1. Promote end E. 2. Action A promotes end E. 3. Perform action A.9 Second, instrumental reasoning must be distinguished from the brute figuring of means to ends or ways of realizing ends. When a person reasons practically, she figures out what she ought to do. When a person reasons theoretically, she figures out what is the case. Clearly, instrumental reasoning is a kind of practical reasoning. When a person runs a means-end syllogism, she figures out what she ought to do in order to promote her goals, ends or desires. The brute figuring of means to ends or ways of realizing ends, in contrast, is a kind of theoretical reasoning. When a person calculates, she figures out which actions, as a matter of fact, further or realize a given goal, end or desire. I stress this point, because calculation is so tightly bound up with instrumental reasoning that we often mistake it for instrumental reasoning itself, assuming that whenever we calculate we reason instrumentally. As a general rule we can say that whenever a person reasons instrumentally, that is, whenever a person runs a meansend syllogism, she calculates. After all, in standard cases theoretical calculation is the sort of reasoning a person uses to establish the minor premise.10 It is a mistake, however, to assume that whenever a person calculates, she reasons instrumentally or does so in the service of instrumental reasoning. An economist, for instance, can figure out whether tax cuts better distribute the wealth without having any interest

162

C. Thorpe

in acting on this information. Moreover, as we will see, calculation can be used in the service of non-instrumental as well as instrumental reasoning. If we insist on equating instrumental reasoning and calculation, these points get lost.

9.4 The Moral Reasoning Mediated by Rules View As I said, on the moral reasoning mediated by rules view, there are four kinds of practical reasoning – instrumental reasoning, substantive reasoning about goals, moral reasoning mediated by rules, and practical judgment. I begin by focusing on moral reasoning mediated by rules. Later, I will turn to practical judgment. There are many kinds of moral reasoning that are mediated by rules. The feature these forms of reasoning have in common is that each is mediated by one or more moral rules. These forms of reasoning differ, among other things, in terms of (1) the number of rules that mediate, (2) the content of the mediating rule(s), (3) whether the mediating rules are taken to be justificatory or simply heuristic and (4) whether the rules mediate directly or indirectly. As to (1) and (2), on a certain interpretation of Kantian ethics, moral reasoning is mediated by one rule, the categorical imperative. Likewise, on a certain interpretation of utilitarianism, the principle of utility is the only rule that guides moral reasoning. According to other views (and on other interpretations of Kantian ethics and utilitarianism), a plurality of rules guide moral thinking. For example, on some understandings of Christian ethics, the Ten Commandments guide moral reasoning, on some interpretations of Kantian ethics, a set of absolute duties guides moral reasoning, on most understandings of act utilitarianism, rules of thumb guide moral reasoning, and according to rule utilitarianism, a set of rules guides moral reasoning.11 As to (3), the categorical imperative and the principle of utility are typically taken to be justificatory, while the rules of thumb utilized by the act utilitarian are merely heuristic. Lastly, as to (4), when a person engages in moral reasoning that is directly mediated by rules, she reasons her way to what she ought to do in particular cases by appealing to a rule. Sometimes, the easiest or most efficient way to determine which actions accord with the rule is to appeal to a decision procedure. The categorical imperative procedure is an example of a decision procedure, advocated by some Kantians, that a person can use to determine whether a proposed action accords with the categorical imperative. Utilitarian calculation is another example of a decision procedure that critics of utilitarianism typically assume a person uses to determine whether a proposed action accords with the principle of utility.12 In contrast, when a person engages in moral reasoning that is indirectly mediated by a rule, that rule plays an indirect or regulative role in her reasoning about what she ought to do in particular cases. Here, the person does not appeal to the rule each time she acts; rather, she appeals to it periodically, when she reflects on what sort of person she ought to be, what general sorts of conduct she ought to engage in, what sorts of motives and dispositions she ought to have and act from, etc.

9 The Limits of Teleology

163

On particular occasions, she acts from these permissible character traits, motives, dispositions, etc.13

9.4.1 Direct Mediation Without a Decision Procedure: The Simple Case Let’s first consider a simple case of moral reasoning mediated by rules. Consider a hypothetical moral theory, according to which there is only one moral rule – ‘Truthtelling is right.’ Implicit in this rule is the idea that people ought to do what is right, and therefore ought to tell the truth. We can therefore restate the rule as follows: ‘Tell the truth.’ Assume that Jack accepts this theory, and that Jack has used indelible ink to make notes in a library book. When Jack returns the book, the clerk notices, and asks Jack whether he defaced the book. Clearly, there is a process of reasoning through which Jack can reason his way from the rule ‘Tell the truth’ to the conclusion to answer the clerk in the affirmative. It might go like this: He says to himself ‘Tell the truth.’ He then examines the facts of the case before him, and considers possible ways of responding to the clerk. He could answer in the affirmative, in the negative, change the subject, or remain silent. The former, he reasons, is the only option that amounts to telling the truth. He therefore concludes that he ought to answer in the affirmative. In the simple case, it is useful to think about moral reasoning directly mediated by rules in terms of a rule-case syllogism.14 The major premise tells us to follow some rule, such as ‘Tell the truth.’ The minor premise tells us which actions fall under that rule in the particular case at hand. The conclusion, an inference from the major and minor premises, tells us to perform the actions that fall under the rule in the particular case at hand. We might represent this pattern of reasoning as follows: 1. Follow rule R. 2. In this case, action A falls under rule R. 3. Perform action A. This is strikingly similar to the form that a person’s reasoning takes when she reasons instrumentally. The main difference is that here the major premise tells us to follow a rule, whereas in a means-end syllogism the major premise tells us to promote a goal, end or desire. How much of a difference does this make? Earlier, I said that instrumental reasoning is intimately bound up with the brute figuring of means to ends or ways of realizing ends. Moral reasoning mediated by rules, at least in the simple case, is similarly bound up with this sort of theoretical calculation. This, however, may not be immediately apparent, because when Jack comes to see that answering in the affirmative counts as an instance of telling the truth, he does not come to see that answering in the affirmative is a means to the end of telling the truth. What he comes to see is that a particular action – saying ‘yes’ in the situation sketched above – falls under or counts as a certain type of action – the truth-telling type. Coming to see this, however, involves the same sort of reasoning

164

C. Thorpe

that a person uses when she comes to see that a particular action is a way of realizing some end. Jogging, for instance, is typically understood to be a way of realizing the end of exercise rather than a means to exercise. Coming to see that jogging is a way of realizing the end of exercise, however, involves coming to see that jogging falls under or counts as a certain type of action – the exercise type. If this is right, then the same sort of theoretical calculative reasoning that is bound up with instrumental reasoning is similarly bound up with moral reasoning mediated by principles. In standard cases, a person calculates in order to establish the minor premise in both a means-end syllogism and in a rule-case syllogism. Does this mean that when a person runs a rule-case syllogism her reasoning takes the same form that it takes when she runs a means-end syllogism? It is tempting to think so. I think we should resist this temptation. The major premise of a rule-case syllogism constrains our actions to those of a certain type. In contrast, the major premise of a means-end syllogism tells us to promote an end. To claim that a person’s reasoning takes the same form when she runs a rule-case and a means-end syllogism is to claim not only that the same sort of reasoning figures in each, but also that (1) the major premise of a rule-case syllogism can be restated in terms of a premise that tells us to promote an end, and (2) the conclusions of the resulting syllogism are in all cases the same as the conclusions of the original rule-case syllogism. I will not attempt to show that this cannot be done. Rather, I will illustrate the difficulties involved by looking at a simple example. Consider Jack’s rule ‘Tell the truth.’ The major premise of a rule-case syllogism that makes reference to this rule can be states as follows: Follow the rule that says ‘Tell the truth.’ Can we replace this premise with a premise in terms of an end to be promoted, and do so in such a way that the resulting means-end syllogism yields the same conclusions, in all cases, as the original rule-case syllogism? The most natural way to do this is to specify the major premise as follows: ‘Promote the end of truth-telling.’ With this in place, Jack asks himself, ‘What action, in this case, promotes the end of truth-telling?’ In the case described earlier, answering in the affirmative promotes truth-telling. In another case, however, Jack’s telling one lie might promote truth-telling by preventing ten other lies from being told. So long as Jack reasons from a major premise that makes reference to the rule ‘Tell the truth,’ he does not reach conclusions that recommend telling lies, even in those cases when doing so prevents other lies from being told. This is because the rule ‘Tell the truth’ constrains his actions to those of the truth-telling type, whereas promoting the end of truth-telling does not.15 Is there a way to specify the major premise of a means-end syllogism so as to constrain the agent’s actions to the truthtelling type? One strategy is to distinguish Jack’s acts of truth-telling from everyone else’s acts of truth-telling, and state the major premise in terms of the promotion of the former. This, however, falls prey to the same problem as the earlier suggestion. In promoting this end, Jack may lie once now, so as to prevent himself from lying twice in the future. Another strategy is to stick to the original formulation of the major premise (‘Promote the end of truth-telling’) but add a provision to the effect that all actions performed by Jack that are violations of truth-telling are to be assigned an infinite negative weight in the promotion of truth-telling.16

9 The Limits of Teleology

165

The last strategy seems to do the trick, until we notice that in some cases Jack may limit his interactions with others, willfully deceive himself, or fail to acquire certain kinds of knowledge, all in the service of promoting the end of truth-telling. Subsuming these same cases under the rule ‘Tell the truth’ is unlikely to yield the same results. Perhaps it will help if we specify the major premise as follows: ‘Promote the end of telling the truth here and now.’ It is hard to imagine cases in which lying, and doing so here and now, promotes the end of telling the truth, and promotes it here and now. Nonetheless, it is easy to imagine cases in which cultivating a disposition to tell the truth, and doing so here and now, promotes the end of telling the truth, and promotes it here and now. Although promoting the end of telling the truth here and now is not likely to lead a person to tell lies, it doesn’t necessarily constrain a person’s actions to the truth-telling type. Lastly, consider this as a specification of the major premise: Promote the end of acting on the rule ‘Tell the truth’. From what I said earlier, it should be clear that a person might violate this rule once to prevent either herself or others from violating it multiple times in the future. Clearly, specifying the major premise in this way does not constrain a person’s action to the truth-telling type. None of this, of course, conclusively shows that the major premise of a rulecase syllogism cannot be stated in terms of an end to be promoted. However, it does point us towards the idea, which perhaps was apparent from the beginning, that a rule-case syllogism is distinct from a means-end syllogism. A person reasons differently when she runs one rather than the other. When a person runs a meansend syllogism, she reasons instrumentally. Her reasoning is teleological, that is, it is directed towards figuring out how to promote an end. Actions performed on the basis of instrumental reasoning are likewise teleological. In contrast, when a person runs a rule-case syllogism, she does not reason instrumentally. Her reasoning is not directed towards figuring out how to promote an end. Rather, she figures out how to constrain her action, in the particular case, in the way recommended by the rule. This is a species of non-instrumental, non-teleological reasoning. Actions done on the basis of it are likewise non-teleological. Furthermore, in standard cases, actions done on the basis of this sort of non-instrumental reasoning produce change in the outside world. Jack, for instance, in answering in the affirmative, produces change in the outside world. If all of this is plausible, then it looks as if in the simple case, at least, moral reasoning mediated by rules, independently of instrumental reasoning, can guide the kind of action at issue in the teleological thesis. And if this is right, it is plausible to think that the teleological thesis is false, even with respect to intentional deliberative action.

9.4.2 The Complex Case The simple case is simple because there is only one moral principle. Now consider a moral theory in which there are several non-lexically arranged principles.17 Here, there is potential for conflict among the principles. In such cases, a person must

166

C. Thorpe

resort to practical judgment, a non-instrumental form of practical reasoning, to discern which principle is overriding. For instance, say that Jack accepts a moral theory according to which there are two non-lexically arranged principles – ‘Tell the truth’ and ‘Help others.’18 Now, consider a case in which these principles conflict. Say that Jack is hiding a man, one who will be framed and executed for a murder he did not commit. Inevitably, the authorities come to Jack, and ask where the man is. Here, Jack says to himself, ‘I ought to follow the rule that says “Tell the truth” and I ought to follow the rule that says “Help others.” In this case, however, I cannot do both. I will therefore follow the rule that says “Help others,” because saving a person from being unjustifiably executed is clearly more important than telling the truth.’ He concludes, ‘In this situation, helping others involves lying to my interrogators.’ It is clear that Jack’s reasoning does not take the form of a rule-case syllogism in which the major premise can be stated: Follow the rules that say ‘Tell the truth’ and ‘Help others.’ If this were this case, he would be led to an impasse, because these rules, when brought to bear on the case, recommend contradictory actions. Rather, Jack figures out what he ought to do by (1) discerning which rule is overriding in the case at hand, and (2) running a rule-case syllogism in which the major premise makes reference to the overriding rule. Discerning which principle is overriding requires practical judgment, a non-instrumental, non-teleological form of practical reasoning that is typically taken to require a special kind of sensitivity, that is, a certain way of seeing things. A person acquires this sensitivity through moral training and immersion in the moral life. If this is right, then in the complex case, a person uses two forms of noninstrumental reasoning. First, she uses practical judgment to figure out which rule plays a role in the major premise of a rule-case syllogism. Second, she uses moral reasoning mediated by principles – which is simply subsuming cases under rules – to figure out what she should do in the case at hand. In standard cases, acting on the basis of practical judgment and moral reasoning mediated by principles, as described above, produces change in the outside world. In standard cases, therefore, practical judgment and moral reasoning mediated by principles, independently of instrumental reasoning, can guide the sort of action that is at issue in the teleological thesis. So long as this is a plausible way to think about practical reasoning, it is likewise plausible to think that the teleological thesis is false.

9.4.3 Direct Mediation with a Decision Procedure The categorical imperative procedure and utilitarian calculation are examples of decision procedures a person may use to determine whether a proposed action accords with the categorical imperative and the principle of utility, respectively. In each case, a person’s reasoning follows the pattern of a rule-case syllogism. In the first case, the major premise can be stated as follows: Follow the rule that says ‘Act so that

9 The Limits of Teleology

167

the maxim of your action can be willed as universal law’. In the second case, the major premise can be stated as follows: Follow the rule that says ‘Maximize the good.’19 In the first case, the minor premise tells us which actions, in the case at hand, have maxims that can be universalized. In the second case, the minor premise tells us which action, of the alternatives available to the person in the case at hand, maximizes the good.20 In both cases, the calculation required in order to establish the minor premise is spelled out stepwise, in the form of a decision procedure. In the first case, a person is to state the maxim of the proposed action; imagine a world in which this maxim holds as universal law; and determine whether there is something contradictory about such a world. If there is not, the minor premise specifies that action and the conclusion of the syllogism is a recommendation to perform the action specified by the minor premise. In the second case, a person is to calculate the good produced by each alternative open to her in the case at hand, and determine which alternative produces the most good. The one that produces the most good is specified by the minor premise, and the conclusion of this syllogism is again a recommendation to perform the action specified by the minor premise. Moral reasoning mediated by a rule and a decision procedure has the same structure as moral reasoning mediated by a rule without a decision procedure. Everything I said about moral reasoning mediated by rules without a decision procedure therefore holds for moral reasoning mediated by rules with a decision procedure.

9.4.4 Indirect Mediation Earlier, I said that in standard cases, actions done on the basis of moral reasoning mediated by rules produce change in the outside world. Now, let’s look at a nonstandard case. A person can figure out what sorts of character traits, dispositions, motives, etc. to have on the basis of moral reasoning mediated by rules. Here, her reasoning follows the pattern of a rule-case syllogism, with a moral rule standing in for ‘R’ in the major premise. The conclusion is a recommendation to cultivate certain character traits, dispositions, motives, etc. When a person acts on the basis of this reasoning, that is, when she cultivates certain character traits, dispositions or motives within herself, she does not produce change in the broader world outside her psychological state. Assume that included among these permissible dispositions are dispositions to engage in characteristic forms of practical reasoning. A person who cultivates such dispositions will, for the most part, engage in those forms of practical reasoning. She may not use them every time she deliberates – most notably, she is not likely to use them when she reflects upon what sort of character traits, dispositions, motives, etc. she ought to cultivate. Nonetheless, when she does use them, her reasoning is indirectly mediated by a moral rule. This is because the disposition to engage in these characteristic forms of practical reasoning is chosen on the basis of some moral rule.

168

C. Thorpe

As we have seen, in the simple case, moral reasoning directly mediated by a moral rule is non-instrumental. Actions based on such reasoning are not teleological and, in standard cases, produce change in the world. We have also seen that in the complex case, moral reasoning directly mediated by rules has two non-instrumental components – practical judgment and moral reasoning mediated by rules. Actions based on such reasoning are not teleological and, in standard cases, produce change in the world. In both the simple and complex cases, actions based on moral reasoning mediated by rules are counterexamples to the teleological thesis. These actions are intentional and deliberative, produce change in the outside world, and yet are not guided by instrumental reasoning, not even in part. In contrast, moral reasoning that is indirectly mediated by a moral rule is not necessarily non-instrumental. In fact, whether actions based on moral reasoning that is indirectly mediated by a moral rule are non-teleological depends upon the moral rule that guides a person in cultivating her character traits and dispositions. Actions based on moral reasoning that is indirectly mediated by moral rules, therefore, are not ready counterexamples to the teleological thesis. I will therefore set them aside.

9.5 The Moral Reasoning Unmediated by Rules View On the moral reasoning unmediated by rules view, there are three kinds of practical reasoning – instrumental reasoning, substantive reasoning about goals and practical judgment. The main difference between this view and the moral reasoning mediated by rules view is that on this view, moral reasoning is practical judgment, through and through. On the moral reasoning mediated by rules view, Jack takes truth-telling and helping others to be relevant to his decision as to what he ought to tell the authorities because he accepts a moral theory that tells him, by putting forth rules, that truthtelling and helping others matter morally. What his theory doesn’t tell him is which matters more in particular cases. This is where practical judgment comes in. On the moral reasoning unmediated by rules view, in contrast, moral reasoning is practical judgment, through and through. In particular cases, a person discerns, through practical judgment, what is relevant to what she ought to do, which feature, if there is more than one, matters more, and finally, what she ought to do. This is all done without appealing to moral rules. Let’s take a closer look at practical judgment. Say that Jack is hiding the same man under the same circumstances described above. When Jack judges what he ought to do in this case, he takes truth-telling and helping others to be relevant to his decision because of the details of the case – an innocent person’s life hangs on whether Jack lies or tells the truth. Likewise, in this case Jack takes helping others to matter more than truth-telling because of the details of the case – an innocent man will be executed if Jack tells the truth. Finally, Jack reasons that in this case, lying

9 The Limits of Teleology

169

to the authorities is a way of helping others and that he therefore ought to lie. Does instrumental reasoning get him to the idea that lying is a way of helping others and that he therefore ought to lie? If it does, then it isn’t so clear that practical judgment, independently of instrumental reasoning, can guide the kind of action at issue in the teleological thesis. Jack does not reason instrumentally when he comes to see that in this case lying is a way of helping others and that he therefore ought to lie. If he did, his reasoning would take the form of a means-end syllogism, with ‘Promote the end of helping others’ standing in for the major premise. As I have described things however, Jack doesn’t come to see that he ought to promote the end of helping others. Rather, he comes to see that he ought to do what counts as helping others, that is, he ought to constrain his actions to the help others type. If this is right, then his reasoning at this juncture follows a pattern that looks more like a rule-case syllogism than a means-end syllogism. The pattern, however, differs from a rule-case syllogism in that the major premise does not specify a rule. Rather, it specifies an action type to be performed in this particular case. There is room for quibbling here. Some may be inclined to say that this bit of syllogistic reasoning is distinct in kind from practical judgment. If this is right, I need to amend the list of the kinds of reasoning there are on the moral reasoning unmediated by rules view. Others may be content to say that this bit of syllogistic reasoning, along with the reasoning that Jack uses to discern which features matter and which matter more, all fall under the head of practical judgment. Regardless of which way we go here, the point remains that no aspect of Jack’s reasoning is instrumental. Jack reasons non-instrumentally when he discerns which features matter, he reasons non-instrumentally when he discerns which features matter more, and he reasons non-instrumentally when he runs this special kind of syllogism. If this is right, then on the moral reasoning unmediated by rules view, non-instrumental reasoning, independently of instrumental reasoning, can guide the kind of action at issue in the teleological thesis. If follows that so long as the moral reasoning unmediated by rules view is plausible, it is likewise plausible to think that the teleological thesis is false. Before leaving this section, I should point out that something like the moral reasoning unmediated by principles view is held by some particularists, most notably Dancy, and has been attributed to Aristotle and even W.D. Ross.21 It may seem odd to include Ross here. Proponents of this view, however, do not necessarily think that there are no moral rules. On Ross’ view, although there are moral rules, they do not play a significant role in moral reasoning.22 Dancy, unlike Ross, thinks that there are no moral rules (Dancy 1993). It is also worth mentioning that the moral reasoning unmediated by rules view leaves room for many interpretations of practical judgment. As I have discussed practical judgment, it has an inferential, syllogistic component. It needn’t, however. Practical judgment can be conceived as an immediate, non-inferential act of moral judgment. So long as we are willing to use the term ‘reasoning’ to refer to noninferential acts of judgment, the moral reasoning unmediated by rules view can accommodate practical judgment conceived in this manner.

170

C. Thorpe

9.6 Conclusion When a person intentionally does something that produces change in the outside world, and she does it on the basis of instrumental reasoning, her action is teleological. She acts in order to promote some goal, end or desire. When a person intentionally does something that produces change in the outside world, and she does it on the basis of instrumental reasoning and substantive reasoning about goals, her action is likewise teleological. In contrast, when a person intentionally does something that produces change in the outside world, and she does it on the basis of moral reasoning that is mediated by rules, her action is not teleological. She does not perform that action in order to promote some goal, end or desire. Rather, she does it because she believes that it is right, good, required by duty or supported by reasons. Here, her action is value-driven rather than goal-directed. Likewise, when a person intentionally does something that produces change in the outside world, and she does it on the basis of moral reasoning that is not mediated by rules, her action is value-driven rather than goal-directed. So long as the moral reasoning mediated by rules view and the moral reasoning unmediated by rules view are plausible conceptions of practical reasoning, it is plausible to think that the teleological thesis is false.23

Notes 1 Here, I have in mind the belief-desire theory of action that we find in Davidson’s early work. See Davidson (2001). 2 See Kant’s Metaphysics of Morals (see pp. 516–517 in Gregor 1996). In the Critique of Practical Reason (see p. 167 in Gregor 1996), Kant claims ‘it is indeed undeniable that every volition must have an object and hence a matter; but the matter is not, just because of this, the determining ground and condition of the maxim. . .’ The corresponding page numbers in the German Academy of Sciences edition of Kant’s collected works are 385–386 (Metaphysics of Morals) and 34 (Critique of Practical Reason). 3 The sort of teleology that Stocker argues against is the view that ‘only an act’s purposes, goals, ends, desires, and the like – taken as psychological “entities” – are critical for understanding action’ (see p. 748 in Stocker 1981). 4 If we take a person’s character to be a complex of dispositions, including dispositions to engage in characteristic forms of practical reasoning, acting out of or from character may sometimes be deliberative. Stocker doesn’t address this possibility. In fact, in describing acts done out of character, Stocker says that they are ‘done “naturally,” simply expressing one’s character.’ He then notes the parallel between his claim and ‘Zen claims about acting naturally and directly’ (see p. 761 in Stocker 1981). This suggests that Stocker doesn’t see acting out of or from character as being deliberative. 5 She may produce a change in her neurophysiology as well. 6 Clearly, if a person goes on to further a newly adopted end, she may produce change in the outside world. And any change she produces in the outside world is, in a sense, the result of her having adopted that end. It does not follow from this, however, that when a person adopts an end, she produces change both in her psychological state and in the outside world. The act of adopting an end, considered independently of any other actions a person may (or may not) perform as a result of having adopted that end, merely produces a change in the person’s psychology. What if one of the causal upshots of adopting an

9 The Limits of Teleology

171

end is that it leads a person to become anxious and break out in a cold sweat? Here, it looks like adopting an end causes both a change in a person’s psychology and a change in the broader world outside her psychology. My response is to say that this isn’t a standard case. In standard cases, when a person adjusts her beliefs, desires, intentions, etc., the only change she produces in the world is a change in her psychology. 7 Instrumentalism, a conception of practical reasoning I will not discuss, does not include non-instrumental forms of reasoning. On this view, all practical reasoning is instrumental. If instrumentalism is correct, then the teleological thesis, with respect to intentional deliberative action, is true. 8 I am not committed to the idea that the best or only way of formalizing instrumental reasoning is as a means-end syllogism. I could make the same points without conceiving of it in this way. Conceiving of it in this way, however, gives me a framework for articulating the differences between instrumental reasoning, on the one hand, and theoretical calculation, moral reasoning mediated by rules and moral reasoning unmediated by rules, on the other. I am indebted to David Chan for pressing me on my use of syllogistic logic. 9 This, of course, isn’t the only way of representing the premises and conclusion of a means-end syllogism. For example, we could state the major premise as: ‘Further end E,’ ‘Satisfy desire D,’ ‘Promote goal G in the most efficient way,’ ‘Further end E in the best and easiest way,’ etc. 10 In some cases, the connection between means and end is so simple and immediate that no calculation is required. For instance, if the end is opening my car door, unless there is some sort of problem with the door – it is locked or frozen shut – I needn’t calculate in order to figure out how to open it. 11 I do not include Ross’ theory of prima facie duties here because Ross claims that these principles do not, for the most part, play a role in moral reasoning. For a defense of this interpretation of Ross, see Dancy (1983, 1991). 12 I know of no act utilitarian who advocates utilitarian calculation as a moral decision procedure. As has been made clear by critics, this procedure alienates the agent from her projects and commitments, including her relationships with others. Instead, act utilitarians insist that utilitarianism is to be understood as a criterion of rightness rather than as a decision procedure, and they advocate some other decision procedure. Smart suggests that we follow rules of thumb, and Railton suggests that we cultivate those dispositions that, over time, lead us to maximize the good. See Smart (1973) and Railton (1984). 13 Some Kantians and consequentialists use this kind of strategy in response to the general worry that the Kantian and consequentialist agent is alienated from her projects and commitments, including her relationships with others. See Railton (1984), Herman (1981) and Baron (1984). 14 I am not committed to the idea that the best or only way of formalizing moral reasoning mediated by rules is in terms of a rule-case syllogism. See note 8 above. 15 Whether this constraint is absolute is a question I will leave unanswered. 16 Nozick, whose discussion I am indebted to here, discusses this sort of move in his discussion of moral constraints and moral goals. See Nozick (1974). Special thanks to David Copp for pointing me to this. 17 Because moral theories with multiple lexically arranged principles are generally thought to be too rigid to be plausible, I will not consider them here. 18 Again, these are restatements of the rules ‘Truth-telling is right’ and ‘Helping others is right.’ 19 Again, these are restatements of the rules ‘It is right to act so that the maxim of your action can be willed as universal law,’ and ‘It is right to maximize utility.’ 20 If two or more actions produce maximal good, both will be specified by the minor premise. 21 For a reading of Aristotle that seems particularly likely to fall under the moral reasoning unmediated by rules view, see McDowell (1996). 22 On Ross’ view, judgments about which features matter in a particular case, and which matter more, carry with them a commitment to underlying general rules. For a defense of this interpretation of Ross, see Dancy (1983, 1991). 23 I am indebted to David Chan and David Copp, who provided extremely useful criticisms and suggestions on an earlier draft of this paper.

