Proto Sociology Vol. 16, 2002 - Understanding the Social: New Perspectives from Epistemology

ProtoSociology

ISSN 1434-4319

An International Journal of Interdisciplinary Research Volume16 – 2002 Understanding the Social: New Perspectives from Epistemology Editor: Gerhard Preyer Editorial of the Vol. 16: Georg Peter Layout and Technical Conception: Georg Peter Editorial Office: ProtoSociology, Stephan-Heise-Str. 56, 60488 Frankfurt am Main, Germany, Phone: (049)069-769461, E-Mail: [email protected] P Homepage: http://www.protosociology.de P P P P

Published Volumes

373

Titel

3

Proto Sociology An International Journal of Interdisciplinary Research Vol. 16, 2002 Understanding the Social: New Perspectives from Epistemology

Contents From Mentalizing Folk to Social Epistemology The Mentalizing Folk Alvin I. Goldman

7

Belief and Acceptance as Features of Groups Margaret Gilbert

35

Collective Agents and Cognitive Attitudes Anthonie Meijers

70

Challenging Epistemic Individualism Deborah Perron Tollefsen

86

Externalism, Events and Institutional Facts From an Externalistic Point of View: Understanding the Social Gerhard Preyer

121

Are “Contexts and Events” Equivalent? Possibilities for an Ontological Symbiosis Steven Miller

152

4

Autor

Institutional Facts and the Naturalistic Fallacy. Confronting Searle (1964) with Searle (1995) Frank A. Hindriks

170

Group Responsibility and the Consensus Account of Justified Belief Group Action and Group Responsibility Pekka Mäkelä and Raimo Tuomela

195

Collective Commitment: A Theoretical Understanding of Human Cooperation Lambèr Royakkers and Vincent Buskens

215

Justification and Consensus: The Peircean Approach Frederick F. Schmitt

241

On Contemporary Philosophy and Social Sciences How I Came to Know Quine: A Reminiscence Roger F. Gibson

287

Responsibility and Globalization Hans Lenk, Matthias Maring

298

Contributors

334

Imprint

335

On ProtoSociology

336

Published Volumes

337

Bookpublications

348

Subscription / Digital Version

362

ProtoSociology: Digital Publications available

363

The Mentalizing Folk

From Mentalizing Folk to Social Epistemology

5

6

Alvin Goldman

The Mentalizing Folk

7

The Mentalizing Folk Alvin I. Goldman

Abstract Three major questions should be answered by any theory of “folk psychology”, or mentalizing. The first question concerns the contents of mental concepts; the second concerns the processes of mental-state attribution (both first- and third-person attribution); and the third concerns the development or acquisition of mentalizing skills. Some major problems are presented for different variants of the “theory-theory” approach, namely, philosophical functionalism, the child-scientist approach, and the modularist approach. The approach favored here is an “introspection-simulation” approach: introspection as an account of first-person mental-state attribution, and simulation as an account of third-person mental-state attribution. The simulationist part of the story is clarified and defended in some detail, including a discussion of the relation between knowledge and simulation. Special attention is given to empirical findings that support the notion that pretend states could really be “facsimiles” of their genuine counterparts, a relationship that seems to be required if simulation is to produce successful mental-state attributions.

1. Multiple Questions Ordinary people are mentalizers. The folk not only have mental states, but they represent themselves – both self and others – as having mental states. Many of these mental states, e.g., beliefs and desires, represent how the world is or ought to be. By representing themselves as having such first-order representational states, the folk engage in second-order representation, or metarepresentation. The class of representations is not exhausted, of course, by mental states; linguistic utterances and inscriptions are also representations though not mental ones. The representation of mental representation is a big enough topic for a single paper, however, and it is all that I shall bite off here. Restricting ourselves to mental metarepresentation, there remains a wide array of questions to which philosophers and cognitive scientists devote their attention. I divide the central questions into three main groups: (1) Questions about the contents of mental concepts. What concepts do the folk have of mental representations? How do they conceptualize or represent to themselves such states as belief, desire, intention, and the other mental attitudes? What properties do they endow these states with, in their fundamental

8

Alvin Goldman

grasp of them? These questions should be contrasted with questions about the essential nature of mental states. As far as scientific ontology is concerned, mental states might turn out to be neural states. But scientifically and philosophically untutored people presumably do not conceptualize beliefs and desires as neural states. How do they conceptualize them? (2) Questions about processes of mental-state attribution. The folk not only possess mental-state concepts, but they deploy these concepts with some frequency, ascribing specific instances of these concepts both to self and others. A person may ascribe to herself a desire that p, or ascribe to another a belief that q. How do they go about selecting these specific ascriptions? What evidence and inferential procedures do they use, or what techniques, routines, or heuristics are deployed? Since mental attitudes are ascribed both to self and others, and since these ascriptions may involve different evidence and different routines, there are two questions to be asked: What procedures are used in making first-person attributions, and what procedures are used in making third-person attributions? (3) Questions about the development or acquisition of mentalizing skills. What are the sources and antecedents of people’s mentalizing facility? What native endowments underpin this cognitive skill, and what modes of maturation or learning generate full competence at mentalizing or metarepresentation? Issues about acquisition are the core issues for many developmental psychologists. As a philosopher, however, I am principally interested in the first two sets of questions. Acquisition is of interest to me only insofar as it sheds light on answers to the first two sets of questions. Obviously I shall not try to answer any of these questions in full detail, but some of the principal answers currently on offer will be reviewed and evaluated, and I shall place my own position against the background of these competing answers. I shall begin by examining the collection of approaches grouped under the heading “theory-theory” (TT), approaches that jointly comprise the dominant perspective on mentalizing. Each of these approaches will be faulted, either for empirical implausibility or for failure to answer some of the core questions about mentalizing. I shall then turn to the approach I find most compelling, a version of the “simulation” theory flavored with an infusion of introspectionism.

The Mentalizing Folk

9

2. Theory-Theories Many researchers claim that the folk (i.e., adult folk) have a rudimentary theory of mind, a theory that gets deployed in mental-state attribution, explanations of actions in terms of mental states, and so forth. A related claim is that mental-attitude concepts are understood wholly in terms of this folk theory. The contents of mental concepts can only be spelled out properly in terms of the theory in which they are embedded. 2.1. Philosophical Functionalism The earliest version of TT is due to philosophers of mind (Sellars 1963, Lewis 1972), who were largely interested in the “analysis” of mental-state concepts. Their leading idea was borrowed from positivist philosophers of science, the idea that theoretical, non-observational concepts in science are definable by means of lawful relationships to observables. They claimed that mentalistic words like ‘desire’ and ‘belief’ are definable in terms of laws relating the intended inner states to stimulus inputs, other inner states, and behavioral outputs. Since these laws are supposedly possessed and deployed by the folk themselves, not just by scientists, this approach is sometimes called “commonsense functionalism” (Lewis 1972, Block 1980). Commonsense functionalism is touted for its ability to account for the layperson’s facility at predicting and explaining the behavior of others. Laws permit users to infer behavior predictively or retrodictively from a third party’s mental states. They also permit users to infer mental states from prior mental states or from stimulus conditions. So commonsense functionalism purports to show how people attribute mental attitudes, viz., by means of inferences employing lawlike premises. Three fundamental difficulties face this functionalist version of TT. First, is there really a set of commonsense psychological laws that laypersons deploy in prediction and explanation? Although philosophers of mind provide a few examples of such laws, difficulties and obscurities surround the attempt to state them carefully (Schiffer 1987, Goldman 1989). Furthermore, it is extremely doubtful that predictions are always based on such laws. I have used Kahneman and Tversky’s (1982) airport example to illustrate this point. Given descriptions of two imaginary travelers who both miss their flights, subjects were asked to predict which traveler would be more disappointed. Ninety-six percent of the subjects said that the traveler who barely misses his flight will be

10

Alvin Goldman

more disappointed than the one who misses it by 30 minutes. Does this consensus result from all subjects’ knowing a law about degrees of disappointment? That is highly dubious (Goldman 1989). A second nest of problems for commonsense functionalism concerns firstperson attributions of mental states (Goldman 1993a). Functionalism is poorly equipped to accommodate people’s special access to their own present attitudes. Since functionalism claims that desires, beliefs, and intentions are understood purely in terms of their causal relations to other states, and ultimately in terms of their causal relations to stimuli and behavior, the only method of identifying one’s current desires and beliefs according to functionalism is by inference from stimuli and behavior. This seems clearly wrong. Surely I can tell that I currently intend to cook shrimp for dinner without knowing any stimulus that has caused this intention nor any behavior that has resulted from it. (I may not have purchased any shrimp yet, nor performed any other item of shrimp-oriented behavior.) Of course, functionalism does not say that a given mental state always produces designated outcomes; it is only a disposition or propensity to produce them if other inner and/or outer conditions are ripe. But it is radically unclear how a person is supposed to be capable of detecting a token state’s dispositional or subjunctive properties. It is therefore unclear how a person can actually go about using functionalist laws to classify their own mental states, especially to classify them in the seemingly “direct” fashion available to us. Functionalism’s postulation of extensive conceptual connectivity among mental states also creates severe computational problems, specifically, the threat of combinatorial explosion (for details, see Goldman 1993a). Thus, the functionalist story has many counts against it. 2.2. The Child Scientist Theory Psychologists as well as philosophers use the label “theory of mind” for their approach to folk mentalization. But psychologists often mean something a bit different by “theory of mind,” and the arguments they adduce for their position are certainly quite different. Several psychologists, however, present a version of TT with some similarities to functionalism. Wellman, Gopnik, and Perner (Wellman 1990, Perner 1991, Gopnik 1993, Gopnik and Wellman 1992) all view folk mental concepts as theoretical constructs, introduced to provide causal explanations and facilitate predictions. They claim these concepts must be theoretical concepts, because they undergo theoretical change just like theoretical concepts in science. Their argument for the theoretical character of

The Mentalizing Folk

11

mental concepts rests on claims about the way these concepts change in early childhood development. The hallmark of this group of psychologists is the view that the growing child is, in essence, a theoretical scientist. Gopnik and collaborators (Gopnik and Wellman 1992, 1994; Gopnik 1996; Gopnik and Meltzoff 1997) view children’s alleged transitions from one theory of mind to another as comparable to transitions in scientific theory, e.g., astronomy’s theoretic transitions from Copernicus to Brahe to Kepler. Children’s early concepts of perception, desire, and belief are said to be non-representational. This primitive form of mentalistic theory excludes the idea of misrepresentation, e.g., false belief. That is why, supposedly, three-year-olds and younger children standardly fail the false-belief task: they don’t yet have a concept of false belief. The richer, genuinely representational concept of belief only comes with the advent of a subsequent, four-year-old theory. This story of alleged transitions, however, including the sub-plot of conceptual deficit prior to four years of age, is empirically quite problematic. In recent studies concerning false-belief tasks, the earlier hypothesis of conceptual deficit prior to age four has been disintegrating. Clements and Perner (1994) found that children from 2 years 11 months demonstrated implicit understanding of false belief even when their explicit answers to false-belief questions failed to evince such understanding. Two papers by Freeman and colleagues strongly suggest that 3-years-olds’ troubles with false-belief tasks rest on performance problems rather than a conceptual deficit. Lewis et al. (1994) found that preschoolers (including 3-year-olds) did better on false-belief tasks when they were allowed to go through a story twice rather than just once, presumably giving them needed help in reconstructing the protagonist’s mental state from memory. The memory explanation for poor performance on false-belief tasks was further confirmed by Freeman and Lacohee’s (1995) finding that children are better able to attribute prior false belief to themselves when an aid to memory is provided. All of these studies shed doubt on the notion that young children have an early conception of belief (or “prelief,” as Perner calls it) that lacks the notion of misrepresentation. What about other mental attitudes in young children? There too the story of early conceptual deficits seems to be falling apart with the help of more sophisticated experimental techniques. Consider the work of Meltzoff (1995) on young children’s grasp of intention. Even 18-month-olds were found to have a grasp of misrepresentation in connection with intentions. In Meltzoff’s experiments, infants observed adults “try” but fail to achieve

12

Alvin Goldman

unfamiliar effects with certain manipulanda (at least this is how the behavior would be described by an adult). The experimenter would pick up a stick-tool and “aim” to push a button with it, but the button was always missed. Would an infant interpret this behavior in goal-oriented terms, despite the fact that the goal was never fulfilled? Dramatically, the answer was “Yes”. Eighteenmonth-olds enacted the intended but failed attempts of the experimenter rather than the actual behavior. Thus, even children of 18 months have a preverbalized concept of intention that permits misrepresentation. Compelling evidence of another variety against the child scientist approach comes from studies of Williams Syndrome patients. Patients with Williams Syndrome are retarded, with an average IQ of about 50, and even adolescents and adults appear to be unable to undergo any of the forms of conceptual change associated with theory building (Johnson and Carey 1996). Nonetheless, children with Williams Syndrome begin to explain actions in terms of beliefs and desires at about normal ages (Tager-Flusberg 1994). This is powerful evidence against the notion that mentalizing is a product of a general theory-building capacity (Segal 1996; Carey and Spelke 1996). In addition to this body of empirical counter-evidence, three theoretical considerations also speak against the child scientist approach. First, as Leslie and Roth (1993) point out, the child scientist approach sees the child not just as a theorist, but as a quite brilliant theorist. After all, the concepts of propositional attitudes are highly sophisticated and logically complex. But children don’t appear to be brilliant theorists in all domains, which suggests that the development of mentalizing has different roots. Second, as Segal (1996) notes, the developmental pattern and end point of mentalizing are remarkably similar across individuals and apparently across cultures. This convergence would be surprising if each child develops their own mentalizing theory like a scientist, because in science different theorists do not normally converge so rapidly on the same theory. (For example, psychologists don’t agree about how the mentalizing facility is acquired.) Third, as Carruthers (1996) points out, scientists modify a theory when they come to believe that it is false, or at least believe that it makes false predictions. But the very young child scientist cannot decide that her hitherto accepted theory of mind is false when, by hypothesis, her only concept of belief up until then precludes the idea of falsity! This is a basic incoherence in the child-scientist version of TT. Finally, for those of us primarily interested in the first two (sets of) questions presented in section 1, the child-scientist version of TT has little to offer. This group of proponents of TT mainly argue that mental concepts must be theore-

The Mentalizing Folk

13

tical constructs of some sort, and inferences deployed in applying these concepts must be theoretical inferences of some sort. They give us few specifics, however, about the exact content of the various mental concepts or the exact inferences used in applying them. From the point of view of our first two questions, little is provided by way of answers. 2.3. The Modularity Theory Let us proceed, therefore, to another version of TT, the one advanced by Leslie (Leslie 1987, 1994; Leslie and Thais 1992; Leslie and Roth 1993). Leslie claims that the ability to engage in mentalizing depends upon possession of a domainspecific mechanism, ToMM. This mechanism is an innate endowment of the human species and begins to develop in infancy. ToMM is an information processing device that computes data structures called metarepresentations. A metarepresentation makes explicit four kinds of information: (1) an attitude (e.g., belief, desire, pretense), followed by three arguments which specify, respectively, an agent, an anchor (some aspect of the real situation, e.g., that the agent is holding a banana), and an imaginary or pretend state (e.g., that the agent is holding a telephone). For example, a metarepresentation might be: Mother pretends (of) this banana (that) ‘it is a telephone’. Leslie claims that “having access to such data structures, together with the inferences they support, constitutes a tacit and intuitive theory of mind, or … a tacit theory of the specific ‘representational relations’ (like pretends, believes, wants) that enter into the causation of agents’ behavior” (Leslie and Roth 1993: 91). Elsewhere he says: “We will make no bones about the fact that we are on the side of theory-theory” (Leslie and German 1995: 123). These and similar statements make it natural to place Leslie in the TT camp. But although Leslie himself classifies his approach (in these passages) as a version of TT, that does not make it so. Leslie and his collaborators clearly offer a cognitive science theory of mentalizing, but a cognitive science theory of mentalizing is not necessarily a TT. Simulation theory is equally a cognitive science theory of mentalizing (or aspires to be), but it is not a specimen of TT. Something qualifies as a specimen of TT if it not only is itself a theory but ascribes to the folk the possession and deployment of a theory. Despite occasional pronouncements to the contrary, Leslie does not seem to attribute to the folk the possession and deployment of a genuine theory, in an appropriate sense of ‘theory’. Furthermore, it is unclear whether Leslie’s approach tries to answer the first two questions on our original list, whether in terms of theory possession or

14

Alvin Goldman

otherwise. But insofar as he hints at answers to them, it is not obvious that the answers embody a TT. One part of Leslie’s work, as we have seen, involves a characterization of the “logic” of pretend representations. According to this logic, “primary” representations such as ‘the cup is full’ can be “decoupled” so as to produce what philosophers call an opaque construction. Thus, there is no contradiction--and the child understands that there is no contradiction--in a representation of the form, ‘I pretend (of) the empty cup (that) it is full.’ Obviously, this logic per se does not constitute a theory in the sense of a set of lawlike principles used by the child for explanation and/or prediction. However, Leslie goes on to say that decoupled expressions can be combined with real world inferences to elaborate on pretend situations (Leslie 1994). For example, a real-world causal inference is: If a container filled with liquid is upturned, then the liquid will pour out and make something wet. This causal inference enables a pretend scenario to unfold. If it is pretended of an empty cup that it is full, and if this cup is upturned, then it is pretended of the cup that it pours liquid out and makes something wet. Leslie has presented this kind of scenario to children, and they indeed respond by saying that a cup upturned over a play character’s head makes that character wet. Leslie further describes the children as inferring what the experimenter was pretending. So here Leslie addresses the issue of inferring the mental states of others. But has any such inference involved a psychological theory possessed by the child? No evidence is presented to this effect. A child in Leslie’s experiment may simply build on the experimenter’s props and actions to construct his/her own pretend scenario, and then infer that the experimenter pretends likewise. If this is what the child does, it involves no psychological theory featuring psychological laws. (It may involve physical laws, however, such as the upturning-causes-spilling ‘law’ mentioned above.) In fact, it sounds very much like the story a simulation theorist would tell about this situation. In fact, Leslie himself concedes that the simulation account is a perfectly acceptable fit with what the child does! [I]f the child can infer that a cup containing water will, if upturned over a table, disgorge its contents and make the table wet, then the same child can also elaborate his own pretence or follow another person’s pretence using the same inference… Now, if someone wants to call the above “simulation”, then they can, but it adds little or nothing to the account to do so. (Leslie and German 1995: 128)

The Mentalizing Folk

15

The last sentence acknowledges that the simulation approach is quite compatible with this kind of case, a paradigmatic one in Leslie’s studies. The previous paragraph indicates that, while Leslie (sometimes) does address the second question posed in section 1, he has no evidence that specifically favors a TT answer to the question over a simulation theory (ST) answer. As to the first question raised in section 1, it does not seem as if Leslie answers it, or has the resources within his theory to answer it. Leslie does not explain how each separate attitude concept is represented, what the conceptual differences are, for example, between belief, desire, fear, and intention. Nor does his theory explain how a person classifies or identifies his/her own mental states (part of the second set of questions). Consider the judgment, “I currently desire a cookie.” How does a person decide that he currently desires a cookie rather than believes that he will get a cookie? Leslie does not tell us. 2.4. Generic Theory-Theory Another group of TT defenders defend it only in the most general or generic terms, without venturing to provide any details of folk theorizing. This is the approach of Stich, Nichols, and colleagues (Stich and Nichols 1992, 1995; Nichols et al. 1996). These writers are deliberately non-committal on the specific content of the folk theory: whether it involves laws (as in functionalism), theoretical constructs as in science, or the like. They mean by ‘theory’ any kind of knowledge structure whatever – either “a body of rules or principles or propositions” (Stich and Nichols 1992, 35). They define the field as a debate between TT and ST, and their chief aim is the negative one of disproving ST rather than providing a specific form of TT. Their way of defining the issue, however, is actually a bit contentious. They often focus on only a “radical” version of ST that involves the wholesale exclusion of “knowledge” in favor of “ability” (Leslie and German 1995). They insist that any use of information by subjects in the mentalizing process is incompatible with ST (Nichols et al. 1996). But these portraits of ST unduly restrict its range. While some friends of simulation might accept these renditions of their view, my own version of ST permits subjects to employ a modest or even substantial amount of information (as explained below). This allows a reasonable contrast between ST and TT, but rejects the framing of the debate as one of some information versus no information.

16

Alvin Goldman

3. Introspection and First-Person Attribution I turn now to my own, positive approach to mentalizing, beginning with the concepts of attitudinal states. I propose to take seriously the role of consciousness in our folk concepts of mental states. This proposal takes its point of departure from three assumptions that should be congenial to cognitive science. (1) There is conscious experience. (2) The occurrent, or activated, forms of the representational attitudes are often found in conscious experience. (3) People, even children, have introspective access to these denizens of conscious experience. With the dramatic growth of consciousness studies in contemporary cognitive science, assumption (1) should now be uncontroversial. Assumption (2) is also mild; it simply denies that conscious experience is exhausted by perceptions, feelings, and emotions. Tokens of the propositional attitudes (at least in their occurrent forms) are also found in experience. If anybody needs convincing that cognitive science regards certain representational states as conscious, consider the fact that memory researchers distinguish explicit from implicit memory by saying that the former involves conscious recollections (Schacter 1989). Assumption (3) is also quite modest, simply saying that conscious events are open to a special form of access: introspective access. Elsewhere (Goldman 1997) I show that reliance on introspection is widespread in cognitive science, and not epistemologically objectionable. So assumption (3) is not alien to cognitive science. My hypothesis is that mental-state concepts such as desire, belief, and so forth are understood at least in part in terms of non-dispositional characteristics of conscious experience, characteristics that can be introspected by the subject of the experience. The most natural account of people’s special access to their mental states is an account that posits some sorts of non-dispositional characteristics that are introspectively available to the subject and distinguishable from one another. My hypothesis is that such characteristics underlie the grasp of mental-state concepts. To say that a characteristic of consciousness is introspectively accessible or discriminable is only to say that it can potentially be identified or discriminated, not that it is or will be identified on every (or any) occasion that it occurs. (The HOT theory of consciousness, by contrast, says that a state is conscious only if its subject actually has a higher-order thought about that state; see Rosenthal 1992). In particular, I am not committed to any thesis of the infalli-

The Mentalizing Folk

17

bility or transparency of conscious characteristics. Property detection requires appropriate conceptual facility, and this facility is not guaranteed by introspective accessibility. So although an infant or child may experience beliefs, hopes, and desires, the child may lack the conceptual facility to classify them in intentional or representational terms. Once the child develops this conceptual facility, however, she can locate instantiations of these concepts squarely within her conscious experience. In this sense, occurrent incarnations of belief, desire, intention, and so forth are not “unobservables,” contrary to most theory-theorists (e.g., Leslie 1994: 212). It is important to appreciate that the introspective or “Cartesian” story is perfectly compatible with the early development of metarepresentational concepts, even along lines that theory-theorists have sketched. Leslie might be right, for example, to postulate an innate module which, when fully deployed, makes available the sort of conceptual structures he postulates. He may also be right in claiming that autism involves an impairment in ToMM which results in autistic children never acquiring the postulated metarepresentional structures, or not applying them, at any rate, to first-order mental states. This is all compatible with introspectivism. What introspectivism says is that acquiring the concepts of mental representational states involves the task of latching these complex conceptual structures onto the right sorts of conscious characteristics, i.e., latching the belief concept onto belief episodes, the desire concept onto desire episodes, and so forth. Although autistics certainly experience belief episodes and desire episodes, they apparently have deficits in representing them as representations, i.e., as states that can be true or false, fulfilled or unfulfilled, and the like. This is compatible with an introspectivist approach. Gopnik (1993; Gopnik and Wellman 1995) argues that an introspectivist or privileged access account of first-person attribution fails to explain certain errors of young children in self-ascription of beliefs. In the Smarties task, threeyear-olds are asked what they first thought was in the candy box before it was opened. In fact, they thought it was candy. But one-half to two-thirds mistakenly say they had originally thought there were pencils (as they subsequently learned). Gopnik thinks that introspectivism cannot accommodate this finding because introspectivism is committed to first-person infallibility. But introspectivism is not committed to infallibility; it can cheerfully grant that three-year-olds might lack a mastery of propositional-attitude attribution, especially verbal mastery. Moreover, in Gopnik’s experiments the target beliefepisodes are in the (recent) past, and hence are not open to immediate introspection. They can only be recalled or not recalled. Memory explanations of

18

Alvin Goldman

errors in these sorts of tasks are indicated by papers cited in section 2.2. It is no objection to introspectivist accounts that young children have memory difficulties even with recently experienced thoughts. When the states are genuinely concurrent, however, the special access to one’s own conscious experience readily accounts for the ease and accuracy of first-person attribution. No version of TT explains this ease and accuracy. A different problem raised against the introspection-simulation approach is how it accounts for developmental differences across mental-state concepts. According to this approach, why is there a lag in deployment of belief as compared with desire in action explanations (Gopnik and Wellman 1994, 1995)? There are several possible answers here. One might appeal to the comparative phenomenological salience of desire as compared with belief. If desire is more salient than belief, it may be easier for children to latch onto the concept of desire than the concept of belief. A very different explanation might invoke differences in the language of desire and belief. As Harris (1996) notes, there may be a social premium on the communication of desire that results in greater conversational deployment of the language of desire. A great deal of childadult activity involves joint or collaborative action, the success of which depends on the communication of the desires of each. The focus on desires, with beliefs left unspoken, might explain the child’s greater preoccupation with desire and mastery of the language of desire. Many cognitive scientists are leery of regarding one’s own conscious episodes as central to our grasp of psychological concepts. In addition to methodological worries about introspection, this denigration of the first-person perspective may stem from an evolutionary perspective. Identifying the mental states of others, as opposed to one’s own mental states, has a clear selectional advantage. There is a definite biological benefit to be derived from accurate beliefs about the inner life of one’s rivals, as a predictor of what they will do next. Given this selectional advantage, shouldn’t we expect mental concepts to be rooted in the behavioral features of others, not in our own conscious states? Humphrey (1983) points out a plausible connection between the two: nature might provide us the means of mindreading others via an ability to read our own minds and thereby simulate the minds of others. [A] revolutionary advance in the evolution of mind occurred when, for certain social animals, a new set of heuristic principles was devised to cope with the pressing need to model a special section of reality – the reality comprised by the behavior of other kindred animals. The trick that nature came up with was introspection; it proved

The Mentalizing Folk

19

possible for an individual to develop a model of the behavior of others by reasoning by analogy from his own case, the facts of his own case being revealed to him through examination of the contents of consciousness. (Humphrey 1983: 30) But aren’t there innate tendencies to detect mental states in others on a purely behavioral basis? And doesn’t this support a third-person rather than firstperson approach to the fundamental grasp of mental-state concepts? BaronCohen (1995) identifies several behavioral cues for mental states and conjectures specialized innate mechanisms for exploiting those cues. An intentionality detector interprets self-propelled motion as a sign of goal or desire. An eye-direction detector detects the direction of another organism’s eyes and infers that that organism sees that thing (where seeing is a variety of mental state). I respond that the use of these behavioral cues is compatible with a firstperson, introspective understanding of mental-state concepts. Baron-Cohen suggests this himself: “I assume that the infant obtains [the knowledge that eyes can see] from the simple contingencies of closing and opening its own eyes … Although this knowledge is initially based on the infant’s own experience, it could be generalized to an Agent by analogy with the Self.” (1995, 43) Another possibility is innate cross-modal knowledge of mental states, analogous to the cross-modal knowledge discovered in infants by Meltzoff. Meltzoff and Moore (1983) found that newborn infants as young as 42 minutes can imitate such facial gestures as mouth-opening and tongue-protrusion. How do these infants correlate visual cues of the other’s movements with motor/proprioceptive information about their own movements? They suggest an innate cross-modal matching between visual and proprioceptive information. I suggest that there may be a similarly innate matching between third-person behavior and mental states understood in introspective terms. The child may be innately disposed to interpret other creatures’ self-propelled motion as stemming from a desire experience familiar from her own case. My position here contrasts sharply with that of some other simulation theorists, notably Gordon (1992, 1995, 1996). Gordon tries to develop an account of first-person attribution that avoids any appeal to introspection or inner recognition. He proposes what he calls an “ascent routine” approach. The way a person answers the question, “Do you believe Mickey Mouse has a tail?” (Q1) is to ask herself the question, “Does Mickey Mouse have a tail?” (Q2). If the answer to Q2 is Yes, then the presumptive answer to Q1 is also Yes. The answer to Q1 is No if either the answer to Q2 is No or no answer is available. What

20

Alvin Goldman

attracts Gordon to this approach is that it allows one to get the answer to a question about oneself, and specifically about one’s mental states, by answering a question that is not about oneself, nor about mental states at all (Gordon 1996, 15). Gordon wisely avoids the implausible claim that ascent routines alone yield genuine, comprehending self-ascriptions of belief. Ascent routines yield such comprehension, he says, only when they are embedded within a simulation, or when relativized to a point of view. I find it difficult to understand, though, how the two elements combine to yield genuine, comprehending self-ascriptions of belief. There is a further difficulty for the ascent routine approach: how can it account for self-attributions of attitudes other than belief? Consider hope, for example. Suppose someone is asked the question, “Do you hope that Team T won their game yesterday?” (Q1). How is she supposed to answer that question using an ascent routine? Clearly she is not supposed to ask herself the question, “Did Team T win their game yesterday?” (Q2), which would only be relevant to belief, not hope. What question is she supposed to ask herself? Gordon fails to address this, nor does he provide any reason to think that, for each attitude type, there is a distinctive non-mental question the response to which would provide an answer to the first-person attitude question. So his account is quite incomplete, restricted, as it is, to self-ascription of belief. Finally, Gordon does not really tell us, in cognitive terms, how people go about answering his de-mentalized questions (the Q2 questions). Introspection may be required to answer these questions, and if so, Gordon has not really avoided the necessity of introspection for accurate first-person attribution.

4. Simulation and Third-Person Attribution This section of the paper concerns the simulation theory (ST) alternative to the TT account of third-person attribution. Since my overall theory of attribution combines a simulation approach to third-person attribution with an introspectivist approach to first-person attribution, it is naturally called the introspection-simulation view. The basic idea of ST is that ordinary folk can attribute representational states to others by first imaginatively adopting their perspective, and then seeing what further states would ensue. For example, to determine what decision or choice another person will make, you pretend to have the desires and beliefs that you antecedently take them to have (on independent grounds). These pre-

The Mentalizing Folk

21

tend desires and beliefs are fed as inputs into your own decision-making system, which is run “off-line”. The output of the decision-making system does not lead you to action; rather, it is used as a basis to predict what the other person will do. At the heart of the simulation hypothesis is the notion that pretend mental states have some sort of intimate similarity, or homology, to their non-pretend counterparts. That is how they can succeed, on many occasions at least, in generating the same outputs as the non-pretend states would do when fed into the same processing system. Pretend states are obviously not identical to their non-pretend counterparts. It is an empirical question, to be discussed below, just what the respects of similarity and dissimilarity turn out to be. The simulation approach can be endorsed in various stronger or weaker variants. The strongest version would say that all third-person mental attributions by the folk involve the simulation routine. Weaker versions would say that some but not all attributions involve simulation. I have never endorsed the strongest version of the approach (see Goldman 1989, 1992a), nor do I wish to do so here. I wish to endorse only the weaker and admittedly vague view that simulation plays an “important” role in third-person attribution. The precise extent of simulation must be determined by future empirical investigation. It is noteworthy that many researchers who began in the TT camp, or still identify themselves as theory-theorists, have now moved over to a hybrid position, either explicitly or implicitly. So ST has already won numerous partial “converts.” Perner, who began as an unqualified opponent of ST (Perner and Howes 1992), now advocates a simulation-theory “mix” (Perner 1996). Carruthers (1996) calls himself a theory-theorist, but acknowledges a “limited” role for simulation. Similarly, Stich and Nichols heretofore among the staunchest foes of ST, now propose a sort of theory-simulation hybrid (Stich and Nichols 1997). Finally, although Leslie considers himself a theory-theorist, he makes substantial concessions to simulation in recent writing. He resists simulationism largely because he mistakenly thinks that it is inconsistent with metarepresentational conceptual structures. Leslie and German (1995) write as follows: In summary, we have few qualms about entertaining the idea that “simulation” may be one of the theory of mind related abilities. What these abilities have in common is that they use structured, systematic metarepresentational knowledge. Access to metarepresentations is required to define the problems to be solved, to initiate and guide the problem-solving process, to select relevant inputs for it,

22

Alvin Goldman

and to encode and interpret its intermediate and final results … [S]imulation needs metarepresentation. (Leslie and German 1995, 133) Moderate simulationism, however, can cheerfully live with metarepresentational concepts. Appropriate concepts are certainly needed to make attitude attributions, and they are needed for simulation-based attributions as much as theory-based attributions. Furthermore, some forms of propositional knowledge are undoubtedly needed both to initiate and to accompany the simulation heuristic. First, to engage in successful simulation, the attributor needs information about the target agent’s initial states on the basis of which accurate (initial) pretend states can be constructed. Second, while running a simulation, the pretend states must be represented or “tagged” as states of the target agent. At a minimum this must be done at the beginning and at the end of the simulation process, if not concurrently during its run. Otherwise, how can the simulator know that the output of the process should be attributed to the target agent, e.g., Jones, rather than someone else? Admitting the attributor’s need for such knowledge or information does not compromise the fundamental distinctness of the simulation approach. ST contrasts with pure TT in its positive claim that some attribution processes involve attempts to mimic the target agent, and in its negative claim that denies the use of theoretical propositions, such as scientific laws, in these attributional activities. Let me be more explicit about these differences. TT maintains that all attribution employs the attributor’s beliefs about the target’s initial states (which include desires, beliefs, and so forth). The attributor’s beliefs take the form: Ba(Dt(p)), Ba(Dt(q)), Ba(Bt(r)), Ba(Bt(s)), etc. These beliefs of the attributor, together with his belief in some (typically) nomological proposition(s) about folk psychology, Ba(N), are fed into the attributor’s theoretical inference mechanism (or a specialized module), which generates an output belief about a further mental state of the target. By contrast, the simulation heuristic allegedly works by starting with pretend states of the attributor. These pretend states are not supposed to be ordinary belief states, desire states, or the like, so let us distinguish them by an asterisk: Da*(p), Da*(q), Ba*(r), Ba*(s), etc. Although some of these states may be pretend belief states, others can be pretend mental states of any variety. To execute a simulation heuristic, the attributor feeds these pretend states into some cognitive mechanism which operates on them and produces a suitable output. In general, this need not be a theoretical inference mechanism. If the attributor wishes to predict a target agent’s decision,

The Mentalizing Folk

23

for example, the attributor will feed pretend desires and pretend beliefs into a decision-making mechanism, which will output a (pretend) decision. In the simulation heuristic, moreover, there is no need for any theoretical premise ‘N’. The prediction is accomplished by utilizing the attributor’s own cognitive mechanism rather than a theoretical premise about the operational pattern of such a mechanism. Having clarified the fundamental contrast between TT and ST, let me address a continuing source of resistance to ST. This stems from the Stich and Nichols contention that the simulation hypothesis is fundamentally at odds with the “dominant explanatory strategy” in cognitive science (Stich and Nichols 1992: 35), that it constitutes a “radical departure from the typical explanations of cognitive capacities” (Nichols et al. 1996: 39). I want to spend a little time now motivating the view that cognitive scientists should feel perfectly comfortable with the simulation hypothesis, because processes akin (in different ways) to mental simulation are already well established by empirical research. The processes I shall discuss in the next segment are not themselves instances of mind-reading, i.e., mental attribution, but acknowledgement of these processes should increase the scientific credibility of simulation as an approach to mind-reading. The essence of third-person mental simulation, in my view, is the deliberate attempt to replicate or reproduce in oneself an event or sequence of events that occurs (or is thought to occur) in somebody else. Even in successful simulation, there may be important differences between the target events and the facsimile in one’s own mind. Pretend beliefs and desires may differ in intrinsic as well as relational properties from natural, non-pretend beliefs and desires. There must be important respects of resemblance if the simulation heuristic is to succeed in making accurate attributions and predictions. It is an open, empirical question, however, just what the points of similarity and dissimilarity may be. At the outset, we can distinguish “pretend” mental states from their “natural” counterparts by noting that, in central cases, the former are produced artificially by the simulator’s act of will, or voluntary choice. Natural beliefs, desires, and other mental states arise spontaneously, not as a result of the subject’s voluntary control. The volitional source of simulation is something it shares with imagination. Seeing an elephant is having a visual experience of an elephant produced spontaneously and involuntarily by the visual system. Imagining an elephant, by contrast, is deliberately creating a vision-like experience of an elephant, at least in typical cases. Other cognitive resources beyond mere willing are

24

Alvin Goldman

undoubtedly utilized in generating a visual image of an elephant. But it suffices for present purposes to highlight the contrast between natural and artificial (voluntary) modes of producing the two events. Imagination also resembles simulation in that the facsimile event (the visual image of an elephant) does not share all cognitive properties with its counterpart (the actual seeing of an elephant). Exactly which properties are and are not shared is a question for cognitive science to resolve, but substantial neuropsychological commonalities are already well established (Currie 1995; Farah 1988; Kosslyn et al. 1993). For example, there are content specific interactions between imagery and perception: imaging an “H” affects detection of visually presented “H”s more than it affects detection of “T”s. And patients who suffer visual deficits from cortical damage suffer parallel imagery deficits as well. For those who doubt that simulation events ever transpire, who doubt that such facsimile events should be countenanced by cognitive science, it is instructive to be apprised of such research findings. The existence of mental events that are deliberately produced facsimiles of other mental events is established by this imagery research, and their existence makes it more credible than it otherwise might be that mental attribution could consist of analogous facsimile events. Currie (1995) rightly emphasizes this parallel with mentalistic simulation. Nichols et al. (1996) challenge Currie’s linkage of simulation and visual imagery. If simulation is to account for visual imagery in a novel fashion, they argue, either a distinct pretend-input generator for imagery must be postulated, or the same device must be used as in the case of decision-making simulation. Neither of these assumptions, they claim, is plausible. If Currie means to be claiming that existing theories of visual imagery are simulationist, Nichols et al. deny this on the grounds that ST is committed to the “no information” assumption, whereas leading theories of visual imagery do appeal to an information base. These criticisms do not seem very damaging to me. Although theory-theorists commonly postulate pretend-input generators in their expositions of ST, that is not really essential to ST. ST can hold that pretend-states are generated by some voluntary marshalling of pertinent informational resources without as yet committing itself to either task-specific or task-general devices. The important point is that there are products that mimic “naturally” produced states, however exactly these are produced. Furthermore, it is no embarrassment to ST for it to acknowledge that the generation of such facsimile events partly involves information. The no-information assumption is one that TT

The Mentalizing Folk

25

proponents foist off on ST, not an assumption that ST proponents are bound (or ought) to accept. A further striking parallel is presented by Currie and Ravenscroft (1997) in a treatment of motor imagery. Here too it is found that subjects can deliberately produce facsimiles of naturally produced motor-controlling cognitions, where the facsimiles bear remarkable similarities to their targets. Let me briefly summarize some of the salient findings reported by Currie and Ravenscroft. Sports psychologists have long recognized the ability of mental rehearsal to enhance athletic performance (Richardson 1967). Athletes who mentally rehearse their performance show marked improvement over those who do not. In “internal” imagery, the athlete experiences sensations of both effort and movement, and this kind of imagery has a more marked effect than so-called “external” imagery, which is picturing oneself as a spectator might see one. This suggests the hypothesis that internal motor imagery is generated by deliberately or artificially running the motor system but without letting it generate its normal behavioral outcome. This hypothesis leads to a number of predictions. First, temporal and kinematic properties of movements should be reflected in motor imagery. Second, when damage to central parts of the motor cortex cause deficits in motor performance, it should also cause impairment in motor imagery. Third, motor control should be improved by motor imagery. Fourth, motor imagery should affect heart and respiration rates, since activating the motor areas of the brain affects these rates. Finally, the neural substrate of movement should overlap the substrate of motor imagery. All of these predictions are borne out to a remarkable extent, as Currie and Ravenscroft document. Yue and Cole (1992) compared the increase in muscular strength among subjects who actually trained with the strength of subjects who merely generated the appropriate motor imagery. Remarkably, while actual training produced a 30% increase in maximal force, motor imagery produced a 22% increase. EMG recordings revealed that the subjects in the motor imagery condition did not make covert muscle contractions during imagery, so the explanation of the effect must be a central explanation. Concerning heart and respiration rate increases, Adams et al. (1987) demonstrated that these changes must originate in the (central) motor system itself. When the motor system instructs the large muscles to begin vigorous action, a parallel signal is sent to the vegetative effectors. We might therefore expect an increase in both heart and respiration rates during motor imagery of vigorous exercise. This is precisely what is found. Decety et al. (1991) report that not merely did heart rate and respiration rate

26

Alvin Goldman

increase during imaged exercise, they increased in proportion to the imagined vigorousness of the exercise! The foregoing material supports the existence of intentionally produced facsimiles of mental events, but it does not address the use of any such facsimiles for interpersonal interpretation. To strengthen the plausibility of this hypothesis, I first adduce evidence of interpersonal behavioral imitation on the part of young children, and then argue that mentalistic imitation would be a natural extension or outgrowth of behavioral imitation. As discussed earlier, Meltzoff and Moore established the existence of infant imitation of facial expressions 15-20 years ago. In more recent work, they show how imitation is central to the infant’s understanding of people. Meltzoff and Moore (1995) provide evidence that 6-week-old infants use facial imitation to differentiate among individuals they encounter and to reidentify them. For example, on day 1 they saw a single adult demonstrate either mouth opening, tongue protrusion at midline, tongue protrusion to the side, or no oral movement (control). On day 2 the same adult first sat and presented a neutral face, then the assigned gesture was demonstrated again. The results showed both immediate imitation (when the gestures were shown) and imitation from memory (when the adult sat with a neutral face). Thus, infants were acting on their remembrance of things past. Meltzoff and Moore suggest that infants are using these gestural signatures to identify the individuals. As discussed earlier, Meltzoff (1995) has shown that by 18 months children go beyond imitating the visible surface behavior of an adult to imitating acts that are intended but not executed. Thus, at this age, children’s replication proclivities extend, as it were, into the adult’s psychology, not simply to his surface behavior. None of this demonstrates, of course, that children and adults go on to deploy a process of mental imitation, in which they try to interpret others’ mental states through the medium of intentionally produced facsimiles in their own minds. (More precisely, models of mental simulation suggest that one intentionally initiates a sequence of mental events intended to mimic another mind, but the whole sequence is not under direct voluntary control. Later stages are controlled by whatever system operates on the initially generated “inputs”.) But we can say this much. Meltzoff’s work suggests that the capacity for intentional imitation of behavior is an innate and fundamental capacity of human nature. Currie and Ravenscroft show that people have capacities for intentional production of mental facsimiles as well, at least facsimiles of their own mental states. It is not too dramatic a leap, therefore, to conjecture that people might intentionally deploy mental facsimiles to interpret the mental

The Mentalizing Folk

27

states of others, as ST proposes. The evidence at least warrants rejection of the claim that ST is a “radical departure” from accepted trends in cognitive science. How can more direct evidence be obtained for or against mental simulation? One type of argument for the existence of mental simulation is the absence of pertinent theoretical information that could guide predictors to their conclusions. Undermining TT in this fashion would indirectly support ST, which is TT’s only competitor. This was the type of argument I offered with the airport example (see section 2.1). Another type of case is Harris’s (1992) hypothetical example in which one person is asked to predict another’s judgments of the grammaticality (or acceptability) of various sentences. Since members of the folk possess comparatively little theoretical knowledge of grammaticality, Harris plausibly contends that a person might execute this task by simulation, by making her own judgments and assuming that the other person would judge similarly. This is not a standard form of simulation, however, as Stich and Nichols (1997) rightly point out. Here the predictor places herself in the actual position of the predictee, rather than pretending to be in the predictee’s situation. A second general type of evidence in favor of ST might exploit what I regard as the basic difference between ST and TT. This consists in the fact that ST posits facsimile tracking of another person’s mental states, at least on some occasions of third-person mental interpretation, whereas pure TT denies that this ever occurs. By “facsimile tracking” I mean the production in the interpreter’s own mind of a sequence of states that resemble, in crucial cognitive respects, the intended target of interpretation or attribution. TT agrees, presumably, that interpretation and successful attribution typically involve some sort of “tracking” of the target’s states, but in this case the tracking would be purely doxastic or intellectual. It would consist of a series of conjectures and/or beliefs about the mental states of the target, and in general these beliefs would not be facsimiles of their counterparts. Instead of replicating or emulating the mental activity of the target, the interpreter would be a pure “observer”or spectator. Given this difference between the views, some observable predictions might be deducible. Facsimile tracking should predict telltale feelings, behavior, or physiological symptoms that intellectual tracking would not predict. Of course, when the target mental activity is merely theoretical inference or decision making, even facsimile tracking would not predict any noteworthy physiological symptoms or unusual feelings. But facsimile tracking might predict such feelings where the target mental activity itself yields affect or other distinctive sensations. If a predictor engaged in tracking

28

Alvin Goldman

such a target herself undergoes facsimiles of such feelings, this might be evidence in favor of ST. So I have argued in linking ST with empathy (Goldman 1992b, 1993b). For example, people asked to predict the feelings of the parents of a kidnapped child might themselves be found to have empathic emotional reactions, especially if the circumstances are described in detail. This would suggest that the predictors have engaged in simulative (role-taking) tracking of the parents’ feelings, not merely intellectual tracking. If the tracking had been purely intellectual, why would the predictors themselves experience homologous emotions? A similar argument could be made for readers of a romantic novel who experience feelings homologous to those of the novel’s heroine. And many people experience feelings of discomfort or queeziness when observing the pierced tongues or pierced lips of contemporary teen-agers. Such feelings are readily explained on the assumption that the observer tracks the imagined experience of having the organ pierced, and tracks it in a simulative or replicative mode. Nichols et al. 1996) reply that empathy can be explained in informationbased terms. “Empathic responses might arise when the subject is reminded of events in her past similar to those of the object of empathy. So, for example, if your friend tells you that her dog has died, you might empathise with her via remembering the death of your own dog” (Nichols et al. 1996: 61). They even cite researchers on empathy (Hoffman 1984, Eisenberg and Strayer 1987) who mention the memory explanation as one possibility. However, memory does not accommodate all cases. You can empathize with the parents of a kidnapped child even if you have never yourself suffered such a tragedy and therefore lack any such memory. Furthermore, even if memory does play a role in generating empathic affect, that is not incompatible with a simulation story. Perhaps memories simply assist the construction of a simulation, which then produces the affect. Remembering different wounds of your own might help you imagine your tongue or lip being pierced. This imagination might be necessary to elicit discomfort or queeziness. A purely factual memory of having suffered past wounds might not suffice to induce queeziness or discomfort. It has already been found, experimentally, that imaginative involvement plays a role in such matters. Stotland (1969) had subjects watch someone else whose hand was strapped into a machine, which was said to generate painful heat. Some subjects were instructed just to watch the man carefully, some to imagine the way he was feeling, and some to imagine themselves in his place. Using both physiological and verbal measures, the experimental results showed that the deliberate acts of imagination produced a greater response than just watching. All

The Mentalizing Folk

29

of this requires more detailed empirical research, but initial evidence seems to support a role for simulative tracking. Among the attempts to produce experimental evidence against ST, the most widely discussed are Stich and Nichols’ experiments based on considerations of “cognitive penetrability” (Stich and Nichols 1992; Nichols et al.1996). Stich and Nichols made the theoretical claim that if ST is correct, what attributors don’t know about psychology won’t hurt them, that is, won’t lead to false attributions. If people’s decision-making propensities have a quirky feature, theoretical ignorance of that quirk shouldn’t prevent simulative predictors of their decisions from predicting correctly, because as simulators they would just use the same psychological mechanisms, which should yield similar outputs given similar inputs. Stich and Nichols proceeded to study people’s predictions of choice behavior shown by psychologists to be quirky, to see if the predictions accord with what ST (allegedly) implies. For example, Langer (1975) reported an experiment in which subjects who agreed to buy a ticket in a lottery were later offered an opportunity to sell their ticket. Langer found that subjects who were offered an opportunity to choose their own ticket set a significantly higher price than those who were simply handed a ticket when they agreed to buy one. Stich and Nichols performed an experiment (reported in Nichols et al. 1996) in which subjects saw videotapes of targets in a Langerstyle situation and were then asked to predict their sell-back price. Some videotaped targets got to choose their own ticket while others were simply given tickets, and subjects were shown either the choice situation or the no-choice situation. Subjects viewing the choice version did not predict higher values than those viewing the no-choice version. Given the background theoretical assumption, Nichols et al. (1996) argued that this evidence undermines ST. There are several reasons why these results are not decisive, however. First, the moderate form of ST that I endorse does not claim that all behavior prediction is simulation based. So even if simulation was not used in the Langer-style task, this does not show that it is never used. Second, it is an open question, as I have emphasized, exactly which properties mental facsimiles share with their natural counterparts. Even the best simulation may not perfectly replicate its target. So it cannot be precluded that observer subjects in the experiment did engage in facsimile tracking but still failed to reproduce the price-setting behavior of their targets. Third, Kuhberger et al. (1995) failed on several occasions to replicate the original findings of Langer, and this raises doubts about all dimensions of this line of experimentation. Nichols, Stich and Leslie (1995) reply to Kuhberger et al., but the dispute is far from settled. Finally, Heal

30

Alvin Goldman

(1996) has challenged the theoretical claim that misinformation cannot penetrate a predictor’s simulation process. In response to Heal’s challenge, Stich and Nichols (1997) now concede that it is a mistake to make the general claim that simulation is cognitively impenetrable. On the other hand, it is not clear how this helps an ST theorist explain what was happening in their Langer-style experiment, because it is not clear what belief might have penetrated the predictors’ simulation process, if simulation is what they were using.

5. Conclusion The inconclusiveness of the experiments just reviewed is typical of the experimental forays made thus far to decide between TT and ST. Pending more conclusive direct experimentation, the prospects for the two approaches must be judged by converging evidence from many sources of the kinds reviewed earlier. This evidence undercuts pure forms of TT and gives substantial support to ST, at least in the weak form that assigns simulation an “important” (not necessarily universal) role in third-person attribution. First, the detailed articulations of TT either face serious empirical problems or turn out on careful inspection not to be so clearly versions of TT rather than ST. Second, TT suffers from a failure to acknowledge or deal adequately with people’s ability to attribute mental states to themselves in a distinctive manner. The introspection-simulation approach deals with this problem quite naturally. Third, evidence from several domains of cognitive research, including visual imagery, motor imagery, neonate imitation and empathy, make it highly plausible that people could have a skill at producing mental facsimiles that is deployed in interpreting the states of others. These kinds of evidence may be among the reasons why many of the most resolute critics of ST have recently moved toward a hybrid position, a position that is difficult to distinguish from weak simulationism. There is, no doubt, much work to be done to articulate ST with greater precision, and to fill in the developmental story in an ST-friendly way. At the present juncture, however, the prospects for ST look extremely bright.

The Mentalizing Folk

31

Acknowledgments I wish to thank Tim Bayne, Gregory Currie, Robert Gordon, Alan Leslie, Shaun Nichols, Joel Pust, Ian Ravenscroft, and Dan Sperber for helpful comments and feedback on earlier drafts of this paper. Reprint from Dan Sperber (ed.), Metarepresentation, Oxford University Press (2000).

References Adams, L., A. Guz, J. Innes, and K. Murphy. “The Circulatory and Ventilatory response to Voluntary and electrically induced Exercise in Man.” Journal of Physiology 383 1987,19-30. Baron-Cohen, Simon. Mindblindness. Cambridge, MA: MIT Press 1995. Block, N. “Troubles with Functionalism.” Readings in Philosophy of Psychology, vol. 1. Edited by N. Block. Cambridge, MA: Harvard University Press 1980. Carey, S., and E. Spelke. “Science and Core Knowledge.” Philosophy of Science 63,1996,515-533. Carruthers, P. “Simulation and Self-Knowledge: A Defence of Theory-Theory.” Theories of Theories of Mind. Edited by P. Carruthers and P. Smith. New York: Cambridge University Press 1996. Clements, N., and J. Perner (1994). Cognitive Development 9 1996, 377-395. Currie, G.“Visual Imagery as the Simulation of Vision.” Mind and Language 10, 1995, 25-44. —, and I. Ravenscroft. “Mental Simulation and Motor Imagery.” Philosophy of Science 64, 1997, 161-180. Davies, Martin and Tony Stone (eds.). Folk Psychology. Oxford: Basil Blackwell 1995a. — (eds.). Mental Simulation. Oxford: Basil Blackwell 1995b. Decety, J., H. Sjoholm, E. Ryding, G. Stenberg, and D. Ingvar. “The Cerebellum Participates in Cognitive Activity: Tomographic Measurements of Regional Cerebral Blood Flow.” Brain Research 535, 1990, 313-317. Eisenberg, N., and J. Strayer. “Critical Issues in the Study of Empathy.” Empathy and Its Development. Edited by N. Eisenberg and J. Strayer. New York: Cambridge University Press 1987. Farah, M. “Is Visual Imagery really Visual? Overlooked Evidence from Neurophysiology.” Psychological Review 95, 1988,307-317. Freeman, N., and H. Lacohee. Making explicit 3-year-olds’ implicit Competence with their own false Beliefs. Cognition 56, 1995, 31-60. Goldman, A. “Interpretation psychologized.” Mind and Language 4, 1989,161-185. Reprinted in Davies and Stone (1995a). —. “In Defense of the Simulation Theory.” Mind and Language 7: 104-119, 1992a. Reprinted in Davies and Stone (1995a).

32

Alvin Goldman

—. “Empathy, Mind, and Morals.” Proceedings and Addresses of the American Philosophical Association 66 (3) 1992b. Reprinted in Davies and Stone (1995b). —. “The Psychology of Folk Psychology.” Behavioral and Brain Sciences 16, 1993, 1528. —. Philosophical Applications of Cognitive Science. Boulder, CO: Westview Press 1993b. —. “Science, Publicity, and Consciousness.” Philosophy of Science 64, 1997, 525-545. Gopnik, A. “How we know our Minds: The Illusion of First-Person Knowledge of Intentionality.” Behavioral and Brain” Sciences 16, 1993, 1-14. Gopnik, A. “The Scientist as Child.” Philosophy of Science 63, 1996, 485-514. —, and Andrew Meltzoff. Words, Thoughts, and Theories. Cambridge, MA: MIT Press 1997. —, and H. Wellman. “Why the Child’s Theory of Mind really is a Theory.” Mind and Language 7, 1992, 145-171. Reprinted in Davies and Stone (1995a). —, and H. Wellman. The Theory Theory. Mapping the Mind. Edited by L. Hirschfeld and S. Gelman. New York: Cambridge University Press 1994. Gordon, R. “The Simulation Theory: Objections and Misconceptions.” Mind and Language 7, 1992, 11-34. Reprinted in Davies and Stone (1995a). —. “Simulation without Introspection or Inference from me to you.” Davies and Stone (1995b). —. “‘Radical’ Simulationism.” Theories of Theories of Mind. Edited by P. Carruthers and P. Smith. New York: Cambridge University Press 1996. Harris, P. “From Simulation to Folk Psychology.” Mind and Language 7, 1992, 120144. Reprinted in Davies and Stone (1995a). — (1996). “Desires, Beliefs, and Language.” Theories of Theories of Mind. Edited P. Carruthers and P. Smith. New York: Cambridge University Press. Heal, J. “Simulation and Cognitive Penetrability.” Mind and Language 11, 1996, 44-67. Hoffman, M. “Interaction of Affect and Cognition in Empathy.” Emotions, Cognition, and Behavior. Edited by C. Izard, J. Kagan, and R. Zajonc. New York: Cambridge University Press 1984. Humphrey, Nicholas (1983). Consciousness Regained. New York: Oxford University Press. Johnson, S., and S. Carey. “Distinguishing Knowledge Enrichment from Conceptual Change: Evidence from Williams Syndrome.” Unpublished paper, Department of Psychology, New York University 1996. Kahneman, D., and A. Tversky. “The Simulation Heuristic.” Judgment under Uncertainty. Edited by D. Kahneman, P. Slovic, and A. Tversky. Cambridge: Cambridge University Press 1982. Kosslyn, S., N. Alpert, W. Thompson, V. Maljkovic, S. Weise, C. Chabris, S. Hamilton, S. Rauch, and F. Buannano. “Visual Mental Imagery Activates topographically organized visual Cortex: PET investigations.” Cognitive Neuroscience 5, 1993, 263287. Kuhberger, H., J. Perner, M. Schulte, and R. Leingruber. “Choice or no choice: Is the

The Mentalizing Folk

33

Langer Effect evidence against simulation?” Mind and Language 10, 1995, 423-436. Langer, E. “The Illusion of Control.” Journal of Personality and Social Psychology 32, 1975, 311-328. Leslie, A. (1987). “Pretence and representation: The ‘origins’ of theory of mind.” Psychological Review 94, 1987, 412-426. —. “Pretending and Believing: Issues in the Theory of ToMM.” Cognition 50, 1994, 211-238. —, and T. German. “Knowledge and Ability in ‘Theory of Mind’: One-eyed Overview of a Debate.” Davies and Stone (1995b). —, and D. Roth. “What Autism teaches us about Metarepresentation.” Understanding Other Minds: Perspectives from Autism. Edited by S. Baron-Cohen, H. TagerFlusberg, and D. Cohen. Oxford: Oxford University Press 1993. —, and L. Thaiss. “Domain specificity in Conceptual Development: Neuropsychological Evidence from Autism.” Cognition 43, 1992, 225-251. Lewis, C., N. Freeman, C. Hagestadt, and H. Douglas. “Narrative Access and Production in Preschoolers’ false belief Reasoning.” Cognitive Development 9 1994, 397424. Lewis, D. “Psychophysical and Theoretical Identifications.” Australasian Journal of Philosophy 50, 1972, 249-258. Meltzoff, A. “Understanding the Intentions of Others: Re-enactment of Intended Acts by 18-month-old Children.” Developmental Psychology 31, 1995, 838-850. —, and M. Moore. “Newborn Infants imitate adult facial Gestures.” Child Development 54 1983, 702-709. —, and M. Moore. “Infants’ Understanding of People and Things: From Body Imitation to Folk Psychology.” The Body and the Self. Edited by J. Bermudez, A. Marcel, and N. Eilan. Cambridge, MA: MIT Press 1995. Nichols, S., S. Stich, A. Leslie, and D. Klein (1996). “Varieties of Off-Line Simulation.” Theories of Theories of Mind. Edited by P. Carruthers and P. Smith. New York: Cambridge University Press. —, S. Stich, and A. Leslie (1995). “Choice Effects and the Ineffectiveness of Simulation.” Mind and Language 10, 1995,437-445. Perner, Josef (1991). Understanding the Representational Mind. Cambridge, MA: MIT Press 1991. — (1996). “Simulation as Explicitation of Predication-implicit Knowledge about the Mind: Arguments for a Simulation-Theory Mix. Theories of Theories of Mind. Edited by P. Carruthers and P. Smith. New York: Cambridge University Press. Richardson, A. “Mental Practice: A Review and Discussion, Parts I and II.” Research Quarterly 38,1967, 95-107; 262-273. Rosenthal, D. (1992). “Thinking that one Thinks.” Consciousness. Edited by M. Davies and G. Humhreys. Oxford: Blackwell. Schacter, D. “On the Relation between Memory and Consciousness.” Varieties of Memory and Consciousness. Edited by H. Roediger and F. Craik. Hillsdale, NJ: Erlbaum 1989.

34

Alvin Goldman

Schiffer, S. Remnants of Meaning. Cambridge, MA: MIT Press 1987. Segal, G. “The Modularity of Theory of Mind.” Theories of Theories of Mind. Edited by P. Carruthers and P. Smith. New York: Cambridge University Press 1996. Sellars, W. “Empiricism and the Philosophy of Mind.” Science, Perception and Reality. London: Routledge 1963. Stich, S., and S. Nichols. “Folk Psychology: Simulation or Tacit Theory?” Mind and Language 7, 1992, 35-71. Reprinted in Davies and Stone (1995a). —, and S. Nichols. “Second Thoughts on Simulation.” Davies and Stone (1995b). —, and S. Nichols. “Cognitive Penetrability, Rationality and Restricted Simulation.” Mind and Language 12, 1997, 297-326. Stotland, E. “Exploratory Investigations of Empathy.” Advances in Experimental Social Psychology, vol. 4. Edited by L. Berkowitz. New York: Academic Press 1969. Tager-Flusberg, H. “Theory of Mind in Young Children with Williams Syndrome.” Paper presented at the Biennial Meeting of the Williams Syndrome Association, San Diego 1994. Wellman, Henry. The Child’s Theory of Mind. Cambridge, MA: MIT Press 1990. Yue, G., and K. Cole. “Strength increases from the motor program: Comparison of training with maximal voluntary and imagined muscle contractions.” Journal of Neurophysiology 67, 1992, 1114-1123.

Belief and Acceptance as Features of Groups

35

Belief and Acceptance as Features of Groups Margaret Gilbert

Abstract In everyday discourse groups or collectives are often said to believe this or that. The author has previously developed an account of the phenomenon to which such collective belief statements refer. According to this account, in terms that are explained, a group believes that p if its members are jointly committed to believe that p as a body. Those who fulfill these conditions are referred to here as collectively believing* that p. Some philosophers – here labeled rejectionists – have argued that collective belief* is not belief but rather acceptance. This paper presents several arguments against rejectionism. One has to do with the proper methodology for arriving at an account of belief. Two address rejectionist claims to the effect that collective beliefs* lack key features of belief in general, the features in question being “aiming at truth” and having a particular relation to the will. A fourth notes that there is a phenomenon more apt for the label of “collective acceptance” than is the phenomenon of collective belief*.

1. Introduction 1.1 Collective belief statements Statements such as the following are commonplace: “The United States believes that those responsible for these dreadful acts must be punished.” “The teachers’ union believes it can negotiate a significant pay raise for its members.” “Our family believes that the best way to spend Christmas is at the beach”. In these and many other cases, a belief is ascribed to something – the United States, the union, our family – which is in some sense made up of individual human beings. For present purposes I shall refer to things of this kind as collectives or groups. I shall refer to such statements as collective belief statements.

36

Margaret Gilbert

1.2 What is collective belief? To what phenomenon are collective belief statements intended to refer? I shall call this phenomenon, whatever it is, collective belief.1 In a classic social scientific discussion, Emile Durkheim includes the beliefs of groups among the phenomena he denominates paradigmatic “faits sociaux” or social phenomena. Among other things, he says that such beliefs are “independent of their individual manifestations” and possess a certain “coercive power” and “authority”.2 Many have found Durkheim’s discussion obscure. It does, indeed, cry out for further elaboration.3 Analytic philosophers, established purveyors of clarity and rigor, have traditionally paid no attention to collective belief. This includes those who focus on epistemology, for which a central issue is the nature of belief. More recently, even those who characterize what they are doing as social epistemology have neglected collective belief in favor of consideration of aspects of the beliefs of individuals. 1.3 The plural subject account of collective belief In a number of places I have been developing and defending an account of collective belief. For reasons that will emerge, I refer to it as the plural subject account.4 In this essay I focus on some claims that have been made concerning collective belief when it is understood according to this account or in a roughly similar way. 1 My writing of the “phenomenon” as opposed to the “phenomena” is not intended to rule out the possibility that there are importantly different types of collective belief. 2 Emile Durkheim, Rules of Sociological Method, ed. S. Lukes (New York: Free Press, 1982). First published in French 1895. 3 For a proposed articulation of the core elements of Durkheim’s account of social phenomena in general and collective belief in particular, see Margaret Gilbert, On Social Facts, (Princeton: Princeton University Press, 1989), Chapter 5; also Margaret Gilbert, “Durkheim and Social Facts” in Debating Durkheim, ed. Herminio Martins and Donald Pickering (London: Routledge, 1994), 86-109. 4 See Margaret Gilbert, “Modeling Collective Belief”, Synthese 73 (1987): 185-204; Social Facts; “Durkheim”; “Remarks on Collective Belief” in Socializing Epistemology, ed. Frederick Schmitt (Lanham, MD: Rowman and Littlefield, 1994), 235-256; “Collective Belief and Scientific Change” in Margaret Gilbert, Sociality and Responsibility: New Essays in Plural Subject Theory (Lanham, MD: Rowman and Littlefield, 2000), 37-49. “Modeling” and “Remarks” are reprinted in Margaret Gilbert, Living Together: Rationality, Sociality, and Obligation (Lanham, MD: Rowman and Littlefield, 1996), whose introduction is also pertinent.

Belief and Acceptance as Features of Groups

37

1.4 “Rejectionism” An antecedently discussed philosophical distinction has been brought to bear on the plural subject account of collective belief. This is a distinction between “belief”, on the one hand, and “acceptance”, on the other. A number of philosophers working in epistemology have made a distinction in these terms. These include Michael Bratman, Jonathan Cohen, Keith Lehrer, Robert Stalnaker, and Bas Van Fraassen. Though they use the same terminology, they do not always have the same distinction in mind. In the case of Lehrer, for instance, “belief” is the broader term, acceptance a special form of belief.5 For others, including Stalnaker, “accept” is the broader term, belief being a special form of acceptance.6 In yet others, there are no mutual entailments between belief and acceptance, though believing a proposition is compatible with accepting it on occasion.7 Referring to a distinction such that acceptance is not a form of belief, a number of philosophers have suggested that collective belief, according to my account of it, is not belief, but rather acceptance.8 For the sake of a label, I shall call their position “rejectionism”. The philosophers I have in mind are not opposed to my account of collective belief. That is, they do not reject the plural subject account of the referent of collective belief statements.9 Rather, they reject the idea that the phenomenon 5 See, for instance, Keith Lehrer, Self-Trust: A Study of Reason, Knowledge and Autonomy (Oxford: Clarendon Press, 1997). 6 Robert Stalnaker, Inquiry (Cambridge, MA: MIT Press, 1984). 7 See Jonathan Cohen, An Essay on Belief and Acceptance (New York: Oxford University Press, 1992); Bas Van Fraassen, The Scientific Image(Oxford: Clarendon Press, 1980). 8 See Anthonie Meijers, “Believing and Accepting as a Group”, in Belief, Cognition and the Will, ed. A. Meijers (Tilburg; Tilburg University Press, 1999), 59-71; K. Brad Wray, “Collective Belief and Acceptance”, Synthese 129 (2001):319-333. See also Raimo Tuomela, “Group Beliefs”, Synthese 91 (1992):285-318; The Importance of Us (Stanford: Stanford University Press, 1995; “Belief versus Acceptance”, in Meijers, ed. Belief, Cognition, 319333, and elsewhere). This suggestion has also come up informally, once in the early 1990s in Dan Sperber’s Princeton graduate class where the students were discussing my account of collective belief. I understand that Professor Meijers has a new discussion of the topic that will be published along with the present essay in Protosociology. He has not, however, given up rejectionism. I limit myself here to discussion of his 1999 article. 9 Meijers and Wray do not question the broad outlines of my account. Wray explicitly accepts it and draws on it. Tuomela, Importance, grants it a measure of plausibility (310), and offers an account of his own that resembles it in various ways (450, note 9. He sees it as a consequence of his own account that so-called group beliefs are not beliefs proper (311). See also his “Group Beliefs”.

38

Margaret Gilbert

the account describes is a form of belief, as opposed to acceptance. Thus K. Brad Wray, “… the phenomenon that concerns Gilbert is a species of acceptance [rather than belief]”.10 For present purposes, then, I shall assume that my account of the referent of collective belief statements is essentially correct. I detail the account shortly. I shall henceforth refer to the phenomenon it delineates as collective belief*. The asterisk is intended to indicate that the status of this phenomenon as a form of belief is in dispute. I shall continue to refer to statements in which so-called beliefs are ascribed to collectives as collective belief statements. If rejectionists are right, these statements are at best misleading, at worse false. For, if they are right, in spite of the way it is referred to in everyday speech, the phenomenon in question is not belief, but rather acceptance. 1.5 Acceptance I shall focus on what rejectionists say about belief, and the question of whether they have shown that collective belief* is not belief. A word might usefully be said at the outset, however, as to the nature of acceptance as they understand it. A canonical example of acceptance, for all concerned, I think, is a case of assuming something “for the sake of the argument”, or “for present purposes”. Thus, in the text above, I explained that for present purposes I shall assume collective belief* is the referent of collective belief statements. As it happens, this is what I believe, irrespective of present purposes. Importantly, one can assume something for the sake of argument without believing it.11 1.6 The argument of this essay In this essay I question rejectionism. My argument proceeds along four different paths. The first concerns a fundamental methodological question for epistemology, in particular: how is one to arrive at an account of belief in general? Evidently this question is important for other fields of philosophy, including the philosophy of mind, as well as for other disciplines also. The second and 10 Wray, “Collective Belief”, 319. See also Meijers, “Believing”, 67-68. 11 Cf. Wray (referring to common ground between Stalnaker and Van Fraasen): “For example, one might accept something that one does not believe for the sake of an argument”, 324; Meijers: “To accept that p is to make the decision to use a certain proposition as a valid premise in one’s deliberations”, 64.