172

C. Thorpe

References Baron M (1984) The alleged moral repugnance of acting from duty. Journal of Philosophy 81:197–220 Dancy J (1983) Ethical particularism and morally relevant properties. Mind 92:530–547 Dancy J (1991) An ethic of prima facie duties. In: Singer P (ed) A companion to ethics. Blackwell, Malden MA Dancy J (1993) Moral reasons. Blackwell, Malden MA Davidson D (2001) Essays on actions and events. Oxford University Press, New York Gregor M (trans, ed) (1996) The Cambridge edition of the works of Immanuel Kant: practical philosophy. Cambridge University Press, Cambridge Herman B (1981) On the value of acting from the motive of duty. Philosophical Review 90:359–382 McDowell J (1996) Deliberation and moral development in Aristotle’s ethics. In: Engstrom S, Whiting J (eds) Aristotle, Kant and the Stoics: rethinking happiness and duty. Cambridge University Press, Cambridge Nozick R (1974) Anarchy, state and utopia. Basic Books, New York Railton P (1984) Alienation, consequentialism and the demands of morality. Philosophy and Public Affairs 13:134–171 Smart J (1973) An outline of a system of utilitarian ethics. In: Smart J, Williams B (eds) Utilitarianism: for and against. Cambridge University Press, Cambridge Smith M (1987) The Humean theory of motivation. Mind 96:36–61 Stocker M (1981) Values and purposes: the limits of teleology and the ends of friendship. Journal of Philosophy 78:747–765

Chapter 10

Motivation to the Means Stephen Finlay

Abstract Rationalists including Nagel and Korsgaard argue that motivation to the means to our desired ends cannot be explained by appeal to the desire for the end. They claim that a satisfactory explanation of this motivational connection must appeal to a faculty of practical reason motivated in response to desire-independent norms of reason. This paper builds on ideas in the work of Hume and Donald Davidson to demonstrate how the desire for the end is sufficient for explaining motivation to the means. Desiring is analyzed as having motivation towards making the end so, which is analyzed as engaging in mental activity aimed at facilitating that end. I conclude that it is constitutive of an agent’s desiring an end that he is motivated towards what he believes to be means. Keywords Desire · Humean theory of motivation · Instrumental motivation · Practical reason I desire a soda, so I feed a vending machine. What could be more mundane? But some perceive a difficult problem in explaining this motivation to the means: a problem for the orthodox Humean Motivation-by-Desire Principle, which holds that all motivation of action – causal influence towards activity exerted by minds1 – requires and is produced by conative psychological states like desire and aversion. How does a desire for one thing motivate action of doing something else? If this motivational transfer cannot be explained satisfactorily, there is license for attributing the motivation to sources other than desire, such as rational acceptance of imperatives of reason. Thomas Nagel, Christine Korsgaard, and others (henceforth ‘rationalists’) have thereby sought to find an Achilles heel in the Humean position.2 Here I argue that the Motivation-by-Desire Principle is not vulnerable to this attack. I first explain the objection, which focuses on the role of normative reasons in instrumental motivation, then examine the Davidsonian response analyzing reasons in terms of desires and beliefs, finding it unsuccessful. Building on moves by Hume and Donald S. Finlay Mudd Hall of Philosophy, University of Southern California, 3709 Trousdale Parkway, Los Angeles, CA 90089, USA, Tel.: 1-213-821-2219, Fax: 1-213-740-5174 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

173

174

S. Finlay

Davidson, I argue that a closer examination of what desiring is enables us to defeat the objection.

10.1 The Explanatory Problem Suppose I am thirsty and desire soda. Believing that in order to satisfy this desire I must insert coins in the vending machine, I do so. Somehow, the combination of desire for soda and belief about the means yields motivation to that means. The Motivation-by-Desire Principle requires that I am motivated to the means by some desire. The obvious candidate is already onstage: the desire for soda. But how does desire for the end motivate me to the means? Nagel writes: It must be realized that the case does require an explanation. Upon reflection, it can seem mysterious that thirst should be capable of motivating someone not just to drink, but to put a coin in a slot. Thirst by itself does not motivate such technical undertakings; an understanding of currency and the protocol of vending machines is necessary. But when these factors have been added to the explanation, we still lack an account of how they combine with the thirst to produce action. (1970: 33)3

The Motivation-by-Desire Principle seems to require that in the presence of a belief that the means to E is M, the motivational influence of desire for E is extended to M. But the interaction of belief and desire to yield such an output involves a mechanism that has yet to be explained.4 At first glance explanation seems easily accomplished, and at least the three following accounts can be offered. (1) Motivation to the means is produced by the desire for the end; human psychology happens to be so structured that the combination of end-desire and means-belief yields such motivation. (2) It is produced by a separate desire for the believed means; we are psychologically constituted such that the combination of end-desire and means-belief motivates not (directly) action, but means-desire, which then motivates action directly. (3) We antecedently possess a dispositional desire to take the means to our desired ends, whatever those ends and means may be. Natural selection would favor creatures possessing this additional desire, so its prevalence would be unremarkable. These options leave much to be explained, with their vague appeals to human psychological structure. But the missing details are for cognitive scientists, not philosophers, and are philosophically uninteresting. So what is the philosophical problem supposed to be? It is not the absence of a neurological or psychological story as such that troubles rationalists, but the absence of a story that makes room for reasons and normativity. Desiring an end is widely believed to provide or entail having a reason for motivation to the means. Recognizing such reasons doesn’t entail being motivated to the means, but does (absent conflicting reasons) entail that one ought to be and would be if rational. Even some neo-Humean champions of the Motivation-by-Desire Principle concur; failure of motivational extension from ends to means is a (or the) manifestation of culpable practical irrationality.5 In being motivated to the means to our desired ends, we therefore sometimes act

10 Motivation to the Means

175

as rational agents motivated by recognition of reasons.6 Rationalists contend that the Motivation-by-Desire Principle cannot accommodate this important dimension of motivation, and that therefore we need an alternative, rationalist, theory of motivation. The natural Humean rejoinder pioneered by Donald Davidson and developed by Michael Smith is to reconcile motivation by reasons with motivation by desire by analyzing reasons in terms of desires and beliefs. Giving the reason why someone acted is a form of explanation, and actions can be explained (according to the Motivation-by-Desire Principle) by citing desire-belief pairs: might reasons then simply be such pairs?7 The desire-based explanations (1)–(3) above would then qualify as instances of motivation by reasons. I shall argue that while this Davidsonian strategy cannot meet the objection, an alternative can. This strategy conflates two different senses of ‘reasons’.8 Appealing to my desires and beliefs as the states causally responsible for my action is to provide explanatory reasons or reasons why I acted, just as observing metal fatigue can provide an explanatory reason for why a plane crashed. But citing these psychological states does not provide normative reasons or reasons why I should act;9 we can have normative reasons where we lack relevant belief (and arguably desire), and we can lack them where we have both belief and desire. Smith responds that with regard to the motivation of action, the relevant kind of reasons must be motivating reasons (1987: 37–41, 1994: 100). Motivation involves causation of action, and can occur in the absence of normative reasons. Motivating reasons are thus causes of action,10 and the Davidsonian solution is back in business, as the causes of action are plausibly desire-belief combinations. However, this fails to parry the thrust of the rationalist challenge. Nagel’s point is that as rational agents, we are capable of being motivated by normative reasons. Normative reasons are sometimes also motivating reasons, when we recognize and respond to their authority.11 This is impossible for Smith, who places normative and motivating reasons in distinct ontological categories: ‘motivating reasons’ are psychological states, whereas normative reasons are propositional (facts), as is widely agreed.12 Therefore, while an agent ‘may have a motivating reason to do what he has a normative reason to do,’13 the normative reason can never be his motivating reason. Rationalists find this gulf between motivation and normativity unsatisfactory; surely we can be motivated by normative reasons. I think we must concede this. We might concede that motivation of action is a form of causation and that normative reasons cannot by themselves cause or explain action.14 The causally relevant psychological state is that of accepting the normative reason. It might seem then that what motivates is this psychological state and not the reason, its content. But the concept of motivation can accommodate both claims; both the state and its content motivate us, in different but related ways. To say that the content (reason) motivates can be to say in part that cognition of it causes certain behavior. This provides an answer to Smith’s objection that we can be motivated by reasons even where no relevant normative reasons exist; given what it means to be motivated by a reason, it is possible to be motivated by a reason that doesn’t exist.

176

S. Finlay

Smith’s error is to identify the motivating state as the motivating reason. Insofar as something motivates us qua reason, it does so by appearing at least minimally normative to us.15 Motivating reasons thus belong in the category of normative rather than explanatory reasons.16 The rationalist challenge retains its force: rational agents are disposed to be motivated to the means by recognition of normative reasons. This failure of the Davidsonian solution does not, however, entail the failure of all Humean analyses of motivation by reasons in terms of desires. Rationalists maintain that ‘acceptance of reasons’ is a belief-like yet intrinsically motivational state. But if normative reasons are true propositions, they could be the contents of the relevant beliefs, while the source of their normativity (status as reasons) is their connection with desires.17 The motivation might then derive entirely from those desires. Motivation by reasons would then be identical with a form of desire-belief causation.18 The rationalist case against this strategy invokes the contingency of psychological causation. In each desire-based explanation (1)–(3) above, motivation to the means is accounted for by hypotheses about human psychology. But psychological facts are contingent and so, Korsgaard writes, ‘it is perfectly possible to imagine a sort of being who could engage in causal reasoning . . . that would point out the means to her ends, but who was not motivated by it.’19 Worse, we can imagine beings with alternative motivational connections, extending (e.g.) motivation from ends to actions believed counter-productive to them. We might say that these people have reasons that are different from ours, but rationalists reasonably contend that this is counter-intuitive. Means-end motivation is self-evidently rationally appropriate, while these alternatives are not. The Humean could stipulate that motivation arising from abnormal causation doesn’t count as reason-based, so that I only qualify as acting on my reasons if my action is normal (i.e., if in light of my belief, I can reasonably expect it to satisfy my desire). But what he apparently cannot do is accommodate the role of this unique rational appropriateness in his tale of the motivational connection: If [such psychological devices] do not yield adequate explanations in the peculiar case [of deranged behavior], there is reason to believe that their analogues are not the basis of intelligibility in the normal case. The analogous hypotheses seem to fill the motivational gap in the normal case only because they are not actually needed to make the behavior intelligible, whereas in the abnormal case, where something more obviously is needed, they do not succeed. That leaves us, if we do not wish to be arbitrary, with the task of dividing the intelligible connections from the unintelligible ones and explaining why the former work and the latter do not. (Nagel 1970: 34)

The proper motivational outputs are not rendered ‘intelligible’ (or, more appropriately, ‘rational’)20 simply by being normal; they are privileged in some other significant way. While motivation theory has no obligation to explain the normativity or rationality of this connection, it is obliged to accommodate its role in instrumental motivation, which is where rationalists contend that the Motivation-by-Desire Principle fails. If our disposition to means-motivation is merely a contingent psychological fact, on a par psychologically with alternative dispositions, then it is merely a fortunate accident. A person might recognize the rational appropriateness

10 Motivation to the Means

177

of end-means motivational transfer, yet be psychologically constituted so as to be unable to effect it. But surely we are not so impaired. It is not merely a fortuitous accident that we are disposed to be motivated to the means; as rational agents, we are capable of being motivated by the recognition of the rational appropriateness of such motivation. The alleged failure of the Motivation-by-Desire Principle is thus to explain motivation to the means in a way that accommodates our status as rational agents who can act in response to our cognition of normative requirements such as those from instrumental reasons. Rationalists exploit the problem as a vulnerability in the Humean conception of agency, pushing for a substantive faculty of practical reason, cognizant of desire-independent rational principles and yielding motivation in response. On this account, a rational agent is sometimes motivated to the means not by the desire for the end or by any other desire, but directly by recognition of a normative requirement. This is an audacious attempt to beard the lion in his den; if even action aimed at satisfying desire draws its motivation from a source other than desire, then the Motivation-by-Desire Principle must be rejected and there is good reason to suspect that moral, prudential, and other kinds of actions can also be motivated without desire.

10.2 Hume and Davidson on Motivation to the Means Can we find any trace of an answer to the rationalists in the classic exponents of the Motivation-by-Desire Principle? Hume and Davidson seem both to adopt positions from which a response can be constructed; they appear to deny the existence of any gap between desire for the end and motivation to the means. This is the solution I shall defend after examining these precedents. In discussing the ‘love of gain,’ Hume maintains that anyone who desires this end necessarily has motivation to the believed means. Because of the superior force of this passion, he writes, There is no passion . . . capable of controlling the interested affection, but the very affection itself, by an alteration of its direction. Now this alteration must necessarily take place upon the least reflection; since ‘tis evident, that the passion is much better satisfy’d by its restraint, than by its liberty . . . (1978: 492)

By ‘direction’ of a passion he means its ‘tendency to action’ (1978: 382) or motivational effect. For a passion to ‘alter its direction’ is therefore for it to acquire a different motivational effect: towards the means. This alteration is brought about by beliefs about means, since it is reason that ‘discovers the connexion of causes and effects, so as to afford us means of exerting any passion.’ (1978: 459).21 Hume therefore looks to the desire for the end as a sufficient source of motivation to the means rather than postulating any further desire. What is provocative here is the claim that this alteration ‘must necessarily’ occur; Hume denies the very possibility of failure in motivational extension to the means. However, he says nothing to clarify

178

S. Finlay

the grounds or nature of this alleged necessity, so refraining from speculation, I turn to Davidson. Davidson states as a necessary condition for a ‘primary reason’ (both a cause and justification for action) the following principle: R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent towards actions with a certain property, and a belief of the agent that A, under the description d, has that property. (1980: 5)

He desires to turn on the light, so performs an action he believes a means to this end, flipping the switch. Is there any need for further explanation of how this action was motivated? No, because the action, which can be given indefinitely many veridical descriptions, is performed under the description, ‘turning on the light’. The desire is for actions possessing a certain property, and the relevant belief is that the action in question possesses that property. The end and the means are in fact identical and hence the desire for the end is simply a desire to perform the means. I will appropriate two elements of this story: the focus on intensionality and the strategy of making constitutive claims about desire. But Davidson’s account is insufficient to counter the rationalist challenge because it considers only desires that take actions as their ends. The means it considers are therefore all constitutive means, inseparable from the end itself. Many desires do not take actions as their ends – Davidson observes that desires ‘are often trained on physical objects’ (1980: 6). I may desire soda, a ’67 Ford Mustang, the delightful person across the room, etc. He seems to suggest that such desires can be reduced to or entail desires to act in some way towards those objects: ‘I want that watch in the window’ is not a primary reason and explains why I went into the store only because it suggests a primary reason – for example, that I wanted to buy the watch. (1980: 6)

Clearly an object itself cannot be the end or goal of desire. My desire for soda has to be distinguished from desires to hold or see soda; it is rather the desire to drink soda. But it doesn’t follow that all desires properly have actions for ends. A desire for a ’67 Mustang, or for the watch in the window, is typically a desire not to buy it, but to possess it (consider that the desire signifies unwillingness to give away the object once purchased). And despite the infinitive, this is not a desire to act, but a desire for the obtaining of a state of affairs: that I possess that object. Many if not all desires have states of affairs as their ends, which may involve no actions by the agent and sometimes in which the agent does not figure: e.g. that the Chicago Cubs win the next World Series. A distinction can therefore be drawn between desireto and desire-that.22 Davidson’s strategy will not work with desire-that (or with desire-to in the case of merely instrumental means), as the actions that satisfy it are not constitutive means, identical to their end, but rather instrumental means only causally related to the end. Desiring these ends therefore doesn’t entail desiring the means,23 and it is possible to desire an end without desiring to perform any of the believed means. A successful execution of such a defense of the Motivation-by-Desire Principle requires more; we must show that an agent’s desire for some state of affairs

10 Motivation to the Means

179

necessarily provides that agent with motivation towards believed instrumental means to that state of affairs. The bizarre motivational failures described by rationalists would then be impossible, and the motivational behavior they deem ‘rationally appropriate’ would acquire that appearance by being necessary. So I shall argue, but how could the motivational output of a psychological mechanism be necessary? Further, isn’t it possible to desire an end while not recognizing any means to it?

10.3 An Argument from the Constitutive Nature of Desire To make my case, we must explore the concept of desire.24 Rather than investigating what ‘a desire’ is, my focus will be on what it is to desire. I shall argue that on the conception of desiring that my rationalist opponents share with me, their challenge to the motivational efficacy of desire is incoherent and there is no difficulty in explaining how desire for an end motivates us to the means. While there is a wide array of competing philosophical views about desire, the rationalists like Nagel and Korsgaard who push this challenge seem to accept a motivation theory of desire: desiring an end essentially involves having motivation towards it.25 After all, their challenge is to explain how motivation is transferred from the end to the means: how ‘desire could extend its motivational influence beyond the scope of its immediate, spontaneous manifestations, through connection with certain beliefs’ (Nagel 1970: 35). Nagel’s question is not ‘How can the desire to drink motivate drinking?’ but rather ‘How can the desire to drink motivate inserting coins in a slot?’ This motivation view of desiring has implications that these rationalists have not fully thought through. ‘Motivation towards the end’ is significantly opaque. Partly this is because the very notion of motivation is unclear, but for now I focus on clarifying what it means for motivation to be ‘towards an end’. To be motivated is always to be motivated towards some kind of behavior or activity. (Causation of such activity is neither necessary nor sufficient, however. Not every kind of causation of activity is motivation, and motivation need not be causally efficacious; it can be defeated by opposing motivation.) Nagel recognizes a limited range of behaviors that are directly and non-problematically involved with desiring – its ‘immediate, spontaneous manifestations’26 – but rationalists deny these include pursuit of means. The puzzle here is: what kinds of activity could motivation towards an end be towards? Given that they apparently see no problem in the desire to drink’s capacity to motivate drinking, the rationalists’ answer is presumably that the immediate manifestation of desiring E is attempting to do E, and hence that motivation towards E is motivation towards doing E. We confront an immediate and obvious problem when we restrict the immediate manifestations of desiring like this; they are only possible for desire-to and not for desire-that. Suppose I desire that my children are happy. That my children are happy is not something I can do; it is not a kind of behavior. It is not plausible that (following Davidson) this is really a misdescribed desire to make my children happy, because it may be just as well satisfied if my

180

S. Finlay

children make themselves happy. Like others,27 Nagel embraces the conclusion that desires are largely impotent: ‘what they can explain is limited, and . . . even in simple cases they produce action by a mechanism which is not itself explicable in terms of desires’ (1970: 27). However, on a motivation view of desiring, this is incoherent. If motivation towards an end is nothing but motivation towards doing the end, then ‘desires’ whose ends are not the kinds of things that can be performed cannot be desires at all. It is absurd to suggest that we cannot desire states of affairs, and so rationalists must either abandon motivation theories of desiring, or allow that the ‘immediate manifestations’ of desiring include more than we have hypothesized. The only viable solution for a motivation view of desiring is that motivation towards an end E is motivation towards making E so. Desiring that I drink involves having motivation towards making it so that I drink; desiring that my children are happy, towards making it so that my children are happy.28 Desiring an end does not necessarily involve motivation towards doing the end, but rather towards facilitating the end. While this account of desire is open to serious objections (to be addressed), my point here is that no other version of a motivation theory could be correct because there is no other behavior that is feasibly the direct output of that motivation. It might be suggested that the ‘immediate manifestations’ of desiring an end might include ancillary behaviors like reflecting on the merits of the end (Scanlon 1998: 38), seeking to know whether the end has obtained (Mele 2003: 22), or fantasizing about the end obtaining (Audi 1973: 4). However, none of these are plausibly direct manifestations of motivation towards the end, which is essential to desiring on the motivation theories under consideration. It would be peculiar if desiring an end could directly motivate these behaviors but not behavior of pursuing means. The rationalist claim, that motivation towards the means cannot be explained by appeal to desire-provided motivation towards the end, can now be shown to be incoherent. To believe some action M to be a ‘means’ to an end E is simply to believe that M is an action (or part thereof) of making E so, or a facilitating condition for E. (Strictly speaking, the notion of a ‘means’ is narrower than this and doesn’t include all facilitating conditions, just as ‘causes’ don’t include every element in a causal chain. I use ‘means’ here for economy.) The motivation involved in desiring E, we have seen, must be motivation towards making E so. In other words, motivation towards the end just is motivation towards means to that end. If desire involves motivation towards the end, then it involves motivation towards the means. Within the framework of a motivation theory of desire, the rationalist objection to the Motivation-by-Desire Principle is incoherent. It doesn’t seem incoherent that I could be motivated towards drinking a soda without being motivated towards inserting a coin in the slot of a vending machine, even though I believe it is a means to my end. The key here is to note that desire, belief, and motivation are all intensional with regard to their content, or sensitive to particular descriptions or aspects of that content. In Davidson’s example, he flicks a switch thereby illuminating a room and thereby inadvertently alerting a burglar. These are three descriptions of one action, involving one bodily motion, bending a finger. But to call his intentional action ‘bending his finger’ is inappropriate. Rather his action is better described as ‘turning on the light,’ a less unwieldy way of saying

10 Motivation to the Means

181

‘making it so that the light is on.’ Any activity can be described in an indefinite number of ways. Because of this, an action can be made to look peculiarly disconnected from the motivational reach of the desire that prompts it, if it is inappropriately (albeit correctly) described. It can seem puzzling how a desire to drink could motivate me to insert coins in a vending machine. But when I insert those coins, from the intentional standpoint I am not in the first instance inserting coins in a machine, any more than I am pushing small metal disks through a plastic slot or redistributing wealth. If asked what I am doing, I will not reply ‘Inserting coins in a vending machine,’ but more likely ‘Getting a drink’. I am motivated to perform this action under the aspect of making it so that I drink soda. As I have argued, desiring to drink soda essentially involves motivation towards making it so that I drink soda. Nagel’s rationale for restricting the ‘immediate manifestations’ of desiring so as to exclude motivation towards the means seems to be that performance of means requires a ‘connection with certain beliefs,’ which must be mediated by something other than desire.29 Two interpretations are possible here. On the broad reading Nagel’s claim is that any collaboration of desire and belief is beyond the explanatory resources of the Motivation-by-Desire Principle. On the narrow reading it is only for particular kinds of belief that this interaction is problematic: pertinently, instrumental beliefs. The broad reading has at least the following to commend it; beliefs and desires are supposed to be distinct mental states, so it may seem reasonable to suspect a global problem in accounting for their interaction. Suppose therefore that the ‘immediate manifestations’ of desiring are motivated behaviors that involve no contribution from belief. This divorce of the normal operation of desire from belief is absurd. Even the ‘immediate manifestations’ of desiring where the means and the end are identical are unimaginable without interaction with belief. Being motivated to drink (swallow liquid) requires first believing that there is liquid in one’s mouth. While behavior can be caused as a direct reaction to sensation, this is not motivation by desire but mere reflex.30 The paradigm case for motivation by desire is when belief ‘excites a passion by informing us of the existence of something which is a proper object of it.’ (Hume 1978: 459) No fuss is raised about such goings-on, but clearly here too desire’s motivational effects are dependent upon belief. It is constitutive of desire to involve belief-guided motivation.31 ‘It is by means of thought only that any thing operates upon our passions,’ Hume writes (1978: 662), hence ‘the moment we perceive the falseness of any supposition . . . our passions yield to our reason without any opposition.’ (1978: 416) The supposition that in their natural element desires operate without interaction with belief is on closer inspection inconceivable nonsense and incompatible with the concept of desire. Particular desires are possible in the absence of particular beliefs, but not in the absence of belief altogether. We must suppose that Nagel and Korsgaard mean to question only the interaction between desire and beliefs of a particular kind: those concerning causal relations and instrumental means. Granting that desiring can motivate action only in conjunction with belief, it may still remain a problem how desiring an end (e.g., drinking) can

182

S. Finlay

motivate action that is not identical with that end (e.g., inserting a coin in a slot). But I have already argued that these scruples are misconceived. Desiring involves motivation towards making its end the case, and that just is to be motivated towards believed means. This account can accommodate the rationalists’ story about the reason-responsiveness involved in instrumental motivation. The instrumental ‘reason’ on which we act when we are motivated to perform some particular means to our desired end is simply the content of the means-belief (e.g., that inserting coins in a vending machine is a way to get a soda), and on a motivation theory, it is constitutive of desiring an end that we have motivation towards means to that end.

10.4 Desires as Causes and the Nature of Motivation It may be thought here that if I am correct, I have inadvertently undermined the Motivation-by-Desire Principle I am trying to defend. Indeed this is what most proponents and critics of motivation theories alike take them to accomplish.32 If desiring to ␸ is simply being motivated to ␸, then the desiring like the action is what gets motivated, and therefore cannot be what does the motivating: something else such as ‘reason’ or normative belief can claim the title of being the source of motivation. Explanations of action in terms of desire would be trivial, claiming that a person acted because he was motivated to act. In denying this consequence, I shall not deny that beliefs and reasons can appropriately be said to motivate us. But I will insist (a) that desire is a causal antecedent of overt action, (b) that motivational explanations of action in terms of desire are informative and nontrivial, (c) that explanations of motivation in terms of desire are informative and nontrivial, and (d) that the activity of being motivated is itself produced by desire, in some sense. Desiring that E essentially involves having motivation towards (making it the case that E), not being motivated to (make it the case that) E. Having motivation towards E is not a matter of being caused to (make it the case that) E, because motivation can be defeated; we can fail to act on our desires, by resisting them or because they are outweighed by conflicting motivations. Consider Jason, who desires (to have an intimate relationship with) his brother’s wife, Libby. Jason is a good brother and a decent person. He never acts on his desire, keeping it suppressed his whole life. He never makes a pass at Libby, flirts with her, makes any effort to undermine his brother’s marriage, or casts her lascivious looks. Nobody knows about Jason’s desire, or has any way of knowing about it, except Jason himself who is painfully conscious of it. What is it that Jason is conscious of? The notion of motivation is rather mysterious: we speak metaphorically of the ‘push’ and ‘pull’ of desire, but this needs cashing out. Motivation in the ‘success’ sense doesn’t seem as hard to explain; to be motivated to ␸ is to ␸ for a motive. But desiring entails motivation only in the non-success sense. What is it to be motivated towards something: to have nonefficacious motivation? Answering this question will help us rebut the charge of triviality.

10 Motivation to the Means

183

The solution to this puzzle lies in the distinction between overt activity and mental activity. Desiring or having motivation does not entail any overt behavior, but I shall argue that it does essentially involve mental activity, which is frequently overlooked. Whenever an agent acts on motivation, his (overt) action is always preceded or accompanied by mental activity. Such activity often occurs where there is no resultant overt action. It is here that I believe we can find the behavior of desiring.33 What kinds of mental activities are these? As manifestations of desiring E, what they have in common is that they satisfy the rough description, ‘seeking to make it the case that E.’ This needs care, however. To seek to make it the case that E is to take steps that you think facilitate its being the case that E. These need not be ‘means’ in a strict sense, but merely facilitating conditions. Intending to E, for example, may not be a means to E-ing, but it is a facilitating condition.34 This highlights a further need for caution. It is unlikely that an agent engages in any such mental activity because she believes the proposition that it facilitates the end. Our knowledge of the facilitating mental steps to action is more plausibly a matter of know-how than of propositional knowledge. The relevant notion of means-belief must be interpreted broadly enough to include this. What are the mental activities that constitute facilitating an end? They lie on a spectrum. On one extreme we find the exercise of volition, the necessary mental component of overt action, and at one (temporal) remove, the formation and maintenance of intention. Intending some action M is one manifestation of desiring some end E. A further remove from overt action we find calculation of means; plotting the best course to an end is often a manifestation of desiring it. Deciding or forming the intention to bring about the end is another mental step of facilitating the end. Of course, often, desiring does not involve any of these mental activities, as in Jason’s case. But there are other, sub-intentional activities of facilitating an end. In deliberation about whether to act (overtly) to promote an end (i.e., in forming our plans), one looks to see whether pursuit of that end is compatible with one’s other values and ends, and can be fitted into one’s plans without unacceptable cost. This exercise is itself a form of seeking to facilitate the end, even if the end is not taken up in intention. If Jason’s desire manifests as temptation, it will take this form. He may weigh a relationship with Libby against his relationship with his brother, or he may scheme how he might have both. But desires may not even develop this far. Motivation towards E may take the form of choosing between other ends and means to them with an eye to minimizing or avoiding harming the prospects for E. This activity of protecting the end is a form of facilitating it. Even if Jason never deliberates about whether or how to seduce Libby, he may seek to avoid courses of action that might lessen her esteem of him, or that would be detrimental to his access to her. (Perhaps he refuses to marry, or turns down career opportunities that would take him abroad, because they would harm the prospects of such a relationship.) These mental activities need not be deliberate or prolonged; they can be fleeting, unbidden, and even unwelcome. Deliberating over a foreign job offer, Jason may momentarily weigh the separation from Libby as a reason not to accept it. Although he

184

S. Finlay

immediately recoils from the thought as irrational and contemptible, this weighing is an activity aimed at facilitating the end and a manifestation of desire. Perhaps the most primitive form that motivation towards an end may take is the activity of (mentally) looking for means or facilitating conditions. Again this need not be deliberate or sustained, and it may even be ‘irrational’. We may automatically do this in a rudimentary way, even where we know full well that bringing about the end is impossible for us (e.g., if the desire is directed at the past); after all, there must be some process of thought that makes relevant and thereby activates that knowledge. This thought process also satisfies the description, ‘seeking to facilitate the end’, and offers a reply to what may be the most threatening objection against motivation theories: that we can desire states of affairs that we believe to be completely out of our control.35 If an agent is engaged in no such form of mental activity, he cannot be occurrently desiring the end. (He may still have a ‘desire’ in the dispositional sense). The difference between being motivated towards an end and being motivated to it is simply that the former involves taking mental steps towards but not all the way to the end. This allows us to explain how an agent can be motivated towards proper means to an end even without having identified any proper means; searching for means is taking a step towards them. I conclude that motivation towards E involves engaging in mental activities that one thinks facilitate E. These mental activities are causal antecedents of overt action, if any eventuates, and therefore it is true on this motivation theory that desiring causes action, although it need not. (I have here offered no defense of the claim that every action is caused by desire. This will be the case if every action must arise from such mental activities, as is highly plausible; see Finlay 2007.) Some will grant this, however, and maintain that explanations in terms of desire remain trivial and uninformative. I agree that it is trivial to explain a person’s action with the claim that the person had a desire. But it is not trivial to explain an action by appeal to a particular desire (i.e., with a specified object). Some proponents of motivation theories simply identify being motivated towards something with desiring it, but this is a mistake. It doesn’t respect the platitude that we can be motivated to do things intentionally that we don’t desire to do. While some think that this platitude is incompatible with the Motivation-by-Desire Principle,36 this also is a mistake. Given my identification of desiring/having motivation towards E with mental activity seeking to facilitate E, desire for E can motivate us to (and towards) the means to E without any need for a desire for the means. Indeed, I believe (although I shall not argue for it here) that strictly speaking desiring E only ever involves intrinsic motivation towards E, or motivation towards E for its own sake.37 It is nontrivial to explain an action by appeal to a particular desire, therefore, because any action can result from any number of distinct desires, and it is informative to be told the motive with which a person acted. Even if desiring is a cause of the action, it can be denied that it motivates it (Dancy 2000: 85–86). If desiring is to be motivated towards action, it is argued, then that which does the motivating of the action must be something else. However this argument seems to assume that if A is motivationally antecedent to B, and B

10 Motivation to the Means

185

is motivationally antecedent to C, then it can only be A and not B that ‘motivates’ C. I see no reason to accept this; the parallel does not hold, for example, between proximate causal connections and being a ‘cause’.38 Desiring causes overt action in a motive-involving way, and this is sufficient for us to maintain that desiring motivates action, even if this isn’t the whole story. The objection re-emerges with regard to the realm of mental activity, however. If desiring or having motivation towards an end is constituted by engaging in facilitating mental activities (planning, deliberating, searching, rehearsing) then it cannot be what motivates or causally explains those activities themselves. Something else must explain both those activities and the desiring, and this must be the source of motivation for the desire (and the ultimate motivator of action). I have four replies.

1. An explanation of motivation in terms of a specific desire need not be a causal explanation at all; it can be an informative constitutive explanation, enabling us to understand the motivation as a desire for E. Analogously, attributing a killing to suicide is not a causal explanation, but is an informative explanation nonetheless. 2. I have been careful to speak of desiring as involving ‘having motivation’ rather than ‘being motivated’. The intended difference is that the former suggests being motive-oriented, while the latter suggests being motive-caused. To be motiveoriented is an active rather than a passive state, and doesn’t entail being motivecaused by anything. It is consistent with this motivation theory to maintain that desire is the fundamental motivator. 3. Granted that desire or motivation is not a ‘first cause’ or causa sui, it doesn’t follow that whatever causes desire motivates it (and thereby action). Dancy writes, ‘if [desires] cannot be explained, then neither can the action that, in desiring as we do, we are motivated to perform’ (2000: 85). We are caused to desire in various ways, but because I am skeptical about our ability to commence desiring at will, I grant that episodes of desiring are never caused or motivated by further desires. The ultimate springs of desire are the territory of neuroscience and psychology, and do not ‘motivate’ us but merely cause us to have motivation. An inability to trace the etiology of actions beyond desires hardly entails that we are unable to explain how those actions were motivated. Rationalists claim that we can be motivated to desire by reasons. I have here argued at least that instrumental reasons do not motivate desires, but merely give them direction (I argue for the general thesis in Finlay 2007). 4. There are two senses in which ‘desire’ does cause us to have motivation towards our desired ends. (i) Desires as dispositions are relevant to the causal (and possibly motivational) explanation of our occurrent desiring activity. Explanations of our being motivated towards certain actions by appeal to the disposition to desire particular ends is nontrivial for the same reason that explanation of action by appeal to occurrent desire for particular ends is. (ii) In a sense desiring is after all a causa sui. Once an episode of desiring has commenced, earlier activities constituting desiring are causal antecedents of later ones. Identifying a means is a cause of intending it, for example. An episode of desiring consists in a causal

186

S. Finlay

chain, so even these mental activities can be causally explained by appeal to desiring, which is a self-motivating (though not self-initiating) process.

10.5 Defending the Account: Loose Ends In this final section, I briefly address three important remaining questions. (i) Should we accept the motivation theory of desire? (ii) Does the account have the resources to explain motivation to the means? (iii) Might the account be compatible with rationalism? i. If motivation theories of desire are mistaken then my explanation of motivation to the means fails, and the rationalist objection to the Motivation-by-Desire Principle may succeed. We might now wonder whether rationalists err simply in assuming motivation theories of desire. In particular, it may be judged implausible, in spite of my efforts, that desiring always involves behavior aimed at facilitating the end (especially where we believe the end to be unobtainable). But then what is going on when we are desiring something? Opponents of motivation theories often offer no account at all, but there are many rival accounts available. There is no space here for a critical investigation of the alternatives, so I will merely register my skepticism that any of them could be right, hopefully eliciting concurring intuitions. The main contenders are phenomenological theories (on which desiring E essentially involves some kind of sensation like pleasure or pain upon representation of E),39 judgment theories (on which it involves a normative judgment about E),40 reward theories (on which it involves having a representation of E function as a reinforcer of behavioral dispositions),41 and disposition theories (on which it involves being disposed to behave in certain ways).42 All deny that motivation is essential to desire and substitute something else in its place. But this something else seems neither necessary nor sufficient. Intuitively, we can desire E without feeling any particular sensations towards E, or judging E good, or being disposed to act in certain ways (other than in motivated ways), or having a representation of E reinforcing our behavioral dispositions. But intuitively, it is inappropriate to ascribe occurrent desire in the presence of these sensations, judgments, dispositions, or behavioral modifications, if there is no motivation at all. It is most plausibly motivation that is essential to our concept of desire, and rival theories are inadequate precisely because they make the motivational role of desire a mysterious, seemingly accidental property. ii. I have preserved the causal efficacy of desire by restricting my constitutive claims about desire to mental rather than overt activities. In doing so, it may be suggested that I have impaled myself on another horn of a dilemma: the constitutive account is too weak to account for rational motivation to the proper means. Engaging in many of the mental activities I have described falls far short of overt action aimed at some sufficient means, or a believed necessary

10 Motivation to the Means

187

means. But, it may be argued, desiring an end provides a reason to perform some sufficient means, or any necessary means, which other things being equal is thus required by rationality. There must therefore be a rational norm requiring this, and we still need a substantive faculty of practical reason that responds by extending motivation to the performance of means. Rationalists will typically agree, however, that we are not rationally required to pursue the satisfaction of our mere desires. If there is a rational instrumental norm (in Finlay 2008 I argue there is not), it addresses intending rather than merely desiring an end. So long as I intend to drink a soda, I will (if not irrational, some will insist) do whatever I believe necessary for that end. But I may desire to drink a soda and yet reasonably fail to do something I think necessary for that end – if, for example, I prefer to retain my prized commemorative quarters rather than exchange them for a drink. To be motivated by an instrumental reason, on the picture I have drawn, is just to have one’s desiring activity channeled by a particular means-belief. Even if there is no viable rationalist challenge here, an explanation may still be demanded for how desire for an end is sufficient to motivate performance of sufficient means to that end. The answer is that the activity of seeking to facilitate an end is progressive: it entails trying to advance towards an end from your current position. Before you know how to E, you are facilitating E by looking for sufficient means. But once you’ve found satisfactory means, you are only continuing to seek to facilitate E if you are doing something to take additional steps towards E. Hence desiring an end can be a sufficient source of motivation to performing sufficient means. This is its natural manifestation when a sufficient and satisfactory means has been identified and there are no inhibiting or competing motives.43 iii. Finally, it may be thought that I have not rejected the rationalists’ normresponsive motivational faculty of practical reason so much as shown how it may be found operating within desire, rather than as something that operates on desire as raw material (Nagel 1970: 31). This result would be compatible with my primary goal of defending the Motivation-by-Desire Principle against the argument from motivation to the means. But the implications for rationalism are more severe. On this account of instrumental motivation, the motivation is provided by the desire. In following instrumental reasons and seeking means, desire is not obeying any authoritative normative principles but simply acting according to its essential nature. There is no place in this account for rational norms or a motivational susceptibility to any. I have argued that if we take seriously the idea that motivation is essential to desire, then the Motivation-by-Desire Principle faces no genuine problem in explaining motivation to the means. I have further argued that we can and should take a motivation theory of desiring seriously. We can, because once we pay attention to mental activity, we find resources for solving the most pressing objections. We should, because there really doesn’t seem to be anything else that desiring plausibly could be.44

188

S. Finlay

Notes 1 It is a matter of substantive dispute whether motivational explanations of action are causal explanations. I assume that the process of being motivated to act is a causal process, although I allow that some explanations in terms of motivation are constitutive rather than causal explanations. 2 Nagel 1970: 33–5, Korsgaard 1986: 12–13, 1997: 220–21, see also Scanlon 1998: 38, Smith 2004. For Korsgaard, reason legislates these norms, whereas for Nagel, Scanlon and Smith, it discovers them. Smith’s inclusion here may raise eyebrows, as he is considered the foremost champion of the Motivationby-Desire Principle. But although Smith thinks the motivation to the means comes from a desire for the means, he argues that it must be practical reason and not desire for the end that generates this desire. 3 Thirst is perhaps not, as Nagel assumes, a desire to drink. Suppose rather it involves having dryness in one’s throat and desiring it gone. There is still an end to which drinking is a means. 4 See also Davidson 1980: 79. 5 E.g., Hubin 1996: 40, Joyce 2001: 57–8, Smith 2004. 6 Action merely in accordance with reasons is also considered ‘rational,’ but only reason-motivated action is relevant here. 7 Davidson 1980: 3, Smith 1992: 327. 8 The tension is evident in Davidson 1980: 10, 84. Dancy (2000: Ch. 1) provides excellent criticism. See also my 2006. 9 By a ‘reason’ we generically mean an explanation, or answer to a ‘why?’ question. The bifurcation into (nonnormatively) explanatory and normative results from there being two sorts of ‘why?’ questions: ‘Why is p the case?’ and (approximately) ‘Why ought p to be the case?’ In my 2006 I observe that the normative question is better framed in terms of value than ‘ought’, since we can have reasons that are outweighed. 10 Davidson maintains this, while Smith is more circumspect (1987: 44), although he clearly favors this view. 11 Smith attributes the distinction between motivating and normative reasons to Nagel (1987: 38n). But on the cited pages (1970: 4, 18), Nagel only distinguishes explanatory and normative reasons. 12 Originally (1987) he simply calls them normative requirements, but in 2004 he observes ‘the considerations that justify do indeed seem to be propositions’. 13 Smith 1987: 39, also 1992: 329, see also Davidson 1980: 9. 14 Care is needed here. Being motivated to act in the success sense involves causation of action, but motivation towards action is merely a causal factor, as I shall argue. 15 Not all normative reasons are good reasons. See my 2006, and also Davidson 1980: 9, Smith 1987: 38–9, Millar 2002, Mele 2003. 16 See also Dancy 2000: Chs. 4, 5. A complicating factor is that we can give explanatory reasons for actions by citing motivating reasons. The explanatory reason is the fact that the person accepts that motivating reason, not the motivating reason itself. 17 Williams 1981, and my 2006, 2007. Velleman suggests this ‘noncognitive’ account of reasons is implausible because the normativity of reasons is intrinsic to their content (2000: 101). However reasons are ordinary facts, and the same fact can count as a reason for and against one and the same action, indicating that their normativity is external to their content. 18 It cannot be any causation by belief plus desire, because of ‘deviant causal chains’ – see Davidson 1980: 79. Trembling may be caused by a desire-belief combination, but is not action for a reason. Davidson pleads ignorance regarding the right kind of causal connection for motivation, but the word itself suggests causation by providing a motive (end). 19 Korsgaard 1986: 12–13. See also Nagel 1970: 34, Korsgaard 1997: 220–21, and Smith 2004: 86–87. 20 We might have no trouble understanding irrational behavior. 21 Note the desire is ‘exerted,’ not merely ‘satisfied’: i.e. it is motivationally active. Hume further writes that ‘aversion and propensity towards any object . . . extend themselves to the causes and effects of that object, as they are pointed out to us by reason and experience.’ (1978: 414, my emphasis).

10 Motivation to the Means

189

22 Alternatively: action vs state desires (e.g. Mele 2003), performative vs optative desires (Green 1986: 120). 23 Davidson may appear to concur: ‘Why insist that there is any step, logical or psychological, in the transfer of desire from an end that is not an action to the actions one conceives as means? It serves the argument as well that the desired end explains the action only if what are believed by the agent to be means are desired.’ (1980: 7) But the context is an account of how we can explain an action by providing just part of its causal etiology, so I read Davidson as claiming that an action is adequately explained by citing desire for a state of affairs although this assumes the (contingent!) derivation of desiring the action (means). The passage then has the opposite signification. 24 Here I controversially assume without argument that we have one univocal concept of desiring. A distinction is commonly drawn between a substantive/narrow conception of desire and a formal/broad conception. In my view, the evidence for a ‘formal conception’ of desire can be explained away as pragmatically licensed loose talk. 25 See also Nagel 1970, Darwall 1983, Audi 1986, Schueler 1995, Lenman 1996, Dancy 2000. 26 In Hume’s words, its ‘first and most natural movements’ (1978: 492) and ‘direct expressions’ (1978: 598). 27 See also Marks 1986: 141, Scanlon 1998. 28 This does not collapse desiring that E into desiring to (or that I) make it so that E. These desires may involve the same activity, but are responsive to representation of different goals. 29 Also Korsgaard 1997: 220–21. 30 See Mele 2003: 7. 31 Velleman rejects the suggestion (2000: 262), but is working with narrower conceptions of desire and belief. There may be some (‘arational’) desires to act that can be immediately executed without instrumental beliefs proper (e.g. to hum), but even these require at least knowing how to act so. 32 Nagel 1970: 29, Platts 1979: 256, Locke 1982, Darwall 1983: 42, Staude 1986, Dancy 2000: 86–7. 33 I have also explored this idea in my 2007: 223–28. It is apparently the theory of motivation favored by psychologists (Mook 1987: 53, discussed in Mele 2003: 5–6). 34 But see Harman 1976. 35 This objection is pressed in Marks 1986: 140, Mele 2003: 22–27, Schroeder 2004: 16. For discussion see my 2007: 226n. Dancy (2000: 85–89) argues similarly, pointing out that the sense of frustration that arises in cases of desiring the unattainable shows that there is ‘motivation looking for an outlet and not finding it.’ However, he thinks this can be accommodated by a dispositional clause, which I think is insufficient and incompatible with a pure motivation theory. 36 For example Schueler writes, ‘Since it is possible to do apparently very sensible things that one has no proper desire to do, it is very implausible to say that having a proper desire of some sort is a necessary condition of having a good reason to perform an action.’ (1995: 59). 37 Finlay 2007: 224–5. See also Audi 1986: 20–21, Marks 1986: 144, Staude 1986: 177, Chan 2004. 38 Dancy seems to make this assumption, although he may rather be conflating motivation towards action with being motivated to act: ‘if we explain an action by specifying what motivated it . . . what motivated it cannot be the desire, for this only consists in the agent’s being motivated to do it. (Though it may be partly caused by this.) What motivates must therefore be that which underpins the desire.’ The parenthetical remark is inconsistent with such a conflation. 39 For example Shaw 1989; 1992, Staude 1986: 182. This view is epitomized by many of Hume’s claims. Rationalists hint at such views by focusing on particularly visceral desires like hunger, thirst, and lust (which are not in my view purely desires). Phenomenological views are comprehensively criticized in Smith 1987, 1994: 104–11. 40 Judgment theories are given by Anscombe 1957, De Sousa 1974, Davidson 1980, Quinn 1995, Millgram 1997, Scanlon 1998. For compelling arguments against judgment theories see Stocker 1979, Velleman 2000. Relatedly, perception theories link desiring with perceptions rather than judgments with normative content; see Stampe 1986, 1987, Hurley 2001. 41 Schroeder 2004. 42 Smith 1987, 1994, Stalnaker 1984: 15, Heuer 2004. Often when we ascribe a ‘desire’ we are ascribing a disposition of some kind. But active desiring involves something going on. Most plausibly, the

190

S. Finlay

disposition we ascribe as a ‘desire’ is the disposition to desire, just as the disposition we ascribe as a ‘belief’ is the disposition to believe. 43 The motivational manifestations of desirings are often limited as a result of their interactive rather than independent operation. Desire aims at its end within the boundaries established by other desires. Like the current of water or electricity, it seeks the path of least resistance. Other desires close off paths, and sometimes may stem the flow, reducing a desire merely to activities such as surveying the scene for some new or unnoticed path. Even if I believe that giving up my quarters is the necessary means to drinking a soda, my desire need not be snuffed out by my preference to retain them: this belief does not preclude hope that it might be false, and so the desire may live on in the form of ‘keeping an eye out’. 44 I thank Aaron James, David Chan, Gideon Yaffe, Janet Levin, Autumn Winters, and audiences at conferences in Southern California and Wisconsin for their helpful comments.

References Anscombe GEM (1957) Intention. Cornell University Press, Ithaca Audi R (1973) Intending. The Journal of Philosophy 70:387–402 Audi R (1986) Intending, intentional action, and desire. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Chan D (2004) Are there extrinsic desires? Noˆus 38:326–50 Dancy J (2000) Practical reality. Oxford University Press, New York Darwall, SL (1983) Impartial reason. Cornell University Press, Ithaca Davidson D (1980) Essays on actions & events. Oxford University Press, New York De Sousa RB (1986) Desire and time. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Finlay S (2006) The reasons that matter. Australasian Journal of Philosophy 84(1):1–20 Finlay S (2007) Responding to normativity. In: Shafer-Landau R (ed) Oxford studies in metaethics, volume 2. Oxford University Press, New York Finlay S (2008) Against all reason? Skepticism about the instrumental norm. In: Pigden C (ed) Hume, motivation and virtue. MacMillan, New York Green OH (1986) Actions, emotions, and desires. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Harman G (1976) Practical reasoning. Review of Metaphysics 29:431–63 Heuer U (2004) Reasons for action and desires. Philosophical Studies 121:43–63 Hubin D (1996) Hypothetical motivation. Noˆus 30(1):31–54 Hume D (1978) Treatise of human nature. Nidditch PH (ed), 2nd edn. Clarendon, Oxford Hurley P (2001) A Kantian rationale for desire-based justification. Philosophers’ Imprint 1(2): www.philosophersimprint.org/001002/ Joyce R (2001) The myth of morality. Cambridge University Press, Cambridge Korsgaard C (1986) Skepticism about practical reason. The Journal of Philosophy 83:5–25 Korsgaard C (1997) The normativity of instrumental reason. In: Cullity G, Gaut B (eds) Ethics and practical reason. Oxford University Press, New York Lenman J (1996) Belief, desire and motivation: an essay in quasi-hydraulics. American Philosophical Quarterly 33:291–301 Locke D (1982) Beliefs, desires and reasons for actions. American Philosophical Quarterly 19:241–49. Marks J (1986) The difference between motivation and desire. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Mele AR (2003) Motivation and agency. Oxford University Press, New York Millar A (2002) Reasons for action and instrumental normativity. In: Bermudez JL, Millar A (eds) Reason and nature: essays in the theory of rationality. Clarendon, Oxford Millgram E (1997) Practical induction. Harvard University Press, Cambridge MA

10 Motivation to the Means

191

Mook D (1987) Motivation: the organization of action. Norton, New York Nagel T (1970) The possibility of altruism. Clarendon, Oxford Platts M (1979) Ways of meaning. Routledge & Kegan Paul, London Quinn W (1995) Putting rationality in its place. In: Hursthouse R, Lawrence G, Quinn W (eds) Virtues and reasons: Philippa Foot and moral theory. Clarendon, New York Scanlon TM (1998) What we owe to each other. Harvard University Press, Cambridge MA Schroeder T (2004) Three faces of desire. Oxford University Press, New York Schueler GF (1995) Desire: its role in practical reason and the explanation of action. MIT Press, Cambridge MA Shaw D (1989) Hume’s theory of motivation. Hume Studies 15:163–83 Shaw D (1992) Hume’s theory of motivation – part 2. Hume Studies 18:19–39 Smith M (1987) The Humean theory of motivation. Mind 96:36–61 Smith M (1992) Valuing: desiring or believing? In: Charles D, Lennon K (eds) Reduction, explanation, and realism. Clarendon, Oxford Smith M (1994) The moral problem. Blackwell, Oxford Smith M (2004) Humean rationality. In: Mele A, Rawling P (eds) The Oxford handbook of rationality. Oxford University Press, New York Stalnaker RC (1984) Inquiry. MIT Press, Cambridge MA Stampe DW (1986) Defining desire. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Stampe DW (1987) The authority of desire. Philosophical Review 96:335–81 Staude M (1986) Wanting, desiring, and valuing: the case against conativism. In: Marks J (ed) The ways of desire. Precedent Publishing, Chicago Stocker M (1979) Desiring the bad: an essay in moral psychology. The Journal of Philosophy 76:738–53 Velleman JD (2000) The guise of the good. In: Velleman JD (ed) The possibility of practical reason. Oxford University Press, New York Williams B (1981) Internal and external reasons. In: Williams B (ed) Moral luck. Cambridge University Press, Cambridge

Chapter 11

Thresholds, Vagueness and the Psychology of Small Improvements Erik Schmidt

Abstract I argue that recent accounts of incomparability rely on empirically inadequate accounts of the psychology of choice. I support that claim by offering a novel analysis of cases of small improvements, which arise when neither of two alternatives (A and B) count as better than the other, although we can identify some small improvement of A that clearly counts as better than A but not better than B. Drawing on the work of economist Herbert Simon, I suggest that cases of small improvements only appear to raise a case of incomparability because of differences in the search costs of making certain comparisons and the way those differences can lead to intransitive preference structures. In Sections 11.1 and 11.2, I introduce the problem of small improvements, explain why it has been thought to establish a case of incomparability, and identify the main responses found in the literature. In Section 11.3, I summarize some recent work on the psychology of comparison and explain how it applies to the case of small improvements. In Section 11.4, I lay out an account based on that research and explain why I find it compelling. In Section 11.5, I conclude by identifying the implications my account would have for discussions of the problem of incomparability and practical reason more generally. Keywords Incomparability · Rationality · Satisficing · Small improvements · Vagueness

11.1 Over the past three decades, cognitive psychologists have examined the way people make decisions under conditions of uncertainty.1 Much of that research calls into question the standard model of decision-making drawn from rational decision theory E. Schmidt Department of Philosophy, Gonzaga University, Spokane, WA 99258, USA, Tel.: 1(509) 323-5975, Fax: 1(509) 323-5718 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

193

194

E. Schmidt

and economics. People simply aren’t rational in the way that traditional models of expected utility predict. Taken on its own, that finding may not be all that surprising. After all, there are many reasons why people consistently fail to act out of calculated self-interest. What is surprising, however, is that research on the psychology of choice has discovered that people deviate from the standard model of rationality in systematic and largely predictable ways. Far from being erratic or unruly, cognitive psychologists have discovered that people violate the standard accounts of rationality in ways that can be described, modeled, and incorporated into our accounts of human choice. But while many philosophers take a curious interest in the apparent irrationality uncovered by cognitive psychologists, few have examined the implications that research might have for philosophical accounts of practical reason. The chief explanation for this neglect can be found, I suspect, in a common commitment to avoiding the naturalist fallacy.2 For while no one can reasonably deny that people frequently fail to conform to the axioms of rational decision theory, it is also true that the way people actually behave does not tell us how we have reason to behave. The fact that our preferences are frequently or even predictably inconsistent does not establish that they should be inconsistent. The problem with this line of thought is that it ignores the ways in which our theories of practical reason and rational action depend on our accounts of the interests we bring to practical deliberation and the way those interests shape both our deliberative practices and our reasons for choice. If we fail to attend to the psychology of choice, therefore, we should expect our accounts of practical reason to misfire. One prominent example of such a misfire can be found, I think, in recent worries about incomparability.3 In this essay, I argue that concerns about incomparability reflect a widespread acceptance of an empirically inadequate approach to practical reason that views our reasons for choice in terms of a simplified, formal model of evaluative comparison. I suggest that a commitment to that account leads philosophers to both misdiagnose what is going on in their own test cases and to accept a problematic set of restrictions on what can count as a reason for choice. To support that thesis, I focus on one argument that has often been thought to establish the evaluative incomparability of certain alternatives: the argument from small improvements.4 After describing that argument and outlining some of the implications it has been thought to establish, I explain how we might use recent work on the psychology of choice to develop an account of cases of small improvements that does not commit us to any controversial views about the indeterminacy of normative language, the vagueness of certain values, or the incomparability of certain alternatives.