Belief and Acceptance as Features of Groups

39

third arguments concern two features that have been claimed to be key features of belief in general: a truth-seeking nature, and a particular relation to the will. Collective beliefs* have been said to lack these features. The fourth argument appeals to the difference between collective belief* and a related phenomenon that naturally attracts the label “collective acceptance”. I hope my discussion as a whole will encourage epistemologists and others to take collective belief statements seriously. I hope, also, that it will clarify in a number of ways the nature of collective belief*. In my view collective belief* is a widespread and consequential phenomenon under any name. That this is so is not disputed by rejectionists. 1.7 Cognitive states and collective mentality Belief and acceptance are what I shall here refer to as cognitive states.12 One rough and ready way of giving a general account of cognitive states is to say that they are states whose specification requires reference to a proposition, such as the proposition that two plus two equals four, or the proposition that it is raining in Dubuque right now. This is not the place to attempt to characterize propositions, so that idea must remain at an intuitive level here. Cognitive states, on this account of them, may vary widely in their specific character. Where the letter “p” stands for some proposition, cognitive states will include believing that p, wondering whether p, doubting whether p, and accepting that p. It is natural to think of cognitive states as mental states, though not as the only species of such states and perhaps not the paradigmatic ones. Now it is quite common for people to argue against the possibility of collective mental states in general or, to use a phrase usually intended to raise hairs, group minds. What is one to make of the rejection of group minds? What bearing does it have, if any, on the present inquiry? Those who argue, or simply state, that there are no group minds often forbear to explain their position. They tend to take it to be obvious. At first blush, however, it appears to fly in the face of everyday attributions of belief and so on to groups. One construal of the no-group-minds thesis that fits at least some versions is a form of reduction. The position is roughly this: insofar as collective beliefs 12 This phrase may sometimes be used more narrowly than I use it here. That does not matter for present purposes.

40

Margaret Gilbert

and so on exist they are a function of facts about individuals. Individuals are, if you like, the bedrock of collective beliefs. If so construed, the no-group-minds thesis is not obviously antagonistic to everyday attributions of belief or other cognitive and mental states to groups. Nor does it clearly stand in opposition to my own views. There is at least a sense in which I would not be averse to describing the bedrock of collective beliefs* as individuals.13 However, to say that is to say very little about them. For present purposes, much more must be said.

2 Collective belief* 2.1 The range of collective belief statements When we consider the collective belief statements of everyday life, the supposed believers are quite diverse, at least on the surface. Two strangers, for instance, having conversed for a while, might say to a third party “We think a defensive war is justified”. This might also be said of a family, a political party, a committee, a governing body. Presumably, all families, political parties, governing bodies – all relatively long-lasting, more or less organized or hierarchically structured groups – have come about ultimately through small-scale, short-term, informal processes involving a few individuals in face-to-face contact. There is reason, then, to consider collective belief statements made on the basis of such processes as the basic cases and to start with them. It can be argued that the account I sketch shortly is adequate both for these basic cases and at least some others.14 2.2 Collective belief statements and the distributive condition Consider the case of the strangers who explain to a third party what “we” think. Suppose Roz and Mark are the two people involved. Some may claim that this is not really what I am calling a collective belief statement. That is, contrary to appearances, it is not the ascription of a belief to a collective comprising Roz and Mark. It is, rather, an ascription of the belief in question to Roz, on the one hand, and to Mark, on the other, expressed elliptically. The 13 On the issue of “individualism” versus “holism”, see Gilbert, Social Facts, Chapter 7. 14 See, for instance, Gilbert, Social Facts, Chapter 5.

Belief and Acceptance as Features of Groups

41

non-elliptical way of expressing this ascription in English would be “We both believe that p.” The only alleged believers in the situation, therefore, are two individuals. One of these individuals is Roz, the other is Mark. There is no further believer, comprised of Roz and Mark.15 Sentences of the form “We believe that p” may sometimes be uttered as elliptical expressions. Suppose Jeb has just said to Roz and Mark “I don’t know what Roz thinks about this, and I don’t know what Mark thinks about it either …?” In responding “We think a defensive war is justified” Roz may mean, and reasonably assume she will be understood to mean, that both think a defensive war is justified. Insofar as, in these cases, sentences of the form “We think that p” are elliptical for “We both think that p”, or “We all think that p” then one is not dealing with what I am referring to as collective belief statements. So we may set them aside for present purposes. A collective belief statement is canonically expressed in English by a nonelliptical sentence of the subject-predicate form where the subject term refers to a collective and the predicate ascribes a belief to that collective. In English, in the case of “We believe that p” the verb is plural, given, perhaps, the overt constitution of the subject by a plurality of persons. In the case of “The union believes that p” and so on, the verb is singular. It can be argued that it is neither necessary nor sufficient for all (or most) of the members of the collective to believe that p. In other words, for these cases fulfillment of what I shall call the distributive condition is neither necessary nor sufficient. I have argued this at length elsewhere, and the interested reader is urged to consult the more extended discussions.16 A brief account of some main points follows With respect to the insufficiency of the distributive condition, a central argument is this. When Roz and Mark understand that “We believe that p” the following is true. If she goes on to say to him, without preamble, that p is not the case, he will judge her as having offended against him in some way. She will understand the appropriateness of this reaction. It is hard to argue that satisfaction of the distributive condition can ground these understandings. The same goes for certain more complex conditions I have discussed elsewhere.17 The necessity of the distributive condition is also open to doubt. The following 15 See Anthony Quinton, “Social Objects”, Proceedings of the Aristotelian Society, 76 (1975): 1-27. 16 See, in particular, Gilbert, Social Facts, Chapter 5; also “Modeling”, “Remarks”. 17 See Gilbert, Social Facts, Chapter 5.

42

Margaret Gilbert

scenario seems possible. Roz personally believes that it is never justified for one country to take up arms against another. When Mark asserts the justifiability of a defensive war, he speaks very forcefully. Rather than argue, Roz decides to agree with him. So she says “Yes, indeed”. It seems that now either of them could properly make the collective belief statement “We believe that a defensive war is justifiable.” And this is true even if Mark was for some reason asserting the opposite of his personal view when he spoke. In other words, it may be true, for Roz and Mark, that “We believe that p” though neither of them personally believes that p. Evidently, the distributive condition is not a necessary condition of the truth of the collective belief statement “We believe that p”. I have developed an account of the referent of collective belief statements that does not require satisfaction of the distributive condition. To this I now turn. 2.3 The plural subject account of collective belief According to my account, and in terms to be further explained: The members of a population, P, collectively believe that p if and only if they are jointly committed to believe that p as a body.18 As presaged earlier I shall, for present purposes refer in what follows to the phenomenon that this account describes – a joint commitment to believe that p as a body – as collective belief*. In other words, by stipulative definition, the members of a population P collectively believe* that p if and only if they are jointly committed to believe as a body that p. 2.3.1 Joint commitment A key element in this account is joint commitment. I have explained what this is at length elsewhere.19 For present purposes a brief account should suffice. It is useful first to introduce the conception of a personal commitment. If Maureen decides to do something, she is now personally committed to do that thing. That is, her commitment is such that she has unilaterally brought it into 18 My earlier formulations of the account were somewhat different. See Gilbert, Living, Introduction, and also the text here, below. 19 See, for instance, Gilbert, Living, Introduction; Gilbert, “Aspects”. I first invoked the relevant idea in “Modeling”.

Belief and Acceptance as Features of Groups

43

being and she can unilaterally get rid of it by changing her mind. In short, she is the sole author, and has sole authority over, the commitment. A joint commitment can be characterized negatively as follows: it is not a sum or aggregate of personal commitments of the parties. Rather, it is a commitment of the parties, of Jim and Lisa, say. The parties are committed through it, but these ensuing commitments are not personal commitments in the sense characterized above. Absent special background understandings, the joint commitment can only be brought into being or rescinded by the parties acting together. Nor does any one bring a part of it into existence. Only when all have acted does it, or any “part” of it exist. All “parts” then come into being simultaneously. 2.3.2 The general form of joint commitment We can think of the general form of a joint commitment as follows: Persons xi, xii, … xn are jointly committed to @ as a body. Substitution instances for “@” will be terms for psychological states, such as “believe that p”, “feel grief over X”, “accept goal G”.20 The general idea of a joint commitment to @ “as a body” is this. The parties take it upon themselves to constitute, as far as is possible, a single body that @s. The term “body” is not sacrosanct. It could be replaced by “person” or “entity”. I say more about what it is jointly to commit to believe as a body shortly. I have given the following technical meaning to the phrase “plural subject”: By stipulative definition, persons xi, xii, … xn form a plural subject of @ing if and only if they are jointly committed to @ as a body. Hence I refer to my account of the referent of collective belief statements as the plural subject account.

20 On collective emotions see, for instance, Margaret Gilbert, “Collective Remorse”, in Sociality, 123-140. On goals and intentions see “Walking Together: A Paradigmatic Social Phenomenon”, in Peter French, Theodore Uehling, and Howard Wettstein, eds.,Midwest Studies in Philosophy, 15 (Notre Dame, Indiana: University of Notre Dame Press, 1990):1-14, reprinted in Living Together, and Margaret Gilbert, “What is it for Us to Intend?” in Sociality, 14-36.

44

Margaret Gilbert

2.3.3 Forming joint commitments How are joint commitments formed? Roughly, each of the parties must express – in conditions of common knowledge –his or her readiness to be jointly committed with the others in the relevant way21 There are different cases `that need to be distinguished, but for present purposes that brief description may suffice. An important aspect of all cases has to do with the way a given joint commitment is understood. The point may be illustrated by reference to the following two cases.22 Case 1: Ellen and Sylvie are fellow-students who have been conversing. Each has expressed her personal readiness to be jointly committed with the other to believe as a body that there needs to be a cafe on campus. Case 2: Phil, Cass, Ben and Ted are the members of the campus improvement committee. In a session of the committee they discuss what new amenities might be needed on campus. Each then expresses his readiness to be committed with the other members of the committee to believe as a body that there needs to be a cafe on campus. These cases appear to differ in at least the following way. In Case 1, the resulting joint commitment holds between certain individuals without qualification. That is, it holds between Sylvie and Ellen, period. In Case 2, the commitment holds between the members of the improvement committee as such. This has consequences for the correct description of the subject of a given collective belief statement. One may baldly say of Ellen and Sylvie that they collectively believe that a new cafe is needed. This would not be appropriate for the parties in Case 2. There one might baldly say that the improvement committee believes a new cafe is needed. In case 2 there may be a temptation to say something like this: Phil, Cass, Ben, and Ted believe that a new cafe is needed, in their capacity as members of the committee. This mode of description may tend to suggest that there is a special mode of believing, “believing in a certain capacity”. That there is such a thing, however, is unclear.23 If one adopts this mode of description I suggest 21 On “common knowledge” see David K. Lewis, Convention: A Philosophical Study, (Cambridge, MA.:Harvard University Press, 1969); Stephen Schiffer, Meaning (Oxford: Oxford University Press, 1972); Gilbert, Social Facts. 22 The second case does not involve a basic case of joint commitment. Informal analogues of Case 2 can be found. See “Remarks”. 23 On this see Gerald A. Cohen, “Beliefs and Roles”, Proceedings of the Aristotelian Society (1966-70): 53-66. Cohen argues that “whatever produces his belief, it is his belief, rather than his only qua stockbroker …” 63.

Belief and Acceptance as Features of Groups

45

that it should always be seen as a cumbersome way of ascribing the belief to the collective in question. One can highlight the contrast between the cases thus. Suppose Phil and Cass – two members of the improvement committee – are friends, and are meeting for lunch near the campus. Implying a negative response to the question, Phil might ask Cass: “Do we really need a cafe on campus?” He poses this question to Cass as an individual, rather than as a member of the improvement committee. The committee is not in session and they are meeting as friends. One might say that a new subject of a new collective belief is being proposed here. The question is: will Cass jointly commit with Phil to believe as a body that a cafe is not needed, contrary to the belief of the committee of which they are both members? Were the same thing to transpire between the parties to Case 1, after they have settled on collectively believing that a cafe is needed, it would have a special character. Ellen and Sylvie collectively believe that p, and Sylvie, say, is proposing that the question be re-opened between them.24 2.3.4 The joint commitment in collective belief* How are people to act so as to constitute, as far as possible, a body that believes that p? As I understand it, they are not required personally to believe that p, nor to try personally to believe that p. Nor are they required to act as if they personally believed that p, whether or not they do. Rather, they are to act as would any one of several mouthpieces of the body in question, thus uttering its beliefs, as opposed to the beliefs of any of its members, including the utterer. This understanding accords with the following observation regarding everyday collective belief statements. While acknowledging that she and her friends collectively believe a certain thing, Rose might add, without a sense of fault, that she, personally, does not believe it. Here she certainly does not act as if she has the belief herself. On the contrary she avows, in the most emphatic way, that she does not have it. 25 2.3.5 A “historical” note In the above account of collective belief* the parties are jointly committed to believe that p as a body. In my earliest formulations of the account, I wrote rather of the parties jointly accepting that p, implicitly parsing this as 24 For related discussion see Gilbert, Living, 353-354; 359n15. 25 See Gilbert, Social Facts, and elsewhere.

46

Margaret Gilbert

jointly committing to accept as a body that p.26 In the present context someone might ask why I originally made use of the term “acceptance”. Was there a conscious choice of this term as more appropriate than “belief”? To the best of my knowledge the answer is negative. There is, I think, a non-technical sense of “accepts that” such that it is more or less synonymous for “believes that”. Without evidence to the contrary, I assume that I was using “accept” in that sense, when talking of joint “acceptance” and so on. There is one reason to prefer the phrase “joint acceptance” to “joint belief” that might have been influential. It has nothing to do with meaning. The former is more euphonious, the latter involving a babble of consonants between the terms. 2.3.6 No circularity According to the plural subject account, a collective belief that p requires a joint commitment to believe that p as a body. The concern may be raised: is there a circularity here? The issue would seem to be this. Is speaking of “believing that p as a body”, or “constituting, as far as is possible, a single body that believes that p” not simply another way of speaking of “believing collectively”? I would say not. Consider the case of an individual human being. If Lois, say, believes that p, she constitutes a single body that believes that p. Which is to say this: a belief that p is correctly ascribable to her. One can understand what it is, then, for a single body to believe that p, without understanding what it is for a collective to believe anything. Suppose now that Tom and Alice Jones are jointly committed to constituting as far as is possible a single body that believes that p. In order to carry out this commitment, they need only to understand the idea of a single body that believes that p. Being jointly committed to constitute as far as is possible a single body that believes that p, they are committed to find ways to bring that about to the extent that it can be brought about. According to the plural subject account, they require just such a commitment in order to constitute a collective body that believes that p.

26 See Gilbert, Social Facts, “Modeling”, Living, Introduction, and elsewhere.

Belief and Acceptance as Features of Groups

47

3 The Question of Method Most theorists of belief focus on the beliefs of individual human beings, or, to put it briefly, on the individual case. While their views of the nature of belief may to some extent be based on intuitive judgments about belief as such, these intuitions, however, are standardly tested out on thought experiments regarding hypothesized cases. For instance, one might ask, “Would I judge that someone believed something if I knew he had no evidence for it whatsoever?” Theorists have focused on the individual case in this respect. In other words, the “someone” in question is standardly an individual rather than a group. It is not surprising, then, that rejectionists draw on an understanding of belief that derives from consideration of the individual case.27 Rejectionism’s reliance on such an understanding, however, is open to the objection that it begs an important methodological question. Why do most theorists of belief focus on the individual case alone? There are several possibilities. One is that a given theorist’s interest is in how human beings, as opposed to human groups, function. The theorist is interested precisely in the beliefs of individuals. Another possibility is that the theorist has not considered that there are – or are commonly said to be – collective cases. Ignoring the collective case is not then a principled matter. Yet another possibility is that the theorist is aware of how common collective belief statements are, but for some reason assumes at the outset that such statements cannot be taken at face value. “Of course there are no genuinely collective beliefs” the implicit argument may run. “So – assuming that they are not deluded – those who make these statements cannot be understood as claiming that there are.” Collective belief statements may then be accounted for in some way – as “facons de parler”, perhaps – but they are not taken seriously in constructing an account of belief. It might be argued that a well-founded account of belief in general will take collective belief statements seriously. The argument I have in mind runs somewhat as follows. 27 Thus Meijers, “Believing” makes use of a list of properties that beliefs have “in the individual case”, using this list as a touchstone for deciding whether or not collective beliefs* are beliefs. See Meijers,63. Meijers cites Pascal Engel, “Believing, Accepting, and Holding True”, Philosophical Explorations I (1998): 140-151, 143-4 and 146-7. Engel, in turn, draws on the work of Michael Bratman in “ Belief and Acceptance in a Context”, Mind (1993).

48

Margaret Gilbert

An important criterion of adequacy for a philosophical account that is not merely stipulative is a degree of responsiveness to the use of the relevant terms. In this case, the central terms are “belief” and its cognates (believes, believer, and so on). An adequate account of belief must, in particular, accommodate all phenomena that are referred to without question as cases of “belief”, unless some good reason can be given for excluding some of these phenomena. Thus, the argument concludes, unless some good reason can be given for excluding collective belief statements, it would be better to consider their referents in formulating an account of belief, and not to pronounce as to the nature of belief as a result of considering only individual belief statements and their referents. A similar argument can be made for the whole range of attributes that are ascribed both to individuals human beings and to collectives.28 One could, of course, decide to focus on the individual case, asking what it is for an individual human being to believe something. Suppose one called the resulting account an account of individual belief. It is not clear that it would then be legitimate to argue that, if the so-called beliefs of collectives lack features of individual belief , then such so-called beliefs are not in fact beliefs. The proper conclusion would surely be that a collective’s believing something differs in important ways from an individual’s believing that thing. A famous example of the more inclusive procedure is that of Plato in the Republic. Plato’s guiding question was the nature of an individual human being’s justice, and whether or not it was of benefit to its possessor. Shortly after raising this question, however, he turns to the question of a just society. His most salient argument for doing so is probably the least persuasive. He says that since a society is larger than an individual, it should be easier to make out the lineaments of justice in a society than in an individual. A related, deeper point he makes is this. We ascribe justice to both individuals and societies. It seems, then, that there is a general idea of justice that applies to both. We can investigate justice in a society independently of any considerations of justice in the individual. Suppose we do so, and come up with a certain account of it. We then investigate justice in an individual human being, and find the results tally nicely. In other words, the account of justice in the society corresponds in a 28 I do not mean to imply that ascriptions to individual humans and to collectives exhaust the field. For present purposes, however, these are the ascriptions at issue. Another case is ascriptions of belief to non-human animals. On the whole there is more receptivity among epistemological theorists to discussing the animal case than to discussing the collective case. See, for instance, Bernard Williams, “Deciding to believe” in his Problems of the Self (Cambridge: Cambridge University Press, 1970): 136-151.

Belief and Acceptance as Features of Groups

49

salient fashion to the account of justice in the individual. These accounts mutually reinforce each other. Were justice in the individual not importantly analogous to justice in society, according to our accounts of them, we would have reason to doubt at least one of these accounts. Plato’s accounts of justice in the individual and in society are indeed analogous. At the same time, they respect obvious differences between societies and individuals. The psyche of a human individual, in Plato’s scheme, has its own elements and ideal structure. A society too, has its own elements and ideal structure. In both cases there are just three such elements, and the ideal structures have the same shape. Obvious differences are not ignored. The elements of a society are themselves individuals, albeit individuals in which one or another of the central psychic elements tends to hold sway. With respect to a society there is room for discussion of such matters as censorship of the arts, or the ideal marriage institution. There is no room for such discussion in the individual case. There is a case, then, for taking collective belief statements seriously as part of an enquiry into the nature of belief itself. One consequence of doing so would be the following. Instead of arguing, where appropriate, that “since individual belief has feature F, and so-called collective belief does not, so-called collective belief cannot be genuine belief”, we may argue “since so-called collective belief does not have feature F, though individual belief does, feature F cannot be a necessary feature of belief as such.” Perhaps there are good reasons for privileging so-called individual belief in the search for an adequate account of belief. I shall not pursue that question here. What I propose is the following. First, this question requires more attention than it has to date been accorded. Second, unless and until a satisfactory answer is available, an account of belief that fails to take collective belief statements seriously cannot be used in support of rejectionism. The broader method envisaged in this section itself suggests the following question, among others. What do the phenomena picked out by collective belief statements, on the one hand, and ascriptions of beliefs to individual humans, on the other, have in common? Much of what follows is germane to this issue.29 29 Gilbert, Social Facts, Chapter 5, 313, briefly notes on some commonalities between the individual and collective cases. Austen Clark argues for parallelisms between the individual and the collective cases of beliefs and desires, with particular reference to Dennett’s analyses (“Beliefs and Desires Incorporated”, Journal of Philosophy (1994), 404-425).

50

Margaret Gilbert

4 Taking collective belief statements seriously Suppose we take collective belief statements seriously, that is, we take them to be ascriptions precisely of belief. What happens to the various assumptions epistemologists and others make about belief, based on their understanding of the individual case? Do these assumptions have to be changed? Assuming the plural subject account of the referent of collective belief statements, do collective beliefs* have the relevant properties, or not? Rejectionists argue, in effect, for particular disanalogies between belief in the individual case and collective belief*. They go on to conclude that collective belief* is not belief. The inference may or may not be defensible on methodological grounds. Setting that question aside, are the rejectionist claims about particular disanalogies correct? In the next two sections I scrutinize two central rejectionist claims. Both are of the form: beliefs as such have feature F, and collective beliefs* lack F. The supposed features of all beliefs that I consider, though often taken to be central, do not exhaust the features that have been ascribed to beliefs as such by those focusing on the individual case. The points made in discussion indicate how other features may be approached. It has to be said that there is continuing controversy among those epistemologists who, focusing on the individual case, have come up with characterizations of the nature of belief. Indeed, the status of the very features to which rejectionists appeal has been questioned. My strategy here will hopefully sidestep such issues. I start by setting out the relevant rejectionist claim to the effect that beliefs as such have a particular feature F. I attempt to arrive at an interpretation of the nature of F such that the rejectionist claim appears to be relatively plausible on the face of it. I then explore whether or not collective beliefs* lack F, as the rejectionist claims. If, contrary to the rejectionist, collective beliefs* have F, then rejectionism will not follow from the premise that beliefs as such have F. Rejectionism may also be forestalled as follows. It may be possible to show that insofar as it is unclear whether collective beliefs* have F, it is equally unclear whether individual beliefs do, and for similar reasons. In other words, it may be possible to show that the claimed disanalogy does not exist. My discussion involves counters to rejectionism of both these kinds.

Belief and Acceptance as Features of Groups

51

5 Is truth the “aim” of collective belief*? Belief, it is often said, “aims at truth”. This phrase may have been introduced by Bernard Williams in his well-known essay, “Deciding to Believe”.30 To say that belief “aims at” something may suggest that believing is a form of goaldirected activity. 31Williams, however, does not obviously suppose this. In spelling out what he means, Williams emphasizes three things. First, “truth and falsehood are a dimension of the assessment of beliefs”. Second, to believe that p is to believe that p is true. Third, to say, “I believe that p” “conveys the message that p is the case”.32 Do collective beliefs* aim at truth in something like Williams’ sense? Anthonie Meijers, for one, argues that they do not.33 He himself summarizes the idea that belief in general aims at truth as follows: “beliefs represent the world as it is; they have a mind-to-world direction of fit”.34 He goes on to make several points in support of his claim that collective beliefs* do not aim at truth. I now consider these one by one. 30 Williams, “Deciding”. 31 Cf. Peter Railton, “Scientific Objectivity and the Aims of Belief”, in Pascal Engel, ed., Believing and Accepting (Dordrecht: Kluwer Academic Publishers, 2000) 179-208. 32 Williams, “Deciding”, 137. Williams says “This trio of points constitutes the [feature of belief] which I vaguely summed up by saying ‘beliefs aim at truth’”. 33 Meijers, “Believing”, 64-66. Meijers is operating with an account of so-called collective beliefs that he calls “contractualism”. This is very close to my plural subject account, and I consider his arguments in relation to that. On the “contractualism” inherent in the appeal to plural subjects see, Gilbert, Social Facts; also Margaret Gilbert, “Reconsidering the ‘Actual Contract’ Theory of Political Obligation”, Ethics 109, (1999):236-260, reprinted in Sociality.] 34 Meijers, “Believing”, 64. The claim that beliefs have a “mind-to-world direction of fit” is common in epistemology, along with the idea that desires, in contrast, have a world-tomind direction of fit. Others have made what look like closely related points, e.g. Stalnaker, Inquiry, 80, “Correct beliefs are beliefs the contents of which are true. A correct desire or hope, however, is not one that will in fact be satisfied.” Note that what Stalnaker says of beliefs here may not be said of acceptance in the sense at issue here. At least, acceptances are not appropriately evaluated by reference to their truth – though they are, of course, truth-evaluable – but rather by reference to the accepter’s purposes. Williams explanation of his first point tends to obscure the distinction between being appropriately evaluated in terms of something and being evaluable in terms of that thing. (Compare the irrelevance of a mathematician’s weight to the evaluation of excellence as a mathematician.) After stating that “truth and falsehood are a dimension of the assessment of beliefs” he goes on “when someone believes something, then he believes something which can be assessed as true or false”. I take it that he means to claim not only that beliefs can be so assessed but also are appropriately so assessed.

52

Margaret Gilbert

5.1 Are collective beliefs* necessarily “context-bound”? First, Meijers claims that collective beliefs*, unlike individual beliefs, are “context-bound”. He writes: “we as a group believe[*] that p… always given a particular situation, role, or point of view. Genuine beliefs, on the other hand, are not context-bound. A person believes that p regardless of the context she is in” (Meijers, 64). One might, I think, restate the point in terms of a contrast with acceptance. Suppose one accepts that p for the purposes of some argument. This is what one sees oneself as doing. One then explicitly or implicitly ties one’s acceptance that p to a context, in this case, the pursuit of a particular argument. This kind of intentional binding to a specified context, it is claimed, shows that one is dealing with acceptance rather than belief. Belief, it is alleged, is never so contextbound. Collective belief*, on the other hand, is said always to involve such binding to a context, so it is never to be identified as belief. Setting aside the claim about belief, we can consider whether collective beliefs* are necessarily context-bound in the way at issue. Surely they are not. We can quite unproblematically suppose, for instance, that the campus improvement committee believes* that there should be a cafe on campus, – or, for short, believes* that c, – period. That is, we can quite unproblematically suppose that the committee neither explicitly nor implicitly ties its belief* that c to a particular context. It should be emphasized that what the committee believes*, along with any associated qualifications on the context of the belief*, is determined by the content of constitutive joint commitment. In the case in question, the joint commitment is, simply, to believe that c as a body. There is no explicit or implicit qualification as to context. Whether or not any such qualification would make sense appears to depend entirely on whether or not such a qualification ever makes sense for beliefs. The joint commitment is, after all, a commitment precisely to believe that c as a body. I shall not attempt to address that issue here. The committee’s situation in the case of its belief*that c appears to be exactly parallel to that of an individual person who believes that c. Suppose that Pam presently believes that c, period. This does not mean that in all – or indeed, in any – future contexts she will continue to believe that c. She may come to doubt that c or simply cease to believe it, for a variety of reasons. It does mean that she cannot later consistently come to believe that it is not the case that c without giving up her belief that c. All this may surely be said, mutatis mutan-

Belief and Acceptance as Features of Groups

53

dis, of the improvement committee’s belief* that c. The committee may, at some time in the future, come to believe* that it is not the case that c. If it has not first given up its belief* that c, it will hold inconsistent beliefs*. I conclude that, contrary to Meijers’s suggestion, collective beliefs* are not always context-bound in the sense at issue. It is unclear, indeed, if they are ever context-bound in that sense. If they sometimes are, that would strongly suggest that belief in general may be sometimes be context-bound also. I cannot enter into that issue here. However it turns out, collective beliefs* certainly cannot be argued to be acceptances, rather than beliefs, on the grounds that they are always context-bound. Before leaving this particular topic, it is worth noting the following. To claim that collective beliefs* are not always, and may perhaps never be, contextbound in the sense at issue, is not to say anything against the point, explained earlier, that the subject of a given collective belief* will depend on the situation or context in which it is formed.35 5.2 Collective beliefs* and evidence Meijers writes “Beliefs ... require that all the relevant evidence be taken into account …36 I am inclined to gloss this as a normative statement, somewhat as follows: the belief that p is open to criticism if it does not reflect all the relevant evidence. Thus, to take an extreme case, if all the relevant evidence points strongly in the direction of not-p, the belief that p is open to criticism for that reason. Meijers goes on (and this is what he sees as his second point in favor of rejectionism): “We may give up our ... collective beliefs[*] for reasons not related to the epistemic evidence we have but, for example, because of prudential concerns.”37 I take it that he means that one is not open to criticism for giving up – or, for that matter, adopting – a collective belief* for reasons that do not relate to the evidence for or against the belief in question, in short, for nonepistemic reasons. Such a belief*, indeed, may fly in the face of the relevant evidence without being open to criticism.38 35 36 37 38

See section 2.3.3 above. Meijers, “Believing”, 64 Meijers, “Believing”, 65 Cf. Wray, “Collective Belief” who claims very clearly that “plural subjects are and should be responsive to different reasons than the reasons individuals are responsive to when they are deciding what to believe” (327). He specifies that – generally – the reasons

54

Margaret Gilbert

In the example with which Meijers illustrates his point, the members of an alliance of states fear the political repercussions at home of a particular bombing campaign, “and consequently the alliance may give up its belief[*]” that the bombing will achieve a certain desirable result.39 What is the import of “consequently” in the statement that “consequently the alliance may give up its belief*”? It could mean “as a causal consequence” or “for this reason”. Meijers is evidently focusing on reasons, the reasons for which the alliance may give up its belief* about the bombing. His claim, then, would seem to be something like this: the alliance may properly – that is, without violating any appropriate standards of assessment –- give up its belief* for the reason that doing so would be politically advantageous to the member states. In that case, he argues, its belief*, whatever it may be called, is not a belief. For one cannot properly give up a belief for purely prudential reasons. The following should be noted, however. Suppose that each member of an alliance of states thinks it will gain an important political advantage if the alliance gives up its belief* that p.40 Each member is, indeed, fearful of the outcome if the collective belief* is not given up. Subsequently, the alliance gives up its belief* that p. That the giving up of the alliance’s belief* follows in the footsteps of these thoughts and fears on the part of its members does not entail that the alliance’s reasons for giving it up had to do with the political advantage of doing so for its members, or with their fears. It does, of course, strongly suggest that the noted beliefs and fears of the members played some role in the process that led to the alliance’s giving up its belief* that p. But that is not enough for them to count as the alliance’s reasons for giving it up. How are we to tell why a collective gives up a certain belief*, or why a collective comes to a given belief*? How do we know what epistemic reasons it takes into account, or what practical reasons, if such reasons are at issue? These are important questions, and it is beyond the scope of this paper to explore them fully. Evidently, one cannot stop with an examination of the reasons, epistemic or practical, that individual members of the group have for trying to create or plural subjects should be responsive to are not “epistemic” but rather “practical” (327). 39 Meijers, “Believing”, 65. 40 In order to keep close to Meijers’s example, I am writing as if that the members of an alliance of states – particular states – think certain things. If that seems bothersome in the present context, the reader is invited to substitute an example in which the members of the relevant group are individual human beings.

Belief and Acceptance as Features of Groups

55

change the group’s belief*. These reasons may certainly be sought and knowledge of them may help to fill important gaps. But they do not speak directly to the issue. 41 Consider the following imaginary scenario. There are three states in a particular alliance. Each is represented by a particular official around the conference table, the officials being Peter, Antoine, and Karl. Previously the alliance had come to believe* that the way to achieve its goal G was to bomb country C, or, for short, it had come to believe* that g. Now each member prefers, for its own reasons, that the alliance’s bombing of C be discontinued. It therefore wants to bring it about that the alliance cease to believe* that g. Karl speaks first, in the name of his own country. He is quite likely to say: “Is bombing C really going to achieve G?” In other words, he is likely to question the truth of g. How might Peter or Antoine react, if Karl asks this question? Like Karl, they both want the bombing stopped, and therefore they both want the alliance to change its mind on whether the bombing will achieve G. Meanwhile, the form of Karl’s question constrains the appropriate responses. Peter might appropriately say “It’s not clear that it is!”. Antoine might add “Hardly!” Peter might push on: “I’d say that bombing isn’t likely to achieve G. Given the people we are up against, it is just as likely to have the opposite effect!” Karl and Antoine might eagerly approve this, thus establishing for the alliance a new collective belief*, the belief* that the bombing is unlikely to achieve G. Any one of them can then reasonably continue: “Then we should stop the bombing …!” It would clearly have been off the point for Peter or Antoine to have said “Well, my country would prefer that we didn’t believe* that g”. Either of them could have said this, but, in the context, it would have had to be taken as an aside. The business of the conversation has already been established. It is to consider the truth or falsity of the alliance’s current belief* that g. The preferences of any member of the alliance are simply irrelevant to that. Given this scenario, on what basis did the alliance give up its belief* that g? It did so for this reason: given the character of the people the alliance is up against, the bombing of C is not – after all – likely to achieve goal G. The alliance discarded its belief* that g, then, on the basis of epistemic as opposed to practical reasons. These epistemic reasons, one might add, were believed* by the alliance to be epistemic reasons for discarding its belief* that g. For, in the course of their 41 Cf. Margaret Gilbert, review of Jon Elster, Ulysses Unbound, Mind (2002, in press).

56

Margaret Gilbert

discussion, the members of the alliance mutually expressed their readiness jointly to commit to believing as a body that the character of the people, etc., was evidence in favor of the claim that the bombing would not achieve G. This was, we may assume, common knowledge. That a group gives up – or takes on – a belief* for epistemic reasons is, I propose, is a very common situation. One could, indeed, predict that this would be so if collective beliefs* are open to criticism if they do not reflect all the relevant evidence. And that they are is, surely, a plausible claim given what collective beliefs* are. A collective belief* is constituted, after all, by a joint commitment to believe something as a body. It is plausible to suppose that it is required of the parties to such a commitment to make some effort to support the giving up or taking on of the relevant collective’s beliefs* with what are collectively believed* to be epistemic reasons.42 Meijers might attempt to counter this by emphasizing that even if the type of situation in my story about the alliance is a common one, it is not the only possible one. Karl might, in principle, have said “Is it in our – the alliance’s – interest to believe* that bombing will achieve G?” He would then have posed a prudential rather than an epistemic question. And perhaps this line of questioning could have resulted in the formation of a collective belief* that accorded with what the alliance believed* was in its interests, without recourse to epistemic reasons. It is should be noted, in response, that individual human beings, also, may question and assess their beliefs from a purely prudential point of view. Thus Joe may wonder if it is in his interests to believe that his wife is unfaithful. He may decide that it is not, since the belief is making him miserable. And perhaps it could come about, in consequence, that he believes the opposite, without recourse to epistemic reasons.43 What conclusion can we draw from Joe’s case? Does it show that the belief of individuals is not always open to criticism on the basis of its relation to the relevant evidence? On the face of it, it does not show this. And this is what seems to be at issue in the passage from Meijers under discussion.44 42 For a detailed exploration of the justification of group beliefs and a “joint account” of group justification see Frederick R. Schmitt, “The Justification of Group Beliefs”, in Frederick R. Schmitt, ed, Socializing Epistemology,(Lanham, MD: Rowman and Littlefield, 1994). 43 On this see also section 6 below. 44 As it is in the passage from Wray quoted above, note 22.

Belief and Acceptance as Features of Groups

57

It is not necessary for present purposes to pursue the question whether Joe’s case shows that belief in the individual case is not always open to criticism on the basis of its relation to the relevant evidence. For suppose, as I am inclined to think, that it does not show this. Then the existence of parallel collective belief* cases does not show that collective beliefs* are not open to criticism on such grounds. Suppose, however, that it does show this. Then the existence of the parallel collective belief* case shows nothing that takes away from the idea that collective beliefs* are any the less beliefs than are the so-called beliefs of individual human beings. I take the claim that belief as such is open to criticism if it does not reflect all the relevant evidence to be quite plausible. It is, I think, relatively weak, not ruling out the possibility of such cases as Joe’s. I cannot, however, attempt to assess it here. Suffice it to say that there appears to be no radical disanalogy between collective belief* and individual belief in relation to this particular matter. 5.3 Would people form collective beliefs* if truth were all that mattered? The joint commitment underlying a collective belief* is, indeed, a commitment. In addition, it allows for claims from any one party against any another for baldly gainsaying what is collectively believed* in the relevant context.45 Meijers argues that, if the truth of a collective belief* were the only thing that mattered, “it would be incomprehensible why there are inter-individual commitments involved in these beliefs …”.46 His concern here seems to be with the motivations of the individuals who participate in the underlying joint commitment. It is unclear, however, what the details of these motivations have to do with the nature of collective beliefs*, once formed. It is worth noting, in this context, that people who are all personally certain that p can collectively believe* that p. Thus a group of mathematicians can collectively believe* that the square on a triangle’s hypotenuse is equal to the squares on the other two sides, while individually being certain of that this is so. In such a case, even if they wish not to be constrained from speaking the truth, the people in question may have no reason for not partici45 For extended discussion of this point see Margaret Gilbert, “Obligation and Joint Commitment”, Utilitas 11 (1999): 143-163, reprinted in Sociality, 50-70. 46 Meijers, “Believing”, 65, see also 68.

58

Margaret Gilbert

pating in the collective belief* that p. There may, meanwhile, be good reasons to do so.47 As I have argued elsewhere, people who, not being certain that p, participate in a collective belief* that p, may indeed find this frustrating, should they later discover that p is not true. This aspect of collective beliefs* can play an inhibitory role in the development of the sciences, among other things.48 Nonetheless, the balance of reasons may justify participating in a potentially false collective belief*. In the case of scientists, one of these reasons may be that participation in a collective truth-seeking enterprise is arguably the best way overall to reach the truth, whatever its hazards and, in particular, the potential of the collective beliefs* that arise for repressing new ideas that may prove to be correct. Indeed, truth may not be the only thing that matters to the individuals who participate in a particular collective belief*. This is a matter of the aims of those individuals. These aims are not strictly relevant to the claim that collective beliefs* aim at. This is, after all, a claim about collective beliefs, not the individuals who make them up. It is, moreover, not really a claim about “aims” but rather about appropriate standards of assessment. 5.4 Collective beliefs* and the beliefs of the participants Meijers argues, finally, that collective beliefs* do not aim at truth because “we as a group may believe[*] a proposition that no-one takes to be true individually” (65). This surely does not rebut the claim that collective beliefs* aim at truth. Suppose the campus improvement committee believes* that a café would improve the campus, or, for short, believes* that c. Suppose, further, that this collective belief* emerged in spite of the fact that no member of the committee personally believed that c, and the members have not personally changed their minds on the matter. What happened was that everyone knew that the Chancellor was wedded to the idea of a campus café, and, in the committee’s discussion, no one dared propose anything contrary to the Chancellor’s view. 47 On reasons for participating in collective beliefs* in general, see Gilbert, “Remarks”, 253, Social Facts, 225, cited and discussed in Wray, “Collective Belief”, 322. 48 For discussion of the potentially unfortunate role of collective beliefs* in the progress of science see Margaret Gilbert, “Collective Belief and Scientific Change”, in Sociality,3749. Wray, “Collective Belief” discusses the role of collective belief* – which he prefers to label “collective acceptance” – in the scientific laboratory, 327-330.

Belief and Acceptance as Features of Groups

59

If one looks at the members of the committee, they do not, indeed, take c to be true. This, however, is not relevant to the stance of the collective they constitute, namely the committee itself. It is a mistake to look at to the personal situation of the members of the committee when what is at issue is the stance of the committee as a whole. How are we to assess the situation of the committee as a whole, when it believes* that c? We would do well to consider what the members express, or might well express, when acting in light of the joint commitment to believe that c as a body. They are then acting, in effect, as the mouthpieces of the whole their commitment tells them to constitute, as far as is possible. If one looks at these expressions as a clue to the committee’s stance, does it seem to take the proposition that c to be true? Well, one might well find expressions of the claim that c. One might well find expressions of the claim that it is true that c. And one might well find expressions of the form “We believe that c”, which, according to Williams, would convey the message that p is the case, at least as far as we are concerned. Presumably the same can be said, mutatis mutandis, of an individual who believes that c. If so, with respect to “taking p to be true”, it is not clear that there is any distinction to be made between the case of a collective that believes* that p, and an individual human being who believes that p. 5.5 Collective beliefs* and truth I have considered the situation of collective belief* in relation to Williams’ criteria for “aiming at truth”, and in relation to several related points from Meijers. The conclusion that suggests itself is that collective belief* holds much the same relation to truth as the beliefs of an individual do.

6 Belief and the will 6.1 “Involuntarism” It is often claimed that belief cannot be willed. That is, I cannot bring a belief of mine into being by an act of will, or not directly.49 I shall call this claim 49 See, for instance, Williams, “Deciding”, 147f. Meijers, “Believing and Accepting”, 64; Engel, 1998, 143; Tuomela, “Belief versus Acceptance”, 86; Wray, 2002. Sometimes the

60

Margaret Gilbert

“involuntarism”. I shall not here examine the reasons people have adduced in favor of involuntarism.50 Nor shall I attempt to assess it in any detail. Involuntarism has provoked the following rejectionist argument: collective belief* is necessarily willed, hence it is not belief.51 I shall question the claim about collective belief on which this argument is premised. 6.2 Collective beliefs* need not be willed by the collective It is crucial to understand that collective beliefs* do not require to be willed by the relevant collective. They are, indeed, constituted by what could be referred to as a collective will, but that is a different matter. I have in mind, of course, that a joint commitment might be referred to as constituting a collective will. To say that there is a sense in which a collective belief* involves a collective will, however, is not to say that a collective belief is necessarily willed into existence by the relevant collective. Compare the case of an individual’s belief. One may suppose that this, also, to some extent is constituted by that individual’s will. There may be something passive about an individual’s beliefs, insofar they are responsive to something outside the control of the believer, in particular, to evidence for the beliefs. Arguably, however, there is always something active about them also. There may also be a need for judgment, for instance, where judging is an active rather than a passive matter. If performing an action, such as raising one’s arm, is completely active, and having a sensation, such as feeling pain, is completely passive, then believing would seem to lie somewhere in the middle. To say that an individual’s belief is in part constituted by an exercise of that individual’s will is not to say that the belief – constitutive judgment and all – is necessarily willed into existence by that individual. Indeed, it could be true that, in some sense, a given individual cannot believe at will, though believing necessarily involves that individual’s will. The same goes for the collective case. Meijers says that the main reason for thinking that involuntarism supports rejectionism is that collective beliefs*: “… require some sort of voluntary asclaim is put in slightly different terms. The details will not matter for present purposes. 50 On this see, for instance, Williams, “Deciding”, 147f. Williams suggests that “… I cannot bring it about, just like that, that I believe something. Why is this? One reason is connected with the characteristic of beliefs that they aim at truth …” (148). 51 See, for instance, Meijers, “Believing”, 64, section (1), discussed in the text below. See also Wray: “the goals of a plural subject determine” what its members collectively believe*, “Collective Belief”, 7.

Belief and Acceptance as Features of Groups

61

sent, agreement, or decision by the members of the group for the belief”.52 Again: “members of a group … express their commitment to certain beliefs in their practices …”53 The focus here is, evidently, on the individual members of the group and the involvement of their wills in the production of a collective belief*. Consider the case of Sam, who helps to create a collective belief* in his department that costs must be cut. On the one hand, his input to the creation of this belief may not be fully voluntary. Conscious that his boss would disapprove of the group’s thinking otherwise, he feels under great pressure to propose that costs must be cut.54 On the other hand, in setting out to help create the collective belief*, he expresses a certain state of his will. He expresses his readiness to be jointly committed with the others to believe as a body that costs must be cut. In order for the department to believe* that costs must be cut, the other members must express themselves likewise, in conditions of common knowledge. In the sense exemplified in this case, then, the wills of the individual group members are involved in the production of a collective belief*. This does not show however, that collective beliefs* must be willed by the collective in question and or even that they can be so willed, contrary to involuntarism. As far as the story goes, the collective belief* in question was not created by the will of the collective whose belief it is. So, contrary to Meijers suggestion, there is no confrontation with involuntarism here.55 I have argued, what seems firm, that a collective belief*, is not necessarily willed by the relevant collective. According to Meijers, acceptance, as opposed to belief, is “voluntary. To accept that p is to make the decision” to use p as a premise in one’s deliberation.56 Contrary to Meijers proposal, then, collective belief* is not a case of acceptance so characterized. 6.3 Can collective belief* be willed by the collective? Someone might raise the question whether a collective belief* can be willed into existence by a given collective. More precisely, can a collective belief* arise directly as a result of the collective’s willing it into existence? 52 Meijers, “Believing”, 64. 53 Meijers, “Believing”, 64. 54 On the possibility of entering joint commitments in coercive circumstances, see Margaret Gilbert, “Agreements, Coercion, and Obligation”, Ethics 103(1993): 679-706. 55 See also Wray, “Collective Belief”, 7, cited earlier. 56 Meijers, “Believing”, 94.

62

Margaret Gilbert

It may be proposed, along lines that Bernard Williams has traced out for the individual case, that insofar as belief aims at truth, there is something rebarbative about the idea of adopting as a collective goal what is, in effect, a joint commitment to believe as a body that p.57 The idea in question, more specifically, is that these people would have as their goal their constituting as far as possible, a body that believed that p, without that body’s having any inkling of the truth of p. There may, then, be a hedge of sorts around collective beliefs* just as there may be such a hedge around individual beliefs, tending to prevent people from wantonly creating or indeed trying to create such beliefs just because they want to have them, or because it would be in their interests to have them. Let us suppose that, in spite of these considerations, and perhaps per impossibile, Fran and her friend Trudy explicitly adopt as their collective goal their collectively believing* that their future is bright, without concern for whether or not their future actually is bright. It would be good for them, they think, to believe* this. Thus they become the plural subject of the goal: our believing* that our future is bright. This situation is interestingly different from that of familiar collective goals where, once the goal is set, all that needs to be done is that the individuals in question personally act in light of the relevant joint commitment so that the goal be achieved. An example might be the goal of walking along the beach side by side for a while; once the parties have jointly committed to adopting this goal as a body, all they have to do is perform relevant actions in light of the joint commitment in question. Thus each must simply walk along the beach monitoring her spatial relationship to the others, if necessary urging the other to walk faster, and so on.58 In the present case, by hypothesis, a new joint commitment must be made. It is true that the parties are creating this commitment in light of another such commitment. Nonetheless, the group’s wanting to believe* that p does not lead to its believing* that p with the same immediacy with which it may lead to the beach walk. One might put the point this way: a group cannot directly bring its beliefs* into being by an act of its own will. In other words, one can argue that this case does not present a counterexample to involuntarism.

57 See Williams, “Deciding”, 147f. 58 For detailed discussion of the case of walking together see Margaret Gilbert, “Walking Together”; also “Acting Together”, forthcoming, Grazer Philosophische Studien (2002).

Belief and Acceptance as Features of Groups

63

5.4 Collective belief* and involuntarism at the collective level In the last two sections, I have argued that the will of the collective – as would be found in a collective goal or a collective decision – is not necessarily involved in the production of a collective belief*. So it is wrong to consider collective belief* a form of acceptance, if that involves something that is necessarily willed by the subject of the acceptance. To see this one needs clearly to distinguish between the individual wills of the members of a collective and the will of the collective itself. That is an important distinction that needs to be made, irrespective of the issue of rejectionism. I have also suggested that there may be something rebarbative about a group’s setting its possession of a certain collective belief as a goal, just like that, just as there may be something rebarbative in the individual case. The two cases would seem to be on a par in relation to this matter, which is the main point for present purposes. Finally, I argued that a collective belief* cannot be directly willed by the collective, even if the collective does, perhaps per impossibile, form the goal of possessing that belief*. There is no reason in particular for me to want this to be true. However, if it is true, it goes against the view that collective beliefs* are not beliefs because – as it is often supposed – genuine beliefs cannot be directly willed into existence by the believer. 6.5 Collective belief* and involuntarism at the individual level I said earlier that the joint commitment involved in collective belief* does not, as I understand it, require the participating individuals personally to believe something, nor even to try personally to believe it. Hence it does not involve the denial of involuntarism at the individual level. It may be worth elaborating upon this point here. I have characterized the joint commitment in question as a commitment to constitute, as far as possible, a single body that believes that p. It may be observed, first, that insofar as a given collective belief* that p exists once the relevant joint commitment does, the existence of a collective belief* cannot not depend on the participants personally believing that p. Even if that were what the commitment required of them, the commitment in and of itself suffices for the collective belief*, whether or not it is conformed to. 59 59 The question of how lack of conformity impacts on the existence of a joint commitment

64

Margaret Gilbert

However that may be, the commitment, as I understand it, does not require that anyone do anything impossible. Suppose Judy is jointly committed with certain others to constitute, as far as is possible, a single body that believes that p. Suppose, further, that involuntarism about belief is true. Then it is not possible for Judy to believe that p at will. It follows that she is not required to believe that p at will. It is often suggested that there are things one can do in order to get oneself to believe something, even if the means of creating the belief are indirect. One can avoid places where contrary evidence is likely to be found, and so on. I said earlier that a joint commitment to believe that p as a body does not require that one do what one can personally to believe that p. I shall assume, for present purposes, that this is not because such an attempt is impossible. Why would a joint commitment to believe that p as a body not require the individual participants personally to believe that p or to attempt personally to believe that p? Presumably this is why: the focus of the commitment is on the creation of a body that believes, not on the components of the body in question, or their personal states of mind. Though this is so, it is presumably easier to conform to a joint commitment to believe that p as a body if one does oneself believe that p. It will be easier, and less internally stressful, to give voice to the view that p when it is one’s personal view. In this sense collective beliefs* are likely to influence what people think: they have what Durkheim called “coercive power”. They are, in short, extremely consequential.60

7 Collective belief* and acceptance 7.1 Collective belief* and individual acceptance If the participants in collective belief* are not required to believe that p, are they required to accept that p in some sense in which acceptance does not entail belief? In other words, when the collective belief* that p is “in full flower”, with the relevant individuals all complying with the joint commitment, are we confronted with a population of individuals all of whom accept that p, in relevant contexts, as opposed to believing that p? cannot be entered here. See Gilbert, Living, Chapter 16, for some discussion. 60 See Durkhiem, Rules, Chapter 1; Gilbert, Social Facts, Chapter 5, and elsewhere.

Belief and Acceptance as Features of Groups

65

Note that whatever the answer to this question, it does not appear to impact on the question whether collective belief* is belief or, rather, acceptance. Suppose, for the sake of argument, that in order to participate in a fully flowering collective belief*, the members of a population must in some appropriate sense accept that p, in relevant contexts, though they need not believe it. This would not seem to show that collective belief* was acceptance, in the relevant sense, and not belief. To say that it did would seem to be confusing what is going on at the level of the individual with what is going on at the collective level. Does collective belief* , in full flower, involve a set of acceptances on the part of the participants? That depends, of course, on precisely what acceptance is, and authors vary in their detailed descriptions of acceptance. For Meijers, following Engel and others, acceptance is a propositional attitude which, unlike belief, has at least the following properties. It is the result of decision, aims at utility rather than truth, need not, even ideally, be shaped by evidence, and is context-dependent. The paradigm case of “accepting for the sake of the argument” would appear to have all of these properties. Now it is not clear that a collective belief* that p involves the participants, personally, in having any particular propositional attitude to p. More precisely, constituting as far as is possible a body that believes that p does not appear to require any such attitudes on behalf of the participants. Their acts and expressions are, in a sense, representational; they act and utter expressions as representatives of the collective. It does not seem that this requires them to have any particular attitude towards the proposition that p themselves. 7.2 Collective belief* as belief as opposed to acceptance The above considerations strongly suggest that there is no particular reason to say that collective belief* is acceptance as opposed to belief. For one thing, it seems that there need not be acceptance of the relevant proposition p at the individual level. That is, there need not be personal acceptance of that proposition, any more than there need be personal belief that p at that level. For another, there is no obvious reason to say that there is acceptance rather than belief at the collective level. Why, though, say that collective belief* is belief? For one thing, it is, I am assuming, the referent of everyday collective belief statements. For another, it involves a joint commitment to believe something as a body, not to accept something as a body. And a joint commitment to accept a proposition, in con-

66

Margaret Gilbert

trast, is perfectly possible. I discuss this shortly. If one has to choose between saying that collective belief* is belief and saying that it is acceptance, I propose that one should say it is belief. 7.3 Collective acceptance Accepting a proposition, where this does not involve believing it, is evidently a useful thing to do. Often one wants to consider the consequences of some state of affairs without investigating whether or not it obtains. This is certainly true in science, it is true in various branches of scholarship, and it is true in everyday life as well. Thus people regularly say things like this: “Suppose she has left already. What would I do then?” Collectives can, surely, accept propositions, in this sense. People make collective acceptance statements just as they make collective belief statements. And I would propose an analogous account of them. Cases are easy to imagine. The campus improvement committee is discussing whether the campus needs a cafe. Cass says, “Look, let’s suppose, for the sake of argument, that Café Sunshine – the only cafe in town – is going to close in the next year.” Phil says, “Okay”, and the others nod in agreement. The committee members have now expressed their readiness to be jointly committed to accept as a body, for the purposes of their discussion, that Café Sunshine is going to close in the next year. And this is – we can assume – common knowledge between them. They have, in other words, now constituted themselves, in their capacity as members of the committee, a plural subject of acceptance, in the sense at issue. It is worth noting that collective beliefs* are unlikely to be brought into existence in an analogous way, that is, with the suggestion “Let’s believe that …” Rather, a proposition is put forward as a candidate for collective belief*: “Nice day!” “Yes, indeed”. It is a proposal for acceptance, I would think, that needs to be specially marked as such. Normally to put a proposition forward in conversation without qualification is an invitation to one’s fellow conversationalists collectively to believe* it. Insofar as there are several viable ideas of acceptance in the individual case, they should all be susceptible of exemplification in the collective case. The whole range of cognitive states possible for individuals would appear to be attributable to groups in relevant circumstances. Groups can surely doubt that p, wonder whether p, be certain that p, consider whether p, and so on. And a distinct plural subject account seems to be available for each

Belief and Acceptance as Features of Groups

67

one of these, the underlying joint commitment having a different content in each case. 61

8 Concluding remarks What is most important about collective belief*? I would say not that it is – or is not – a kind of belief, but rather the following. It is a phenomenon of which one might say, in Durkheimian terms, that it has “coercive power” over individuals. Once again in Durkheimian terms, in the context of collective beliefs* the individual is liable to be pressured by other group members, acting as such, into speaking, acting, and, ultimately, thinking in certain ways. Indeed, the individual is liable eventually personally to believe what the collective believes* – whether the collective is properly said to believe it or not. I have argued for all these things elsewhere.62 The issue discussed in this paper, then, does not impinge on the importance of collective beliefs* under whatever label, nor does it throw their existence into doubt. The present discussion suggests that ordinary language, in referring to them as beliefs has not been shown to be in error. It is not clear that it could be shown to be in error, if, as I have suggested, collective belief statements are part of the data for an account of belief in general. If, for some reason, one refuses to allow this, it appears that, nonetheless, collective beliefs* behave much like beliefs in the individual case, and certainly more like such beliefs than rejectionists have supposed. It is hard to see why the learned should not speak with the vulgar, and go ahead and call them beliefs.63

61 Cf. Gilbert, “Remarks”. 62 See, for instance, Gilbert, Social Facts, Chapter 5; “Durkheim”. 63 Selected sections of this essay were presented to the Department of Philosophy at the University of Connecticut, Storrs, November 2001. I am particularly grateful to Jerome Shaffer for his contribution to the lively discussion. I have in response made more of the distinction between reasons for and causes of belief. Ruth Millikan helpfully emphasized the intrinsic interest of analogies and disanalogies between the individual and the collective cases. I am most grateful also to Donald Baxter, Paul Bloomfield, Simon Evnine , Anne Hiskes, Saul Kripke, Scott Lehmann, Giovanna Pompele, K. Brad Wray and Nick Zangwill, for related discussion. Finally, I thank those authors I have discussed, Anthonie Meijers, Raimo Tuomela, and K. Brad Wray, for provoking me to consider the topic. Though I resist their conclusions, I have found much of merit in their discussions.

68

Margaret Gilbert

References Michael Bratman, “Belief and Acceptance in a Context,” Mind (1993). Gerald A. Cohen, „Beliefs and Roles“, Proceedings of the Aristotelian Society (196670): 53-66 Jonathan Cohen, An Essay on Belief and Acceptance (New York: Oxford University Press, 1992) Austen Clark, “Beliefs and Desires Incorporated”, Journal of Philosophy (1994): 404425. Emile Durkheim, Rules of Sociological Method, ed. S. Lukes (New York: Free Press, 1982). Pascal Engel, “Believing, Accepting, and Holding True”, Philosophical Explorations I (1998): 140-151. —, ed., Believing and Accepting (Dordrecht: Kluwer Academic Publishers, 2000). Margaret Gilbert, “Modeling Collective Belief”, Synthese 73 (1987): 185-204. —, On Social Facts, (Princeton: Princeton University Press, 1989). —, “Walking Together: A Paradigmatic Social Phenomenon”, in Peter French, Theodore Uehling, and Howard Wettstein, eds.,Midwest Studies in Philosophy, 15 (Notre Dame, Indiana: University of Notre Dame Press, 1990):1-14, reprinted in Gilbert, Living Together. —, “Agreements, Coercion, and Obligation”, Ethics 103(1993): 679-706. —, “Durkheim and Social Facts”, in Debating Durkheim, ed. Herminio Martins and Donald Pickering (London: Routledge, 1994), 86-109. —, “Remarks on Collective Belief”, in Socializing Epistemology, ed. Frederick Schmitt (Lanham, MD: Rowman and Littlefield, 1994), 235-256; —, Living Together: Rationality, Sociality, and Obligation (Lanham, MD: Rowman and Littlefield, 1996 —, “What is it for Us to Intend?” , in Raimo Tuomela and Ghita Holmstrom-Hintikka, Contemporary Action Theory, vol. 2, Social Action (Dordrecht; Kluwer Academic Publishers, 1997): pp-pp, reprinted in Gilbert, Sociality, 14-36. —, “Reconsidering the ‘Actual Contract’ Theory of Political Obligation”, Ethics 109, (1999):236-260, reprinted in Gilbert, Sociality. —, “Obligation and Joint Commitment”, Utilitas 11 (1999): 143-163, reprinted in Sociality, 50-70. —, Sociality and Responsibility: New Essays in Plural Subject Theory (Lanham, MD: Rowman and Littlefield, 2000), 37-49. —, “Collective Belief and Scientific Change”, in Gilbert, Sociality,37-49. —, “Collective Remorse”, in Gilbert, Sociality, 123-140 —, “Acting Together”, forthcoming, Grazer Philosophische Studien (2002). —, review of Jon Elster, forthcoming, Ulysses Unbound, Mind (2002, in press). Keith Lehrer, Self-Trust: A Study of Reason, Knowledge and Autonomy (Oxford: Clarendon Press, 1997). Anthonie Meijers, „Believing and Accepting as a Group“, in Belief, Cognition and the

Belief and Acceptance as Features of Groups

69

Will, ed. A. Meijers (Tilburg; Tilburg University Press, 1999), 59-71. Anthony Quinton, „Social Objects“, Proceedings of the Aristotelian Society, 76 (1975): 1-27. David K. Lewis, Convention: A Philosophical Study (Cambridge, MA.:Harvard University Press, 1969). Peter Railton, “Scientific Objectivity and the Aims of Belief”, in Engel, 2000, :179-208. Stephen Schiffer, Meaning (Oxford: Oxford University Press, 1972). Frederick R. Schmitt, “The Justification of Group Beliefs”, in Frederick R. Schmitt, ed, Socializing Epistemology (Lanham, MD: Rowman and Littlefield, 1994): pp.-pp. Robert Stalnaker, Inquiry (Cambridge, MA: MIT Press, 1984). Raimo Tuomela, “Group Beliefs”, Synthese 91 (1992):285-318. —, The Importance of Us (Stanford: Stanford University Press, 1995). —, “Belief versus Acceptance”, in Meijers, ed., Belief, Cognition, 319-333 Bas Van Fraassen, The Scientific Image(Oxford: Clarendon Press, 1980). Bernard Williams, “Deciding to believe”, in his Problems of the Self (Cambridge: Cambridge University Press, 1970): 136-151. K. Brad Wray, “Collective Belief and Acceptance”, Synthese 129 (2001):319-333.

70

Anthonie Meijers

Collective Agents and Cognitive Attitudes Anthonie Meijers

Abstract Propositional attitudes, such as beliefs, desires, and intentions, can be attributed to collective agents. In my paper I focus on cognitive attitudes, and I explore the various types of collective beliefs. I argue that there is a whole spectrum of collective beliefs, and I distinguish between two extremes: the weak opinion poll conception and the strong agreement-based conception. Strong collective beliefs should be understood in terms of the acceptance of a proposition rather than of belief proper. They are not purely epistemic and involve practical considerations. To believe that p collectively in the strong sense is to adopt a policy to use that proposition in the group’s deliberations about future actions.

1 Introduction The issue of cognitive attitudes such as believing, holding true and accepting a view is usually discussed in the context of attitudes held by individual persons. In this article, I shall widen the scope of the discussion by analysing a widespread phenomenon: the attribution of cognitive attitudes to groups or collective agents. Let us consider some real-life examples: (1) The editorial board believes that blind reviewing improves the quality of the journal. (2) Microsoft believes that it should settle the antitrust case. (3) The cabinet believes that genetically modified food is safe and that it should not be forbidden. (4) The British believe that the euro will eventually be introduced in the UK. How are we to analyse such beliefs? Is our talk in terms of collective beliefs only metaphorical, or is it literal, and, if so, in what respect? Groups evidently do not have a mind or consciousness, so the analogy between groups and persons must break down somewhere. But if we attribute beliefs to groups, how do they function in the causal explanation of group behaviour? Are these beliefs purely epistemic, or do they involve a pragmatic aspect as well? Is there a

Collective Agents and Cognitive Attitudes

71

distinction to be made here between believing and accepting a proposition? And is it possible for groups to have a “will to believe”, to use William James’s phrase? The concept of belief cannot be taken for granted, even in the individual case. Several competing conceptions are on offer. For example, the idea of beliefs as “maps by which we steer” (Ramsey, Amstrong, and Dretske) or as “global states or attitudes of persons, manifested by what they think, say and do” (Baker). There are two elements that probably need to be taken into account by any theory of beliefs: their representational and behavioural aspects. On the one hand, we want to say that beliefs represent states of affairs in the world. On the other hand, we want to say that beliefs, together with desires and intentions, cause human actions. And this applies to individual as well as to collective beliefs. Several authors have recently developed accounts of collective beliefs which identify this type of belief as a belief sui generis. Margaret Gilbert and Raimo Tuomela, among others, defend the idea that collective beliefs are not reducible to individual beliefs.1 I am sympathetic to their account. Before discussing some of the arguments in favour of the non-reductionist position, I shall introduce two conceptions of collective beliefs.

2 Two Conceptions of Collective Beliefs The first conception of collective beliefs makes a modest claim. To say that a group believes that p is to say that its members, as individuals, believe that p. This weak conception – which is sometimes called summative – is reductive in that it tries to account for collective beliefs in terms of individual beliefs. It can be formulated as follows: weak conception (1): ‘we believe that p’ = ‘we, the individuals in the group, as individuals, believe that p’ In this conception, not all the individuals in the group need to believe that p. The distribution of the beliefs among the individuals in the group may be dif1 Margaret Gilbert, On Social Facts (Princeton, 1992); Raimo Tuomela, The Importance of Us. A Philosophical Study of Basic Social Notions (Stanford, 1995).

72

Anthonie Meijers

ferent in different situations. Sometimes all the members of the group need to believe that p, and sometimes it is enough that the majority of its members believe that p. Conception (1) typically applies to the fourth example of the British believing that the euro will eventually be introduced in the UK. I shall call this account of collective beliefs the opinion poll conception. Though this conception applies to a number of cases where we attribute beliefs to groups, it is too weak as a general account. There are cases where a more comprehensive conception of collective beliefs is necessary. For example, cases in which the individuals involved need to know that the other members of the group believe that p (and vice versa). One way to specify this additional requirement is in terms of common knowledge or common belief. For a belief to be a collective belief it is necessary, then, that the members of the group believe that the other members believe that they believe (and so on) that p. Such an account can, of course, be refined and improved; I only want to give the general idea here. The opinion poll conception will, when combined with the common knowledge approach, be adequate in many cases. There are, however, cases that require a much stronger conception of collective beliefs. Such a conception will be at the other end of a continuum of approaches to collective cognitive attitudes. The strong conception of collective beliefs that I want to present here as my second conception takes collective beliefs to inhere in social groups, where a group is taken to be different from the simple summation of the individuals belonging to it, be it the total number or the majority. I shall call this conception the agreement-based conception. It can be formulated as follows: strong conception (2:) ‘we believe that p’ = ‘we, the group as a whole, believe that p’ This conception may seem somewhat mysterious in its idea of beliefs that inhere in groups, where a group is taken to be different from the summation of its members. The distinction that we need to make here is between a group as a structured whole and a group as an aggregate. The passengers boarding a train form an aggregate unless they have decided to travel together. Such a decision transforms an aggregate into a structured whole which has a collective intention to travel together. The distinction between a structured whole and an aggregate is a relative one. The same individuals may form an aggregate in some respects, while in others they are a structured whole. For example, the individuals working at the same university form a structured whole from the point

Collective Agents and Cognitive Attitudes

73

of view of their work, while they are an aggregate from a political point of view. As employees they (usually) do not have political views or perform political actions. Although the idea of a group as a structured whole does not involve the rather exotic ideas of a group mind or a group consciousness, it does assume that such a group is different from an aggregate. The third example of the cabinet believing that genetically modified food is safe and should not be forbidden is covered by this second conception of collective beliefs. Here we have a structured group, even institutionalised by law, where the relations between the members of the group are carefully defined. In such a case, we attribute beliefs to the group as a whole provided that the appropriate procedures have been followed (usually a discussion in that group and the joint acceptance of the view). From what I have said it follows that the same group of individuals can have several types of collective beliefs, corresponding to the opinion poll conception, the agreement-based conception, or some conception in between. Cabinet members, for example, in addition to the beliefs which they hold as a structured group, probably also collectively believe (in the opinion poll sense) that Paris is south of Brussels, or that Galileo should not have been sentenced by the Inquisition. The cabinet, however, usually does not have an opinion on these matters.

3 Non-reductionism and the Necessity of a Pluralist Approach An important question is why the opinion poll conception outlined above cannot account for all the relevant cases of collective beliefs, and, in particular, for the cabinet example. Why are we not able to analyse this case in terms of individual beliefs, a certain distribution of these beliefs in the group, and common knowledge about these beliefs? Let me review some of the arguments as to why this cannot be done. A first argument is that collective beliefs in the cabinet case have a binding and engaging effect on the participants. Members of the cabinet qua members are supposed to defend their belief when genetically modified food is being discussed or decisions are being made about food production. They also have the right to claim that the other members do the same: their collective belief creates reciprocal commitments. In addition, they are expected not to change their views on genetically modified food unless some procedure has been followed by the cabinet.

74

Anthonie Meijers

I take it to be evident that the opinion poll conception cannot account for this mutually binding effect. Neither individual beliefs nor common knowledge have these normative consequences. The opinion poll conception applies rather to the case where all the members of the cabinet express their beliefs on genetically modified food individually on television so that all the other members of the cabinet know each other’s beliefs with respect to genetically modified food. This case, however, is different from the one which we are discussing here. In that case we attribute a collective belief to the cabinet as a whole, based on some (explicit or implicit) form of decision making. Such a belief has mutually binding effects for its members, and it motivates future cabinet’s actions. A second and related argument is that commitments to beliefs held collectively are conditional. They depend upon the willingness of the other members of the group to accept the belief as the belief of the group, and to be committed to it. If some members of the cabinet were to change their minds about genetically modified food, this could free the other member from their commitments. In more formal cases, such as the one discussed here, there will be official procedures for releasing members from these commitments. Until these procedures have been followed, however, members are committed to their collective belief. If some members do not live up to these commitments, then the others are no longer committed either. The idea of a conditional commitment as part of what constitutes a collective belief is incomprehensible from the point of view of the opinion poll conception. A third argument as to why the weak conception cannot account for all the relevant cases is that collective beliefs attributed to structured groups need not be held by any of the individual members of the group. Consider the following rather extreme example: The cabinet believes that genetically modified food is safe and that it should not be forbidden, though this is no longer believed by any of its members individually. Such a situation might arise if new evidence becomes available about the potential effects of genetically modified food. Cabinet members may then start doubting whether their collective belief that such food is safe is true. The official view would still hold, however, and commit the members in what they say and do. In such cases a distinction would arise between what the cabinet believes and what the individual members as individuals believe.2 Such a collec2 It is necessary in such cases to make a distinction between roles. In their role as cabinet members they believe something different from what they believe in their role as indi-

Collective Agents and Cognitive Attitudes

75

tive belief would still influence subsequent cabinet actions in all kinds of situations even though none of the individual members actually holds this view. Another example is a committee that needs to decide on the best candidate for a particular job.3 If nobody’s first choice is shared, the members of the committee may agree on a candidate who is second best. None of the individual members believes, then, that (s)he is the best candidate for the job, whereas the committee as a whole does. Again we seem to have a case where a belief is attributed to a collective agent, but is not believed by any of its individual participants. The opinion poll conception can handle majority or consensus cases, but not those where the belief in question is not present at all among the individuals making up the group. I conclude from my discussion that the first conception of collective beliefs distinguished above (the weak opinion poll conception) fails to capture cases where the collective agent or group is not simply an aggregate of individuals but a structured whole. The weak conception can thus only account for part of the spectrum of collective beliefs. In many situations we shall have to rely on the strong agreement-based conception of collective beliefs. Neither conception, however, exhausts the total spectrum of collective cognitive attitudes. In addition, the strong conception cannot be reduced to the weak one. In my discussion below, I shall focus mainly on beliefs that fit the agreementbased conception. These beliefs are the more interesting ones as they cannot be reduced to individual beliefs, and require additional conceptual tools for their analysis. I shall argue that the distinction between belief and acceptance is useful for their analysis and that these collective beliefs are never purely epistemic. They are to be analysed using the model of acceptance rather than of belief. This makes voluntarism with respect to this type of collective beliefs a true thesis. As individuals we may not be able to believe at will, but we are able to accept a proposition at will, and, thus, to believe at will collectively.