11.2 The argument from small improvements rests on cases where neither of two alternatives, A and B, count as better than the other, although we can clearly identify some small improvement of A, call it A+, that counts as better than A. Despite the

11 Thresholds, Vagueness and the Psychology of Small Improvements

195

clear superiority of A+ over A, in some cases we may still find ourselves unable to say that A+ is better than B, which implies that A cannot be equal to B. Given our initial assessment that A is neither better nor worse than B, we must now conclude that A counts as neither better than, worse than, or equal to B. Therefore, if the three comparative relations of superiority, inferiority, and equality exhaust the possible range of comparative relations, we must conclude that A and B count as incomparable. To take Joseph Raz’s example, imagine a set of choices between two alternatives (A or B: clarinetist or lawyer) that differ along a number of evaluative dimensions (e.g., money, creativity, job security, etc.) and some third alternative (A+: Clarinetist plus $1,000) that is slightly better than one of the original alternatives (A) with respect to one of the evaluative dimensions (money) used in the initial comparison of A and B. Raz’s argument for incomparability depends on the recognition that even though the improvement is both small and one-dimensional (i.e., a slight increase in financial gain), it provides us with a clear reason for thinking that the improved alternative is better than one of the original alternatives, although it does not provide us with a reason for thinking that it is better than the other. That implies that the two original or unimproved alternatives cannot be equal in value. For if they were, an improvement over one would entail an improvement over the other. One is neither better than, worse than, nor equal to the other. They are, Raz concludes, incomparable. Perhaps the most natural response to Raz’s argument would be to say that it presupposes a level of precision that is rarely, if ever, found within the domain of evaluative judgment. We might agree that if the two original or unimproved alternatives (A and B) are strictly equal – equal, we might say, with mathematical precision – then the case he offers falls outside of the three comparative relations of equality, superiority, and inferiority. But we might think that our judgments are rarely that precise. Don’t we usually allow for some degree of imprecision? Where should we locate the source of that imprecision? To search for an answer to that question, recent discussions of the problem of incomparability have turned to various accounts of vagueness and as a result, most recent accounts of incomparability can be aligned with one of the three ways that philosophers approach the more general topic of indeterminacy. They usually appeal to some kind of semantic, epistemic, or ontological account of vagueness.5 Semantic accounts of vagueness focus on the way that predicates sometimes lack the specificity required to pick out unique states of affairs in the world.6 They contend that cases of incomparability arise when evaluative terms fail to fix the truth of the claim that one of the comparative relations holds between the available alternatives. John Broome, for example, argues that cases of incomparability arise because of the semantic indeterminacy of the dyadic predicate, ‘better than’ (Broome 1999). As a result of this kind of semantic indeterminacy, the evaluative concepts we use to make a comparison can fail to fix the relations of superiority, inferiority or equality between the options. To take a simple example, the concept of artistic originality may be thought to be vague in a way that makes it neither true nor false that Picasso was more, less or equally original than Bach.

196

E. Schmidt

The epistemic approach to vagueness, meanwhile, attributes indeterminacy to ignorance. It says that while predicates have well-defined extensions, we sometimes lack the discrimination required to recognize where those boundaries lie.7 Accounts of incomparability based on epistemic accounts of vagueness argue that while one of the three comparative predicates determinately applies to each pair of alternatives, we sometimes fail to recognize it (Regan 1997). Finally, ontological approaches to vagueness focus on how objects or states of affairs might be vague (Tye 1990, Parsons and Woodruff 1995). Accounts of incomparability based on ontological accounts of vagueness argue that values can lack the structure required to result in a determinate comparative ranking. They argue that incomparability arises because no comparative predicates apply with respect to some non-trivial evaluative consideration. So on this account, incomparability arises not because we lack the capacity to make a comparison but because there is no non-trivial comparison to be made. For example, I may value reading novels and listening to jazz in ways that leaves it indeterminate whether I value one more than the other or whether I value them equally. Each of these three accounts finds a different location for the imprecision behind the argument from small improvements. The semantic account locates the indeterminacy in our normative language. Even if the comparative relations among the alternatives are determinate, our normative language fails to identify them precisely. The epistemic account locates it in our powers of discrimination. The comparative relations may be determinate, but our lack of discrimination leaves us with only a vague idea of what they are. The ontological account locates the indeterminacy in the values. Values are vague in a way that can leave us with indeterminate comparative relations. In the next section, I will argue that we can account for the indeterminacy found in cases of small improvements without committing ourselves to any general conclusions about our capacities of discrimination, the indeterminacy of normative language, or the vagueness of our values.

11.3 One important observation made by recent research on the psychology of choice is that the descriptions we use to represent prospects do more than pick out states of affairs in the world. They also highlight different features of a choice context and elicit different strategies or heuristics for incorporating those features into our decision. Different descriptions, therefore, can elicit different preference structures by encouraging us to frame them in different ways. For example, since people tend toward risk-aversion in cases of gains and riskseeking in cases of loss, our preferences tend to vary with respect to whether the options are represented as gains or losses (Tversky and Kahneman 2000). Or to take another example, our preferences among a set of alternatives may be influenced by the addition of further members to the set of available alternatives. So even though

11 Thresholds, Vagueness and the Psychology of Small Improvements

197

the relative value of two alternatives, A and B, is not affected by the addition of another alternative, our preference between A and B may, in some contexts, be sensitive to additions or subtractions from the set of available alternatives.8 The principle source for these ‘framing effects’9 is the fact that people compare alternatives in terms of their subjective rather than their absolute value and the fact that assessments of subjective value tend to be a function of perceived differences in value rather than total amounts of wealth. We tend to measure the value of the available alternatives in terms of losses, gains, or the maintenance of the status quo.10 But the sensitivity of our subjective estimates of those differences in value is not entirely uniform. In some contexts, we are more sensitive to differences in value than we are in others. As a result, our preferences may be sensitive to small differences in values in some contexts but sensitive only to fairly large differences in others. Let’s call a measurement of the minimum amount of evaluative difference required to register a difference in subjective value its threshold. By lowering a value’s threshold, we reduce the amount of the evaluative difference required to register a difference in subjective value. By lowering thresholds, therefore, we increase the sensitivity or precision of the comparisons we make with respect to that value. The concept of a threshold clarifies what is going on in Raz’s argument from small improvements. In the case Raz presents, we use a different set of thresholds to apply the same set of values to the two comparisons he asks us to make. The thresholds we use to compare A with A+ are lower than the thresholds we use to compare either A with B, or A+ with B. As a result, an increase in value feels like a significant difference when it comes to the comparison of A and A+ even though it feels insignificant with respect to a comparison of A+ and B. In order to defend this account of the argument from small improvements, we need to have a reason to think that we apply a conflicting pair of thresholds in just this way. Such a reason can be found, I think, in Herbert Simon’s account of satisficing rationality (Simon 1983). The central idea behind Simon’s account of satisficing is that due to a variety of limitations in areas like computational skill and memory, or what he calls the conditions of bounded rationality, some of the information that is relevant to the choices we make may be costly to obtain. As a result, it may be expensive in terms of time and energy to determine which alternative counts as optimal with respect to our criteria for choice. Simon argues that it may often be rational, given our overall goals, to economize on our processes of deliberation. It may sometimes be rational, in other words, to avoid optimizing our decision by seeking the best alternative available. In particular, it is more rational to be a ‘satisficer’ and not optimize one’s choice in cases where the advantage one stands to gain by choosing the best may not be worth the cost of figuring out which alternative that would be.11 Given the constraints on our computational ability, the costs of discovery can outweigh the marginal gain of choosing a better alternative. To take just one example, imagine a company that makes steel tubing for bicycles. Such a company may choose to use a blend of alloys that provides an acceptable balance between weight and stiffness even though it fully recognizes that

198

E. Schmidt

further research would probably reveal an alloy that exhibits an even better balance of those properties. The reason why the company may stop its research at some point is because it might have reason to think that any further gain it could achieve through a better alloy would be offset by the research costs required to discover it. Simon’s observation that additional information often costs more than the benefits that information would provide offers an explanation for why the thresholds we use to make assessments of subjective value tend to vary across multiple choice contexts. For our choice of thresholds will be partly a function of both the costs we would incur by lowering our threshold and the value we can expect to gain by doing so. For example, we should expect people to increase their sensitivity to differences in relative value in cases where they have reason to expect that the cost of that sensitivity will be compensated by a corresponding increase in the amount of value they will realize through the corresponding increase in the sensitivity of their choice. Alternatively, we should expect people to increase their sensitivity to differences in relative value in cases where that increase can be gained at little or no cost, because comparisons are relatively easy to make. But in choice contexts that require people to make complex trade-offs, it will be difficult and therefore costly to make comparisons with any degree of precision. In those contexts we should expect people to use fairly high thresholds, unless they have a strong reason to think that they would gain a lot by making more fine-grained comparisons. This is why it is important to notice that the improvement found in Raz’s version of the argument from small improvements is small but also that it is onedimensional. A and A+ differ only in terms of a discrete amount of money. The fact that the difference between A and A+ is one-dimensional is important. For although it is not impossible to make comparisons along multiple dimensions, it is usually more difficult to make fine-grained comparisons along multiple dimensions. The presence of multiple evaluative dimensions drives up the relative cost of lowering our thresholds. The cost of lowering our thresholds in the context of one-dimensional differences, in contrast, will tend to remain quite low. This will be especially true in cases like the one described by Raz, where the one-dimensional difference involves a difference in something like money, which already has a built in cardinal scale.12 In cases involving one-dimensional differences, the cost of lowering the threshold may diminish almost to the point where it will be non-existent. Other things being equal, for example, we can lower the threshold to differences of pennies without increasing our decision costs at all.13 This suggests that the argument from small improvements is partly a product of our sensitivity to the costs associated with increasing or decreasing a threshold. When presented with the task of comparing two alternatives, A and B, which differ along many relevant dimensions (e.g., a career as a clarinetist and a career as a lawyer) together with the task of comparing one of those initial alternatives with some additional alternative that realizes a small improvement along one dimension of evaluation, there will be an enormous asymmetry in decision costs associated with the thresholds we use to frame each of those choices. In the choice between

11 Thresholds, Vagueness and the Psychology of Small Improvements

199

A and A+, there will be no great cost in using a threshold that registers the small evaluative difference between them. When it comes to the choice between B and either A or A+, however, the decision cost of employing a threshold that would register that evaluative difference would be enormous, due to the decision costs associated with making the complex trade-offs that comparison requires us to perform. But now it looks like cases of small improvements do not offer a clear case of incomparability. The important issue does not seem to be that some alternatives are difficult to compare because the normative language we use to make those comparisons is indeterminate or because the values we use are vague. Rather, the problem results from differences in the difficulty of making certain comparisons and the way those differences encourage us to frame those comparisons in ways that lead to the impression of incomparability. In this way, the puzzle raised by cases of small improvements appears to be a product of human psychology and as such, it may have nothing important to teach us about the vagueness of our values or the indeterminacy of our normative language.

11.4 In the last section, the difference in the thresholds that I claimed was responsible for the phenomenology behind the argument from small improvements was a product of the fact that the improvement contained a cardinal scale. The improved career as a clarinetist carried a $1,000 bonus. The account I offered highlighted the way that the presence of a cardinal scale makes precise comparisons easier or less costly. Nothing I have said should be taken to imply that a cardinal scale is required for comparison. After all, most cases of comparison do not involve the application of a cardinal scale to calibrate our choice. In addition, nothing I have said should be taken to deny that considerations other than the cost of precision might affect the costs of making certain comparisons in a way that leads to the appearance of incomparability. There may, for example, be psychological costs associated with making certain comparisons. Concerns of etiquette or concerns about being too stingy or profligate, for instance, may motivate us to raise our thresholds when making certain choices. Most of us would be reluctant, for example, to allow small differences in size to guide our choice when selecting a pastry from a conference party tray even if there would be little difficulty in pulling out a tape measure to measure that difference.14 The account I have offered attempts to explain the phenomenology of cases of small improvements by appealing to the impact that differences in decision costs can have on the way we frame certain comparisons. It leaves open the question of whether some differences in decision costs may be the product of the fact that some alternatives are incomparable. Of course, if two alternatives were literally incomparable, we should expect that fact to be reflected in the way we would frame a choice between those alternatives. But if my analysis of the argument from small improvements is correct, we can allow that as at least a possibility, without committing ourselves to the view that cases of small improvements count as such a case.

200

E. Schmidt

So unlike the epistemic account, my account recognizes that there may be cases where there simply is no fact of the matter with respect to whether or not one of the alternatives counts as at least as good as the other, and unlike the semantic or ontological approaches, my account does not commit us to any further claims about the indeterminacy of normative predicates or to any meta-ethical claims about the nature of the values we use as criteria for making comparisons. In addition to that parsimony, the chief advantage of the account I have offered is that it locates the intransitivity that drives the argument from small improvements in the thresholds or choice procedures that we apply to our evaluative comparisons. Recent accounts of the psychology of choice explain why that is the most plausible source for intransitive preferences. Those accounts suggest that we tend to make decisions by tracking assessments of subjective value. This explains why our preferences occasionally violate the principle of transitivity. Instead of tracking the contribution each alternative would make to the total amount of value realized in our lives, our preference structures track subjective assessments of value. Flexibility within our decision procedures, therefore, can lead to intransitivity in our judgments. Say, for example, that as a result of framing effects, someone prefers (i) (A over B), (ii) (B over C), but also (iii) (C over A). When offered a choice between C and B, she would take B, following preference (ii), and if offered a choice of keeping B or taking A, she would take A, following preference (i), which amounts to a preference of A over C that violates preference (iii). Acting on her first two preferences, in this way, threatens to frustrate her third preference. It is not difficult to see how framing effects could lead someone to employ a set of intransitive preferences. For the conflict that arises in cases of intransitivity, we might notice, is not directly between two binary preferences: (C over A) and (A over C). Rather, it is a conflict between a binary preference (C over A) and a transitive preference (B over C and A over B) that is equivalent to a further binary preference (A over C) that conflicts with the initial preference (C over A).15 So while the axioms of rational decision theory express the conditions we must satisfy in order for our preferences to be consistent with respect to the values each option would realize, their value is being considered independently of the description by which each option is represented. That is why the fact that we tend to make decisions by tracking assessments of subjective value explains why our preferences occasionally violate the principle of transitivity. Instead of tracking the contribution each alternative would make to the total amount of value realized in our lives, our preference structures track subjective assessments of value. So we occasionally fail to treat the preference ordering (B over C) and (A over B) as equivalent to (A over C) because we track subjective value and our judgment is subject to influences that do not affect the amount of value each alternative would realize. When faced with a choice between A and C framed in one way, we choose C because we perceive that it has greater relative value, even though when faced with a different framing of a choice between A and the winner of a choice between B and C, we choose A because it has greater subjective value. But why should we think that intransitivity arising through flexible features of the process of decision-making is any better than the original incomparability found

11 Thresholds, Vagueness and the Psychology of Small Improvements

201

in the argument from small improvements? After all, there are clear benefits to avoiding intransitive preference structures, since conforming to the principle of transitivity (and by extension, the other rules of rational decision-making) insures that we avoid giving up the values we prefer without being adequately compensated in some other way. It prevents us from being ‘money pumped’ or ‘Dutch-booked.’ One reason why I think it is better to accept the intransitivity of our deliberative procedures is that we have some reason for thinking that the features of the psychology of decision-making I have identified would not have dire implications for our prospects of becoming rational creatures. We need to remember that the intransitivity arising from the ways we frame certain sets of comparisons does not result from the intransitivity of some stable set of preferences that hold across multiple choice contexts. It is a product of the fact that the psychology of choice is characterized by its flexibility. After all, the central observation of the literature on the cognitive psychology of decision-making is that our preferences are not simply elicited or revealed by the choices we make, they are partly constructed during the process of deliberation (Slovic 1995, Tversky 1996). But that flexibility is precisely what provides us with a way of guarding against the threat of being ‘Dutch-booked’ or ‘money-pumped’ through the apparent intransitivity of our preferences in cases of small improvements. In cases where someone attempts to exploit the weakness created by certain framing effects, we are likely to recast the choice they present in ways that prevent those strategies from being successful. Recent empirical research backs up this entirely reasonable expectation (Kelman et al. 2000). The fact that we employ a set of flexible decision frames in ways that lead to violations of transitivity, therefore, may simply show that we rarely operate in contexts filled with ‘Dutch-bookies’ or their equivalent. If we were faced with the possibility of being pumped out of our resources at every turn, then we should expect people to develop choice strategies that protect them against that hazard. The fact that we employ strategies that may lead to intransitivity in our revealed preferences, therefore, gives us some reason to think that the hazards identified by the defenders of decision theory do not count among our chief concerns. It may be far more important, in the end, that we be able to make choices in cases of incomplete information, or in cases where we must make difficult trade-offs among values that are important to us. If that’s true, we have some reason for thinking that the phenomenon of small improvements does not have dire implications for practical reason. It simply reveals a certain degree of flexibility in the way we make comparisons across a range of contexts.

11.5 The central goal of this essay has been to defend the suggestion that cognitive psychology has something to contribute to our accounts of practical reason and rational action. I have argued that worries about incomparability reflect an acceptance

202

E. Schmidt

of an empirically inadequate account of practical reason that views our reasons for choice in terms of a simplified, formal model of evaluative comparison. That simplified account leads philosophers both to misdiagnose what is going on in their own cases and to accept a problematic set of restrictions on what can count as reasons for choice. Once those restrictions are in place, certain types of choice strike them as unreasonable (or more precisely, non-reasonable) because the comparisons they employ fall outside the restrictions they have placed on what can count as a reason for choice. This results in the allegation that those cases present us with instances of incomparability since they appear to present cases where the alternatives cannot be reasonably compared. The picture that emerges from recent research on decision-making, however, suggests that individuals make choices on the basis of reasons that systematically violate the maximizing logic of good/better/best, since the descriptions we use to represent different prospects do more than pick out states of affairs in the world. They also frame the choices we make with respect to those prospects by highlighting different features of the choice context and by eliciting different heuristics for incorporating those features into deliberation. By taking a closer look at research on the strategies we use to make decisions, therefore, I believe that we will be able to develop a clearer picture of our interests and concerns, and the way they figure into our reasons for choice.

Notes 1 Many early important contributions to that research have been collected in Kahneman, Slovic and Tversky (1982). 2 For further discussion of this claim, see Held (1998). 3 Following Chang (1997), I will use the term ‘incomparability’ here rather than the more traditional label, ‘incommensurability.’ Using the term ‘incommensurability’ suggests that the source of the problem is the lack of a common measure (or unit of measurement) in terms of which a comparison might be made. That confuses the problem, an inability to make certain comparisons, with one alleged source of that problem, the absence of a common measure. 4 Discussions of the argument from small improvements can be found in Chang (2002), De Sousa (1974), Raz (1988), Sinnott-Armstrong (1988), and Temkin (1996). 5 For the sake of brevity, I will set aside the recent defense by Chang (2002) of a fourth comparative value relation of parity, which appears to count as a variation on the three-valued logic strategy developed by Michael Tye and Nicholas Rescher. Parfit (1984) offers a similar strategy by arguing that while both the ‘better than’ and ‘worse than’ relations can readily encompass vagueness, equality cannot, since it would lead to violations of transitivity. 6 This approach is defended in Russell (1923), Dummett (1975) and Evans (1978). 7 The epistemic approach is defended by Williamson (1992) and Cargile (1969). 8 For a discussion of this issue as it arises within the context of medical decision-making, see Redelmeier and Shafir (1995). 9 For a general summary of the psychology of framing, see Thaler (2000), Kahneman and Tversky (2000b). 10 The implications this has for discussions of rationality and choice have been recognized for some time. Perhaps the first person to appeal to this fact in order to explain the psychology of decision-making was

11 Thresholds, Vagueness and the Psychology of Small Improvements

203

the 18th century mathematician Daniel Bernoulli, who wanted to explain why risk aversion decreases with wealth. Bernoulli (1954) suggests that our preferences depend upon the subjective value of the amounts of money involved and so as the amount of money increases, the subjective value of the difference between the two amounts offered in the choice will decrease. For example, the differences between the subjective utility of receiving $20 and receiving $100 is greater than the difference between receiving $1,020 and receiving $1,100 and so we will tend to be more risk averse in the first case than we are in the second. 11 The situation here is analogous to the problem of optimizing a search for a parking space, where the time spent searching for a closer spot can exceed the time one saves as a result of finding a closer spot. 12 The presence of a built-in cardinal scale is conspicuous in many of the central examples found in the literature on the argument from small improvements: Anderson (1993), Broome (1997), Raz (1988), and Chang (2002) 13 Of course, other things are rarely equal and many considerations other than decision costs will be relevant to decisions about where to set a threshold. For example, we may think that reducing our thresholds in Raz’s case of a career so that we register differences measured in terms of pennies would amount to a form of penny pinching that carries some kind of social or psychological cost. If that is true, then we may sometimes be reluctant to set the threshold as low as we would if decision-costs were our only concern. But that only shows that there will often be many considerations that play a role in the way we frame a choice. 14 The account I have offered here might, in this way, apply to arguments in favor of normative or constitutive incomparability (i.e., to the thought that some forms of comparison are ruled out by the norms that govern appropriate attitudes toward certain goods, like friendship and human life). Instead of considering those norms to be constitutive of the value that is involved, we may consider them to be rules that govern the way in which choices that raise those values ought to be framed. 15 For further discussion of this point, see Schick (1986) and Velleman (1993).

References Anderson E (1993) Value in ethics and economics. Harvard University Press, Cambridge MA Bernoulli D (1954) Exposition of a new theory on the measurement of risk. Econometrica 22: 23–36 Broome J (1997) Is incomparability vagueness? In: Chang R (ed) Incommensurability, incomparability, and practical reason. Harvard University Press, Cambridge MA Broome J (1999) Ethics out of economics. Cambridge University Press, Cambridge Cargile J (1969) The sorites paradox. British Journal for the Philosophy of Science 20:193–202 Chang R (1997) Incommensurability, incomparability, and practical reason. Harvard University Press, Cambridge MA Chang R (2002) The possibility of parity. Ethics 112:659–88 De Sousa R (1974) The good and the true. Mind 84:547–48 Dummett M (1975) Wang’s paradox. Synthese 30:301–24 Evans G (1978) Can there be vague objects? Analysis 38:208 Held V (1998) Whose agenda? Ethics versus cognitive science. In: May L, Friedman M, Clark A (eds) Mind and morals: essays on ethics and cognitive psychology. MIT Press, Cambridge MA Kahneman D, Tversky A (eds) (2000a) Choices, values, and frames. Cambridge University Press, Cambridge Kahneman D, Tversky A (2000b) Prospect theory: an analysis of decision under risk. In: Kahneman D, Tversky A (eds) Choices, values, and frames. Cambridge University Press, Cambridge Kahneman D, Slovic P, Tversky A (eds) (1982) Judgment under uncertainty: heuristics and biases. Cambridge University Press, Cambridge

204

E. Schmidt

Kelman M, Rottenstreich Y, Tversky A (2000) Context-dependence in legal decision making. In: Sunstein C (ed) Behavioral law and economics. Cambridge University Press, Cambridge Parfit D (1984) Reasons and persons. Oxford University Press, New York Parsons T, Woodruff P (1995) Worldly indeterminacy of identity. Proceedings of the Aristotelian Society 95:171–91 Raz J (1988) The morality of freedom. Oxford University Press, New York Redelmeier D, Shafir E (1995) Medical decision making in situations that offer multiple alternatives. Journal of the American Medical Association 4(273):302–305. Regan D (1997) Value, comparability, and choice. In: Chang R (ed) Incommensurability, incomparability, and practical reason. Harvard University Press, Cambridge MA Russell B (1923) Vagueness. Australasian Journal of Philosophy and Psychology 1:84–92 Schick F (1986) Dutch bookies and money pumps. Journal of Philosophy 83:112–19 Simon H (1983) Reason in human affairs. Stanford University Press, Stanford Sinnott-Armstrong W (1988) Moral dilemmas. Blackwell, Oxford Slovic P (1995) The construction of preference. Journal of American Psychology 50:364–67 Temkin L (1996) A continuum argument for intransitivity. Philosophy and Public Affairs 25: 175–210 Thaler R (2000) Mental accounting matters. In: Kahneman D, Tversky A (eds) Choices, values, and frames. Cambridge University Press, Cambridge Tversky A (1996) Rational theory and constructive choice. In: Arrow K (ed) The rational foundations of economic behavior. Cambridge University Press, Cambridge Tversky A, Kahneman D (2000) Loss aversion in riskless choice: a reference-dependent model. In: Kahneman D, Tversky A (eds) Choices, values, and frames. Cambridge University Press, Cambridge Tye M (1990) Vague objects. Mind 99:535–57 Velleman D (1993) The story of rational action. Philosophical Topics 21:229–54 Williamson T (1992) Vagueness and ignorance. Proceedings of the Aristotelian Society Supplement 66:145–62

Part V

Freedom and Moral Agency

Chapter 12

Taking Liberty with Humean Necessity: Compatibilism and Contingency Troy L. Booher

Abstract This essay considers what David Hume’s views on free will and responsibility, properly understood, can tell us about current debates over the adequacy of compatibilist views on free will. Following Hume’s lead, I argue that questions concerning whether determinism threatens free will and responsibility cannot be answered in the abstract. Instead, answers to such questions depend upon contingent considerations involving moral sentiments. The contingency present in debates over compatibilism explains why disputes concerning free will that are set forth in absolute terms (e.g., does determinism destroy the possibility of free will?) always end in stalemate. Understanding the contingency present in disputes over free will not only sheds light on the nature of the disputes themselves, but also explains why more specific current debates over, for example, genetic predispositions, should be understood as presenting moral, not metaphysical, questions. Keywords Agency · Determinism · Free will · Hume · Responsibility Many scholars consider compatibilists to argue persuasively against so-called libertarians (roughly, those who deny that causal necessity applies to actions we consider free), but fail to address seriously the concerns of so-called hard determinists (roughly, those who deny that there are free actions because causal necessity applies to actions we consider free). David Hume is taken to exemplify this shortcoming in compatibilist accounts. In his chapters on liberty and necessity, Hume argues that we cannot plausibly deny that causal necessity applies to actions we consider free, but provides no reason to believe that the actions we consider free actually are free. In other words, Hume argues that the libertarian notion of freedom is implausible, but does not provide a positive account of free action that explains why hard determinists are mistaken when they claim that causal necessity precludes actions from being free. This apparent shortcoming leaves hard determinists understandably unsatisfied. T.L. Booher Snell & Wilmer L.L.P. Law Offices, 15 West South Temple, Suite 1200, Beneficial Tower, Salt Lake City, Utah 84101-1004, USA e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

207

208

T.L. Booher

This traditional reading of Hume is mistaken, and understanding why sheds light on the perennial debate between compatibilists and hard determinists. Hume not only provides an adequate positive account of freedom (i.e., of the kind of freedom required for moral accountability and agency), but, once properly understood, also provides insight into why arguments between compatibilists and hard determinists cannot be resolved in the abstract and seem to end in stalemate. In short, Hume’s account reveals that the question of whether causal necessity precludes the kind of freedom required for moral responsibility and agency depends upon contingent, empirical facts, and therefore the answer to this question could change over time. If this account is correct, the ‘winner takes all’ framing of the disputes between hard determinists and compatibilists is misguided and misleading. To reach this conclusion, however, it is important first to understand why the traditional reading of Hume is incorrect. Toward that end, this essay first describes an interpretive puzzle generated by Hume’s chapters on liberty and necessity and then provides a solution to that puzzle. The solution reveals that Hume’s account of freedom is not contained in his chapters on liberty and necessity as the traditional reading of Hume supposes, but instead is contained in Hume’s discussion of moral sentiments. Hume’s positive account of freedom places him in the ‘reactive attitudes’ camp of compatibilists, but with a recognition that reactive attitudes can change. This recognition permits Hume’s account to explain the fact that views on agency and responsibility have changed throughout history, reveals how disputes between hard determinists and compatibilists can be resolved, and, perhaps, sheds new light on contemporary disagreements over whether, for example, genetic predispositions are relevant to moral responsibility, and if so, how.

12.1 The Interpretive Puzzle Hume describes his Treatise discussion of liberty and necessity as putting ‘the whole controversy in a new light, by giving a new definition of necessity’ (see p. 661 in Hume 1978). According to this new definition, necessity consists in ‘the constant union [of objects] and the inference of the mind’ (see p. 400 in Hume 1978). Paul Russell has argued that many traditional interpretations of Hume’s discussion of liberty and necessity fail because they provide a role for Hume’s definition that is merely consistent with the rest of his discussion instead of essential to it (Russell 1995). While Russell’s criticism is well taken, surprisingly, Russell’s own interpretation of Hume is subject to the same criticism. Russell’s interpretation does not explain what Hume meant when he said that his ‘new definition of necessity’ put ‘the whole controversy in a new light.’ One of the great merits of Russell’s book is that it reveals a major interpretive puzzle. Different interpretations of Hume’s discussion of liberty and necessity can be understood as attempts to solve this puzzle. The following three claims are all persuasively attributed to Hume by Russell (see pp. 11–12 in Russell 1995):

12 Taking Liberty with Humean Necessity

209

Humean Necessity: Necessity consists in the constant union of objects and the inference of the mind. Anti-libertarian Claim: Any liberty that means a lack of necessity does not exist and would make morality impossible. Internal Cause Claim: Free actions are distinguished from unfree ones, not by the absence of a cause, but rather by the type of cause.

Conjoining these three claims leads to an interpretive puzzle, which is described below. According to one possible interpretation of Hume, someone acts freely just whenever the cause of her action is of the right type, for instance, her willing.1 On this interpretation, recognizing the Internal Cause Claim is all that is essential to understanding when actions are free. But notice that if freedom requires compliance only with the Internal Cause Claim, then it seems that Humean Necessity is not essential to Hume’s discussion. If we distinguish free actions by the type of cause, then it is irrelevant what account we give of the necessity involved in the causal relation.2 Hume’s announcement that his new definition of necessity puts the whole controversy in a new light, however, requires an essential role for Humean Necessity to play. This interpretation seems inadequate because Humean Necessity is merely consistent with the overall discussion rather than essential to it. Call this the ‘Not Essential Problem’. It is tempting to think that this way of reading Hume can easily be amended to provide an essential role for Humean Necessity: it is essential because a ‘stronger’ kind of necessity would threaten freedom.3 After all, Hume believes that a ‘weaker’ kind of necessity destroys freedom, as the Anti-libertarian Claim shows. Hume also believes that mistaking mere causation for compulsion—in Hume’s terms, for ‘force, and violence, and constraint’—is a major confusion that leads others to find necessity a threat to freedom (see p. 407 in Hume 1978). So perhaps the essential role that Humean Necessity plays is that it allows one to respect the Anti-libertarian Claim while distinguishing actions that are caused from actions that are compelled. This amended interpretation suffers from three problems. First, it seems to ignore Hume’s Internal Cause Claim, which states that we identify actions as free by reference to the type of cause. Second, it is a mistake to think that Humean Necessity aids in distinguishing causation from compulsion. Whether an action is compelled depends upon the type of cause rather than the type of necessity involved in the causal relation itself. The difference between a person going across the room because she is pushed and because she wants a drink of water requires reference to types of causes and not the types of necessity involved in the causal relations. Third, while some contemporary compatibilists may want to hold a position akin to the amended interpretation, such a position cannot be attributed to Hume. A compatibilist could maintain that in all possible worlds where the type of necessity is ‘stronger’ than Humean Necessity, there are no free actions. Hume, however, believed that a ‘stronger’ necessary connection would not threaten freedom. For Hume, an action must be free for it to be an action for which one could properly be held responsible. But whether one can properly be held responsible depends upon dispositions to praise and blame. Thus, unless believing that there was a

210

T.L. Booher

‘stronger’ causal relation would alter dispositions to praise and blame, the ‘stronger’ causal relation would not affect our practice of holding each other responsible; nor should it affect our practice for Hume as long as our relevant sentiments continue to approve of our practice overall. But we should not expect such effects, or anyway, we should not attribute the expectation to Hume. In the Enquiry, Hume argues that even if one were convinced that (i) the Creator initially caused whatever originates all actions and (ii) every deed was contributing to an overall good, these convictions would have little affect upon our moral ascriptions because ‘[t]he mind of man is so formed by nature that, upon the appearance of certain characters, dispositions, and actions, it immediately feels the sentiment of approbation or blame’ (see p. 102 in Hume 1975). Hume did not believe that moral ascriptions would be affected by belief in a ‘stronger’ causal connection. The amended interpretation of the role that Hume provides for Humean Necessity is mistaken, not simply because it misunderstands what it takes for an action we consider free to be compelled, but also because it wrongly interprets Hume as thinking that a ‘stronger’ causal relation would threaten freedom. Call this the ‘Not Hume Problem’. The interpretive puzzle has now emerged. How can one provide an interpretation of Hume that distinguishes free actions by the type of cause, and yet provides Humean Necessity an essential and acceptable role to play? Paul Russell has attempted to solve this interpretive puzzle. Russell correctly points out that (i) moral praise and blame are, according to Hume, felt rather than judged, (ii) one must have access to another’s character to have the necessary moral feeling, and (iii) one can have access to another’s character only by making a causal inference. The necessity involved in causation is consequently needed for the moral sentiments of praise and blame (see p. 64 in Russell 1995).4 Thus, according to Russell, Humean Necessity is essential because, without it, Hume has no basis to claim that we have access to another’s character to judge her. Russell’s interpretation avoids the Not Essential Problem and the Not Hume Problem. Russell’s interpretation fails. Although Russell has outlined a position Hume holds, Russell has failed to explain why Hume thinks that he has shed ‘new light’ on the ‘long disputed question of liberty and necessity’ in the Treatise ‘by giving a new definition of necessity.’ It is essential to judging someone responsible that one make inferences between character and actions. It is also essential to making those inferences properly that one experience a constant conjunction between types of character and types of actions, which certainly essentially involves Humean Necessity. However, this does not explain why Hume thinks that his definition of necessity sheds any new light on the problem. The inference between character and actions requires at least Humean Necessity, but it does not require Humean Necessity specifically. The much ‘stronger’ understandings of necessity that Hume is arguing against in the Treatise would allow inferences from actions to character just as well. In fact, any account of causal necessity that implies causal connections are inferentially detectable, or at least have constant conjunction as a consequence, will work. Russell has found an essential role for Humean Necessity, but has failed to find the essential role that Hume claims puts the ‘whole controversy in a new light.’

12 Taking Liberty with Humean Necessity

211

12.2 The Solution to the Interpretive Puzzle To solve the interpretive puzzle, the controversy Hume thought he was shedding light upon must be considered more carefully. Plainly enough, the controversy Hume was addressing was the controversy during the time he wrote, which perhaps is best understood by examining the debate concerning liberty and necessity between Thomas Hobbes and John Bramhall.

12.2.1 The Hobbes / Bramhall Debate Bramhall claims that real necessity and true liberty are incompatible with one another. Bramhall makes this claim by drawing upon the common intuition that ‘if it be inevitably imposed upon me . . . [it is] impossible for me to choose whether I shall undergo it or not’ (see p. 43 in Chappell 1999). Bramhall, however, does not deny that free actions have causes. Instead, he merely denies that these causes necessitate free actions in the same way as causes that do not involve free actions necessitate their effects. One way Bramhall does this is by drawing a (somewhat unclear) distinction between ‘absolute necessity’ and ‘necessity upon supposition.’ Bramhall claims that only the latter type of necessity applies to free actions. Hobbes refuses to draw this distinction, thereby provoking Bramhall to explain the importance of the distinction: [T]here is a great difference between determining and being determined. If all the collateral causes concurring to the production of an effect were antecedently determined, what they must of necessity produce and when they must produce it, then there is no doubt but the effect is necessary. But if these causes did operate freely or contingently, if they might have suspended or denied their concurrence, or have concurred after another manner, then the effect was not truly and antecedently necessary, but either free or contingent. . . . So [human] necessity is no absolute, no antecedent, extrinsical necessity, but merely a necessity upon supposition. (See p. 44 in Chappell 1999.)

While this paragraph contains a number of claims worth discussion, for understanding the ‘controversy’, it is important only to note that Bramhall believes his distinction between two kinds of necessity is required to explain both how actions are free and how free actions involve contingency. In contrast, Hobbes argues for the familiar position previously outlined as having the Not Essential Problem. Hobbes claims that if the will causes an action, then the action is free, and that this kind of freedom is compatible with there being only one kind of necessity (see p. 73 in Chappell 1999). Only the type of cause determines whether an action is free for Hobbes, and thus, the ‘strength’ of the necessity involved in the causal relation is irrelevant. There is no reason to recognize two kinds of necessity because the existence of free actions does not depend upon it. Ultimately, the debate between Hobbes and Bramhall boils down to disagreement over whether more than one type of necessity is needed to account for free actions. This was the controversy at the time Hume wrote.

212

T.L. Booher

This controversy is one on which Hume’s new definition of necessity does shed light. Hume admits that it is conceivable that given any cause the effect will fail to occur (see p. 95 in Hume 1978). Hume nonetheless believes that the same kind of necessity applies to all cause/effect relations. Hume accounts for the common intuition Bramhall relies upon—that no ‘absolute necessity’ applies to free actions—but only because Hume finds no ‘absolute necessity’ in any causal relation. As Hume states, ‘I do not ascribe to the will that unintelligible necessity, which is suppos’d to lie in matter [; instead] I ascribe to matter, that intelligible quality. . .which [all] must allow to belong to the will. I change, therefore, nothing in the receiv’d systems, with regard to the will, but only with regard to material objects’ (see p. 410 in Hume 1978). Hume also agrees with Hobbes, however, that there is no reason to posit more than one kind of necessity. Thus, Hume’s new definition of necessity resolves the Hobbes/Bramhall controversy by acknowledging contingency in free actions without positing two kinds of necessity. For Hume, ‘there is but one kind of necessity, as there is but one kind of cause, and. . .the common distinction betwixt moral and physical necessity is without any foundation in nature’ (see p. 171 in Hume 1978). The controversy that Humean Necessity resolves is whether free actions are necessitated just as other natural events, not whether or how actions are free. Because both Hobbes and Bramhall simply assume that there are free actions (actions for which one can be held accountable), there was no controversy on that issue.5 Interpreting Hume as referring to the Hobbes/Bramhall controversy in the Abstract explains the essential role that Humean Necessity plays in Hume’s discussion of liberty and necessity while avoiding both the Not Essential Problem and the Not Hume Problem. The Not Essential Problem is that if free actions are distinguished by the type of cause, then Humean Necessity is inessential to Hume’s account of liberty. To say that Humean Necessity is inessential to Hume’s account of liberty, however, is not to say that it is inessential to resolving the controversy at the time. Humean Necessity resolves the controversy over whether the same kind of necessity applies both to free actions and to other natural events. It is this controversy that Hume’s new definition of necessity plays an essential and acceptable role in resolving. The Not Hume Problem is that providing an essential role for Humean Necessity seems to imply that a ‘stronger’ causal relation would rule out freedom, and this is not Hume’s view. While Hume argues that there is no ‘stronger’ causal relation, Hume also believes that free actions would not be compromised if there were a ‘stronger’ causal relation. Hume does appear to believe that the contingency recognized by Humean Necessity explains the intuition exploited by Bramhall that there is contingency in human action, but this does not mean that Hume must agree with Bramhall that ‘stronger’ necessity would threaten free actions. Taking Hume to refer to the Hobbes/Bramhall controversy, when he says that his ‘new definition of necessity’ sheds ‘new light’ on the ‘long disputed question of liberty and necessity,’ solves the interpretive puzzle.

12 Taking Liberty with Humean Necessity

213

12.2.2 Hume’s Discussion of Liberty and Necessity Taking Hume to refer to the Hobbes/Bramhall controversy also makes better sense of Hume’s sections on liberty and necessity in both the Treatise and the Enquiry. Hume’s argument in the Treatise is structured in two stages. Hume first points out that (i) necessity is the constant union of objects and the inference of the mind and (ii) free actions are united together just as regularly and produce inferences in the mind just as often as other natural events. As a result, there is no reason to deny that free actions are necessitated because (i) we can predict free actions just as well as other natural events; (ii) predictions are based upon cause/effect reasoning; and (iii) necessity and cause/effect relations are inseparable. Because Hume has previously argued that Humean Necessity applies to all natural events, Hume’s argument straightforwardly implies that there is only one kind of necessity. Hume’s new definition shows that the same kind of necessity applies both to free actions and to other natural events because, given Hume’s definition, nobody would deny that Humean Necessity applies to both. In this way, Hume’s new definition directly resolves the Hobbes/Bramhall controversy. This is the bulk of Hume’s argument in the Treatise (see pp. 399–407 in Hume 1978). In what remains, Hume explains why people mistakenly deny that necessity applies to free actions. First, people confuse liberty of spontaneity (an internal cause) with liberty of indifference (a lack of necessity) (see p. 407 in Hume 1978). Hume points out that free action requires only the former. Second, people have a false sensation of liberty of indifference when they act and therefore believe they could have chosen whether to do otherwise (see p. 408 in Hume 1978). The sensation is false however, because given enough information about circumstances, character, and motives, human behavior is just as predictable as that of a watch. Finally, people mistakenly assert that morality and responsibility require liberty of indifference (see p. 409 in Hume 1978). Hume shows that responsibility instead requires Humean Necessity because one needs to infer motive and character from action in order to judge another properly. So Russell is correct that this is an essential role of Humean Necessity, but he is incorrect that this is the essential role that sheds ‘new light’ on ‘the controversy.’ Hume’s argument in the Enquiry differs only slightly. Hume argues that because causal necessity is at most union and inference, it therefore applies to free actions just as it applies to other natural events. Hume then defines ‘liberty’ as ‘a power of acting or not acting, according to the determinations of the will’ (see p. 95 in Hume 1975). In the Treatise, Hume defined ‘will’ as ‘the internal impression we feel and are conscious of, when we knowingly give rise to any new motion of our body, or new perception of our mind’ (see p. 399 in Hume 1978). Thus, Hume is referring to liberty of spontaneity in the Treatise. Hume thinks that no one would deny that we have this kind of liberty, the kind required by the Internal Cause Claim and the Anti-libertarian Claim. Hume then argues that there is no liberty of indifference because a lack of necessity implies a lack of cause, which reduces the occurrence to mere chance. This is

214

T.L. Booher

a reductio for Hume because he had previously argued that no philosopher accepts that chance occurs in this sense (see p. 87 in Hume 1975; pp. 403–404 in Hume 1978). Again, the bulk of Hume’s argument is designed to show that one cannot plausibly deny Humean Necessity applies to free actions, just as it does to other natural events. Hume’s sections on liberty and necessity in both the Treatise and the Enquiry are designed to show both that actions we consider free involve Humean Necessity just as other natural events, and this result is required for actions to be free. Hume’s discussion resolves the Hobbes/Bramhall controversy because (i) Humean Necessity consists only of constant conjunction and inference, both of which are obviously present in actions we consider free, and (ii) Humean Necessity is consistent with the intuition that there is contingency in these actions.

12.3 Where Hume Fits in Current Debates The fact that Hume was primarily addressing the controversy at the time also explains why Hume’s chapters on liberty and necessity, standing alone, contribute little to current controversies concerning free action. Neither Hobbes nor Bramhall doubts that there are free actions; they instead disagree over how to account for the fact that there are free actions. Thus, Hume had no reason to address the specific concerns of hard determinists in these chapters.6 Yet Hume’s chapters on liberty and necessity are incomplete precisely because he does not take the concerns of hard determinists seriously. Hume argues that the actions we consider free must involve causal necessity because Humean Necessity (i) applies to all events and (ii) is required for agency and moral responsibility. Hume does not seem to consider the possibility that causal necessity also precludes actions from being free in the sense we care about, that is, in the sense relevant to moral accountability and agency.7 In other words, perhaps regardless of what account we provide of causal necessity, if the same necessity applies to actions we consider free as applies, for example, to falling rocks, then moral accountability and agency are precluded. Hume has argued that if there are free actions, then they must involve causal necessity, but Hume has not argued that there are free actions (actions for which one can be held morally accountable). Yet this is precisely what the hard determinist denies. In fact, the only positive account of free action that Hume provides in his chapters on liberty and necessity is that to be free actions, actions must have the right type of cause (they must be caused by one’s will). Yet such an account is clearly inadequate to address the concerns of the hard determinist: an action caused by one’s will is not sufficient for the action to exhibit the type of freedom needed for moral accountability. Some actions caused by the will are not free in the relevant sense, namely actions performed by kleptomaniacs or small children. Despite such obvious difficulties, numerous scholars have assumed that the Internal Cause Claim represents Hume’s positive account of freedom in its entirety (see pp. 181–182 in

12 Taking Liberty with Humean Necessity

215

Berofsky 2002).8 We should avoid attributing such a superficial account of freedom to Hume if possible. Taking Hume’s positive account of freedom to come from his chapters on liberty and necessity not only saddles Hume with a superficial account, but it also locates Hume within a debate he would have considered misguided. Hume is typically considered a ‘traditional compatibilist,’ belonging to a line of argument later developed by Moore (1912), Hobart (1934), Ayer (1954), Schlick (1966), and Davidson (1973), who advocate a hypothetical or conditional account of freedom. Very roughly, these conditional accounts of freedom hold that if one had chosen to do (or desired or willed) X, then he or she would have done X. Incompatibilists typically respond to traditional compatibilists by arguing that the concern is not that one’s choice (or desire or will) does not cause the action, but rather that if causal necessity applies universally, one has no control over the choice (or desire or willing) in the first place (Kane 1996; Klein 1990). Philosophers then engage in what is at base a dispute over the extent to which conditional accounts can survive logical analysis and yet account for whatever ‘control’ is needed for moral accountability (Lehrer 1966; van Inwagen 1983; Berofsky 2002). For example, Keith Lehrer argues that the analysis of ‘S can X’ cannot be the conditional ‘If C, then S X’s’ because this conditional (whatever C is) is logically consistent with ‘S cannot X’ (Lehrer 1966). Hume’s response to such an argument would not be to provide a more sophisticated analysis of the relevant conditional, which is the response of the so-called traditional compatibilists (e.g., Chisholm 1964; Davidson 1973), but would be to show that such considerations are beside the point. Such logical analyses, on Hume’s view, have no direct implications for moral responsibility: For Hume, what looks like reasoning on moral matters is not a sequence of logically valid inferences, but is directing attention to matters that engage the moral sentiments (see Appendix I in Hume 1975).9 The logical implications of the relevant conditionals simply have no direct moral implications for Hume, but instead have moral relevance only insofar as they affect moral sentiments. Hume argues that it is not relations detected by reason—such as similarities between parricide by oak trees and parricide by humans or between incest involving humans and incest involving other animals (see pp. 466–468 in Hume 1978)—which explain the viciousness of an action or character, but rather ‘a feeling or sentiment of blame from the contemplation of it’ (see p. 469 in Hume 1978). Hume would think that both sides of the debate over the plausibility of traditional compatibilism simply ‘over-intellectualize the facts’ (see p. 73 in Russell 1995; see p. 78 in Strawson 1962). Recall Hume’s claim in the Enquiry that even if we were convinced that the Creator initially caused whatever causes all actions and every action was contributing to an overall good, it would have little effect upon our moral ascriptions because ‘[t]he mind of man is so formed by nature that, upon the appearance of certain characters, dispositions, and actions, it immediately feels the sentiment of approbation or blame’ (see p. 102 in Hume 1975). It must first be determined whether it matters to moral accountability that the Creator is the ultimate cause of actions before it becomes necessary to find creative metaphysical routes to avoid implications for

216

T.L. Booher

moral accountability. Unless incompatibilists’ arguments cause us to disapprove of our overall practice of holding each other (or ourselves) accountable (or at least cause us to do so in individual instances), then such arguments simply have little to do with moral accountability for Hume.10

12.4 What We can Learn from Hume 12.4.1 The Reactive Attitudes Approach There is, however, a contemporary strand of compatibilism that is consistent with Hume’s views: the reactive attitudes approach developed by Peter Strawson (1962). For Strawson, the question of under which conditions an agent is morally accountable is best understood as the question of under which conditions it is appropriate to hold an agent morally accountable. And because the latter conditions are explained in terms of ‘natural human reactions’—that is, susceptibility to reactive attitudes— moral accountability consists in adopting these attitudes toward one another (see p. 67 in Strawson 1962). Strawson argues that we cannot give up such natural reactive attitudes because they play such a central part of our lives, and we should not give up these attitudes because it would impoverish human life, especially interpersonal relationships. Some scholars have responded to Strawson’s arguments by pointing out that, as a matter of fact, (i) human life would not be as impoverished as Strawson claims because he has exaggerated the extent to which our reactive attitudes would have to be altered (Pereboom 1995), and (ii) we could at least give up certain reactive attitudes in appropriate circumstances (Russell 1992). Jay Wallace defends Strawson in light of these attacks (1996). Wallace concedes that we could abandon the relevant reactive attitudes (for Wallace, resentment, indignation, and guilt) but argues that Strawson’s second claim remains persuasive: even if causal necessity does apply to all actions we consider free, there is no reason that we should give up our reactive attitudes or, consequently, our practice of holding agents morally accountable (see pp. 12–17 in Wallace 1996). Wallace points out that incompatibilists make unwarranted assumptions about moral responsibility. Specifically, incompatibilists assume that if causal necessity applies to actions we consider free, then it is thereby unfair to hold agents accountable for such actions (see pp. 85, 110 in Wallace 1996). Wallace argues that in the conditions under which we typically find it unfair to hold agents accountable—for example, insanity, addiction, hypnotism, torture—it is the absence of the ability to grasp and act from moral reasons (either generally or in a specific instance) that explains the unfairness (see pp. 135–194 in Wallace 1996). Wallace claims that agency, which most hard determinists also consider to be threatened by causal necessity, is essentially constituted by the ability to grasp and act from moral reasons (see p. 180 in Wallace 1996). If correct, then the generalization made by incompatibilists from typical excusing or exempting conditions to all action involving causal necessity is

12 Taking Liberty with Humean Necessity

217

unwarranted because causal necessity does not preclude one from grasping or acting from moral reasons. Hume belongs to this strand of compatibilism. As Paul Russell correctly observes, for Hume ‘[o]ne knows an agent is responsible only if one is aware of that person’s causing a certain sentiment of approbation or blame’ (see p. 64 in Russell 1995). Because ultimately all moral assessments concerning the fairness of holding one accountable are derived from moral sentiments, the fairness of holding one accountable for actions involving causal necessity depends upon how this affects moral sentiments. Hume’s account is consistent with much of what Wallace claims.