4 The Analysis of Collective Beliefs There are certain asymmetries between individual beliefs and collective beliefs of the strong agreement-based type. Let me mention one that immediately comes to mind. Groups obviously do not have a mind or consciousness. If vidual persons. I shall argue below that such cases are better conceived of as cases of accepting instead of believing a proposition. 3 I owe this example to Lynne Rudder Baker.

76

Anthonie Meijers

accessibility to consciousness is conceived of as part of the notion of belief, as defended by John Searle (among others),4 or if beliefs are conceived of as dispositions to have certain feelings of truth and falsity when one is attending to issues raised or items referred to by a proposition, as defended by Jonathan Cohen,5 then it is not evident how this could apply to collective beliefs. The explicit comparison of individual and collective cognitive attitudes is, however, helpful to further refine and develop our notion of collective beliefs. This will be my strategy in this section. Following Bratman, Pascal Engel lists six “obvious” properties that individual beliefs have and compares these properties with those of the mental state of accepting.6 I shall not question these properties, but simply assume that they apply to the individual case. Engel uses these properties to distinguish between the often-confused propositional attitudes of believing, holding true, and accepting. That these attitudes are not the same is shown by the fact that we may hold true sentences without believing what they say. For example, I may hold true the sentence ‘all equivalent infinite sets have the same transfinite cardinal’ on the authority of a mathematician. But if I do not understand what it says, I cannot believe its content. We may also accept a proposition in our practical deliberations without believing it. For example, I may accept for prudential reasons the proposition that it will rain tomorrow, and act on it, without really believing it. The discussion below will show that the distinction between believing and accepting is important for the analysis of collective beliefs of the agreementbased type. The notion of accepting has at least two connotations. It may refer to the mental act of assenting to a sentence or proposition, or to the mental state of having a certain policy. Though both connotations are relevant for understanding collective beliefs, I shall initially concentrate on acceptance as the state of having a certain policy. Let me start with the first property that distinguishes beliefs from acceptances: (1) beliefs are involuntary, whereas acceptances are not Individual beliefs are not subject to direct control. For example, I cannot believe at will that I shall never die, or that Dutch will become the universal lan4 See John Searle, The Rediscovery of the Mind (1992). 5 See Jonathan Cohen, An Essay on Belief and Acceptance (1992). 6 Pascal Engel (1998), “Believing, Accepting, Holding True”, 143-4 and 146-7.

Collective Agents and Cognitive Attitudes

77

guage of the sciences. Acceptance, on the other hand, is voluntary. To accept that p is to make the decision to use a certain proposition as a valid premise in one’s practical deliberations. It is to adopt a certain policy. If we apply the voluntary/involuntary distinction to collective beliefs of the agreement-based type, it follows that they are more like acceptance than belief. The main reason is that these collective beliefs require some sort of voluntary assent, agreement, or decision by the members of the group for the belief to become a collective belief.7 This is clearly illustrated by the cabinet example. Only if the appropriate decision procedures have been followed may a belief properly be called a belief of the cabinet. In other cases, individual members speak for themselves. The cabinet case is misleading, however, in that it suggests that an explicit or even a formal act of agreement or decision is always necessary for agreement-based collective belief. Usually agreement is informal and often it is implicit. It is shown in the way in which members of a group behave. They express their commitment to certain collective beliefs in their practices. (2) beliefs aim at truth, whereas acceptances aim at utility or success This all-important property of beliefs expresses the fact that beliefs aim at representing the world as it is; they have a mind-to-world direction of fit. Wishful thinking, as opposed to belief, is about the world as we would like it to be. Acceptance, on the other hand, is a pragmatic notion. To accept a proposition is to adopt a policy and use the accepted proposition as a premise in one’s practical deliberations about future actions. Acceptance is usually informed not only by epistemic considerations, but also by prudential ones. We may even accept and act on a proposition that we believe is likely to be false. An example would be to act on a worst-case scenario. Do collective beliefs aim at truth, or at utility and success? Let me distinguish the truth view of collective beliefs from the pragmatic view. The distinction is between a view that takes into account only epistemic considerations and aims at truth, and one that also takes into account practical or action-oriented considerations and aims at success. An indication that collective beliefs are best understood as acceptances (the pragmatic view) is that these beliefs tend to be context bound, whereas genuine beliefs and their truth-values are not. We, as a group, usually believe that p in 7 As opposed to collective beliefs of the weak opinion poll type.

78

Anthonie Meijers

the strong agreement-based sense given a particular situation, role, or point of view where practical considerations are part of the reasons for jointly believing that p. The cabinet’s belief that genetically modified food is safe and should not be forbidden is typically context bound. Genuine beliefs, on the other hand, are not context bound. A person believes that p is true (for example, that the earth is round) regardless of the context (s)he is in.8 Beliefs also require that all the relevant evidence be taken into account, whereas in a given context a less inclusive point of view may suffice for the acceptance of a proposition. It is accepted because of its practical utility in a certain situation. A second indication is that we may give up our strong collective beliefs for reasons which are not related to the epistemic evidence we have, but because of prudential concerns. Some members of the cabinet may fear the political repercussions of their collective view that genetically modified food is safe, and be willing to change it. If the truth of a belief were the only thing that mattered in collective cognitive attitudes, it would be incomprehensible how such a change could take place. It would also be inexplicable that there are inter-individual commitments involved in these beliefs, and that these commitments, as we have seen above, are conditional upon the attitudes of the other members of the group. A third indication that the pragmatic view is correct, which I mentioned earlier, is that we, as a group, may believe a proposition that no one takes to be true individually. The committee selecting a candidate illustrated this point: no single member of the committee believes that the candidate agreed upon is the best candidate, though it may be said that the committee does believe this. If my analysis is correct, it means that collective beliefs of the agreement-based type are to be characterized as acceptances rather than as beliefs. It may be argued against the pragmatic view that there are collective beliefs of the agreement-based type to which the truth view does apply. Imagine a religious community having the explicit goal of developing true collective beliefs. This community of truth seekers does not care about the practical value of the beliefs which they collectively hold. They just want the truth and nothing but the truth. Is this a counterexample to the pragmatic view? I believe it is not because these religious beliefs turn out, on closer inspection, to be different from collective beliefs of the agreement-based type. 8 A group can, of course, also believe that the earth is round in an epistemic sense, but this collective belief is of the opinion poll type as opposed to the agreement-based type. Everybody in the group believes that the earth is round, as opposed to us, as a group, believing that the earth is round.

Collective Agents and Cognitive Attitudes

79

Firstly, it seems to be a necessary requirement for these beliefs that all the members of the religious community individually believe what they jointly take to be the truth. There is no room for compromise or majority rule in religious affairs, and the extreme case of a view that nobody individually believes, but which the community as a whole believes to be true, is out of the question. Secondly, the idea of inter-individual commitments among members of the religious group is problematic. Can somebody else rightfully claim that I stick to my religious beliefs if I personally no longer believe this? Is my commitment to these beliefs conditional upon the other members of the community believing the same? It may be argued that authority plays an important role in religions, and that some religious communities may have leaders who determine truth and falsity. Such a leader may claim that I stick to certain beliefs, which shows that commitments are part of religious beliefs. The answer to this objection is that if I stick to these beliefs because of this claim, and not because of evidence given to me, my religious beliefs have become beliefs which no longer aim solely at truth – which was the assumption the counterexample was based on. If I stick to these beliefs because I regard the utterances of the religious leader as epistemic evidence ((s)he may have a privileged relation with God), my belief is an individual belief, and not a collective belief of the agreement-based type. So, either these collective religious beliefs are of the opinion poll type, and aim at truth, or they are of the pragmatic type, and are not only based on evidence, but on authority as well. I conclude from this discussion that the truth view of collective beliefs does not hold. Religious beliefs do not form a threat to the pragmatic view as these beliefs are best analysed as beliefs that conform to the opinion poll conception. Though they are believed by every member of the religious community they should be analysed as individual beliefs. And, as such, they do indeed aim at truth. Let me turn to the third and the fourth characteristics Engel uses to distinguish beliefs from acceptances: (3) beliefs are shaped by evidence, whereas acceptances need not be so (4) beliefs are subject to an ideal of integration, whereas acceptances require only coherent policies If beliefs aim at truth, their credibility is proportional to the evidence which we have for what is believed. An agent’s belief also needs to be consistent with his or her other beliefs, while inconsistency would count as evidence against that

80

Anthonie Meijers

particular belief. Acceptances, on the other hand, need not be shaped by evidence. Though, of course, the truth of a proposition counts as a strong argument for accepting it. Non-evidential considerations, however, also play a role in what we are prepared to accept as premises in our practical deliberations. These premises need to be internally coherent. If they are in conflict, it is hard to base our future actions upon them. When applied to collective cognitive attitudes, it can said that collective beliefs of the opinion poll type are shaped by evidence, and subject to theoretical integration, as they are nothing but individual beliefs and a certain distribution of these beliefs in the group. Collective beliefs of the agreement-based type, on the other hand, are, if my pragmatic view is correct, shaped by utility considerations, and subject to integration in a practical sense. Again, the suggestion here is not that acceptances are not shaped by evidence. The point is, rather, that they are not solely grounded on evidence. In practical situations where action is necessary there will often be a mixed situation where the evidence is not sufficient or conclusive, and practical considerations (e.g. prudence) will be brought in to decide on which view to use in our deliberations about possible actions. To use our earlier example, we expect the cabinet to have coherent policies about genetically modified food and to have beliefs supported by as much evidence as possible. In addition, there will be propositions which are taken for granted when it comes to practical action. The fifth property that distinguishes beliefs from acceptances is: (5) beliefs are context-independent, whereas acceptances are context-dependent In my discussion of the second property (beliefs aim at truth, acceptances aim at utility or success) I have already referred to the importance of the context for the analysis of collective cognitive attitudes. Collective beliefs of the agreement-based type turned out to be context-dependent and, therefore, more like acceptances than genuine beliefs that aim at truth. Let me add a specific sense in which collective beliefs may be called context-dependent. In the case of individual beliefs, the meaning of the proposition that is believed is usually evident. An agent knows what he believes. This, however, cannot be taken for granted in the case of a collective agent. Shared understanding of the proposition at stake is required. What is meant, for example, by the simple expression that genetically modified food “is safe”? Initially, this may not be clear to the members of the cabinet and settling this issue will be part of the agreement

Collective Agents and Cognitive Attitudes

81

which is constitutive of the collective belief. Furthermore, the interpretation of phrases that were initially taken for granted may cause serious conflict at a later time The process of collectively accepting a proposition as a shared view is, at the same time, a process of developing a shared understanding of what that proposition means. In this important sense, collective beliefs are also contextdependent. Their meaning is tied to the context in which the group acquires them. I now turn to the last property, which is that (6) beliefs come in degrees, whereas acceptances are qualitative. Individual beliefs are, according to Pascal Engel’s Bayesian view, quantitative; they have subjective probabilities. Acceptance, on the other hand, is a qualitative phenomenon; it is a matter of all or nothing. If we apply the distinction between quantitative and qualitative to the notion of an agreement-based collective belief, it is evident that it is a qualitative notion. As we have seen, collective beliefs of this type require an act of (explicit or implicit) assent or agreement, and agreement is an on/off matter. We agree or we do not. This, again, strengthens the idea that this type of collective belief should be understood in terms of the acceptance of a view rather than of a belief proper. Let me summarize what has been accomplished in this section. I have made a comparison between the individual attitudes of belief proper and acceptance on the one hand, and collective cognitive attitudes on the other. The comparison of properties has shown that collective beliefs of the agreement-based type are similar to acceptances, while collective beliefs of the opinion poll type are similar to beliefs proper. Agreement-based collective beliefs are never purely epistemic and will always involve practical and contextual considerations. To accept that p collectively is to adopt a policy of positing or postulating that p, given a certain context, and to use the proposition that p in the group’s practical deliberations. My conclusion here is largely in line with Raimo Tuomela’s analysis of collective beliefs, whereas another key author, Margaret Gilbert, does not explicitly address the issue of believing versus accepting as a group.9 Gilbert’s Durkheimian view, however, that the existence of group beliefs is logically independent of the correlative individual beliefs suggests that she is sympa9 See Tuomela (1995) and Tuomela (2000).

82

Anthonie Meijers

thetic to the position defended in this paper according to which collective beliefs should be understood in terms of the acceptance of a proposition. Tuomela makes a distinction between experiential beliefs (for example, perceptual beliefs) and acceptance beliefs, where the latter can be divided into narrow and wide acceptance beliefs. Narrow acceptance is truth-oriented, whilst wide acceptance involves pragmatic virtues. He argues that collective beliefs proper (positional beliefs in his terminology) always involve wide acceptances. His analysis is part of a much broader analysis of the notion of acceptance, which covers several other cognitive attitudes that fall under the everyday concept of collective belief. Agreement-based collective beliefs – being acceptances of a proposition – are voluntary and, as such, subject to the will of the members of a group. In this sense, voluntarism is trivially true of this type of collective beliefs. Collective beliefs of the opinion poll type, on the other hand, need to be analysed as beliefs proper. Voluntarism with respect to this type of belief is much more controversial. This issue, however, falls outside the scope of this paper. Taking agreement-based collective beliefs as acceptances also helps to understand how commitments between members of a group are part of this type of belief. As we have seen, if a group believes that p, individual members are committed to and bound by this belief and cannot suddenly change their minds about the issue; others may even claim that they stick to the belief of the group. Furthermore, their commitment to accept this view is conditional upon the commitment of the other members of the group to do too. These binding phenomena are completely incomprehensible using the model of belief proper, whereas they make sense when analysed using the model of acceptance. Interindividual commitments arise from agreeing on or jointly accepting a view. A few important characteristics of this notion of joint acceptance will be explored in the next and final section.

5 The Joint Acceptance of a View There are several asymmetries between jointly accepting a view and individually accepting it. Two of these I have already mentioned. The first is that acceptance by the members of a group is conditional. I am only prepared to accept a certain view as the group view – with all its binding effects – if other members of the group do the same. Furthermore, if they fail to live up to commitments that follow from such a jointly accepted view, then I am no longer committed.

Collective Agents and Cognitive Attitudes

83

The second asymmetry is that what I am prepared to accept in my role as a member of a group may differ from, or even conflict with what I am prepared to accept individually. The compromise case of the committee determining the best candidate for a particular job illustrated this point. Joint acceptance is relative to one’s role as a member of a group. The third asymmetry regards the public character of accepting. Acceptance of or assent to a proposition in the individual case may be a silent or a mental act. This kind of acceptance is not adequate in the context of collective beliefs. What is needed here is an overt act of accepting – usually a speech act. The others need to know, implicitly or explicitly, that I am willing to accept a certain proposition as the view of the group, and to use this proposition as a premise in our subsequent deliberations. Because of the public nature of acceptance in the collective case, it also becomes common knowledge among the members of the group that members accept this view as the view of their group. A fourth asymmetry has to do with the qualitative nature of acceptance. Whereas acceptance in the individual case is a clear qualitative notion, matters are more complicated in the collective case. Collective beliefs of the agreement-based type figure as premises in practical deliberations within a group, and are used to determine future actions. It becomes important, then, what kind of support a particular view has. A group may accept a proposition in various ways: by complete consensus, by a qualified majority, by a simple majority, or by delegating the decision to a subgroup or even a single expert. The amount of support a view has – or, to put it differently, its degree of acceptance – will be an important factor in determining subsequent actions. What is legitimised by the cabinet’s belief that genetically modified food is safe depends, among other things, on the support this particular belief has in the cabinet. So, in the case of agreement-based collective beliefs we seem to have a notion of acceptance that has quantitative features too. Above a certain threshold determining whether a view is accepted or not, the degree of acceptance plays a role in subsequent practical deliberations. The last asymmetry I want to mention here is that the joint acceptance of a view can be what Margaret Gilbert calls ‘collectivity involving’.10 In the act of jointly accepting a view, a group may be constituted that did not exist before. The joint view then becomes the defining characteristic of the group. This point is related to a question which I raised at the beginning of my paper: are attributions of collective beliefs literal or metaphorical? Groups are different 10 Gilbert (1992), 256.

84

Anthonie Meijers

from persons; they do not have a mind or consciousness. So, what could such an attribution to a group mean? As we have seen, agreement-based collective beliefs are best understood as acceptances rather than as beliefs, or, to put it differently, they are the adoption of a policy by a group. The groups to which we attribute such propositional attitudes are certainly not to be conceived of as rather mysterious super-individual subjects which exist independently of their members. At the same time, a summative approach to groups – according to which a group is a simple summation of individuals – will not work either. Such an approach cannot account for the normative relations between the members of the group. The analysis outlined here suggests that collective beliefs inhere in groups, where a group is taken to be an association of individuals. Members are minimally bound by their joint acceptance of a proposition. To attribute a collective belief, then, is to literally attribute the acceptance of a proposition to a group, consisting of associated or related individuals.11

6 Conclusion Collective cognitive attitudes come in various types. Two types have been distinguished in this paper: the weak opinion poll type, and the strong agreementbased type. In between these two types there is a continuum of cognitive attitudes. The analysis has shown that the strong agreement-based type cannot be reduced to the weak one because of the various commitments involved. It has also shown an important and related asymmetry between individual and collective cognitive attitudes. While weak collective beliefs can be understood in terms of individual beliefs and a certain distribution of these beliefs among the members of a group, strong collective beliefs are different. They require an act of acceptance by the group. Such an act cannot be an act in foro interno; it has to be a public act which is transparent to the members of the group. The result of this analysis is twofold. On the one hand, the analysis implies that voluntarism is true with respect to collective beliefs of the strong agreement-based type. These beliefs depend on the will of the members of the group. Together they decide on the beliefs collectively held. On the other hand, from the analysis it follows that strong collective beliefs will never be purely epistemic, and are unlike individual beliefs. They always have a prag11 See also my “Collective Beliefs and Practical Realism: Giving Relations their Proper Metaphysical Due”, in Meijers (2001).

Collective Agents and Cognitive Attitudes

85

matic and contextual dimension and should, therefore, be understood in terms of acceptances rather than of beliefs. To accept a proposition as a group is to adopt a certain policy to use that proposition as a premise in the group’s deliberation about future actions. Given a situation, the specific roles of the members of the group, and their shared understanding of what is at stake, they are jointly prepared to accept a certain proposition as their collective belief. That such a belief has a pragmatic dimension is not to say, of course, that it is purely pragmatic. The truth of a proposition is usually a strong argument for accepting it.12

Bibliography Baker, Lynne R. Explaining Attitudes; A Practical Approach to the Mind. Cambridge: Cambridge University Press, 1995. Bratman, Michael. “Practical Reasoning and Acceptance in a Context”. Mind 102 (1993), 1-15. Cohen, Jonathan. An Essay on Belief and Acceptance. Oxford: Oxford University Press, 1992. Engel, Pascal. “Believing, Accepting, Holding True.” Philosophical Explorations I (1998), 140-151. Gilbert, Margaret. “Modelling Collective Belief.” Synthese 73 (1987), 185-204. – , On Social Facts. Princeton: Princeton University Press (1992). Chapter V: “After Durkheim: Concerning Collective Belief”, 237-314. Meijers, Anthonie. “Collective Beliefs and Practical Realism: Giving Relations their Proper Metaphysical Due.” In Explaining Beliefs; Lynne Rudder Baker and her Critics, edited Anthonie Meijers. Stanford: CSLI Publications, 2001, 163-182. Searle, John R. The Rediscovery of the Mind. Cambridge (Mass.): MIT Press, 1992. See Chapter 7: “The Unconscious and its Relation to Consciousness”. Tuomela, Raimo. The Importance of Us. A Philosophical Study of Basic Social Notions.+ Stanford: Stanford University Press, 1995. Chapter 7: “Group Beliefs”, 306-332. Tuomela, Raimo. “Belief versus Acceptance”. Philosophical Explorations III (2000), 122-137.

12 This article is a revised version of “Believing and Accepting as a Group”, originally published in Anthonie Meijers (ed.), Belief, Cognition and the Will (Tilburg: Tilburg University Press, 1999), 59-71. The Press was closed down shortly after the publication of the book. I would like to thank Luc Bovens, Randall Dipert, and Monica Meijsing for their stimulating comments on previous versions of my paper.

86

Deborah Perron Tollefsen

Challenging Epistemic Individualism Deborah Perron Tollefsen

Abstract Contemporary analytic epistemology exhibits an individualistic bias. The standard analyses of knowledge found in current epistemological discussions assume that the only epistemic agents worthy of philosophical consideration are individual cognizers. The idea that collectives could be genuine knowers has received little, if any, serious consideration. This individualistic bias seems to be motivated by the view that epistemology is about things that go on inside the head. In this paper I challenge this type of epistemic individualism by arguing that certain groups are intentional agents and that cognition occurs at the level of groups. In section I, I extend a plausible and well-defended account of individual intentionality to organizations. In section II, I appeal to research on distributed cognition in order to challenge the view that cognition occurs exclusively in the individual mind-brain. Having established that groups can be believers and cognizers, I turn to the issue of the justification of group belief. In section III, consider whether a reliabilist account of justification can be extended to groups.

Contemporary analytic epistemology is preoccupied with individual cognition. The standard analyses of knowledge found in current epistemological discussions assume that the only epistemic agents worthy of philosophical consideration are individual cognizers. Aside from Popper’s attempt to eliminate the knowing subject (Popper: 1972), the idea that collectives could be genuine knowers has received little, if any, serious consideration.1 This individualistic bias seems to be motivated by the view that epistemology is about things that go on inside the head. As one prominent epistemologist puts it, “Knowers are individuals, and knowledge is generated by mental processes and lodged in the mind-brain.” (Goldman: 1987) In this paper I challenge this type of epistemic individualism2 by arguing that certain groups are intentional agents and that cognition occurs at the level of 1 There have been recent attempts to understand “social” knowledge but these accounts still do not acknowledge groups as legitimate epistemic agents. See, for instance, Corlett (1996) and Lehrer (1987) 2 There are other ways in which epistemology has exhibited an individualistic bias. For instance, traditional epistemology has generally given the testimony of others a second class justificatory status. Traditional epistemology has also been criticized for focusing on the epistemic agent in isolation from social contexts. Give Schmitt sources.

Challenging Epistemic Individualism

87

groups. In section I, I extend a plausible and well-defended account of individual intentionality to organizations. In section II, I appeal to research on distributed cognition in order to challenge the view that cognition occurs exclusively in the individual mind-brain. This research strongly suggests that cognition can also take place within groups. Having established that groups can be believers and cognizers, I turn to the issue of the justification of group belief. In section III, consider whether a reliabilist account of justification can be extended to groups.

I Epistemic individualism has already been challenged to some extent by contemporary accounts of collective intentionality. Consider, for instance, Margaret Gilbert’s account of group intentionality.3 According to Gilbert, under certain circumstances, a group of people can form a unique subject to which the ascription of intentional states is appropriate. Members form a plural subject of belief, for instance, when they openly express their willingness to be jointly committed with others in believing that p as a body. If groups are the legitimate subject of intentional state ascription then this would appear to challenge the view that groups can not be epistemic agents. After all, if groups are believers then it is natural to ask the further question of whether group beliefs are justified and whether they constitute knowledge. The challenge will be easily dismissed, however, unless we can show that plural subjects are truly intentional subjects. Gilbert provides us with an account of how plural subjects are formed and an analysis of our group intentionality ascriptions but what we need to show is that group beliefs are beliefs.4 Otherwise, our ascriptions of intentionality to groups might be thought of as merely an abbreviated way of referring to a set of complex individual intentional states rather than referring to genuine intentional states of the group. If groups are not true believers then they are not epistemic agents. I have argued elsewhere (2002a), (2002b), (2002c), that interpretationism provides the perfect framework for understanding how certain groups can be 3 Gilbert has developed her account in various places. See in particular, (1988) and (1996). 4 Gilbert herself does not seem to take a definitive stand on the issue of whether group beliefs are genuine beliefs and whether plural subjects are literally intentional agents. She notes only that there is an analogy between individual and group belief. See 1988: 313 for her discussion of the analogy.

88

Deborah Perron Tollefsen

true believers. Interpretationism is the view that if an agent is interpretable then they are an intentional agent. This view has been most notably defended by Daniel Dennett and Donald Davidson.5 Interpretationists claim that we can reach an understanding of the nature of the mental by reflecting on our practice of interpreting others. Rather than asking what intentional states are, the interpretationist argues that we must begin by looking closely at our common sense practice of attributing intentional states and ask, “ what are the constitutive features of this practice such that our practice of making sense of others is successful?” The study of the conditions or constraints on interpretation will yield metaphysical insights. These constraints are not merely methodological but constitutive of the mental. Reflection on our practice of making sense of others suggests that intentional states are not, as the Cartesian picture of mind has it, internal entities in the head or mind. When we attribute beliefs and intentions to individuals we attribute states to them. Our practice makes sense of persons or unified subjects, not simply behavior or utterances. When we say of Susan that she believes it will rain and desires to stay dry this explains why Susan carries an umbrella to work. I want to know why she behaved in this manner. So my primary interest is in understanding Susan and not simply explaining her bodily movements or what happened in her brain. Intentional states, then, are states of persons or systems. Reflection on our practice also reveals that the content of intentional states is determined in various ways by external rather internal aspects. Whether someone believes that what is wet will depend on whether she has been in causal contact with water and what ‘water’ means in her socio-linguistic environment. Interpretationists, then, are externalists about mental content. When we offer an explanation of an agent’s behavior, we provide reasons for their behavior. Explanation in terms of beliefs and desires is rationalizing explanation and not merely causal. To provide a rationalizing explanation rather than a mere rationalization is to provide a particular type of justification for the agent’s behavior. We are not looking simply for reasons why someone would behave in that way (any number of reasons may rationalize the behavior) but the reasons for which she acted the way that she did. Again, we are interested in her reasons not simply reasons per se. This, in turn, requires that 5 Davidson develops his version of interpretationism in various places. See especially, (1963), (1970), (1973), (1982), (1993). Dennett’s version of interpretationism can be found in (1971), (1979), (1987a), (1987b) and (1991).

Challenging Epistemic Individualism

89

when we attribute beliefs, intentions, and desires to an agent in order to provide them with reasons for their behavior these must be reasons that can be seen as intelligible from the agent’s perspective as well as our own. The connection between intentional states and the notion of intelligibility is one source of the interpretationist’s principle of charity: interpret in such a way as to minimize inexplicable error. To minimize inexplicable error, we must assume that the agent shares our norms of rationality, i.e. that the agent is rational. This means that we must suppose that they have a pattern of interrelated attitudes that is similar to the pattern of our own beliefs and attitudes. This pattern is essentially governed by the norms of consistency, coherence, and truth. The constraints we recognize as applying to our own beliefs, constraints that we could loosely call epistemological, are constraints on interpretations of others. This does not mean that we cannot attribute false beliefs nor does it mean that we can never attribute attitudes in such a way that would reveal the agent as disagreeing with us. On the contrary, the very possibility of disagreement depends on a backdrop of shared beliefs and attitudes. The fact that widespread agreement is presupposed by our disagreement makes clear that, in addition to the rationality constraint on interpretation, there is a holistic constraint. We can not intelligibly attribute beliefs in an atomistic way. The content of belief, is in part, a function of its place in a network of other attitudes. The intelligibility of attributing some thought to another rests upon the possibility of attributing to him or her lots of other thoughts most of which are true and intelligible by our lights. If our explanation of the agent in terms of beliefs and desires makes her behavior intelligible to us only if it can be seen as intelligible from the agent’s perspective, then our practice of interpretation involves the positing of an alternative point of view; a rational point of view from which the intelligibility of one’s own behavior can be assessed. When we make sense of others we often project ourselves into their rational point of view in order to better provide reasons that are intelligible in the way the reflexive constraint requires. Agents could not be assumed to be governed by our norms of rationality if we did not also assume that they had a rational point of view from which they can assess their beliefs, desires, and intentions for consistency, truth, and intelligibility. In The Bounds of Agency (1998) Carol Rovane suggests a similar account of the notion of a rational point of view. The rational point of view, according to Rovane, need not be tied to a particular soul, animal, or phenomenological view point. A rational point of view is a view from which a person deliberates. According to Rovane, we engage in the following sorts of activities from the

90

Deborah Perron Tollefsen

rational point of view: accept the implications of one’s attitudes, resolve contradictions and conflicts, rank preferences, assess opportunities for action, determine means for arriving at ends, consider the consequences of certain actions and attitudes. But the rational point of view is more than simply a perspective from which one deliberates. It is also a locus of power and responsibility. When a person deliberates from the rational point of view the outcome of their deliberations have a certain immediacy for them. They are motivated to change or alter other attitudes they have. It is from this perspective that reasons have the immediacy characteristic of reasons. It is for this reason that the rational point of view is often associated with the first person singular. Without a first person perspective or rational point of view, assessments of rationality would make explicit no immediate reason to alter the commitment being evaluated because the judgment is not attached to a judge.6 The first person concept marks this locus of power. Once we can assume that a subject of interpretation is rational, in the sense that their actions and beliefs originate from a rational point of view, we attribute to the agent intentional states that a rational agent ought to have. We begin by attributing beliefs that the agent ought to have given her environment and function in the world. Then we figure out what desires she ought to have given the same considerations. Once we have attributed to the agent the beliefs and desires a rational agent ought to have, we can predict what a rational agent with those beliefs and desires ought to do or appeal to these beliefs and intentions in order to explain the agent’s behavior. To the extent that our practice is successful and the basis of a great deal of successful explanation, our assumption of rationality is justified. And when the assumption of rationality is justified then we are dealing with an agent who has a pattern of belief and thought that is similar to our own and that is governed by the same norms of intelligibility. If our best efforts to make sense of an individual fail then there is no reason to believe that we are dealing with a rational agent and hence there is no reason to believe we are dealing with an intentional agent. If this is so then the nature of intentionality is revealed in our practice of making sense of others. According to interpretationism, then, “… all there is to really and truly believing that p (for any proposition p) is being 6 This line of argument is developed by Burge in (1998). His aim, though, is not to establish the necessity of the first person concept for reasoning but the necessity of the first person concept for understanding reason and reasoning. I make the stronger claim here that such a concept is necessary in order for reasons to have the immediacy characteristic of reasons.

Challenging Epistemic Individualism

91

an intentional system for which p occurs as a belief in the best (most predictive) interpretation” (Dennett 1987: 29). One of the most striking features of our practice of attributing intentional states is that it is not limited to individuals but extends to other sorts of subjects. Pick up any newspaper and you are likely to find passages like the following: Microsoft believes that its status as an accused monopolist entitles it to use the judicial process to delve into its competitors’ most sensitive commercial information. (USA Today, May 9, 1999) The lawsuit filed by Washington Attorney General Christine Gregoire accuses the tobacco industry of conspiring to violate antitrust and consumer-protection laws, restraining and intentionally suppressing health research, manipulating cigarette nicotine levels without disclosure and threatening retaliation against companies that marketed safer cigarettes. (Associated Press, June 9, 1999) Even more interesting is the role these ascriptions play in social scientific research. Even the most ardent individualists routinely apply the intentional idiom to groups like corporations, nations, and committees. Consider, for instance, Jon Elster’s illumination of the structure of choice between alternative modes of energy production facing most Western nations. Elster writes, On the basis of pure self-interest, reinforced by moral reasoning, a nation that has not already acquired nuclear weapons should abstain from nuclear power, so as not to have to hand down to later generations a potential for starting or precipitating a nuclear war … (Elster 1983: 200) Elster arrives at this conclusion by applying rational choice theory to nations rather than to individuals – a theory that attributes beliefs and desires to agents in an effort to determine the rational way of making choices. Although he does not explicitly attribute group intentional states, the use of rational choice theory presupposes them. The use of the intentional idiom is also implicit in his talk of a nation’s self-interest and its ability to engage in moral reasoning.7 In both the colloquial contexts and in the social scientific contexts these attributions are made in an effort to explain and predict the behavior of groups. Prima facie, then, we are engaged in the same interpretative enterprise we are 7 Elster would, no doubt, argue that these ascriptions can be reduced to statements ascribing intentional states to individuals. My point here is simply that social scientists and philosophers of social science apply the intentional idiom at the level of groups.

92

Deborah Perron Tollefsen

engaged in when we attempt to make sense of other people. If this is so, the constraints that govern our attempts to make sense of other people also govern our attempts to make sense of the actions of groups. If groups are interpretable, they are rational agents. But how are we to understand groups as rational agents? The assumption of rationality involves the positing of a rational point of view. What sense is there to be made of the rational point of view of a group or organization? And when assume an organization is rational aren’t we simply assuming that all or some of its members are rational? As I argued in “Organizations as True Believers,”(2002b) groups like organizations are governed by the norms of consistency and coherency and form a unique rational perspective from which organizational members deliberate.8 An organization is a complex social unit that exists on a relatively continuous basis and engages in activities that are usually related to a set of goals. The rules and policies of an organization present a unified perspective from which decisions can be made, information assessed, inferences drawn, consequences considered, means to ends determined, inconsistencies avoided and joint actions performed. The structure of the organization provides a way of synthesizing the disparate perspectives of individuals into a unified perspective from which goals and subgoals can be set and achieved. This rational point of view of an organization is expressed in organizational policy, in the structure of the organization, and in the history of its decision making. Recall, that a rational point of view is a perspective from which deliberation takes place. When individuals deliberate in an organizational setting they adopt the rational point of view of the organization. It is from the point of view of the organization rather than their own personal point of view that deliberation ought to take place.9 Organizations establish hierarchies of goals. Means for achieving general goals become themselves subgoals. Each level is considered an end relative to the levels below it and as a means relative to the levels above it. An organization’s hierarchy can be viewed as a congealed set of means-ends chains promoting consistency within the organization. Rationality, then, resides in the 8 The view that organizations can be intentional agents has also been defended by French (1984). French, however, focuses on a particular type of organization-corporations. In (2002a) I build on the work of Philip Pettit (2000), (2001) and show how the norms of rationality apply at the level of groups. In particular, I focus on small decision making groups that are often found within larger organizations. 9 In fact, one explanation of the failure of individuals to act jointly and the failure of group decision making is that individuals are deliberating from their own rational points of view rather than adopting and sharing the perspective of the organization.

Challenging Epistemic Individualism

93

structure of the organization itself and not merely in the individual participants. Our attempt to make sense of the actions of organizations by ascribing to them intentional states would fail unless we assumed that the organization itself is rational. This involves assuming that the organization has a rational point of view from which members engage in the same sorts of cognitive activities individuals engage in and that the organization is governed by the same norms of rationality. When we assume that an organization is rational, then, we assume that it, rather merely the members in it, are governed by norms like consistency and coherency. Once the assumption of rationality is in place we attribute beliefs, intentions, and desires to organizations in the same way we do to individuals. We ascribe those beliefs that an organization ought to have if it wants to achieve its goals efficiently and given its function in the environment. We attribute desires in the same manner. And such attributions are made holistically as well. They presuppose other beliefs, goals, and intentions. These attributions are then the basis for explanation and prediction of the actions of organizations.10 The success of practice of interpreting organizations provides strong justification for our assumption that we are dealing with a rational agent. If our assumption of rationality is justified in the case of organizations then this is all we need to know in order to count organizations as intentional agents. If all there is to being a believer is to be a system whose behavior is interpretable from the intentional stance then organizations are true believers. Although I have provided a full defense of this line of reasoning elsewhere (2002c), it will be helpful to consider and reply to a few standard criticisms. First, a critic might challenge my claim to have defended the view that organizations really have beliefs. Interpretationism, according to some crit10 Do we actually apply this method in our explanation of organizational behavior? That is, do we adopt this interpretative stance when we try to make sense of organizations? I think reflection on our everyday practice reveals that we do. Consider the predictions we make about car companies. What will Ford Motor Company do in response to the rise in gas prices? If Ford is rational then it will act so as to maximize its profits. Because Ford wants to avoid losing money on their line of large vehicles and believes that individuals are less likely to buy large vehicles during a time at which gas prices are high, we can predict that Ford will discount its large vehicles. Clarke (1992) provides this good example. Suppose we want to explain the incidents of gunfire occurring during a naval blockade. We read the official rules of engagement published to govern naval operations. These rules embody the Navy’s rational point of view. We can then explain why boats sometimes shoot at other boats. Under certain conditions, specified in its rules of engagement, the navy will fire on anything it believes to have hostile intent

94

Deborah Perron Tollefsen

ics11, is just a form of instrumentalism and if all I am claiming is that from a certain perspective one can view organizations as having intentional states then organizations don’t really have intentional states we just treat them as if they do. Therefore, the position I have defended is the opposite of what I intended: that our ascriptions of intentional states to groups are useful fictions but, strictly speaking, false. A brief look at Dennett’s discussion of realism will, I think, provide some response to this concern. Interpretationism is a “mild and intermediate sort of realism” (Dennett 1991: 29). The patterns of human behavior discernible from the intentional stance are objectively out there to be interpreted. Dennett develops this point by appealing to Gregory Chaitin’s definition of mathematical randomness. A series (of dots or numbers or whatever) is random if and only if the information required to describe (transmit) the series accurately is incompressible: nothing shorter than the verbatim bit map will preserve the series. Then a series is not random – has a pattern – if and only if there some more efficient way of describing it. (Dennett 1991: 32) A pattern, then, is real, if there is a description of the data that is more efficient than the bit map. There may, of course, be several different ways to efficiently describe the data but that fact does not mean that the pattern is not really out there to be described. Dennett applies these considerations to the arena of intentional state attribution. The patterns of human behavior are real patterns because there is a description of human behavior that is more efficient than a description that cites micro physical properties. The more efficient description is the one offered at the level of folk psychology. It is only from the intentional stance that we can discern these real patterns. There may be several ways of interpreting an agent, each equally successful, but this fact does not mean that these patterns are any less real. We can extend Dennett’s point to organizations. The interpretative stance we take towards organizations allows us to discern real patterns of social behavior, patterns that are missed if one attempts to explain the social world by appealing only to individual intentional states. Interpretationism, then, allows us to be ‘mild and intermediate” realists about the intentional states of organizations.12 We need not posit a group mind or consciousness in order for our 11 See, for instance, Fodor (1985), and Searle (1992) 12 Interpretationism can take on an even more realist face when coupled with broad func-

Challenging Epistemic Individualism

95

ascriptions of collective intentionality to have legitimacy nor do we have to think of these ascriptions as mere metaphors. Organizations and individuals really have beliefs but their ontological status is more akin to centers of gravity then tables and chairs. A second objection involves the distinction between acceptance and belief. Brad Wray (2000), Anthonie Meijers (this volume), and to some extent Raimo Tuomela (1995) have argued that given the differences between belief and acceptance, groups cannot be believers. Groups accept propositions but do not believe them. For instance, in “Collective Belief and Acceptance”(2000) Wray draws on the work of Stalnaker (1984), Cohen (1992), and Bratman (1999), and suggests that belief and acceptance differ in the following respects: 1. You can accept things that you do not believe, whereas you cannot believe something that you do not accept. 2. We often accept a proposition in order to fulfill some goal. We might accept a certain proposition, for instance, in order to test some scientific theory. According to Wray, acceptance is tailored for our purposes, but belief is not like that. 3. Belief is a disposition to feel that something is true. 4. Acceptance is voluntary, whereas belief is not. Given these differences, Wray argues that group belief is a species of acceptance. When a group believes something it is because the group adopts a certain position. They do so voluntarily and for the sake of pursuing some goal. Consider Wray’s example: The scientists working in the laboratory together first set their goal of working together on a research project. This shared goal then leads them to adopt the view that whatever sub-field they choose to work in is a subject worthy of investigation. There are probably a number of sub-fields that a groups of scientists working in a lab together could pursue research in, but a conscious choice must be made. (Wray 2000: 7) Wray also suggests that the ways in which we attempt to change the groups view reveals that we are dealing with the phenomenon of acceptance rather than belief. According to Wray, beliefs are typically changed by eroding the subjects disposition to feel that what they believe is true. This is done by providing counter evidence. But when we try to persuade a group to change their tionalism. This is the version of interpretationism I develop and defend in (2002c) and one that has been developed and defended by Pettit (1993).

96

Deborah Perron Tollefsen

view, we usually do so by showing them that the view is counterproductive to achieving their goals. We might also appeal to pragmatic, ethical and political concerns. Persuading groups to change their views, according to Wray, involves appealing to extra-veristic concerns. Truth and falsity is not an issue. Indeed, Wray goes so far as to say that most groups should not be judged by epistemic standards at all but by practical standards. There is much that can be said in response to this line of argument.13 The first thing to note is that even on a hard-core realist view of intentionality like that espoused by Jerry Fodor, belief is primary. One cannot have desires, intentions, etc. if one does not also have beliefs. Davidson describes the primacy of belief in the following way: “… belief is central to all kinds of thought. If someone is glad that, or notices that, or remembers that, or knows that, the gun is loaded, then he must believe that the gun is loaded. Even to wonder whether the gun is loaded, or to speculate on the possibility that the gun is loaded, requires the belief, for example, that a gun is a weapon, that it is a more or less enduring physical object, and so on.” (1991: 364) Belief is a prerequisite for having any other intentional states. If groups are not believers, then they are not intentional agents at all, not even acceptors of propositions. Now, this is not a problem for individualists but it is a problem for Wray and others who want to acknowledge some intentional phenomenon at the group level. So the need to respond to Wray’s argument is pressing. Lets begin by focusing on the issue of voluntarism. The other features Wray identifies as associated with the notion of acceptance rest on the issue of voluntarism so if we can deal with this issue we will have made some headway.14 First, we need to establish what it means to say that belief is involuntary. The idea is that an individual cannot simply will herself to believe something. I may want to believe that God exists, for instance, because I believe it will make me happy to be a theist. But just wanting to believe is not enough to make me believe. What may be enough, however, is providing evidence that challenges the atheistic or agnostic beliefs I now hold. I might solicit the testimony of others, read the Medieval proofs for the existence of God, talk with scientists who claim to have found the divine in quantum physics, etc. Providing 13 For a full response to the Wray, Meijer, Tuomela position see Tollefsen (2002c). 14 The idea that you can accept things that you do not believe and the idea that acceptance is essentially tied to the pursuit of some goal rest on the notion that belief is involuntary. If belief were voluntary one could believe something they did not accept, and believe something in order to pursue a goal.

Challenging Epistemic Individualism

97

myself with counter evidence may induce a change of belief. So in this respect my control over my beliefs is somewhat voluntary. I can control the evidence and the epistemic practices that may induce a change of belief. Now, there seems to be a weak and a strong version of the claim that belief is involuntary. The strong claim is that it is logically impossible to will to believe. There is something about the nature of belief that makes it impossible. The weak claim is that it is not possible for human beings to will to believe – its metaphysically or empirically impossible to will to believe. On the weaker claim there is something about the nature of human beings, rather than the nature of belief, that makes it impossible.15 Here is a plausible argument for the weaker claim: Try as hard as you can to will to believe that p and you will find its not possible. This is a simple but surprisingly convincing argument. Unfortunately, there is no simple and convincing argument for the stronger claim. Indeed, though there have been attempts,16 the stronger claim remains unsupported. Cohen can be viewed as providing an argument for the stronger claim. According to Cohen, belief is a disposition to feel that p is true. But this simply defines belief in a way that excludes it from being voluntary and thus begs the question with respect to this issue. Obviously if belief is a disposition to feel that p it can’t be voluntary. We can’t make ourselves feel certain ways. But there are serious problems with defining belief in such phenomenological terms. What does one feel when they feel that p is true?17 There seems to be no feelings associated with my belief that today is Monday. Its being Monday may give rise to feelings – disappointment, anxiousness, stress, etc. But the belief itself seems to have no associated phenomenology. Without a defense of the stronger claim, then, I don’t see a problem for groups having beliefs. It may be true that as individual human subjects we are incapable of willing to believe that p. We are limited in that way. But plural subjects are different sorts of subjects and may be able to will to believe. Plural subjects have control over their doxastic states. The same control they exhibit 15 See Alston (1989) for a discussion of the weak and strong version of the claim that belief is involuntary. 16 See for instance Bernard Williams (1972). 17 Perhaps Cohen has something like the following in mind, to believe that p is to assign p a high probability of being true. If this is so it is not clear what the difference is between acceptance and belief. When I accept, for the sake of argument, that God exists but do not believe that God exists, I am assigning the proposition “God exists” a certain probability of being true. It is, though, a very low probability but high enough for me to consider the consequences that follow from it.

98

Deborah Perron Tollefsen

when they choose to accept a proposition. This sort of control allows them to believe voluntarily that p. The difference lies in the subjects, not the nature of belief. Thus, I see no reason to think that groups cannot be believers. A final objection can be formulated in the following way:18 (1) If interpretationism is correct then organizations really have intentional states. (2) But it is ridiculous to think that organizations really have intentional states. (3) Hence, interpretationism must be false. According to this objection, my thesis constitutes a reductio of interpretationism. This argument loses its force when we uncover some of the presuppositions that motivate premise two. The idea that organizations could have beliefs and desires is often rejected because organizations do not have minds or brains. But the possibility of group intentionality is not so easily dismissed. Whether something has a mind depends on whether it has intentional states and not the other way around.19 Further, most philosophers of mind now reject any crude type-identity theory so the fact that organizations lack grey matter is not evidence that they lack intentional states. The fear of group minds, however, is not the only thing motivating premise two. Traditionally, intentionality has been linked to notions like that of a first person perspective, consciousness, and self-awareness and these concepts just do not seem applicable to organizations. It is the fact that groups seem to lack these capacities and properties, rather than simply the absence of a mind or brain, that motivates skepticism with respect to group intentionality. According to some, groups do not seem to have the features necessary for intentionality. Hence, they cannot be true believers. We have seen, however, that organizations do share one important feature with individuals. They can, under certain conditions, form a rational point of view. This is not a first person singular perspective but a first person plural perspective. But what about consciousness and self-awareness? The fact that organizations lack consciousness and self-awareness need not lead to skepticism about collective belief. There is another conclusion that can be drawn. Consciousness and self-awareness are plausibly thought of as properties of individual subjects. Because groups do not have these properties the conclusion one should draw is that groups do not have the features necessary for 18 Clarke (1992) considers this argument as well. His response to it differs from my own. 19 A similar point is made by David Velleman in “How to Share an Intention” (1997).

Challenging Epistemic Individualism

99

individual subjecthood. This, of course, is not a strange conclusion to draw. Groups are not individual subjects, they are groups!20 Note that this does not rule out the possibility that groups have the features necessary for intentional agency. If the differences lie in the nature of the subjects rather than in the nature of intentionality, there is reason to think that groups can be intentional subjects, just subjects of a different kind. The fact that my version of interpretationism acknowledges certain groups as true believers should not, then, be seen as a reductio. On the contrary, it should be seen as evidence of its strength as a theory of intentionality. It provides a univocal concept of intentionality that is applicable to different kinds of subjects.

II Acknowledging that certain groups are intentional agents puts epistemic individualism in an uncomfortable position but the challenge is not yet complete. Epistemic agents are cognizers. They perceive, the infer, they remember. They engage in cognitive processes. Do groups engage in these processes? We have already seen that groups deliberate. They consider consequences, establish means to ends, change their attitudes, etc. But a skeptic may claim that all this amounts to is individual cognition in a social or group setting. Cognition, asserts the individualist, is not a group property. But recent research in social psychology and related fields strongly suggests otherwise. The idea of group cognition or group minds was prevalent in the early days of social psychology. The earliest texts in social psychology, Le Bon’s (1896) Psychology of Crowds and McDougal’s (1920) The Group Mind, espoused the view that group behavior could not fully be understood by studying individual behavior. Instead, one must appeal to group or aggregate level properties and forces. To this extent they were in agreement with sociologists and social philosophers like Mead (1934) and Durkheim (1953), (1982). The development of behaviorism all but banished the idea of the group mind, however. Although interest in group level phenomenon has waxed and waned (see Steiner, 1974, 1986; Moreland, Hogg, and Hains, 1994 for reviews) the focus has been primarily on individual cognition in a social or group context. Like social epistemology, social psychology has been individualistic in the sense that the subject of study and assessment has remained the individual. 20 In Gilbert’s words, they are plural subjects.

100

Deborah Perron Tollefsen

Social psychology, however, has witnessed a resurgence of interest and research on cognition at the group level. The backdrop for much of the research is artificial intelligence. For instance, some social psychologists and organizational theorists have adopted the use of connectionist networks to make sense of group cognition. A connectionist network is a computational model that involves the synchronous parallel processing among many interrelated unreliable and/or simple processing units. The basic idea is that knowledge in very large networks of very simple processing units resides in patterns of connections, not in individuated local symbols. Any units activity is dependent upon the activity of neighboring units, connected to it by inhibitory or excitatory links whose strength can vary according to design and/or learning. In a connectionist network any one unit can represent several different concepts and the same concept in different context may activate a slightly different network of units. Edwin Hutchins has used connectionist networks such as the “constraint satisfaction network” to model how groups form interpretations. A constraint satisfaction network is a network in which each unit represents a hypothesis and in which each connection between the units represents constraints among the hypotheses. For instance, if feature B is expected to be present whenever feature A is, there will be a positive connection from the unit that represents the hypothesis that A is present to the unit the represents the hypothesis that B is present. If constraints are weak then the weights assigned to the connections will be small. If the constraints are strong the weights assigned to the connections will be large. “Similarly, the inputs to such a network can also be thought of as constraints. A positive input to a particular unit means that there is evidence from the outside that the relevant feature is present. A negative input means that there is evidence from the outside that feature is not present.” (Rummelhart, et al. 1986: 9). Each unit will adjust its activation on the basis of the activation of the units connected to it and the strengths of the connections to those units. The network will then eventually settle into a state in which as many constraints are met as possible. A constraint satisfaction network models, among other things, the way in which a person might arrive at an interpretation or judgement. Hutchins’ work reveals how the constraint satisfaction network model can be used to understand the way that groups arrive at interpretations or judgements. A group level constraint satisfaction network will involve many sub networks. These subnetworks represent individuals in the group. The individuals are then connected in various ways and these connections are

Challenging Epistemic Individualism

101

given weights depending on whether the connection is one that inhibits or excites. The connections between individual networks represent communication between individuals in the group. There are several communication parameters including: the pattern of interconnections among networks (who talks with whom), the pattern of interconnectivity among the units of communicating nets (what they talk about), the strengths of connections between the nets that communicate (how persuasively they are) and the time course of communication (when they communicate). Altering these patterns have significant consequences for how quickly and efficiently groups arrive at interpretations. In “The Social Organization of Distributed Cognition” (1991) Hutchins argues that confirmation bias, the propensity to affirm prior interpretations and to discount, ignore, and reinterpret counter evidence to an already formed interpretation, occurs at the level of groups. But by manipulating the communication parameters his connectionist network models have shown that confirmation bias can be mitigated. For instance, when individual networks have overlapping rather than mutually exclusive task knowledge, they have less of tendency to ignore counter evidence. When individuals are aware of each other’s expertise and their responsibilities within the group there is a less of a tendency to stick to an interpretation and more discussion of alternative interpretations.21 Another line of research that supports the view that cognition can take place outside the individual mind, is the work on collective memory. Although memory is usually viewed as a process that occurs within individuals, recent research by social psychologists strongly suggests that memory is also a process found in groups. One of the most intriguing and well-developed accounts of collective memory can be found in the work of Daniel Wegner (1987), (1995). At the individual level memory involves three different stages: encoding, storage, and retrieval. Information is encoded and stored in long or short term memory so that it may be retrieved at a later time. Items are not stored independently. Whole sentences may be stored or connected with other items. Bananas, for instance, may be stored with the color yellow. Retrieval of one 21 For other examples of connectionism applied to the group level see Kashima et al, (2001). Connectionism as a theory of the architecture of the human mind has many critics. My discussion remains agnostic with respect to that debate. If one is not sympathetic to connectionism, computational and information processing models of mind have also been extended to groups. For a review of research on the extension of these models to groups see Hinsz, et al. (1997).

102

Deborah Perron Tollefsen

item is often accompanied by retrieval of many other items. Recalling that I have a dinner party to go to, for instance, might also retrieve the information that I need to get a sitter. Memory theorists have also introduced the notion of a metamemory (Flavell & Wellman 1977). Metamemory refers to the beliefs people have about their own memory facilities. Consider a case of one person who reads a text and reports that she knows and remembers what the text says. Another person reads the same text and says that she does not remember what the text says. The two people may have the same pattern of retrieval when tested, however, and so they differ in what they meant by saying they “knew” the material. This is a matter of memory about memory. As Wegner points out metamemory is very useful. If we know how much we have to study before something if fully encoded, for example, we are in a better position for retrieval than someone who has little inkling of how much or what kind of studying to do. In essence, a person with a strong metamemory is able to take full advantage of whatever memory capacities he or she has. Wegner has argued convincingly that certain types of groups engage in processes that are identical to the stages of memory found at the individual level. At the individual level information is stored both internally and externally. Indeed, many of our everyday activities and tasks would be much more difficult if we did not rely on external storage. “Our walls are filled with books, our file cabinets with papers, our notebooks with jottings, our homes with artifacts and souvenirs, our floppy disks with data records, and at times, our palms with scribbled answers to a test. Quite simply, we seem to record as much outside our minds as within them” (Wegner 1987: 187). It is important to note that external storage is not simply an aid in retrieving information that is stored internally. Although we might leave a note on the fridge that says ‘garbage’ in order to aid us in retrieving the information that today is the day the garbage must go out, external storage is not simply about triggering retrieval. It is often the central storage area for large bodies of information. “The scrawlings one may make in a diary or daily log, for example, typically become the only record of many of the day’s mundane activities. Internal storage of many details is rather sketchy, and although we may be able to reconstruct our day in general without recourse to a diary, the external source provides some with startling reminders” (Wegner 1987: 188). Externally stored information is retrievable only if we know where to retrieve it. The successful retrieval of something from external storage requires

Challenging Epistemic Individualism

103

that we encode a retrieval cue or label for the item and a notion of the location of the item. For instance, I will not be able to retrieve the things my son said when he was two unless I encode and store internally the information that I wrote down these things in the leather journal and where that journal is at the present time. The necessity of having a label and information about location suggests the external storage is much different from internal storage. But the notions of a memory item, its label, and location have analogs at the level of individual memory. We can retrieve items from internal storage only if they are labeled as well. I cannot recall Mary Lou’s phone number unless I label it as such. The cues or labels are often vary broad. I may be reminded of the number simply by thinking of Mary Lou rather than having the label “Mary Lou’s number” in mind. And Although we do not need to know the location of information in order to retrieve the information from internal storage, we do have a type of location information in the form of meta memories we have about our own memory capacities. We have opinions about what we know and do not know, and these allow us to judge whether an item or set of them is to be found in our own memories. Using this facility we can report, for instance, that we know Rudy’s number without even reviewing it; or, without trying to reproduce the number, we can assert that we do not have it. Such judgments of location allow us to make both general and specific assessments of our internal information stores (Wegner 1987: 188). The processes of encoding, retrieval, and storage, then, have both internal and external manifestations. We can encode and store information either internally or externally. Successful encoding of either type requires that we label the information or associate it with some cue and that the label itself be encoded and stored internally. In the case of internal encoding, the item itself must be encoded with the label. In the case of external encoding we need not encode the item itself we need only to internally encode the item label and the location of the item. For instance, I need not encode the full text of Wegner’s article. I can encode his views externally by writing a set of notes and encoding internally a label for these notes and the location of the notes. The notebook serves as an external storage area. Just as notebooks, computers, journals, etc. can be external storage areas, so too can people. I might tell my husband that I have an appointment on Tuesday knowing that this information will be retrievable on Monday night. I have access to information in his memory in virtue of knowing that he is the loca-

104

Deborah Perron Tollefsen

tion of an item with a certain label. As Wegner suggests this interdependence often produces a “knowledge-holding” system that is larger and more complex than either of the individuals’ own memory. Each individual may know the system from one perspective-having in internal storage many items, labels, and locations, and knowing that the locations are in the other’s memory. However, this individual usually will not be aware of the complementary location information held internally be the other, information that the other uses to keep track of what the individual knows. The transactive system, in short, is more than its individual component systems. (Wegner 1987: 190) A transactive memory system involves the operation of individual memory systems and the interaction between them. This system goes through the same stages that occur at the individual level: encoding, storage, and retrieval. Encoding at the collective level occurs when members discuss information and determine the location where it will be stored in the group and in what form it will be stored. Transactive encoding sometimes takes the simple form of direct instruction for one group member to encode information internally (e.g. “Lulu, remember this number”), but more often involves complex negotations regarding the common labels that should be assigned to items (e.g., “What was that?”), the matter of responsibilites for internal storage (e.g., Isn’t this your bailiwick?”), the preferred locations of items (e.g. “I’ll take care of that”), and the like. In this process, the very nature of incoming information can be changed, translated into a form that the group can store. (Wegner 1987: 190) Retrieval involves identification of the location of the information. Retrieval is transactive when the person who holds an item internally is not the person who is asked to retrieve it. A professor, for instance, may ask a colleague who she knows is knowledgable about the work of P. F. Strawson, to recall the name of an early article. If the colleague retrieves the information internally and passes it along then transactive retrieval is successful. Often times, however, retrieval will involve more complex interactions. The colleague may not know the name of the article but may know someone who does. This in turn will require interaction with another colleague and he may know only part of the title. Full retrieval may come only after various sources are checked. At this point it seems that merely asking someone else a question is enough to establish a collective memory system. But transactive memory is much

Challenging Epistemic Individualism

105

more complex. First, individuals in a transactive memory system must know something about each other’s domains of expertise. They need to be aware of where information is stored and the storage capabilities of individuals in the group. Known experts in the group are held responsible for the encoding, storage, and retrieval of domain specific information. Other members contribute to the storage of information by directing new information to the appropriate expert. When there are no clear experts other ways of assigning responsibility for the information are used. The person who first introduces the information may, by default, be held responsible for encoding, storing and retrieving the new information. Overall, then, there are two sources of information people use to decide who is to be the acknowledged location of a set of labeled knowledge in the group. Individuals are seen as linked to knowledge on the basis of their personal expertise, or through circumstantial knowledge responsibility that accrues as a result of how the knowledge has been encountered by the group. (Wegner 1987: 192) And athough Wegner does not say this explicitly, transactive memory systems are formed through a process of interaction over time. If I ask a policeman for directions, we have not formed a transactive memory even though I have retrieved information from him. Individuals must participate in the encoding, storage, and retrieval of information. They must be involved in the allocation of information to specific experts, for instance, and in the determination of what information will be stored. Further transactive memory arises in a group when they are engaged in a common goal or share a common perspective; the purpose being to share cognitive responsibility and work. As a rule we can say that transactive memory systems are properties of groups. Since the policeman and I do not form a group it is not surprising that we do not form a transactive memory system. Transactive memory, then, provides us with another example of cognition at the group level. Transactive memory is group process not simply something individuals do in a group. It is not to be found in any particular individual but is a result of individual memory systems interacting in various ways. It involves the same stages of encoding, storage, and retrieval found at the individual level thus it is functionally equivalent to individual memory. The extension of theories of individual mind and cognition to groups is likely to be met with skepticism from the individualist. The extension of connectionism, for instance, may seem to be unjustified given connectionist’s claims that their work is grounded in the brain’s microanatomy. But connec-

106

Deborah Perron Tollefsen

tionists themselves admit that their models are metaphorical. They repeatedly refer to “neurological plausibility” (Quinlan 1991: 41), “neuron-like units” (Churchland 1993: 32) and “brain-style processing” (Rumelhart 1992: 69). Although neural networks are modeled after networks of neurons in the brain, connectionists are open to the possibility that neural networks are multiply realized. Indeed, they must be in order to avoid the charge of chauvinism. The use of individual cognitive models to understand group cognition is no more problematic than using computational models to understand the way that the human brain processes information. The computational model is in the first place a description of how computers process information. It has been adopted in the last 50 years or so as a model for understanding how humans think and process information. The description of how a computational machine processes information is at a very abstract level of description and these models were adopted, in part, because they allowed for multiple realizability. As in the case of transactive memory, there are functional analogs of individual level processing at the collective level. It may have seemed odd at the infancy of the computational approach to mind to think of the human mind as a computer but the approach has been explanatorily fruitful. Likewise, it may seem odd at first to think of certain groups as connectionist networks or information processing units but the approach has proven to be explanatorily fruitful. The research from social psychology and related fields is not, of course, conclusive. We need to explore more closely the functional analogs of cognition at the group level and the relationship between group and individual cognition. But this research coupled with the argument in section I does challenge the assumption that “knowers are individuals, and knowledge is generated by mental processes and lodged in the mind-brain.” (Goldman1987: 109) One can no longer simply assume that cognition only occurs at the individual level. Individualists will have to show that the extension of theories of individual mind and cognition to groups is unjustified. And they will have to do so with out invoking physical differences. That is, the individualist cannot argue that cognition is a phenomenon that occurs only within the brain. This sort of chauvinism is indefensible given the possibility that cognition is multiply realized. In order to support epistemic individualism, one must argue that the nature of cognition itself rules out the possibility that it take place in groups. The burden of proof, then, has been shifted.

Challenging Epistemic Individualism

107

III If groups like organizations are true believers it is natural to ask whether they have knowledge. I do not have the space here to develop an account of group knowledge. Instead, in this section I focus on the issue of the justification of group belief. Frederick Schmitt (1994) has argued that reliabilist accounts of justification are the most likely candidates for extension to groups. I agree and in this section I respond to a criticism that, if sound, would cast doubt on this extension. Reliabilism is motivated in part by the intuition that one need not know or have access to the grounds for our beliefs in order for them to be justified and in order to have knowledge. We often forget, for instance, the reasons we have for our beliefs. Still we feel justified in our beliefs. Imagine a mathematician that works out an elaborate proof. Later he forgets the steps of the proof but remembers the conclusion. It seems right to say that the mathematician is still justified in believing the conclusion. Reliabilist accounts of justification attempt to preserve this intuition by accounting for justification in terms of the reliability of the process that produces the belief. What matters for justification, as the mathematician example suggests, is not the ability to cite or access one’s justification but whether the belief was produced or is sustained by a reliable processes or mechanisms (e.g. five senses, memory, introspection, testimony reports, and ability to make valid inferences.) Consider Goldman’s (1996) reliabilist account of justification: Agent A’s belief that p (Abp) is justified, if: 1. The belief that p is a result of a reliable process; 2. Given A’s relevant alternatives there are no other reliable or conditionally reliable processes that would lead A to believe not p and there is no other evidence that the subject possesses that would undermine the belief. In other words, A has no defeater for his or her belief. The reliability specified in the first condition is judged in terms of the processes’s ability to produce more true beliefs than false beliefs. As Goldman puts it, “a justificationally permitted process must be one that yields a high truth ration, higher – perhaps appreciably higher – than .50.” (Goldman 1986: 103.) And reliability is not judged simply in terms of how the process works in the actual world but in the set of normal worlds. This normal world chauvinism, as Goldman puts it, allows us to make sense of our intuition that if we were in a Cartesian demon world are beliefs would still be justified.

108

Deborah Perron Tollefsen

Suppose it turned out that the actual world is not as we think it is. Suppose it is a Cartesian demon-style world, in which we are radically deceived about ‘external’ things (though we are not deceived about the nature of our cognitive processes). This means that our cognitive processes are not reliable in the actual world. At least our perceptual processes are unreliable, and perhaps our inductive processes as well. It does not follow, on my theory, that our actualworld beliefs are unjustified. The cognitive processes we use may still be reliable in normal worlds, in world similar to the presumed actual world. (Goldman 1986: 108). Further, the reliability of a process should be construed globally rather than locally. Cognitive processes, such as perception and memory, are locally reliable when they are reliable in terms of what a cognizer believes at t, namely p. Processes are globally reliable when they are reliable not simply at particular time but also in regards to that cognizer’s belief-forming system over time. The second condition invokes the notion of defeasibility. Suppose, for example, that I hear from my next door neighbor that my car lights are on. My neighbor is normally a reliable source of information. I would not be justified in accepting that testimony, however if there is another process available to me that would defeat that testimony, such as looking out my window and seeing that my car lights are not on. Or suppose that I have the belief that my car lights are on and it has been produced by a reliable process (perception). But I also have the belief that there are other cars just like mine on the street, that my husband has my car this afternoon, etc. These beliefs act as defeaters for the belief that my car lights are on and thus I would not be justified in believing that my lights are on.22 A straightforward application of Goldman’s theory to groups would look like this: A group G is justified in believing p just in case: 1. The belief that p is a result of a reliable process; 2. Given G’s relevant alternatives there are no other reliable or conditionally 22 The issue of what counts as a relevant alternative is a difficult one. Since my aim is to extend an account of individual justification to groups, I do not have the space to develop an account of relevant alternatives. I will say, however, that it is not just a problem for Goldman but for any epistemology that appeals to relevant alternatives. Further, though there may be no hard and fast way of distinguishing between relevant and nonrelevant alternatives, intuition seems to be a fairly good guide. We can tell in standard cases what is relevant and what is not.

Challenging Epistemic Individualism

109

reliable processes that would lead G to believe not p and there is no other evidence that the group possesses that would undermine the group belief. In other words, G has no defeater for its belief. Given that there are group level cognitive processes, the extension of reliabilism seems to be unproblematic. Angelo Corlett (1996) has argued, however, that “there are problems with group decision making that make problematic any group-based account of social knowledge which is based on the reliability of group decision making.” (Corlett 1996: 33). The problem according to Corlett is that group decision making is inherently flawed and thus is not a reliable producer of true beliefs. Group decision making is flawed because groups affect the decision individual cognizers make. He discusses three ways in which the process of group deliberation is undermined: the group polarization effect, pressure toward group consensus (groupthink), and deindividuation. Group polarization occurs when members adopt a view that is more extreme than the view they held prior to group deliberation. The more extreme view may not be the view that is closest to the truth. Indeed, given the powers of persuasion found in some groups the extreme view may be the view that is most likely to be false. The problem of pressure toward group consensus occurs when their are pressures, either from inside or outside the group, to arrive at a decision quickly. Deviation from consensus is often given negative feedback or sanctioned in various ways. Individuals feel pressured to go along with the majority view. Consequently, a unanimous decision might be made in a timely manner but it may not be truth-conducive. The most familiar problem that arises in group decision making contexts is the phenomenon of Groupthink. In highly cohesive groups there is a tendency to ignore relevant evidence that goes against the current group belief or view. “Factors that influence and help create the groupthink problem include personality and cultural factors, the size and unanimity of the group, the degree to which group members are isolated from outside influences, and, again, a high degree of group cohesiveness.” (Corlett 1996: 34) Highly cohesive groups also promote the phenomenon of deindividuation. E. Diener describes deindividuation in the following manner: People who are deindividuated have lost self-awareness and their personal identity in a group situation. Because they are prevented by the situation by awareness of themselves as individuals and from attention to their own behavior, deindividuated person do not have capacity for self-regulation and the ability to plan for the future. Thus … they become more reactive to immediate stimuli and

110

Deborah Perron Tollefsen

emotions and are unresponsive to norms and to the long-term consequences of their behavior. (1980: 210) When deindividuation occurs members are less likely to respond to outside influences and alternative sources of evidence. Given these well-established psychological phenomenon, Corlett concludes that decision making in groups is unreliable and thus any account of group justification or knowledge in terms of the reliability of group belief forming processes is doomed to failure. If group decision making and by extension group processes like transactive memory are unreliable it would certainly challenge the extension of reliabilism to groups. But the existence of phenomenon like group think no more doom a reliabilist account of group justification than the existence of confirmation bias, stereotyping, emotions, etc. doom a reliabilist account of individual justification. Cognitive processes will be unreliable in certain circumstances. Perception will not produce more true beliefs than false beliefs if the perceptual system is placed in a dark environment. Inference will not reliably produce true beliefs if it is influenced by emotion rather than evidence. Memory will be unreliable if it is asked to recall only items that were stored a long time ago. But acknowledging this does not mean that our cognitive processes are unreliable. Likewise, acknowledging that in certain contexts group decision making is not likely to produce true beliefs does not mean that group decision making is unreliable. It simply means that the process needs to be assessed globally not locally. Corlett could strengthen his argument by establishing that group decision making is never reliable but this would be a difficult thing to establish especially since current research in social psychology has identified ways to mitigate the effects of group polarization, group think, and deindividuation. We have already seen in Hutchins work, for instance, that confirmation bias in a group can be mitigated by altering the lines of communication between individuals and by developing networks that have overlapping rather than mutually exclusive task knowledge. Confirmation bias at the group level can be seen as a form of group polarization or group think. Further, current research suggests that group polarization and group think are a function of unshared information. Unshared information is often information that would lead the group to reconsider their decision and consider alternative possibilities and interpretations. For various reasons individuals often withhold information that is relevant to group decision making and the information that is held in common is the basis of decision making. Sometimes there are in- group norms that discourage individuals from contributing

Challenging Epistemic Individualism

111

information that they have. The tendency of groups to focus on shared rather than unshared information, is referred to as the common knowledge effect. But the common knowledge effect and therefore, group polarization and group think, can be attenuated. Larson, Foster-Fishman, and Keys (1994) have shown that unshared information becomes more prevalent in groups that engage in extended discussions. Groups that were put under strict time constraints and pressured to come up with a decision quickly often exhibited group polarization. Groups in these conditions focus on fewer alternatives and place more emphasis on information that members share in common. Recent work by Sawyer (1997) and Sheffey, Tindale, and Scott (1989) has shown that when group members have access to informational files during discussion there is also more unshared information introduced into discussion. Further, Stasser (1992) found that when members knew which other members of the group had expertise in which specific knowledge domain, the amount of unshared information they actually discussed increased significantly. Deindividuation can also be mitigated by stressing individual autonomy within a group decision making scenario. By distributing tasks and identifying areas of expertise individuals maintain their own unique identities. If individuals contribute information that was not previously shared by other members the contribution allows them to be in control of the decision making in ways that allow them a level of autonomy. Research on transactive memory also supports the view that group level processes are reliable. Wegner’s early research focuses on intimate couples. Couples in this study had been seeing one another exclusively for at least 3 months and were ask to together remember 64 items of information. The items were taken from certain areas of expertise and embedded in a context that made it clear even to non-experts what area of expertise they represented. For instance, “The Kaypro II is a personal computer.” Each partner was asked to review the 64 items and identify in each case whether the item was more likely to be within their own domain of expertise, their partner’s, both their own and their partner’s, or neither. Each subject was then given four items at a time to study for 1 minute and then were instructed to pass them on to their partners who reviewed them for 30 seconds. Each partner was disadvantage (or advantaged) in terms of encoding time and these disadvantages (or advantages) were known to the other partner. The experiment then involved both the storage of information based on expertise and on the basis of circumstantial responsibility. If an item was viewed longer by one partner than that partner would take on responsibility for that item even if it was not in their domain of expertise.