12.4.2 Compatibilism and Contingency Despite the similarities between Hume and Wallace, however, the differences are significant, and it is beneficial to examine them in some detail. Wallace believes that it is the fact that one grasps and acts from moral reasons that explains (or partially constitutes) the conditions in which our moral sentiments11 are appropriate; whereas Hume believes that grasping and acting from moral reasons is significant only because our moral sentiments so inform us. In other words, for Hume there is no reason why objects that do not grasp or act from moral reasons are inappropriate objects of praise and blame, as it is simply a psychological fact that certain objects in certain circumstances arouse our moral sentiments. Any relationship between moral responsibility and grasping and acting from moral reasons is a contingent one for Hume. For Wallace, this relationship is essential to his accounts of agency and moral responsibility. Hume’s account of agency and moral responsibility has distinct advantages over Wallace’s account. Hume’s account makes better sense of historical changes in the practice of identifying agents and free actions, and permits us to diagnose, and perhaps end, the perpetual stalemate between compatibilists and incompatibilists. Wallace holds that only creatures with an ability to grasp and act from moral reasons qualify as agents (see pp. 159–66 in Wallace 1996). History does not reflect this view. Oliver Wendell Holmes traced the forms of liability from ancient societies through the English common law (1991). As Holmes describes it, many societies have attached liability to inanimate objects and animals when they were ‘the immediate cause of offense’ (see p. 10 in Holmes 1991). The motivation for these practices was ‘vengeance on the immediate offender’ and the resulting judicial process was ‘expressly directed against the object, animate or inanimate’ (see p. 10 in Holmes 1991). For instance, a tree that fell and killed a person ‘was delivered to the relatives, or chopped to pieces for the gratification of the real or simulated passion’ (see p. 11 in Holmes 1991). Similarly, in Southern Asia, if a tiger killed a man, his family was disgraced until ‘they had retaliated by killing and eating the tiger,’ and ‘if a man was killed by a fall from a tree, his relatives would take their revenge by cutting the tree down, and scattering it in chips’ (see p. 19 in Holmes 1991). Under English common

218

T.L. Booher

law at the time of King Alfred, if a horse caused a man to drown, then the horse was surrendered; but if the horse caused the man to drown because the man compels ‘his horse to take the water,’ the horse was not surrendered because the death was the man’s fault; not the horse’s fault (see p. 21 in Holmes 1991). Even later, those in medieval England believed that ‘no difficulty was felt in treating animals as guilty’ (see pp. 300–302 in Blackstone 1765–1769; p. 22 in Holmes 1991). Consider the case of the dog Provetie. In 1595, a dog bit a child’s finger while the child was playing with a piece of meat in his hand (see p. 246 in Arthur and Shaw 2006). ‘The child later died and the dog was arrested, imprisoned, and condemned to death after a trial. The court held that the dog be hanged by “a rope until death ensues” and then taken to the gallows field, “to the deterring of other dogs and to all as an example.” The dog’s goods, should he have any, [were to be] confiscated and forfeited for the benefit of the countship.’ These few accounts from history demonstrate that the objects qualifying as agents (in the sense relevant to morality), as well as the actions that trigger moral accountability, has not remained constant.12 On Hume’s account, these differences should not be surprising. Because our moral sentiments are, at base, a biological reaction rather than a reaction constituted by reason alone, it would seem more surprising if they had not changed over time. Nietzsche makes the point in a different way using social changes to provide an explanation: the cause of the origin of a thing and its eventual utility, its actual employment and place in a system of purposes, lie worlds apart; whatever exists, having somehow come into being, is again and again reinterpreted to new ends, taken over, transformed, and redirected by some power superior to it; all events in the organic world are a subduing, and becoming master, and all subduing and becoming master involves a fresh interpretation, and adaptation through which any previous ‘meaning’ and ‘purpose’ are necessarily obscured or even obliterated. (See p. 77 in Nietzsche 1967.)

Whatever the explanation, history seems to confirm that what (we believe) qualifies as an agent or free action has changed over time. These considerations suggest that it was a mistake for Wallace to identify grasping and acting from moral reasons (abilities other animals and trees do not seem to have) as the essential aspect of agency and moral accountablity. In response, Wallace could simply hold that the examples cited are nothing more than examples of moral mistakes, and thus, such evidence does not negatively impact his views. In fact, Wallace would likely use a distinction he draws between mere causal responsibility and moral responsibility to make this claim (see p. 52 in Wallace 1996). But this distinction merely explains why we (currently) would not consider inanimate objects and other animals morally accountable, as the examples cited indicate that historically, the distinction Wallace draws has been rejected. It is a virtue of Hume’s account that it straightforwardly explains these historical practices for what they appear to be on their face—reflections of moral views now rejected, not moral mistakes now corrected. There is no reason to believe that, upon reflection, these societies would have rejected their practices were they made aware of the distinction Wallace draws.

12 Taking Liberty with Humean Necessity

219

A second consideration in favor of Hume’s account is that by focusing upon grasping and acting from moral reasons, Wallace is open to the following response by incompatibilists: acting from moral reasons (just like having an action caused by one’s will) is not sufficient for an action to be free because insofar as causal necessity applies universally, it is unfair to hold anyone responsible for anything they do regardless of whether they possess these capacities An example of one advancing this type of incompatibilist response is Ishtiyaque Haji. Haji argues that insofar as Wallace’s theory makes a claim about the fairness of holding agents accountable for their actions, it adopts the principle that ‘ought implies can,’ which is not satisfied if causal necessity applies (see p. 209 in Haji 2002). Specifically, Haji argues that the following is a requirement for an action to be morally wrong in the first place: ‘It is morally wrong for S to do [not to do] A only if S can refrain from doing [do] A’ (see p. 209 in Haji 2002). From this, Haji claims that if causal necessity ‘eradicates genuine alternatives, and wrong actions require such alternatives, then the truth of determinism implies that no actions (where action is broadly construed to include choices) are wrong’ (see p. 209 in Haji 2002). For this reason, Haji claims, it is unfair to hold agents accountable if causal necessity applies to their actions. Haji’s argument is nearly identical to incompatibilist arguments advanced in response to traditional compatibilists: Haji’s argument can be summarized as roughly, ‘for grasping and acting from moral reasons to have moral relevance, the supposed agent must have some control over what she grasps or over whether she acts on the moral reasons’; whereas the typical response to traditional compatibilists is roughly, ‘for the (type of) cause of the action to have moral relevance, the supposed agent must have some control over the cause itself.’ A response is available to Wallace: Wallace’s account better captures our concrete moral judgments concerning excuses for and exemptions from moral responsibility, which means that the principle involving alternative possibilities that Haji cites requires independent justification; a justification, Wallace would claim, Haji cannot provide (see p. 196 in Wallace 1996). For the point made here, it is not important whether Haji or Wallace ultimately has the better argument. Instead, it is important to note only that if Wallace is correct, then the fact that causal necessity applies to actions we consider free has no moral relevance whatsoever because it does not affect whether one grasps and acts from moral reasons. Yet this conclusion fails to account for the common intuition Bramhall exploited (and almost all incompatibilists share) that causal necessity is at least relevant to assessing moral responsibility. Wallace attempts to explain this intuition as part of a moral mistake, but, much like Wallace’s response to the historical examples, this response does not seem to take the intuition (which lingers nonetheless) seriously enough. The inadequacy of Wallace’s responses becomes more apparent in light of the fact that the change in which objects society considers agents has nothing to do with grasping and acting from moral reasons. Instead, it is likely the ability to understand the behavior of inanimate objects and other animals mechanistically that explains the historical changes. Any account that ultimately concludes causal

220

T.L. Booher

necessity is wholly irrelevant to moral responsibility and agency fails to capture the very intuition upon which the compatibilist/incompatibilist debate rests. For these reasons, such compatibilist solutions seem open to the charge that they simply beg all of the important questions. Much as Hume’s account mediates the Hobbes/Bramhall controversy, Hume’s account permits us to mediate the Wallace/Haji controversy. Disputes over which objects are agents and which actions are free cannot be resolved in the abstract, but instead present a contingent moral question dependent upon our moral sentiments. Considerations relevant to moral responsibility are those considerations that arouse our moral sentiments when we contemplate them. Recognizing that causal necessity applies to all actions we consider free (just like recognizing the Creator ultimately caused all our actions) has little effect upon our moral ascriptions because ‘[t]he mind of man is so formed by nature that, upon the appearance of certain characters, dispositions, and actions, it immediately feels the sentiment of approbation or blame’ (see p. 102 in Hume 1975). For this reason, Hume believes that, as a matter of fact, causal necessity is not a threat to moral responsibility or agency. By Hume’s own standards, however, this is a contingent fact. For Hume, disputes over whether causal necessity precludes moral responsibility are (at base) not different from disputes over whether a certain upbringing excuses subsequent bad acts in our criminal justice system or whether a genetic predisposition excuses certain behavior. People of good faith can and do disagree about such matters. Even if we conclude that one’s upbringing does not excuse, it remains a morally relevant consideration. Similarly, even if we conclude having a genetic predisposition to engage in certain behavior does not excuse, it also remains a morally relevant consideration. In the same way, causal necessity remains morally relevant even if we conclude that it is not an excuse and does not exempt one from moral responsibility. Wallace is correct that, for example, insanity is different from causal necessity because the former exempts while the latter does not. Wallace is incorrect, however, in his view that to consider the latter an exemption is to make a mistake of reason by employing irrelevant considerations. Haji is correct that Wallace simply ignores morally relevant considerations—roughly, the eradication of genuine alternatives or the power to pursue them—but is incorrect that these considerations preclude moral responsibility and agency altogether. Haji is entitled to claim only that these considerations are morally relevant insofar as they affect moral sentiments, not that these considerations are decisive. Causal necessity is compatible with moral accountability and agency because moral sentiments survive contemplation of the implications of causal necessity. However, this ‘victory’ by compatibilists is not absolute, and cannot be. Because our moral sentiments change over time, the relationship between causal necessity on the one hand, and agency and actions we consider free on the other hand, could also change over time. Thus, while compatibilists seem to have the upper hand, the most they can claim is that hard determinists are currently mistaken, which, after all, is all that is needed.

12 Taking Liberty with Humean Necessity

221

Notes 1 This position is advanced by Thomas Hobbes. Hobbes argues that my actions are free when ‘I can do as I will,’ which is compatible with necessity in free actions (see p. 16 in Chappell 1999). Hobbes does not subscribe to Hume’s definition of necessity, and thus it seems unlikely that Hume’s new definition plays an essential role for this type of compatibilism. 2 Following Hume, I assume throughout that causal relations are ‘inseparable’ from necessity. Because of this, I will sometimes speak of causal relations and other times of necessary connections, but both are implied whenever either is used. I also assume that puzzles involving determinism and free will or free action do not differ significantly from puzzles involving causal necessity and free will or free action. 3 This position is advanced by John Stuart Mill (1874: VI, ii, 3). 4 Russell’s actual argument is as follows (see p. 64 in Russell 1995):

1. 2. 3. 4. 5. 6.

Approval and disapproval are essential to morality. Only character traits or mental qualities arouse our moral sentiments of approval and disapproval. Knowledge of a person’s character traits or mental qualities requires inference. A person or thinking being is held responsible if we regard her as an object of a moral sentiment. Regarding an agent as responsible is, therefore, a matter of feeling not judgment. Without inference to character (i.e. necessity), no such feeling could, as a matter of psychological fact, be aroused in us, and therefore no one could be regarded as responsible.

It is an empirical psychological fact that without necessity morality would be impossible. 5 Anthony Collins and Samuel Clarke had a debate similar to the one between Hobbes and Bramhall. Collins and Clarke made only slightly different distinctions, and Hume’s new definition of necessity similarly resolves their controversy (Clarke 1975; Ferguson 1974; O’Higgins 1976). 6 Although libertarians, as incompatibilists, share some concerns with hard determinists, there was little reason for Hume to take hard determinism itself seriously because the only hard determinist at Hume’s time appears to have been Baron d’Holbach. 7 While there are metaphysical puzzles involving free will and free action that are interesting for their own sake, the primary motivation of most scholars who attempt to solve such puzzles is the perceived threat to moral responsibility and agency (Dennett 1984). As Jay Wallace recognized, ‘[w]hat we want is not freedom of the will per se, but the kind of freedom that makes us persons, or deliberators, or autonomous valuers, or morally accountable agents; a desire for freedom that floated loose from all such contexts would be a kind of fetish’ (1996: 2–3). Galen Strawson shares this view: whether human beings are ever morally deserving of praise or blame or punishment or reward ‘is the only really troublesome question when it comes to the problem of free will’ (2002: 441). Hume was not content merely to show that Bramhall could not posit different kinds of necessity to distinguish free actions. Rather, Hume further argued that a lack of Humean Necessity would undermine moral responsibility. Hume recognized that an account of free will or free action that had no implications for moral responsibility is of little interest. For this reason, I also discuss freedom only insofar as it affects agency and moral responsibility. If it turns out that we have no free will, but this fact has (and should have) no impact on our lives as moral agents, then there remains little reason to continue to argue over whether we have free will. 8 This point is made in detail by Paul Russell (1995). 9 It is noteworthy that Hume’s account, despite its use of the term ‘character’, is consistent with recent claims that predicting behavior is better done by reference to circumstances than enduring traits of the person (Doris 2002). 10 Hume likely would have responded similarly to a related debate over whether alternative possibilities are necessary for moral responsibility (Ekstrom 2002; Frankfurt 1969; Kane 1996). The use of counterexamples in moral debate suggests that issues of moral responsibility admit a precise rational structure, a claim that Hume would deny (see Appendix I in Hume 1975). Perhaps when we consider Frankfurt counterexamples, our moral sentiments disapprove of the conduct described, but when we consider the further information that causal necessity applies to all actions we consider free, our moral sentiments change. Any resemblance between the two cases that may suggest a reasonable person must treat them

222

T.L. Booher

the same way is beside the point for Hume—again, as his parricide and incest examples are designed to show (see pp. 466–68 in Hume 1978). It is a mistake to make rigid distinctions and then to impose a rational structure using those rigid distinctions upon something that has none, namely the conditions admitting of moral responsibility. 11 I use ‘moral sentiments’ only to refer to the relevant moral sentiments corresponding to the reactive attitudes at issue (see p. 33 in Wallace 1996). Wallace recognizes that such a move to what he calls ‘dispositional’ accounts could result in an account ‘only notationally different’ from the one he advances (see n. 11 on p. 33 in Wallace 1996). 12 Bernard Williams explains such differences as follows: ‘We have conceptions of legal responsibility different from any such conception the Greeks had, but that is because we have a different conception of law—not, basically, a different conception of responsibility’ (1993: 65). Wallace also claims that there are instances in which we ‘hold people legally responsible without holding them morally responsible’ (1996: 70 n. 27). While there clearly is a difference between legal and moral responsibility—e.g., holding mentally impaired people civilly liable for damages to ensure that those with a stake in their estate will keep them from harming others—the examples Holmes cites do reflect views on moral responsibility, specifically views about fault. A desire to make ancient practices comport with our own is not a reason to refuse to take these examples at face value.

References Arthur J, Shaw WH (eds) (2006) The case of the dog Provetie. In: Readings in the philosophy of law. Pearson Prentice Hall, New Jersey Ayer AJ (1954) Freedom and necessity. In: Philosophical essays. St. Martin’s Press, New York Berofsky B (2002) Ifs, cans, and free will: the issues. In: Kane R (ed) The Oxford handbook of free will. Oxford University Press, Oxford Blackstone W (1765–1769) Commentaries on the laws of England Vol 1. Clarendon Press, Oxford Chappell V (1999) Hobbes and Bramhall on liberty and necessity. Cambridge University Press, Cambridge MA Chisholm R (1964) Human freedom and the self. In: Watson G (ed) Free will. Oxford University Press, Oxford Clarke S (1975) An essay concerning human understanding. Nidditch PH (ed). Oxford University Press, Oxford Davidson D (1973) Freedom to act. In: Honderich T (ed) Essays on freedom of action. Routledge & Kegan Paul, London Dennett D (1984) Elbow room. The MIT Press, Cambridge MA Doris JM (2002) Lack of character: personality and moral behavior. Cambridge University Press, Cambridge Ekstrom LW (2002) Libertarianism and Frankfurt-style cases. In: Kane R (ed) The Oxford handbook of free will. Oxford University Press, Oxford Ferguson JP (1974) The philosophy of Dr. Samuel Clarke and its critics. Vintage Press, New York Frankfurt H (1969) Alternative possibilities and moral responsibility. Journal of Philosophy 66:829–839 Haji I (2002) Compatibilist views of freedom and responsibility. In: Kane R (ed) The Oxford handbook of free will. Oxford University Press, Oxford Hobart RE (1934) Free will as involving determinism and inconceivable without it. Mind 43:1–27 Holmes OW (1991) The common law. Dover, New York Hume D (1975) An enquiry concerning human understanding. Selby-Bigge LA (ed) Nidditch PH (rev). Clarendon Press, Oxford Hume D (1978) A treatise of human nature. Selby-Bigge LA (ed) Nidditch PH (rev). Clarendon Press, Oxford Kane R (1996) The significance of free will. Oxford University Press, Oxford

12 Taking Liberty with Humean Necessity

223

Klein M (1990) Determinism, blameworthiness and deprivation. Oxford University Press, Oxford Lehrer K (1966) An empirical disproof of determinism. In: Freedom and determinism. Random House, New York Mill JS (1874) A system of logic. Harper & Brothers, New York Moore GE (1912) Free will. In: Ethics. Oxford University Press, Oxford Nietzsche F (1967) On the genealogy of morals. Kaufmann W, Hollingdale RJ (trans). Vintage Books, New York O’Higgins J (1976) Determinism and freewill: Anthony Collins’ A philosophical inquiry concerning human liberty. Martinus Nijhoff, The Hague Pereboom D (1995) Determinism al dente. Nous 29:21–45 Russell P (1992) Strawson’s way of naturalizing responsibility. Ethics 102:287–302 Russell P (1995) Freedom and moral sentiment. Oxford University Press, New York Schlick M (1966) When is a man responsible? In: Berofsky B (ed) Free will and determinism. Harper & Row, New York Strawson P (1962) Freedom and resentment. Proceedings of the British Academy 48:1–25 van Inwagen P (1983) An essay on free will. Clarendon Press, Oxford Wallace RJ (1996) Responsibility and the moral sentiments. Harvard University Press, Cambridge MA Williams B (1993) Shame and necessity. University of California Press, Berkeley

Chapter 13

Rational Choice and Evolutionary Fit Malcolm Murray

Abstract Despite their irrationality, moral agents are more likely to pass on their genes than non-moral agents. This news is spearheading a shift from classical rational choice models to a resurgence of evolutionary ethics. Ignoring its unfavorable association with Social Darwinism, evolutionary ethics is not without its critics. As Maynard Smith observes, ‘A scientific theory – Darwinism or any other – has nothing to say about the value of a human being.’ My role in this paper is to say otherwise. Specifically, I argue against the following charges: (i) Evolutionary models show the success of the wrong thing: heritability, not preferences. For agents who have no specific interest in the benefits of their future progeny, morality will not be motivating. (ii) Computer simulations show ‘success’ under very limiting initial conditions. The probability of those initial conditions is exceedingly low. (iii) The ‘winner’ in artificial games does not sufficiently capture moral agency. (iv) Evolutionary ethics commits the is-ought problem. Morality is a normative notion, whereas evolutionary accounts can only speak in terms of description. Keywords Contractarianism · Evolution · Game theory · Rationality Few believe that morality is a bad thing but there is no consensus on what makes it a good thing. In this paper, I wish to look at two accounts. One is the rationalists’s view. They hold that morality is a rational choice. Agents prone to behave in ways operationally defined as moral do better in terms of net utility maximization compared to those agents prone to behave in ways operationally defined as non-moral or immoral. The second is the evolutionists’s view. They maintain that moral strategies have better fit than immoral strategies. Moral agents are more likely to pass on their genes than non-moral agents and this is so for reasons that have nothing to do with rational choice. I side with the evolutionists. The first part of this paper offers reasons for why the rationalist model fails. The second part of this paper addresses the criticisms aimed at evolutionary ethics. M. Murray Department of Philosophy, University of Prince Edward Island, 550 University Avenue, Charlottetown, Prince Edward Island, Canada C1A 4P3, Tel.: (902) 566-0575, Fax: (902) 566-0420 e-mail: [email protected]

D.K. Chan (ed.), Moral Psychology Today,  C Springer Science+Business Media B.V. 2008

225

226

M. Murray

13.1 The Rational Model Rational moralists argue that adhering to moral constraints is rational. By rational, it is meant that one’s subjective utility is thereby maximized. Generally, it is conceded that morality’s role is to thwart our rational self-interested actions, so the rational moralists must solve this apparent paradox. The motive, by the way, in identifying morality with rationality is to avoid begging the question. Morality would be groundless if we cannot justify our moral conceptions by appealing to anything other than those very moral conceptions. Unlike those who disdain all foundationalist approaches who have no problem with this, moral rationalists are committed to non-question-begging reductions. In other words, rational moralists are committed to claiming that if the conditions under which it is rational to be moral are not met, so much the worse for morality (see, for example, p. 20 in Gauthier 1991). The difficulty the rational model faces is best seen in terms of a Prisoners’ Dilemma (PD). In a PD, the rational move is defection, not cooperation. The problems with Hobbes’s external solution are well documented.1 David Gauthier offers an internal solution by shifting our focus. Instead of choosing to cooperate or defect in a PD, our choice should be between adopting a disposition prone to defect and a disposition prone to cooperate only with other cooperators. The disposition prone to always defect we will call an Unconditional Defector (UD). The disposition prone to cooperate only with other cooperators we will call a Conditional Cooperator (CC).2 To determine which disposition it is rational to adopt, we pit the two dispositions together in PD competitions and see which disposition yields the highest utility payoff for the individual players. The winner is CC. From this, Gauthier asserts that morality is rationally justified. We can ground morality without appealing to moral presuppositions. This bold assertion depends on (i) the robustness of the conditions under which CC prevails, (ii) whether CC is really a rational strategy, and (iii) whether CC is a sufficient representation of moral agency. (i) For CC to prevail, the following conditions must be met. (a) There are enough other CCs with whom to interact. (b) There are no computation costs that CCs have compared to UDs. And (c) the ability for CCs to correctly identify UDs is sufficiently accurate. If these conditions are not met, it would be irrational to adopt the CC disposition. The response to (a) is to note that since CCs do at least as well as UDs, the likelihood of a CC invasion increases, and this invasion will exponentially increase since the more CCs there are, the better CCs will do. The response to (b) and (c) is to show that CC can still beat out UDs even with computation and detection costs. For example, assuming an initial population of 20% CCs and 80% UDs, and a 20% error in CC’s correctly identifying a UD, it will still be more rational to be a CC than a UD.3 If errors in detection are set at 10%, CC’s can beat out UDs with an initial population proportion of only 10% CCs to 90% UDs.4 Notice however that the reply to (a) depends on our accepting the reply to (b) and (c), and the reply to (b) and (c) depends on our accepting the reply to (a). Enough

13 Rational Choice and Evolutionary Fit

227

CCs can get into the mix only if the computation and detection costs are not penal, and that computation and detection costs are not penal so long as there are enough CCs in the mix. The problem has not gone away. How does one get enough CCs in the mix in the first place? Assuming even minimal computation and detections costs, those initial CC agents would have to be irrational: they will do worse than being a UD. CC’s rationality is predicated on the irrationality of the first CC agents. (ii) This leads to the second worry. If the success of CC depends on initial irrationality, in what sense can we call CC agents rational? Even ignoring the problem of the first CC agent, it is always irrational to cooperate with another cooperator when unilateral defection will earn the defector greater individual utility.5 The rational move is precisely what CC agents prevent themselves from doing. Gauthier’s response to this charge is that traditional theories of rationality have been developed within parametric choice situations: they are inapplicable under strategic choice situations. The rationality of CC is a logical extension of theories of rationality under strategic choice (see p. 183 in Gauthier 1986). Peter Danielson takes another route. He argues that theories of rationality are themselves open to refutation or modification, and if CC is shown to do better than the disposition that the accepted theory of rationality would select, then that would be a reason to abandon or revise our accepted theory of rationality (see pp. 62–3 in Danielson 1992). The worry about whether CC is rational in the traditional sense is however irrelevant. Even by Gauthier’s own model of rationality, CC is irrational. This follows once we admit the possibility of further dispositions than UD and CC. For example, let us introduce into the mix two more strategies, Unconditional Cooperators (UC), and Reciprocal Cooperators (RC). A UC will cooperate no matter what the other player will do. An RC will cooperate only on the condition that the other player’s cooperation is conditional upon her own cooperation. That is, RC will cooperate with CC agents (and other RC agents), but not with UD nor – and this is the important part – with UCs. UC will cooperate no matter what. But if the reason to constrain oneself to cooperate depends on future cooperative ventures with this player, the motive to cooperate with UC vanishes. In other words, RC will exploit UCs, defect against UDs, and cooperate with CCs and other RCs. Precisely because of RCs’s exploitation of UCs, RC will do better in terms of utility than CC. Now this result would not matter if we deemed RC at least as moral as CC. But since we are not likely to call moral a strategy whose success depends upon the exploitation of innocent parties, we discover a case where the rational strategy veers from the moral strategy. The desire for a rational morality fails. (iii) The third worry concerns the morality of CC. That is, even if CC were a rational strategy and we are happy to accept the initial conditions that lead to CC’s success, all is for naught if we are unlikely to accept that CC represents moral agency. This worry may be presented in two ways. First, imagine a case where Dick, who is drowning, asks passerby Jane for help. Suppose, too, that both Dick and Jane are CC agents. If Jane is a CC, we know that once Jane accepts Dick’s request, Jane will be bound to carry through on her promise (assuming Jane is able to rescue Dick without harm to Jane). But merely being a CC agent does not commit Jane to agree to help Dick in the first place. Although CCs are committed to keep