112

Deborah Perron Tollefsen

“Items has been sorted by each participant into four categories of personal expertise (self, partner, both, or neither), and then were encountered under circumstance that led either self or partner to be more responsible for them – in that either self viewed the items longer than partner or vice versa.” (Wegner 1987: 193) The experiment yielded several interesting results. First, across the board subjects remembered items from areas that they had judged to be their own areas of expertise. This is in keeping with the theory of metamemory. People know which areas they are more familiar with and therefore which items will be more easily encoded, stored, and retrieved. A more interesting result was that subjects that were circumstantially responsible for a topic more often recalled items from that area when they believed their partners were not experts in that area. In other words, members of these couples accepted the responsibility for information placed upon them by their circumstance, even when they were not expert, but did so only when they knew their partner had no history of personal expertise in the area. This, then, is a case in which an individual’s internal memory retrieval is affected by transactive constraints – knowledge of what someone else’s memory can or cannot do. (Wegner 1987: 194) Individual memory recall of items actually improved when items were encoded and retrieved by the partners together. That is, the individual would have remembered less and been less accurate about the items had she not been forced by the social situation to take responsibility for certain items. Finally, the most relevant results for our purposes is the finding that group memory is generally better than individual memory. Partners were able to recall more information and to do so more accurately than individuals. In an early study of group memory Yuker (1955) had individuals recall a certain story they had read. He then placed them in small groups and had them discuss the story and generate group recall protocols as well. The group versions were more accurate than the average individual recall in 38 of 40 groups tested. This research has been confirmed by more recent experiments (Clark and Stephenson, 1989; Hartwick, Sheppard, and Davis, 1982; Hinsz, 1990; Volrath et al., 1989). Not only is transactive memory a reliable belief forming process it is more reliable then individual memory. Transactive memory is not, however, without its flaws. Information may be changed or altered. During retrieval group discussion may change the nature of the information or in the process of encoding the information may apply a

Challenging Epistemic Individualism

113

label that is misleading and may lead to the retrieval of information that does not have a correspondence to reality. This is particularly likely because individuals tend to encode items well when there is elaborate rehearsal – reviewing of the item’s organizational ties to other items. Group commentary on incoming memory items provides just such rehearsal, and for this reason, individuals are more inclined to remember items as discussed than items as perceived. The group-acknowledged label thus can become more than a tag, growing to provide the major portion of what is remembered of the item by all group members. The term “UFO”, for instance, might serve as a rallying point for a wide array of specific memories held by individuals who witnessed an event one evening. The label suggests a common experience and provides a common foundation for explanation and elaboration. “Without this label, individuals might variously speak of seeing, say ‘funny lights’ or a ‘bright object,’ but with the label, a core memory is formed that provides an interpretative scheme for many different items of information.” (Wegner 1987: 194) But these problems are present in individual memory as well. There is no guarantee that an item presented to an individual will be retrieved in the same form. Individual recall of certain events is notoriously unreliable. But this does not mean that memory at the individual level is unreliable. Nor does the occasional unreliability of collective memory mean that the process itself is unreliability. Reliability is judged globally, not locally and reliability, as Goldman notes, is a matter of degree. There are certain contexts in which memory will be more reliable and certain ways of improving the reliability of both individual and collective memory. Identifying those contexts and those methods is the purview of individual and social psychology. As Schmitt notes there are many questions raised by the extension of reliabilism. What I have tried to do in this section is eliminate one major obstacle to this extension. Corlett’s conclusion that group processes are unreliable is too hasty. At the individual level there will be psychological phenomenon and environmental and social contexts that influence the reliability of individual cognitive processes but this does not mean that these process are unreliable. It simply means that reliability must be judged globally and with reference to the set of normal worlds. Likewise, group decision making and other cognitive processes will be unreliable in certain circumstances but to the extent that phenomenon like group polarization can be mitigated there is no reason to think these processes are globally unreliable. Indeed, one would think that if this were true we would have eliminated group decision making,

114

Deborah Perron Tollefsen

for instance, a long time ago. On the contrary, a substantial portion of our everyday lives are spent engaged in joint or group decision making. ****** The idea that groups can be legitimate epistemic agents has never been taken seriously by contemporary epistemologists. But epistemologists can no longer ignore the possibility and plausibility of collective epistemic agents. If I am correct groups like organizations are literally intentional agents and the models of mind and cognition used to understand individual epistemic agents can be extended to groups. The extension of theories of mind and cognition to groups is not metaphorical or at least it is no more metaphorical than saying that the human mind is like a computer or that our mind is a network of interconnected “neural-like” units. To understand this it is helpful to keep in mind that even at the individual level virtually everything we know about cognition is inferred from observations of behavior. As Larson and Christensen point out, The structure of long-term memory, the abstractness and stability of cognitive schema, and the use of various decision rules are all inferred from the behavioral responses of experimental subjects. Indeed, we cannot know that individual level cognition has taken place at all except by inference from behavior. (1993: 7) In a sense, then, cognition is a theoretical construct developed to understand and make sense of individual action. Viewed in this light, it does not seem unreasonable to extend this set of explanatory constructs to make sense of group actions like group decision making and problem solving.

Acknowledgments The research for this paper was funded in part by a summer grant from the Ohio State Center for Cognitive Science. I am very grateful for their support. I would also like to thank Dr. Marilynn Brewer for directing my research o social cognition. Thank also to David Eng, Sondra Bacharach, and George Pappas for insightful comments on earlier drafts of this paper.

Challenging Epistemic Individualism

115

References Alston, W. Epistemic Justification. Ithaca, NY: Cornell University Press 1989. Bratman, M. (1999). “Practical Reason and Acceptance in a Context.” Faces of Intention. Cambridge: Cambridge University Press 1999, 15-34. Burge, T. “Reason and the First Person.” Knowing Our Own Minds. Edited by B. S. C. Wright and C. MacDonald. Oxford: Oxford University Press 1998, 243-70. Churchland, P.M. “A deeper unity: Some Feyerbendian themes in neurocomputational form.” Connectionism: Theory and Practice. Edited by S. Davis. New York: Oxford University Press, 1992, 30-50. Clark, A. “Beliefs and Desires Incorporated.” The Journal of Philosophy, XCI, 8, 1994, 404-25. Clark, N. K., & Stephenson, G.M. “Group remembering.” Psychology of Group Influence. Edited by P. Paulus. 2nd edition. Hillsdale, NJ. Erlbaum, 1989, 357-91. Cohen, L. J. An Essay on Belief and Acceptance. Oxford: Clarendon Press 1992. Corlett, A. Analyzing Social Knowledge. Maryland: Rowman and Littlefield, 1996. Davidson, D. “Actions, Reasons and Causes.” Journal of Philosophy, 60, 1963, 685-700. —. “Mental Events.” Experience and Theory. Edited by L. Foster and J. Swanson. London: Duckworth, 1970, 79-101. —. “Radical Interpretation.” Dialectica, 27, 1973, 313-28. —. “Rational Animals.” Dialectica, 36, 1982, 318-27. —. “Thought and Talk.” The Nature of Mind. Edited by D. Rosenthal, 1991, 363-71. —. “Thinking Causes.” Mental Causation. Edited by J. Heil and A. Mele. Oxford: Clarendon Press 1993, 3-17. Deiner, E. “Deindividuation: The Absence of Self-awareness and Self-regulation in Group Members.” Psychology of Group Influence. Edited by P. B. Paulus. Hillsdale, NJ: Lawrence Earlbaum Associates 1980. Dennett, D. “Intentional Systems.” Journal of Philosophy, 68, 1971, 87-106; repr. Dennett (1987a), 3-22. —. “True Believers: The Intentional Strategy and Why it Works.” Scientific Explanation. Edited by A. Heath. Oxford: Oxford University Press, 1979, 53-75. —. The Intentional Stance. Cambridge: MIT Press 1987a. —. “Instrumentalism Reconsidered.” Dennett (1987a), 69-81. —. “Real Patterns.” Journal of Philosophy, 1991, 27-51. Durkheim, E. The Rules of Social Method. New York: The Free Press 1982. —. Sociology and Philosophy. Glencoe, Illinois: The Free Press 1953. Elster, J. Explaining Technical Change. Cambridge: Cambridge University Press 1983. Flavell, J. H., and Wellman, H. M. “Metamemory.” Perspectives on the Development of Memory and Cognition. Edited by R. V. Kail and J. W. Hagen. Hillsdale, NJ: Erbaum, 1977, 3-33. Fodor, J. “Fodor’s Guide to Mental Representation.” Mind, XCIV, 1985, 76-100. French, P. Collective and Corporate Responsibility. New York: Columbia University Press 1984.

116

Deborah Perron Tollefsen

Gilbert, M. On Social Facts. New York: Routledge 1988. —. Living Together. Maryland: Rowman & Littlefield 1996. Goldman, A. Epistemology and Cognition. Cambridge, MA: Harvard University Press 1986. —. “Foundations of Social Epistemics.” Synthese, 73, 1987, 109-44. Hartwick, J. Sheppard, B. H., & Davis, J. H. (1982). “Group Remembering: Research and Implications.” Improving Group Decision Making in Organizations. Edited by R. A. Guzzo. San Diego, CA: Academic Press, 41-72. Hinsz, V. B., Tindale, R.S., and Vollrath, D.A. “The Emerging Conceptualization of Groups as Information Processors.” Psychological Bulletin, 121, 1997, 43-64. Hinsz, V. B. “Cognitive and Consensus Processes in Group Recognition Memory Performance.” Journal of Personality and Social Psychology, 59, 1990, 705-18. Hutchins, E. “The Social Organization of Distributed Cognition.” Perspectives on Socially Shared Cognition. Edited by Lauren Resnick, John Levine, and Stephanie Teasley. Washington, DC: American Psychological Association, 1991, 283-307. Kashima, Y., Kashima, E., & Aldridge, J. (2001) “Toward Cultural Dynamics of SelfConceptions.” Individual self, Relational Self, Collective Self. Edited by C. Sedikides and M. Brewer. Philadelphia: Psychology Press, 2001, 277-98. Larson, J. R., & Christensen, C. “Groups as Problem-Solving units: Toward a new Meaning of Social Cognition.” British Journal of Social Psychology, 32, 1993, 5-30. Larson, J. R., Foster-Fishman, P.G., and Keys, C.B. “Discussion of shared and unshared Information in Decision-making Groups.” Journal of Personality and Social Psychology, 67, 1994, 446-61. LeBon, G. The Crowd. London: T. Fisher Unwin 1896. (Originally published 1895 in French) Lehrer, K. “Personal and Social Knowledge.” Synthese, 73, 1987, 87-107. McDougal, W. The Group Mind. New York, London: G. P. Putnam’s Sons 1920. Mead, G. H. Mind, Self, and Society from the Standpoint of a Social Behaviorist. Chicago: University of Chicago Press 1984. Meijers, A. “Believing and Accepting as a Group.” In this Vol. Moreland, R. I., Hogg, M. A., and Hains, S.C. “Back to the Future: Social Psychological Research on Groups.” Journal of Experimental Social Psychology, 30, 1994, 527-55. Pettit, P. The Common Mind. Oxford: Oxford University Press 1993. —. “The Discursive Dilemma and Social Ontology.” Presented at the Erasmus School of Social Ontology, Rotterdam, The Netherlands 2000. —. “Groups with Minds of their Own.” 2001, unpublished. Popper, K. “Epistemology Without a Knowing Subject.” Objective Knowledge. Oxford: Clarendon Press 1972. Quinlan, P. Connectionism and Psychology. Chicago: University of Chicago Press 1991. Rovane, Carol. Bounds of Agency. New Jersey: Princeton University Press 1997. Rummelhart, D. E., Smolensky, P., McClelland, J. L., and Hinton, G. E. “Schemata and sequential Thought Processes in PDP Models.” Parallel distributed processing: Explorations in the microstructure of cognition: Vol.2 Psychological and biological mod-

Challenging Epistemic Individualism

117

els. Edited by J. L. McClelland, D. E. Rummelhart, and the PDP Group. Cambridge, MA: MIT Press, 1986, 7-57. —. “Towards a Microstructural Account of Human Reasoning.” Connectionism: Theory and Practice. Edited by S. Davis. New York: Oxford University Press, 1992, 69-83. Sawyer, J. E. (1997). “Information Sharing and Intergration in Multifunctional Decision making Groups.” Paper presented at the Society of Judgement and Decision Making Annual Meeting, Philadelphia, PA. Schmitt, F. “The Justification of Group Beliefs.” Socializing Epistemology. Edited by Frederick Schmitt. Maryland: Rowman & Littlefield, 1994, 257-88. Searle, J. The Rediscovery of the Mind. Cambridge: MIT Press 1992. Sheffey, S., Tindale, R.S., and Scott, L.A. “Information sharing and Group Decision Making.” Paper presented at the Midwestern Psychological Association annual convention, Chicago, IL 1989. Stalnaker, R. Inquiry. Cambridge: MIT Press 1984. Stasser, G. “Pooling of unshared Information during Group Discussion. Group processes and productivity. Edited by S. Worchel, W. Wood, and J. Simpson. Newbury Park, CA: Sage, 1992, 48-57. Steiner, I. D. “Whatever happened to the Group in Social Psychology?” Journal of Experimental social psychology, 10, 1974, 94-108. —. “Paradigms and groups.” Advances in experimental social psychology. Edited by L. Berkowitz. Vol 19. New York: Academic Press, 1986, 251-89. Tollefsen, D. “Collective Intentionality and the Social Sciences.” Philosophy of the Social Sciences, 2002a, forthcoming. —. “Organizations as True Believers.” Journal of Social Philosophy, 2002b, forthcoming. —. “Making Sense of Groups.” Dissertation. The Ohio State University 2002c. Tuomela, R. “Corporate Intention and Corporate Action.” Analyse und Kritik, 15, 1993, 11-21. —. The Importance of Us. Standford: Standford University Press 1995. Vellman, D. “How to Share an Intention.” Philosophy and Phenomenological Research, LVII, 1997, 29-50. Vollrath, D. A., Sheppard, B. H., Hinsz, V. B., and Davis, J. H. “Memory Performance by decision-making groups and individuals.” Organizational Behavior and Human Decision Processes, 43, 1989, 289-300. Wegner, D. M. “Transactive Memory: A contemporary Analysis of the group mind.” Theories of group behavior. Edited by B. Mullen, and G. R. Goethals. New York: Springer-Verlag, 1987, 185-208. — (1995) “A Computer Network Model of Human Transactive Memory.” Social Cognition, vol 13, 1995, 319-39. Williams, B. “Deciding to Believe.” Problems of the Self. Cambridge: Cambridge University Press 1972. Wray, B. “Collective Belief and Acceptance.” Synthese, 2000, 1-15. Yuker, H. E. “Group Atmosphere and Memory. Journal of Abnormal and Social Psychology, 51, 1955, 17-23.

118

Gerhard Preyer

From an Externalistic Point of View: Understanding the Social

Externalism, Events and Institutional Facts

119

120

Gerhard Preyer

From an Externalistic Point of View: Understanding the Social

121

From an Externalistic Point of View: Understanding the Social Gerhard Preyer Abstract The circle between belief and meaning has revolutionized our understanding of communication. In this article I will sketch from the triangulation model of radical interpretation an answer to the question: “How does semantics take effect in social science?”, that is, it is to show “How do we link the mental, language, and the social?” I resystemize the principle of charity by the principle of natural epistemic justice and complete triangulation with the social frame of reference. Therefore, to break into the circle between belief and meaning leads us to the structure of elementary systems of communication as the framework within which we ascribe attitudes and actions. In this framework, together with the reformulation of the task of RI, we find an answer of the Burge-problem that the intelligible redescription and explanation of behavior take also effects on social considerations. In the procedure of RI also the problem of background theories emerges, and we border on the epistemic capacity of speakers that limits our understanding in general. In the last step it is shown – looking back to the controversy between individualism and holism (collectivism) in social science – that group-predicates can not instantiate ontologically to individual entities.

J. R. Searle (1998) claims to answer the question: How do we link the mental (attitudes), language and the social? This is surely a fruitful question for interdisciplinary discovery in particular in the philosophy of social science. This problem is of a particular significance if we make the assumption that communication does not reach (directly) the mental (consciousness). I will not discuss Searle’s account here. My intention is to give a sketch how we link the mental, language and the social from the triangulation model of radical interpretation (RI).1 This shows us how semantics projects into social sciences. It was at first G. Macdonald and P. Pettit (1981) who have shown that radical interpretation (RI) as a semantics of natural language takes effect in the philosophy of social science: “We believe that 1 The paradigm of Davidson’s unified theory of meaning and action is continued from E. Lepore and K. Ludwig to an interpretive fulfilment theory applied to declarative and nondelarative sentences; Lepore and Ludwig 2002a, Ludwig 1997. On Davidson’s philosophy G. Preyer 2002b, 2000, Preyer and M. Roth, 1998, Preyer, F. Siebelt, A. Ulfig 1994. On Searle see Frank A. Hindriks in this vol.

122

Gerhard Preyer

the elaboration of Davidson’s understanding of semantics, the pursuit of the so-called Davidsonian program, has thrown new light on the old project of making sense of people, and that in this light we can gain fresh insight into the issues raised by the philosophy of social science.” (3) The theory of interpretation and communication has a mirror image feature because communication is to analyze not from the actor’s point of view, but – like language in the framework of RI – from the addressee’s side. Davidson’s epistemological turn of RI – the triangulation model or RI – is giving up traditional epistemology. His view of truth and meaning is a rejection of facts, the scheme-content distinction, representationalism (the myth of subjectivity) and sense-dates as foundation in epistemology (sensations connect the world and beliefs; classical empirism, phenomenalism; about that there is no difference to Putnam 1992).2 The first task is to analyze why the externalistic foundation of communication followed from RI has revolutionized our understanding of the social. But in the process of RI the problem of a correspondence of the background theories between speaker and interpreter emerges. The claim is to resystematize charity by the principle of natural epistemic justice, and to complete the triangulation model of RI with the social frame of reference. Thereby, it is to give a response to the Burge-problem without going along with social externalism in general. To handle the restrictions of RI we border on the epistemic capacity that limits our understanding in general.3 In this context we give a further turn to RI linking semantics and social science. This proposal shows us, why epistemology takes effect on social considerations. Yet the step to social epistemology is born from triangulation. The externalistic foundation of the social is another turn making social behavior intelligible as we know from the tradition of sociological thinking, for example, in M. Weber’s sinnverstehender sociology or in A. Schütz’s phenomenological approach of Der sinnhafte Aufbau der sozialen Welt because communication is based by the social frame of reference in triangulation.4 From 2 On a detailed examination of Davidson’s ontological and epistemological consequences of his truth-centered semantics see S. Neale 2001. 3 The epistemic capacity is a matter of (our) richness of expression of power leveling language that takes epistemic effects in interpretation. 4 The terminus externalism goes back to D. Armstrong 1993. On the classical debate on externalism and internalism (reliabilism ) see H. Kornblith (ed.) 2001. Reliabilism is not an internalism as a physicalistic non-relational account of mental states. Davidson’s externalism itself is to distinguish from this externalism as a foundationalism. The problem is to relate the system of beliefs to external circumstances not merely semantically but epistemologically. The certitude of beliefs of our own propositional attitudes is no foundation

From an Externalistic Point of View: Understanding the Social

123

there it follows that all communications are events which happen temporally and decay at the same moment. Therefore there is a fundamental functional imperative of social systems that they have structures that cope with such events. In the following of Popper’s critique on the historicism there have been extensive writings about the conceptualization of social science (J. W. Watkins 1959, 1973, J. Agassi 1960, S. Lukes 1960, M. Brodbeck 1966, M. Mandelbaum 1959, 1966 and others, on further research looking back to the new debate on collective agents beginning with M. Gilbert 1989 see P. Pettit 2001). From our contemporary point of view the controversy has led to clarifications on conceptualizations and explanations of social science, and in the end a unified theory of meaning and action establishes a conceptual and explanatory dualism and therefore a methodological separatism between (human) social and natural science because all intelligible redescription is founded by coherence, consistency and true-beliefs (yet this is not to be misunderstood because semantics works within both). Individualism in social sciences goes back to Schumpeter, Popper and Hayek. In the third step I am recalling an interesting question for us today: “What is the universe of discourse we presuppose, if we ascribe G-predicates (group-predicates)?” The result is that the opposition between individualism and holism (collectivism) has to be given up and this is a well-understood consequence of the controversy.

1. The Externalistic Foundation of Communication (a) How do we break into the Circle of Belief and Meaning? What is the link between the semantics for natural language born from RI and communication? Normally we make the assumption that the ability to master a language is that we are able to speak. In contrast to that, in Davidson’s theory of language, the RI is the basic concept of a language. Therefore the essential feature of language is that the speaker himself is an interpreter of another speaker (Davidson 1975: 157). To solve the task of RI a detailed knowledge of speaker’s intentions and beliefs does not count as evidence. All strategies to define or of knowledge. Therefore, for Davidson the foundation of interpretation, but not the foundation of knowledge, is that the interpreter has to treat first person authority of speakers (1982a:169, 175).

124

Gerhard Preyer

explain linguistic meaning on the basis of non-linguistic intentions, uses, purposes, functions, activities are not only difficult, but impossible: We generally find out exactly what someone wants, prefers or believes only by interpreting his speech. (Davidson 1980: 4) For Quine the token of linguistic behavior does not distinguish between the net-meaning and the beliefs of a speaker about the situations of utterances. This is the distinction between the meaning of uttered words, the indicated beliefs, and the consequences of such beliefs expressed in the speaker’s words (communicative intentions). In other words, meaning is not to select from the communicative content of utterances as a pure entitative meaning. Understanding linguistic behavior is therefore confronted with the circle between belief and meaning. The dilemma of interpretation following Quine is therefore: we cannot interpret linguistic activity without knowing what a speaker believes, and there is no way to found a theory of interpretation on a primary subsententional discovery of his beliefs and intentions; … we cannot infer the belief without knowing the meaning, and have no chance of inferring the meaning without the belief. (Davidson 1974a: 142, 144) The contribution of speaker intentions for RI is that we interpret the utterance, for example, as an act of assertion. But in the first place meaning is not a result caused from such intentions. How, then, do we cope with the interdependence between belief and meaning without resting on a knowledge of meanings or beliefs and intentions of a particular speaker? The attitude of holding a sentence true seems to be without any epistemic restrictions of an evidence of a theory of interpretation because this belief is to apply to all sentences and to all utterances, even though each utterance is to be interpreted in a different way. Such evidence allows us to construct an interpretative theory of truth, that is, a satisfaction-theory of belief and meaning. This is the hard core of Davidson’s semantics. This method is intended to solve the problem of the interdependence of belief and meaning by holding belief constant as far as possible while solving for meaning. This is accomplished by assigning truth conditions to alien sentences that make native speakers right when plausibly possible, according, of course, to our own view of what is right. (Davidson 1973a: 137) Such attitudes, they may be true or not, are always the same. They immediately correspond with the beliefs of speakers. On this basis the interpreter con-

From an Externalistic Point of View: Understanding the Social

125

structs a theory about the beliefs of the respective speaker and his verbal behavior, that is, from the interpreter’s point of view beliefs and desires are theoretical constructs. Yet a speaker can only hold a sentence true if he disposes about the concept of truth. It may be that a speaker has a false belief but this does not mean that a radical interpreter makes the assumption that most of his beliefs are false. It is this pattern that fixes the truth or falsehood of a single belief (statement). The attitude of holding a sentence true is a fundamental relationship between the beliefs of a speaker and the interpretation of his utterances. This explains the particular role of truth conditions – or fulfilment conditions in general – for RI because the interpreter can know the holding true of sentences from the speaker’s point of view without having a knowledge about the meaning of the occasional utterances or the particular beliefs in a given case. But a theory of interpretation “will not stand alone” because holding a sentence true can only interpret utterances if we find the intentions of the speaker that are connected with our ascription of his desires and actions (Davidson 1975: 162). Therefore, the unified theory of meaning and action is a “total theory” of an interpretation of action and the assignment of beliefs and desires, that is, a “composite theory” of both (Davidson 1975: 163). Therefore the total theory of behavior deals directly with the relationship between actions and propositional attitudes (Davidson 1974b: 234, he calls all such attitudes thoughts). The problem of evidence goes back to radical translation. For Quine such evidence is reaching our surfaces of the neutral unconceptualized information that are the pattern of sensory stimulations (proximal stimulus) caused by physical receptors. Quine has called such information evidence. Therefore this epistemology makes a fundamental distinction between scheme and our content of knowledge that is refuted as the third dogma of empiricism (on Quine’s naturalism P. K Moser and D. Yandell 2000). What is the connection between beliefs, their truth and their linguistic expressions? With this connection we find the framework to link the mental, language and communication (the social). The interaction explains – that is the claim –, why we apply the principle of compositionality on all levels of RI. But it leads us also to the individuation of the content of propositional attitudes by distal stimuli, consequently by things and events. Yet the truth-centered theory of RI borders on the problem that the semantical redescription as an identification of truth-condition presupposes the so-called referential transparence of semantical context. This leads us to the Burge-problem (1979) because we need an answer for the problem – in Burge’s counterfactual story – that the expres-

126

Gerhard Preyer

sion in a referential non-transparent position like, for example, “arthritis”, does not allow us to ascribe individual propositional attitudes.5 This is the link to partially connect semantics and social science. First we can clarify the above-mentioned connection by the circumstance that a true utterance depends on the concept of true belief. A belief may be idiosyncratic, but we cannot understand it just as a private state. We grasp what it means that someone has a belief, if we recognize beliefs in their role of interpreting linguistic behavior. Therefore, we can have beliefs if we are familiar with the concept of truth and therefore with the concept of error. Now we can see why from holding a sentence true, as evidence of RI, the epistemological step to triangulation follows: (The) main point is that our basic methodology for interpreting the words of others necessarily makes it the case that most of the time the simplest sentences which speakers hold true are true … For the sentences that express beliefs, and the beliefs themselves, are correctly understood to be about the public things and events that cause them, and so must be mainly veridical. (Davidson 1982a: 174) Yet this does not mean that a speaker has no false beliefs. In the contrary, the distinction between holding a sentence true and its being true gives error its point. A radical interpreter himself makes the assumption that most beliefs (statements) he ascribes to any speaker are true; but within this frame he fixes the possibility that single beliefs of the speaker could be false. Holding a sentence true as the typical sort of evidence of RI is the fundamental relationship 5 This was the level that emerged in our Frankfurt group in the mid 1980s, Michael Roth 1986: pp. 92-97. Michael has argued in particular that pairing the referential transparence of an individual utterance with the evidence of holding a sentence true does not lead to a right translation/interpretation by intelligible redescription in general. This is reasoned by Burge’s thought-experiment because it has shown that the ascription of attitudes is not based by the obvious caused reactions state from the radical interpreter in general. Yet the principle of epistemic justice gives us a partial solution of this problem because we apply the principle not only to the individual speaker but also within our social frame of reference. I do not discuss Burge’s thought-experiment in detail here. It is well-known. But it is to mention that in difference to Putnam’s “Twin-Earthlings” the physiological descriptions (conditions) in the two stories of the Burge-problem are the same. The difference in the stories is not a matter of the physiological conditions that cause the “pain”, but the point is that within the division of linguistic labor the reference of the expression “arthritis” is fixed. On Burge see Davidson 1987: 25-28, 32-33, 1990a: 198-204. The challenge of Burge’s thought-experiment is that he has argued that all materialistic identity theories are “rendered implausible” for non-individualism; on Putnam see Davidson 1987: 18-22, 2832.

From an Externalistic Point of View: Understanding the Social

127

between the beliefs of the speaker and the interpretation of their utterances (but the ascription of beliefs does not presuppose propositional knowledge from the speaker’s side). This explains the particular role of truth conditions in the procedure of RI because an interpreter himself knows the holding a sentence true without simultaneously knowing what the speaker actually means or which belief he utters in given circumstances. Yet the evidence for RI explains also why all beliefs are a network and not to grasp as single attitudes because a correct understanding of the speech, beliefs, desires, intentions, and other propositional attitudes of a person leads to the conclusion that most of a person’s beliefs must be true, and so there is a legitimate presumption that any one of them, if it coheres with most of the rest, is true. (Davidson 1983: 127-8) The role of evidence for RI and the unlimited application of the principle of charity are the guarantee to grasp the quantificatory structure of attitudes. Also it cannot be concluded from the evidence that the beliefs of a speaker and an interpreter are corresponding; and it can also not be excluded that there is a change of beliefs. Holding a sentence true enables us to interpret an utterance only if we are able to ascribe intentions that are connected with desires as well. The distal theory of meaning (reference) leads us to the new role of an interpreter. The frame of ascription of meaning to utterances is our evidence to explain and understand the given behavioral items. For Quine, the meaning of words is conditioned by our sensory stimuli and the evidence is given in the description of conditions of utterances in a stimulus-response scheme (see on Quine and Davidson in this respect R. F. Gibson 1994). The understanding of attitudes (the mental) is only possible on the level of linguistic behavior. But in difference to Quine a distal theory of meaning characterizes the role of the interpreter in a new way because a radical interpreter concludes the speaker’s attitudes from his own reaction to the behavior of the speaker. The new active role of the interpreter is the ground that the interpreter refers to the obvious causes that give the speaker’s beliefs their content. Therefore It is not the speaker who must perform the impossible feat of comparing his belief with reality; it is the interpreter who must take into account the causal interaction between world and speaker in order to find out what the speaker means, and hence what that speaker believes. (Davidson 1982a: 174) Communication and higher level of social systems are emerged only on this basic line of the same reactions of speaker and interpreter. I will call this prin-

128

Gerhard Preyer

ciple of RI – given by the new role of the interpreter – the principle of identity of reference of verbal reactions. Davidson calls that the principle of correspondence of true beliefs and their individuation. Yet, what is the epistemological role of the causation of attitudes? It is to emphasize that causal relations are not “relations” of con- or disconfirmation. Such relations do not build a block (building block theory). Causes are not propositions or beliefs, but events that cause us to believe something, that is, events itself are not the evidence but “they cause us to belief something”, and “it is the belief that is properly called the evidence, not the event” (173). The subject of a theory of interpretation is behaviorally given and the theory informs us what an utterance means, that is, enables us to ascribe propositional attitudes. Yet we need an epistemological answer for the individuation of entities of these attitudes (their content). This is the step from the truth-centered theory of interpretation to the distal theory of meaning and cognition: triangulation is the frame of reference of the ascription and individuation of the (objective) content of attitudes. RI claims to attribute attitudes because both beliefs and distal stimuli take effect in the process of RI in the way that they “take up the slack” between objective truth, holding a sentence true, and communication. This leads us to another understanding of communication. It seems that mental states (or ordinary propositional attitudes) cannot be individuated by both: inner states and external facts (social or otherwise) – the “inner” is identified with neuronal processes that happen in our body6. The expansion of RI by triangulation leads to the epistemological result to refute the myth of subjectivity: thoughts have no mental objects. Davidson, Burge, Putnam, Dennett, Fodor, Stich, Kaplan and Evans and others agree about that: “propositions can’t give the content of our thoughts and be subjectively assured” (Davidson 1987: 37). Meaning is not a mental reality; and in the first person authority we have no knowledge about ourself. Yet, why is the link between the mental and language also shown by triangulation? RI has a mental counterpart. We link both by the asymmetry of RI. The redescription of linguistic behavior by a radical interpreter makes the assump6 For the anomalous monism and the triangulation-model of RI both conditions are satisfied: mental states are ‘inner’ as being identical with states of the body and non-individualistic to individuate because we identify such states by causal relations to objects and events outside the subject (Davidson 1987). On the problem of this answer in principle see Rogler, Preyer (2001).

From an Externalistic Point of View: Understanding the Social

129

tion that the ascribed belief is coherent to the set of the speaker’s true beliefs (therefore there are no de re beliefs, on critique R. Schantz 1996: 140-42). But RI is to be characterized by an asymmetry: the speaker does not interpret himself. From his point of view there is no RI. He also does not compare his uttered sentences with something that is given outside of him in the world to determine their empirical content. A speaker can express the truth or falsehood of his beliefs with his uttered sentences only: “Our sentences provide the only measure of the mental.” (Davidson 1995: 6) Thus, we can infer what a speaker means and intends to do, but the speaker himself stays in a black box, or in other words, the non-transparence of all ascriptions of attitudes is not excluded in principle.7 RI and radical translation take place in the third person. This is a point constantly misunderstood. The third person approach of RI is the guarantee that a speaker does not have to know the sentences that the interpreter redescribes. If this were a requirement of RI, then persons would not be successful with their linguistic communication. It is an essential feature of language that this is not necessary. The first person authority is an unavoidable part of RI because without such authority the ascription of attitudes and their content would be empty. If we ascribe a belief to a speaker on the basis of evidence, we cannot avoid making the assumption that the speaker knows what he has said and what he believes. However, the speaker knows his own thoughts in a way no interpreter can know them. He knows these without any evidence. When we apply the procedures of RI it is obvious why the utterances of our simple sentences are normally true. It shows us also how the authority of first person is connected with the principle of the autonomy of meaning. The first person authority also gives us an answer of the problem of self-deception. An absolute self-deception is not possible. RI is basically an asymmetric situation between speaker and interpreter, but it is symmetrical in respect to the truth conditions of sentences that a 7 It may be that this is a little bit ‘bulky’ to grasp. In a reversal of Quine’s philosophy – from the ontological point of view reference is scrutable, from the radical translation point of view (epistemological, behavioral) it is not – : we are ontological black boxes, but not epistemologically, that is, if the ascription of attitudes works, the indeterminacy between the ascription of attitudes and the given interpretation of the speaker’s words is limited (also in case there are discrepancies), but we can not discover from the RI point of view the attitudes themselves because we conclude from our theory of interpretation what attitudes are. To be a black box ontologically takes also effect in rationalization behavior–the “climber-example” – because there are no ways to explain how our attitudes cause an action that may be also caused in another way (Davidson 1973b: 79).

130

Gerhard Preyer

speaker ascribes beliefs to himself and also the sentences that an interpreter ascribes propositional attitudes to the speaker. This explains the role of truth conditions of semantics for natural languages as a theory of individual speaker utterances. It also explains us why the paratactical account of indirect discourse and the concept of samesayer are parts of RI and its truth-centered theory of interpretation (on the paratactic account and mixed quotation see Cappelen and Lepore 1999 and Elugardo 1999). The evidence of the attitude of holding true directed toward sentences gives the theory of interpretation its quantificatory form. And this is just the case “… although most utterances are not concerned with truth, it is the pattern of sentences held true that gives sentences their meaning.” (Davidson 1975: 162) (b) The Principle of Epistemic Justice To answer the question “What is the guarantee of the independency of evidence of RI?” has shown its possibility. In other words, understanding linguistic behavior requires a unification because only listing individual cases gives us no generalized answer. Holding true of sentences is introduced as the guarantee for the independent availability of evidence. Without this assumption the solution of the task of RI could not begin. For RI charity is the bridging principle of grasping objective truth and interpersonal understanding. I will systemize it in the next steps. It consists of some assumptions (principles) that are combined in RI. I will name this systematization the principle of natural epistemic justice, that is, the speaker is right, so long as he is not shown the opposite (therefore, I take into account the Burge-problem). At the first glance this is not obvious from the evidence of RI, but it is a result of a resystematization of charity that takes effect in the intelligible redescription of behavior in general. In this way, we recognize how and why the individuation of the content of attitudes goes along with the solution of the task of RI. It must be emphasized that the principle of natural epistemic justice also gives the principle of charity another turn. Epistemic justice is a guarantee for a correspondence of the background theories between speaker and interpreter. Breaking into the circle between belief and meaning always means breaking into the background theories of the speaker. Yet from the line of communication and the role of the second person also the social frame of reference emerges within which we ascribe the local content of attitudes (on the local content see Bilgrami 1992). This turn is

From an Externalistic Point of View: Understanding the Social

131

in harmony with the inscrutability of reference and the indetermination of translation. 1. Principle of coherence: The interpreter optimizes the self-consistency of the speaker. If he did not, he could not understand him. This is an application of a two-value logic as a component of the theoretical framework of RI. Other versions of logic do not give us a powerful instrument for analyzing attitudes. The ascription of attitudes and the redescription of behavior as actions do not avoid the assumption that propositional attitudes cause an extensive consistent pattern of behavior. Consequently we presuppose a coherent (consistent) set of speaker dispositions and could only decide within this framework on the divergences in cases under study. The first point is to complete with the following point. 2. Principle of correspondence: The interpreter assumes that there is a correspondence between his and the speaker’s view of world. The occasion of successful interpretations and communications are evidence of a shared world view. The general features of language will give us the general features of reality. It is the claim of the truth-centered theory of RI to show this. The dependence of belief and meaning is concerned with three problems: the ascription of attitudes, the interpretation of sentences, and the individuation of the content of attitudes. The guarantee of the independence of evidence and the individuation of the content of attitudes is given by the application of two principles: the optimization of true beliefs and the identity of reference of caused verbal reactions by distal stimuli. a) Holding a sentence true as evidence of RI is a matter of the principle of optimizing true beliefs. RI applies a theory of interpretation, in the style of Tarski, relativized to speaker and time point. The formal feature is that the biconditionals are true and conclude from the axioms of the theory. Furthermore empirical sentences are to exclude that are coherent with the formal features but are incomprehensible making the behavior intelligible. It is obvious that the speaker holds true sentences in given circumstances. Therefore he optimizes a correspondence in respect to true beliefs, and he identifies such beliefs only within a right sort of pattern of beliefs. If we would not make this assumption, our discussion about error could make no sense. Someone cannot have a belief unless he understands the possibility of being mistaken, and this requires grasping the contrast between truth and error – true and false belief. But this contrast, I have argued, can emerge only in the context of interpretation, which alone

132

Gerhard Preyer

forces us to the idea of an objective, public truth. (Davidson 1975: 170) Therefore the radical interpreter is forced to make the following assumption: We cannot intelligibly attribute the thought that a piece of ice is melting to someone who does not have many true beliefs about the nature of ice, its physical properties connected with water, cold, solidity, and so forth. The one attribution rests on the supposition of many more – endlessly more. And among the beliefs we suppose a man to have, many must be true (in our view) if any are to be understood by us. The clarity and cogency of our attributions of attitude, motive and belief are proportionate, then, to the extent to which we find others consistent and correct. (Davidson 1974c: 302) b) The optimization of true beliefs is consistent with the principle of similar reactions. Both are systematically connected. Therefore the interpreter presupposes that the speaker is disposed to response in the same or a comparable way toward things and events like he himself does. The interpreter ascribes to the speaker true beliefs that are directly caused by distal stimuli. The principle of similar reactions is the guarantee of the public channel of the causation of beliefs, and therefore a condition of objective truth that comes into play for the ascription of attitudes in the triangle, that is, perceptual beliefs: the beliefs that are directly caused by what we see and hear and otherwise sense. These I hold to be in the main true because their content is, in effect, determined by what typically causes them. (Davidson 1999a: 18-19) The principle formulates an identity of reference of caused verbal reactions. This is presupposed for an application of truth conditions to linguistic behavior and for the ascription of attitudes. And this simultaneously frames the intelligible behavior in our given environment by triangulation as a “prelinguistic” and “precognitive” situation, that is, it is a matter of mutual (similar) reactions that are caused from the external world that guarantees the objectivity of interpretation.8 The similar reactions among humans and the 8 Triangulation “… is the result of a threefold interaction, an interaction which is twofold from the point of view of each of the two agents: each is interacting simultaneously with the world and with the other agent. To put this in a slightly different way, each creature learns to correlate the reactions of the other with changes or objects in the world to which it also reacts… . The triangle I have described stands for the simplest interpersonal situation. In it, two (or more) creatures each correlate their own reactions to external phenom-

From an Externalistic Point of View: Understanding the Social

133

external causation of perceived shared stimuli are necessary conditions to individuate the content of attitudes within the framework of the truth-centered theory of RI. Therefore, it follows that the “second person” is to be included in the solution of the task of interpretation. But the assumption of mutual standards makes no sense because there is a reciprocal interpretation only. The question is not to ask for mutual standards but for the objectivity of interpretation as a guarantee of successful interpretation. 3. Principle of continence: The interpreter assumes a coherence between propositional attitudes and intentional actions, that means: “One should prefer (act on) the judgement based on all considerations deemed relevant” (Davidson 1985: 350) or in other words: “perform the action judged best on the basis of all available relevant reasons.” (Davidson 1970: 41) Objective irrationality is an intra-inconsistency of otherwise consistent sets of beliefs. Ultimately we are unable to act against our strongest beliefs. Weakness of the will is in principle a borderline case although in ordinary life we are well-known with weak-willed people (Davidson 1970; 1985). It is the claim of the principle of continence to give a particular answer to the relationship between the intending and intentional actions. Forming an intention is an “all-out judgement” or “unconditional judgement” like “This action is desirable.” (Davidson 1978: 98) “Self-deception”, “overpowering desire”, “lack of imagination” and so on are faults in the building of such judgements. Prima facie judgments – judgments that are desirable in respect to certain attributes of an action – could be reasons for acting but an intentional action is caused by a stronger judgment. This is Davidson’s (final) explanation of the relationship between intentions and intentional actions (on the problems of Davidson’s way to this solution, see Preyer 2002b). Wants, desires, principles, prejudices, felt duties, and obligations are reasons for actions and intentions we express in prima facie judgments, but intentions and the judgments that cause an intentional action are a result of an all-out judgment: pure intending is an intention as an all-out judgement that is a result of the relevant beliefs of an agent. I come back to such judgments to find the relationship between attitudes of agents and social units. ena with the reactions of the other (or others). Once these correlations are set up, each creature is in a position to expect the eternal phenomenon when it perceives the associated reaction of the other. What introduces the possibility of error is the occasional failure of the expectation; the reactions do not correlate… . If this is right, then not only language, but thought itself, is necessary social.” (Davidson 1999b: 128-129, see also 1991b)

134

Gerhard Preyer

The entities of RI – the uninterpreted utterances of natural speakers – are given us directly as intelligible behavioral items in this framework.9 Therefore the theory of interpretation informs us what the utterance of speakers means, that is, it has to enable us to ascribe propositional attitudes and to individuate the content of such attitudes without an empirical researching of those attitudes, that is, “identifying the causes with the objects of his beliefs” (Davidson 1983: 152). Therefore, the distal theory of meaning and cognition does not need propositions as entities of propositional attitudes. Yet, holding true and the principle of identity of reference of behavioral reactions increasingly take an epistemic effect in the procedure of RI, and this results in the problem of the selection of truth (fulfilment)-conditions from the radical interpreter’s side. In general the assumption is to be made that the radical interpreter perceives that the speaker’s holding true is caused by the distal stimuli. Thereby this is strengthened because the interpreter constructs his theory on the speaker’s beliefs on the basis of holding true and by the obvious reactions caused by distal stimuli. From my point of view, it is essential therefore to ascribe to the speaker not only a holding true of sentences but also a specific knowledge. This explains us why RI requires a convergence of background theories. If an interpreter makes these assumptions for the solution of the task of RI, we have to consider an expanded frame of reference for RI. The task of RI is to solute without knowing something about the particular beliefs of the speaker, that is, “the theory describes conditions under which an utterance of a speaker is true, and so says nothing directly about what the speaker knows.” (Davidson 1990b: 312) From my view, we can circumvent epistemical restrictions as an option of the theory of interpretation only if we presuppose the background theory as well known. Yet we do not completely get out of background theories; even though it may be that such theories are difficult to explore. This shows that Davidson’s charity leads us to restrictions of making behavior intelligible in principle. In general, the radical interpreter has to 9 But from the triangulation model point of view we do not identify attitudes with behavioral items. “At one time behaviorism was invoked to show how it was possible for one person to know what was in another’s mind; behaviorism was then rejected, in part because it could not explain one of the most obvious aspects of mental states: the fact that they are in general known to the person who has them without appeal to behavioristic evidence. The recent fashion, though not strictly behavioristic, once more identifies mental states partly in terms of social and other external factors, thus making them to that extent publicly discoverable. But at the same time it reinstates the problem of accounting for first person authority.” Davidson 1987: 23. On holding true as intensional attitudes toward sentences and behaviorism see Davison 1979: 231.

From an Externalistic Point of View: Understanding the Social

135

project his analytical-empirical knowledge to the speaker (on a externalistic concept of narrow content Schantz 1996: 405-12). The interaction of the principles of RI means: coherence (consistency), true beliefs, the identity of reference in respect of caused reactions and relevant judgments are essential features of the network of our system of attitudes (we are not successful stating attitudes one by one). These are prerequisites of the emergence of an elementary social system of communication as the framework within which we ascribe attitudes (whatever its internal differentiation and organization may be). Such a system could not exist without the interaction of these principles. With the second and third point, triangulation is introduced because the content of beliefs is caused by things and events, and this is obvious in the triangle. Thereby the basic line of communication is found: Communication begins where causes converge: your utterance means what mine does if belief in its truth is systematically caused by the same events and objects. (Davidson 1983: 151) With that we give an answer also on the significance of the “second person” in the triangulation model of RI. We can ascribe meaning and attitudes only if people reciprocally interpret their linguistic behavior. The only way of knowing that the second apex of the triangle – the second creature or person – is reacting to the same object as oneself is to know that the other person has the some object in mind. But then the second person must know that the first person constitutes an apex of the same triangle another apex of which the second person occupies. (Davidson 1992: 121) This shows how and why the steps to triangulation are made through applying the principle of epistemic justice (I do not discuss the intra-theoretical step from the truth-centered theory of RI in detail here). But we also recognize why the conditions (assumptions) for RI are to be systematized by the principle of epistemic justice: The intelligible redescription of speech and actions cannot avoid presupposing that the speaker (actor) is largely right in respect to the truth of their beliefs and other fulfilment conditions. But this is not just a mere charity, we have to assume it: “We must, if we are to interpret at all.” (Davidson 1976: 290) This strategy allows us to pair the sentences of the speaker and his reactions with ours that we also hold true. The particular authority of the speaker and his background is the restriction of RI. The epistemic effects of the application of the principles shows us also that, if beliefs are a condition of the ascription of attitudes, then such beliefs are not only behavioral items. This is comparable with Davidson’s view:

136

Gerhard Preyer

There are conceptual ties between the attitudes and behavior which are sufficient, given enough information about actual and potential behavior, to allow correct inferences to the attitudes. (Davidson 1982b: 100) If the theoretical framework of a radical interpreter is presupposed as true for the speaker, then we have to specify for the ascription of attitudes also the epistemic restrictions from the speaker’s side. The application of the principle of epistemic justice leads us to the peculiarity of the evidence of RI. I agree with the paradigmatic inside of RI: it is just not possible to state at first what a speaker believes, desires, intends, and in the next step to give an answer to the question what the uttered words refer to. Therefore RI states the inscrutability of reference in a new way, because if we have collected all evidence for our theory of interpretation, it is not possible to separate the contributions of beliefs, desires, thoughts, actions for RI step by step: we need “total theories” (Davidson 1979: 241). The task is to give us a theoretically revised version of the inscrutability of reference as not to eliminate combinations of propositional attitudes. The principle of epistemic justice gives this version a further turn. But from my point of view, the principle is not an epistemic norm (see also D. K. Henderson 2000). We use it for descriptive and theoretical characterizations. But I think that we cannot naturalize epistemic predicates. (Here, I do not analyze the role of norms in communication. From my point of view, social norms are not the foundation of social systems but have their function at a derivative place. This does not mean that norms do not play a role in such systems. But this is not my subject here.) How far does RI revolutionize our understanding of communication? At first glance this is not obvious. The answer to this question leads us back to the circle between belief and meaning and shows us the new role of an interpreter that emerges from RI. The so-called “revolution” of our understanding of communication is comprehensible if we make clear three points that are fundamental for every communication: 1. Systems of communication can only exist if we – as participants – are able to distinguish between holding true and being true (and the same distinction for other fulfilment conditions); in general we must be able to evaluate propositional attitudes; 2. on the other hand, communication is only possible if the content of our beliefs is caused by distal stimuli and such causation has to be obvious in the triangle. 3. There is a social frame of reference within which we ascribe attitudes that cannot be disputed from the point of view of different observers. A. Avramides has summarized the problem of the individuation of propositional attitudes in respect to the

From an Externalistic Point of View: Understanding the Social

137

circle between belief and meaning in order to clarify why thoughts are dependent on communication: … in communication one speaker understands the words and sentences of another; that is, in communication we find ourselves in the position of interpreters of the words and sentences of another. As interpreters we are faced with the following problem: what a person means is in part a function of their belief. Yet we have just been told that what persons believe is a function of what their words and sentences mean. Our dilemma is how to break into this circle of belief and meaning. (1999: 146; see also K. Ludwig 1999: 41-44). Communications are an upshot of successful tries to communicate something, that is, to take an effect on the addressee’s side. Therefore, such actions are intentional or instrumental doings (see also M. Ulkan 1997). It is an essential feature of communication to decide about the continuation of the social intercourse. Without such a decision there is no communication. It is the distal stimulus that makes the intented effects transparent, and, if the shared reactions work, there is a frame of reference to understand and to continue the process of communication. This also explains the role of language in the universe of the social. It is a requirement of communication that the participants identify (reidentify) the same stimulus and the same communicative event. This is only possible if they can describe (express) these in a language (and there is a class of expressions, the singular terms, that are the guarantee to recognize that). On this level an elementary system of communication has emerged. What does the triangulation model of RI imply for the concept of language? Applying the strategy to cope with the circle of belief and meaning, we find a basic property of linguistic behavior. The role of parts (words, sentences) can only be understood “in the context of the “system (language)”. To understand a language means “translating it into our own system of concepts” (Davidson 1974b: 239). In the case of beliefs, desires, and actions we proceed in the same way. The correspondence of true beliefs and the shared reactions among different individual speakers are the prerequisites for a common language. “Beliefs cannot be ascertained in general without command of a man’s language; and we cannot master a man’s language without knowing much of what he believes.” (Davidson 1974b: 238) If this were not the case, then there would be neither a common language nor a public verification of the components 1 - 3 of the principle of epistemic justice. RI applies the truth conditions to the behavioral reactions of individual speakers. This is its fundamental procedure. For RI language always means

138

Gerhard Preyer

speaker’s language (idiolect theory). The correspondence in the use of language of individual speakers is a result of the development of similar reactions caused by stimulus-responses in different situations. This is a concept of language that is reasoned by a theory of RI and an externalism of triangulation. A common language or a linguistic community only exists if the speakers have the same theory of interpretation. It is a basic feature of linguistic behavior that the participants of the triangle can verify their occasional sentences (perceptual beliefs). This is an essential feature of human language. Until a base line has been established by communication with someone else, there is no point in saying a person’s… . words have a determinate content.”(Davidson 1991b: 213) “Still, we can ask why language is essential to thought. The reason, stated briefly, is that unless the base line of the triangle, the line between the two agents is strengthened to the point where it can implement the communication of propositional contents, there is no way the agents can make use of the triangular situation to form judgments about the world. Only when language is in place can creatures appreciate the concept of objective truth. (Davidson 1999b: 14) The triangulation model of RI also claims to answer the question of the function of communication in epistemology: We can only ascribe thoughts to creatures that are able to communicate. Therefore communication is the guarantee of objectivity: A non-communicating creature may be seen by us as responding to an objective world; but we are not justified in attributing thoughts about our world (or any other) to it. (Davidson 1991a: 201) This is the prerequisite for someone being right or wrong in his beliefs. With respect to this, the interpretation of speakers presupposes language, or depends on language. We can only attribute thoughts on the basis of given languages and no interpretation is language independent. But it may be that we have thoughts that remain solely our own, and we are unable to express them. Everyone knows a situation where he has a thought and it is difficult for him to express it in words. Thoughts and expressions in someone’s language are factually interdependent, but there is no reduction of one to the other. This also holds in the case of reciprocal reinforcement. In sum: we link attitudes, language and the social in the triangle from the line that beliefs and distal stimuli take effects in RI. Triangulation links the social directly with perceptual externalism, that is, the second person (the social) is located within the causal nexus. Yet the problem is not only the causation of

From an Externalistic Point of View: Understanding the Social

139

the content of our thoughts but it is also to find the correct sort of causal ascription and it is to be taken into play to re-interpret linguistic behavior, the epistemic restrictions (background theories). RI requires a social frame of reference of the ascription of attitudes that makes behavior comprehensible. With this step we know how to handle Burge’s problem (1979) that the autonomy of meaning does not correspond to an essential relationship between an individual belief and sentence meaning (on Burge’s and Kripke’s social externalism see also Bilgrami 1992). I make a distinction between such a “frame” and a semantics of community because it is not the autonomy of meaning that is the problem, but fixing reference of expressions within the epistemic differentiations in the social universe of discourse. I introduce the social frame of reference to continue RI as a feature of the ascription of comprehensible epistemic attitudes that take into account social considerations. And this leads us to the epistemic restrictions of RI, not only as of individual speakers but also as a part of the social frame. This is the link between semantics and social science.

2. The Extension of G-Predicates (a) The Meaning of G-Predicates We can conclude from the restrictions of RI that the ascription of attitudes requires a social frame of reference. This leads us back to the debate between methodological individualism versus holism; and to the problem of the ontology we presuppose for the ascription of group-predicates (G-predicates). I will show that the social frame of reference, the universe we presuppose, explain us the ascription of G-predicates. This universe is the membership condition in social systems. No P-predicate (predicates of individuals) can instantiate that. Now methodological individualism certainly does not prohibit attempts to explain the formation of psychological characteristics; it only requires that such explanations should in turn be individualistic, explaining the formation as a result of a series of conscious or unconscious responses by an individual to his changing situation. (J. W. Watkins 1959: 508) Social situations and events are, for this account, particular configurations of individuals, their dispositions, beliefs, situations of acting, and physical resources. Such individualism is committed to a theory of meaning that makes

140

Gerhard Preyer

the following assumption: every statement about the social is a statement about individuals; if this is not the case, the statement would not be meaningful. All social predicates refer to individuals, that is, all statements on social states of affairs are to translate, without any loss, to statements about individuals. For example, an army would be nothing more than the number of its soldiers, and the truth condition of all statements on an army were to reduce to statements on particular soldiers that are members of the respective army. Yet it should be emphasized that there are two versions of methodological individualism: 1. There are no undefined G-predicates; this is the conceptual theory of methodological individualism, and 2. the explained laws of groupbehavior are to reduce to individuals in principle (M. Brodbeck 1966: 325 shows that the first assumption does not imply the second). Firstly I will recall the difference between individualism and holism by the following comparison (Agassi 1960: 244 see also 246):

(a) Holism

(b) Individualism

1. Society is the ‚whole‘ which is more than its parts (holism). 2. ‚Society‘ affects the individuals‘ aims (collectivism).

Only individuals have aims and interests (individualism). The individual behaves in a way adequate to his aim, given his circum stances (rationality principle). The social set-up is changeable as a result of individuals‘ action (institutional reform).

3. The social set-up influences and constrains the individuals‘ behavior (institutional analysis).

To analyze social systems as an organism goes back to the 19th century’s antiquated concept. Mandelbaum argues, for example, that social concepts, such as “contract”, “marriage”, “company” and so on, are not defined by psychological concepts, therefore by predicates we apply to persons (P-predicates) (M. Mandelbaum 1966: 483). The meaning of expressions like “the President of the United States of America” or “an unmarried man” are determined by the constitution of a state or the structure of the kinship system (Mandelbaum 1959: 461). Therefore we have evidence against the assertion that all social phenomena have to be explained by attitudes, decisions and actions of individuals. The truth condition of, for example, “Peter signs the contract” is true if there is a

From an Externalistic Point of View: Understanding the Social

141

certain event that makes happen the personal signature of the contract is done and presupposes a social frame of reference and a particular knowledge to identify one’s signature to agree with a contract. Therefore such identification is a matter of specific conditions of fulfilment like the person who signs the contract is of age, he/she is able to keep the contractual duties, the contract does not offend common decency and so on. This argument against methodological individualism does not mean that social concepts do not have individuals as conditions of fulfilment. Normally we have no doubts that groups and organizations with their regularities, social norms, procedures and traditions face us from outside (on the theory of institutions see S. Lukes 1973). Yet the ontology of such entities is not obvious (on this problem E. Gellner 1959). This is also the case if the entities exist in a temporal continuation. We describe such entities as expectations, norms, regulations, procedures to coordinate actions and co-operation that are valid in groups, and the members of groups are more or less willingly to agree with these restrictions of their behavior. Yet such restrictions are also changeable. The framework to cope social change are “institutional reforms” as a result of actions from individual actors. It is not to be disputed that social change may be a result of individual actors; the reality of an institutional analysis, however, I will not discuss here. b) The Universe of G-Predicates What are the elements of the universe of discourse which we presuppose for the ascription of G-predicates? Are we forced to assume that the elements of the universe of the social are individuals? The use of descriptions of groups respectively, the extension of G-predicates, is not to explain by P-predicates, for example, “he is a member of the football team Frankfurter Eintracht” is not semantically equivalent to any Ppredicate. It is not possible to give a reduction by definition of such predicates to individual persons. Entities like social systems and organizations do not have the status of individualistic entities to which we attribute certain properties or deny them. Nevertheless, such entities have a temporal continuation, and we would not say that such entities do not exist. A part of the set of predicates which we ascribe to groups can be ascribed to individuals, for example the utterance “The football team shoots a goal” and its specification to the scorer, to “Peter Miller”, that is, the predicates have individuals and their status-positions as their conditions of fulfilment. For other predicates this is not the case, for example, 1. “has members”, the predicates

142

Gerhard Preyer

that 2. state a relationship between a group and their members, for example, “is superior of …” but also with the predicates that 3. state a particular feature of the group, for example, “the organization has existed for 100 years” (MacDonald, Pettit 1981: 118-19). Social systems consist of members. With respect to this, it seems to be a truism that the application of G-predicates refers to the members of such systems. The specification of P-predicates therefore have G-predicates and their specification to positions of membership (roles) as their counterparts, such as “ordinary member”, “chairman”, “employee”, and so on (the social positions are instances of information for the ascription of attitudes). The individualistic understanding of the social has asserted that we have to instantiate all grouppredicates to individual members of the given group, or respectively to the relationship of such individuals. Popper argues that institutions or associations are to be described as models of abstract relationships between individuals. Yet, against this proposal it is to argue that there is a mixed extension of the ascription of P- and G-predicates, for example; “Peter Miller, the dangerous forward of the Eintracht Frankfurt …”. Such extension has a divided reference. Yet the mixed extension does not decide on the ontological status of such predicates in principle (that is, the elements of the presupposed universe). A methodological individualist would not contest such mixed extension either. The so-called mix means: if, in one case, we apply G-predicates like, for example, “… forward of the Eintracht Frankfurt”, but not in another case, some predicates are associated with the reference to P-predicates, therefore in the first case we could instantiate G-predicates but not in the second – “dangerous” is a P-predicate – and quite on the contrary. Social systems are determinated by their conditions of membership and the positions of their members. The decision on membership is a Letztelement (ultimate element) of social systems. Such decisions are to be made not only in continuation in social systems, but social systems are emerged by decisionmaking on membership (on the primitive level of social evolution as structural selections of kinship systems). This is valid for social systems in general. This distinction is especially suitable to grasp the ontological commitments of Gpredicates, that is, the universe of discourse that is presupposed. We find the presupposed universe by answering the question “How do we specify P- to Gpredicates?”, for example, “Peter Miller, the gifted football player and forward of the Frankfurter Eintracht”. The extension of the predicates is fixed thereby if we determine the application of G-predicate with the distinction “member” versus “non-member”, “superior” versus “subordinate” and so on. The mem-

From an Externalistic Point of View: Understanding the Social

143

bership conditions are the elements of the presupposed universe of discourse for our instantiation of G-predicates: An institution cannot be identical with an individualistic entity because it is considered as remaining one and the same through a turnover of membership and such an assumption cannot be maintained of any individualistic entity. (122) The membership condition is the framework within which we ascribe groupattitudes (intentionality). The mixed-extension between the two sorts of predicates can also apply in other cases, such is, we use G-predicates on the basis of an incomplete list of Ppredicates and we cannot fix definitely the disjunction between the application of both sorts of predicates. We fix the mixed extension between the both sorts of predicates if we make the assumption that there are expectations in social systems or applications of procedures, for example, the regulation of the admission of new members, orders of the day and so on. The fulfilment of conditions of membership, expectations, or the conformity to regulations enable us to determine the mixed extension between G- and P-predicates, for example, “Peter Miller, the dean of the faculty, correctly opened the session”. The application is also successful in the case, if the list of G- and P-predicates is incomplete, for example, “Peter Miller, the chairman of the football association, opens the meeting, …” and the associated list of P-predicates like “he makes no fault in the procedure”, “he is not drunk”, “he is happy” and so on is not finished (119). G-predicates are therefore applicable without a particular specification of the P-predicates because such predicates are instantiated to membership positions. The significance of P-predicates, we associate in particular circumstances, are secondary extensions of the G-predicates. If we use pure P-predicates, like, for example, “Peter Miller, the dangerous forward …”, then we change our universe of discourse. Such predicates may associate with G-predicates as in case of a “forward”, but “dangerous” is not an intrinsic property of any football player. MacDonald and Pettit reconstruct the meaning of G-predicates with their supervenience over P-predicates: Supervenience means that if the group predicate applies in one case but not in another, there must be some associated individualistic predicates that apply in the first case but not in the second, or vice versa. In other words, it means that there is no group difference without a distinction in individuals: specifically, a difference capable of being marked with individualistic predicates. That there is no

144

Gerhard Preyer

group difference without a difference in individuals is guaranteed by the fact that a group is a whole of which individuals are the parts, for it is scarcely conceivable that a whole should change without any alteration in its parts.” (119) I agree with MacDonald and Pettit with their objection to methodological individualism and that the extension of G-predicates is to fix by membership. Yet I think the supervenience of G-predicates over individualistic predicates needs a modification. Let us begin with the presupposed universe to determine the extension of Gpredicates. The membership condition of a social unit is not primary defined by individualistic predicates, but by social status, origin, and also by functional imperatives with which the social unit has to cope. Therefore there is no ontological reduction of G-predicates to P-predicates, and we do not fix the basic concepts of our theory of interpretation by P-predicates to understand, for example, sentences like “the dean opens the meeting”. For the extensional interpretation of such sentences we presuppose a universe of membership conditions (as a class of entities). Yet, what are the instances over which such extensions are supervenient? The link between RI, language and communication are regularities of shared reference caused in the triangle. Parallel to that, the same behavior is redescribed within the procedures in a social frame of reference. In other words, the use of G-predicates is supervenient over such behavioral items that have fulfilled the procedures (call this procedural sociology). The “dean” makes a fault in the procedure to open the meeting, for example, not all members are present; it is a mistaken opening and the colleagues disagree to continue with the meeting. This is not supervenient on an individualistic predicate, although in this case we give an individualistic explanation of the fault. If the “dean” makes the fault when he opens the meeting and the “dean” is “Peter Miller”, then we can apply the principle of extensionality. Therefore we see not only how the divided reference of the use of mixed descriptions of P- and G-predicates works but also that the individuation of the content of the description fits in the framework of triangulation. It is the same response, the same action, and the same event under a certain description that makes communication possible in the social intercourse, and this works only by the principle of the shared reference of reaction, be it verbal or not. The supervenience of G-predicates over procedures and regularities of behavior furthermore explains why RI is to expand for different observers by a social frame of reference within which we individuate behavior. But this is not a turn to social externalism in general.

From an Externalistic Point of View: Understanding the Social

145

Yet one other problem is worthwhile to mention. What is the link between attitudes of agents and the characterization of social units? If we characterize social units by agent attitudes, “we are not referring in the strict sense to those institutions themselves, but rather to the attitudes” (128). We are always participants in social units by status positions and we as human beings are not a part of social systems.10 Our attitudes come in our mind, change in time, are not present for us in every situation, and we are black boxes for observers. Communication does not reach the mental (consciousness) of participants. I think we can bridge the gap between attitudes and social units in the context of their ascription not only by behavioral regularities of shared reference (the identity of reference of caused reactions) but also by the principle of continence. Prima facie judgments are reasons for actions, but our intentional actions are born by stronger judgments. With such judgments we redescribe the functional imperatives of social units. They are parts of the social framework and not of the action. We bridge the gap between attitudes and the social unites therefore without any research of attitudes itself. The divided reference of the mixed extension of the ascription of P- and G-predicates works in such cases. If communication is successful, then there are desirable features from our unrestricted judgments like “this action is desirable” that are desirable in our social intercourse. It was my intention to show how we link the mental (attitudes), language and the social from the triangulation model of RI. This lead me to the problem why triangulation has revolutionized our understanding of social. Now we see why the theories of interpretation, communication and epistemology work together. Is that a social epistemology? The externalistic foundation of the new 10 It is one of the merits of N. Luhmann to reformulate modern individualism (1993: 149258) In pre-modern societies individuals are determinated by inclusion and internal differentiation of such social systems works by exclusion, that is, the membership condition is restricted to exclusive social units, for example the household. The individuality of the members of a social system is founded by social status in principle. In the modern social system individuality is exterritorial (exclusive), and it is a functional imperative of the differentiation of the societal system (Gesellschaftssystem) to develop forms of inclusion (160). The fundamental problem under such conditions is that the differentiated systems have their own conditions of membership. There is no social system without inclusion , but there restrictions are determinated by the evolution of the societal system. I have reconstructed Luhmann’s theory of inclusion and the classic theory of integration as a sociology of membership, but I do not go along with radical constructivism and the theory of autopoietical systems, see, par example, Preyer 1998.

146

Gerhard Preyer

role of interpreter explains us why communication works without the participants reaching their consciousnesses reciprocally. We do not observe the second person’ attitudes directly, but we attain their behavioral items (and utterances). There are different versions of social epistemology. But from my point of view the “social” is connected elementarily with epistemology, if we explain the ascription of thoughts by triangulation because the triangle exemplifies the elementary social relationship. Introducing the social frame of reference shows us that we border on the epistemic differentiation making behavior intelligible. If this is right, we are motivated to recall the Burge-problem because we re-interpret behavior also in respect to the frame of reference in a linguistic (social) community. The ultimate element of this social frame is the membership condition as we presupposed for the ascription of G-predicates, that are not to reduce to individualistic entities. My analysis procedures from RI at first to the epistemic restrictions of intelligible redescription and then from the specification of P- to G-predicates. In sum, it is right to say: Some social phenomena simply can be observed (as both trees and forests can): and indeed, many features of social phenomena are observable (e.g. the procedure of a court) while many features of individuals are not (e.g. intentions). Both individual and social phenomena have observable and non-observable features. If it means that individual phenomena are easy to understand, while social phenomena are not (which is Hayek’s view), this is highly implausible: compare the procedure of the court with the motives of the criminal. (S. Lukes 1973: 123)

Appendix Steve Miller has replied to me that my epistemic charity principle may have been too “charitable” because we need to assume that the body of beliefs attributed to the other are, in the main, true, but problems then arise when we are asked to provide for evidence outside of the triangle, especially adequate empirical evidence. Steve notes down a real problem. Firstly, in my book on Davidson I have discussed this problem in the context of Henderson (2000), on a first version Preyer (2000). Secondly, from my contemporary point of view, I see a solution of this question in leveling epistemic capacities – with different expressibility (also as epistemical means) –; yet, I have not restacked RI in this way in my article. But one of the main point of the principle of epistemic jus-

From an Externalistic Point of View: Understanding the Social

147

tice is – familiar to that –, that background theories take effect to solute the task of interpretation, I have considered. This is the bridge to go outside of the triangle. From the social scientist point of view Steve brings into the play a further point: when I gather qualitative data, what am I to assume about the “truth” of what my respondents are telling me? To border on that in factual systems of communication, we have to cooperate within from different social roles and positions in social stratification, we have to adjust our theories of interpretation about the participants. Many restrictions of different sorts are given on these levels. That is common among social scientists. We need something like the whole picture of the actor (-s) within which we order our intelligible redescriptions. But also a layer or a phoney cannot change the meaning of the words they utter, and we can catch on to them. Michael’s (Roth) response goes in another direction. The difference between us is not significant in principle. He has replied to me: „Gerhard among other things – correctly – makes clear the following: firstly, belief is to be conceived functionally via its role in interpreting linguistic behavior, secondly, to relate beliefs to reality is the interpreter's business, not the speaker’s. Then he considers higher levels of the social to be based on the principle of identity of reference of verbal reactions, That turns out to be his version of Davidson's view of correspondence of true beliefs and their individuation. In this line of argument – following a truth-centered theory of meaning – he has to vote for a distal theory of meaning, because there are no other concepts licensed by this approach that can be considered able to identify the content of propositional attitudes. (Correctly anyway: inner states nor external facts can do so.) For RI of a speaker Gerhard proclaims a principle of natural (!) epistemic justice: the speaker is considered right as long as not proven wrong. “Natural” meaning naturalistic or just the common way of assigning knowledge/ability/ competence, or the like to other fellows? First person authority appears essential to RI, because the speaker, contrary to the interpreter, knows what he believes without evidence. Ascribing propositional attitudes and individuating their content does not require identifying their objects as causes. In that, Gerhard and me like to follow Davidson. But to assume that speaker’s holding true and interpreter’s RI are due to the same distal stimuli is to assume a certain knowledge on the speaker's side, that is, we can not completely escape background theory on what causes linguistic behavior. I like to follow Gerhard in that, too.