228

M. Murray

those agreements they make, nothing dictates they make them. As a CC agent, the only agreements Jane is willing to make are those in which she stands to benefit. If there is no benefit in entering the agreement, rationality dictates she decline. Merely being a CC does not commit one to accept the terms of every agreement proposal. Thus, Jane’s being a CC does not commit Jane to accept Dick’s request for help. If we want an argument that commits Jane to help Dick in this sort of case, appealing to the rational merits of CC is not equal to the task.6 CC may capture some aspects of what we mean by moral agency, but not all. Secondly, we may return to the introduction of UC and RC. UCs are not rational since they set themselves up for exploitation by UDs. UCs will certainly do worse than CCs, but one could argue that UC better captures our understanding of moral agency than CC. After all, Jesus recommended turning the other cheek, and although we can doubt both the divinity and rationality of Jesus, few will argue that Jesus is less moral than the rest of us.7 Matters get worse. In terms of Gauthier’s and Danielson’s understanding of rationality, RC agents are the most rational. For the argument that rationality and morality are integrally linked, we should expect that the most rational agent will also be the most moral. In the case of RC, this condition fails. One might try to rescue the rational morality model by arguing, as Danielson did, that RC is in fact more moral than CC or UC. The existence of UCs allows UDs to prosper. Since UDs are operationally defined as immoral, UCs abet immorality. It is not uncommon to treat abetters of crimes as criminals themselves. Therefore, moral agents should punish both UDs and UCs (see pp. 114–18 in Danielson 1992). Only RCs satisfy this condition. This may seem harsh. Normally, punishing the innocent to get at the guilty is a miscarriage of justice. It is akin to stealing from the innocent in order to remove the temptation for thievery.8 On the other hand, ‘Stop the Silence Against Violence’ is a common slogan used in the campaign against violence to women. This message is aimed, not merely to the abusers, but also to those who remain silent about the abuse. The moral response is to boycott all interaction with abusers, yet UCs continue to cooperate with abusive UDs. It is part of our role as moral agents to not tolerate immorality even when the immorality has no direct effect on us. This is a case where UC’s ‘doing nothing’ is immoral. This reply strikes me as illegitimate, however. It imputes a moral motive to an action we are supposed to believe is motivated by rationality alone. Even if punishing UCs successfully rids UDs, this cannot be the motive behind RC’s action. In terms of rational self-interest, since RCs are immune to UD’s defection, RC could not possibly care whether UDs exist. Therefore, the motive to punish UC to get at UD is a motive beyond RC’s rational arsenal.9

13.2 The Evolutionary Model The failure of the rational model should not surprise us. Rational strategies do not generally prevail in nature. Evolution tends to favor simpler – often patently irrational – strategies. The evolutionary model of ethics maintains that despite their

13 Rational Choice and Evolutionary Fit

229

irrationality, moral agents are more likely to pass on their genes than non-moral agents. As Brian Skyrms demonstrates, evolutionary dynamics favors operationally defined moral agents who use ‘weakly dominated, modular irrational strategies’ (see p. 32 in Skyrms 1996). Admitting that CC is not technically rational has no impact on the evolutionary model: they never make such a claim. Evolutionary success of irrational strategies is the norm, not the exception, in nature. There are countless cases where the ideally rational move is not what evolution sells. In any event, the rationality of morality is not what we really need to show in order to understand the origins of morality in terms of non-moral presuppositions. Beyond Skyrms’s results, the findings of Alexander (2001), Axelrod (1984), Danielson (1998) independently show how moral development has evolutionary success, despite the seeming paradox that morality curtails (some of) our purely self-interested actions. However, similar problems that face the rational model may be applied to the evolutionary model, and so before we rest content with the evolutionary game theory’s successes, we need to show why the evolutionary model can adequately respond to these charges in ways that the rational model cannot. I shall consider the following four problems with connecting morality to evolution. (i) To advocate irrational strategies seems patently incoherent. Besides, evolutionary models show the success of the wrong thing: heritability, not preferences. For agents who have no specific interest in the benefits of their future progeny, morality will be unmotivating. (ii) The problem concerning favorable initial conditions that plagued the rationality model has not been avoided by the evolutionary model. (iii) The problem of whether the ‘winner’ is really moral has also not been avoided by the shift to evolutionary models. (iv) Evolutionary ethics commit the is-ought problem. Morality is a normative notion, whereas evolutionary accounts can only speak in terms of description. (i) Dismissing the notion of rationality may seem wildly premature. The received view of rationality entails maximizing one’s personal utility. Therefore, acting rationally will secure for me more of what I prefer. So to say I might not be interested in following the advice of rationality is plainly to confess that I do not prefer that which I prefer: an incoherence. Reply. I accept this. Still, to be rational requires brain power. To have sufficient brain power requires having the requisite brains. To have the requisite brains requires having the requisite bodies to maintain it. To have the requisite bodies, requires having the requisite resources. These are scarce, and in any event fluctuate. A cheaper model may be adequate. In nature, rational-like individuals may evolve, while ideal rational agents may not. So long as certain heuristics can develop which work in normally occurring situations, the speed and convenience of cheap heuristics may benefit agents who have developed them more than agents who have not. The cost of ideal rational models may not be sufficiently compensatory, given the environments they generally inhabit. A caveat is needed. A danger of appealing to simple strategies in social interactions is that they tend to be less sensitive to their external environment. Since environments are variable in both the short term and long term, plasticity is necessary if organisms are to survive to reproduce. The danger of relying too much

230

M. Murray

on heuristics is to rely too much on the chances that the environment in which the heuristic works will remain. If biology teaches us anything, that assumption is not to be expected. In fact, precisely because environments are not stable, variability is essential to biological species. This does not mean we will abandon heuristics. It is not a matter of choosing a context-insensitive heuristic over an ideal rationality. Successful heuristics will be those more sensitive to environmental fluctuations. Let us say that to be fully sensitive is to be rational in the ideal sense. But sensitivity to one’s environment lies on a continuum. Fitness is a relational concept. Saying ‘Trait A has fit,’ means only that trait A has greater fit than other available traits, not that it is the best conceivable trait. In card games, both a king and a three beats a two: to say the king beats the two better than the three is to misunderstand the relation. Fitness can be achieved without reaching an ideal sensitivity the way normative rational models demand. By the way, if evolutionary fit does not favor moral development, there is nothing in the evolutionary model to justify our crying about it. As with the rational model, we would simply shrug our shoulders and say, so much the worse for morality. This may seem unbearably harsh. The difficulty in accepting that evolutionary fit should trump morality should the two conflict stems, I suspect, from the Ancients’ drive to show morality has intrinsic worth. An evolutionary account can only understand moral development as having instrumental benefit, if it is to have any benefit at all. Even if a trait produces a preference such that the satisfaction of that preference yields an intrinsic benefit to the individual carrying that trait, the intrinsic benefit is itself understood in terms of its instrumental fitness. A similar conceptual snag prevents people from appreciating how damaging the problem of evil is to religion. One of the usual religious responses is that suffering, whether by sin or natural disaster, allows us to adopt the moral response, a thing good in itself, a thing that outweighs the suffering to innocents. But the alternative is to understand morality as a successful strategy in the face of suffering, so that in the absence of suffering, we would not require morality. If morality is viewed as a response that benefits human agents, removing that benefit is to remove the value of morality. So although morality is not rational on the evolutionary model, morality is still seen as instrumentally beneficial – and that is as sufficient a demonstration as we need. Some will object that the thing instrumentally benefited in the evolutionary model is not the same thing instrumentally benefited in the rational model. Specifically, evolutionary fit has to do with success in trait propagation, which may have nothing to do with one’s preferences. My preferences are not satisfied merely on the basis that I behave in such a way that my progeny will prosper. Therefore, evolutionary models show the success of the wrong thing: heritability, not preferences.10 We may call this a motivation problem, but if what I have shown in section 1 is right, the ‘rational’ solution is not motivating either. Cooperating with a cooperator does not track one’s preferences. In a single PD interaction, defection will track one’s preferences better than cooperation. If one prefers to cooperate, that preference will have to be built into the utility stipulated by the structure of the game, and if this is done, one is no longer playing a PD. The rationalists require one to adopt a disposition that will avoid succumbing to the temptation of always following one’s occurrent

13 Rational Choice and Evolutionary Fit

231

preferences. Rationalists will argue that it is at least the long term benefit to the individual that the rationalist’s model tracks, whereas it is the propagation of a trait that the evolution model tracks. In other words, there is a closer relation between the future individual to the present individual than grandchildren to grandparents. But as far as tracking current preferences go, the evolutionary model does better: one’s preferences are part of one’s heritage. The issue is no longer whether you ought to adopt a certain strategy to best track your future preferences, as if you have a choice. The question is whether your ancestors have survived partly because of this hereditary trait, passed on through convention. Therefore, the evolutionists can say moral behaviour actually tracks the preferences of those in which moral traits have evolved.11 Whereas rationalists cannot say moral behaviour tracks current preferences of rational agents, only that agents would do better in the long term if they constrain preference satisfaction under strategic choice situations – if even then. (ii) The ‘success’ of evolutionary models occurs under very narrow conditions. In Skyrms’s Ultimatum Game, for example, moral agents do better than immoral agents so long as we begin with a greater percentage of moral agents. Starting with equal population proportions, immoral strategies win. In fact, altering Skyrms’s initial population distribution by even .001 is enough to tip the scales back in favor of immoral agents.12 Skyrms has identified a condition in which fair minded persons will have evolutionary fit, but he has also (unwittingly) shown that the odds are not in its favor. All that we can so far say then is that moral-like behavior has evolutionary fit, or that moral behavior has evolutionary fit in exceedingly rare cases. These are not overwhelmingly favorable results for tracking the evolution of moral behavior. Reply. As already noted, evolutionary models need not rely on the rationality of becoming a CC agent prior to there being enough CC agents for CC to prosper. Random mutation is all that is required for an answer. Analogously, the sheer unlikeliness that you will win the lottery does not undermine the odds of someone winning the lottery. The sheer unlikeliness of earth supporting life does not undermine the odds of some planet supporting life. Likewise, the sheer unlikeliness of the favorable initial conditions supporting moral generation is not itself a refutation of the explanation. Why should we anticipate favorable initial mixes? ‘Fluke’ is as sufficient an answer as we need. After all, that’s an acceptable answer in general talk of evolution. We do not require a reason for a mutation. Natural selection entails the interrelation between fitness and heritability. Trait x is fitter than trait y if and only if x has a higher probability of survival and/or a greater expectation of reproductive success than y. The fitness of an organism concerns some characteristic that makes a difference in its ability to survive and reproduce (see p. 70 in Sober 1993). Mere survival is not sufficient if the organism does not reproduce. Moreover, the reproduction must carry the fit trait. Heritability concerns some way to ensure the offspring resemble their parents in terms of fitness. Survival in one generation must carry over to future generations before we say that trait has fitness. A problem with rational models is their focus is too narrow. It is rational to defect in a Prisoner’s Dilemma. It is rational to demand 9 pieces of the 10-piece pie for oneself in the Ultimatum Bargaining Game. Neither of these traits are what we would call moral

232

M. Murray

strategies. But the success in terms of survival of these immoral strategies backfires in terms of heritability. The greater proportion of Defectors in the population means the fitness of Defectors decreases. Strategies that prosper among their own kind will do better in terms of reproductive advantage than strategies that do not. The ‘fluke’ answer may still appear too glib for many. Another kind of reply is available. The above games rely on regulated encounters: that is, each individual player will interact with every other individual player in the competition exactly once per round. In nature, no such regulated interaction exists. Rather, interactions are, to use the language of Aumann (1974, 1981), ‘correlated’. That is, individuals interact with only a select number of other individuals. It will not matter in terms of end results whether the selection criterion is in terms of correlating with like-strategies or simple geographic convenience, or both. Such differences will affect the time it will take to reach a Nash equilibrium. As with the pure fluke reply though, strategies that can prosper among their own kind will do better in terms of reproductive advantage than strategies that do not. So long as interactions in nature are more correlated than random, as seems to be the case, evolutionary models of Prisoner’s Dilemmas and Ultimatum Games favor moral strategies, not immoral strategies, even when the immoral strategies outnumber the moral strategies in the initial population proportions (Skyrms 1996, Alexander and Skyrms 1999). (iii) Evolutionary models can no more support our robust understanding of morality than the rationalist model. Although moral agents win in Skyrms’s Ultimatum Game, UCs are losers and RCs winners when applying replicator dynamics to PDs. Reply. Assuming equal initial population frequencies across the four dispositions (UC, UD, CC, and RC), after 1,837 generations, an equilibrium is reached in which CCs consume 42% of the population and RCs 58%.13 Superficially, RCs still ‘win’, but notice that both UDs and UCs have become extinct. If UCs are extinct, our intuitions against RCs’s exploitation of UCs is contaminated. The evolutionary model predicts that we have no UCs, and so any moral indignation we feel is mis-directed, or at least suspect. At any rate, without UC, RC and CC will not differ in practice. Furthermore, so long as CCs outnumber RCs and UDs outnumber UCs in the initial population mix, CCs will in fact do better than RCs. For example, if the initial population consists of 20% UCs, 30% CCs, 20% RCs, and 30% UDs, an equilibrium is reached after 1,965 generations in which CCs consume 53% of the population and RCs 47%. Evolutionary fit is in fact better aligned with moral strategy than the rational model. Some will say that this hardly helps the evolutionary model. The success of RC is not a demonstration that RC captures our full conception of morality. In other words, objectors might insist that if there were UCs, we would think it morally wrong to exploit them. Therefore, the evolutionary account cannot capture morality. We should be cautious, however, against dismissing theory whenever theory rubs against our intuitions. Our intuitions may be our problem. How one ought to behave within a community of CCs does not tell us how to behave in some other community. That we might think so is due to our making a useful strategy in one context an overarching strategy in all contexts. Let us call such generalizations overextended

13 Rational Choice and Evolutionary Fit

233

heuristics. The moral intuition against exploiting UCs may be nothing more than an overextended heuristic. Examples of overextended heuristics in nature abound. The program in a greylag goose to retrieve stray eggs to her nest can get fooled by footballs, skulls, and lightbulbs. The drive for male red-winged blackbirds to defend their territory is triggered not merely by the red wings of invading male red-winged blackbirds, but red balls and red shirts of passing joggers. The sexual attraction to bigger claws in Fiddler crabs results in offspring that are both more attractive to and less able to avoid predation. Vibrations on a web will trigger eating behaviors in female spiders that need to be circumvented by amorous male spiders (by repeated tapping on the web until the eating response is saturated). The success of a hawk in a world of doves is not carried over in a world of Hawks. In cases of humans, our poor use of probability reasoning is a prime example of an overextended heuristic in action. So long as the heuristic normally benefits agents, or normally does not impede agents, the heuristic will thrive. This is the same with certain of our moral intuitions. We may well end up with a distinction between ideal morality and evolutionary ethics, but nothing commits us to move toward the ideal morality merely by pointing out this gap. If the ‘successful’ disposition does not perfectly match the going account of morality, we could conclude either that evolutionary models have nothing to teach us about morality, or that our ideals of morality are overextensions of the merits of moral heuristics. Taking the evolutionary model seriously commits us to the latter. We do not say, ‘Because the egg retrieving heuristic has fit, there is fit in retrieving broken beer bottles.’ Nor should we infer that since moral dealings is good (due to its fitness) in certain circumstances, its extension to all circumstances is also good. That we develop an overextended heuristic that presents itself as an intuition for how we ought to behave in counterfactual conditions cannot thereby be particularly telling. (iv) A descriptive account of evolutionary fitness cannot possibly provide any insight into the normativity of ethics. Even Maynard Smith observes, ‘A scientific theory – Darwinism or any other – has nothing to say about the value of a human being’ (see p. 374 in Smith 1998). Evolution seems to show that groups are differentiated and some are destined for extinction. To infer from this that it is therefore right that they be so extinguished (as the Social Darwinist movement professed) is to unfoundedly assume that what is is also what ought to be. Reply. It is certainly true that evolutionary ethicists wish to distinguish themselves from Social Darwinists. Under the Social Darwinism of Herbert Spencer,14 for example, any non-fit allele is deemed immoral because it either fails to contribute to fitness of the group, or it is counter to fitness of the group. Thus, we ought to scrap welfare programs and health programs because those who are poor or sickly are hardly fit. To have the fit take care of the unfit is to artificially bolster the continuance of unfit alleles. Whereas if the goal of evolution is to reach a pinnacle of supremely fit genes, then we should not thwart that process by artificially allowing the propagation of unfit alleles. The problem with the Social Darwinists is not (just) with making bad normative conclusions from science: they do not get their science right either. The concept of progress in evolution must not be confused with intentionality. The norm in evolution is extinction, not a continuous direction to

234

M. Murray

perfection. Rare alleles become common; common alleles become extinct.15 There is no direction or purpose in evolution. A duty to remove that which does not fit the norm presupposes there is a norm. The Social Darwinists made another mistake. They accord rights to those who are fit: the unfit have no rights to protect them. But this gets the picture backward. It is not the case that moral rights are accorded to those who are fit, as if morality were a prize awarded to the most deserving. Rather, those who employ conditional moral strategies are those who tend to have more fitness. In the same way, we do not accord tool use to only those with opposable thumbs. Rather, those with opposable thumbs are able to use tools, and this enhances fitness. Similarly, the Social Darwinists failed to consider our ethical beliefs as themselves part of our evolution. To say we ought to excise the sick and infirm since they do not contribute to good offspring is to miss the fact that we tend to find such sentiment morally repugnant. From where did this repugnance originate? It strikes me that a larger account of evolutionary dynamics should have at least something to say about this, and the work of Skyrms, Binmore, Alexander, and Danielson, to name a few, have tracked it in terms of conditional, strategic interaction. Two cautionary remarks are needed to avoid misunderstanding. The first is to admit that not all features of a given species necessarily contribute to that species’s survival. For example, hair growing out of one’s ears does not obviously confer fitness. Since nature generally abhors unnecessary appendages, however, the continued existence of a strategy tends to speak for itself. Therefore, given the fact of our moral sentiments, the odds favor the attribution to morality of a contributing role for our kind. Evolutionary fit is the right ordering of the right amount of a vast number of characteristics and dispositions – including social behaviors. If inherent value is something we ascribe to human beings, then it is not improbable that that ascription has evolutionary fit. To speak of the ascription of value to human beings is to speak about one of the features (our propensity to ascribe value to human beings) that may be part of the Homo Sapiens arsenal for survival. If so, one would think evolutionary theory can account for it. After all, if the moral values we routinely ascribe to Homo Sapiens has no connection to the evolutionary stability of Homo Sapiens, this would need explanation. The second caution concerns the charge that evolutionary theory is tautologically vacuous. Fitness is defined in terms of actual survivability, while survivability is predicted on the basis of fitness. But such a tautology is no more problematic to evolutionary theory than the assumption embedded in scientific research that everything has a natural cause. This statement needs to be accepted in order to pursue scientific research; it is not itself an outcome of scientific research. It is the same with the link between fitness and survivability. The worry dissolves, I think, once we jettison the notion that fitness has any non-relative significance. Fitness cannot be understood in a vacuum. Moreover, X’s current survival does not entail X’s fitness, since fitness concerns both survivability and heritability. A mutant may exist now that is unable to propagate, hence its current survival does not entail its fitness. A trait now at fixation in some population may have reached fixation for any number of reasons (see p. 70 in Sober 1993). And even if trait X’s survival is due to its past fitness, it does not

13 Rational Choice and Evolutionary Fit

235

guarantee it will have continued fitness. Having fitness in one environment does not guarantee its fitness in a changed environment. And what would have fitness now may have become extinct before an environment favorable to it arose. Still, the concept of a trait surviving that has no fitness makes no sense. If moral behavior really disadvantages owners of that disposition in terms of evolutionary fit, then we should predict the weeding out of morality. Conversely, if morality has remained, we should presume our ability to track its evolutionary benefits. That fitness is a relative notion means that if morality has fit, we can only predict its fitness under the narrow conditions in which we live. Altering any of those conditions undermines the prospect that morality may have fit. This may not strike one as the right sort of move. It resurrects the complaint concerning our hypothetical treatment of UCs. It is precisely because normative talk is not apt in biology that evolutionary models are not apt in morality, say the critics. But one need not deny that is cannot by itself imply ought to suggest that evolutionary models can have something to tell us about ethics. The idea that morality must have normative force may itself be a nomological dangler that should be eradicated along with the notion that moral behavior is rational. If so, the suggestion that evolutionary models cannot accommodate what we mean by ‘real morality’ need not bother us. After all, when we explain a conjurer’s trick, the reply, ‘But that isn’t real magic,’ is ignored. That we have not explained real magic will not bother those of us who dismiss its existence. We have explained the semblance of magic, and that is enough.16 The complaint that the evolutionary model fails to accommodate what we really mean by morality should, perhaps, receive the same rebuttal.17

Notes 1 Hobbes is thought to have provided the insight needed to tie morality to rationality. According to Hobbes, even for self-interested reasons, people would be willing to do whatever it takes to escape an amoral existence. Since it will always be rational to renege on one’s promises, promises to cooperate without the sword are but empty words. To solve this stalemate, Hobbes inserted a sovereign to punish defectors. Without a sovereign, the Nash Equilibrium point (the point at which any unilateral motion costs the mover) in Prisoner’s Dilemmas (PD) is mutual defection, but with a Sovereign who punishes defection, the Nash Equilibrium point is mutual cooperation. Under a sovereign, no one can do better than mutual cooperation by unilateral movement. Problems with the sovereign solution are many and well documented. Here are three. First, an incoherence exists. The claim is that we cannot keep agreements without the sovereign, but presumably to employ the sovereign we have to keep our agreement in electing the sovereign: an impossibility. Alternatively, we can keep our agreements without a sovereign, in which case the sovereign solution must be unnecessary. As an aside, Hobbes must accept this latter position, since he speaks of both confederacy and family bonds in the State of Nature. (‘We each have strength enough to kill another through confederacy’ (see XIII, para. 1, 63 in Hobbes 1988), and families may be formed and maintained ‘through lust’ (see XIII, para. 11, 65 in Hobbes 1988). A second worry with the sovereign solution entails costs. The cost of supporting a sovereign may outweigh the benefits gained. Depending on the tax rate, mutual cooperation under a sovereign may yield the same utility as mutual defection in the State of Nature. A third worry, raised by Gauthier, is that erecting a sovereign is merely a political solution, not a moral solution (see p. 164 in Gauthier 1986). For a moral solution, we need to show self-constraint. We need a way of showing why internalizing the disposition to cooperate in PDs is rational.

236

M. Murray

2 Gauthier calls the first a Straightforward Maximizer (SM) and the second a Constrained Maximizer (CM) (see p. 167 in Gauthier 1986). 3 In 74 generations, CC will take over the population: UDs will become extinct. 4 CC will take over the population in 173 generations in this case. 5 Binmore is, perhaps, the most insistent about this point (see pp. 27, 103, 114 in Binmore 1994). 6 Gauthier, by the way, does not think this is a real problem. Along with asking whether it would be prudent to adopt a disposition to comply to agreements one thinks it is rational to make, the standard contractarian talk appeals to ex ante agreements which encompass all of our social rules, including the rule to be charitable (For example, see p. 23 in Gauthier 1991). 7 Assuming Jesus is a UC, Gauthier would say Jesus is not moral, since he believes morality requires constraint (see p. 2 in Gauthier 1986), and there is no constraint in UC agents. 8 Or see Gauthier’s reply: ‘If the farmer, in order to protect his chickens from the foxes, eliminates the local rabbits, thus reducing the foxes’ overall food supply and encouraging them to go elsewhere, it does seem rather hard on the innocent rabbits.’ (See p. 401 in Gauthier 1988). 9 On the other hand, since the existence of UC benefits UD, and the more UDs there are, the worse RC does, it might be maintained that RC’s exploitation of UC is rationally motivated. I would rather say RC’s selfish motive to exploit UC has a beneficial side-effect of reducing the UD population. 10 Chris Tucker and Stephen Finlay have independently made this objection in correspondence. 11 So long as choice reveals preferences. More is needed to make this case, admittedly. 12 The Ultimatum Game has one player offer a division of a 10-piece pie and the second player accept or reject the offered division. If the second player rejects the offered division, neither player gets any of the pie. Limiting the offers to a 9-1 split and a 5-5 split, and assuming a strategy governs what one should do when one plays first as well as what one should do when one plays second, there will be the following eight strategies: (S1) Demand 9; Accept All. (S2) Demand 9; Reject All. (S3) Demand 9; Accept 5, Reject 9. (S4) Demand 9; Accept 9, Reject 5. (S5) Demand 5; Accept All. (S6) Demand 5; Reject All. (S7) Demand 5; Accept 5, Reject 9. (S8) Demand 5; Accept 9, Reject 5. Only S1 and S7 need interest us. S7 represents moral behavior, while S1 captures what rationality would demand, and so here is a case where the rational choice is not the moral choice. The only condition in which S7 wins, however, is when the initial population proportions across the eight strategies are as follows: S1 = .32, S2 = .02, S3 = .1, S4 = .02, S5 = .1, S6 = .02, S7 = .40, S8 = .02. With these initial population proportions, S7 ends up with 56% of the total population along with another moral-like strategy S5 at 44%, while the rational S1 strategy goes extinct. If we alter S1 to .321 and S7 to .399, while leaving unchanged the remaining population proportions, the immoral S1 completely takes over the population in 78 generations. For a note on the formula for deriving these results, see pp. 51–53 in Skyrms (1996). 13 The formula for tracking the Replicator Dynamics for strategy A is U(a)p(a)/U. U(a) = the average utility of strategy A. p(a) = the current proportion of players using strategy A. U = the average utility of all players in the competition (hence: Σ[U(ai )p(ai )]). Let us call p(A)i the next generation proportion for A. Then p(A)i = U(a)p(a)/U. To put the formula into practice, we need a table (Table 13.1) charting the reverse ordinal ranking for our payoff matrix for the four strategies, UC, UD, CC, and RC, showing individual scores for row while playing against column.