148

Gerhard Preyer

As speaker and interpreter are then considered to be involved respectively into the causation of true beliefs – and as this appears so basic to RI – the second person is brought into the picture – that is, two persons interpreting each other – not themselves – reciprocally under a necessary principle of natural epistemic justice. It is now, that it is clear to me, there is a naturalistic reading of the principle intended. Attributing judgements on the world (third apex of triangulation approach) hinge on the interpretational relation of speaker hearer, that is, apex one and two; so is the whole concept of being about, as thoughts and the like are concerned. Aboutness of linguistic behavior hinges on the irreducible second person. Here comes my scepticism: why make the social irreducibility of aboutness argumentatively depending on the natural stimulus in a descriptive sense instead of the knowledge/background theory “causing” linguistic meaning in a non-descriptive sense? Why let yourself been driven into naturalistic commitment rather than into making the first person’s (understanding her- or himself) depending on socially differentiated knowledge, as well? As with the principle of charity this still includes a functional concept of belief and a coherence view on speaker’s beliefs – as being coherent with the socially differentiated and individually accessible knowledge. In conjunction with arguing for irreducibility of G-predicates a less naturalistic version of epistemic justice seems at hand: hearers/interpreters assign epistemic authority in a social frame of referring. That is, they interpret the other person with regard to the knowledge plausibly available to them. This appears a non-individualistic solution to the Burge-problem to me as looked at on the background of two irreducibilities of the social – on any level: (a) aboutness of linguistic behavior and (b) the use of G-predicates.” But why do we bring into play “natural” epistemic justice to solute the task of interpretation? I have mentioned already – in the reply to Steve – that I have not systemized RI by leveling language. This would be the step from RI to an epistemical reinterpretation of behavior, so-called. Yet interpretation can only begin if our response in a similar way is caused by a shared world, we perceive directly and just this is obvious in triangulation. This is not in conflict because of that we make behavior intelligible in a social frame of reference. The externalistic base of our elementary beliefs – expanded by the social frame of reference – is, in my view, a guarantee of our common knowledge, and also science cannot work without that. This is the ground for that why communications work without reaching the mental of the participants. But this is not the last word leveling

From an Externalistic Point of View: Understanding the Social

149

ontology. This is in harmony, that epistemic differentiations in the social universe links semantics and social science.

Bibliography Agassi, J. 1960. “Methodological Individualism.” British Journal of Sociology. Amstrong, D. 1973. Belief, Truth and Knowledge. Cambridge. Avramides, A. 1999. “Davidson and the New Sceptical Problem.” In Zeglen ed., 1999. Bennett, J. 1982. Sprachverhalten (1976). Frankfurt am Main. Bilgrami, A. 1992. Belief and Meaning. The Unity and Locality of Mental Content. Oxford. Brodbeck, M. 1966. “Methodological Individualism: Definition and Reduction.” In Philosophical Analysis and History, ed. H. Dray. London. Burge, T. 1979. “Individuation and the Mental.” In Midwest Studies in Philosophy Vol. V. Studies in Metaphysics. Minneapolis. Cappelen, H. and E. Lepore. 1999. “Semantics for Quotation.” In Zeglen ed., 1999. Davidson, D. 1974. “Paradoxes of Irrationality.” In Freud. A Collection of Critical Essays, ed. R. Wollheim and J. Hopkins. New York. —. 1980. Essays on Actions and Events. Oxford. —. 1980a. “Toward a Unified Theory of Meaning.” Grazer Philosophische Studien 11. —. 1984. Inquiries into Truth and Interpretation. Oxford. —. 1985. “Incoherence and Irrationality.” Dialectica, Vol. 39, No. 4. —. 1990. “The Structure and Content of Truth.” The Journal of Philosophy, Vol. LXXXVII, No. 6. —. 1994. “Davidson, Donald.” A Companion to the Philosophy of Mind. Edited by Guttenplan. Cambridge USA. —. 1999. “Is Truth a Goal of Inquiry? Discussion with Rorty.” In Zeglen ed., 1999. —. 2001. Subjective, Intersubjective, Objective. Clarendon Press: Oxford. Elugardo, R. 1999. “Samesaying.” In Zeglen, 1999. Gibson, R. F. 1994. “Quine and Davidson: Two Naturalized Epistemologies.” In Preyer,Siebelt, Ulfig eds., 1994. Gilbert, M. 1984. On Social Facts. Princeton, N. J. Henderson, D. K. 2000. “Epistemic Rationality, Epistemic Motivation, and Interpretative Charity.” In Preyer and Peter eds., 2000. Hindriks, F.A. “Institutional facts and the naturalistic fallacy.” in this Vol. Kornblith, H. ed. 2001. Epistemology: Internalism and Externalism. London 2001. Lepore, E., and K. Ludwig. 2002. “What is Logical Form.” In Preyer and Peter eds., 2002. Ludwig, K. 1997. “The Truth about Mood.” Protosociology Vol. 10: Cognitive Semantics I - Conceptions of Meaning. —. 1999. “Theories of Meaning, Truth and Interpretation.” In Zeglen, 1999.

150

Gerhard Preyer

Luhmann, N. 1989. “Individuum, Individualität, Individualismus.” In Gesellschaftsstruktur und Semantik. Studien zur Wissenssoziologie der modernen Gesellschaft, Vol. 2 (4 Vols.). Frankfurt am Main. Lukes, S. 1973. “Methodological Individualism Reconsidered.” In The Philosophy of Social Explanation, ed. by R. Ryan. London. MacDonald, G. and Philip Pettit. 1981. Semantics and Social Science. London. Mandelbaum, M. 1966. “Societal Law.” In Philosophical Analysis and History, ed. W. H. Dray. London. —. 1956. “Societal Facts.” In Theories of History, ed. P. L. Gardinier Glencoe. Moser, P. K. and D. Yandell. 2000. “Against Naturalizing Rationality.” In Preyer and Peter, 2000. Pettit, P. 2001. A Theory of Freedom. From the Psychology to the Politics of Agency. Cambridge. Neale, S. 2001. Facing Fact. Oxford. Preyer, G, F. Siebelt, A. Ulfig. 1994. “On Donald Davidson’s Philosophy.” Language, Mind and Epistemology. On Donald Davidson’s Philosophy. Edited by G. Preyer et al. Synthese Library. Dordrecht. —, and M. Roth. 1998a. “On Donald Davidson’s Philosophy: An Outline.” Protosociology Vol. 11: Externalism in Debate. —. 1998b. “Mitgliedschaftsbedingungen. Zur soziologischen Kerntheorie einer Protosociologie.” In Strukturelle Evolution und das Weltsystem. Theorien, Sozialstruktur und evolutionäre Entwicklungen. Hrsg. von G. Preyer. Frankfurt am Main. —, and G. Peter eds. 2000. The Contextualization of Rationality. Problems, Concepts and Theories of Rationality. Paderborn. Preyer, G. 2000. “Interpretation and Rationality: Steps from Radical Interpretation to the Externalism of Triangulation.” In Preyer and Peter, 2000. — and G. Peter (eds.). 2002. Logical Form and Language. Oxford GB. Preyer, G. 2002. Donald Davidsons Philosophie. Von der radikalen Interpretation zum radikalen Kontextualismus. Frankfurt am Main. Putnam, H. 1992. “Replies.” Philosophical Topics 20 (1992). Rogler, E., G. Preyer. 2000. Materialismus, anomaler Monismus und mentale Kausalität. Zur gegenwärtigen Philosophie des Geistes bei D. Davidson und D. Lewis. Frankfurt am Main. Searle, J. R. 2001. Geist, Sprache und Gesellschaft. Philosophie in der wirklichen Welt (1998). Frankfurt am Main. Roth, M. 1986. Bedeutungsverstehen und Sozialwissenschaft. Zu einem sozialen Rahmen der Einstellungszuschreibung. Frankfurt am Main. Schantz, R. 1996. Wahrheit, Referenz und Realismus. Eine Studie zur Sprachphilosophie und Metaphysik. Berlin. Ulkan, M. 1997. “Kommunikative und illokutionäre Akte.” In Intention, Bedeutung, Kommunikation. Kognitive und handlungstheoretische Grundlagen der Sprachtheorie. Opladen. Digital version Verlag humanities-online Frankfurt am Main 2000.

From an Externalistic Point of View: Understanding the Social

151

Watkins, J. W. 1973. “Ideal Types and Historical Explanation.” In The Philosophy of Social Explanations. Edited by A. Ryan. London. —. 1959. “Historical Explanation in the Social Science.” In Theories of History. Edited by P. L. Gardiner. Glancoe. Zeglen, U. M. (ed.). 1999. Donald Davidson. Truth, Meaning and Knowledge. London. Ziff, P. 1960. Semantic Analysis. Ithaca.

152

Steven Miller

Are “Contexts and Events” Equivalent? Possibilities for an Ontological Symbiosis Steven Miller Abstract The article is an attempt to analyze the central but troublesome term “context.” While the concept has important empirical implications for the special sciences (indeed, they may not be possible without it), the focus here is to identify and assess its meta physical and ,especially, ontological status. After reviewing a number of perspectives directed to understanding the origins of the term’s ambiguity and vagueness, it is suggested that the metaphysical domain of “events-theory” may have interesting parallels with what constitutes a “context.” However, although some aspects of context may be assimilated by the events-theory perspective, others are more resistent to such an interpretation. Implications for further investigation are suggested.

I would like to examine an issue that is especially important to the special sciences, but is also relevant to a variety of other positions such as epistemological contextualism and naturalism (DeRose 1999; Kornblith 1999). What I wish to examine, but from an ontological perspective primarily, is what is meant by a “context.” For the special sciences “context” is the foundation for a large range of claims that involve putative explanations of social actions and social behavior. To cite one prominent example in this tradition: sociology argues not only that there are different and unique social configurations (i.e., “social structures”) which can be differentiated on supposedly natural and objective characteristics and, moreover, produce understandable and different social behaviors in virtue of “coming from” or being “located in” these particular configurations (Boudon, 1981). For example, a social structure such as a school is given a special characterization by noting the proportion of its members coming from a given socioeconomic classification. Thus, the social organization itself is given a unique label, here a “high-SES context school,” which is supposed to typify it uniquely and, importantly, produce “outcome effects” different from a “low SES-context” school. Outcome effects being, for instance, some average measure of school performance. The variation in some characteristic is then evidence of a specific context. In a more directly philosophical perspective the evolving epistemological area of “contextualism” gives yet another perspective on the term. Briefly,

Are “Contexts and Events” Equivalent?

153

epistemological contextualism attempts to examine the possibility that there may be alternative ways to justify beliefs that are neither foundationalist or coherentist. As DeRose (1999: 187-188) states it: Contextualism refers to the position that the truth-conditions of knowledge ascribing and knowledge denying sentences vary in certain ways according to the context in which they are uttered. What so varies in the epistemic standards that S must meet in order for such a statement to be true. In some contexts, “S knows that P” requires for its truth that S have a true belief that P and also be in a very strong epistemic position with respect to P, while in other contexts, the vary same sentence may require for its truth, in addition to S’s having a true belief that P, only that S meet some lower epistemic standards. Thus, the contextualist will allow that one speaker can truthfully say, “S knows that P”, while another speaker in a different context where higher standards are in place can truthfully say “S doesn’t know that P”, though both speakers are talking about the same S and the same P at the same time. A host of other contexualist views are available (Annis 1978; DeRose 1992; Dretske 1970) with varying perspectives, but none of them explicitly examines what a “context” is; at best it seems to be a change in the discourse of various agents. It appears a bit strange, then, to note that while the term is so central to the special sciences and this particular form of epistemology, its meaning or internal structure continues to remain mysterious. Below, I wish to list some of the more-or-less commonly related semantic issues and ambiguities associated with the term, just as a backdrop to what I wish to develop later. What I want to eventually show is the possibility that a “context” may in some way be viewed as (if not equivalent to) an “event.” In particular I will look at some of the recent theorizing on the “ontology of events” as proposed by Kim (1999) and Lombard (1999). The general idea is to shift the discussion about “context” away from the epistemological to the ontological. The idea is not simply to attempt to define the term but to see if it can be anchored in some way to the debate on “events.” While itself controversial, the theories concerning event-structures are foundational, I believe, to a clearer understanding of what the special sciences are about. In a very basic sense the special sciences are lacking in epistemic purpose unless they are clear about the ontological status of “context.” If contexts are, for example, equivalent to events (e.g., in terms of an identity-structure), the social sciences become studies of “events,” in addition to whatever else they purport to do. On the other hand, if contexts are not

154

Steven Miller

events they (contexts) will either have to remain in their present ambiguouslystipulative ontological realm, or at least some clarification may be afforded, if one can precisely say why they are not events. Before I try to develop this context-event position, below are some examples, illustrating the ambiguity issue related to a “context.”

Defining Context What, then, is (a) “context”? We have such expressions as the “context of discovery” and the “context of justification,” as well as, in the contextualist approach itself, “conversational” and “issue” contexts. But what, actually, is a context? Does anything really hinge on the concept? Well, for contextualism itself, everything hinges on the concept, since the position would be incoherent without it. But there still is the nagging question of what (a) context is, if anything. I am not sure there exists a non-question begging definition of the term, but a rough beginning place might be to examine it as a category with some possibly relevant essential features. One such feature might be found in the commonly used expression, “you took that out of context.” Leaving aside for now what type of semanticallylinked entity or token the term “that” stands for, let’s focus on the possibility of taking something “out” of context. Assuming this to be more than a metaphor (although it could certainly also be this), an essential feature of a context must be a relevant comparison category, or a relevant reference-category. If I accuse another of quoting me out of context, there must be a relevant state-ofaffairs such that I have knowledge of it and can “show” that what I meant and did say is not what I am now being accused of. Such a “context” could be said to be referentially-antecedent in the sense I am now (presumably) in another context whose meaning is dependent on the previous context. Now, if we label such contexts CB (before) and CA (afer), CA entails CB, but not the other way, for this example; although CB, perhaps in some related sense, for example, could not only entail some other reference category but CA itself, given that certain conditions applied to CB in virtue of which a situation like CB would have a high probability of occurring. In other words, perhaps the semantic content of CA is such that it is highly probable that I am likely to be quoted and interpreted “out of” context. Although some relevant reference-category seems to be the basis for a context, more needs to be said. Can a context be such without a reference-cat-

Are “Contexts and Events” Equivalent?

155

egory? This is conceivable if something is so designated and remains so in some type of spatial-temporal order. Here one would say something like, “here is ContextD and here it still is … and so on.” A context could be invariant, although our intuitions normally are that “a” context can best be understood only in the plural. Indeed, it is unclear whether the term could even be applied to any type of description in the singular. Something like the phrase, “this is my context” would seem odd, although another might speak of your “having” or “being” in a type of context. Thus, the self-referential use of the term would seem to be out of place, although one could have or be in one and only one context – at least for a given time. However, the invariance feature of a context, while logically possible, seems only possible under the following conditions: (1) All essential features remain the same – indefinitely and perhaps infinitely, (2) some, but not all, such features vary, or (3) at least one such feature remains invariant, and only this feature is also considered to be central or primary in some sense. On the face of it (1) seems to be impossible given the features of our world; (2) may seem to be more plausible given (1) but still faces the problem of what even (some) limited features would consist of and why these would be so selected; and (3) suffers the similar fate of support ability given the problems of (1) and (2). Of course (1) could be given a type of support by arguing for Platonic-like entities or prime numbers or certain logical forms or pretheoretical (foundational) perceptive states and the like, but even these would presuppose other contexts, or if not, a non-self referential, non-contextual class of entities somehow “existing.” But a non-contextual class of this type would appear to be self-contradictory. A context, then, whatever it is, would seem to require the possibility of at least one other “context” to make the very concept minimally coherent. And here, one probably doesn’t want to say that a context is simply what one stipulates it to be, since such stipulations would still need some type of minimal justification. Now, what this “other” context would need to have to make it relevantly distinguishable from the claim that one has the “original” one is problematic. Perhaps, it need only be one feature that is both necessary and sufficient, e.g., CA is a context because it differs from CB in terms of some other feature not possessed by CB. But, again, as indicated by (2) above, what the distinguishing feature in the relevant sense should be is unclear. If it is anything this would be too wide of the mark, while if it is only a range or degree of some feature this would seem too limiting – although certainly possible, and indeed, the stipulative basis of what counts as a “context” in much of the social sciences.

156

Steven Miller

For example, some types of organization may have differing degrees of “bureaucracy” – of course with some agreement (by a relevant community) that some thing is an organization. However, if we want to go in this direction what a context is must be an empirical matter, or at least have the possibility of being so. In this case, we naturalize the meaning of context and in so doing imply that, for instance, contextualism as an (alternative) epistemological theory can only be addressed in this way, since what a context is, is given by such a description. In fact, such an attempt to naturalize “context” may turn out to be a plus for epistemological contextualism, or at least not harm it in any fundamental way. Naturalizing may be helpful in its delimitation of what, once and for all, is a context, and in so doing assist the contextualizing-epistemologist by showing she was right all along: our attributions about “knowing” statements are predisposed to act in particular ways given differing contexts. The variety of considerations discussed above still leave the very concept in doubt: it may or may not have various features, but still in the most basic ontological sense the term remains question begging: we still don’t know what one “is” or “looks like.” To try to get some possible perspective on it, I will try to look at it from an “event” framework, asking, “could contexts be events?”

Ontology of Events Why “event?” Events seem to be a continuing source of metaphysical interest because they appear to be so basic in explaining especially our social existence. Events seem to be the anchors we rely on to gain a sense of place, time, and causality. Theories of events are myriad and complex, the most prominent being these (in no special order) of Chisholm (1976), Quine (1960), Kim (1999), and Davidson (1980) to mention a few. A complete and balanced overview of these can be found in the recent State of the Art Essay by Lombard (1998), and I will refer to Lombard also for his particular take (especially in reference to Kim) on the issue of “exemplification.” The exemplification notion is especially relevant, I believe, in trying to establish possible links with contexts and events as they apply to the special sciences. I chose Kim [1976, page references in S. Laurence and C. Macdonald (eds.) 1988], to analyze events since his basic structure has the twin virtues of elegance and simplicity. Additionally, however, his format sets the stage for some extensions and revisions by others which, in turn, may help us understand the context-event relation. To begin, Kim holds that, “the term ‘event’

Are “Contexts and Events” Equivalent?

157

ordinarily implies change, and most changes are changes in a substance” (310). He goes on to remind us that, “by ‘substance’ I mean things like tables, chairs, atoms, living creatures, bits of stuff like water and bronze, and the like; there is no need to associate this notion with a particular philosophical doctrine about substance” (310). Substances, in turn, have properties, and the event’s “change” may be a result of something that the property undergoes in terms of loss or gain, or it may be the “having” of that property at a given time. Events, then, as Kim says are “exemplifications by substances of properties at a time” (311). The term “exemplification”, as can be expected, is key in trying to see how properties relate to events. Generally, in Kim’s so-called “property-exemplification account” any given event has three distinctive constituents: a substance (the ‘constitutive’ object of the event), a property it exemplifies (the ‘constitutive property’ or ‘generic event’), and a time (311). These three elements are given as [x, P, t] and can be defined as, Existence condition: Event [x, P, t] exists just in case substance x has property P at time t, and Identity condition: [x, P, t] = [y, Q, t´] just in case x = y, P = Q, and t = t´ The identity condition, motivated in important ways by Quine’s dictum of “No entity without identity”, is central to the understanding of events and perhaps “contexts.” Its importance lies in attempting to specify the nature (or “essential nature”) of the concept of “property” itself, since it is this dimension that ultimately figures into what an event must be. Here, I will simply adopt the notation for an identity relation given by Lombard (1998: 279) and the comment on it to reflect the concern about contexts. Thus the criterion of identity may be given as: (CI) Q (x)(y) ( I x + I y > (x = y) R (x, y))), where I is some type of entity and R refers to an expression about the relation that x and y bear to each other if they are alike in terms of some subclass of properties, all properties of which I’s can have. One of the important elements of the identity concept actually lies in its counter intuitive and, perhaps, paradoxical function. On the one hand it is needed to argue against the possibility of properties, and on the other, as the only way to show that properties must

158

Steven Miller

figure (in one way or the other) in the explanation of events. One upshot of this, as Lombard (280) indicates in his Principle of the Indiscernability of Identicals, is that in saying that identical things share all properties in common, if a true criterion for identity is one that picks out some properties for I’s, then it follows if I’s are alike for these, they must be alike for all their properties. With the addition of other types of restraints on the identity model, the issue becomes one of not only trying to account for how some type of “metaphysically interesting” identical, but also showing how to be an I depends on those very properties that are picked out by I (Lombard: 280-81). That is, in a noncircular and non-vacuous way, can specific properties, or perhaps a range of relevant properties, be identified so that it can be determined what properties events have “essentially?” Thus for the special sciences whose foundation consists of legitimatizing the nature of “contexts”, the task is to determine if any given context is such because of an identity criterion. The issue of an identity criterion is tied in with how events, in terms of their property dimensions, are conceptualized. For Chisholm (1976) events are considered to be abstract universals because it is possible to conceive of a given event as occurring, in reference to some property it has, more than once for a given individual (universal), and different events can occur to the same subject at the same time (abstract). On the other hand, several theorists construe events, in one form or another, as particulars: for Quine as concrete particulars in the sense that as physical objects occupy one and only one spatio-temporal location so must events; while for Kim, for instance, events are abstract particulars in the sense that an event can be considered as a type of “object,” but one whose particular properties can be “exemplified” in potentially countless ways as explanations for an event. Because events are the exemplifications of property-specific occurrences, there is, in principle, an unlimited number of events which can simultaneously occur in a given space/place. Events are identical, then, if and only if the exemplification of a given property at a given time is the same for that property at another time. The key term in all of this appears to be “exemplification.” For Kim, it is the explanation of why some object comes to have a particular property. Lombard (285), in his review, suggests that Kim’s way of putting it makes objects and their properties (o, P, t), and (o´ p´ t´) for the identity condition, appear to be propositions. If they are propositions, they are not events, but propositions about events, and in any case, something more like universals rather than particulars, supposedly then vitiating Kim’s view of events as abstract particulars. But more directly, Lombard claims (on a distinction made by Bennett, 1988)

Are “Contexts and Events” Equivalent?

159

that if “exemplification” is used in Kim’s sense as “O’s exemplifying Pat t” it is an example of an imperfect nominal which refers possibly to “facts,” while a perfect nominal (e.g., “O’s exemplifying of Pat t”) picks out events. Now, Lombard (285-286) seems to make much of the proposition/fact issue, and while I do want to use his definition of events for my own analysis of “contexts”, I do not believe his criticism of Kim on this issue is completely correct. Again, he seems to place a great weight on the idea that if propositions are involved in some way in Kim’s analysis of events, then by that fact alone Kim’s analysis is flawed because propositions are not about “particulars” but about “universals.” If propositions are about “universals”, then Kim’s analysis would become Chisholm’s. Lombard’s more direct concern is the necessity of keeping a sharp distinction between propositions and events, for if there is none, the ontology of events is assimilated into that of propositions. To show how this may happen, Lombard (285-286) uses the following illustration: Thus, it should come as no surprise to find that, on Kim’s theory there is a match between events and certain true propositions. But clearly it would be wrong to assimilate judgements concerning the identity of propositions to the judgements that we would, apart from Kim’s theory, make concerning when events are identical (surely, my running was my escape from the mugger, but the proposition that I escaped from the mugger is not the same as the proposition that I ran). So, Kim’s theory turns this around and has judgements of identity concerning events assimilated to judgements concerning the identity of propositions (285). On Lombard’s first point about facts being universals, I believe he is using it in the sense that if the propositions “snow is white” and “grass is green” express states of affairs that are true, then, indeed, both are universal in the sense that they hold given what they express. They also, however, refer in particular ways. Thus, while (x) Fx e ( x) Fx, it is the instantiation of an (abstract) particular that makes a purported event an event. While such a view implies that events are propositions, in a “minimal (abstract) particular sense,” there is nothing to fear. By this I mean, that while events are particularly referring propositions, the propositions that they do “exemplify” or are “exemplifications of” are the actual properties that must be included in the very definition of an event itself, for it to be an event. Thus property exemplifications of “objects” are themselves propositions of a certain type-particularly referring ones. Their status as universal or not does not undercut Kim’s position in

160

Steven Miller

any substantial way. A particular property exemplification for that property at that time in that particular spatio-temporal setting may even be considered “universal” then (e.g., it encompasses everything that could only be there, then). For the quote above, “running to escape the mugger,” “escaping from the mugger” and “running” are believed to be, by Lombard, different things. The first is an “event”, while the other two are “propositions.” His concern is that Kim, because his view purportedly involves proposition-stating, is not correctly characterizing the uniqueness of an event. This may be, but rather than assuming that judgement of event identity are assimilated into judgements of proposition identity, it should be noted that what we have above are simply three different kinds of propositions. But if so, are purported events any different from propositions in general? Davidson, for instance, believes events can only be individuated in terms of their identity condition, if and only if they have the same causes and effects. And although begging the question of what an event is, the cause-effect criterion may be important in that there is a (at least implicit) request for an explanation. However, events-as-propositions and propositions in general may or may not involve causal type explanations. A different direction, but one not taken here, is that events are “tropes,” abstract particulars that can be unique characterizations of properties (Macdonald 1998). If they can, properties, or at least our talk about them, may well be about propositions. I am not certain, then, if an event can ever be clearly and absolutely defined. But what is needed, I think, is simply another way of keeping the conversation going. And for this, I believe Lombard’s (289) definition is the most appealing. He argues that the central concept in the talk about events must be that of change-motivated in part by Davidson’s general concerns. Lombard says: “And so it seemed to me that it would be fruitful to think of events as the changes that objects undergo when they change.” He commits himself, broadly, to events being property exemplification accounts but in the following way: An object changes just in case it goes from the having of one property to the having of another, contrary property. An event, then, is not an object’s having of a property (as Kim would have it) or a (particularized) property (as Bennett would have it), but an objects going from the having of one to the having of another property. An event is not, I think, the exemplification of a property, but the exemplifying of a dynamic property.

Are “Contexts and Events” Equivalent?

161

The key idea is that, for Lombard, a property does not remain static; that it changes (becomes an exemplification of) into its contrary. This happens, also, in what he describes as a “quality space”, a class of properties which change into their contraries. I am not exactly certain what this means, but an important point is that Lombard acknowledges the possibility of more than one property changing, and supposedly simultaneously although this is not specified. With these comments, what I would like to do is to return to the original issue of examining a “context.” From Kim, I believe it is useful to retain, in a general sense, the ideas of events as referring to an object’s “substance” as the exemplification, or exemplification of a property, P, at t. From Lombard, the basic definition of event as a describable change, and from Davidson’s original view the “cause-effect” view, not necessarily as the identity criterion alone, but as a way of describing the “dynamic change” concept of Lombard. That is, my view is based on the belief that, indeed, events are about explanations, and specifically the cause-effect variety. The “dynamic change” position also requires it, I believe. However, the issue becomes central for the special sciences in that any “context” is just so in virtue of the fact that it produces a demonstrable “effect.” It is this special nature of contexts that makes them interesting to investigate within the discussion of events.

Is a Context an Event? Does anything philosophically important hinge on the above question? As I mentioned in the beginning, I think it does in virtue of the fact that the special sciences – the area I am especially interested in here – owe their very existence to maintaining the term “context.” Without it virtually no empirical claims can be made in terms of explaining socially produced actions and behaviors. Contexts are foundational, then, for the social sciences since without them these areas of inquiry serve no purpose. By this I mean that the special sciences depend in very basic ways on the possibility of comparison, and comparison is not possible unless there are contexts, that not only logically are, but that differ also. Put differently, these sciences depend on the possibility that behaviors and actions can be shown to vary; without variation, no empirical social science. But the variation, in turn, must have origins, which turn out to be contexts. Contexts, then, serve a bi-valanced function: they are responsible for both producing and explaining variation. But in order for this to happen, con-

162

Steven Miller

texts must first be identified, and then, presumably like events, be shown to have properties that change, or are capable of changing. The matter would seem to be rather straightforward if we could simply say that contexts are equivalent to events. If this could be done the advantage for the special sciences would be in their having at least a stronger presumptive ontological claim for their being legitimately unique methods of inquiry; although, at best, this might have to be a minimalist claim given the controversies still existing in this branch of metaphysics. However, one could still claim that at least it is an important and very fundamental issue. Moreover, if a “context” identity relation is equivalent to an event identity relation, we will at least know that the special sciences are involved in event-explanation and, as such, the more mysterious “context” claim could be replaced or abandoned. Whether or not it can will be the focus of the analysis below. I would like to return to the example of “contextual effects” initially introduced. In many of the social sciences a contextual effect is an outcome of some type that empirically attempts to explain some social action or behavior, or classes of these, by virtue of being in some type of “context.” A given context may be conceptualized from the standpoint of an abstract particular in that a very large number of them, of different types, can be located in principle in the same place at the same time. Thus, a context may simultaneously be an historical-spatio-temporal, cultural, societal, organizational, psychological, and linguistic occurrence. Moreover, the number of “properties” for a given context will depend on the characterization of the context and what (in this case) is worth emphasizing empirically. From an empirical point of view, a context is better understood as the context in the sense that it “exists” in virtue of how it is “placed together,” or what assumptions are made given the beliefs about what is important to highlight and how it should be highlighted. In other words, to use Kim’s language, the “constitutive substances” of contexts, let alone their properties, may go beyond a “generic event” identification. For example, the “cognitive status” (i.e., attitudes, beliefs, abilities, etc.) of individuals within an empirically determined context (e.g., a department of philosophy) may be of a certain type – they are all brilliant! – or they may vary in some particular way. Now, how they are may be the result of some specific historical-cultural process and may (then) constitute a general class (brilliant) because of this process. However, the “context” of the philosophy department may vary for other reasons and thus other contexts do occur, and all of these can be studied empirically in one way or another.

Are “Contexts and Events” Equivalent?

163

What I wish to point out is that there seems to be an initial distinction between events and contexts: Events seem to occur, while contexts are created. There are, of course, linguistic expressions of the sort, “that event was staged (i.e., artificially constructed),” but this seems a bit wide of the mark because, I think, we are still more likely to characterize events as somehow more “naturally” occurring, either in terms of behaviors or dispositions. Contexts, on the other hand, seem to have a type of artificiality about them: they are “constructed” with “ends” in mind, whether those “ends” are the objects of a certain type of explanation, e.g., “why are philosophy departments the way they are?”, or whether they are “constructed” in a rather strict way, as when experimental psychologists “create” an experimental design where, in effect, “control” and “experimental” groups are contexts, and systematically different ones (McCall 1970). Perhaps a rough initial characterization may view events that seem to be naturally occurring as first-order events (e.g., my falling, my escaping from the mugger, etc.), while a first order context may be something like the set up of a “classical” experimental design. However, especially with contexts, there are borderline cases. To return to our early example, an organization like a school may be defined as a context (or, “having” a context – perhaps a type of exemplification) based on it occurring “naturally”, e.g., it just happens that this school has a predominance of low “ability” students, or we may create a “highability” context school, for whatever policy purposes we have in mind. The first is in one sense a more “naturally” occurring, leaving aside that its occurrence – this type, here and now – is the basis for a broader contextual explanation. Of more general interest, however, may be the possibility of an event being described as such by virtue of its occurring in a context. By this I mean that although all events must occur in some context, however broad and vague they are described, there is the possibility that a given specific event could not occur in the absence of a rather specific context – here contexts may be viewed as being necessary conditions for events. But, perhaps, they are merely sufficient. If, for example, the student experiences (or has) the event of doing well on a standardized test, and she is in a “high-ability” context school, the event of her doing well may or may not be the fact of the context. If it is, this is what empirical social science must determine. If there is no empirically determined connection to a particular context, the event may still be context-related in another fairly specific context of another kind. I do not believe, therefore, there is something like the first order type event described above. Events, like contexts, require explanation, although events

164

Steven Miller

seem to be capable of explanation only if fairly specific contexts are available. The need to understand both events and contexts by “explaining” them seems basic. Figure 1 illustrates this. Events

Context

Causal

(a)

(b)

Functional

(c)

(d)

A context, especially the type related to the empirical sciences, provides for a functional explanation (d), roughly of the following kind, “the function of this context is to …,” where what is predicated is commonly some outcome. Whereas, if the context is believed to be causal (b), it posits a particular feature of the context (e.g., “high-ability”) as being determinative of some outcome. As mentioned, the context is defined as a context by some special feature(s) it has. The feature, then, is in some ways similar to the “constitutive object” of Kim’s analysis, while its “properties” (here) may be a range of empirical indicators. For events, the functional (c) interpretation seems linguistically odd in the sense that if a characteristic of an event is simply that it occurs(ed), there is no point in why (functionally) it did so. If the event is “strolling through the park” then that’s what it is. However, if we say that events, to be events, require some type of explanation, then a functional explanation may not seem that strange: “Why were you running to escape the mugger? So that I wouldn’t get hurt”, seems perfectly adequate. Likewise, if the element of change is allowed, or even required, to understand events, the very notion of “exemplification” may serve as a form of causal explanation: the exemplification of a property in a particular way suggests that some account of how the change has occurred over time may entail a causal explanation; thus (a), at least in a broad sense, is possible. Contexts may not, then, be clearly differentiated from events on these versions of “explanation” in regards to the empirical social sciences. And, yet, there seems to be something basically different between the two. For instance, if we simply stipulate that the term “context” must, in intent and use, always be replaced by “event”, there still appears to be something unnatural about this, e.g., “In this context (event) moral relativism is perfectly understandable,” or “You took my comments out of context (event),” or “These two different or-

Are “Contexts and Events” Equivalent?

165

ganizational contexts (events) can predict the degree of alienation.” Of these, the last example still echos the possibility of some type of equivalence being possible between the two. I think that if any case is possible along these lines, it must be located, as Lombard, Kim and others point out, in the idea of change.

Changing Contexts and Events If the defining characteristic of an event is, as Lombard says, the going from one property to its contrary, it would be interesting if this requirement holds for contexts as well. There is also Lombard’s notion that this exemplifying of a property is to be viewed as an example of “dynamic change.” I am not sure what “dynamic” exactly is, but certainly the “change” element is important. Now, for the empirical social sciences, contexts do have a similarity with events in terms of change. However, for many cases of contexts the definition of a “context” is not determined by its change from one condition to another, but rather by either a stipulation or empirical determination that (at least) two different contexts exist. Let me clarify. If a social organization (e.g., a school) is claimed to be of a particular type, such as a “high-SES context” school, it obtains this designation in virtue of the fact that (1) a claim is made that such social organizations can be correctly “typed” or labeled by such a designation, and (2) that there are reliable and valid empirical indicators attesting to this fact (generally, Little 1991). The empirical indicators are needed to first establish what the distribution of persons coming from high-socioeconomic status backgrounds is within the school and what proportion is sufficient to label it in this way. Secondly, and crucial for the context argument, is that the characteristic in question (i.e., SES) can vary. That is, other social organizations can have other SES distributions, e.g., a “low-SES context” school. But what constitutes a context, then, for this and most other examples in the special sciences, is a type of classification scheme; a “placeholder” whereby a given context is a kind of place, setting or linguistic phenomenon which describes a socially constructed entity. An important point, I think, is that the exemplification of a characteristic (call it a “constitutive substance”) is based on the (empirical) possibility that such a characteristic (e.g., “ability”) has “properties” (i.e., specific empirical attributes) which imply the possibility of other contexts, given the distribution of these properties. Thus, a high-SES context school will imply a low-SES context school. However, contexts which

166

Steven Miller

can be juxtaposed in this way do not, necessarily indicate that a true change has occurred: a high-SES context school does not change into a low-SES context one. Events, on the other hand, imply a qualitative change, where an event is determined to be so only if some property “dynamically” alters itself over time. It is in this sense that properties associated with events are thought to be “exemplifications.” There is no such qualitative change in the present example of a context, since there is no real property exemplification occurring, as Lombard (290) puts it, “in the same quality space.” A context is in a quality space but is not of it. At least on this characterization of contexts, then, they are not identical with events. Within the empirical social sciences, however, contexts have yet another feature, and it is this feature that in some sense at least brings them closer to events. As used in the social sciences, contexts are meant to simply divide social reality “at the joints” (Hirsch 1993). That is, their role as descriptors is simply a prelude to their role as predictors of actions and behaviors. A context with this characterization is meant to develop a causal or functional explanation scheme. As such, contexts are associated with predicted outcomes. Thus, a high-SES context school is postulated to predict, let us say, corresponding high standardized test achievement scores, in and of itself, but also as compared to a low-SES context school. The addition of an “outcome measure” also changes the analysis of what a context is. Now, the context describes a type of event: the “event structure” is just the case where the context is (causally) associated with a predicted outcome. A unique event is described – the event of the school context producing (or not producing) some outcome. Such a “context-event” can, broadly, be associated with the dimensions of events: substance, property, time, and quality space.

Concluding Comments The type of equivalence between context and event specified above at least suggests a limiting case for the empirical social sciences, but one in which the term “context” is given justification within the ontology of events. Thus, there are some substitution instances where, for the criterion of identity (CI) that Lombard formulates, contexts and events are symmetrical. This type of symmetry, however, does not apply to all cases. That it doesn’t suggests that contexts may have additional features that at least need mention. One is that events supervene on contexts, but the supervention is inversely related. That is, all

Are “Contexts and Events” Equivalent?

167

events are “context-specific” but as the generality of the context increases its possible explanatory force (e.g., in terms of causality) of the event is diminished. For example, within the special sciences, the historical event, “the bombing of Pearl Harbor on December 7, 1941” may be located in the context of “Japanese cultural beliefs.” Now, if the view is taken that events are what they are irrespective of, but occurring in, a “larger” context, then explanatory accounts are irrelevant, or at least irrelevant at this level. The potential explanatory efficacy of the event becomes more probable, presumably, as the context becomes more specific. However, this characterization shifts the analysis to the possibility of contexts causing events as opposed to contexts and events having separate and individual causal properties. And it is exactly here that the empirical social sciences must maintain the context-event distinction. Again, slightly differently, the special sciences “construct” the explanatory contexts needed to explain events. These contexts are mostly a posteriori attempts to account for events, although, of course, they may be predictive, in the sense that perhaps given knowledge of “Japanese cultural beliefs,” as a context, one could have predicted the subsequent attack. I do not want to enter into those debates which consider the possibility of events being partially or wholly socially constructed (for various views, Finn 1997, Goldman 1999, Lynch 1998, Searle 1995), but only to note the asymmetry that occurs when contexts are considered to be causally efficacious for events, but not vice versa. A further distinction between contexts and events is that the former are bounded by normative or rule considerations in ways that events are not. To say that an organization has a context is just to say it can be specified by empirically relevant criteria. Contexts are bound by methodological rules which both define and constrain their range of meaning(s). Moreover, while rules and empirical criteria both bound the definition and evaluation of a context, the evaluation of events is somewhat different. It seems that truth-functional criteria for events, at least those emphasizing the exemplification point of view, i.e., in property change, minimally, consists of affirming that a property has changed to its contrary. This is an internal logical criterion and appears to be primary in these discussions. Of course, once this criterion has been established, other external standards can be applied. For instance, if the bombing of Pearl Harbor on December 7, 1941 is taken as an event because of some change in a previous state-of-affairs, the truth of the statement expressing the event can likewise be established by applying, say, the criteria of historical analysis. Contexts, on the other hand, at

168

Steven Miller

least those for the empirical social sciences, satisfy both the internal and external truth functional criteria by application of empirical criteria themselves. In sum, the theories on the ontology of events do play a pivotal role in understanding the functions of the special sciences. What has not been previously explored, I believe, is both how events may be related to contexts in these sciences, and, more generally, what constitutes the possibility of an “ontology of contexts.” What I have tried to allude to is that the relation between contexts and events is complex. The mistake is to think of them as somehow equivalent. There are similarities of course, but a context, to be a context, involves the empirical specification of background conditions that are linked to some type of outcome or effect. When contexts are so defined they share similarities with events. When they do not, a context may range from any arbitrary way of dividing social reality by empirical classifications which, purportedly, do this objectively. Contexts also appear to be more about the “division problem” than do events (Hirsch 1993). That is, the issue for determining a “context” is to argue that some term within a natural language more accurately individuates some aspect or “portion” of reality than does a comparable term drawn from that language or an “artificial” one, e.g., something that is “green and circular” versus “gricular.” On the other hand, what an “event” is appears to be more about determining an “internal criterion” rather than how of two different events, one “divides” reality more accurately than the other. Events, then, can be viewed as occurrences which go in the “opposite direction” from contexts. Events are about the criteria needed to say that a given characterization (of an event) is a unique one; one that goes beyond a variety of other possible “events” related to it, e.g., “running to escape the mugger” versus “running” versus “escaping the mugger,” etc. Contexts go the other way: the justification of any and all divisions of reality; “social” reality for the special sciences. Contexts and events are then, ontologically symbiotic: A context becomes interesting to the extent it is seen as (non-circularly) producing and interpreting and event, and events (aside from their “internal” criteria) are interesting to the extent they can be shown to be either context-free or contextdependent.

Are “Contexts and Events” Equivalent?

169

References Annis, David. “A Contextualist Theory of Epistemic Justification.” American Philosophical Quarterly 13 1978, 213-219. Bennett, Jonathan. Events and Their Names. Indianapolis, IN: Hackett Publishing 1988. Boudon, Raymond. The Logic of Social Action: An Introduction to Sociological Analysis. London: Routledge and Kegan Paul 1981. Chisolm, Roderick M., Person and Object. LaSalle, IL: Open Court Publishing 1976. Davidson, Donald. “The Individuation of Events.” Essays on Actions and Events. Oxford: Oxford University Press 1980, 216-234. DeRose, Keith. “Contextualism: An Explanation and Defense.” The Blackwell Guide to Epistemology. Edited by J. Greco and E. Sosa. Oxford: Blackwell Publishers 1999, 187-205. — . “Contextualism and Knowledge Attributions.” Philosophy and Phenomenological Research 52 1992, 913-929. Dretske, Fred. “Epistemic Operators.” Journal of Philosophy 67 1970, 1007-1023. Finn, Collin. Social Reality. London: Routledge, 1997. Goldman, Alvin I. Knowledge in a Social World. Oxford: Clarendon Press 1999. Hirsch, Eli. Dividing Reality. Oxford: Oxford University Press 1993. Kim, Jaegwon. “Events as Property Exemplifications.” Contemporary Readings In The Foundations of Metaphysics. Edited by S. Laurence and C. Macdonald. London: Blackwell 1999, 310-326. Kornblith, Hilary. “In Defense of a Naturalized Epistemology.” The Blackwell Guide to Epistemology. Edited by J. Greco and E. Sosa. Oxford: Blackwell Publishers 1999, 158-169. Little, Daniel. Varieties of Social Explanation: An Introduction to the Philosophy of Social Science. Boulder, CO: Westview Press 1991. Lombard, Lawrence B., “Ontologies of Events: States of the Art Essay.” Contemporary Readings In The Foundations of Metaphysics. Edited by S. Laurence and C. Macdonald. London: Blackwell 1999, 217-294. Lynch, Michael P. Truth in Context: An Essay on Pluralism and Objectivity. Cambridge, MA: The MIT Press 1998. Macdonald, Cynthia. “Trepass and Other Things: State of the Art Essay.” Contemporary Readings In The Foundations of Metaphysics. Edited by S. Laurence and C. Macdonald. (London: Blackwell 1999, 327-350. McCall, Robert B. Fundamental Statistics for Psychology. New York: Harcourt, Brace and World, Inc. 1970. Quine, W.V. Word and Object. Cambridge, MA: MIT Press 1960. Searle, John. The Construction of Social Reality. NewYork: Free Press 1995.

170

Frank A. Hindriks

Institutional Facts and the Naturalistic Fallacy. Confronting Searle (1964) with Searle (1995) Frank A. Hindriks

Abstract In 1964 Searle argued against the naturalistic fallacy thesis that an ought-statement can in fact be derived from is-statements. From an analysis of this argument and of Searle’s social ontology of 1995 – which includes a full-blown theory of institutional facts – I conclude that this argument is unsound on his own (later) terms. The conclusion that can now be drawn from Searle’s argument is that social or institutional obligations are epistemically objective even though they are observer-dependent. I go on to argue that the strength of such obligations depends on the strength of the underlying collective acceptance, which is a kind of collective intentionality. I also point out that all normativity in Searle’s framework can be traced back to either individual or collective intentionality. This bars him from providing an account of intention-independent moral facts.

1 Introduction In 1964 John R. Searle argued that the thesis that an ought-statement cannot be derived from is-statements is mistaken. Supporters of the thesis claim that those who believe that the thesis is mistaken commit the naturalistic fallacy. Because of this, I will refer to the thesis as the naturalistic fallacy thesis. Searle argues against it that an ought-statement can in fact be derived from is-statements. His argument depends crucially on an account of institutional facts. However, at the time Searle wrote his paper, he did not have a full-blown account of institutional facts. It is only since 1995 – when Searle’s book on social ontology The Construction of Social Reality was published (henceforth CSR) – that such an account is available. This account has been summarized in the last two chapters of Mind, Language, and Society (Searle 1999, henceforth MLS). I will use it to shed new light on his argument of 1964. The outcome of my analysis consists of two claims concerning the implications of Searle’s recent social ontology for his earlier attack on the naturalistic fallacy thesis. The first claim is a negative one: Searle’s full-blown account of institutional facts in CSR and MLS implies that his 1964 attack on the natural-

Institutional Facts and the Naturalistic Fallacy …

171

istic fallacy thesis fails on his own (later) terms. The claim that Searle’s attack fails is not new. Richard M. Hare (1969 [1964]), for instance, claims the same. However, the way in which I argue for it is new, as I rely on Searle’s recent ideas for evaluating his earlier argument. The second claim is a positive one: Searle’s full-blown conception of institutional facts can be used for deriving a genuine implication from his argument, which is the thesis that statements about institutional facts and the concomitant rights and obligations are epistemically objective. A claim to this effect was already present in his 1964 paper, but recent additions to Searle’s conceptual apparatus, such as the notion of collective acceptance and the distinction between epistemic objectivity and subjectivity, open up the possibility of making the earlier conclusion more precise. A corollary of the claim is that not all evaluative matters are necessarily subjective in the sense of being merely matters of taste or opinion. On Searle’s full-blown account of institutional facts, they require collective acceptance for their existence. Collective acceptance is a form of collective intentionality, which is a notion that captures the idea that people can share intentional states. Collective acceptance is the kind of collective belief on which the rights and obligations of the relevant institutional facts depend. These social rights and obligations are objective. Their objectivity provides a direct basis for the positive claim. The fact that the obtaining of collective acceptance is an empirical matter provides the basis for the negative claim, as this implies that one of the premises of Searle’s argument is both synthetic and evaluative and hence not an is-statement. A further claim for which I argue is that the strength of the social obligations that are involved depends on the strength of the underlying collective intentionality. Finally, I point out that all normativity in Searle’s framework depends on either individual or collective intentionality. Because of this, it cannot accommodate intention-independent moral facts. I start section 2 by discussing the naturalistic fallacy thesis and the requirements that a counter-argument to it has to meet. I go on to present Searle’s 1964 argument and the view of promising on which he relies there, putting it in perspective by shortly discussing some ideas of Anscombe and Austin that inspired him. I also discuss the wider significance Searle believed his argument had. Section 3 discusses Searle’s current views on institutional facts generally and promising in particular. It focuses on the role of collective acceptance and its relation to social obligations and argues that the alleged objectivity of collective acceptance implies that his earlier argument was unsound (the negative claim; section 3.2). Finally, I argue that the social obligations that exist in virtue of collective acceptance are epistemically objective (the positive claim; section

172

Frank A. Hindriks

3.3). Section 4 discusses the place of normativity in Searle’s framework and the relation between the strength of collective intentionality and that of the related social obligations. Section 5 summarizes the main conclusions.

2 The naturalistic fallacy thesis and Searle’s 1964 argument about promising 2.1 The naturalistic Fallacy Thesis According to Searle, the naturalistic fallacy thesis is the view that ‘no set of descriptive statements can entail an evaluative statement’ (1964: 43; emphasis in original). He presents an argument concerning promising as evidence against the soundness of this view and concludes that the naturalistic fallacy thesis is mistaken. Before considering his argument, we need to know what it takes to show the naturalistic fallacy thesis to be unsound. The heart of this matter consists of a grasp of what it is for a statement to be descriptive or evaluative. In order for a statement to be evaluative, it must satisfy two conditions. First, it must contain evaluative terms such as ‘good’, ‘obligation’, or ‘ought’. Second, it must be synthetic. Moore (1903) held that there are no analytic statements concerning the term ‘good’. He regarded it as an indefinable predicate. The contrary claim that the term ‘good’ can be defined he dubbed the naturalistic fallacy (ibid.: 7 and 13). He also held that the property ‘good’ is not identical to any other property. What Searle calls the naturalistic fallacy is a corollary of this view. If ‘good’ cannot be defined, all statements about it must be both synthetic and evaluative. It follows that any argument with an evaluative conclusion concerning the term ‘good’ requires a synthetic statement that contains that evaluative term as one of its premises. Moore regarded the claim that ‘good’ is indefinable and hence that statements concerning the term must be synthetic as obviously true. If this view is generalized to all evaluative terms, it follows that an evaluative conclusion can only be derived from a set of premises at least one of which is evaluative. (See also Brink 1989: 147-49 and Frankena 1939.) Evaluative statements are not necessarily moral statements. Searle’s formulation of the naturalistic fallacy in terms of evaluative statements leaves it open whether or not he takes his argument to be related to morality. Only replacing ‘evaluative’ by ‘moral’ would commit him to the moral version of the natural-

Institutional Facts and the Naturalistic Fallacy …

173

istic fallacy thesis. Many philosophers who subscribe to the moral version subscribe to the non-moral version as well. Hare (1969 [1964]) takes Searle to be concerned with the moral version. In his original statement and defense of the argument, Searle is not explicit about this matter. He does claim that the ought of the conclusion of his argument is categorical, which can be taken to imply that a moral obligation is involved. In 1969, however, he distanced himself from the view that the institution of promising is a moral one by claiming that ‘the obligation to keep a promise probably has no necessary connection with morality’ (Searle 1969: 188). I will assume that Searle is concerned with the non-moral version of the naturalistic fallacy thesis and hence that the relevant obligations are merely social or institutional. In order to show that the naturalistic fallacy thesis is unsound, one needs to find an analytic statement about an evaluative concept. Furthermore, one needs to derive an evaluative statement from it, and show that none of the synthetic statements required for the derivation is evaluative. In other words, apart from the validity of the argument, the core issue is the status of the premises: are they analytic or synthetic, and if the latter are they evaluative or not? I will refer to the premises and the conclusion as the narrow argument. I will talk of the broad argument when I intend to include Searle’s claims concerning the status of the premises and the conclusion. 2.2 Searle’s argument about promising Searle’s narrow argument consists of the following series of five statements: (1) Jones uttered the words ‘I hereby promise to pay you, Smith, five dollars.’ (2) Jones promised to pay Smith five dollars. (3) Jones placed himself under (undertook) an obligation to pay Smith five dollars. (4) Jones is under an obligation to pay Smith five dollars. (5) Jones ought to pay Smith five dollars. (1964: 44) Searle regards (1) as a descriptive statement, and (5) as an evaluative statement. Given certain amendments, he takes the relation between consecutive statements to be one of entailment. He believes that none of the amendments introduces an evaluative statement. If he is right, he has shown the naturalistic fallacy thesis to be unsound. The amendments pertaining to statements (3) and (4) consist of ceteris paribus clauses that have to be satisfied in order for the argument to go through.

174

Frank A. Hindriks

Searle added them in order to take into account the fact that obligations are often defeasible. These clauses have been the target of severe criticisms. Because of this, Searle has reformulated the argument without them in his 1969 and reformulated statements (4) and (5) accordingly. I will assume that the steps from (3) to (4) and from (4) to (5) are valid, either in their original formulation or in the amended form, and do not involve hidden evaluative premises. My concern is not with defeasibility. Instead, I will focus on statements (1) through (3), (roughly) the first half of the argument. As Searle (1969: 179n1) notes, statement (3) is already evaluative because of the term ‘obligation’. This provides a reason for regarding the first half as the core of the narrow argument. Including the amendments, the core of the narrow argument consists of the following six statements: [1] Jones uttered the words ‘I hereby promise to pay you, Smith, five dollars.’ [2] Under certain conditions C anyone who utters the words (sentence) “I hereby promise to pay you, Smith, five dollars” promises to pay Smith five dollars. [3] Conditions C obtain. [4] Jones promised to pay Smith five dollars. [5] All promises are acts of placing oneself under (undertaking) an obligation to do the thing promised. [6] Jones placed himself under (undertook) an obligation to pay Smith five dollars. (See Searle 1964: 44-46; the numbering is mine) [1], [4], and [6] correspond to (1), (2), and (3); [2], [3], and [5] are the amendments that have been added to make the relation between the original statements one of entailment. 2.3 The status of the statements and the institution of promising We arrive at the (core of the) broad argument by complementing the narrow argument with claims about the status of the premises and the conclusion. Statements [1] through [3] are premises from which statement [4] is derived. In his 1969 Searle formulates the form of the constitutive rule as ‘X counts as Y in context C’ (35). Statement [2] has this form, and hence is (at least part of) the constitutive rule of the institution of promising. Context C corresponds to the conditions C, the phrase ‘anyone who utters the words (sentence) “I hereby promise to pay you, Smith, five dollars”’ is the X of the constitutive rule for

Institutional Facts and the Naturalistic Fallacy …

175

promising, and the phrase ‘promises to pay Smith five dollars’ is the Y. This enables us to see that [1] states that X obtains and that [3] states that context C obtains. Searle regards both statements as empirical claims. It is uncontroversial that the claim that Jones uttered certain words is an empirical claim that does not contain evaluative terms. Hence, statement [1] is non-evaluative and synthetic. Searle claims that statement [3] is non-evaluative as well. Searle claims that the constitutive rule of promising is a meaning rule and hence an analytic statement. This implies that statement [2] is non-evaluative. He also describes the statement as expressing a fact about the English usage of certain words, which supports his view of the constitutive rule as a non-evaluative and analytic statement. The constitutive rule together with the two allegedly empirical claims imply that Jones has indeed made a promise, as statement [4] has it. Searle regards statement [5] as an analytic statement that together with statement [4] entails statement [6]. Statement [6], the conclusion, is an evaluative statement. This completes the broad argument. Whereas it is relatively clear that statements [1] and [2] are non-evaluative, the status of statement [3] is more controversial. Searle claims that the conditions of statement [3] are straightforward empirical conditions. He gives several examples of the conditions he has in mind, which include the condition that the speaker (Jones) has to be in the presence of the hearer (Smith), and that they are both speakers of English. I grant that these are empirical conditions. What is missing, however, is a general argument for the claim that all conditions including those that Searle does not make explicit are empirical. As said in the introduction, Searle did not have a full-blown theory of institutional fact at the time he put forward his attack on the naturalistic fallacy thesis. This made it impossible for him to prove his contention. In section 3, we will consider whether it holds on his recent theory of institutional facts as proposed in CSR. Here I will present the most important claims Searle made about promising in 1964 and 1969 after I have put his views in a historical perspective. G.E.M. Anscombe and J.L. Austin have been important sources of inspiration for Searle. Searle took the distinction between brute and institutional facts from Anscombe (1981 [1958]). Anscombe’s main example of an institutional fact is that of owing money to a grocer. The example concerns an institutional fact that obtains because the grocer supplied potatoes under certain conditions. The range of conditions that suffice for the institutional fact to obtain cannot be spelled out completely, but only indicated roughly. Anscombe regards the distinction between brute and institutional facts as a relative one. The fact that the grocer supplied me with potatoes is a brute fact relative to me

176

Frank A. Hindriks

owing him money. Supplying potatoes in turn is an institutional fact relative to the grocer ‘carting a quarter of potatoes to my house and leaving them there’ (ibid.: 24). The conditions mentioned reappear in Searle as conditions C of the form of the constitutive rule ‘X counts as Y in C’. In contrast to Anscombe, however, Searle regards the distinction between brute an institutional facts as an absolute one. Brute facts are those facts that do not presuppose the existence of certain institutions (Searle 1969: 51). The implication is that on Searle’s conception, the X of a certain institutional fact can itself be an institutional fact (see CSR: 80). Austin (1962) was Searle’s source of inspiration for his analysis of speech acts. Austin argued that we do things with words, i.e. we perform acts by our utterances. Describing reality or asserting something is only one possible action. In relation to this, Austin introduced the notion of illocutionary force. The underlying idea is that one can perform different acts with respect to one and the same proposition. I can assert that it is time for lunch, and I can hope it is time for lunch. In certain circumstances I can even declare that it is time for lunch. Searle used these ideas in his analysis of promising. He argued that to make a promise is not to assert or describe something, nor to evaluate something or voice one’s opinion about it. Instead, it is a speech act by which an institutional fact is created. In addition to this, Anscombe and Austin may have inspired Searle with regard to his attack on the naturalistic fallacy thesis. Anscombe claimed that Hume’s objection to ‘passing from ‘is’ to ‘ought’ would apply equally to passing from ‘is’ to ‘owes’ or from ‘is’ to ‘needs’’ (1969 [1958]: 176). She also wrote that she does not find accounts of the naturalistic fallacy thesis coherent. Austin wrote that his classification of speech acts is ‘quite enough to play Old Harry with two fetishes which I admit to an inclination to play Old Harry with, viz. (1) the true/false fetish (2) the value/fact fetish’ (1962: 150; cf. Flew 1969 [1964]). I suggest taking a closer look at some of the things Searle writes about promising in order to appreciate the fact that he regards the constitutive rule of promising as an analytic statement. The relation between constitutive rules and obligations needs our attention as well. Consider the following passage: In certain circumstances, uttering the words in quotation marks in (1) is the act of making a promise. And it is part of or a consequence of the meaning of the words in (1) that in those circumstances uttering them is promising. ‘I hereby promise’ is a paradigm device in

Institutional Facts and the Naturalistic Fallacy …

177

English for performing the act described in (2), promising. (Ibid.: 44; emphasis added) Searle writes here about the speech act of promising and its relation to the meaning of words. Later on he claims that in uttering the phrase ‘I hereby promise …’ one performs but does not describe the act of promising. Once promising is seen as a speech act of a kind different from describing, then it is easier to see that one of the features of the act is the undertaking of an obligation. But if one thinks the utterance of ‘I promise’ or ‘I hereby promise’ is a peculiar description – for example of one’s mental state – then the relation between promising and obligation is going to seem very mysterious. (Ibid.: 51) From this passage, it can be gathered that Searle argues in effect that the performance of a certain act has normative ramifications. This should make us suspicious of his claim that he derives an evaluative statement from a descriptive statement, as it raises the suspicion that there is a hidden evaluative premise after all. Finally, consider the following passage: It is a consequence of [the meaning of the phrase ‘I hereby promise …’] that the utterance of that phrase under certain conditions is the act of promising. Thus by presenting the quoted expression in (1) and by describing their use in (1a) we have as it were already invoked the institution of promising. (Ibid.: 49) One thing these passages do is relating the meaning of certain words to the constitutive rule. Apparently, what Searle called a fact about English usage of certain words is a fact about the meaning of the relevant words (Anscombe makes the same claim in a later article 1981 [1978]: 102). This suggests that the statement of the constitutive rule is an analytic statement. Searle also characterizes the constitutive rule as a meaning rule and calls it a tautology (1969: 185 and 1964: 56-57 respectively). Furthermore, he writes that promising is by definition an act of undertaking an obligation (Searle 1969: 178). These statements show that Searle regards the constitutive rule as an analytic statement and that social or institutional obligations owe their existence at least in part to the constitutive rule. 2.4 Searle’s general diagnosis Searle claims that his example of promising shows that there is something fundamentally wrong with what he calls the classical empiricist picture. Accord-

178

Frank A. Hindriks

ing to this picture, he claims, there are objective facts that can be reported in descriptive statements that are true or false. On the other hand, there are subjective evaluations, which are matters of opinion or taste. Statements that express such subjective matters are not true or false. Searle objects to this picture, because it leaves out institutional facts. These facts are similar to the facts of the empiricist in that they are objective. At the same time, they resemble subjective statements because they have evaluative features. Although Searle balks at the claim that ‘promise’ is an evaluative rather than a descriptive term when this is seen as a claim that undermines his argument, in the end he is willing to grant that the term is in some sense evaluative even though it is descriptive as well. This shows, he claims, that ‘the whole distinction needs to be re-examined’ (Searle 1964: 58). Searle regards the distinction between descriptive and evaluative as a conflation of two distinctions, one between different kinds of speech acts such as describing and evaluating, and one between statements whose truth or falsity is an objective matter and statements for which this does not hold. The former is usually seen as a special case of the latter, which in turn is taken to imply that factual premises cannot entail evaluative conclusions. Searle denies this and points to institutional facts as objective but evaluative at once. The main message of his article is that ‘[i]t is often a matter of fact that one has certain obligations, commitments, rights, and responsibilities, but it is a matter of institutional fact, not brute, fact’ (ibid.: 56).

3 Institutional facts and the objectivity of collective acceptance 3.1 The three illocutionary forces of promising Searle’s later ideas about promising require a basic understanding of his theory of speech acts. Searle distinguishes between five kinds of illocutionary forces and hence between five kinds of speech acts. The five kinds of speech acts are assertives, directives, commissives, declarations, and expressives. Assertives are descriptive statements; directives are statements we use to get the hearers to do things; commissives are statements we use to commit ourselves as speakers to doing certain things; declarations bring about changes in the world in virtue of the utterances themselves; finally, we use expressives to express our feelings (see for instance Searle 1983:166 and MLS: 148-50). Searle claims that promises are assertives, commissives, and declarations all at once (MLS: 150). In other

Institutional Facts and the Naturalistic Fallacy …

179

words, promising has three illocutionary forces and in making a promise one performs three illocutionary acts. We will consider these three illocutionary acts in turn. That Searle regards promising as an assertive makes it possible to clear up some confusions that may arise from reading Searle (1964). There he said that ‘[i]n uttering [the words “I hereby promise”] one performs but does not describe the act of promising’ (ibid.: 51). At the same time, he ended up saying that promising is both evaluative and descriptive, as we saw above. Searle has changed his mind and has come to see promising not only as a declaration but as an assertive as well. Because of this, we can now conclude that at least in some sense one is describing the act of promising (rather than one’s mental state) when one utters the relevant words. According to Searle, each speech act has a sincerity condition. The kind of sincerity condition depends on the intentional state that it expresses. The sincerity condition of assertives is belief. In other words, assertives normally express beliefs and the speaker is committed to the truth of the proposition expressed. One of the conditions of making a promise is that the person who makes the promise believes she does so in virtue of uttering the appropriate words and that she has actually done so. Searle regarded promising as a commissive already in the sixties – at least implicitly. This comes out in his discussion of objections to his argument against the naturalistic fallacy thesis. He writes for instance: ‘By definition, promising is undertaking an obligation or commitment to do something.’ (1969: 193; see also CSR: 146 and MLS: 147-8) This resembles statement [5] to a large extent. Hare (1964) regards the combination of statements [2] and [5] as the constitutive rule and there is reason to believe Searle would go along with this. This can be gathered from one of Searle’s own glosses on the constitutive rule of promising, namely, ‘to make a promise is to undertake an obligation’ (1964: 56). (Once statements [2] and [5] have been combined, statements [4] and [6] should also be combined as the overall conclusion of the argument.) If statement [5] is seen as part of the constitutive rule, committing oneself to the promise one makes should be seen as one of the conditions C of the rule of promising. Something like this is suggested by Searle’s reaction to an objection he considers that has the following as its core: ‘[B]y literally and unreservedly using that word we are committing ourselves to the institution of promising. But that involves an evaluation’ (1969: 194). Searle reacts by saying that in a sense this is a restatement of his argument. He accepts it, as long as it is not taken to undermine the objectivity of the promise made.

180

Frank A. Hindriks

The sincerity condition of commissives is intention. In other words, one expresses one’s intention when one makes a promise. Intentions involve commitments: ‘Every commissive is a commitment by the speaker to undertake the course of action represented by the propositional content.’ (MLS: 149) Arguably, commissives involve evaluativity. If, as was suggested in the previous paragraph, the fact that making a promise requires the speaker to commit himself is part of the conditions C, we have located the evaluative premise in Searle’s argument – statement [3], the claim that conditions C obtain. Considering Searle’s work up to MLS, however, there is no firm basis for attributing this view to Searle. I will argue, however, that an evaluative condition can be identified in what Searle says about promising as a declaration, the third kind of speech act to which the act of making a promise belongs. ‘Hereby’ is the obvious word by which many declarations can be recognized. It follows that the utterance of statement [1] is a declaration, given the truth of statement [2]. Declarations ‘bring about a change in the world by representing it as having been changed’ (ibid.: 150). Whereas all speech acts are acts by themselves, ‘[d]eclarations are unique among speech acts in that they actually make changes in the world solely in virtue of the successful performance of the speech act’ (ibid.; emphasis added). Declarations are possible only because of the existence of extralinguistic institutions. My strategy for showing that Searle’s argument does not undermine the naturalistic fallacy thesis consists of arguing that statement [3] is an evaluative synthetic statement according to Searle’s own theory of institutional facts. Searle (1964: 49-50) considers the objection that statement [3] is evaluative and claims it begs the question as long as the evaluativity is not located precisely. The next section is devoted to locating the evaluativity, i.e. to making explicit an evaluative condition that is part of the conditions C of statement [3]. 3.2 Searle’s building blocks, the 1964 argument, and the negative claim In 1964 Searle did not have a full-blown theory of institutional facts. He developed such a theory only recently in CSR. According to this theory, institutional facts involve normativity. Searle presents the theory as having three building blocks: the assignment of function, collective intentionality, and the constitutive rule. I will consider them in turn, as they will play a crucial role in the subsequent argument for the two main claims of this paper mentioned in the introduction. Starting with the first, Searle circumscribes the notion of a

Institutional Facts and the Naturalistic Fallacy …

181

function by explicating some of its features. A function is something that is observer-dependent, because it does not exist independently of intentional agents. A function is something that can be assigned to or imposed on an object or a phenomenon. Some kind of purpose is involved in all assignments of function and the performance of a function is conducive to achieving that purpose. (CSR:14; MLS: 121-22.) Although something is observer-dependent because it depends on intentional agents, intentionality itself is observer-independent, because its existence is not dependent on anybody’s attitude towards it (MLS: 117). Functions can be assigned to natural phenomena and to objects that are specifically created to serve a certain purpose. Furthermore, functions may be agentive or non-agentive, depending on whether a practical or a theoretical purpose is involved. In case of agentive functions, ‘the ascription of function ascribes the use to which we intentionally put these objects’ (CSR: 20; emphasis in original). A chair is an example of agentive functions, as is a log that is used as a bench, because the objects are used for a certain practical purpose. An example of non-agentive functions is a heart that pumps blood in order to promote life and survival. This ascription of a function is theoretical. (Note that one can accept Searle’s account of agentive functions without accepting his account of non-agentive functions. The latter implies that functions did not exist before there were intentional agents, a conclusion that some will find implausible. Only agentive functions play a substantial role in Searle’s social ontology.) One of the functions that can be assigned is intentionality or representation. If that is done, derived intentionality is created, which – in contrast to underived or intrinsic intentionality – is observer-dependent. Assigning the function of representation results in meaning, according to Searle. Meaning is a form of derived intentionality: ‘Sometimes the agentive function assigned to an object is that of standing for or representing something else. […] in this case we have intentionally imposed intentionality on objects and states of affairs that are not intrinsically intentional.’ (CSR: 21; MLS: 141) The function of representation is a specific kind of agentive function. The second building block of Searle’s social ontology is collective intentionality. Collective intentionality is a notion that is meant to capture the idea that people can share intentional states with one another and constitute a social group in virtue of that. The form of a collective intention may simply be ‘we intend’ or ‘we are doing so-and-so’ (CSR: 26). If the collective intentionality is shared between two or more persons, a social fact exists (MLS: 121). Searle regards shared collective intentionality both as a necessary and a sufficient

182

Frank A. Hindriks

condition for a social fact. An example of a social fact is two people who go for a walk together. The intentional state that is shared here is an intention in action. Collective intentionality can be used for assigning functions, just as individual intentionality can (CSR: 39). For instance, two people may use a log as a bench, and collectively think about it as a bench. In doing so they assign the function of being a bench to it. Just as in the case of sharing an intention in action, a social fact comes into existence in virtue of this shared assignment of function. Searle takes collective intentionality to be psychologically primitive. Because of this, collectively sharing an intentional state is something different from having the same individual intentional state as someone else. We have seen that the function of representation is a special agentive function that is constitutive of meaning. Such a function may be recognized or accepted by a collective instead of merely being assigned by an individual. A collectively accepted representation function is constitutive of an institutional fact. Collective acceptance is a special case of collective intentionality; it is a kind of collective belief. This implies that institutional facts are combined occurrences of special cases of the two building blocks of social reality, the assignment of function and collective intentionality. Searle introduces the notion of a status function for this combination of collective acceptance and the function of representation (although he is less explicit than I am about the relation between the function of representation and the status function; I defend my reading in Hindriks 2003). A fact is an institutional fact just if it involves the assignment or performance of a status function. Searle provides two explications of the kind of collective intentionality involved in institutional facts. I will call them the function account and the deontic power account. In the function account, Searle relates the collective intentionality required for institutional facts to the “counts as” locution: the form of the assignment of the new status function can be represented by the formula “X counts as Y in C.” This formula gives us a powerful tool for understanding the form of the creation of the institutional fact, because the form of the collective intentionality is to impose that status and its function, specified by the Y term, on some phenomenon named by the X term. (CSR: 46; emphasis in original) Furthermore, it is said that ‘we have good reasons to believe that the “counts as” locution specifies a form of collective intentionality’ (CSR: 95). Searle uses the “counts as” locution in order to explain the notion of a status function. In addition to this, he uses it to provide an account of the notion of a constitutive

Institutional Facts and the Naturalistic Fallacy …

183

rule. This prompts the question how the status function relates to the constitutive rule. Searle regards the constitutive rule as the third building block of social reality. Up until CSR, it was the key notion he used for explaining what institutional facts are. In my summary of Searle’s theory of CSR thus far I have explained what institutional facts are without referring to constitutive rules. The “counts as” locution is the form of the constitutive rule, but we have found that it is the form of the status function as well. Thus far, all this suggests that conforming to constitutive rules is nothing but assigning or performing status functions. The account in terms of status functions surpasses the previous account in terms of the constitutive rule, because it provides a more extensive analysis – notably, it includes collective intentionality. One may want to conclude from this that the constitutive rule has become redundant in Searle’s framework. There are, however, a few statements in CSR that block this conclusion. The crucial passage is this one: [W]here the imposition of status function according to the formula [the counts as locution] becomes a matter of general policy, the formula acquires a normative status. It becomes a constitutive rule. This is shown by the fact that the general rule creates the possibility of abuses that could not exist without the rule, such as counterfeit money […]. (CSR: 48; emphasis in original) Searle subsequently relates the possibility of abuse and the normative status the “counts as” locution acquires to codification. The possibility of abuse is not relevant for all institutional facts – although Searle seems to think it is – as not all institutional facts involve codification – a claim on which Searle agrees. The implication is that on pain of inconsistency the constitutive rule is not relevant for all institutional facts on Searle’s current views. Notably, it is not relevant for promises anymore, as the institution of promising has not been codified. On this interpretation, I should switch from talking about constitutive rules in the context of promises to talking only about status functions. What follows, however, does not depend on this interpretation of Searle. As a consequence, I will abstract from these problems and talk about constitutive rules and status functions interchangeably. Those who accept my interpretation, which I defend in Hindriks (2003), can replace ‘constitutive rule’ by ‘status function’ in the remaining part of this paper. In addition to the function account, Searle offers what I will call his deontic power account of the collective intentionality involved in institutional facts.

184

Frank A. Hindriks

According to this account, the structure of the collective intentionality that constitutes an institutional fact is ‘We accept (S has power (S does A))’ (CSR: 104 and 111). This implies that institutional facts involve power (CSR: 96; MLS: 132). The power involved is conventional or deontic power and may be positive or negative, enabling or requiring actions of an individual or a group. In other words, institutional facts confer rights and obligations to people (CSR: 100). For our purposes, these accounts can be regarded as complementary (but see Tuomela 1997 and Hindriks 2003). The deontic power account provides the basis for Searle’s claim that ‘where human institutions are concerned, we accept a socially created normative component’ (CSR: 146). For the case of promising, this normative component is explicated thus: ‘It is precisely because of the rule that making a promise counts as undertaking an obligation that we recognize that certain kinds of behavior within the institution of promising are acceptable and certain other kinds are remiss.’ (CSR: 147) According to this passage, there is a close link between the constitutive rule and obligations. It is important to appreciate the difference between uttering the relevant words within and outside the institution of promising. Imagine that Jones utters the words ‘I hereby promise to pay you, Smith, five dollars’ as an actor in a play (the example is from Searle 1964: 45). It would be awkward if Smith would come up to Jones after the play and request the five dollars. Only if the words have been uttered within the institution of promising the speech act is collectively recognized. Only if the speech act is collectively recognized has a promise really been made. The difference between uttering the words within or outside of the institution resides in the presence or absence of collective acceptance. The obligation to do as promised depends at least in part on such collective acceptance. These insights support the two main claims of this paper, the negative and the positive one. The negative claim is that Searle fails to show that the naturalistic fallacy thesis is mistaken on his own (later) terms. The positive claim is that statements about institutional facts and the related rights and obligations are epistemically objective. The remainder of this section is devoted to the negative claim, whereas the positive claim is the focus of the next section. Recall that Searle concluded in 1964 that the naturalistic fallacy thesis is mistaken because he had argued that within systems of constitutive rules one can derive an ought-statement from is-statements (57). It is now clear that this conclusion does not follow given Searle’s current conception of institutional facts and collective acceptance. The reason for this is that institutions are themselves

Institutional Facts and the Naturalistic Fallacy …

185

normative. If for example a promise is made, the normative nature of promising is already taken for granted. The collective recognition or acceptance that is constitutive of the obligation is already presupposed. The obtaining of collective acceptance is an empirical matter. All this can be appreciated by making it explicit as one of the conditions C of the constitutive rule. A premise [3a] should be added according to which Jones’ utterance is collectively accepted as a promise with the concomitant rights and obligations. A premise of Searle’s broad argument is that statement [3] – and by implication statement [3a] – is non-evaluative. It will now be clear that this premise is false. Hence, Searle’s broad argument is unsound. The narrow argument is valid and sound. However, statement [3] covers several premises at once, not all of which had been made explicit by Searle. This made it hard to see whether one of them is evaluative. The (negative) conclusion that can be drawn is that Searle’s full-blown account of institutional facts in CSR and MLS implies that his 1964 attack on the naturalistic fallacy thesis fails on his own (later) terms. This means that the example concerning promising is not a counter-example to the naturalistic fallacy thesis. 3.3 Collective acceptance, epistemic objectivity, and the positive claim This does not mean, however, that Searle’s argument failed to show anything at all. Another main conclusion Searle drew from his 1964 argument is that ‘[i]t is often a matter of fact that one has certain obligations, commitments, rights, and responsibilities, but it is a matter of institutional fact, not brute, fact [sic]’ (56). The core of the positive claim of this paper is that Searle’s argument does indeed imply this claim. The preceding discussion of Searle’s current conception of institutional facts makes it somewhat clearer what it might mean: rights and obligations are matters of deontic power constituted by collective acceptance of a status function. The full positive claim presented below can only be understood once further aspects of Searle’s current views are appreciated, notably those concerning his notions of collective acceptance and epistemic objectivity. In addition to distinguishing between observer-independence and observerdependence, Searle distinguishes between epistemically objective and epistemically subjective matters. This latter distinction allows us to see more clearly what Searle’s 1964 argument implies. We saw that something is observer-dependent if it depends on human beings and their intentional attitudes for its existence. Searle lists the following examples of observer-dependent features:

186

Frank A. Hindriks

the features of ‘being money, property, a knife, a chair, a football game, or a nice day for a picnic’, and adds that generally the natural sciences deal with observer-independent features, whereas the social sciences are concerned with observer-dependent features (MLS: 116-17). Features of the world that do not depend on intentionality are observer-independent. Examples of observer-independent features are mass and molecular structure. The distinction between epistemic objectivity and epistemic subjectivity does not concern features of the world, but applies to judgments or statements and their truth or falsity. In CSR, Searle writes the following: Epistemically speaking, “objective” and “subjective” are primarily predicates of judgments. We often speak of judgments as being “subjective” when we mean that their truth or falsity cannot be settled “objectively,” because the truth or falsity is not a simple matter of fact but depends on certain attitudes, feelings, and points of view of the makers and the hearers of the judgments. (8) Similarly, in MLS Searle writes: ‘[A] statement is considered objective if it can be known to be true or false independently of the feelings, attitudes, and prejudices of people. A statement is epistemically subjective if its truth depends essentially on the attitudes and feelings of observers.’ (44) In contrast, something is epistemically subjective if it is a matter of taste or opinion, if its ‘truth or falsity depends on the attitudes, preferences, and evaluations of observers.’ (MLS: 44) Searle uses the qualification ‘epistemic’ in order to distinguish the distinction from that between ontologically objective and subjective states. Searle takes mental states to be ontologically subjective (CSR: 8-13 and MLS: 44). The distinction between epistemically objective and subjective is not entirely unproblematic. I will mention two problems. First, the distinction is defined in an asymmetric way. Whether a statement is epistemically objective is said to turn on whether it can be known to be true, whereas whether a statement is epistemically subjective is said to turn on whether it is true or false (in)dependent of feelings, attitudes, and prejudices. This leaves one wondering whether or not the distinction is a distinction between different kinds of justification. Second, it is left implicit what kinds of attitudes are relevant to the distinction. One may wonder, for instance, whether they are propositional or non-propositional attitudes, or whether they are individual or collective attitudes. The fact that Searle claims of observer-dependent features that they depend on our intentional attitudes as well only adds to the confusion.

Institutional Facts and the Naturalistic Fallacy …

187

We can make some headway with respect to this last issue by considering some of Searle’s examples. Searle regards the statements ‘Rembrandt is a better painter than Rubens’, and ‘the moon is beautiful tonight’ as epistemically subjective (CSR: 8 and 12-13). These examples show that what is at stake are evaluations that are matters of taste or opinion. Furthermore, they refer to the tastes and opinions of individuals. In contrast, Searle considers statements about such diverse facts as that someone is in pain or thirsty, that a stone is used as a paperweight, that something is a screwdriver, and that a piece of real estate has a certain value as epistemically objective (CSR: 10-13; MLS: 44-45 and 94). The first two examples pertain to a personal experience or sensation. The third and fourth examples pertain to assignments of causal agentive functions, either by an individual or by a collective. The fifth and final example involves an institutional fact. It is difficult to extract an exact positive definition of epistemic objectivity from these examples, i.e. a definition that goes beyond the claim that a statement is epistemically objective if it is not epistemically subjective (even this characterization is not unproblematic because of the asymmetry in Searle’s formulations mentioned above). Rather than trying to extract such a definition, I will try to gain further understanding of the notion by making explicit certain definite features of it. Subsequently, we can focus on the one relevant to institutional facts. Three things are clear about the notion of epistemic objectivity. First, both facts about individuals and facts about collectives may be epistemically objective. This follows from the fact that for instance the function of a stone of being a paperweight can be assigned by an individual, while the institutional fact of a piece of real estate having a certain value requires a collective. Second, both facts that involve and facts that do not involve a function can be epistemically objective, as follows from the example of pain that does not involve a function. Third, statements about institutional facts are epistemically objective. Searle claims this explicitly in MLS: ‘[A] set of observer-relative institutional phenomena [involving deontic power] can have an epistemically objective existence even though their ontology is observer-dependent’ (117). It is this third conclusion that is important for what follows. (Note that Searle switches here from using the term ‘epistemic objectivity’ for statements to using the term to characterize the way in which institutional phenomena exist. I allow myself to do the same in the remainder of this paper.) In as far as the ‘ought’ of the conclusion depends on collective acceptance (in contrast to the extent to which it depends on personal commitment on the side

188

Frank A. Hindriks

of the speaker; see section 3.1), the obligation is a social one. To this extent, the conclusion of Searle’s (narrow) argument can be paraphrased as: Jones ought to pay Smith five dollars because the collective of which Jones is a member accepts or agrees that Jones ought to pay Smith five dollars. Whenever there is collective intentionality that is satisfied in fact, there is a social group (see Hindriks 2001 and 2002). Hence, the obligation is at least in part an obligation that is constituted by the collective attitudes of the members of a social group. This analysis provides the material for the (full) positive claim. This claim is that Searle’s 1964 argument implies the following thesis (formulated in terms of his 1995 social ontology): institutional rights and obligations – that are matters of deontic power constituted by collective acceptance of a status function – are epistemically objective even though they are observer-dependent in that they depend on the attitudes of a group of people. With respect to Searle’s guiding example, it follows that making a promise is an institutional fact that due to its dependence on collective acceptance involves normativity. Collective acceptance and the related normativity are epistemically objective. This means that the true import of Searle’s argument is that something can be evaluative and objective at the same time; evaluative matters are not necessarily matters of taste or opinion. In other words, a statement can be at once both descriptive and evaluative.

4 Intentionality, normativity, and the strength of social obligations Having argued for the two main claims of this paper, I will end by analyzing the relation between intentionality and normativity in Searle’s framework somewhat further. The first part of this section will focus on collective intentionality and normativity, whereas the second part looks at intentionality and normativity more generally. Searle uses a wide range of expressions for describing the collective acceptance that is required for institutional facts throughout CLS and MLS. Apart from the term acceptance, he frequently uses words like agreement, recognition, and authorization. In addition to these, Searle talks of going along with, permitting, and acknowledging. All of these terms describe collective intentional states that are constitutive of deontic power a person or a group of persons may have. The key thing to notice is that there is a difference in strength between these words. Going along with is weaker than agreeing, and accepting is weaker than

Institutional Facts and the Naturalistic Fallacy …

189

authorizing. It is clear that Searle himself believes there to be a difference in strength between these terms. This is apparent when he discusses money as an example of an institution, and claims that ‘there has to be collective agreement, or at least acceptance, both in the imposition of that status on the stuff referred to by the X term and about the function that goes with that status.’ (CSR: 44; emphasis added) I suggest using this observation as the basis for a thesis that I take to be rather plausible: the strength of a social right or obligation is proportional to the strength of the collective intentionality involved. I propose to call this the proportionality thesis. Searle does not make a claim like this. However, the difference in strength of the words he uses to characterize the collective acceptance of different institutional facts suggests it. In contrast to Searle, Gilbert (1989, 1996, 2000) and especially Tuomela (1995, 2000) distinguish explicitly between different degrees of normativity and relate this to collective intentionality. Perhaps they would subscribe to the proportionality thesis. A further difference between them is that Gilbert and Tuomela take the domain in which social obligations exist to be wider than Searle does. Searle believes that all social obligations are institutional obligations. Recall that Searle regards all facts that involve shared collective intentionality as social facts, whereas institutional facts require a special kind of collective intentionality: collective acceptance (of a status function). Only facts that depend on collective acceptance involve obligations. This implies that a mere social fact such as going for a walk together does not involve social obligations on Searle’s view. On Gilbert’s and Tuomela’s views it does. They believe that social facts involve joint commitment or agreement – notions closely related to collective acceptance. As an example of a social but non-institutional obligation consider Gilbert’s claim that the parties of a duo that is walking together have ‘a right to the other’s attention and corrective action’ (1996: 180). Social rights and obligations depend on collective intentionality. There is, however, a more general relation between normativity and intentionality in Searle’s framework. In fact, all normativity can be traced back to intentionality (in the sense of aboutness) in Searle’s framework. This holds for the sincerity conditions of speech acts (including commissives), for the individual or collective assignment of causal agentive functions, as well as for the collective acceptance required for institutional facts. Normativity does not occur in Searle’s framework in any other place than these three. Satisfying sincerity conditions is intentional and therefore not choice or action-independent. Take for example the case of assertives, which have belief as

190

Frank A. Hindriks

their sincerity condition. Even though having a certain belief is involuntary, expressing it is not. An assertive may involve a commitment to its truth, but again this is an intentional matter. Assignment of causal functions and acceptance of status functions are choice or action-dependent. The implication is that Searle’s framework does not allow for intention-independent moral facts. Arguably, this can be seen as an implicit denial of the existence of such moral facts.

5 Conclusion The main goal of this paper has been to analyze the implications of Searle’s recent social ontology with regard to his earlier argument about promising and the naturalistic fallacy thesis. I have argued for four claims, which I will summarize here starting with the two main ones. First, Searle’s attack on the naturalistic fallacy thesis fails on his own later terms. One of the conditions C of the “counts as” locution of promising is collective acceptance. According to statement [3], conditions C, including the requirement of collective acceptance, are satisfied. This is an empirical claim. Collective acceptance is constitutive of social obligations. This implies that statement [3] is evaluative, in contrast to what Searle claimed. It follows that he has failed to show that the naturalistic fallacy thesis is mistaken. Second, the other main conclusion Searle drew from his argument does follow from it – given his theory of institutional facts. On that theory, it is often a matter of objective institutional fact that one has certain obligations. Using the terms Searle deploys in presenting his social ontology, this means that the argument supports the thesis that statements about institutional facts and the concomitant rights and obligations are epistemically objective. This is the case in spite of the fact that the institutional facts as well as the rights and obligations exist in virtue of collective acceptance, which is observer-dependent. The two subsidiary claims that I have defended are as follows. First, the strength of social obligations is proportional to the strength of the underlying collective intentionality. Second, all normativity in Searle’s framework can be traced back to individual or collective intentionality. Searle cannot accommodate intention-independent moral facts within his framework.

Institutional Facts and the Naturalistic Fallacy …

191

Acknowledgements I gratefully acknowledge comments from Garrett Cullity, John Haldane, Uskali Mäki, Emma Rooksby, Raimo Tuomela, and the members of the philosophy reading-group Gavagai in Rotterdam. I also wish to thank Allan Gibbard for a conversation on the topic of this paper at the Erasmus Summer School on Social Ontology of 2000.

References Anscombe, Gertrude E. M. “On Brute Facts.” 22-25 in Ethics, Religion and Politics. Collected Philosophical Papers of G. E. M. Anscombe. Oxford: Basil Blackwell, 1981 [1958]. —. “Modern Moral Philosophy.” 175-95 in Hudson 1969 [1958]. —. “Rules, Rights and Promises.” 97-103 in Ethics, Religion and Politics. Collected Philosophical Papers of G. E. M. Anscombe. Oxford: Basil Blackwell, 1981 [1978]. Austin, John L. How to Do Things with Words. Oxford: Clarendon Press, 1962. Brink, David Moral Realism and the Foundations of Ethics. New York: Cambridge University Press, 1989. Flew, Antony G. N. “On Not Deriving ‘Ought’ from ‘Is’.” 135-43 in Hudson 1969 [1964]. Frankena, William K. “The Naturalistic Fallacy.” Mind 48 (1939): 464-77. Gilbert, Margaret. On Social Facts. Princeton (NJ): Princeton University Press, 1989. —. Living Together. Rationality, Sociality, and Obligation. New York: Rowman and Littlefield Publishers, 1996. —. Sociality and Responsibility. New Essays in Plural Subject Theory. New York: Rowman and Littlefield Publishers, 2000. Hare, Richard M. “The Promising Game.” 144-56 in Hudson (ed.). London: MacMillan, 1969 [1964]. Hindriks, Frank A. “Social Ontology, Collective Intentionality, and Ockhamian Skepticism.” Grazer Philosophische Studien (2002). —. “Social Groups, Collective Intentionality, and Anti-Hegelian Skepticism.” In Realism in Action: Essays in the Philosophy of Social Science, Matti Sintonen, Petri Ylikoski, and Kaarlo Miller (eds.), Dordrecht: Kluwer Academic Publishers, 2002. —. “The New Role of the Constitutive Rule.” American Journal of Economics and Sociology (2003). Hudson, William D. (ed.) The Is/Ought Question. A Collection of Papers on the Central Problem in Moral Philosophy. London: MacMillan, 1969. Moore, George E. Principia Ethica. London: Cambridge University Press, 1903. Searle, John R. “How to derive “ought” from “is”.” Philosophical Review 73 (1964): 4358.

192

Frank A. Hindriks

—. Speech Acts. An Essay in Philosophy of Language. Cambridge: Cambridge University Press, 1969. —. Expression and Meaning. Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press, 1979. —. Intentionality. An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press, 1983. —. The Construction of Social Reality. New York: Free Press, 1995. —. Mind, Language, and Society: Doing Philosophy in the Real World. London: Weidenfeld and Nicolson, 1999. Tuomela, Raimo. “Searle on Social Institutions.” Philosophy and Phenomenological Research 67, 2 (1997): 435-41. —. The Importance of Us. A Philosophical Study of Basic Social Notions. Stanford (CA): Stanford University Press, 1995. —. Cooperation. A Philosophical Study. Dordrecht, Kluwer Academic Publishers, 2000.

Group Action and Group Responsibility

Group Responsibility and the Consensus Account of Justified Belief

193

194

Pekka Mäkelä and Raimo Tuomela

Group Action and Group Responsibility

195

Group Action and Group Responsibility Pekka Mäkelä and Raimo Tuomela

Abstract In this paper a social group’s (retrospective) responsibility for its actions and their consequences are investigated from a philosophical point of view. Building on Tuomela’s theory of group action, the paper argues that group responsibility can be analyzed in terms of what its members (jointly) think and do qua group members. When a group is held responsible for some action, its members, acting qua members of the group, can collectively be regarded as praiseworthy or blameworthy, in the light of some normative standard, for what the group has done. The paper gives a necessary and sufficient conditions analysis of a group’s responsibility for its actions and their outcomes, and the conditions can be cashed out in terms of the group members’ joint (and other) actions.

Introduction We commonly attribute actions to social collectives and especially to proper social groups. Thus, we use locutions like ‘firm F produced the goods G’, ‘nation N1 attacked nation N2’, ‘the board dismissed Jones’, ‘the team scored’, and so on. Not only do we attribute actions to groups but we ascribe responsibility for (their) actions to social groups as well. Locutions like the following are quite common. “The Exxon Corporation was responsible for the worst oil spill in American history when the Exxon Valdez ran aground in Alaska”, or “Germany was responsible for the Second World War”, or “The club as a whole is to blame for being relegated”. On the basis of examples like these it seems reasonable to accept this common-sense view at least in part and to think that true statements of the above kind can be made. We will do so here and investigate some central philosophical and conceptual problems related to actions performed by groups and responsibility for such actions. Our main interest is in the (retrospective) responsibility for actions performed by groups. Roughly, when we hold a group retrospectively responsible for some action we take the members of the group, qua members of the group, to be praiseworthy or blameworthy for what the group has done in light of some normative standard. Moral standard is of course the paradigm but an action can also be appraised from other evaluative perspectives, e.g. from the point of view of legal norms, etiquette, or rational long term interests.

196

Pekka Mäkelä and Raimo Tuomela

In this paper we consider only cases of blameworthiness. To begin with, a group is blameworthy for performing X only if there is an acceptable normative standard which prohibits X. Due to the possible asymmetry between praiseworthiness and blameworthiness, our analyses may need some tinkering when praiseworthiness is at stake. It is fair to hold an individual agent responsible for some action only if the agent has certain capacities (e.g. rationality, capability of intentional action) and the agent freely exercised these capacities in acting. Thus, there are both internal and external factors to be taken into account here when deciding whether the action was “up to the agent” in the right way. In this paper we will look for precise criteria for group action and, especially, for the group’s responsibility of such action. We will focus on the following questions. a) When is an action attributable to a group as its action? b) Under what conditions do the actions by the members of the group generate action attributable to the group? c) What does it mean that a group bears responsibility for its action? d) Under what conditions is it justified to hold the group responsible for the actions performed by the members of the group? Questions a) and b) will be answered in terms of the account given in Tuomela (1995), Chapter 5, and our discussion will mainly summarize that account. As to c) and d), we will have new things to say, and in part our answer to a) and b) will be used to generate answers to the latter two questions. We claim that the strongest case for group responsibility is the responsibility in the following case concerned with a voluntary group. The responsibility concerns an internally and externally free group action, or the outcome of such action, as evaluated on the basis of a “relevant” normative standard (viz. evaluative perspective) that the group was aware of and accepted as its standard or at least in some sense ought to have accepted as its standard. The group action we are speaking about here is action that the group performed as an agent and that therefore binds the group members. When this group action is viewed from the perspective of the group members’ action which constituted it we can require that the latter action be we-mode action. By “wemode” we, roughly, mean thinking and acting as a group member with a collective commitment (for more see below, section 2). In our view group responsibility for an action performed by the group can be accounted for in terms of responsibility ascribed jointly to (individual) members of the group acting qua members of the group. Group responsibility understood along these lines is a central notion of collective responsibility if not the notion of collective responsibility.

Group Action and Group Responsibility

197

Group action is “internally free” on such action if it is performed by the members of the group, G, that none of the members of G was forced or coerced by other members of the group to agree with the group decision, or to act in accordance with it. A group action was “externally free” if there was no external pressure on the members of the group to adopt the goal, to make the plan and to perform the action. By external pressure we mean pressure exerted on the group by other groups or agents outside the group. If the action performed by the group was both internally and externally free we can say, in somewhat metaphorical terms, that the group acted out of its own free “group will”. In our account central criteria for a legitimate ascription of group responsibility to a member qua a member of the group are collective commitment and collective acceptance of the action (acceptance belief, intention or what have you) in question. Thus, collective commitment and collective acceptance are entailed by the proper we-mode. Our notion of we-mode action captures the core of group agency, or at least so we claim. (More precisely we-mode is primarily meant to concern attitudes and derivatively actions ensuing from attitudes held in the we-mode.) Social groups will here be assumed to be collectives capable of action and possessing authority systems (see Tuomela 1995, Chapter 4). Group membership here need not involve more than that an agent regards himself as a member of the group and that the other members of the group tend to regard him as a member of the group. Thus legitimate ascription of group responsibility is not based on or justified by mere membership. In our view group responsibility is, or can be, legitimately ascribed to both operative and non-operative members of the group. The distinction between the operative and the non-operative agents of a group is important here, and it applies also to the formation of group attitudes. Operative members are the actively acting members in virtue of whose action the group’s action comes about (cf. Tuomela 1995, Chapters 5, 6 and 7).1 Here we take it for granted that the notion of group responsibility admits degrees pretty much the same way as the notion of moral responsibility does. A non-operative member who learned only afterwards of the action performed by his group and accepted it retrospectively may bear less group responsibility than an agent who acted as an operative member. The strength of the group responsibility borne by non1 In the case of intentional actions performed by collectives also the non-operative agents, so to speak, “passively” participate in virtue of tacitly accepting (or at least being obligated so to accept) what the operative members do for the group (for more see Tuomela 1995, Chapter 6, and Tuomela 1992).

198

Pekka Mäkelä and Raimo Tuomela

operative members depends on their awareness of what the group is doing, their possibility to control or influence the group’s action, and so on. Our primary focus will be on voluntary and internally authorized groups. By voluntary groups we mean groups the membership of which is up to the members. We presuppose that the agents have had a choice whether to become a member of the group in question or not, as opposed to nations for instance where citizenship typically is not a result of the agent’s choice. Non-voluntary groups are obviously more problematic from the point of view of the collective responsibility of the non-operative members. Internally authorized groups as opposed to externally authorized groups are groups in which operative agents, either operative members or representatives, get their authority to act on behalf of the group from the members of the group in question. (For more on authorization see Tuomela 1995, Ch 5.)

1 Group Action In this section we will give an account of the conceptual nature of actions performed by social groups and do it primarily by investigating under what conditions attributions of actions to social groups can correctly be made. One of the central theses below will be that the actions performed by social groups are “made up” of, or “constituted” by, joint actions of persons. This thesis will be discussed and made precise below. Here is a simplified formulation of this thesis: If a group (with the agents A1,...,An as its members) does something X then at least some of its members, say A1,...,Am (m≤n) must, in the right circumstances, do something X1,...,Xm, as their parts of a joint action X (or of a joint action generating X); and in normal circumstances these parts serve to generate or “make up” X. Here, X1,...,Xm will be parts of a joint action of A1,...,Am which need not be of the type X but which still generates or brings about a token of X. In the case of intentional group action intentional joint action, and therefore we-mode “we-intentions” by the agents will be involved (see Tuomela 1995, Ch 2, Sections III and V for the notion of joint action)2. 2 We can say, roughly, that a member of a group we-intends in the we-mode to do X if and only if (i) he intends to do his part of X (as his part of X), (ii) has a belief to the effect that the joint action opportunities for an intentional performance of X will obtain; and, furthermore, (iii) believes that there is (or will be) a mutual belief among the participating members of the group or – at least among those participants who do their parts of X – intentionally as their parts of X – to the effect that the joint action opportunities for an

Group Action and Group Responsibility

199

Consider an example, say, a hockey team’s scoring. Some player, or perhaps players, did the scoring. We may say that it was the operative members of the team who did it and define that the operative members in the group relative to an action, X, are those in virtue of whose acting the performance of action X can be attributed to the group in this situation. We can also say in our example case that the team’s scoring was constituted by these operative members’ actions. More generally, it is a common situation in the case of groups with normatively specified positions that some of those positions and their holders are related to action in the analogous way and that the position holders thus designated as operative members are indeed operative members for a large range of actions, perhaps all the actions that the group in question is capable of undertaking. In such cases we need not always speak of specific actions in the context of group action, but may speak of operative members for all the activities or for some broad subclass such as decision making (or plan formation) and for carrying out decisions and plans. As actions by groups will in the core case be analyzed in terms of joint actions, some (additional) comments on them may start our discussion. An intentionally performed joint action must come about because of a weintention.3 Joint actions in our sense will include joint task-performances based on (at least believed) explicit or implicit agreement or joint plan. Thus typically these qualify: carrying a table jointly, playing tennis, satisfying a (many-person) contract, and sometimes, questioning-answering and conversing.4 Actions by groups are connected closely (and in a precise sense) to joint actions by the members of the group performed qua members of the group (see section 2). The basic content of our a general thesis on the nature of actions by groups is this: a group's intentional action requires (ultimately on conceptual grounds) that at least some of its members suitably act and that as a consequence the group will have acted. Indeed, a joint action appropriately performed in the we-mode in the right social and normative circumstances by the intentional performance of X will obtain. Shared we-intentions about which the participants are mutually aware are called joint intentions. (For discussion see Tuomela 1995, Chapter 3.) 3 Also such non-intentional joint actions as mistakes are included in our account (as long as the agents think that they are intentionally performing something together). 4 Joint actions in this sense should be distinguished from actions performed by (single) groups such as a group’s solving a problem, a bank staff’s performing its daily routines, a community’s electing a leader, a nation’s declaring war, and so on.

200

Pekka Mäkelä and Raimo Tuomela

operative members of the collective will be redescribable as a group action. Accordingly, if a group can be taken to be responsible for its actions, by way of this analysis we can speak of the members of the group being jointly responsible for such group actions. This thesis analyzes intentional actions performed by social groups. Such actions are obviously as central in the case of groups as they are in the case of single individuals. Thus, for instance, a group can be regarded as legally and morally responsible for its intentional action. The concept of joint action that our thesis on group action relies on requires some further remarks. First, as an important special case we technically allow actions performed by one single individual, in order to have unified terminology (cf. the President representing the country). Furthermore, we accept the following liberal usage: in the present context we need not be able to say that the operative agents jointly do X (even in the above wide technical sense) but only that they jointly (in the indicated technical sense) do something which will bring about X and is believed by them to bring about it. What they thus perform could be a joint action Y, non-identical with X. To see the reason for this consider a case of a state's entering a treaty where the operative agents jointly ratified the treaty and did whatever was needed; but of course they did not jointly enter the pact even if they jointly brought about that the state entered the pact.5 Our view of group action implicitly contains the idea that the non-operative members minimally tacitly accept or at least ought to accept the fact that the operative members perform X. Tacit acceptance here means not only acceptance as true but – what is central here – acceptance in the sense of not very seriously disagreeing with what the operative members do. The present requirement applies to all cases with internal commitment authorization, thus also to e.g. societies, which are collectives with involuntary group membership. On the other hand, in the case of teams, for instance, we need to require more, for in them either all the members are operative members or the non-operative members stand “in reserve” in the strong sense that they may be called in at any moment and could equally well in principle have been selected as operative ones. In this kind of strong cases we must require, it seems, that all group members 5 Note that in the present context of group actions the joint actions involved may be institutionalized. This does not remove the need for shared we-intentions for intentional action, as institutionalized practices still are based on we-intentions when intentionally performed (cf. Tuomela, 2001, Chapter 4).

Group Action and Group Responsibility

201

accept or at least ought to accept the we-intentions and group beliefs needed for group action.6 Given the above, there will be a shared we-intention to perform X (in our schematic case), and that involves the idea that the members of the group are collectively committed to bringing about X (see Tuomela 1995, Ch 3). But in the general case this does not require that all the members strongly participate in the group will in question. Thus, in the case of non-operative members this involves not a commitment to doing one’s part of X but the commitment to supporting passively – or at least not overtly opposing – what the operative members are doing when carrying out their commitments, given, of course, that the non-operative members are adequately informed about what the operative members are doing. Those non-operative members who also endorse “We will do X” are, however, committed to contributing, actually or potentially, to X. We can now spell out our preliminary thesis in a more precise, although stylized from and arrive at the following formulation, assuming that – on conceptual grounds – acting for the group to be a task rather than only a right of the operative members (we draw on Tuomela, 1995, p. 242f.): (IGA) A group, G, performed an action X intentionally in the social and normative circumstances C if and only if in C there were operative agents A1,...,Am of G such that 1) A1,...,Am, acting qua group members (and thus performing their positional tasks due to their exercising the relevant decision making system of G) intentionally jointly brought about X (viz., there was an action Y such that the operative agents intentionally jointly performed Y and this performance of Y generated, and was believed and purported by the operative members to generate, the result-event of X); 6 Here are a couple of remarks pertaining to all cases with internal authorization. First, concerning the ought-requirement, which is our minimal requirement, it does not come from the notion of group action but from the notion of an authority system (joint intention or plan formation system) of a social group. An exercise of an authority system results in a shared group-intention to perform some action as a joint action (giving group action as a result). More specifically, an exercise of an authority system will in the present kind of context be taken to consists either of the formation of a new group will or of applying (or “transferring”) a previously formed group will to specific cases by means of practical reasoning. Think, for instance, of a U.S. company’s decision to expand its business activities to Europe: specialization of this group will may result in the transferred group will to build a factory in Lisboa. (See Tuomela 1995, Chapters 4-5.)

202

Pekka Mäkelä and Raimo Tuomela

2) because of 1), the (full-fledged and adequately informed) non-operative members of G, as members of G, tacitly accept the operative agents’ intentional bringing about of X – or at least ought to accept it; 3) there was a mutual belief in G to the effect that there was at least a chance that 1) and to the effect that 2).7 Accordingly, given the right C, we claim that (IGA) is acceptable as an account of intentional group action. It also analyzes the sense in which individuals can bring about a group action and also the sense in which group actions can be said to be constituted by the we-mode actions (essentially: proper actions qua a group member) by their members and authorized representatives (for a longer discussion, e.g. of the notion of correct social and normative circumstances, see Tuomela 1995, Ch 5). What we have been analyzing above in (IGA) is group action in a sense involving the group as a whole. The exercise of the authority system was claimed to involve the whole group. Especially, in (IGA) the non-operative members are assumed to have the obligation to (tacitly) accept what the operative ones did. In our view, proper group action requires proper “groupness”, viz., at least implicit agreement or plan about relevant goals and views.

2 We-mode and Acting qua a Group Member As already mentioned, in our account of group responsibility we avail ourselves of the closely related notions of we-mode and acting qua a group member.8 For the purposes of the present paper it suffices to say that the group members act in the we-mode if and only if they act fully qua group members and are 7 Less than fully intentional joint actions can possibly be allowed here. Clause 2) presupposes that the non-operative members must have the right beliefs about what the operative members performed (viz., X): acting intentionally because of something p can only take place via correct beliefs about p. If they do not have any beliefs or only incorrect beliefs, we should not speak of a group action in a categorical sense – but at best of a kind of “conditional” or, better, condition-involving group action (see more on this Tuomela 1995, Chapter 5). Clause 2) (and mutual awareness in G concerning it) is a conceptual presupposition underlying situations where authority systems are being exercised. The phrase "there was at least a chance" in clause 3) is prompted by cases in which a group tries to do something chance-involving and will have done it intentionally if at all. 8 These notions have been discussed at length in many previous works by Tuomela (e.g. Tuomela 2000, 2001, 2002a,b, and Miller and Tuomela 2001).

Group Action and Group Responsibility

203

collectively committed to the group action in question. In the more general case we can speak of the group’s constitutive goals, norms, and standards instead of group action. Those constitutive goals, beliefs etc. can be called the ethos of the group. Given this it remains to say what acting qua a group member amounts to. For our present purposes we can say the following. We assume that the group in consideration is concerned with certain specific topics in it activities. Let us speak of the group’s realm of concern. Obviously the topics its ethos is about must at least belong to this realm, if not exhaust it. Considering the general case to be that of a structured group, we can classify the types of actions within the group’s realm of concern as follows: 1) positional actions related to a group position, which include i) actions (tasks) that the position holder in question ought to perform in certain circumstances and ii) actions that he may perform in some circumstances; 2) actions which the ethos of the group requires or allows; 3) actions which are based on agreement makingetc. which have not been codified by the group but which still are consistent with actions in 1) and 2); 4) freely chosen activities which include actions and activities not within classes 1)-3) but which, although not incompatible with them, still are actions within the realm of concern of g and rationally (understood broadly to amount to reasonably) collectively accepted by, or acceptable to, the members of G as such actions. Basically, acting as a group member is to intentionally act within the group's realm of concern. It can be either successful action or an unsuccessful action. What is required is that the group member in question will intentionally attempt to act in a way related to what he takes to be the group’s realm of concern such that he does not violate what he takes to be the group’s ethos. Thus full success will not be required. There may thus be failures due to false beliefs about the group’s norms and standards, due to lack of skill, or due to environmental obstacles. Thus acting as a group member in the positional case, viz. in a structured group, is equivalent to acting intentionally in one of the senses 1) – 4) or attempting so to act. Below this notion will be meant when we speak of acting or functioning qua a group member. Note especially that one can act within the realm of the group's concern but intentionally fail to obey the norms and standards of the group. Then the group may not be responsible for this group member’s action, except perhaps when it has promised – or has otherwise obligated itself – to see to it that its members do not violate its ethos. Considering the group’s, E, we assume that the group members have collectively accepted it and are collectively committed to it. We can say that E is

204

Pekka Mäkelä and Raimo Tuomela

internal to the group as it depends on the group members collective acceptance and commitment. Note that E can be in contradiction with the normative standard, N, in light of which the praiseworthiness or blameworthiness of the group’s action is to be judged. N is assumed to be external to the group in the sense that it is not, at least not solely, dependent on the collective acceptance and commitment of the members of group the action of which is to be appraised. N offers an external yardstick against which the group’s action is measured. We may say that E should, and of course could, be compatible with N but is not necessarily so. (E.g. criminal groups can be proper groups.) Actions in 1) are of course typical positional actions that accordingly qualify as acting qua a member of g in one’s position. Subclass ii) of 1) thus consists of actions that the holder of position may choose from. However, classes 2)-4) also can occur in the positional case and also in other cases. Note that in the case of unstructured groups, class 1) is empty.

3 Group Responsibility In this section we will discuss responsibility for group action. To recapitulate we claim that the strongest case for group responsibility is responsibility of a voluntary group for an internally and externally free group action, or the outcome of such action, performed in the we-mode, given that the action, or the outcome, is relevant with respect to some normative standard. In our view group responsibility is to be understood in the sense of responsibility ascribed jointly to (individual) members of the group qua members of the group for an action performed by the group, that is by at least some of the members of the group acting qua members of the group in the we-mode. 3.1 Structured Groups We will first discuss the central case of structured groups, viz. groups with normatively governed positions. Our treatment will assume internal authorization and voluntary group membership. It seems that our account can rather easily be modified to account also for groups with involuntary membership. The removal of the assumption of internal authorization again seems to involve considerable new problems. Authorization is highly central in our present model. The operative members for decision and/or acting have been authorized by the group members. It

Group Action and Group Responsibility

205

is in virtue of operative members activities that activities are attributed to the group, be these activities, for instance, plan formation or carrying out those plans. In general, the group is responsible, and its members are jointly responsible, for what it does – and especially for what it intentionally does. Even in the case where the operative members have been internally authorized by the group members to act concerning a certain scope of issues, there are many cases to consider. In the case of a task group (team), for instance, we can think that all the members are collectively responsible to an almost equal degree for the team’s action, say X, as the non-operative ones can be regarded as being in reserve – in the sense that they might equally well (excluding perhaps some contingent practical issues from this equality claim) has been chosen to be operative members. However, even here those members who actually participate in the means-action are more responsible than those in reserve because they were put in a situation where they could have “second thoughts” before action and because their acting causally brought about the state which was describable as the group’s action or a purported consequence of the group’s action. Sometimes a group may use non-member operatives (e.g. hired agents) to act for the group. In such a case the operative members are responsible only for the particular actions they were hired to perform, but not for the group action that they served to bring about or to get performed. At the other extreme we may have groups in which the internally authorized operative members have been authorized for a large range of activities and perhaps in non-specific ways. That is, the range of actions that they have been authorized to decide about or to carry out have been specified in general and perhaps vague terms. In such cases the operatives have much power to act for the group, and this power is reflected in much responsibility attributable to them as compared with the responsibility attributable to the non-operative members. This responsibility of the operative members consists of the component that applies to every group member and the special component that relates to their personal and joint decision making (and/or other acting) when they have to select one out of several possible actions and to use certain specific means to carry out the selected action. (Cf. the operative members of Sonera would have to decide whether or not they should buy a 3G-license in, say, India.) In general, we may say that even when the operative members are assumed to be internally authorized by the group members, this authorization may be of many kinds and this is reflected in the amount of control that the non-operatives have over the operatives’ actions. This amount of control contingently depends in an obvious way on the range of topics the authorization concerns,

206

Pekka Mäkelä and Raimo Tuomela

the range of possible alternatives within each topic within this scope, and the specificity or concreteness of the actions involved, on the means and manners used for carrying out the activities. Furthermore, the type of accountability of the operative members vis-à-vis the non-operative members is a relevant variable here. Thus, for instance, the operative members might have to account for every action right after its performance, at the one extreme, or they might only be accountable in a wholesale way about their totality of activities during a perhaps relatively long span of time (say, so and so many years). As these matters affecting the issue of controllability and accordingly the relative degrees of responsibility of the operative versus the non-operative members can be taken to have (and perhaps also the relative degree of individual members, if they play different roles), are contingent and highly contextual, we, qua philosophers, even in principle cannot go far into this territory without the collaboration of social scientists. It is quite often the case in structured groups that the position holder has certain collectively accepted goals to satisfy. The way in which these goals are to be reached is left for the position holder to consider and decide. In such cases the group bears responsibility for the action performed by its member in accordance with collective acceptance but also the individual responsibility of the operative member is bigger than in the case of precisely specified tasks. The group is responsible for the state of affairs brought about but the individual is responsible for the way the state of affairs was brought about. Openly defined positions set more of the moral burden on the shoulders of the operative member than narrowly defined positions. There may be a “grey area” of action even in the case of structured groups. Some tasks may be at best vaguely stated and the means to be used may not be normatively specified. In such a situation a group member may act in the belief that he acts as a group member (recall Section 2). If he really acted “correctly” as a group member the group would be responsible, as the activity would be attributable to it (or, better, would in conjunction with what some other operative members do be attributable to it). This kind of case is to be decided by contextual collective acceptance in which the group – or its operative members, if they are authorized also for this type of task – decides whether the member in question really acted as a group member in the situation in question. Our collective acceptance criterion as such holds here too. In some cases it may not be clear whether some action is in line with what the group has collectively accepted. Then it may not be solely up to the group to decide whether some action performed by a member was in accordance with what

Group Action and Group Responsibility

207

was collectively accepted. This is analogous with the individual agent case where it is not up to the agent to decide whether he was sane when he committed a crime. When it is open what has been collectively accepted with respect to a particular case, the group members and possibly “outsiders” too have to interpret the rules and norms of the group. They have to decide in light of what was collectively accepted in the past whether the action in question was performed qua a group member or not. Rational collective acceptance, or rather acceptability, is the ultimate criterion here, anyway. The reason for this is simply that the groups we are concerned with, viz. internally and externally free groups with voluntary membership, can be regarded as autonomous concerning their own affairs. When the freedom and voluntary participation involved here is combined with the kind of group-binding group action that we have been concerned with here, this gives a strong reason to assume that the group be able to control its own affairs on an autonomous basis (viz. based on its own decisions and views).9 Such acceptance can be agreement-based or it can be based on mutual (or shared) belief or shared tacit understanding about the case. No communication is necessary if the belief is based on similar previous cases and when there is communication it can be of the nonverbal kind. As an example let us consider a highly organized group like a criminal M.C. group, say Los Patheticos, of twenty members. Membership in the group is voluntary and the members have elected a board of three to run the group among themselves (internal authorization). The board is authorized to adopt strategies, make plans and form intentions for the group and also make the decisions concerning the execution of the plans. The idea of declaring a war against the rival group The Merciless has been discussed in Los Patheticos. Now the board, consisting of the members A, B, and C, decides to do this. They discuss the means of declaring a war. The best idea they can come up with is assassination. For simplicity’s sake they decide to execute the plan themselves. They make an agreement to assassinate a member of The Merciless, say F. The board informs the rest of the group about their decision. A, B, and C make a plan: A is supposed to find a getaway car, B to get the guns, C to be the 9 Note that there is weaker kind of group action based on we-attitudes (especially we-goals, sometimes we-beliefs only; cf. below for the notions) that can be called we-action. For instance, the group members may act towards the same we-goal (or base their action on the same normative we-belief that they ought to act towards some goal or whatever). This is I-mode action and does not make the group qua a group responsible. In other words, this supports our idea about connecting group responsibility to we-modeness.

208

Pekka Mäkelä and Raimo Tuomela

one who shoots. A is the driver, B is the back up guy who will shoot if something goes wrong, and C just concentrates on shooting two deadly shots at F. A, B, and C successfully execute this plan. In the above example, A, B, and C have a collective goal, more precisely a wemode goal, with the content ‘Our goal is to declare a war against The Merciless’. As a means to their goal A, B, and C see the joint action, namely the assassination of F. The agents have formed a we-intention with the content ‘We will assassinate F’. Additionally we can say that both the goal and the intention are accepted’for the group’ in the following sense. First, it is true that A, B, and C we-accept the goal and the intention as ‘our goal and intention’. That is, they accept the goal and the intention in the following we-attitude sense. We-attitudes are attitudes involving social beliefs. We consider a person’s we-attitude related to a “plain” attitude which has a certain content p. This attitude can be a plain want, goal, intention, belief, wish, etc. A we-attitude in its core sense is defined as follows: The person a) has the attitude towards p and b) believes that also the others in the group have the same attitude and also c) believes (or at least is disposed to believe) that it is mutually believed (or in a weaker case, plainly believed) that the members have this attitude. If the we-attitude so defined expresses a goal, it can be either the person’s I-mode goal or his we-mode collective goal. Second, in addition to we-acceptance of the goal and the intention, the agents are also collectively committed to both goal and intention. Furthermore, collective commitment here is two-dimensional in the sense that the agents are not only 1) committed to the content of the intention, for instance, and 2) believe the others are as well and 3) believe that there is mutual belief to the effect that 1) and 2). Additionally, there is a social normative element as they are committed to each other to hold both the goal and the intention, so that they are not allowed to drop or give up the intention or the goal but together. In our example the operative members perform a joint action, assassination, which brings about (or constitutes) group action of Los Patheticos declaring a war against The Merciless. The action in question satisfies our analysis of group action. Thus, the operative members perform joint action Y which brings about X and is believed by the operative members to bring about X. Non-operative members have authorized the operative members and thus are committed to passively support, or accept, what the operative members are doing, given that non-operative members are adequately informed, that is they have correct beliefs about what the operative members are doing. Here the non-operative members are adequately informed.

Group Action and Group Responsibility

209

Here we have a paradigm case of group responsibility. The group is responsible for (performing) war declaration. Responsibility can justifiably be ascribed to each member of the group qua a group member. None of the members bears individual responsibility for X, this is obvious as it was not even possible for any one of them to perform X alone. Responsibility is ascribed to all of the members jointly. In addition to group responsibility for war declaration, the operative members bear joint responsibility for assassination (at least to a much greater extent than the non-operative members), and each of the operative members bears individual responsibility for their part task in performing assassination, qua a group member though. 3.2 Unstructured groups Let us next consider unstructured groups. Here again, rational collective acceptance is the basic measure stock. Even if the group is unstructured it may have previously accepted some constitutive goals and relevant beliefs (views) that guide its activities. Here we take an unstructured group to have a group ethos. Acting as a group member then could not violate what the group’s ethos entails. The group could, alternatively, be one which does not have such underlying goals and beliefs, but is formed only for “togetherness”, for the members being able to do whatever they jointly decide to do together and to “be together”. Collective acceptance by the group is required in principle to determine whether an action by a group member is acting qua a group member. It might e.g. violate the underlying society’s or community’s moral values and norms (which in addition have to be “pre-accepted” by the group or at least in the group) and thus be subject to collective rejection as a proper group member’s action. Whatever the reason for acceptance or rejection would be, collective acceptance would be the ultimate criterion. Some amount of collective commitment to what has been collectively accepted must be required both in the case of structured and unstructured groups. Functionality for group activities in general is a central criterion for how much collective commitment must be present. This, then, is our rather schematic and “mechanical” underlying model for rationally making judgments about collective responsibility: Given internal authorization and voluntary group membership, all group members are jointly responsible for the group’s action. There may be differences in the degrees of responsibility, as noted, due to the different roles the members play in the action (e.g. operative versus non-operative member, leading operative

210

Pekka Mäkelä and Raimo Tuomela

member versus “mere” operative member, etc.). In this paper we do not go deeper into these problems, which, furthermore, seem to contain strong contextual and contingent elements. What about we-mode and the I-mode thinking and acting in this context. Basically, acting qua a group member here is centrally connected to we-mode acting. Recall that rational collective acceptance was taken as the ultimate criterion for the notion of acting as a group member and that collective commitment was also required. So we have the elements that Tuomela’s strongest notion of we-mode requires. As the notion of acting as a group member was seen to be criterial for the attribution of group responsibility, we can now see that we-modeness is about equally central for group responsibility – both in the case of structured and unstructured groups. I-mode behavior does not play any role for the attribution of group responsibility. In Tuomela’s positional model of group attitudes and action the non-operative members are assumed to tacitly accept or at least to be obligated to accept what the operative members do for the group. It seems that we must say that this feature, due to internal authorization, makes the non-operative members responsible as well, although in general to a lesser degree than are the operative ones. 3.3 The Central Analyses In light of what has been said above we formulate the following “analysis” stating sufficient conditions for group responsibility. However, before giving the analysis a few short remarks are, perhaps, appropriate. We are talking about retrospective responsibility here, that is, responsibility for something that took place in the past. Thus, X is assumed to obtain. Furthermore, the analysis is “designed” for the case of blameworthiness, that is, for the case where the group has either performed X or brought about X and according to the normative standard, N, not-X ought to be the case. (RIGA) A group, G, is responsible for X relative to a normative standard N, where X is either an intentional action or an intended consequence of such action, in the social and normative circumstances C, if X is relevant to N and the members of G knew that or they should have known that, and in C there were operative agents A1,...,Am of G such that 1) A1,...,Am, acting qua group members (and thus performing their positional tasks due to their exercising the relevant decision making system of

Group Action and Group Responsibility

211

G) intentionally (jointly) brought about X (viz., there was an action Y such that the operative agents intentionally jointly performed Y and this performance of Y generated, and was believed and purported by the operative members to generate, the result-event of X); 2) because of 1), the (full-fledged and adequately informed) non-operative members of G, as members of G, tacitly accept the operative agents’ intentional bringing about of X – or at least ought to accept it; 3) there was a mutual belief in G to the effect that there was at least a chance that 1) and to the effect that 2). As our account has already been discussed at reasonable length above (RIGA) does not seem to need many further comments. (RIGA) formulates our analysis of the paradigm case of group responsibility, namely, group’s responsibility for a full-blown intentional group action or intended outcome of such an action. As already mentioned, in our analysis the group, G, is assumed to be voluntary and internally authorized, and the action performed by G is assumed to be both internally and externally free. (RIGA) offers the sufficient conditions of the strong case of group responsibility, that is, the case in which the ascription of group responsibility is justified if it ever is (and we of course believe it sometimes is). Our account and the discussion about it has mainly been concerned with structured groups, however, we want to emphasise that in our account unstructured groups can be taken as a special case of structured groups. (For arguments see Tuomela 1995.) Even though (RIGA) completes the primary task we set ourselves in the beginning of this paper, namely, offers only a sufficient condition under which it is justified to hold a group responsible for its action, we want to go a bit further and somewhat tentatively build a necessary and sufficient condition analysis of group responsibility on (RIGA). Here, again we assume the group to be voluntary and the action or omission to be internally and externally free. One might argue that this assumption makes our analysis subject to counterexamples. However, we want to analyze full responsibility here, and weakening the conditions of voluntariness of the group or internal or external freedom of action would lead to less than full responsibility, as lack of freedom is a factor undermining responsibility. So below we are offering necessary-and-sufficient condition analysis of full group responsibility. Note that our analysis allows for the possibility that the group, G, shares the full responsibility for X with some individual agent or with another group.

212

Pekka Mäkelä and Raimo Tuomela

Now the following suggestion for a necessary-and-sufficient condition analysis of group responsibility can be proposed: (RGA) Group G is responsible for X (where x is an action or an outcome of an action) relative to a normative standard N if and only if 0) X is relevant to N and the members of G knew that or they should have known that, and 1) G performed X intentionally or brought X about intentionally; or 2) G intentionally omitted preventing X (outcome), or X was an intentional omission of an action G should have performed; or 3) G brought about X as a foreseen but not an intended consequence of G’s intentional action ; or 4) G brought about X as a consequence (of G’s intentional action) which it did not foresee but which it should have foreseen; or 5) G performed X by mistake when intentionally trying to perform Y and there was either recklessness or negligence involved in its so doing; or 6) X was brought about (possibly by a member of G) and G had an obligation to see to it that not-X, and G was aware of the obligation. Let us comment briefly the clauses 1)-6), especially, how they relate to (RIGA). Obviously, 1) is already analyzed in (RIGA). 2) is not problematic either, there the operative members of G have intentionally omitted doing something they should have done in light of N and their omission satisfies the analysans of RIGA with the obvious modification that ‘action’ is replaced by ‘omission’. In 3) the operative members have performed intentional action Y in accordance with the analysans of (RIGA) and there was a mutual belief among them to the effect that X is a consequence of their performing Y. Here we can say that bringing about X by G was intentional in a much weaker sense than in (RIGA) as X was not an intended consequence of Y. We can say that X was not non-intentional or even unintentional, as it was foreseen. 4) is similar to 3) but there was no mutual belief among the operative members to the effect that X will be a consequence of their performing Y, but there should have been such a mutual belief. Here we have in mind a case where the operative members have been negligent or reckless in their deliberation; they did not take all the relevant facts into account when deciding the course of group’s action. 5) seems to explain itself, for analogously with the individual agent case culpable ignorance is not an efficient excuse for a group either. In 6) the operative members have done something in accordance with our

Group Action and Group Responsibility

213

analysis of group action as a result of which there is an obligation of G to see to it that not-X.

Conclusion In this paper we have discussed group responsibility. According to our account the paradigm of group responsibility is responsibility of a voluntary group for an internally and externally free group action performed in the wemode, given that the action is relevant to some normative standard N. We have defended collective acceptance and collective commitment as the core elements of group agency, or central analytic notions of group agency if you like. According to our view a group can legitimately be held responsible as a group for some action or outcome only if at least some members of the group performed some action in the we-mode. It does not follow from our analyses that a group is always responsible for actions performed by its members. For instance, there may be an intentional deviator in the group who intentionally deviates from the group’s ethos or on purpose acts against what has been collectively accepted in the group, and thus does not act in the we-mode. In a case like this, according to our account, the deviator primarily bears individual responsibility for such action and other members of the group qua members of the group are not to blame. However, if there is collective commitment to satisfy and/or uphold the ethos, as indeed we have claimed, then the group members are responsible for “correcting” deviators’ behavior. How “far” they should go in this correction is a moot point depending e.g. on the normative and factual cohesion of the group.

References Feinberg, J. 1970. Doing and Deserving: Essays in the Theory of Responsibility. Princeton, New Jersey: Princeton University Press. Fischer, J. M. and M. Ravizza. 1988. Responsibility and Control – a Theory of Moral Responsibility. New York: Cambridge University Press. French, P. 1984. Collective and Corporate Responsibility. New York: Columbia University Press. Gilbert, M. 2000. Sociality and Responsibility: New Essays in Plural Subject Theory. Lanham, Maryland: Rowman & Littlefield Publishers. Haji, I. 1998. Moral Appraisability. New York: Oxford University Press.

214

Pekka Mäkelä and Raimo Tuomela

May, L., and Hoffman, S. (eds.). 1991. Collective Responsibility. Maryland: Rowman & Littlefield Publishers, Savage. Mellema, G. 1997. Collective Responsibility. Amsterdam: Value Inquiry Books Series, Rodopi. Miller, K. and R. Tuomela. 2001. “What are Collective Goals?” In Explanatory Connections ,ed. M. Kiikeri, P. and Ylikoski. http://www.valt.helsinki.fi/kfil/Matti/ Oshana, M. 1997. “Ascriptions of Responsibility.” American Philosophical Quarterly 34: 71-83. Strawson, P. F. 1962. “Freedom and Resentment.” Proceedings of the British Academy 48: 1-25. Tuomela, R. 1995. The Importance of Us: A Philosophical Study of Basic Social Notions. Stanford, California: Stanford University Press. — 2000. Cooperation – A Philosophical Study. Philosophical Studies Series 82. Dordrecht: Kluwer Academic Publishers. — 2002a. “The We-mode and the I-mode.” In Socializing Metaphysics: The Nature of Social Reality. Ed. F. Schmitt. Lanham, Maryland: Rowman and Littlefield. Forthcoming. — 2002b. The Philosophy of Social Practices: A Collective Acceptance View. Cambridge: Cambridge University Press. Forthcoming. Tuomela, R., and K. Miller. 1988. “We-Intentions.” Philosophical Studies 53: 115-137. Wallace, R. J. 1996. Responsibility and the Moral Sentiments. Cambridge, Massachusetts: Harvard University Press.

Collective Comitment

215

Collective Commitment: A Theoretical Understanding of Human Cooperation Lambèr Royakkers and Vincent Buskens Abstract Organizations can be seen as a collection of interacting agents to achieve a certain task: a collective task. Since such a task is beyond the capacity of an individual agent, the agents have to communicate, cooperate, coordinate, and negotiate with each other, to achieve the collective task. In distributed artificial intelligence (DAI) theories of organizations, it is emphasized that ‘commitment’ is a crucial notion to analyze a collective activity or the structure of an organization. In this paper, we analyze the notion of commitment to gain more insight in the social interactions between the agents in an organization. Many social interactions between agents demand the use of commitments to reach socially efficient or avoid socially inefficient outcomes. Commitments express the desires, goals, or intentions of the agents in an interaction. Using a game-theoretic model, we will show that, depending on the incentive structure, different interactions require different types of commitments to reach socially efficient outcomes. Based on these results, we discuss whether existing (or slightly adapted) logical formalizations are adequate for the description of certain types of commitments and which formalization is suitable for reaching a socially efficient outcome in a specific interaction.

1 Introduction: Organizational Behavior and the Rational Choice Perspective In recent years, the role of commitment in collective action has received increasing attention from the social sciences. Collective action refers to activities that require cooperation among individuals to achieve desirable outcomes. In particular, this attention is apparent in two distinct literatures, namely, the literature based on in the field of organizational behavior and the rational choice literature (cf. Robertson and Tang 1995). The field of organizational behavior examines behavior in and of organizational systems. Much of this literature focuses on factors that influence the quality of individuals’ involvement and performance in the organization. A primary focus on theory and research has been individual’s work attitudes and the individual and organizational characteristics that shape these attitudes. In these analyses, organizational commitment has been identified as an important work-related attitude: an individual’s

216

Lambèr Royakkers and Vincent Buskens

attitude regarding his or her relationship to the organization. The most popular framework of organizational commitment is based on the work of Mowday et al. (1982). They conceptualized commitment as an individual’s identification with and involvement in an organization.1 While the organizational behavior literature includes multiple perspectives on the concept of commitment, the rational choice literature shares a more consistent view about the concept. The latter focuses on how an individual’s choice of behavior depends on the way he or she perceives and weighs the expected costs and benefits of alternative course of action. It contrasts with the dominant organizational behavior literature emphasis on commitment as a psychological attachment to the organization. In the rational choice literature, the concept of commitment is frequently used in conjunction with the concept of credibility to denote a general theoretical puzzle: Given a group of selfinterested individuals who need to engage in collective action, what are the most effective means for binding them to cooperative arrangements across space and time? Given this theoretical puzzle, commitment can be seen as an actor’s resolve to carry out a promise to perform in a specified fashion. Besides the high degree of consensus about the definitions of basic concepts and the focal theoretical puzzles of the rational choice literature (which contrasts with the organizational behavior literature, that includes a wide diversity of definitions and theoretical questions), two other strengths of the rational choice theory are its consistent effort in applying the concept of credible commitment to a wide variety of collective action arenas (such as hierarchical organizations, contracting, and local common-pool resources) in contrast with the organizational behavior literature that focuses predominantly on work organizations, and its use of gametheory for theory development. In this paper, we will focus on the rational choice perspective of commitment.2 1 More specifically, commitment in this framework is characterized by three factors: (1) a belief in and acceptance of organizational goals and values, (2) a willingness to exert effort toward organizational goal accomplishment, and (3) a strong desire to maintain organizational membership. 2 The assumption behind this perspective is that individuals are susceptible to succumbing to a weakness of will (Elster 1979) or the temptation (Williamson 1983) to renege on previous promises. In the former case, it is argued, weakness of will may decrease the likelihood of a person following through on a course of action even though it would be selfbeneficial, as when a smoker fails to complete a quit-smoking program (Schelling 1984). Credible commitment is established when a person chooses to submit to rules or punishments that make it impossible (or at least more difficult) to deviate from previous resolutions. In the case of temptation, a person may be tempted to break promises when expect-

Collective Comitment

217

From perspective, credible commitment is not purely an individual’s subjective state of mind, that is, whether he or she is psychologically committed to a certain course of action. What is important for successful collective action is not simply whether each individual is personally committed to the collective arrangement. More important is whether conditions exist that can either discourage or prevent a participant from reneging on a promise. The underlying assumption is that although parties may have incentives to reach an agreement, the incentives ex post may be incompatible with maintaining it. As argued by Shepsle (1991), a commitment is credible in two possible senses – motivationally or imperatively. A commitment is motivationally credible if the participants expect to receive sufficient rewards for them to honor the commitment at the time of performance. The commitment is incentive-compatible and thus self-enforcing. It is imperatively credible if the participants are unable to act otherwise because they will be coerced to do so by an external authority or they will lack the means to do otherwise. In this paper, we restrict ourselves to the motivationally credible commitment. The formalization of the notion of commitment is a topic of continuing interest in AI for several reasons. In organization theories of Distributed Artificial Intelligence (DAI), negotiation systems, and cooperative software agents, it is emphasized that “commitment” is a basic ingredient to analyze joint intentions and teamwork, and to analyze a collective activity or the structure of the organization (cf. Gasser 1991). In Belief-Desire-Intention systems (BDI), important contributions have been made on motivational attitudes such as commitments and obligations to specify, analyze, and reason about the behavior of rational agents. BDI-agents are characterized by a “mental state” described in terms of beliefs (viewed as informational attitudes), corresponding to the information the agent has about the environment; desires (viewed as its goals), representing options available to the agent; and intentions (viewed as motivational attitudes) representing the chosen options. Consequently, by formalizing commitment in a logical way or to propose a descriptive ontology for commitment (and other motivational attitudes), it is possible (1) to reason about commitments to achieve tasks; (2) to gain some insights in the fundamental notions of motivational attitudes; and (3) to analyze collective activity. By formalizing the notion of commitment, we think that this could help both in the theory of Organizations and Groups, and in the computational supporting of ing additional gains from doing so, as when someone acts opportunistically in a business transaction. (Robertson and Tang 1995)

218

Lambèr Royakkers and Vincent Buskens

cooperative work, which may contribute to a theoretical understanding of human cooperative activity. However, if we want to represent an interaction with commitment in a logical system, only the possible actions are described. Commitment is then introduced as an elementary proposition (cf. Shoham 1993). This implies that the commitment is a fact that does or does not occur. More sophisticated theories (Dunin-Kęplicz and Verbrugge 1996, Meyer et al. 1999) describe a formalization of motivational attitudes such as intentions, goals, and wishes to explain why agents behave the way they do. Within the logical systems there is nothing that drives the motivational attitudes. It is only stated that if certain attitudes are present, commitments are used without explicit reasoning why and when a certain attitude leads to a commitment. Our main criticism of these logical systems is that they, strictly speaking, do not explain but only describe actions by agents, probably including the use of commitments. Logical systems fail to distinguish between conditions that provide arguments why agents have incentives to use commitments under some conditions and not under other conditions. Moreover, logical systems cannot distinguish which commitment is or is not credible in a given interaction. The reason is that logical systems generally neglect the incentives related to various combinations of actions and the strategic interdependence between different agents. In game theory, however, motivational attitudes are represented by the payoffs agents receive at the end of an interaction, based on their combination of actions. By distinguishing different commitment types using game theory, we are able to specify conditions for which logical formalizations of commitments are appropriate using the goals and intentions of the agents. The following section demonstrates how commitment can be formalized and provides a descriptive ontology. Section 3 demonstrates how commitments are introduced in game-theoretic systems and illustrates how different commitment regimes can lead to different outcomes in interactions among agents. The subsequent section translates the outcomes of the game-theoretic analysis in a logical formalization. Finally, we conclude and summarize.

2 Logical Analysis of Commitment Intuitively, the relation between the members of the group and their collective activity is shaped by the commitment of the group to accomplish the activity according to some plan and structure of the group. In this section, we discuss the notions of social and collective commitment (as a logical de-

Collective Comitment

219

scriptive ontology) that are crucial for the understanding of the relation between the individual agents and their collective activity to accomplish some goal. Castelfranchi (1995) claimed that a notion of commitment is needed as a mediation between the individuals and the collective one: ‘Commitment’ is seen as the glue of the group, of collective activity: it links the agent with the joint goal and the common solution, it links the members’ actions with the collective plan, it links the members with each other. From the above description, we infer three notions we will use for the analysis of the collective commitment: joint goal, collective plan, and relation between the agents. We will discuss these three notions. We will show that for gaining insight in the notion collective commitment it is necessary to have a better understanding of social commitment, which is the basis for a definition of collective commitment, since the bilateral aspects of collective commitment are reflected in social commitments (cf. Castelfranchi 1995; Dunin-Kęplicz and Verbrugge 1996; Royakkers and Dignum 2000). 2.1 Collective Commitment We assume that to commit to an action necessarily implies committing to some result of that action. Conversely, to commit to a goal always implies the commitment of at least one action that produces such a goal as result. Thus, we consider the action/goal pair τ = (α,g) as the real object of commitment, which we call ‘task’. By means of τ, we will refer to the action α, to its resulting goal g, or to both (cf. Castelfranchi and Falcone 1998). To achieve a collective task τ the group/organization has to decompose the task τ (to achieve τ) into a number of subtasks over the agents in management positions. For example, if the program committee has the task to notify the authors of the submitted papers of acceptance before a certain deadline, this task can only be achieved if the work that has to be done, is shared out in several tasks over the members of the program committee. Therefore, the organization needs a plan: a concrete manner to achieve the collective task. A plan includes task division and task allocation. Task division is the decomposition of a complex task τ into subtasks τ1,...,τn. Task allocation indicates which agent of the group has to achieve which subtask of the complex task. This results in a set of individual subtasks that realizes the complex task (the collective goal). We can represent a plan as P(, τ , X), with1 1 We do not account for the fact that some individual tasks might depend on other ones. So,

220

Lambèr Royakkers and Vincent Buskens

• il ∈X for l ∈ {1,...,n}, and X represents a non-empty set of agents; i.e., the agents of the organization/group; • τl is the subtask of agent il belonging to her job description or her task repertoire: tl ∈ T with T the set of the individual subtasks established by the plan; i.e., {τ1,...,τn}; • is the set of individual subtasks il: τl with 1 < l < n; • τ is the collective task which will be achieved by the achievement of all the subgoals (i.e., τ1,...,τn-1, and τn). According to Dunin-Kęplicz and Verbrugge (1998), collective commitment is meant to reflect the way a task is to be achieved according to a plan. This can only be established or maintained if the group has the associated collective intention. In addition, for every one of the subtasks τ1,...,τn that together constitutes the task t in question reflected in the plan P, there should be an agent in the group who is socially committed to at least one (mostly other) agent in the group to fulfill the subgoal. Moreover, there should be a common belief in the whole group that the plan will be entirely realized; i.e., that all the subtasks have been adopted by committed members of the group. In their formalization of collective commitment, all these characteristics are reflected:2

∨ COMM(i, j, i:τ )) ∧ C-COMM(X, X: τ | P) := C-INT(X, X: τ) ∧ ( τ ∧ l ∈T i , j∈ X l

∨ COMM(i, j, τ ))) C-BEL(X, X: ( τ ∧ l ∈T i , j∈ X l

In this formalization of collective commitment the basic ingredient is social commitment: the commitment of one agent to another.3 It expresses a relation a plan must also determine the order of subtasks: the temporal structure of the individual subtasks. For example, the notification of acceptance of a certain paper can only be done if the delegated members of the program committee review it. Thus, an individual task can sometimes only be achieved if other tasks are already achieved. So, a plan does not only contain a set of individual subtasks, but also contains an order of the individual subtasks, i.e., a sequential composition of the individual subtasks. For convenience, we will not discuss the order of subtasks, which obscures the points we want to make in this paper. 2 C-INT(X, X: τ) means that group X has the collective intention to achieve task τ; CBEL(X, X: ϕ) is meant to be true if everyone in group X believes ϕ, everyone in X believes that everyone in X believes ϕ, etc. 3 In Castelfranchi (1995), the social commitment is considered as a relation between two or three agents. The third agent is the witness, who has a very crucial role in normative contexts (norms efficacy) (cf. Conte and Castelfranchi 1995) and in contractual contexts implicating free riders and cheaters. For convenience, we omit considerations about a third agent. Note

Collective Comitment

221

between two agents, more precisely, social commitment is a 3-argument relation: COMM(i, j, i: τ) meaning that i (the committed agent) is committed to j (the agent to whom i is committed) to achieve τ. In this case, only agent j is in the position (has the authority unilaterally) to rescind it. Thus, agent j can rely on the persistence of agent i’s commitment to achieve τ. However, the core of the social commitment relationship is the delegation/adoption relation (cf. Castelfranchi and Falcone 1998), which is completely ignored in this definition. The delegation/ adoption relation is a basic ingredient for joint intentions, true cooperation and team work: in collaborative activity each agent relies on the other agents delegating to them some tasks, and, vice versa, each agent adopts some tasks from the other agents. In fact, the huge majority of DAI and negotiation systems, communication protocols, cooperative software agents are based on the idea that cooperation works through the allocation of some (sub)task by a given agent to another agent, via some “request” (offer, proposal, announcement, etc.) meeting some “commitment” (bid, help, contract, adoption, etc.). Informally, in delegation an agent i needs or desires an action of another agent j and includes it in its own plan, i is trying to achieve some of its goals through j’s actions; thus i has the goal that j performs a given action. The possibility of delegation/adoption depends on the authority-relation within the organization/group, usually reflected in organigrams. We introduce an operator of delegation with three parameters: DEL(i, j, τ), meaning that i delegates the task τ to j.4 For convenience, we presume that the interaction between the delegating agent (client) and the delegated one (contractor) is based on explicit agreement, which is called contract. In a contract, the delegated agent knows that the delegating agent is relying on it and accepts the task. If the agreement is settled, the delegated agent j is socially committed to the delegating agent i to achieve the agreed task τ . Formally: DEL(i, j, τ) → COMM(j, i, j: τ) The delegation/adoption relation based on the authority-relation can specify more accurately the social commitments between the agents in an orgathat the professor in our example in the introduction can be considered as the third agent. 4 For formal semantics of the operator DEL, we refer to Santos, Jones and Carmo (1997).