Table 13.1 Payoff matrix UC UD CC RC

UC

UD

CC

RC

avg

3 4 3 4

1 2 2 2

3 2 3 3

1 2 3 3

2 2.5 2.75 3

Let u, d, c, and r represent the population proportion of UC, UD, CC, and RC respectively. Using the default payoff matrix from Table 13.1, the formula for determining the average individual scoring (U(a)) is as follows:

13 Rational Choice and Evolutionary Fit

237

U = 3(u + c) + 1(d + r) D = 4((u) + 2(d + c + r) C = 3(u + c + r) + 2(d) R = 4(u) + 3(c + r) + 2(d). Notice this does not exclude self-play. So long as the raw numbers of players is sufficiently large, this omission should be negligible. Beginning with equal population proportions (.25 for each of our four strategies), the following average utilities (U(a)) emerge: U = 3(.25 + .25) + 1(.25 + .25) = 2 D = 4(.25) + 2(.25 + .25 + .25) = 2.5 C = 3(.25 + .25 + .25) + 2(.25) = 2.75 R = 4(.25) + 3(.25 + .25) + 2(.25) = 3 The average utility of the population (U) is determined by the following formula: U = U (u) p(u) + U (d) p(d) + U (c) p(c) + U (r) p(r). Given the equal initial population proportions, the average utility of the population (U) is so: U = 2(.25) + 2.5(.25) + 2.75(.25) + 3(.25) = 2.5625 Now, we have all the ingredients for the replicator dynamics to determine the population proportions for each of our dispositions in the second generation p(A)2 . Using our formula: U(a)p(a)/U, gives us the following: p(UC)2 = U (u) p(u)/U = 2(.25)/2.5625 = 0.195 p(UD)2 = U (d) p(d)/U = 2.5(.25)/2.5625 = 0.244 p(CC)2 = U (c) p(c)/U = 2.75(.25)/2.5625 = 0.268 p(RC)2 = U (r) p(r)/U = 3(.25)/2.5625 = 0.293 The cycle is repeated using these new population proportions for the subsequent generation, and the process continues until an equilibrium is reached, when P(A)i = P(A). 14 Spencer (1857, 1862). For a revival of Spencer by appealing to Lamarckianism as opposed to Darwinianism, see Richards (1988). 15 ‘Evolution is the process by which rare alleles become common, possibly universal, and universally distributed alleles become totally eliminated’ (see p. 392 in Hull 1998). 16 Daniel Dennett offered this rebuttal to those who criticize his explanations of consciousness and free will (Dennett 2003). I apply it to morality. 17 I very much appreciate the commentary I received on this paper by David Chan, Stephen Finlay, Chris Tucker, David Velleman, and Paul Viminitz, as well as from many in attendance at the Conference on Values, Rational Choice, and the Will at the University of Wisconsin – Stevens Point, April 3, 2004. A much expanded and more detailed version of this paper appears in Murray 2007 as Chapters 4 and 5.

References Alexander J (2001) Group dynamics in the state of nature. Erkenntnis 55:169–182 Alexander J, Skyrms B (1999) Bargaining with neighbors: is justice contagious? Journal of Philosophy 96(11):588–598

238

M. Murray

Aumann R (1974) Subjectivity and correlation in randomized strategies. Journal of Mathematical Economics 1:67–96 Aumann R (1981) Correlated equilibrium as an expression of Bayesian rationality. Econometrica 55:1–18 Axelrod R (1984) The evolution of cooperation. Basic Books, New York Binmore K (1994) Game theory and the social contract, volume 1: playing fair. MIT Press, Cambridge MA Danielson P (1992) Artificial morality: virtuous robots for virtual games. Routledge, New York Danielson P (1998) Evolutionary models of co-operative mechanisms: artificial morality and genetic programming. In: Danielson P (ed) Modeling rationality, morality, and evolution. Oxford University Press, Oxford Dennett D (2003) Real consciousness, real freedom, real magic. Julian Jaynes lecture, University of Prince Edward Island, 17 Oct 2003 Gauthier D (1986) Morals by agreement. Oxford University Press, Oxford Gauthier D (1988) Moral artifice. Canadian Journal of Philosophy 18:385–418 Gauthier D (1991) Why contractarianism? In: Vallentyne P (ed) Contractarianism and rational choice. Cambridge University Press, Cambridge Hobbes T (1988) The leviathan. Prometheus Books, Buffalo Hull D (1998) On human nature. In: Hull D, Ruse M (eds) The philosophy of biology. Oxford University Press, Oxford Murray M (2007) The moral wager: evolution and contract. Springer, Berlin Heidelberg, New York Richards RJ (1988) The moral foundations of the idea of evolutionary progress: Darwin, Spencer, and the neo-Darwinians. In: Nitecki MH (ed) Evolutionary progress. University of Chicago Press, Chicago Skyrms B (1996) Evolution of the social contract. Cambridge University Press, Cambridge Smith JM (1998) Science and myth. In: Hull D, Ruse M (eds) The philosophy of biology. Oxford University Press, Oxford Sober E (1993) Philosophy of biology. Westview Press, Boulder Spencer H (1857) Progress: its law and cause. Westminster Review 9:445–485 Spencer H (1862) First principles. Williams and Norgate, London

Index

Accountability, 3, 6 and choice, 76, 81 Hume and, 208, 212–219 and responsibility, 89n1, 89n2 Actions, 2–4, 25, 55, 148, 149 causal theory of, 2, 143 and character, 210 deliberative, 157, 158–160, 171n7 explanations of, 60–61, 175, 178, 182, 185, 188n1 free, 4, 11, 207, 209–220 and intelligibility, 126, 129–138 and intention, 57, 60, 61, 142, 143, 144, 145, 153, 157, 158–160, 165–168, 180–181 mental, 159, 160 moral, 12 and motivation, 175, 183–187 value-driven, 158, 170 ‘Actions, Reasons, and Causes’ (Davidson), 143 Action theory, 1–3, 141–146, 151, 152, 158, See also Philosophy of action Agency, 60, 124, 143, 216 Agnosticism, 27, 30, 31–32, 33, 34 Akrasia, 23–24, 71, 79, 107n11 See also Will, weakness of Alexander, J., 229, 234 Anomalous monism, 153n4 Anscombe, G.E.M. (Elizabeth), 1, 9, 56, 123, 124, 133–134, 135, 141–154 Anti-libertarian Claim, 209, 213 Aquinas, Thomas, 1, 9, 141, 146, 148–151 Argument from Queerness, 17 Aristotle, 8, 124, 137, 138, 169 and action theory, 1, 146 and choice, 64, 78, 79, 85 ethics, 9, 141, 148, 150 and teleology, 129–132, 135–137

Arpaly, Nomy, 126–128, 129 Attitudes, 41–42, 43, 46, 48, 79–80, 81, 82–83 pro-, 5, 37, 49n3, 100, 143 propositional, 30 reactive, 82, 208, 216 See also Fitting-Attitude Analysis (FA) Augustine, 8, 125–126, 133–134, 136 Aumann, R., 232 Autonomy, 28, 120, 133 Axelrod, R., 229 Ayer, A.J., 215 Bad, 123–140 Baker, Lynne Rudder, 26 Beardsley, M.C., 39 Behavior, 9, 78, 84, 87, 144 and action theory, 1, 2 cognitively-based, 56–63 and desire, 115, 186 and evolutionary fit, 231–235 Hume and, 213, 219, 220 and Meta-Ethical Pyrrhonism, 31 and moral criticism, 79, 82, 83 and motivation, 6, 56–64, 175–183 Beliefs, 88, 89, 101, 182 de dicto, 126–128, 129, 136 de re, 120, 126–128 and intention, 143, 153 moral, 120, 126–128, 129, 136 and motivation, 173–175, 179, 181 and reasons, 175, 177 Bernoulli, Daniel, 203n10 Binmore, K., 234 Blame, 76, 77, 79, 81, 82, 98, 209–210, 215, 217, 220 Blinov, Arcady, 35n16 Boyle, Joseph, 91n32 Brady, Michael, 47

239

240 Bramhall, John, 211–212, 214, 219, 220 Bratman, Michael, 9, 112, 141, 144–148, 150, 152 Broome, John, 195 Brouwer, L.E.J., 29 Buck-passing, 37–51 Burge, Tyler, 26–27 Burgess, J.A., 25 Buridan cases, 103, 107n13 Calculation, 161–162, 163–164, 166–167 Cardinal scales, 198, 199 Categorical imperative, 146, 162, 166 Causal relation, 209, 211, 212 Causation, 175, 176, 179, 209–210 mental, 26, 27 Causes, 209, 213 desires as, 182–186 and free actions, 210, 211, 212, 214 mental, 153 Chance, 213–214 Choice, 3, 4, 5–7, 11, 55–73, 144, 215 and incomparability, 193–202 insignificance of, 75–92 rational, 225–237 Christman, John, 116–117 ‘Civilization and its Discontents’ (Freud), 132 Cognition, 57–61, 63, 69–70 Coherence, 99, 112 Commensurability, see Incommensurability Comparisons, 63, 65, 68, 70, 89, 193, 195, 197–202 Compatibilism, 207–223 Compulsion, 209 Conscience, 8, 19, 134, 135 Consciousness, 12 Consequentialists, 171n13 Constructive Empiricism, 34 Contingency, 11, 176, 207–223 Contractarianism, 4, 225, 236n6 Cooperation, 226, 227, 230 Correlation, 232 Costs, 193, 197–198, 199, 226, 227 Counterfactuals, 101–102 Dancy, J., 169, 185, 189n38 Danielson, Peter, 227, 228, 229, 234 Darwinism, see Social Darwinism Davidson, Donald, 30, 151, 152, 215 and actions, 143, 144, 189n23 and motivation, 171, 173–179 and will, 56–57, 68, 70–72 DDE, see Doctrine of Double Effect (DDE)

Index Decision-making, 11, 24, 34, 120, 193, 196, 197, 200–202, 202n10 Decision procedures, 162, 163–167 DEFA, see De Facto Attitude Analysis of Personal Value (DEFA) De Facto Attitude Analysis of Personal Value (DEFA), 39–41, 45 Defection, 226, 227, 231, 232 Deliberation, 7, 11, 27, 150, 183 and choice, 55, 58, 71, 85 and desires, 62–70, 106 and indeterminacy, 95, 96–98 and motivation, 10, 100 phenomenology of, 102–103 Dennett, D., 221n7, 237n16 Deontologists, 146, 147, 153 Desirability, 69–72, 124, 134, 136, 150 Desires, 3–4, 7–8, 38, 45, 46, 57, 58, 90n18 alien, 109–120 as causes, 182–186 and choice, 62–64, 68, 85, 86, 91n33 conflicting, 62–64 constitutive nature of, 179–182 extrinsic, 153n11 first-order, 110–120 for the end, 173–191 indeterminacy of, 95–107 induction of, 116–117 intelligibility and, 123–140, 126 and intention, 141, 142, 144, 145, 148, 150, 151, 152, 158 irresistible, 114–115 and motivation, 8–9, 59, 78, 173–191 objectively external, 114–117 and reasons, 123, 175 second-order, 110 theories of, 186 uncontrolled, 115–116 and values, 133 Desire-that, 178, 179 Desire-to, 178, 179 ‘Desiring the Bad’ (Stocker), 124 Determinism, 207–223 Determinists, 11–12, 207, 208, 214, 216, 220, 221n6 Direction of fit, 100, 101 Disposition, 46, 101, 186 Doctrine of Double Effect (DDE), 145–148, 149 Doing, 56–57 Double effect, see Doctrine of Double Effect (DDE) Dreier, James, 90n21

Index Dummett, Michael, 29 Dutch-booked, 201 Ekstrom, Laura Waddell, 111–112 Electio, 149 Emotions, 24, 42, 67, 72, 77, 80, 90n18 Engaging Reason: On the Theory of Value and Action (Raz), 123 Enquiry Concerning Human Understanding (Hume), 210, 213, 214, 215, 220, 221n10 Epiphenomenalism, 26, 27 Epistemology, 18, 20, 24, 26, 27, 89, 100–101, 102–103 Equality, 11, 67, 195 Error Theory, 24, 26 Ethics, 1–4, 8, 9, 124, 151, 162, 235 evolutionary, 225, 229, 233 and intentions, 141, 142, 145–148, 150, 152 virtue, 151, 153 See also Morality Ethnocentrism, 8, 132, 133, 138 T he Eustace Diamonds (Trollope), 115–116 Evil, 230 Evolutionary model, 225–237 Externalists, 23, 24, 98, 100 Favoring, 43, 44, 46, 48 Final value, 43, 49n4 Finnis, John, 153n10 Fitness, 230, 232, 233, 234, 235 See also evolutionary model Fitting-attitude analysis (FA), 37, 38–39, 40, 41–44 Fitting-attitude analysis of personal value (FAP), 45–46, 47 Foot, Philippa, 123, 129–130 Framing, 197, 198, 199, 200, 201, 202 Frankfurt, Harry, 7, 8, 90n4, 106, 109–120, 221n10 Freedom, 3, 6, 11–12, 67, 207–215 ‘Freedom of the Will and the Concept of a Person’ (Frankfurt), 109 Free will, 4, 70, 113, 114, 119, 207–223 Freud, Sigmund, 9, 132 Functionalism, 144–145 Game theory, 225, 229 Gauthier, David, 12, 226, 227, 228 Good, 64, 66, 69–71, 89, 123–138, 146, 150, 153, 167 conditionally, 135, 136

241 and motivation, 61–62 presumptively, 133–136 Good-for, 47–48 Goodness, 125–126 and intelligibility, 132–138 moral, 129 natural, 129–132 Haji, Ishtiyaque, 114, 219–220 Hampshire, S., 56 Hard determinists, see Determinists Hard Problem, 5, 17, 18, 20, 23, 24, 26, 27, 29, 35n3 Hare, R.M., 48 Hate, 135 Heritability, 225, 231, 232, 234 Heuristics, 162, 196, 202, 229–230, 233 Hobart, R.E., 215 Hobbes, Thomas, 124, 132–134, 211–212, 214, 220, 221n1, 226 Holmes, Oliver Wendell, 217–218 H uckleberry Finn (Twain), 126–128 Human agency, 2–3, 7–9, 57, 71, 151 classical versus rationalist conceptions of, 123–124 See also Agency Human development, 130–131 Hume, David, 1, 4, 10, 11, 88, 100, 124, 141, 157, 158, 173 and free will, 207–223 and motivation, 157, 158, 173–191 and necessity, 207–223 Hurley, Susan, 20, 23 Hypnosis, 116–120 Identification, 109–122 attitudinal analysis of, 110–111, 114, 116, 118, 120 structural analysis of, 11–14, 119, 120 Identity: personal, 112, 121n6 practical, 80, 82, 105–106 Ignorance, 196 I liad, 135 Imagination, 7, 57, 95, 97, 98 Immorality, 7, 22, 45, 62, 69, 70, 85, 225, 228, 231, 232, 233 Incommensurability, 11, 55, 65, 66, 67, 69, 70, 72, 202n3 ‘Incommensurability and Agency’ (Raz), 123 Incomparability, 193–204 Indeterminacy, 7, 95–107, 195, 199, 200 implications of, 105–106 kinds of, 103–104

242 Indifference, 213 Inferiority, 11, 195 Instrumentalism, 86, 87, 88 Intelligibility, 8, 29, 123–140 ‘Intending’ (Davidson), 143 Intensionality, 178, 180 I ntentio, 9, 141, 149–151 I ntention (Anscombe), 124, 133, 141–154 Intentional action, 57, 60, 61, 142, 143, 144, 145, 153, 157, 158–160, 165–168, 180–181 Intentionality, 180, 233 Intentions, 3, 4, 8–9, 141–154 and ethics, 141, 142, 145–148, 150, 152 future-directed, 144–145 See also Intention skeptics Intention skeptics, 141, 143–144 Internal Cause Claim, 209, 213, 214 Internalism, 57, 95, 96, 98 Interpretive puzzle, 208–212 Intransitivity, 200–201 Intuition, 56, 219–220 Intuitionism, 29–30 Irrationality, 85, 86, 88, 91n30, 99–100, 174, 184, 194, 226, 227, 229 Is-ought problem, 225, 229 Jackson, Frank, 23, 29–30, 31 Judgments, 43, 44, 48, 186, 195, 221n4 moral, 4, 5, 12, 18–27, 31, 32, 33, 34, 169 practical, 10, 71, 160, 162, 166, 168, 169 value, 55–73 Justification, 99 Kagan, S., 39 Kant, Immanuel, 28, 99, 105, 146 and action theory, 1, 158 and ethics, 124, 141–142, 147, 162 Kim, Jaegwon, 27 Knowledge, 39 and intention, 143, 144, 150 and motivation, 59 propositional, 55, 57, 60–61, 183 Korsgaard, Christine, 90n7, 96, 105–106, 139n7, 173, 176, 179, 181 Lamarckianism, 237n14 Lehrer, Keith, 215 Leon, Mark, 114 Leviathan (Hobbes), 124, 133 Lewis, David, 101–102 Liability, 217–218 Libertarians, 11, 207 Liberty, 207–215

Index Logic, 29 Love, 135, 136 MacIntyre, Alasdair, 141 Mackie, J.L., 4–5, 17–19, 22–30 Manipulation: and identification, 115, 116, 118, 119 Means: motivation to, 173–191 Means-end syllogism, 161, 163, 164, 165, 169 Mental, 151 Mental activity, 173, 183, 184, 185, 187 Mental events, 143 Mental states, 144 Meta-Ethical Pyrrhonism (MEP), 17–36 Meta-ethics, 17–51 Metaphysics, 4, 18, 26–27, 34, 87, 88 Mistakes, 8, 18, 66, 89, 101, 130, 138 ‘Modern Moral Philosophy’ (Anscombe), 134, 135, 147–148, 151 Modularity of mind, 22–23 Moore, G.E., 215 Moral evaluations, 26–30 Morality, 6, 12, 75–89, 207–220, 225–236 folk, 5, 17–34 and internalism, 96 and practical reason, 105 See also Ethics Moral norms, 5, 17, 27, 28, 31, 32, 33 T he Moral Problem (Smith), 98–100 Moral realism, 5, 17–36 Moral sentiments, 12, 20, 207, 208, 215, 217, 218, 220, 221n2, 221n4, 221n10, 222n11, 234 Moral skepticism, 29, 30, 31 Moral theories, 4, 19, 20, 26–30, 34, 146, 147 Motivation, 55–56, 69–70, 71, 78, 85, 87, 99, 100, 101, 105, 106 and choice, 67, 68 conflicting, 62–64, 68, 69, 182 and desires, 8–9, 59, 78, 173–191 and human agency, 151 and Hume, 157, 158, 173–191 and indeterminacy, 95–96 instrumental, 173, 176, 187 and intention, 149 to the means, 173–191 moral, 23–24, 25, 225, 229 rational, 186 and reasons, 175, 176, 182, 185, 187 theory of, 6, 57–61, 62–64 towards the end, 179–180, 183–185, 187 Motivational blocks, 62–64, 68–69 Motivational economy, 111, 115, 120

Index Motivational set, 95, 116, 119 Motivation-by-Desire Principle, 173–187 Nagel, Thomas, 5, 17, 18, 20, 25–34, 173–176, 179 Narcissism, 8, 132, 133, 138 Nash Equilibrium, 32, 232 Natural Goodness (Foot), 129–130 Naturalism, 17–36, 129 and objective prescriptivity, 18–25 Natural Reasons (Hurley), 23 Nature, 132–133, 229, 232, 233 Necessity, 11, 207–223 Nietzsche, Friedrich, 218 Normativity, 83–84, 85, 87, 88, 89, 90n17 and motivation, 175, 176 Nozick, R., 171n16 Nurture, 132–133 Objective impermissibility, 20 Objective prescriptivity, 17, 32, 35 and naturalism, 18–25 Olsson, Jonas, 47 Pain, 133 Pareto situations, 28, 31 Parfit, D., 45 Pascal, Blaise, 33, 35n40 Passion, 88, 118–119, 177 Payoff matrix, 33, 236n13 PD, see Prisoner’s Dilemma (PD) Phenomenology, 100, 101, 186, 199 of deliberation, 102–103 Philosophical Investigations (Wittgenstein), 151 Philosophy of action, 1, 2–3, 141–143, 146, 149, 151–153 See also Action theory Planning Theory, 144 Plato, 28, 124 Pleasure, 133, 134, 136 Praeter intentionem, 146, 148 Praise, 76, 77, 79, 81, 82, 209–210 Predictions, 71, 145 Preferences, 11, 28, 65, 111–112, 196–197, 200–201, 225, 231 Principles, 29, 56, 97, 98, 100, 105, 127, 128 moral, 65, 164–166, 169 normative, 187 rational, 177 Prisoner’s Dilemma (PD), 28, 31, 32, 33, 226, 232 Proairesis, 149, 150 Protagoras (Plato), 89

243 Psychology: cognitive, 145, 193 empirical, 1, 2, 221n4 philosophy of, 144, 148, 150–151, 153 Pyrrhonism, see Meta-Ethical Pyrrhonism (MEP) Quine, W.V.O., 29, 30 Rabinowicz, Wlodek, 39 Radical interpreter, 24 Railton, P., 90n12, 171n12 Rationalism, 12, 186, 187 Rationalists, 124, 157, 158, 186–187, 225, 231 and motivation, 173–187 Rationality, 6–7, 30, 84–89, 98, 99, 176, 194, 201, 202n10 bounded, 197 instrumental, 147, 149, 150 and mental states, 144 and morality, 225–236 practical, 33, 75, 84–85, 86, 89, 96, 144, 147 satisficing, 197 Rationalization, 103, 119, 132, 133, 138 Raz, Joseph, 8, 123–140, 195, 197, 198 Reality, 83, 91n23 Reason, 4, 10, 37, 88, 215 practical, 7, 9–11, 95–107, 173, 177, 193, 194, 201–202 Reasoning, 89 and action, 157 instrumental, 158–170, 185 moral, 10, 148, 160, 162–169 non-instrumental, 158–170 practical, 4, 129, 144, 145, 148–150, 153, 157n4, 158–162, 166–168, 170 substantive, 160, 161, 162, 168, 170 Reasons, 124, 127, 153, 175, 177 and actions, 142, 143 explanatory, 175, 176 and intention, 151 and motivation, 175, 176, 182, 185, 187 normative, 173, 175, 176 pro-tanto, 48 Reductionism, 144 Regress-cum-Incompleteness objection, 109, 110, 113, 114, 119, 120 Religion, 134–135, 230 Representation Problem, 19 Respect, 28, 42, 45, 98 Responsibility, 76–77, 82, 207–223 Revenge, 135 Reward theories, 186 Rhetoric (Aristotle), 124, 138

244 Ring of Gyges objection, 28, 33 Ross, W.D., 38, 169 Rule-case syllogism, 163–167, 169 Rules view: and moral reasoning, 160, 162–169 Russell, Paul, 208–209, 210, 213, 217 Satisficing, 197 Scanlon, T.M., 37, 38, 188n2 Schlick, M., 90n3, 215 Schueler, G.F., 189n36 Sea Wolf (London), 138 Second nature, 8, 123–125, 130–132, 135 Seeings, 128 Selection, 6, 55, 63, 64, 68, 69, 71 Self-consciousness, 105 Sensations, 100–101 Sensitivity, 230 Side-effects, 66, 145, 146, 147, 148, 153 Simon, Herbert, 193, 197–198 Skepticism, 26, 30 Skyrms, Brian, 229, 231, 232, 234 Small improvements, 193–204 argument from, 194–195, 198 Smart, J., 90n3, 171n12 Smith, Maynard, 225, 233 Smith, Michael, 7, 96, 98–100, 139n3, 157, 188n2 and motivation, 171, 175 Social Darwinism, 138, 225, 234 Socialization, 23, 24, 132 Socrates, 89 Spencer, Herbert, 233 Stocker, Michael, 157, 158 Strawson, Galen, 221n7 Strawson, Peter, 79, 90n8, 216 Sub specie boni, 123–126, 129–131, 134, 136 Sub specie mali, 124, 129 Suffering, 20, 133, 134, 230 Superiority, 11, 195 Supervenience, 18, 19 Teleological thesis, 9–10, 157–158, 160–161, 165, 166, 168–170 Teleology, 129, 157–172 Aristotelian, 129–132, 135–137 Thresholds, 193–204 Transitivity, 200–201

Index T reatise of Human Nature (Hume), 177, 208, 210, 213, 214, 222n10 Truth, 57, 61, 101–102 Truth-telling, 163–165, 166, 168–169 Ultimatum Game, 231, 232 Universalizability, 96, 99 U nprincipled Virtue (Arpaly), 126 Utilitarianism, 162, 166 Vagueness, 66, 193–204 Value-for, 44, 48, 49 Values, 3, 4–5, 7, 10, 12, 19, 24, 33, 46, 136, 137 ascription of, 234 and desires, 133 extrinsic, 40, 41, 46 impersonal, 39, 40, 48 and incomparability, 196–201 intrinsic, 41 moral, 22, 23, 24, 27, 29, 235 nature of, 4–5 personal, 37–51 subjective, 11, 24–25, 197, 198, 200, 203n10 and will, 23 Van Fraassen, Bas, 21, 34 Van Gulick, Robert, 26 Velleman, David, 121n12 Virtue, 126, 127, 148, 150 Virtue ethics, 151, 153 Volition, 55–73, 183 and weakness of will, 70–72 Wallace, R.J., 129, 216–220, 221n7 Wanting, 124, 150 Welfare, 44–45, 47 Wiggins, David, 99 Will, 5, 24, 149, 150, 213 and values, 23 weakness of, 56, 57, 70–72 See also Akrasia, Free will Williams, Bernard, 7, 95, 96–98, 222n12 Willing, 56–57, 60, 61, 62, 69, 209 and intention, 143 Wishing, 45, 62, 70 Wittgenstein, Ludwig, 151 W rong Kind of Reasons problem, 49n3