222

Lambèr Royakkers and Vincent Buskens

nization. We will see the collective commitment as the internal commitment of a group (a collective agent) X given the underlying structure of the group (i.e., the authority-relation) with respect to the achievement of t according to a plan. Consequently, we represent the collective commitment as a conditional one: C-COMM(X, X: τ | P, Struct(X)), meaning that the group X has a collective commitment to achieve τ (X: τ) given the plan P and the underlying structure of the group or organization X (Struct(X)). The additional aspect in this representation of collective commitment allows us to specify the delegation/adoption relations, and consequently indicates the possible social commitments between agents. However, it does not give any insight in the notion of social commitment, which still is the basic ingredient of the collective commitment. 2.2 Social Commitment Most approaches formalizing commitments focus on internal or personal commitments (e.g., Cohen and Levesque 1990, Dignum et al. 1996, Meyer et al. 1999, Rao and Georgeff 1991): a relation between an agent and a task: an agent who is committed to a task has promised himself to achieve the task.5 Rao and Georgeff (1991) consider beliefs, goals and intentions to be primitive, and define a notion of internal commitment in terms of these by treating intentions as a commitment to the achievement of current tasks.6 However, these notions (and their semantics) are very fruitful as a basis for the formalization of social motivational attitudes. Rao and Georgeff provide the following two axioms to capture the interrelationships among an agent’s beliefs, goals, and intentions: 1. The axiom of belief-goal compatibility: GOAL(i, i: τ) → BEL(i, i: τ),

5 An agent is subject to an internal commitment if and only if she is the sole author of a commitment, and has the authority unilaterally to rescind it (Gilbert 1999). The social commitment cannot be composed of two internal commitments of two agents; it is created by two agents together. 6 For formal semantics of these primitive notions we refer to Rao and Georgeff (1991).

Collective Comitment

223

which states that if an agent has a goal to achieve task τ, he also believes it; i.e., there is at least one path in all the belief-accessible worlds in which task τ can be achieved. The formula (i: τ) stands for the proposition that agent i achieves τ (or that agent i sees to it that τ will be accomplished). 2. The axiom of goal-intention compatibility: INT(i, i: τ) → GOAL(i, i: τ), which states that if an agent intends to achieve a task, he also has the goal to achieve that task. The two above-mentioned formulas imply that if an agent intends to achieve task τ, he also believes it. Based on the above-mentioned notions, DuninKęplicz and Verbrugge (1996) give the following definition of (social) commitment inspired by Castelfranchi (1995): COMM(i, j, i: τ) := INT(i, i: τ) ∧ GOAL(j, i: τ). If agent i is committed to agent j to achieve something, then i should have the intention to achieve that and j is interested in i fulfilling of i’s intention. The last condition can be seen as a goal adoption: the achievement of the task is a goal of j. Since we restrict ourselves to rational agents, we state that an agent i agrees with the commitment entered by agent j to achieve some task if and only if the achievement of the task by j is a goal of i.7 So goal adoption implies agreement or acceptance of the commitment. The definition of commitment gives rise to some critical notes, which we will discuss below. Commitments require that an agent j to whom an agent i is committed is aware of i’s intention. There has to be a mutual knowledge about the intention of i. In daily life, this is done by expressions in conditions of common knowledge (if it is “out in the open”) as far as the two agents are concerned (cf. Gilbert 1992), and in a business transaction, e.g., by a contract. Therefore, we will extend the above definition as follows: COMM(i, j, i: τ) := INT(i, i: τ) ∧ K(j, INT(i, i: τ)) ∧ GOAL(j, i: τ), 7 In the game-theoretic sense, goal adoption implies that the payoff for j is larger if i performs the action specified in the commitment rather than the other action. Moreover, if j has committed to an action himself, goal adoption should be consistent with the payoffs related to the interaction that incorporates the commitment made by j.

224

Lambèr Royakkers and Vincent Buskens

where K(j,φ) stands for the fact that agent j knows φ.8 According to Meyer et al. (1999), it is rather doubtful whether notions as goals and intentions are primitive, since motivational processes are stemming from internal drives, and are experienced by human as conscious desires. Wishes (or desires) constitute the primitive motivational attitude that models what an agent likes to be the case, and therefore they take wishes as primitive, and define goals by means of these, as follows:9 GOAL(j, i: τ) := WISH(j, i: τ) ∧ ¬(i: τ) ∧ ◊(i: τ) ∧ SELECT(j, i: τ), meaning that goals are defined as selected wishes (WISH(j, i: τ) ∧ SELECT(j, i: τ)) that are unfulfilled (¬(i: τ)), and can be implemented by i (◊(i: τ)): “it does not make sense for an agent to try and fulfill a wish that already has been fulfilled or for which fulfillment is not a practical possibility for that agent.” With the help of the notion of goal, they define (possible) intentions to achieve all the tasks that are correct and feasible with respect to some of their goals: INT(i, i: τ) := CAN(i, i: τ) ∧ K(i, GOAL(i, i: τ)), where CAN(i, i: τ) states that for agent i it is a practical possibility to achieve τ. In the sequel, we use for convenience the terms goal and intention to formalize commitment with the above-mentioned interpretations. However, this logical approach of commitment helps us to make clear what exactly is meant by the notions of commitment, goal, and intention, which are otherwise vague and ambiguous terms. According to the formal definition of commitment, which refers to a relation between two agents and a task, it is necessary that the agent who is committed has the goal that the committed agent achieves his intended task. However, we state that the commitment of one agent does not need the consent (or the acceptance) of the other agent to become effective. For example, to make a commitment and then cut off communication (hang up the phone, and take it off 8 K(j,φ) is interpreted in the Kripke-style possible worlds semantics. 9 This is a simplified representation adapted for social commitment, since Meyer et al. (1999) define their notions for internal commitments in a dynamic framework, which offers the possibility to consider how motivational attitudes change over time. Although this is a very useful extension of the logical approaches of commitment, we do not consider the dynamic elements of this approach because of its complexity, which would obscure our main point.

Collective Comitment

225

the hook).10 We will see in the examples of the next section that, depending on the interaction, commitments that can be made by one agent without the consent of the other agent can lead to outcomes that are worse for the last agent than the outcome if no commitment would have been made. Consequently, the definition of commitment is too strong for using it a priori in all situations. From this remark, it follows that we need different formalizations of commitment that can systematically be linked to intentions and goals of the agents involved. Using the game-theoretic analysis in the next section, we will provide such a link for formalizations of commitments. Besides intentions and goals, commitment also involves normative notions. For example, according to Castelfranchi (1995), a social commitment implies an obligation: an agent ought to do what he is committed to, which can be formalized as follows: COMM(i, j, i: τ) → OUGHT(i, i: τ) This implies that task τ is more than an intention of agent i, it is a duty. However, we think that the above formalization is too strong. The mistake of confusing commitment with obligation arises directly out of a conflation of the two distinct senses of the term commitment: the formula COMM (i, j, i: τ) only involves (unambiguously) the intentional sense of commitment (agent i undertakes an obligation to achieve τ from j), and the formula OUGHT(i, i: τ) the extensional sense of commitment (agent i has an obligation to achieve τ). So, in fact, COMM(i, j, i: τ) need not imply OUGHT(i, i: τ); we can undertake obligations that we do not have (cf. Vogel 1975). A social commitment cannot imply a juridical norm, but at most a directed norm: COMM(i, j, i: τ) → OUGHT(i, j, i: τ) The directed obligation OUGHT(i, j, i: τ) is read as ‘agent i (the addressee) has an obligation towards j (the counter party) that τ will be achieved.11 Several authors (Herrestad and Krogh 1995; Kanger 1971, Lindahl 1977; Makinson 1986) analyzed or described the types of rights relationships between the ad10 Another example is the run away buyer who tries haggling over the purchase price, which he in principle wants to pay for the traded product. If the seller prefers to sell the product for a lower price than to be stuck with the product, he will give in, call back the potential buyer and settle for a lower price. So, the seller is in the sphere of power of the buyer: the buyer can unilaterally commit without the agreement of the seller.

226

Lambèr Royakkers and Vincent Buskens

dressees (or bearers) and counter parties from the classic work by Hohfeld (1964). The counter party is an agent or a group of agents that has a ‘right’ against the addressee, who has a ‘duty’ (an obligation). ‘Right’ and ‘duty’ are correlative terms, i.e. when a right is invaded, a duty is violated.12 An example of a directed obligation is that John is obliged towards Paul that he shall stay away from Paul’s land. In other words, Paul has a right against John that the latter will stay away from Paul’s land, the correlative (and equivalent) being that John is under a duty toward Paul to stay away from the place. The ‘right’ of agent j consists of the entitlement to control, exact/require or complain/ protest the task τ .

3 Game-theoretic Analysis of Social Commitment One of the recurring themes in the analysis of game theory has been that in many situations agents can benefit from the opportunity to make a (binding) commitment. By committing to a given action an agent may be able to alter the play of his opponents to reach socially efficient or avoid socially inefficient outcomes. The way to model the possibilities of commitments (or related moves like “promises”) is to explicitly include them as actions the agents can take (cf. Schelling 1960), which is possible within a game-theoretic setting. We will start with an example. Assume you want to write an article together with a colleague. You are both convinced that joining forces will produce a better product than writing two articles separately. However, you as well as your colleague cannot be sure that the other will actually invest his fair share in this joint project (cooperate). Still, if both of you work hard, you will both be satisfied. You realize that if the colleague sits back (defects) while you do the job, he is even better off and you would have preferred to write an article alone. Clearly, your colleague also fears that you sit back and profit from his effort.

11 For the semantics of the directed obligation we refer to Royakkers (1998). 12 In law, the word ‘right’ is often used to designate ‘legal power’, ‘legal claim’, and ‘immunity granted by an authority’.

227

Collective Comitment

Agent 2 Defect

Cooperate

Defect

2,2

4,1

Cooperate

1,4

3,3

Agent 1

Figure 1: Strategic Form of the Prisoner’s Dilemma Game

The “game” described above (without commitments) is called a Prisoner’s Dilemma Game (cf. Luce and Raiffa 1957). In strategic form,13 the game is shown in figure 1. The values in the cells of the matrix indicate the payoffs for each agent related to a combination of actions of the two agents, which are strictly ordered (from 1 to 4) and which do not represent the exact payoff but the degree of utility. The expected action in this game is “defect” by both agents, because independent of the action of the other agent, each agent is better off by defecting. Consequently, both agents receive 2, although they could obtain 3 if they both would cooperate. Thus, the expected outcome (2,2) is socially inefficient. However, by committing to cooperation, e.g., by mutually informing the responsible professor who can incur sanctions on the researcher who does not work on the joint paper, cooperation becomes the best option for both agents. Hence, a mutual commitment leads to a better outcome for both agents in this situation. This is only one example of a situation in which a commitment can change the expected outcome for an interaction between two agents. Likewise, the usefulness of commitment systems can be investigated for many social and legal interactions. For now, we will describe for the game-theoretic analysis commitment as an action by one agent before an interaction with other agents that signals to the other agents the intention to perform a particular action later on in the interaction. In this paper, we restrict ourselves to commitments that ensure that the agent who commits to a certain action will execute this action (binding commitments). We will show that this broad description of commitment provided here includes several types of 13 For all basic game-theoretic terminology and aspects we refer the reader to Rasmusen (1994) or Binmore (1992).

228

Lambèr Royakkers and Vincent Buskens

commitments that can have different consequences depending on the goals and intentions of the agents in the situations in which the commitment is used. In the following section, we will extend the formalizations with logical systems that correspond with the game-theoretic analysis. The example above has shown that adding incentives to the different actions enables to explain why a commitment is effective in an interaction. In such a “game”, the goal or wish of the agents is to maximize their “utility” at the end of the game. This implies that we only consider rational players (agents). One could say that a commitment signals the intention to perform a specified action. The commitment can only be successful if the value of the sanction applied if an agent deviates from the action specified in the commitment is large enough. As indicated before, we assume that this value indeed is large enough such that an agent will always perform the action intended with the commitment. Within a game-theoretic setting, we can define a commitment as an additional move/action14 in a game. The two agents in the game above have to decide whether or not they commit to one of the two possible actions before they play the underlying game. The payoffs related to the game will not change if none of the agents commits to one of these actions (cooperation or defection in the example). However, if one or both agents commit to one to the actions, the payoffs will change depending on who has committed and the rules that are related to the commitment regime that is chosen in a particular situation. Besides explaining the use and effectiveness of commitments, game theory can help to distinguish between different commitment regimes. We discuss four types of commitment regimes that can be distinguished. Such a regime determines which agent is allowed to commit, if both agents are allowed to commit and whether this has to be done simultaneously or sequentially. Then, if commitments are made sequentially, it has to be determined who goes first. Another aspect that has to be determined in the regime is whether the commitment of one agent needs the consent of the other agent to become effective. A last aspect of the regime can prescribe that a commitment of one agent becomes effective only under that condition that the other agent commits to a specified other action. We describe now four (basic) types of commitments that can occur under the various regimes and that can have different implications in the example we discuss thereafter.

14 The term “move” in game theory corresponds with the term “action” used in logical approaches. In the sequel we will use the term “action”.

229

Collective Comitment

1. Unilateral commitment without agreement: one agent expresses that he intends to perform an action and this commitment becomes effective without the need of the others agent’s agreement (or consent) (since the first agent, e.g., cuts off the communication); 15 2. Unilateral commitment with agreement: one agent expresses that he intends to perform an action and this commitment becomes effective only if the other agent (explicitly or implicitly) agrees with this commitment; 3. Bilateral commitment without agreement: both agents express that they intend to perform an action and these commitments become effective without the need of the other agent’s agreement;16 4. Bilateral commitment with agreement: both agents express that they intend to perform an action and the commitments become effective if both agents agree on the combination of commitments. Agent 2 Left

Right

Top

a1, a2

b1, b2

Bottom

c1, c2

d1, d2

Agent 1

Figure 2: A 2 × 2 Game with General Payoffs

On the basis of the representation of 2 × 2 games, we are able to discuss these commitment strategies. Figure 2 represents a 2 × 2 game (a game with 2 agents each choosing between two possible actions) with general payoffs for both agents. Agent 1 chooses between Top and Bottom, while agent 2 chooses between Left and Right. Both agents obtain the payoffs that belong to their combination of choices. E.g., if agent 1 chooses Top and agent 2 chooses Left, agent 1 receives a1 and agent 2 receives a2. We consider only 2 × 2 games with pref15 In Buskens and Royakkers (2002), this classification of commitment is treated for all 8 different types of 2×2 games. 16 This commitment system can never change the outcome of the game because if the action of both agents is determined as soon as they have placed the commitment, placing the commitment is just equivalent to playing the underlying game itself.

230

Lambèr Royakkers and Vincent Buskens

erences over the four possible outcomes that are strictly ordered for both agents. Because only the ordering of the payoffs is important for the analyses, they can be labeled as 1, 2, 3, and 4.17 Each of the four outcomes represents a possible goal state for the agents. The goal states for the two agents do not need to coincide. Agent 2 Left

Right

Top

4,4

3,3

Bottom

1,1

2,2

Agent 1

Example (1)

Example (1) illustrates a situation in which both agents do not want or need to commit to any of the two actions. Agent 1 has a dominant strategy. An agent has a dominant strategy if there is one action the agent can perform that gives him a higher payoff for each of the actions the other agent can perform. The other agent optimizes his payoff given the dominant strategy of the first agent, and both agents cannot do better using a commitment for some other strategy.18 In the example, this implies that both agents obtain 4. Clearly, none of them can do better whatever commitment regime would be chosen.

17 Rapoport et al. (1976) show that there exist 78 distinct 2 × 2 games with strictly ordered payoffs. In Royakkers and Buskens (2001), we classify these 78 games in 8 groups, such that for all games in a group the same arguments hold if commitment regimes are considered. 18 Readers interested in the precise classification of all the games can contact the authors for an overview.

231

Collective Comitment

Agent 2 Left

Right

Top

2,4

4,1

Bottom

1,2

3,3

Agent 1

Example (2)

In the game of example (2), agent 1 has a dominant strategy, because whatever agent 2 does, he is always better off playing Top. This implies that the agent 2 will choose Left, which leaves himself with 4, but agent 1 only with 2. Since agent 2 reaches his most preferred outcome without commitments, a commitment regime that ask for his consent will never lead to an actual commitment unless the commitments do not change the outcome of the game. However, agent 1 wants to commit to playing Bottom, which would result in a payoff 3 for both agents if he can commit unilaterally. Consequently, the only commitment regime that can change the outcome of this game is the unilateral commitment without agreement of agent 2 by agent 1. Such regime might be considered socially desirable because it ensures a more equal distribution of the payoffs among the agents. Of course, the problem is whether such a commitment without the consent of agent 2 is possible. An interesting suggestion (see the previous section) is to make a commitment and then cut off communication (hang up the phone, and take it off the hook). Agent 2 then has the choice of giving in, or getting an even worse outcome (cf. Schelling 1960). Agent 2 Left

Right

Top

2,4

3,1

Bottom

1,2

4,3

Agent 1

Example (3)

232

Lambèr Royakkers and Vincent Buskens

In the game of example (3), agent 1 prefers to play the game without a commitment, rather than that agent 2 commits (without agreement) to playing Left, while this is the best solution for agent 2. Without commitment, the outcome is obtained by randomization.19 On the other hand, both agents prefer to play the game while agent 1 commits to playing Bottom over playing the game without a commitment. Consequently, a unilateral regime will always work if only agent 1 can commit, but only if agent 2 is allowed to commit, a unilateral commitment with agreement of agent 2 is necessary to reach the outcome (4,3). Agent 2 Left

Right

Top

3,3

1,4

Bottom

4,1

2,2

Agent 1

Example (4)

Example (4) is the Prisoner’s Dilemma game. This is the very special game that is also illustrated in the introduction. In this game, the game-theoretic solution predicts that both agents obtain 2, while they both would prefer to obtain 3. However, this would imply that both agents have to deviate from their dominant strategy. The only commitment regime that can work in this game is the bilateral commitment with agreement, implying that both agents should commit to not playing the dominant strategy and if one agent does not commit, a commitment of the other agent will not materialize. The reason for this is that the agents do not want to commit unilaterally to Top or Left, respectively, because the other agent then certainly plays the dominant strategy leaving the first agent with the worst outcome possible. The bilateral commitment with19 Randomization indicates that an agent chooses with some probability p one action and with probability 1-p the other action using some kind of randomization device, e.g., flipping a coin. Expected outcomes can be calculated if we know the probabilities and assume cardinal payoffs for a moment. If both agent use a probability ½, the expected payoff for both agents is (1+2+3+4)/4 = 2.5. It is easy to show that the expected payoff should lay between 2 and 3 in the general case.

233

Collective Comitment

out agreement does not work, because both agents are not willing to commit to the dominated strategy – which leads to the outcome (3,3) – if this commitment remains effective also if the other agent does not commit to his dominated strategy. In this last situation, both agents cannot be sure that they will not be exploited by the other agent and end up with a payoff 1, while the other agent receives 4. Agent 2 Left

Right

Top

3,4

2,1

Bottom

1,2

4,3

Agent 1

Example 5

Example (5) represents a “coordination” game. In these games, there are more equilibria. One equilibrium involves randomization between the actions by both agents. This leads to payoffs between 2 and 3 for both agents. Therefore, both agents want to coordinate on one of the equilibria without randomization, i.e., Top and Left, or Bottom and Right. However, without a commitment they do not have a clue about what the other agent will choose. Coordination on one of the preferred equilibria is possible if one of the agents can commit unilaterally with agreement: the other agent will agree because he is worse of in the situation without commitment. The agent who unilaterally commits first is best off because he can commit such that he will obtain 4 and the other agent 3. This is sometimes called a first-mover advantage. Bilateral commitments (with simultaneous decisions on committing to one of the actions) are problematic here, since this requires solving the same coordination problem while choosing on commitments rather than on actions. Now we will indicate how we can use this game-theoretic analysis in formalizing logical systems and show that specifying the incentive structure that lies beneath a certain formal specification including the related commitment system can increase the usefulness of logical systems. As long as the incentive structure beneath the interaction between two agents is unknown, it is unclear

234

Lambèr Royakkers and Vincent Buskens

whether the specified commitment system will work and will be efficient. In the following section, we will discuss some ideas about how these game-theoretic results can be integrated in logical formalizations.

4 Logic Formalizations Based on the Game-Theoretic Analysis From the game-theoretic analysis follows that the four commitment regimes mentioned before represent crucial distinctions and cause different outcomes in various interactions among agents. Now, we will provide logical formalizations for these regimes. Unilateral commitment without agreement: COMM1w(i, j, i: τ) := INT(i, i: τ) ∧ K(j, INT(i, i: τ)) In this commitment regime, agent j cannot withhold this commitment. If there was a need of consent by agent j, agent j might not give his consent, since the commitment might not lead to a satisfied outcome for agent j. Then he prefers to play the game without commitment. In example (2) of the previous section, we have shown such a commitment. Since the agreement has to be missed, the commitment does not contribute to a goal of the agent who is not committing. Unilateral commitment with agreement: COMM1a(i, j, i: τ) := INT(i, i: τ) ∧ K(j, INT(i, i: τ)) ∧ GOAL(j, i: τ) In this commitment regime, agent j has to agree upon the commitment of agent i. The commitment is only effective under the condition that agent j agrees upon the commitment. This requires (in the game-theoretic sense) that the resulting payoff is larger for agent j if this commitment is effective than the expected payoff without a commitment. In other words, the commitment has to lead to a goal of agent j. Bilateral commitment without agreement: COMM2w(i, j, i: τ1, j: τ2) := INT(i, i: τ1) ∧ K(j, INT(i, i: τ1)) ∧ INT(j, j: τ2) ∧ K(i, INT(j, j: τ2)) This commitment is composed of the two unilateral commitments without agreement. Both agents commit themselves to a certain task regardless the

Collective Comitment

235

agreement of the other agent and regardless the outcome for the other agent. For example, an agent i commits to agent j to make diner, and agent j commits to agent i to do the laundry (agent j makes a “counter commitment”). If agent i drops his commitment, this does not affect the status of the j ’s commitment as a standing commitment (Gilbert 1999), which differs for the bilateral commitment with agreement. Bilateral commitment with agreement: COMM2a(i, j, i: τ1, j: τ2) := COMM1a(i, j, i: τ1) ∧ COMM1a(j, i, j: τ2) meaning that “on the condition agent j commits to agent i to achieve τ2, i commits to j to achieve τ1.” Consequently, if j drops his commitment, this affects the status of i’s commitment: i’s commitment stands no longer. Due to the symmetry in the formulation, it also hold that agent j’s commitment is only effective under the condition agent i made the related commitment.20 We will now reconsider the five examples of the previous section to use them for specifying logical formalizations based on the game-theoretic analysis. For example (1) it is impossible to formalize a commitment that affects the behavior of the agents. I.e., any commitment the agents want to make leads to the same behavior, as they would execute if there was no commitment. Example (2) shows that both agents have not the same goal state. The outcome (3,3) is the goal state of agent 1 while (2,4) is the goal state of agent 2. Moreover, without commitment the outcome will be (2,4). Consequently, agent 1 wants to commit to play Bottom. Because this is not the goal state of agent 2, this can only be reached as agent 1 can commit unilaterally without the need of agreement. This can be formalized as: COMM1w(1, 2, 1: Bottom) In example (3), a commitment of one agent is again necessary to coordinate on one of the equilibria. The best option for agent 2 is to commit unilaterally to Left: COMM1w(2, 1, 2: Left). The best option for agent 1 is to commit unilaterally to Bottom: COMM1w(1, 2, 1: Bottom). However, if the commitment 20 Although only assuming a one-sided condition of i’s commitment on j’s commitment might seem to be a “weaker” form of commitment, the fact that the commitments will only be effective for one specified combination of commitments ensures that this symmetric formulation is equivalent to a one-sided conditional formulation for the types of games considered in this paper.

236

Lambèr Royakkers and Vincent Buskens

needs to be agreed upon by the other agent, agent 1 will not accept the commitment of agent 2, because he is better off without a commitment. On the contrary, agent 2 will accept agent 1’s commitment because the related outcome is still better than playing without a commitment. This analysis suggests that COMM1a(1, 2, 1: Bottom) is the preferred formalization of a commitment in this situation. In example (4) the social efficient outcome (3,3) will be reached by a bilateral commitment with agreement: agents 1 and 2 have to commit to Top and Left, respectively. So to reach the social efficient outcome (3,3), the commitment regime should be formalized as: COMM2a(1, 2, 1: Top, 2: Left) In example (5), the definition of the unilateral commitment with agreement is a suitable formalization for a commitment that leads to a socially efficient outcome. There is a complication because both agents might commit, but they should not commit simultaneously (thus not by a bilateral commitment without agreement). Therefore, a suitable commitment system should prescribe which agent is allowed to commit. Both agents want to commit because the committed agent receives 4, while the other agent receives 3. The system can be formalized by the convention: (COMM1a(1, 2, 1: Bottom) ∨ COMM1a(2, 1, 2: Left)) ∧ ¬COMM2w(1, 2, 1: Bottom, 2: Left). What we learn from the game-theoretic analysis of simple 2 × 2 games is that the formal definition of (social) commitment provided in section 2 leaves too many essential dimensions of a commitment unspecified. If the commitment has to be agreed upon by the non-committed agent, the committed agent will commit in other situations than if the committed agent can unilaterally commit without agreement. It might be crucial whether one or both agents have an option to commit to an action and in which order the agents obtain the opportunity to commit. In game-theoretic terms, these options can be formalized by adding moves to the game that implement the possibilities for the agents to commit and, eventually, to accept the commitment of the other agent. These moves might be specified simultaneously or sequentially. Using game-theo-

Collective Comitment

237

retic reasoning, solutions of these extended games can be calculated, which provides predictions about whether or not commitments will be used and what the consequences of these commitments are depending on the chosen commitment system. As a result, insides are obtained about whether a commitment system is socially efficient or favors one of the two agents.

5 Conclusions First of all, there are complex and interesting questions in the logical theory of collective activity: From the point of view of methodological individualism, the action of a collective is in some sense composed of or determined by individual actions performed by the members of the collective. The general study of the nature of that composition, of the dependence of collective actions on individual ones, could be said to constitute the theory of collective action in a narrower sense. (Pörn 1970) A first attempt is made to ‘constitute the theory of collective action in a narrower sense’ by the introduction of commitment and to define a goal of a collective as a set of individual subtasks. The notion of commitment is very useful to determine which plan will be followed to achieve a collective task (collective commitment). In this article, we have shown that game-theoretic reasoning provides new insight with respect to commitments for logical systems. Specifically, our analysis demonstrates that the distinction between unilateral and bilateral commitments has been obscured in existing logical formalizations, but this distinction is important for the effectiveness of commitments in certain situations. The possibility of making a commitment without the necessity of the other agent’s agreement is also neglected in existing logical formalizations. To gain insight into the different types of commitment we introduced different commitment regimes. The game-theoretic analysis showed that the different regimes could lead to different outcomes in interactions between two agents. Consequently, the usefulness of logical systems formalizing commitment would be increased by specifying the incentive structure that lies beneath a certain formal specification and adapting the commitment regime such that an underlying “social goal” (social efficiency or equal distribution) can be realized. As long as the incentive structure beneath the interaction between two agents is unknown (i.e., without explicit reasoning why and when a certain motivational attitude leads to a commitment), it is unclear whether the speci-

238

Lambèr Royakkers and Vincent Buskens

fied commitment regime will work and will lead to an efficient outcome. We have shown that the four regimes can smoothly be formalized in a logical framework based on existing primitive notions as intentions and goals, in which we can distinguish whether or not a commitment is credible in a given interaction. This leads to a better understanding of the interaction of agents in an organization. So far, we have limited the game-theoretic analysis to a small sample of games. In further research, the analysis can be extended to 2 × 2 games without strictly ordered payoffs, or groups of agents who have to accomplish a task. This could lead a more general notion: a collective commitment. Such a commitment, we have defined as the internal commitment of a group (a collective agent) given a selected plan (in the line of the work of Bratman et al. 1988) and the underlying structure of the group with respect to the achievement of the collective task. An extension to games in which agents perform sequentially including actions for the commitment decisions seems interesting, which gives the possibility to analyze one-sided conditional commitments. Analysis of such games can be combined with formalizations using dynamic logic and conditional logic, which provides new insight in the dynamics of rational behavior. In the logical approach, we stated a relation between agreement and goal. This demands further elaboration. Finally, more normative issues have to be considered to model real organizations, especially if we view organizations as instances of normative systems, and describing agents’ interactions in terms of what those agents are permitted or obliged to do (allowing also the possibility that their behavior may deviate from the ideal), and in terms of other complex normative relations between them. In that case, it is essential that such concepts as right, permission, obligation, authority, responsibility and delegation, are precisely understood and defined, and not simply treated informally (cf. Jones and Sergot 1996). We want to register the opinion that too little has so far been done in Organization Theory to supply precise qualitative models of norm-governed interaction, and we think that some techniques drawn from a logical and game-theoretic setting may help to fill that gap.

Acknowledgement The article was supported by a grant of the Niels Stensen Foundation.

Collective Comitment

239

Bibliography Binmore, K. Fun and Games: A Text on Game Theory. Lexington: D.C. Heath and Company 1992. Bratman, M. E., D. Israel and M.E. Pollack. “Plans and Resource-Bounded Practical Reasoning.” Comput. Intelligence 4 (1988): 349-355. Buskens, V., and L.M.M. Royakkers. “Commitments: A Game-Theoretic and Logical Perspective.” Submitted to Cognitive Science Qarterly (2002). Castelfranchi, C. “Commitments: From Individual Intentions to Groups and Organizations.” Pp. 41-48 in Proceedings First International Conference on Multi-Agent Systems. Edited by V. Lesser. San Francisco: AAAI-Press and MIT Press, 1995. Castelfranchi, C. and R. Falcone. “Towards a Theory of Delegation for Agent-Based Systems.” Robotics and Autonomous Systems 24 (1998): 141-157. Cohen, P .R, and H. J. Levesque. “Intention is Choice with Commitment.” Artificial Intelligence 42 (1990): 213-261. Conte, R., and C. Castelfranchi. Cognitive and Social Action. London: UCL Press 1995. Dignum, F., J.-J. Ch. Meyer, R. J. Wieringa, R. Kuiper. “A Modal Approach to Intentions, Commitments and Obligations: Intentions plus Commitment Yields Obligations.” Pp. 80-97 in Deontic Logic, Agency and Normative Systems. Edited by M.A. Brown and J. Carmo. Berlin: Springer 1996. Dunin-Kęplicz, B., and R. Verbrugge. “Collective Commitments.” Pp. 56-63 in Proceedings Second International Conference on Multi-Agent Systems. Edited by M. Tokora. California: AAAI-Press, 1996. —. “Collective Motivational Attitudes in Cooperative Problem Solving.” Pp. 22-41 in Proceedings of CEEMAS’99. Edited by V. Gorodetsky. 1998. Elster, J. Ulysses and the Sirens: Studies in Rationality and Irrationality. New York: Cambridge University Press 1979. Gasser, L. “Social Conceptions of Knowledge and Action: DAI Foundations and Open Systems Semantics.” Artificial Intelligence 47 (1991): 107-138. Gilbert, M. On Social Facts. Princeton 1992. Gilbert, M. “Obligation and Joint Commitment.” Utilitas 11, no. 2 (1999): 143-163. Herrestad, H. and C. Krogh. “Deontic Logic Relativised to Bearers and Counterparties.” Pp. 453-522 in Anniversary Anthology in Computer and Law. Edited by J. Bing and O. Torvund. Oslo: COMPLEX - TANO 1995. Hohfeld, W. N. Fundamental Legal Conceptions. Edited by W. W. Cook. New HavenLondon: Yale University Press 1964. Jones, A. J. I., and M. J. Sergot. “A Formal Characterization of Institutionalized Power.”Journal of the IGPL} 4 (1996): 429-445. Kanger, S. “Foundations for Ethical Theory.” Pp. 1-35 in Deontic Logic: Introductory and Systematic Readings. Edited by R. Hilpinen. Dordrecht: Reidel 1971. Lindahl, L. Position and Change. Dordrecht: Reidel 1977. Luce, R. D., and H. Raiffa. Games and Decisions. New York: Wiley 1957.

240

Lambèr Royakkers and Vincent Buskens

Makinson, D. “On the Formal Representation of Rights Relations.”Journal of Philosophical Logic 15 (1986): 403-425. Meyer, J.-J. Ch., W. van der Hoek and B. van Linder “A Logical Approach to the Dynamics of Commitments.” Artificial Intelligence 113 (1999): 1-40. Mowday, R. T., L. W. Porter, and R. M. Steers. Employee-Organizational Linkages: The Psychology of Commitment, Absenteeism, and Turnover. New York: Academic Press 1982. Pörn, I. The Logic of Power. Oxford: Basil Blackwell 1970. Rao, A., and M. Georgeff. “Modelling Rational Agents within a BDI-Architecture.” Pp. 473-484 in Proceedings of the Second Conference on Knowledge Representation and Reasoning. Edited by R. Fikes and E. Sandewall. Morgan Kaufman 1991. Rapoport, A., M. J. Guyer, and D. G. Gordon. The 2 × 2 Game. Ann Arbor MA: University of Michigan Press 1976. Rasmusen, E. Games and Information: An Introduction to Game Theory (2nd), Oxford: Blackwell 1994. Robertson, P. J., and S.-Y. Tang. “The Role of Commitment in Collective Action: Comparing the Organizational Behavior and Rational Choice Perspectives.” Public Administration Review 55 (1995): 67-80. Royakkers, L. M. M. Extending Deontic Logic for the Formalization of Legal Rules. Dordrecht: Kluwer Academic Publishers 1998. —, and V. Buskens. “Logical Formalizations Built on Game-Theoretical Argument about Commitments.” Pp. 92-97 in Intelligent Agent Technology: Research and Development. Edited by N. Zong et al. New Jersey: World Scientific 2001. —, and F. Dignum. “Organizations and Collective Obligations.” Pp. 302-311 in Database and Expert Systems Applications. Edited by M. Ibrahim, J. Küng and N. Revell. Berlin: Springer 2000. Santos, F., A. J. I. Jones, and J. Carmo. “Responsibility for Action in Organizations: a Formal Model.” Pp. 333-350 in Contemporary Action Theory, Vol II (Social Action). Edited by G. Hölstrom-Hintikka and R. Tuomela. Dordrecht: Kluwer Academic Publishers 1997. Schelling, T. C. The Strategy of Conflict. Cambridge MA: Harvard University Press 1960. —. Choice and Consequences: Perspectives of an Errant Economist. Cambridge MA: Harvard University Press 1984. Shepsle, K. A. “Discretion, Institutions, and the Problem of Gorverment Commitment.” Pp. 245-265 in Social Theory for a Changing Society. Edited by P. Bourdrieu and J.S. Coleman. Boulder: Westview Press 1991. Shoham, Y. “Agent-Oriented Programming.” Artificial Intelligence 60 (1993): 51-92. Vogel Carey, T. “How to Confuse Commitment with Obligation.” The Journal of Philosophy 72 (1975): 276-284. Williamson, O.E. “Credible Commitments: Using Hostages to Suport Exchange.” American Economic Review 73 (1983): 519-540.

Justification and Consensus: The Peircean Approach

241

Justification and Consensus: The Peircean Approach Frederick F. Schmitt Abstract It is commonly recognized that the justified beliefs of an individual subject can be supported or undermined by a consensus on the proposition in the subject’s community. A more controversial view is that justified belief turns on consensus in a deeper respect: justified beliefs are correlated with consensual beliefs in a way to which we must attend when we evaluate or theoretically describe justified belief. Call this a consensus account of justified belief. C. S. Peirce proposed such an account, deriving from a more basic fixation theory of justified belief. I regard Peirce’s proposal as the most plausible consensus account now available, and his fixation basis for the account as the strongest available basis for a consensus account. In this paper, I describe Peirce’s fixation account in detail, and I elaborate his argument for the account. I defend the argument up to a point. In the end, however, the argument rests on an implausible empirical psychological assumption about belief-fixation. I list reasons for rejecting this assumption. The result should discourage enthusiasm about a consensus account of justified belief.

Social epistemology asks, among other questions, whether the conditions of individual knowledge include social conditions. One potentially relevant social condition is consensus in the subject’s community. It has long been recognized that a consensus on an issue can support or undermine a subject’s justification for belief – a point made explicit in Condorcet’s jury theorem.1 I take Peirce to have been the first philosopher to propose more boldly that the justified beliefs of individual subjects are correlated (contingently) with consensual belief, and that this correlation requires attention in the practical activities of seeking and evaluating justification, and in the theoretical enterprise of explaining the nature of justification.2 I will call such a proposal a consensus ac1 A more adventurous view is that consensus does not merely play a supporting or undermining role, but provides an autonomous source of individual justification, on a par with sense perception. To accord consensus this much epistemic power is to treat it in a manner analogous to Thomas Reid’s treatment of testimony. Indeed, the most promising basis for this view of consensus would assimilate it to Reid’s treatment of testimony, on the ground that consensus is a special case of testimony. I find the view attractive, but I will pass over it here in favor of the still more adventurous view that individual justification turns largely on consensus. 2 Peirce takes several routes to consensus that I will pass over here:

242

Frederick F. Schmitt

count of justified belief.3 Perhaps the simplest consensus account would say: a subject is justified in a belief p just in case all (or most) people in the subject’s community who hold a doxastic attitude as to whether p believe p. Though no one ever held such a simple view, more sophisticated consensus accounts of justification and rational belief have become increasingly popular in recent years. Peirce does not propose his consensus account as a basic theory of justified belief but derives it from his fixation account. Nevertheless, consensus plays a pivotal role in justification, on Peirce’s view, by constraining the justifying methods of belief-formation. Peirce’s consensus account is not only the first such account but in my view the most promising. It deserves careful atten(a) Truth: Truth is long-run consensus. Knowledge entails true belief (Peirce 1992: 24, line 8 up). So knowledge entails belief that conforms to long-run consensus. However, it is not clear that for Peirce truth entails an actual long-run consensus, as required by a consensus account of knowledge. First, he sometimes defines truth counterfactually, as what one would believe in the long run (123, 139, 230). Second, the subject in terms of which truth is defined is sometimes a counterfactualized individual, not a community (e.g. 230). (b) Cumulative Knowledge: Knowledge accumulates when beliefs are not only true (i.e., conform to long-run convergent opinion) but actually permanently fixed in all subsequent beliefs (that result from scientific method). Accumulated knowledge is in this sense part of actual convergent opinion. However, Peirce does not say that permanent fixation is a necessary condition of current knowledge. Current knowledge does not have to be part of the accumulation. In other words, knowledge can be lost. So current knowledge does not have to be part of actual convergent opinion. (c) Inductive Justification: Induction is rational for an individual because the conclusions of inductive inferences converge on the true population frequency in the long run. But these conclusions are guaranteed to converge only if the long-run use of inductive inference is communal use, rather than individual use. So the individual’s inductions are rational only if they are “altruistic” – i.e., contribute to the long-run communal use (or so the individual must presuppose). (d) Innate beliefs and beliefs from universal experience are universal, hence consensual. Perhaps these universal beliefs are justified because consensual. 3 I assume that, on a consensus account, a subject’s belief is justified only if others (if there are any others) actually hold beliefs of a related, consonant kind (or at least fail to hold beliefs of a dissonant kind). Thus, a consensus account involves something stronger than conformity to Bayesian convergence results. The Bayesian results entail that all possible subjects who assign probabilities in conformity with the probability calculus and update probabilities by Bayesian conditionalization (under independence assumptions) will assign convergent probabilities. This does not entail that any one subject assigns probabilities properly only if others’ probabilities actually agree. Nor do we employ the convergence result to evaluate the propriety of a subject’s assignments. (The result might enable me to judge whether others’ assignments are proper by checking to see if they agree with those of some designated subject, but only if I already know independently that the designated subject’s assignments are proper.)

Justification and Consensus: The Peircean Approach

243

tion; for its failure would signal that consensus plays no greater role in justification than the support and undermining commonly recognized. I will need to offer an interpretation of Peirce on these matters, but my interpretive aim will be modest. I will focus exclusively on the role of consensus elaborated in sections III-V of “The Fixation of Belief” (Peirce 1992) and not attempt to reconcile what I find there with the remainder of Peirce’s work. I will skirt Peirce’s fundamental concerns in that article, which surround the propriety and presuppositions of scientific method, and instead search for suggestions for social epistemology. I interpret Peirce as a naturalistic social epistemologist in this material, but I will relegate defense of that interpretation to Appendices 1 and 2. My attention in this paper is riveted on the epistemological role of consensus. My philosophical aim is to evaluate the consensus account I find in Peirce. I will object to Peirce’s argument for his account on the ground that the empirical premises on which the argument rests are implausible. Once his argument is rejected, his account itself appears implausible. Before getting underway, let me highlight four features of the consensus account I attribute to Peirce that make it worth examining: (1) Peirce argues for his consensus account by appeal to an end conception of justification, and such a conception of justification is more plausible than alternative conceptions. According to an end conception, a justified belief is a belief aimed (or “virtually aimed”) at a characteristic epistemic end. An argument for a consensus account by appeal to an end conception is, in my opinion, more promising than alternative arguments for a consensus account – e.g., arguments by appeal to a subject’s attitudinal commitments (such as respect), to cognitive practice, to the public character of justification, or to the need for objectivity.4 (2) Peirce’s end conception makes an argument for consensus possible because it makes consensus a derived, not a basic epistemic end. An end conception of justification could support a consensus account of justification simply 4 For an attitudinal commitment case for a consensus account (in particular, an appeal to the attitude of respect), see Keith Lehrer and Carl Wagner, Rational Consensus in Science and Society: A Philosophical and Mathematical Study (Dordrecht: Reidel, 1986). I criticize Lehrer and Wagner’s argument for the respect account, and the account itself, in “Consensus, Respect, and Weighted Averaging,” Synthese 62, 25-46. I criticize the publicity argument for a consensus account in “Consensus Theories of Justification: The Publicity Argument,” Esperienza e cononscenza: Introduzione all’epistemologia sociale, ed. Guido Piazza (Milan: LCS, 1995), 147-172.

244

Frederick F. Schmitt

by identifying the epistemic end with a consensus.5 However, a bald identity claim of this sort is intuitively implausible. Consensual belief is desirable for coordinating behavior, but this gives consensus a point only for a limited range of beliefs, and there is no obvious case along these lines for the claim that consensus is an epistemic aim. We need a different argument for this claim, and Peirce provides such an argument. (3) Indeed, Peirce offers the optimal argument for consensus from the epistemic end. His choice of basic epistemic end is the one most promising for delivering consensus – that of fixing belief. It is more likely that we can derive consensus from fixation than from other popular candidates for the epistemic end – e.g., true belief, informativeness, coherence, explanatory power, or simplicity. Thus, Peirce offers the best case for an end consensus account. (4) Peirce expressly designs his account to answer some of the troubling objections to consensus accounts, and his account answers these objections, up to a point, given the success of his argument for the consensus account. In this respect, it differs from all other proposed consensus accounts, which succumb to one or another of these objections. However, I lack space here to discuss this advantage of Peirce’s account. These four features of Peirce’s account together make it worth examining closely.

I. A Fixation Account of Justified Belief Peirce, I have noted, subscribes to an end conception of justified belief: justified belief is belief aimed at the basic (i.e., underived) epistemic end. My term “justified” here is simply a stand-in for Peirce’s term “good,” which he applies to both beliefs and belief-forming methods – it will carry no freight beyond this. Peirce’s particular choice of basic end is the fixation of individual (not common or communal) belief. Peirce’s consensus account of justified belief is then derived from his fixation account. The fixation account is itself embedded in a broader project of uncovering the presuppositions of inference. Peirce’s ultimate object in “The Fixation of Belief” is to make a qualified case for scientific method on the basis of the fixation account, within an account of the presuppositions of inference. Uncovering the presuppositions of scientific method is important because it provides the best available case for the hypoth5 See Ziman (1978) for an example of such an consensus account.

Justification and Consensus: The Peircean Approach

245

esis of reality. I will say a thing or two about scientific method and its presuppositions, but my main concern throughout is the derivation of the consensus account from the fixation account. As I read him, Peirce attempts to infer the basic end of inquiry from the nature of inquiry alone.6 He argues that inquiry begins only with doubt and ends with belief. Doubt is an uneasy and dissatisfied state from which we struggle to free ourselves and pass into the state of belief; while the latter is a calm and satisfactory state which we do not wish to avoid, or to change to a belief in anything else. On the contrary, we cling tenaciously, not merely to believing, but to believing just what we do believe. (114) Doubt is an uneasy, disequilibrium state that we seek to eliminate, preferably in favor of what we already believe, otherwise in favor of belief in an alternative (perhaps contrary) proposition. Belief, by contrast, is an equilibrium state accompanied by satisfaction. “The irritation of doubt,” Peirce says, “causes a struggle to attain a state of belief. I shall term this struggle inquiry...” (114). It is in the nature of inquiry to seek to terminate doubt with permanent belief. Thus, the end of inquiry is the fixation of belief. Peirce considers for a moment the view that true belief is an end of inquiry independent of fixation, but he rejects that view. His stated reason against it is unconvincing,7 but it is reasonable to assume that he has an unstated motive: he 6 My proposal here contrasts with Hookway’s (1985) interpretation, on which Peirce argues that the basic end is fixation because as a matter of human psychology, we naturally aim at fixation in inquiry and can have no other basic end. 7 Peirce argues against the idea that the aim of inquiry is true belief in this passage: With the doubt, therefore, the struggle begins, and with the cessation of doubt it ends. Hence, the sole object of inquiry is the settlement of opinion. We may fancy that this is not enough for us, and that we seek, not merely opinion, but a true opinion. But put this fancy to the test, and it proves groundless; for as soon as a firm belief is reached we are entirely satisfied, whether the belief be true or false. And it is clear that nothing out of the sphere of our knowledge can be our object, for nothing which does not affect the mind can be the motive for mental effort. The most that can be maintained is, that we seek for a belief that we shall think to be true. But we think each one of our beliefs is true, and, indeed, it is mere tautology to say so. (114115) Here is one way to read this argument: We can aim at something, X, as distinct from Y, only if when in state Y we do not necessarily think we are in state X. But when we are in a state of fixed belief, we do necessarily think we are in a state of true belief, since tautologously we hold our beliefs to be true. So we cannot aim at true belief as distinct from fixed

246

Frederick F. Schmitt

regards true belief, conceived as an end independent of fixation, as too difficult a target for individuals to achieve in any systematic way and therefore inappropriate as an end of individual inquiry. On Peirce’s theory of truth, a truth is a proposition belief in which will (or would) eventually be permanently fixed for the community of inquirers (as I will say, a proposition on which there will be convergent opinion). But individuals have too meager a chance of hitting on beliefs in propositions that will eventually be permanently fixed (no matter what methods they use). Though we can have an inductive basis for thinking that this or that proposition we now believe is already permanently fixed and thus true, we have such a basis for too few propositions. The closest that individuals can come (on average) to achieving true beliefs systematically is an end one might hope to be contributory to the distal end of convergent opinion. Peirce proposes that this achievable contributory end is individual fixed belief (i.e., persistent, though perhaps not permanent, belief). This proposal needs to be vindicated by establishing (by way of a presuppositional case) that the method that best achieves the end of fixed belief (namely, scientific method) is also the one that leads to the distal end of convergent opinion. Peirce makes a case that we must presuppose that scientific method leads to convergent opinion if it is to fix individual belief. In this way, the choice of individual fixation as an end is vindicated: this end contributes in the long run to true belief. I will return to these thoughts below. Meanwhile, I will assume that fixation is the individual end. The next step is to consider how this end is related to the justification of belief. It is clear that a justified belief is not merely a fixed (i.e., persistent) belief. Many tenacious beliefs are fixed, but they’re no good. It is plausible that, for Peirce, the goodness of a belief and of the method that forms or maintains the belief are analytically related. This leads me to attribute to Peirce a methodological end conception of justified belief: a belief is justified just in case it is managed or formed by a method that tends to satisfy the epistemic end. On this view, we broadly satisfy the epistemic end by routinely using methods that tend to satisfy the end, and good beliefs are the products of this arrangement. Inserting Peirce’s particular choice of end in the methodological end conception, we arrive at a fixation account of justified belief: belief. The mistake lies in the first premise. I can aim at my own happiness (X), as distinct from thinking that I’m happy (Y), even if when I am in Y (i.e., I think I am happy), I necessarily think I am in state X (i.e., I necessarily think I am happy).

Justification and Consensus: The Peircean Approach

247

(FIX1) A belief is justified just in case it is formed (or managed) by a belief-fixing method. But what exactly is a belief-fixing method? Peirce embraces a simple trackrecord account: a method is belief-fixing just in case it forms beliefs in such a way that its average belief-persistence and doubt-avoidance is high across all uses of the method.8 For the remainder of the paper, I will assume that this fixation account of justification is correct.

II. Fixation and Tenacity Peirce’s derivation of consensus from the fixation account begins by considering a putative counterexample to the fixation account: If the settlement of opinion is the sole object of inquiry, and if belief is of the nature of a habit, why should we not attain the desired end, by taking any answer to a question which we may fancy, and constantly reiterating it to ourselves, dwelling on all which may conduce to that belief, and learning to turn with contempt and hatred from anything which might disturb it? This simple and direct method is really pursued by many men. I remember once being entreated not to read a certain newspaper lest it might change my opinion upon free-trade. “Lest I might be entrapped by its fallacies and misstatements,” was the form of expression. “You are not,” my friend said, “a special student of political economy. You might, therefore, easily be deceived by fallacious arguments upon the subject. You might, then, if you read this paper, be led to believe in protection. But you admit that free-trade is the true doctrine; and you do not wish to believe what is not true.” (115-116) Peirce’s ultimate interest is in eliminating this and other methods by appeal to the fixation theory, in favor of scientific method. For the moment, however, he 8 FIX1 should be read as: necessarily, a subject’s belief is justified just in case it results from a method that fixes belief in actual subjects of the same psychological type as the subject. Without the qualification “of the same psychological type as the subject,” FIX1 is hostage to the psychology of any actual aliens there may be. Hookway (1985: 47-8) has suggested that Peirce values methods that would fix belief in a self-conscious subject. But this is textually implausible, for much the same reasons that the rationalist interpretation is textually implausible (see Appendix 1).

248

Frederick F. Schmitt

considers whether we have here a counterexample to the fixation account: the method of tenacity, as he calls it, is a belief-fixing method; so by the fixation account, tenacious beliefs are justified; yet intuitively this method is not a justifying method; hence the fixation account is mistaken.9 Peirce defends the fixation account from this charge by denying that the method of tenacity is really belief-fixing. He makes the point by appeal to social psychology, perhaps inspired by the Pyrrhonian tradition:10 But this method of fixing belief, which may be called the method of tenacity, will be unable to hold its ground in practice. The social impulse is against it. The man who adopts it will find that other men think differently from him, and it will be apt to occur to him, in some saner moment, that their opinions are quite as good as his own, and this will shake his confidence in his belief. (116) For Peirce, human sociality has a doxastic dimension. People who employ the method of tenacity observe that others hold beliefs contrary to their own; yet they deem these beliefs of others to be as good as their own. This Pyrrhonian observation, as I will call it, engenders enough loss of confidence in their own tenacious beliefs to undermine these beliefs (i.e. reduce their strength or cause their retraction). Thus, the method of tenacity is not belief-fixing, and so it is not a counterexample to the fixation account. Peirce appeals to undermining by the Pyrrhonian observation in order to force the fixation account into extensional equivalence with an account on which a belief is justified only if it is deemed by the subject better than contrary beliefs. This prevents the fixation account from sanctioning intuitively unjustified beliefs. We might say that 9 Must tenacity be bad? Couldn’t it be good to close your ears to protests when you expect that you will be unduly influenced by them? Couldn’t this be so even when you expect the protests to contain some good counterevidence? The answer seems to be Yes, but Peirce could assimilate tenacity in these cases to scientific method – it is based on observations of past reliability. 10 The Pyrrhonian source is Sextus Empiricus, Outlines of Pyrrhonism I 88, the second oppositional mode of Aenesidemus: But if the same objects affect men differently owing to difference in the men, then, on this ground also, we shall reasonably be led to suspension of judgement. For while we are, no doubt, able to state what each of the underlying objects appears to be, relatively to each difference, we are incapable of explaining what it is in reality. For we shall have to believe either all men or some. But if we believe all, we shall be attempting the impossible and accepting contradictories; and if some, let us be told whose opinions we are to endorse. (Sextus Empiricus 1976: 53) Hume carries on the Pyrrhonian tradition in his discussion of the uneasiness of doubt arising from contradictions (Hume 1974: 205-6).

Justification and Consensus: The Peircean Approach

249

Peirce’s appeal to undermining by the Pyrrhonian observation underwrites a sort of bootstrapping that secures for the account of justification an intuitively correct extension, without having to write conditions that secure this extension into the account explicitly. How exactly does the Pyrrhonian observation that the contrary beliefs of others are as good as one’s own undermine belief? Various lines of reasoning from the observation might drive one to retract one’s belief. For Peirce’s purposes, these lines of reasoning need not themselves be justifying; it is enough if they instill a doubt leading to retraction. (By the same token, they must all be preempted by a method if the method is to fix belief.) The various lines of reasoning will differ, depending on what is meant by saying that others’ contrary beliefs are as good as my own. I will list several possible interpretations, without attempting an exhaustive survey. On one reading, the observation is that others hold contrary beliefs (e.g., believe that protectionism is more productive than free trade) as likely to be true as my own. I infer from this that my own belief is very unlikely to be true, and this conclusion undermines my belief. There is reason to doubt that this is what Peirce has in mind. The fact that my own belief is unlikely to be true (in any sense of “true”) should undermine my belief only if “unlikely to be true” means epistemically unlikely or unjustified, rather than objectively unlikely; for I can surely be highly justified in believing an objectively unlikely proposition. Understood as unjustified, however, the inference fails: I cannot infer from the fact that others’ beliefs in protectionism are as justified as mine that my belief in free trade is unjustified. For I can perfectly well recognize that others are as justified in their beliefs as I am in mine without feeling that my reasons for belief are in jeopardy. Let me, therefore move on to two more successful readings of the observation. I will not try to choose between these but employ them disjunctively in what follows: (a) The Available Methods Reading: On what I will call the available methods reading, the observation says: others hold contrary beliefs as justified as my own in a special way. For example, they possess justifying reasons for their protectionist beliefs; these reasons are available to me, and thus I ought to take them into account; and were I to possess them, I would no longer be justified in my belief. We may work here with a methodological variant of this idea: others use methods as justifying as my own to manage protectionist beliefs, and their methods are available to me for forming such a belief. I might think, for example, that by reading the forbidden newspaper, exposing myself to

250

Frederick F. Schmitt

doubts about free trade and to counterreasons against the reasons that have convinced me to accept free trade, and by shunning my old free trade friends, I can end up with a protectionist belief as justified as my current free trade belief. This unsettling thought then undermines my belief. I will call a method as justifying as my own that yields a contrary belief and is available to me for use (on an occasion) an available competing method (with respect to my belief via the method). I take it that if there is an available competing method, then I ought not to employ the method I do to arrive at my belief. The thought that there is an available competing method undermines my belief. (b) The Total Reasons Reading: On a second reading, a “total reasons” reading, the Pyrrhonian observation does not assume that others’ methods are available to me. Rather, I believe that I ought not to believe p if I think that the totality of reasons for and against p may well not favor p. And I observe that other subjects manage contrary beliefs in protectionism by methods as justifying as my own. From this I infer that the total reasons may well not favor free trade. Thus, my observation leads me to the conclusion that I ought not to believe in free trade. This conclusion then causes me to retract my belief in free trade.11 In fact, we need not read the Pyrrhonian observation in any one way. Diverse observations could undermine belief. It is important to remember that at this point we seek a plausible psychological account of undermining, not an account of when my justification for my belief is undermined by these sorts of observations. It is enough if Peirce has one or more plausible stories to tell of the psychology of undermining.12 For simplicity, however, I will generally as11 The total reasons Pyrrhonian observation is related to Peirce’s fallibilism. The observation says: my belief p may well not align with the total reasons. Fallibilism says: many of my beliefs may well not be true – i.e., align with the total reasons. The former has enough power to undermine my belief p; the latter does not. The difference in power lies in the relative specificity of the Pyrrhonian doubt. 12 Here is yet another story. Suppose I believe that methods should conform to the constraint of yielding the same belief in relevantly similar circumstances. And suppose I think that others’ circumstances are relevantly similar to mine, though tenacity yields contrary beliefs in these circumstances. This undermines my faith in tenacity and thus my belief. This story does, however, need to be supplemented with an account of “relevantly similar circumstances” and an explanation of why (in my view) others are in relevantly similar circumstances. Why say that the difference in what we will and the consequent difference in the doubts we are exposed to does not make our circumstances relevantly dissimilar? The answer may be that the circumstances of others are relevantly similar to mine because I could, in a certain sense, be in their circumstances and obtain their belief. I could alter my circumstances to resemble theirs sufficiently to obtain their

Justification and Consensus: The Peircean Approach

251

sume the available methods reading of the Pyrrhonian observation (that others use available competing methods) and return to the total reasons reading only when the difference between the two readings matters. To summarize, Peirce responds to the alleged counterexample of the method of tenacity by admitting that this method is unjustifying, but denying that it is belief-fixing. The method is vulnerable to undermining induced by the Pyrrhonian observation.

III. Consensus and Authority With these remarks about Peirce’s fixation account and the method of tenacity behind us, let us see how Peirce proceeds to the consensus account of justification. According to Peirce, the method of tenacity fails to fix beliefs because those who use the method eventually recognize that others reach contrary beliefs via available competing methods, and this recognition undermines tenacious beliefs. Thus, Peirce derives from the fixation account the following condition: A method M is belief-fixing only if the beliefs formed (or managed) by M are not undermined by this Pyrrhonian observation. But beliefs formed by M are not so undermined only if it is not salient to subjects who use M that others use available competing methods with respect to these beliefs and M. Thus, M is belief-fixing only if, for most beliefs formed by M, there are no (or few) occasions in the community on which people use available competing methods with respect to those beliefs and M. As we may put it, M is belief-fixing only if it generally has no competition in the community. This “no competition” condition is satisfied if M is the best available method on the question on which it is used. Peirce is not, however, in a position to judge yet which method is the best available, because on his fixation account, the best method is the most belief-fixing, and which methods are belief-fixing is precisely what is at stake here. So Peirce naturally turns to a different way in which a method M can satisfy the no competition condition. M satisfies the no competition condition if no one in the community uses methods that are availcontrary belief via a method as justifying as my own. Given this answer to why our circumstances are relevantly similar, the story is equivalent to the available methods reading of the Pyrrhonian observation.

252

Frederick F. Schmitt

able and yield beliefs contrary to those produced by M. And this happens if the method M squelches the competition. That is, the use of M generally prevents the use of competing methods that would otherwise be available – for simplicity, we may say, prevents the use of available competing methods. If M prevents the use of available competing methods on average, we may say that M is a consensus-producing method. Peirce thus infers a consensus account of justification from the fixation account: (SOC1) A subject’s belief is justified only if it is managed by a method M that produces a consensus on most questions on which it is used by any community member, and does this within most communities in which it is used, where A method M produces a consensus within a community on a question when the use of M prevents people in the community from using available competing methods with respect to beliefs on the question and M.13 The method of tenacity fails SOC1, and this leaves room for Pyrrhonian observations, which do arise and prevent the method from fixing beliefs. At this point, a new counterexample to FIX1 comes into view. The method of authority satisfies both SOC1 and FIX1; yet intuitively it is less than justifying: Let the will of the state act, then, instead of that of the individual. Let an institution be created which shall have for its object to keep correct doctrines before the attention of the people, to reiterate them perpetually, and to teach them to the young; having at the same time power to prevent contrary doctrines from being taught, advocated, or expressed. Let all possible causes of a change of mind be removed from men’s apprehensions. Let them be kept ignorant, lest they should learn of some reason to think otherwise than they do. Let their passions be enlisted, so that they may regard private and unusual opinions with hatred and horror … This method has, from earliest times, been one of the chief means of upholding correct theological and political doctrines, and of preserving their universal or catholic character. (117) 13 This account is obviously circular: it employs the notion of “good methods” to characterize a justifying, i.e., good, method. But the circularity can be discharged by replacing “good methods” with “belief-fixing methods.”

Justification and Consensus: The Peircean Approach

253

The method of authority evidently includes indoctrination by such means as catechism, civic education, public exhortation, and censorship. It may also include acculturation by parents and teachers. The beliefs produced by authority may include those passed down by tradition – e.g., beliefs in the classification of animals, plants, natural objects, and artifacts, to the acceptance of proverbs and folk maxims, a knowledge of etiquette, and other such shared beliefs within a community.14 However, I assume that Peirce intends basic, rather than derivative, reliance on authority. Where there is independent firsthand reason to believe that an authority (a witness to a crime, an expert in medicine) is reliable (by induction from perceptually acquired beliefs), the method is really scientific method, not authority. (It should be noted that Peirce allows common sense beliefs to play a crucial role as background for the abductions and qualitative inductions of scientific method. Many of these common sense beliefs are innate, and others are products of acculturation involving authority. However, the features of these beliefs that entitle them to play this role owe nothing to authority (i.e., they owe nothing to being products of the method of authority). From the standpoint of Peirce’s concerns in “The Fixation of Belief,” their key features are their universality and persistence over generations – features that permit their use without threatening the belief-fixing powers of scientific method.) The method of authority Peirce allows to be much more efficacious than the method of tenacity. When used in a community, the method generally does produce a consensus, by preventing people from using available competing methods. Peirce notes that the method is limited to fixing beliefs on a few questions – specifically, questions of theology and politics. For “no institution can undertake to regulate opinions upon every subject” (118). On most questions, however, scientific method is employed instead: “Everybody uses scientific 14 There is some question just what kind of method the method of authority is. Is it a beliefforming method used by individuals on various occasions – namely, the method of believing the pronouncements of an authority? Or is it a method that forms beliefs wholesale – a belief-profile-forming method, if you like, as with mass media? Or is it not a belief-forming method at all but a style of social institution (or a procedure for establishing social institutions) that regulates the use of belief-forming methods? Peirce treats the method in all three ways, but the first is his main understanding. In his proposal that the method of authority is best for most people, Peirce treats the method as an individual method involving believing pronouncements of an authority. Since some people can abstain from using it – can abstain from uptake from authority – the method is not a belief-profile-forming method or style of institution. I will assume this understanding of the method of authority.

254

Frederick F. Schmitt

method about a great many things, and only ceases to use it when he does not know how to apply it” (120). The fixation account thus entails that the method of authority is justifying on few questions. Nevertheless, the method poses a problem for FIX1 and SOC1, since it is not intuitively fully justifying even on these few questions. Peirce responds to this alleged counterexample to FIX1 and SOC1 with ambivalence. On the one hand, he insists that the method of authority is the best most people can do and thus justifies their beliefs: “For the mass of mankind, then, there is perhaps no better method than this” (118). On the other hand, he maintains that the method is not the best of all possible methods, and it does not justify the beliefs of all subjects. There are people, independent thinkers, for whom it forms unstable beliefs: These men possess a wider sort of social feeling: they see that men in other countries and in other ages have held to very different doctrines from those which they themselves have been brought up to believe; and they cannot help seeing that it is the mere accident of their having been taught as they have, and of their having been surrounded with the manners and associations they have, that has caused them to believe as they do and not far differently. (118) Peirce appeals once more to a Pyrrhonian observation. This time, however, the observation is cross-cultural in scope: “others” include individuals beyond the subject’s community. A standard catechism for the young or a papal encyclical may lead to orthodox beliefs about the Trinity or about the death penalty within a religious community, but it does not secure uniform belief across communities. And there are, within communities, elite minorities who observe that in other communities the method of authority gives rise to contrary beliefs as good as their own. This cross-cultural Pyrrhonian observation percolates a doubt that undermines belief. Thus, the method of authority fails to fix beliefs in all individuals even in its home territory of theology and politics. The conditions of justification we have so far developed, FIX1 and SOC1, are clearly too crude to capture Peirce’s ambivalence about authority. They do not distinguish “the mass of mankind” from those with “a wider sort of feeling”: they make authority equally justifying for all (on those topics on which it produces a consensus). How, then, might we refine these conditions to reflect Peirce’s ambivalence? The simplest refinement that comes to mind is unfortunately inadequate. It is to attribute to him a subject-specific fixation condition – i.e., a method justifies a subject’s belief because it fixes that subject’s beliefs – as opposed to fixing the beliefs of subjects in general. On this ap-

Justification and Consensus: The Peircean Approach

255

proach, we would say that the method of authority justifies the beliefs of most subjects because it fixes their beliefs; but it fails to justify the beliefs of sophisticated subjects because it does not fix their beliefs. But the subject-specific condition is ruled out here. For Peirce claims that scientific method is better than authority. This means that scientific method is more belief-fixing than authority. But if fixation is specific to a subject, then this isn’t so. Authority fixes beliefs in most people just as well as scientific method fixes beliefs in sophisticates. On the subject-specific condition, authority is better than scientific method for most people, while for sophisticates, it is vastly inferior. But this not what Peirce claims. He claims that authority is rather good, though scientific method is better. For most subjects, authority, which is highly fixing, is the best they can do, though scientific method, which is perfectly fixing (with respect to the Pyrrhonian observation), is better. Sophisticates can do somewhat better than authority, by using scientific method. These claims rule out the subject-specific condition in favor of the general condition – a method must fix the beliefs of subjects in general. Thus, we cannot understand Peirce’s criticism of the method of authority by appeal to the subject-specific condition. What is the alternative? I read Peirce as introducing here, albeit implicitly, a comparative condition on justification. The method of authority is the best most people can do, and so it justifies their beliefs. The method is belief-fixing, and for most people, it is better (i.e., more generally fixes belief) than any available method that yields contrary beliefs. In other words, the method of authority meets not only an absolute but a comparative condition of belief-fixation. This assumes a comparative account of justification: (FIX2) A belief is justified just in case it results from a belief-fixing method, and there is no available competing method. This comparative condition affords a qualified criticism of authority. Authority, though best for most people, is not best for sophisticated subjects, since they can use scientific method, which is more generally belief-fixing.15 15 Could Peirce have something different in mind when he refers to sophisticated thinkers – that they foment change and thus undermine authority in general? On this view, the method of authority is not in the long run consensus-producing or belief-fixing. Unfortunately, this option does not do the trick. Peirce does recognize that the method of authority does not prevent beliefs from undergoing gradual change. Referring to the organized faiths, he says: If we scrutinize the matter closely, we shall find that there has not been one of their creeds which has always remained the same; yet the change is so slow as to be imperceptible during one person’s life, so that individual belief remains sensibly fixed. (118)

256

Frederick F. Schmitt

Sophisticated subjects given to the cross-cultural Pyrrhonian observation, then, will abstain from using the method of authority because there is a more potent method available to them that yields contrary beliefs. The a priori method, as Peirce calls it, allows natural inclinations to reign, and produces a consensus through discussion. In this way, it may be as fixing in individuals for a period of fashion as authority is over generations, but the cross-cultural Pyrrhonian observation reminds us that natural inclinations are as accidental as will is capricious. Scientific method is more potent than either authority or the a priori method because it escapes the undermining effect of cross-cultural Pyrrhonian observations. For, we may presuppose, it settles belief “by nothing human, but by some external permanency – by something upon which our thinking has no effect” (120). (For our purposes, we may simply define scientific method as a method that, we may presuppose, determines belief by reality. Of course, the claim that perception, induction, and abduction have this feature would require further argument.) No method that determines beliefs by will or circumstance, as tenacity and authority do, can avoid cross-cultural uses that lead to contrary beliefs. Scientific method alone, then, can escape undermining by the Pyrrhonian observation in sophisticated subjects. The best method presupposes the hypothesis of reality.16

IV. Consensus and Scientific Method Thus FIX2 forces a new consensus account, applicable to sophisticated subjects only. A method will escape all Pyrrhonian observations only if it produces consensus not only within communities, but across them as well. This leads us to the stronger consensus account (SOC2) A subject’s belief is justified only if the belief results from a method M that produces consensus on most questions on which it is used by anyone, But the present suggestion requires that the danger posed by independent thinkers is revolutionary rather than evolutionary, since only that danger threatens individual belief. Peirce, however, recognizes only gradual change, which threatens not the fixation of individual belief but rather consensus over time. 16 This is not Peirce’s only presuppositional case for the hypothesis of reality in “The Fixation of Belief.” He also argues that the dissatisfaction with contrary propositions that underlies the dissatisfaction of doubt presupposes the hypothesis of reality (120).

Justification and Consensus: The Peircean Approach

257

where A method M produces a consensus on a question when it prevents people everywhere from using available competing methods with respect to beliefs on the question and M. A method must satisfy the condition of SOC2 in order to satisfy FIX2 for sophisticated subjects. As a matter of human psychology, only scientific method satisfies SOC2 because only scientific method forswears reliance on human will or circumstance, a reliance that will lead to contrary beliefs in others. Appeal to the Pyrrhonian observation thus allows Peirce to argue that subjects can reasonably use a method to manage their beliefs only if they believe that their method determines belief by a nonhuman reality. Unless subjects (at least those who make the cross-cultural Pyrrhonian observation) presuppose (hence believe) that scientific method determines belief in this way, the method will not fix beliefs. I have just exhibited the reasoning that is supposed to show that scientific method satisfies the consensus condition, but it must be admitted that it is far from clear just how this reasoning is supposed to work. Suppose I do not believe that my method is hostage to human will but believe instead that it determines my belief by reality. How does that prevent a Pyrrhonian observation from undermining my belief? Scientific method certainly does produce contrary beliefs in others. Past science contradicts present science, though, on Peirce’s view, it sometimes used the same methods. And similarly, present science is controversial. Since I observe this fact, I do make a Pyrrhonian observation: others hold contrary beliefs via a method as good as my own. Peirce must explain why this observation does not undermine belief, as in the case of tenacity or authority. To fill in Peirce’s reasoning, let us consider what can be said in defense of scientific method on each of the readings of the Pyrrhonian observation mentioned earlier. (a) The available methods reading: The Pyrrhonian observation might undermine my belief by leading to the metabelief that there is available to me a competing method. In the case of tenacity, I form this metabelief because I recognize that I might alter my will so that tenacity manages a contrary belief. Peirce might claim, however, that in the case of scientific method, the method is not available to form a contrary belief. It is a formal constraint on scientific method that in using it, I have already taken into account all the counterreasons possessed by others that are available to me. If I am an astronomer

258

Frederick F. Schmitt

developing views about planetary motion, I have already taken into account Galileo’s reasons for his contrary views and the reasons of my contemporary colleagues for their contrary views. The ground for this requirement on scientific method is that (we may presuppose) it uses reality to determine belief, and it must therefore avoid rude surprises by taking available reasons into account at the earliest feasible moment. So the belief I form using scientific method is the one any use of the method available to me would form (assuming only that my reasons uniquely determine a proposition to believe, under the method). Thus, as far as the “available methods” reading of the Pyrrhonian observation goes, scientific method does forestall undermining by satisfying SOC2. (b) The total reasons reading: The second way the Pyrrhonian observation might undermine my belief p is by leading to the metabelief that the totality of reasons anyone could have for and against p may well tell against p. In the case of tenacity, I form this metabelief because I recognize that will affects the output of tenacity, and this makes an alignment between what my reasons indicate (p) and what the total reasons indicate no more likely than an alignment for any other subject who uses tenacity. Now, the situation may be much the same for scientific method: reality might affect the output of scientific method so that the reasons possessed by diverse individuals are equally unlikely to be aligned with the total reasons. Since I recognize this, I must conclude that the total reasons may well tell against p. But, Peirce might claim, in the case of scientific method, this metabelief need not undermine my belief p. For I might conceive of my use of scientific method as a contribution to a communal project of eventually fixing individual belief in the human community over the course of history. That is, I might think of what I am doing in using scientific method as contributing to making scientific method maximally belief-fixing, by contributing to its leading to eventual individual fixation (hence convergent belief). It might be important for such a project that I maintain my beliefs even in the face of contrary beliefs, until reasons possessed by diverse subjects can be amalgamated and employed to revise my beliefs and those of others. Others may need their contrary beliefs, and the stimulus of my contrary beliefs, to find and publish reasons. So I should maintain my beliefs until I come into possession of enough counterreasons to require retraction. I should do this, despite my awareness that the total reasons may well tell against p, because doing so makes scientific method the best method to use. This is a case in which acting the way I do makes my belief justified, or so I hope. This point against undermining entails that my use of scientific method has a certain presupposition. It presupposes that scientific

Justification and Consensus: The Peircean Approach

259

method contributes to long-run fixation – fixing individual beliefs in the human community over the course of history. Thus, while scientific method allows a “total reasons” Pyrrhonian observation, this observation does not undermine belief when subjects presuppose the hypothesis of reality. Perhaps these points are enough to make a case that scientific method is less vulnerable to undermining by a Pyrrhonian observation, in either its available methods or its total reasons readings, than tenacity and authority are. (I discuss the matter further in Appendix 3). Of course, this shows at best that scientific method satisfies SOC2 as far as the Pyrrhonian observation goes. It does not show that scientific method satisfies SOC2 without restriction. For we have only presupposed the hypothesis of reality, and satisfying SOC2 requires that the hypothesis of reality be true. If scientific method does not really fix beliefs by reality, then we have no reason to expect that it will avoid contrary beliefs across cultures. Showing that scientific method satisfies SOC2 thus requires showing that the hypothesis of reality is true. However, Peirce appears to conceive of presupposition itself as a belief-fixing method – indeed, as a method that permanently fixes belief. But on Peirce’s convergent opinion account of truth, any belief that is permanently fixed is true. So Peirce’s account of truth entails that the hypothesis of reality is true. Thus, scientific method does satisfy SOC2. With these points behind us, we can return for a moment to Peirce’s motive for selecting fixed belief as the basic end. I earlier attributed the following motive. True belief is not an end individuals can achieve by themselves or even in concert within a single generation. For, on Peirce’s definition of truth, a truth is a proposition belief in which will (or would) be permanently fixed for the community of inquirers, and individuals have too meager a chance of hitting on beliefs in such propositions (no matter what methods they use). The most individuals can achieve is an end that contributes in the long run to convergent opinion. Fixation is that end. As Peirce argues, the method that best fixes individual beliefs also leads to convergent opinion, or so we must presuppose if the method is to fix belief. It is the job of the Pyrrhonian observation to make the case that this presupposition is needed for fixation. And this vindicates the choice of fixation as the end. The end of fixation in the short-run can now coincide with the end of true belief, so long as we use scientific method. This vindication at the same time provides Peirce with a way of defending his convergent opinion account of truth. One difficulty with Peirce’s definition of truth is that it allows truth to result from any method that happens to lead to convergent opinion. And this is counterintuitive. But having established that

260

Frederick F. Schmitt

scientific method both fixes belief and (by presupposition) leads to convergent opinion, Peirce can now rewrite the definition of truth to say: a truth is a proposition belief in which would eventually be permanently fixed as a result of scientific method. This rewritten definition rules out crazy convergent opinion as the truth. To summarize our discussion so far, I have interpreted Peirce as deriving a consensus account of justification from his fixation account and employing it to fend off counterexamples to the fixation account. The method of tenacity is ruled out because tenacious beliefs are undermined by the Pyrrhonian observation. Tenacity fails to fix beliefs because it fails the consensus condition SOC1 – it fails to prevent the use of available competing methods in the subject’s community. The method of authority does much better: it satisfies this condition. For some subjects, however, authoritative beliefs are undermined by a cross-cultural Pyrrhonian observation. These subjects need scientific method – method that, we may presuppose, determines belief by reality – to fix their beliefs. Scientific method fares better than authority in satisfying a cross-cultural consensus condition, SOC2, as far as the Pyrrhonian observation goes. Thus, there is a case that it satisfies the fixation account. Users of scientific method must presuppose the hypothesis of reality. This presupposition vindicates the choice of fixation as the basic epistemic end and provides Peirce with a defense of his convergent opinion account of truth.

V. Evaluating Peirce’s Argument for the Consensus Account I have taken the exposition of Peirce’s consensus account as far as I am able, and it is time now to gauge its merits. I want to concede that if Peirce’s argument for the consensus account works – if Peirce succeeds in deriving the consensus condition from the fixation account via the Pyrrhonian observation – then the account has an advantage over other consensus accounts in answering the standard objections to consensus accounts. Most importantly, it answers the objection that consensus accounts make consensual beliefs that result from authority or indoctrination justified, though these beliefs are intuitively unjustified. Indeed, Peirce’s account is expressly designed to yield the result that authority is less than fully justifying on the fixation account of justification. The consensus condition is designed to render the method of authority less than fully justifying, as intuition requires.

Justification and Consensus: The Peircean Approach

261

The trouble, however, is that Peirce’s argument for his consensus account does not work. It does not establish that consensus is necessary for fixation. In particular, it rests on an implausible empirical psychology. Peirce’s argument fails, and without the argument, tenacity and authority reemerge as a counterexample to the fixation account and the consensus account. Peirce’s case for requiring a consensus-producing method, to review, is that, on the fixation account, a justification-conferring method on average yields persistent beliefs; yet subjects’ beliefs are undermined by the Pyrrhonian observation, which arises when they are exposed to contrary beliefs. I will continue to assume the fixation account, as I have all along. I grant that Peirce’s best bet for showing that the methods of tenacity and authority fail to fix beliefs is to appeal to undermining by the observation that there are available competing methods. I grant, too, that this observation is stimulated only by exposure to the contrary beliefs of others. Consequently, preventing exposure to these contrary beliefs would forestall the observation. The methods would fix belief if they satisfied the consensus account – a result that Peirce wants. The difficulty for Peirce’s argument, however, is that methods can perfectly well fix beliefs even when they fail to satisfy the consensus account. In particular, tenacity and authority do fix beliefs despite failing to satisfy the consensus account – i.e., despite the presence of contrary beliefs and even Pyrrhonian observations. Thus, consensus is not necessary for fixation. Exposure to living examples of contrary beliefs often falls short of stimulating observations that undermine tenacious and authoritative beliefs. Subjects often do not notice that others use competing methods. And when they do make Pyrrhonian observations, these often leave beliefs intact. Peirce’s argument rests on a questionable empirical social psychology of doubt. I list four problems with Peirce’s psychology of doubt: (i) Not given to the Pyrrhonian observation: People are not as given to the Pyrrhonian observation upon exposure to the equally good contrary beliefs of others as Peirce assumes. One relevant point from recent cognitive psychology is that people are not as cognitively egalitarian as aptness to the Pyrrhonian observation requires. They are not as given to thinking that the beliefs of others are as good as their own as Peirce assumes. People tend to be overconfident in their beliefs (at least in those beliefs that answer moderately difficult questions involving common knowledge).17 They tend to accord their beliefs more 17 The psychological literature on overconfidence concentrates on the degree to which probability judgments are miscalibrated – in particular overshoot their target frequencies. But I take this to correlate with mismatches between judgments of reliability and

262

Frederick F. Schmitt

confidence than they deserve – where confidence is an assignment either of reliability or of weight of reasons. People are not, however, similarly overconfident about the beliefs of others. To be sure, if people are faced with contrary beliefs of others that result from methods they deem as good as their own, then they ought to have a confidence in others’ beliefs equal to their confidence in their own beliefs. But this suggests that people do not typically deem the methods of others as good as their own. Peirce himself makes this observation in the second sentence of “The Fixation of Belief” (109), but he does not notice its relevance to his argument for a consensus condition. The upshot is that people are less given to the Pyrrhonian observation than Peirce assumes.18 (ii) Beliefs not undermined by Pyrrhonian observations: Even when people make the Pyrrhonian observation, their belief is less likely to be undermined by it than Peirce assumes. Teachers commonly lament the fact that exposure to foreign systems of thought – whether those of neighbors with a different religion, or those of another culture – has very little tendency to undermine the entrenched beliefs of students. It does tend to make them skeptics or relativists, but their conviction in these matters is a theoretical second-order conviction, and as soon as they return to first-order discussion they are as obstinate as ever. In this connection, I note that the Pyrrhonists recommended a regimen of Pyrrhonian observations, not merely unsystematic observations, in order to stimulate doubt, and no one thinks they ever achieved epoche even with this regimen. Even immersion in another culture does not commonly reverse lifelong beliefs. The point here is consonant with what we know of historical episodes – the Hellenistic period in Greece or the European exploration of America – in which cultures first come into prolonged contact. In some cases, one culture comes to dominate or absorb another, but this is often a matter of communal, not individual, belief giving way, and when individual belief does give way, it is the result of arduous missionary work or brutal indoctrination. Beliefs seem to have built into them some degree of resistance to doubt, including doubts inspired by the Pyrrhonian observation. In recent cognitive psychology, it has been argued (by Lee Ross and others) that beliefs perseweight of reasons, and between actual reliability and weight of reasons. Studies in this literature include Oskamp (1982) and Lichtenstein et al. (1982). 18 Admittedly, the empirical findings are consistent with the idea that in explicit conscious reflection on the reliability of their beliefs or the weight of reasons for their beliefs (Peirce’s “saner moment”), people tend to show modesty, claiming no more for their own beliefs than for others’. But how large a role in belief revision do the products of explicit conscious reflection play?

Justification and Consensus: The Peircean Approach

263

vere.19 For one thing, a belief tends to persist after the reasons for which it is believed are lost, even when subjects are apprised of that loss. This is a cognitive mechanism that maintains beliefs in the absence of reasons. More pertinent here, however, is that a belief tends to persist in the presence of doubts, both rebutting doubts (i.e., counterreasons) and undercutting doubts – a recognition that reasons for the belief are lacking. The Pyrrhonian observation is an undercutting doubt, and belief perseverance may have some force against it. Moreover, the mechanism of belief perseverance is not the only force that maintains belief in the absence of reasons. It is a commonplace that motivational and emotional states have this effect, as in wishful thinking and selfdeception. And it is plausible that they too maintain belief in the presence of an undercutting doubt. Finally, the phenomenon of group polarization applies to belief as well. Exposed to the beliefs of other groups, individuals who belong to a group tend to develop beliefs contrary to those of other groups.20 This is one mechanism by which groups increase their internal solidarity and their differences from other groups. It is worth noting that Peirce himself at one point refers to psychological forces that maintain belief: “we cling tenaciously, not merely to believing, but to believing just what we do believe” (114). It is possible that he has in mind here only precautions we take to avoid exposure to doubts that might undermine our belief, and not psychological mechanisms that protect belief. Such precautions are not, however, our only protection against undermining in the absence of reasons. In any case, Peirce overestimates the capacity of undercutting doubts like the Pyrrhonian observation to undermine our belief. (iii) Not exposed to contrary opinions of others: People are not as widely exposed to the contrary opinions of others as Peirce assumes. There are fewer occasions for making the Pyrrhonian observation than one might think. While contrary opinions (via methods deemed as good as one’s own) may be widespread, the question is how widely subjects are exposed to these opinions in a way that would trigger the Pyrrhonian observation, and in such a way as to undermine belief. And the answer seems to be that we are not as widely exposed as Peirce assumes. A subject’s belief is more likely to be undermined by exposure to confrontation and disputation than by mere recognition that others have contrary opinions. By “confrontation,” I mean another person’s knowing expression of 19 See Nisbett and Ross (1980, ch. 8). 20 For discussion of group polarization, see Turner (1987) and Wetherell (1987).

264

Frederick F. Schmitt

opinion contrary to the subject’s own, directed toward the end of undermining the subject’s belief. And by “disputation,” I mean confrontation accompanied by an expression of reasons for the contrary opinion or doubts that rebut the subject’s reasons for belief. Confrontation and disputation may have an undermining effect, but they are not pervasive in everyday life. Mere knowing expression of contrary opinion is more common than confrontation and disputation. Admittedly, it too can have an undermining effect on a subject’s belief. But even knowing expression of contrary opinion is not as common as one might assume. People do not generally make a habit of knowingly expressing contrary opinions in opposition to the confident beliefs of hearers. Quite the reverse: people generally make a habit of avoiding expressing opinions contrary even to the less confident beliefs of hearers, and when they do express such opinions, they often express them too timidly to precipitate undermining. The degree of confidence expressed and the vigor of expression are as likely to be inversely proportional to the perceived degree of conviction of the target as they are to be proportional to it.21 There are prudential reasons for this diffidence. The knowing expression of contrary opinion in everyday life often intimates disapproval, or is taken to do so. The expression of contrary opinion is often taken to intimate a mild criticism of cognitive character or at any rate of cognitive action. What is disapproved is, typically, the subject’s lack of diligence in considering available counterreasons – counterreasons that, had they been taken into account and been given the proper weight, would have caused the subject to abstain from belief and indeed have led to the contrary belief. And the prudent desire to avoid intimating such disapproval inhibits the liberal expression of contrary opinion. When the issue is not urgent enough to risk raising hackles, people often avoid revealing their contrary opinions. They will instead whisper them to their companions as they walk away from the conversation. I add to this that there is a norm of speech etiquette according to which, in speaking to anyone with whom one is not very familiar, frankly expressing contrary opinion requires a tone of modesty, disclaimers of confidence, or expressions of reasons one thinks ought to convince one’s audience. I do not wish to exaggerate the strength or extent of the inhibition of knowingly expressed contrary opinion. Nor do I say that the speech etiquette 21 True, we tend to be more open with family members and close friends. But it is also true that we tend to agree with our intimate acquaintances more than with others, and disagreement with them is less likely to undermine belief.

Justification and Consensus: The Peircean Approach

265

against it has as much force as, perhaps, a prohibition on blunt talk about a stranger’s pimples. I do not deny that there is great variation among individuals and cultures in sensitivity to expressed doubt, inclination to express it, and the frankness of its expression. These variations may correlate with the gender, social position, or the age of conversants, and with the medium of expression – face to face or by email. But it must be admitted that the phenomenon of knowing expression of contrary opinion is relatively limited in everyday life. To be sure, there are special forms of discourse in which the expression of contrary opinion and controversy and disputation are encouraged: the discourses of teaching, formal debate, newspaper editorials, academia, science, democratic electoral politics, and art and cultural criticism, to make a short list. But these, while important, are exceptions to the rule. Nothing in this concession changes the fact that in common life people are not overly given to expressions of contrary opinion. The exceptions to the rule – academia, science, democratic politics – make an interesting study. In these enterprises, there are institutional norms and mechanisms that function, and in some cases are designed, to overcome the common inhibition and abridge the etiquette against expressed contrary opinion. Political debates, symposia in academia, and peer review of journal articles and grant proposals all have as an immediate aim facilitating and sanctioning the expression of doubts. In this connection it is worth noting that academic referees are often kept anonymous not only when they disapprove the refereed work, but even when they approve it. Perhaps this protection is extended against the abnormally grudging author inclined to revenge for any criticism, however favorable the overall judgment. But another motivation for protection is that even the most generous and self-effacing people may, unpredictably, against their will, and often unknown to themselves, hold a grudge for contrary opinions or doubts expressed. The recognition that offense at expressed contrary opinion is common gives some prudential motive to avoid it, and some need for protections where expressed contrary opinion is desirable. In short, people are often inhibited from expressing contrary opinions, and countervailing institutional norms and incentives are needed to overcome it. The upshot of all this is that undermining exposure to contrary opinions is not as common as Peirce assumes. (iv) Countervailing infectiousness of neighbors’ beliefs: As a last point against Peirce’s case for the undermining effect of the Pyrrhonian observation, I observe that there is a social mechanism that countervails the undermining effect of exposure to contrary opinion. While Peirce is no doubt right that ex-

266

Frederick F. Schmitt

posure to contrary beliefs sometimes undermines belief to some extent (when it occurs), it is also true that beliefs are infectious, at least in the sense that anyone’s expression of belief in a proposition generally enhances the chance that hearers will believe the proposition. The infectiousness of belief, as well as doubt, is recognized by Hume (1974: 592-3). Peirce’s formulation of the influence of others’ opinions on our own (116-117) is in fact general enough to subsume both belief infectiousness and undermining, and he explicitly recognizes infectiousness in his remark about reading the newspaper (115). But he does not notice the bearing of infectiousness on his treatment of justification. Infectiousness may be underwritten by an innate disposition to credulity, as Reid believed. Or it may be underwritten by social pressures for conformity of opinion, as in the phenomenon of group polarization I referred to above. These underwriting mechanisms may in turn result (via biological or cultural selection) from the utility or necessity of consensus for coordinating behavior. It is a commonplace that people tend to like and get along better with those who share their opinions than with those who do not, and productive coordinated behavior requires common belief. The utility of consensus may drive the infectiousness of expressed belief. But whatever the origin of infectiousness, it is a social psychological phenomenon with the same order of strength as the undermining effect of expressed contrary opinion. And in extent it would appear to be much greater – at least in its intragroup operation. It causes belief and enhances the persistence of belief even in the absence of a complete consensus. The key point is that expressed contrary opinion stands side-by-side with the infectiousness of belief. Peirce loads the dice in favor of the undermining effect of expressed contrary opinion when he cites undermining without mentioning the infectiousness of expressed belief. Points (i)-(iv) should put to question Peirce’s assumption that expressed or manifested contrary opinion extensively undermines belief. I have questioned Peirce’s social psychology of doubt – his assumptions that people are given to the Pyrrhonian observation, that the Pyrrhonian observation undermines belief, that expressed contrary opinion is widespread, and that there is no countervailing infectiousness of belief. Without these assumptions, we lose Peirce’s case that the fixation of belief requires consensus-producing methods. Tenacity may resist undermining by the Pyrrhonian observation and make a counterexample to both the fixation account and the consensus account.22 22 Peirce’s case against authority may fare slightly better than tenacity here. Since authority fails to fix belief only in elite subjects, and elite subjects are defined as those who attend

Justification and Consensus: The Peircean Approach

267

Of course, I don’t deny that consensus can contribute to fixation by minimizing exposure to living examples of contrary beliefs via methods deemed as good and thus by preventing Pyrrhonian observations. For Pyrrhonian observations can undermine belief. Consensus may at any rate serve as a sort of insurance against potential undermining by Pyrrhonian observations. What I deny is that consensus is needed for fixation. Tenacity remains a counterexample to the fixation account because beliefs are not so easily undermined by a failure of consensus.23 To summarize my efforts, I have found in Peirce’s text a methodological consensus account of justification that I regard as a serious contender among consensus accounts. I have, however, objected to his argument for the consensus account on the ground that the social psychology on which the argument rests is implausible. Peirce’s appeal to the fixation account doesn’t in the end support a role for consensus.

Appendix 1: Naturalism versus Rationalism in “The Fixation of Belief” On the interpretation of Peirce I offer in this paper, his view is naturalistic in two ways. It is psychologistic: the conditions of justification are wholly psychological, involving belief-forming methods and the fixation of belief. And the view assesses methods by their contingent belief-fixing properties in human beings: a belief is justified, at least for a human subject, just in case it results from a method that fixes beliefs in human beings. The methods of tenacity and authority are rejected because, as a matter of human psychology, they fail to fix beliefs. to cross-cultural disagreement, Peirce gets around point (iii) above, that people are not widely exposed to the contrary opinions of others. 23 Might Peirce retreat from the claim that fixation generally requires consensus-producing methods to the claim that such methods are required only in special circumstances? In particular, might Peirce argue that consensus-producing methods are needed where doubts are frequently expressed – in academia, science, and democratic politics? Indeed, with the ever-increasing cultural significance of science and with ever-expanding higher education in developed nations, might the disputational culture of academia become the common culture and the consensus account a fully applicable account of justification? But this gets things backwards. This line of reasoning would have Peirce arguing that because there are doubts, belief-fixing methods must produce consensus. But recall that these doubts are artificially induced. If belief-fixation is the aim, then it would make more sense to eliminate the institutional mechanisms that induce doubt than to impose a consensus requirement.

268

Frederick F. Schmitt

This naturalistic interpretation faces four questions, which I will endeavor to answer here. (1) Isn’t Peirce opposed to psychologism? Doesn’t he reject the idea that the conditions of justification are psychological and that justification turns on contingent features of human psychology? Answer: I can respond here only to the narrow textual question whether Peirce is opposed to psychologism in “The Fixation of Belief,” leaving unsettled whether he is so opposed in other works. I say he is not opposed to psychologism in “The Fixation of Belief.” Peirce does expressly reject one version of psychologism – the “impulse theory” that reasoning is good just in case “when the premises are accepted by the mind, we feel an impulse to accept the conclusion also” (112). But there is no reason to think that he rejects this proposal because it employs psychological terms, or because it makes satisfying the conditions of good reasoning turn on whether the subject possesses a contingent psychological tendency. I propose this interpretation: Peirce is trying to derive the conditions of justification from the nature of inquiry alone. He argues that the aim of fixation is constitutive of inquiry. Thus, successful inquiry involves satisfying the aim of fixation. Peirce assumes, plausibly, that successful inquiry would be haphazard without the use of methods that tend to fix beliefs. So proper inquiry involves the use of such methods. Justified belief is the outcome of proper inquiry. It is thus legitimate to refer to the psychological state of fixation and the use of belief-fixing methods in characterizing justification. Once the conditions of justification are seen to be psychological, contingent psychological facts are relevant to justification because they determine whether the conditions of justification are satisfied. In particular, they determine which methods are belief-fixing. Thus, the disposition to undermining by the Pyrrhonian observation is relevant because it determines which methods are belief-fixing. The impulse power of an argument is not relevant because it does not determine which methods are belief-fixing. It determines only what one will tend to believe. These remarks explain how Peirce can reject the impulse theory without rejecting psychologism. They do not, however, explain why Peirce insists, in the same place, that good reasoning turns on fact: “the question of its validity is purely one of fact and not of thinking” (111-112). Here Peirce seems to contrast questions purely of fact with questions of psychology and thus reject any psychologistic account of good reasoning. The question of fact that Peirce has in mind is the conditional reliability of the argument form – its tendency to take true premises into true conclusions. This is not a psychological condition

Justification and Consensus: The Peircean Approach

269

of good reasoning. However, there is a way to reconcile Peirce’s insistence that good reasoning turns purely on fact, and specifically on conditional reliability, with a psychologistic account of good reasoning. It could be that good reasoning can be defined at different levels, at one level in terms of pure fact, at another in terms of psychology. In particular, good reasoning could be defined basically in terms of psychology and derivatively in terms of fact. I believe that this is just what Peirce is up to. If we distinguish a basic condition of justification from a derivative condition, we can reconcile Peirce’s overtly psychologistic account of justification in sections III-V with his insistence on pure fact in section II. Arguably, it is one central point of “The Fixation of Belief” to derive the factual condition of good reasoning from the psychologistic account of justification. The argument is that only conditionally reliable reasoning affords the convergence required by fixation (the convergence we must presuppose if undermining is to be avoided). Reasoning that often takes true premises into false conclusions will never converge but be perpetually corrected by perception. Thus, the requirement of conditional reliability follows from the requirement of fixation. (Of course, the derivation goes through only under contingent psychological assumptions about perception and undermining. However, these are permitted because they are made relevant by the psychological conditions of justification.) We need to assume that Peirce has something like this derivation in mind to make sense of the overall argumentative structure of “The Fixation of Belief.” If he does have this in mind, then his insistence on fact is compatible with the psychologism I attribute to him. (2) Doesn’t Peirce hold an eschatological fixation theory of justification, and can’t he use that theory to argue directly for the presupposition of the hypothesis of reality, without assuming any empirical psychology? Answer: According to the eschatological fixation theory, a subject is justified in a belief just in case the belief results from a method that leads to convergent opinion. On this interpretation, the basic end for Peirce is not individual belief but communal fixation. But there are several textual objections to this interpretation. (i) The method of tenacity clearly fails the eschatological theory because it does not lead to long-run consensus. But this isn’t Peirce’s criticism of the method of tenacity. His criticism is that tenacity fails to fix beliefs on average in individuals – just what we would expect if he holds the average fixation theory. (ii) Peirce recognizes that the method of authority does not lead to long-run consensus (there is intergenerational drift, 118). But again, this isn’t his criti-

270

Frederick F. Schmitt

cism of authority. Moreover, authority is to some degree good because it fixes beliefs in individuals, despite its failure at long-run consensus. (iii) The eschatological theory lumps tenacity and authority together as equally bad methods. But Peirce prefers authority to tenacity. (iv) When Peirce praises scientific method, he does so not because it leads to long-run consensus, but because it fixes current beliefs. (v) Peirce’s case for the presupposition that scientific method converges is that we must believe this if scientific method is to be justifying. On the eschatological interpretation, this means that Peirce’s case for the presupposition of convergence is that we must believe in convergence if scientific method is to lead to convergence. While this is true (by Peirce’s lights), it is textually more plausible to suppose that Peirce’s case for the presupposition of convergence is that we must believe in convergence if scientific method is to fix beliefs now and on average, as the average fixation interpretation proposes. (vi) The eschatological fixation theory requires us to presuppose the hypothesis of reality in order to form beliefs at all, not in order to fix beliefs. So it doesn’t distinguish belief-formation from belief-fixation, as Peirce does. The second question above is whether Peirce could not argue more directly for the presupposition of the hypothesis of reality by adopting the eschatological theory and arguing that in using a method that satisfies the theory one must presuppose the hypothesis of reality. This would avoid an appeal to the psychology of doubt and undermining by the Pyrrhonian observation. But there are two obstacles to such an easy route to the hypothesis of reality. First, individuals cannot take convergent opinion as their end, since its achievement is vastly beyond their means. At best, they can take as an end doing what would contribute to convergent opinion if others were to behave appropriately. But this may be nothing other than the end of fixing belief. In any event, a method that would contribute to convergent opinion does not automatically presuppose achieving convergent opinion. Second, using a method that leads to convergent opinion does not obviously require one to presuppose achieving convergent opinion in the same way that scientific method presupposes the hypothesis of reality. The latter method fails to achieve the end of fixation unless the user believes the hypothesis of reality. But a method leading to convergent opinion does not require, for achieving the end of convergent opinion, that the user believes that the method achieves convergent opinion. At best, the user need only believe that it is feasible for the method to achieve this end.

Justification and Consensus: The Peircean Approach

271

(3) Isn’t Peirce a rationalist, not a naturalistic epistemologist? Doesn’t he hold this view: a subject’s belief is justified just in case it results from a method that would fix belief in the subject if the subject were rational (alternatively: would fix belief in any rational thinker)? Answer: No. There are the following points against a rationalist interpretation: (i) Peirce says that tenacity would be justifying (although irrational – out of line with reason or scientific method) if the subject lacked the human disposition to undermining by the Pyrrhonian observation and tenacity therefore succeeded in fixing belief (116). But under these conditions, tenacity would not fix belief in the subject if the subject were rational (or fix belief in any rational thinker). For a rational thinker’s belief would be undermined by the Pyrrhonian observation. Thus, for Peirce, a subject can be justified by a method that would not fix belief if the subject were rational. Peirce rejects the rationalist account of justification. For another example, authority is not the most justifying method because it does not fix beliefs in those who possess “a wider sort of social feeling,” not because it does not fix beliefs in rational subjects. (ii) Peirce has no resources to characterize a rational thinker prior to discovering that scientific method satisfies his account of justification. Rational thinking is simply reasoning in accordance with scientific method. It is desirable because scientific method is justifying. Thus, Peirce cannot define justified belief as what a rational thinker would believe. And he cannot appeal to what a rational thinker would believe to argue that scientific method is more justifying than alternative methods like tenacity and authority. (iii) The rationalist account categorically rejects methods that would not fix belief in the subject if the subject were rational (or would not fix belief in any rational thinker). But Peirce does not categorically reject these methods – authority has positive value. Moreover, tenacity and authority are ranked on the same scale as scientific method. But the rationalist account does not permit this. (iv) On the rationalist account, there is no difference between belief-formation and belief-fixation. What would fix belief in a rational thinker is just what would produce belief in a rational thinker. But for Peirce, there is an important difference between these. Tenacity and authority fail because they fail to fix belief, not because they fail to produce belief. (4) Can’t a rationalist account of justification do as well as the naturalistic account I attribute to Peirce? Doesn’t it have as much power to deem scientific method justifying, and tenacity and authority unjustifying?

272

Frederick F. Schmitt

Answer: The rationalist account has as much power to deem scientific method justifying as the naturalistic account does. However, it has no power to deem tenacity and authority unjustifying. Of course, I have argued that the naturalistic account has no power to do this either, so the two accounts are in the same shape in this regard. (i) The rationalist account can deem scientific method justifying. For under the presupposition of the hypothesis of reality, no Pyrrhonian observation can undermine a belief that results from scientific method. (ii) The rationalist account does no better than the naturalist account on tenacity and authority. Would any rational thinker abstain from belief via these methods because of a Pyrrhonian observation regarding the use of the methods? To be more accurate, the rationalist account does not require that a rational thinker would abstain from belief via these methods because of the observation that actual subjects have contrary beliefs as good. A rational thinker need only abstain because of reflection that there are available competing methods, or some similar reflection. Nothing guarantees that any rational thinker would make an observation of actual subjects using these methods. A rational subject need not be embedded in an environment in which actual subjects use these methods. Moreover, nothing in the notion of a rational thinker guarantees that such a thinker would make any observation of his or her environment; Peirce’s notion of rationality is a notion of reasoning in a certain way and does not entail powers of observation. But the rationalist account need not rely on observations of actual subjects. It need only claim that a rational thinker considering the use of these methods would reflect on whether there are available competing methods. Of course, this means that the rationalist account has no case for a social condition.24 Thus, the rationalist account rules out tenacity and authority only if: (a) any rational thinker would (in relevant circumstances) make the Pyrrhonian reflection; and (b) if a rational thinker made this reflection, he or she would not form the belief via the method. 24 Of course, it might be that human subjects would count as rational only in social conditions only because human beings make such a reflection only in social conditions. But this does not mean that for human beliefs to satisfy the rationalist account of justified belief, they need social conditions. All that is needed is that they use methods that would fix belief in a rational person, regardless of whether these methods do fix belief in the subject.

Justification and Consensus: The Peircean Approach

273

The rationalist account requires only that (a) and (b) are satisfied on some reading of the Pyrrhonian reflection, whether it is an available methods reading or a total reasons reading (as discussed in section III above). Let us consider first the available methods reading. On this reading, (b) poses no problem for the rationalist. The explanation of why a rational thinker would abstain from belief if he or she made the Pyrrhonian reflection might parallel the available methods story about the undermining of tenacity I told for the naturalist interpretation (section II). (b) follows from two plausible assumptions: a formal constraint on a rational thinker, that a rational thinker would not believe p via a method if it were not rational for her to believe p via that method; and a formal constraint on a rational belief, that it is not rational for a thinker to believe p via a method if the thinker rationally believes that there are available competing methods with respect to believing p via the method. Any rational thinker will satisfy the formal constraints on a rational thinker, and any belief that a rational thinker would form will satisfy the formal constraints on a rational belief. So we should grant the rationalist (b) on the available methods reading. Unfortunately, (a) makes trouble on this reading. Would a rational thinker make the Pyrrhonian reflection for tenacity and authority? It is plausible enough that a rational thinker would use a method only after having reflected on whether there are available competing methods. The question is whether a rational thinker would be bound to reflect that there are available competing methods in the cases of tenacity and authority. In the case of tenacity, it seems that the only candidate for an available competing method is tenacity itself. For this is the only method that a rational thinker can be sure is as good as tenacity. To know that authority or scientific method is as good, the rational thinker would have to know whether they equally fix beliefs in a rational thinker, and that would require going through the same reasoning we are going through now, introducing a circularity. It would require knowing whether there are available methods that compete with authority and scientific method, and hence knowing whether tenacity is as belief-fixing as these methods – precisely what the rational thinker is now trying to determine. Even if a rational thinker can know that scientific method is maximally belief-fixing, she can’t know yet that tenacity is not.25 So she can’t know whether scientific method 25 Of course, this lack of certainty about which methods are best might by itself cause a rational thinker to abstain from using tenacity. But if the rationalist relies on this to reject to tenacity, the result will be skepticism, since the same line of reasoning applies to scientific method and all other methods.

274

Frederick F. Schmitt

fixes belief, since it might be trumped by tenacity. In short, for the rational thinker, tenacity is the only candidate for an available method competing with tenacity. The question, then, is whether a rational thinker would reflect that tenacity is available to produce a contrary belief. There is no problem with the rational thinker concluding that tenacity is available to her; for if it were not, then it would not fix her belief in the first place and thus would not be a candidate for a proper method. The problem is rather whether the rational thinker is bound to think that it is feasible for her to use tenacity to manage a contrary belief in the rational thinker.26 I suppose that tenacity could be defined so that it is feasible to use it under a change of will, and from this definition the rational thinker could conclude that tenacity can manage a contrary belief – it isn’t tenacity if it can’t. But it should be clear that a definitional proof is not good enough here. Tenacity poses a problem even when defined differently, in Peirce’s way (“constantly reiterating it to ourselves, dwelling on all that may conduce to that belief, and learning to turn with contempt and hatred from anything which might disturb it,” 115), and this definition doesn’t entail that it is feasible to use tenacity under a change of will. Peirce’s definition allows tenacity to be used by those who are unable to change their will when it is used. It is a matter of contingent psychology whether or not it is feasible to use tenacity to manage a contrary belief. Tenacity in this sense is still intuitively irrational and must be excluded by the account of rational belief. The problem for the rationalist account is that an arbitrary rational thinker need not believe that under her psychology it is feasible to use tenacity to manage a contrary belief. So (a) is mistaken on the available methods reading: in the cases of tenacity and authority, not just any rational thinker would make the Pyrrhonian reflection. Hence, the rationalist account does not rule out tenacity and authority on this reading of the Pyrrhonian reflection. The alternative reading of the reflection is the total reasons reading. Here the most pressing problem is (b). On this reading, the rational thinker must reflect that tenacity manages beliefs in such a way that the total reasons 26 Of course, one would only need tenacity to protect one’s beliefs from undermining if unpossessed counterreasons and other circumstances could change them. But tenacity could still be used by a person lacking this trait, so the question of whether the rational thinker has reason to reject it still arises. One might say that tenacity would no longer be intuitively justifying for such a person. But the method is not justifying for us, and what makes a method justifying for us, on the rationalist account, is whether it would fix belief in an arbitrary rational person.

Justification and Consensus: The Peircean Approach

275

may well not favor p. Presumably, the reflection is that it is possible for tenacity to manage contrary beliefs depending on will, and from this the rational thinker infers that the total reasons may well not favor p. This is supposed to establish (a). But the rational thinker believes that she ought not to believe p if the total reasons may well not favor p. Thus, she infers that she ought not believe p, and this causes her to abstain from belief. This is supposed to establish (b). The question for (b), however, is whether any rational thinker must believe that she ought not to believe p if the total reasons may well not favor p. Unfortunately, this is not a formal constraint on rational belief (as were the analogous claims on the available methods reading). The proposition assumes a substantive end of belief – alignment with total reasons. And it is not obvious that any rational thinker must embrace this end. There is, indeed, a case for disallowing belief in the proposition. The proposition connects what one ought to believe with reasons. But it is one point of Peirce’s validation of scientific method to establish that what I ought to believe must depend on reasons. If we had already reached the end of the validation, then we could assume that any rational thinker is committed to some connection between what she ought to believe and reasons (although it would still not be obvious that she must embrace the end of alignment with total reasons). But at the present moment, we are in the middle of the validation; we are in the process of trying to establish it by eliminating a nonreasoning method, tenacity, and we have not reached the required conclusion. So we cannot assume that a rational thinker believes the proposition. Hence, (b) is mistaken.27 Tenacity is not ruled out on the total reasons reading of the Pyrrhonian reflection. 27 It is not entirely clear that the rationalist is entitled to (a) either. Must a rational thinker reflect that it is possible for tenacity to manage contrary beliefs depending on will, or infer that the total reasons may favor p? Whether tenacity manages contrary beliefs in such a way that the total reasons may well not favor p depends on the rational thinker’s psychology and environment. No doubt the rational thinker can know that this holds for some possible subjects by reflecting on possible psychologies and environments. The question, however, is whether a rational thinker has knowledge of her own contingent psychology and environment. The rationalist account can of course be understood to grant this, but if it does so, the account is brought closer to one on which rational belief is belief managed by a method that would fix beliefs in the subject of justification. These accounts explain the impropriety of tenacity in a similar way – by appeal to rational subjects with a human-like psychology. They differ only in that the subject-relative rationalist account makes propriety vary from one subject and psychology to the next, while the non-subject-relative account makes propriety hang for all possible subjects on whether a method fixes belief in any rational subject, with whatever psychology.

276

Frederick F. Schmitt

I conclude, tentatively, that the rationalist account lacks the resources to rule out tenacity and authority as proper methods. Of course in saying this I am not promoting the naturalist account as an adequate account either, since I argued that it too, for different reasons, lacks the resources to eliminate these methods. Neither account can do the job. These points together defend the naturalistic interpretation I propose.

Appendix 2: A Methodological Fixation Account versus A Belief-Doubt Model of Inquiry Those who ascribe to Peirce a certain “belief-doubt model of inquiry” (Levi 1995) may be inclined to reject my methodological fixation interpretation on the ground that it is inconsistent with the belief-doubt model. On the beliefdoubt model, a subject holds a belief as a result of inquiry only if doubt is terminated. But then, once inquiry is terminated, there is no serious possibility of falsehood – the belief gets probability one. So there is no question of needing reasons for belief, and no question of its being justified in a sense that requires reasons. On this interpretation, Peirce makes no distinction between justified and unjustified beliefs. In particular, the status of a belief does not depend on where it came from – on whether it was formed by a belief-fixing method. On a related note, knowledge is not justified true belief for Peirce. Answer: I see no inconsistency between a methodological account and any belief-doubt model Peirce plausibly holds. (i) There is no question that in “The Fixation of Belief,” Peirce accepts some belief-doubt model of inquiry as an analytic account of the nature of inquiry (whether or not the inquiry conforms to norms), and he accepts it as a psychological account of how we inquire. I take the psychological belief-doubt model simply to say that we believe only when doubts have been put to rest, and we have no motive to retract until further doubts arise. In this sense, we countenance no serious possibility of error in our beliefs and we assign them probability one. But that does not entail that there is no normative distinction between beliefs. We psychologically can’t overturn beliefs until further doubts arise, and so there is no question about whether we should do so. But it doesn’t follow from this that there is no distinction between beliefs in normative status. The distinction might consist of the fact that for some beliefs we should actively seek doubts to overturn them and for others we should not. This is

Justification and Consensus: The Peircean Approach

277

Peirce’s substitute for needing reasons for the belief. And the difference between beliefs in this regard might be a consequence of their methodological status. This difference has consequences for how we should treat beliefs – in particular, whether we should rely on them in reasoning. In this sense, then, we do countenance a serious possibility of error in some beliefs and assign probabilities less than one. The belief-doubt model, then, should not be interpreted as barring our seeking doubts about beliefs. This is shown by an additional point. If you read the belief-doubt model as barring our seeking doubts, then it is inconsistent with Peirce’s idea that we ought to aim at convergence on the truth. If the claim that a subject does not allow for the serious possibility that a belief is false means that the subject should not seek doubts about it, then the subject will have an incentive not to engage in behavior that leads to retracting beliefs and thus an incentive to avoid convergence toward the truth. The right response to this is to say that the subject does and should countenance a serious possibility of error in the sense that he or she seeks doubts about the belief. (ii) In my view, the idea that there is an inconsistency between a normative belief-doubt model and a methodological account of justified belief would involve a level confusion. The normative belief-doubt model (so far as Peirce accepts it) belongs to the account of good reasoning (assuming for the moment that the account of good reasoning gives us an account of how we ought to form and revise beliefs). The account of good reasoning is derivative from the methodological account of justified belief. In particular, it emerges from the choice of scientific method as the optimal justifying method. We ought to reason in accordance with scientific method. Scientific method tells us when we should seek doubts to overturn our beliefs and when not. It tells us when we should regard our beliefs as having a serious possibility of falsehood. The epistemic status of a belief after inquiry for purposes of good reasoning is determined by scientific method, since good reasoning just is scientific method, and scientific method determines the preferred features of its premises. However, scientific method recognizes the manner in which a belief was formed as relevant to these features. On this story, the relation between a belief’s being justified (by whatever method) and how the belief ought to be treated in good reasoning is complex, as it intuitively is. (iii) The methodological account of justification is not an account of how we ought to proceed in supporting and revising beliefs. So there is no immediate clash with a normative belief-doubt model of inquiry. The distinction between justified and unjustified belief I attribute to Peirce isn’t a deontic distinc-

278

Frederick F. Schmitt

tion but an evaluative one. Justified belief is Peirce’s “good opinion” and justifying method is Peirce’s “good method.” Take goodness to be endowed by method and to be something one can obtain regardless of how the subject ought to proceed in supporting or revising the belief. Justification or lack of it does not immediately imply that one should behave one way or another toward a belief. That said, I think that it is natural to read Peirce as holding that how you should behave toward a belief is dictated by which method is good for you. The goodness of a belief is relevant to the extent that it affects how the method treats the belief – e.g., the desirability of seeking new doubts that might undermine the belief. (iv) For the above reasons, the belief-doubt model is consistent with the idea that for Peirce knowledge is justified true belief. Peirce is committed elsewhere to the claim that knowledge entails true belief (vol 1: 24, line 8up). There is textual evidence in “The Fixation of Belief” in favor of a justified true belief account of knowledge. The only relevant passage in the work, as far as I can see, is this: “the object of reasoning is to find out, from the consideration of what we already know, something else which we do not know. Consequently, reasoning is good if it be such as to give a true conclusion from true premises, and not otherwise” (111). On a natural reading, this passage entails that (under one qualification) knowledge is true belief that results from good reasoning. I interpret the first sentence as entailing that the object of reasoning is satisfied on an occasion when the subject does arrive at knowledge of the conclusion from (a consideration of) knowledge of the premises. I take the second sentence as assuming that good reasoning is by definition reasoning that satisfies the object of good reasoning. It follows from these points that, for Peirce, knowledge isn’t mere true belief. It also follows that knowledge that results (by whatever means) from a known premise is true belief that results from good reasoning that takes knowledge as premise. For the first point entails that the object of reasoning is satisfied whenever a known conclusion results from a known premise (by whatever means). And the second point entails that the object of reasoning is satisfied only when the conclusion results from good reasoning. These points together entail that the known conclusion must result from good reasoning if it results from a known premise. This is a qualified justified true belief account of knowledge. On my reading, knowledge cannot be mere true belief. For suppose it were. Then, according to my reading of the first sentence, satisfying the object of reasoning would not require good reasoning. Good reasoning would merely facilitate satisfying the object of reasoning by providing a means of satisfying

Justification and Consensus: The Peircean Approach

279

it that is repeatable on various occasions. And that contradicts my reading of the second sentence, according to which good reasoning is what satisfies the object of reasoning. I don’t say that my reading of the passage is the only possible one, but I think it is the most straightforward reading. (v) If the belief-doubt model were inconsistent with the idea that knowledge is justified true belief, it would also be inconsistent with the idea that knowledge entails true belief – an idea Peirce clearly endorses. For suppose the belief-doubt model rules out the idea that knowledge is justified true belief. This would presumably be on the ground that the question of a need for reasons cannot arise: all beliefs are doubt-free and thus have equal status. This assumes that knowledge is the relevant status. Otherwise, there would be no reason to say that knowledge does not require justification. But then the beliefdoubt model also rules out the idea that knowledge entails true belief, since false beliefs should have equal status with true ones. (For the same reason, knowledge cannot be rejection of a serious possibility of falsehood, or any other status beliefs share. For knowledge entails truth, and not all beliefs are true.) I take this to be good reason either not to attribute the belief-doubt model or to see it as consistent with the justified true belief account of knowledge.

Appendix 3: Does Peirce’s Support for Scientific Method Succeed? There are five sources of doubt about Peirce’s claim that scientific method is more belief-fixing than authority: (1) Isn’t scientific method just as vulnerable to a total reasons Pyrrhonian observation as authority is? For abduction and qualitative induction depend on imagination and common sense background beliefs, and these vary across cultures; so we cannot expect scientific method to converge. Answer: Peirce thinks that there are enough innate background beliefs (about force, matter, space, time, etc.) and background beliefs from universal experience (fire burns, etc.) to afford convergence. Peirce comes slowly to a recognition of the dependence of abduction and qualitative induction on common sense. See Hookway (1985) for helpful discussion of this issue. (2) Don’t the underdetermination and incommensurability of scientific theories prevent convergence and underwrite a new undermining Pyrrhonian

280

Frederick F. Schmitt

observation?28 Answer: A Peircean answer to long-run underdetermination is that it is ruled out by pragmatism, since theories underdetermined by all possible data are equivalent upon clarification. (But Peirce himself may be barred from this answer, on pain of circularity, since he wants to use his account of scientific method – in particular, the claim that proper method is exhausted by deduction, induction, and abduction – to establish pragmatism.) The answer to incommensurability should be that one who is worried about it will not be worried by the Pyrrhonian observation that a belief may well not align with total reasons, since there are no total reasons; there are at best total reasons relative to a paradigm, and scientific method does allow convergence to belief that aligns with such relative total reasons. (3) Even waiving troubles with undermining by the Pyrrhonian observation, does scientific method really satisfy the fixation account? For belief-turnover for scientific method might well exceed that of authority, and thus scientific method might be less fixing on average than authority, hence less justifying than authority. Answer: This is a problem for Peirce, but I see two possible answers on behalf of the average fixation account: (i) According to the presupposition of the hypothesis of reality, scientific method eventually permanently fixes beliefs in future generations, so its average fixation over the course of an infinite history is the highest possible, while the average fixation of authority, though high, isn’t perfect. Hence, scientific method has an average fixation superior to that of authority. However, this appeal to convergence to support the highest average fixation for scientific method might seem to run into the same philosophical difficulty as the one posed in Appendix 2 for the eschatological fixation account: it undermines Peirce’s case for the presupposition of the hypothesis of reality. If scientific method has high average fixation because it leads to convergence, then we don’t need to worry about whether it fixes beliefs now – its current belief-fixation or lack thereof will have no effect on its overall average. Thus, we don’t need to worry about whether it prevents undermining now by the Pyrrhonian observation. All that matters is what happens in the distant future. So scientific method can fix beliefs even if we don’t presuppose the hypothesis of reality or convergence. We don’t have to presuppose reality in order for 28 For evidence that Peirce recognized underdetermination and incommensurability, see Skagestad (1981: 181) and (1995).

Justification and Consensus: The Peircean Approach

281

scientific method to be justifying on the average fixation theory, so long as it actually leads to convergent opinion. We can put the problem this way: if the hypothesis of reality is true and there really is convergence, then scientific method doesn’t need the presupposition of the hypothesis of reality in order to fix beliefs or be justifying. So there is a contradiction between the truth of the hypothesis of reality and the claim that scientific method must presuppose the hypothesis of reality if it is to justify. This undermines Peirce’s case for the presupposition. The answer to this on behalf of the average fixation theory is that the Pyrrhonian observation must be preempted in the future if not in the present. So scientific method must presuppose the hypothesis of reality in infinitely many of its uses. (ii) We can also respond to the problem of rapid turnover for scientific method by altering our idea of what is required for fixation. We can say that a belief counts as fixed, not in isolation, but as a contribution to long-run fixation. Fixation isn’t defined simply as the persistence of a belief but as either its persistence or its contribution to long-run persistence. (4) Why should an individual now use scientific method when authority currently does better in fixing individual beliefs – there is less current belief-turnover for authority than for scientific method? Answer: Peirce can appeal to his device of altruism: an individual should conceive of his/her use of methods as a contribution to the long run. (5) Isn’t there a contradiction between the hypothesis of reality and the Peircean explanation of why scientific method fixes belief? The hypothesis of reality says that scientific method determines belief by nothing human. By contrast, the explanation of fixation says that, by the laws of human psychology, belief is not undermined by a Pyrrhonian observation when people presuppose that the method fixes belief by nothing human. But then the process of fixation involves laws of human psychology. There is a contradiction between the content of the hypothesis and the content of the explanation. Since there is a contradiction here, Peirce can’t both save scientific method from the Pyrrhonian observation and establish the hypothesis of reality as a presupposition. Answer: By “determines belief by nothing human” Peirce need only mean: determines belief in such a way that a Pyrrhonian observation does not arise or undermine belief (hence, does not determine by human will). This is enough to avoid undermining by a Pyrrhonian observation. It isn’t necessary that there be no laws of human psychology operative in preventing a Pyrrhonian obser-

282

Frederick F. Schmitt

vation by these means. Yet another idea is that the laws of psychology are operative only in any finite stretch of the convergence. In the limit, psychology drops out and reality alone does the work.29

References Hookway, Christopher (1985). Peirce. London: Routledge. Hume, David (1978). A Treatise of Human Nature. Edited by L. A. Selby-Bigge, 2nd edn. Edited by P. H. Nidditch. Oxford: Oxford University Press. Kahneman, D., P. Slovic and A. Tversky, eds. (1982). Judgment Under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press. Lehrer, Keith and Carl Wagner (1986). Rational Consensus in Science and Society: A Philosophical and Mathematical Study. Dordrecht: Reidel. Levi, Isaac (1995). “Induction According to Peirce.” Ketner (1995). Ketner, Kenneth, ed. (1995). Peirce and Contemporary Thought: Philosophical Inquiries. New York: Fordham University Press. Lichtenstein, S., B. Fischhoff, and L. Phillips (1982). “Calibration of Probabilities: The State of the Art to 1980.” Kahneman, Slovic, and Tversky (1982). Nisbett, Richard and Lee Ross (1980). Human Inference: Strategies and Shortcomings of Social Judgment. Englewood Cliffs, N. J.: Prentice-Hall. Oscamp, S. (1982). “Overconfidence in Case-Study Judgments.” Kahneman, Slovic, and Tversky (1982). Peirce, Charles Sanders (1992). The Essential Peirce: Selected Philosophical Writings Volume 1 (1867-1893). Edited by Nathan Houser and Christian Kloesel. Bloomington, Indiana: Indiana University Press 1992. Piazza, Guido, ed. (1995). Esperienza e conoscenza: Introduizione all’epistemologia sociale. Milan: LCS. Pitz, G. F. “Subjective Probability Distributions for Imperfectly Known Quantities.” L. W. Gregg (1974). Schmitt, Frederick F. (1985). “Consensus, Respect, and Weighted Averaging.” Synthese 62: 25-46. — (1995). “Consensus Theories of Justification: The Publicity Argument.” Piazza (1995). Sextus Empiricus (1976). Sextus Empiricus I Loeb Classical Library. Trans. R. G. Bury. 29 Versions of this paper were read at the University of Melbourne and Indiana University and at the APA Eastern Division meetings (December, 2000), with Miriam Solomon commenting. I would like to thank audiences at these events for helpful comments. Special thanks to Miriam Solomon, to Robert Almeda, and to Tony Coady for their insightful comments. Work on the paper was supported by a Research Fellowship at the University of Melbourne (1995).

Justification and Consensus: The Peircean Approach

283

Cambridge, Mass.: Harvard University Press. Skagestad, Peter (1981). The Road of Inquiry. New York: Columbia University Press. — (1995). “Discussion: Peirce and the History of Science.” Ketner (1995). Turner, J. C. (1987). “The Analysis of Social Influence.” Turner et al. (1987). — , M. A. Hogg, P. J. Oakes, S. D. Reicher, and M. S. Wetherell, eds (1987). Rediscovering the Social Group: Self-Categorization Theory. Oxford: Basil Blackwell. Wetherell, M. S. (1987). “Social Identity and Group Polarization.” Turner et al. (1987). Ziman, John (1978). Reliable Knowledge. Cambridge: Cambridge University Press.

284

Roger F. Gibson

How I Came to Know Quine: A Reminiscence

285

On Contemporary Philosophy and Social Sciences

286

Roger F. Gibson

How I Came to Know Quine: A Reminiscence

287

How I Came to Know Quine: A Reminiscence Roger F. Gibson

Willard Van Orman Quine, the most distinguished analytic philosopher of the latter half of the 20th century, was born in a modest frame house on Nash Street in Akron, Ohio on June 25, 1908. After graduating from Akron’s West High School in January, 1926 he enrolled in Oberlin College in the fall. His studies at Oberlin centered on mathematics and mathematical logic. It was during his studies at Oberlin that he first encountered Alfred North Whitehead and Bertrand Russell’s three-volume masterpiece Principia Mathematica. PM was to have a life-long influence on Quine’s thinking about logic. Another lingering influence traceable to his Oberlin days was the behaviorism of J. B. Watson as expounded in his Psychology from the Standpoint of a Behaviorist. Quine graduated summa cum laude from Oberlin in 1930. In the fall of 1930, Quine enrolled as a graduate student in philosophy at Harvard University. He applied to Harvard because Whitehead was teaching there. At Harvard he studied with C. I. Lewis and H. M. Scheffer, and wrote a dissertation entitled The Logic of Sequences: A Generalization of ‘Principia Mathematica’ under the nominal direction of Whitehead. Quine took his Ph.D. in just two years at the age of twenty-three. He was awarded Harvard’s Sheldon Traveling Fellowship for 1933. He used the fellowship to visit Vienna (where he attended meetings of the Vienna Circle), Prague (where he met Rudolf Carnap), and Warsaw (where he met Stanisław Leśniewski, Jan Łukasiewicz, and Alfred Tarski, among other eminent logicians). Quine’s Sheldon year had a profound and lasting impact on his intellectual development. Upon his return to the United States he was awarded a three-year fellowship as a Junior Fellow in Harvard’s newly formed Society of Fellows. In 1936 Quine was appointed to the faculty of Harvard’s philosophy department. Except for a tour of duty with the U. S. Navy served in Washington, D. C. during WWII, Quine remained on the faculty of Harvard, eventually as Edgar Pierce Professor of Philosophy and Senior Fellow in he Society of Fellows, until his mandatory retirement in 1978 at the age of seventy. During his remarkably productive sixty-five year career as an academic philosopher, including thirty-some years teaching philosophy and logic at Harvard, Quine authored 23 books and countless journal articles. His books have undergone some fifty translations ranging over sixteen language. With the

288

Roger F. Gibson

knowledge of a professional geographer and an unbelievable gift for languages, Quine traveled the world, lecturing on six continents and in six languages. Along the way he garnered eighteen honorary degrees, he won a number of medals and prize, including Japan’s prestigious (and lucrative, $460K) Kyoto Prize. Numerous national and international conferences have been held on Quine’s philosophy, and books and articles written by his critics and commentators are too numerous to list. Yet, for all this, Quine remains an unknown figure to the overwhelming majority of the American public. More surprising is the fact that despite being widely acclaimed in philosophical circles for his eloquent and parsimonious prose, Quine’s philosophy is not always well understood by philosophers. So why is it that so many Americans have never heard of Willard Quine? First, the man in the street rarely reads analytic (i.e., scientific) philosophy so Quine’s brand of philosophy isn’t for everyone. In an article he wrote for Newsday titled “Has Philosophy Lost Contact with People?” Quine put the point as follows: “think of organic chemistry; I recognize its importance, but I am not curious about it, nor do I see why the layman should care about much of what concerns me in philosophy.”1 The truth is, Quine was a philosophers’ philosopher. Most of his writings are aimed at an audience of professional philosophers and logicians, and so, many of his writings have a forbiddingly technical content. Second, personality wise, Quine lacked the flamboyance of a Russell (“Better Red than dead.”) and the mysticism of a Wittgenstein (“What we cannot speak about we must pass over in silence.”), so these two major avenues into the popular mind – flamboyance and mysticism – were not open to Quine. He was a modest man who did not seek the limelight, but then neither did he shun it, even when he harbored misgivings about the context; for example, he wrote: “If instead of having been called upon to perform in the British television series ‘Men of Ideas’ I had been consulted on its feasibility, I should have expressed doubt.”2 Perhaps, but he did appear on the show and allow himself to be interviewed by the host, Bryan Magee. A more difficult question is why do so many of Quine’s philosopher readers fail to understand his philosophy. I think there is enough “blame” to go around. For example, on the one hand, many of Quine’s readers fail to see the systematic unity of his thought, a failure that has all of the ill effects of taking 1 W. V. Quine. “Has Philosophy Lost Contact with People?” in Theories and Things (Cambridge: Harvard University Press, 1981), pp. 192-193. 2 Ibid., p. 193.

How I Came to Know Quine: A Reminiscence

289

things out of context. On the other hand, sometimes Quine’s prose can be deceptively simple – enough so as to engender an unwarranted confidence in his readers’ minds that they have understood him correctly. *** I came to know Quine in the 1970s when I was a graduate student in philosophy at the University of Missouri-Columbia. In those days no course on Quine was offered at Mizzou (in spite of the fact that Quine’s magnum opus, Word and Object, had been in print since 1960). However, in 1972 a congenial faculty member in the department agreed to direct an independent studies on Word and Object for two of my fellow graduate students and myself. We read Word and Object with great enthusiasm but with little comprehension. Our erratic meetings with our director amounted to little more than the blind leading the blind. Somehow, though, I learned enough about Quine’s views to write a master’s thesis in 1973 on his conception of philosophical analysis (explication). By the fall of 1975 I was eager to find a suitable Ph.D. dissertation topic; Quine came to mind. In an impulsive moment in mid-November of 1975, I wrote to the Great Man at Harvard inquiring if it were possible for me to sitin on any courses he might be scheduled to teach in the spring of 1976. Luck was with me, for a few days later I received a letter from Quine informing me that his one course offering for the spring was to be Word and Object, and that I was welcome to sit-in, if I wished “to migrate for such slender faire.” He continued, “a more formal arrangement is possible but that would lead to a needless expense of $1,000.” I was struck by his thoughtfulness, humility, and generosity. I resolved to migrate, but it would be three long months before I would meet the Great Man face to face. *** I arrived in Boston via TWA on Saturday morning, January 31, 1976. I had never been to Boston before, nor did I know anyone who lived there. After retrieving my sole piece of luggage, a seabag dating from my days in the Marine Corps a decade earlier, I hailed a taxi that delivered me to the Holiday Inn in Sommerville. I checked in for one night’s stay. The room rate was a staggering $34 a night. Too rich for my graduate student blood. I would have to find a place I could afford, and soon. However, that would have to wait until tomorrow, for that afternoon I was anxious to see Harvard. I caught a taxi to Harvard Square.

290

Roger F. Gibson

The weather in Cambridge was overcast and bitingly cold. The low lying areas of Harvard College were sheets of ice. My first impression of the campus was deflationary. Most of the buildings on campus stood empty and unlit, awaiting the beginning of the spring semester. At the moment the campus resembled a ghost town. I was virtually alone as I traversed the icy sidewalks of a bleak Harvard Yard in search of Emerson Hall, home to the philosophy department. After locating Emerson Hall, I wandered around campus and Harvard Square before making my way back to my hotel around dusk. Later, as I lay in bed at the end of my first day in Boston I began to wonder if I hadn’t made a big mistake leaving my wife and my three-month old daughter in Missouri, giving up my teaching assistantship for the spring semester at Mizzou, coming to a city where I knew not a soul, thinking I could cut the mustard at Harvard, and believing that I could master enough of Quine’s philosophy in three months to float an ambitious dissertation to be titled “The Logical Structure of Quine’s Philosophy.” I spent a restless night. The next morning, February 1, my nocturnal woes subsided as things began to fall into place. In the “apartments for rent” section of the Sunday Boston Globe I found a one-room, ground level apartment(?) listed (and at an affordable $145 a month) in Jamaica Plain. I called the number in the ad to see if this jewel were still available. It was, and the voice on the other end of the line gave me clear instructions on how best to get from here to there. I checked out of the Holiday Inn and an hour later I met the owner of the voice on the phone. The apartment was on S. Huntington Avenue; a jewel it was not. Inexplicably, a mountain of uncollected trash had accumulated behind a number of apartment buildings in the area. Nevertheless, the price was right so I rented it for February, March and April. The apartment was furnished with a cot, a kitchen table and chair, a compact refrigerator, and a small stove having two electric burners. I couldn’t help noticing that the exterior of the stove behind the burners had melted. Puzzling; but not for long; the explanation was elementary: the apartment had baseboard heat along just one wall. The landlord kept the room just warm enough to prevent the water pipes from freezing during the winter. It didn’t take winter tenants long to realize, however, that the two burners on the stove, if left on “high” for hours, helped to take the chill off the room. The stove registered years of such abuse. The truth is that the three months I lived there were long, lonely, dreary, and cold. However, on the afternoon of February 1, fresh from the high rent Holiday Inn, I was thrilled and relieved to have a place to stow my seabag.

How I Came to Know Quine: A Reminiscence

291

Late that afternoon I took a trolley from “home” to the Park Street Station where I caught the Red Line subway which zoomed to Harvard Square – a practice run in preparation for the real thing once the semester began. On this my second visit to campus I noted that the largest of Harvard’s thirty-some libraries, the Widener Library (named for a victim of the Titanic’s tragic voyage) was roughly fifty paces from Emerson Hall. The authorities had granted me library privileges, and I spent many memorable days reading and writing at one of the huge tables in Widener. *** On Monday morning, February 2, I made my way to campus again, this time with the purpose of introducing myself to Quine. His office was on the second floor of Emerson Hall. I ascended the two flights of stairs as if in slow motion. I was about to come face to face with the Great Man. As I centered his closed door I drew a deep breath before knocking; maybe he’s not in. A voice responded, “Come in.” When I opened the door I found Quine standing before his desk, halfway to the door. He was a good six feet tall, trim, and sporting a marvelous tan he had acquired while vacationing in Mexico during the semester break. He was wearing a coat and tie. At the time he was 67 and I was days away from being 32. As we shook hands I mumbled something stupid like, “Professor Quine, I presume.” He had a firm grip and large hands. I explained who I was and that I had come from Columbia to sit-in on Word and Object. He heard ‘Columbia’ and naturally thought I was from Columbia University. It took an awkward moment or two to clear up this “minor” confusion. Obviously, he didn’t remember our exchanging letters in November of 1975. He volunteered some remarks regarding his recent visit to Mexico – which I would learn later was one of his favorite places to vacation. Getting down to business, he told me that Word and Object would meet on Tuesdays and Thursdays from 10:30 to noon in a classroom in Emerson Hall once the new semester began in a week or so. I thanked him for his time and told him how much I was looking forward to Word and Object; then I left. The meeting that I had anticipated for nearly three months was over in less than ten minutes. Still, it was a success, for I had made contact with the Great Man and he didn’t bite my head off. In fact he was downright friendly. I felt a great wave of relief come over me as I bounded down the same two flights of stairs that minutes before had been so difficult to ascend. I emerged from Emerson Hall into the bright sunshine of Harvard Yard. Of course, it never crossed my mind that my ten-minute meeting with

292

Roger F. Gibson

Quine marked the beginning of a twenty-four year friendship that would change my life. The first class meeting of Word and Object took place in Emerson Hall in early February. The room’s thirty-some chairs were occupied, and there were chairless students lining three walls. By the end of the semester there were less than twenty students who finished he course. However, on the first day of class many of the students who were present had no intention of enrolling in the course. Rather, they had come to get a glimpse of the Great Man. After class several students asked Quine to autograph their newly purchased, pristine, and probably never to be read, copies of Word and Object. Quine cheerfully and without fanfare complied. A further testament to Quine’s celebrity status on campus was the abundance of Quine graffiti etched on various surfaces in the men’s room in the basement of Emerson Hall. Quine taught Word and Object purely as a lecture class. Remarkably, as king questions in class and engaging in discussion were not permitted. Any questions students might have were to be written and handed to Quine at the end of class. Quine would respond to them at the beginning of the next class. There were few questions. Making matters worse, Quine was not a gifted lecturer. He relied on a stack of five-by-seven note cards covering virtually everything in Word and Object and Roots of Reference. Occasionally he would break free of the cards, and when he did his lecturing improved greatly. I recall one class meeting in April when Quine lectured without notes for the last fifteen minutes of class. He so mesmerized the students that when the class was over no one stirred until Quine had left the room. My conjecture is that his Draconian teaching method derived from the fact that he had so much material to cover and so little time to do so. However, there may have been a second consideration, for according to a rumor going around at the time, he once offered the course as a seminar and the students weren’t up to it. *** On Wednesday, March 10, I had my first hour-long meeting with Quine in his office to discuss the feasibility of my dissertation topic – the logical structure of his systematic philosophy. I ask if it were okay for me to record our discussion; he said it was. I was sitting on a couch to the left of Quine’s desk. I placed my tape recorder on his desk. But when I tried to plug in the microphone I found that my hand was shaking so badly that I had to steady it by resting my wrist on the edge of Quine’s desk. From his position seated behind his desk,

How I Came to Know Quine: A Reminiscence

293

he surely noticed this symptom of my nervousness, but he said nothing. The meeting began with my requesting that he not hold back any criticisms he might have regarding my proposed dissertation. I then read him a two-page précis of the dissertation. He liked what he heard for the most part, but he thought that I needed to say more about his behaviorism. One thing led to another and soon my hour was up. As I was preparing to leave, Quine gave me two of his recent (1975) articles I hadn’t known about: “The Nature of Natural Knowledge” and “Mind and Verbal Dispositions.” I was pleased to have them but even more pleased by Quine’s positive response to my project. I had a viable dissertation topic! Now all I had to do was to write it. I celebrated this milestone with a cup of coffee at a café in Harvard Square. Later in March we met again in his office to discuss my progress. After an hour of discussion I began to gather my things to leave; Quine said, “Sit down. You have come a long way.” We discussed for another hour. Quine was a generous person. My writing progressed apace. I worked all day everyday. I copied in longhand every word of Word and Object. And I read and re-read Roots of Reference. I allowed my self no diversions. Getting Quine right became an obsession. By the time my days at Harvard came to a close in April, I had written all but the last chapter of my dissertation. That final chapter was eventually completed, and I took my Ph.D. at Mizzou in 1977. That same chapter was my first publication; it appeared in Erkenntnis. Quine was one of the referees of that paper for Erkenntnis. In 1982 a revised and expanded version of my dissertation was published under the title The Philosophy of W. V. Quine: An Expository Essay, with a foreword by Quine. *** The Boston/Cambridge area was Quine’s home base for seven decades, from 1930 to 2000, but I always thought of him as a Midwesterner. Certainly, many of his character traits were instilled in his youth in Ohio. I have already noted his thoughtfulness, humility, and generosity, but he was also honest, frugal, industrious, sociable, and conservative. As a rule, Quine was patient and polite regarding his critics, but now and again someone would get his goat and suffer his verbal wrath. For example, in response to an irksome remark of a critic, Quine wrote: so-and-so “predicts that I will pretend not to understand what he means by his ‘assertions about the spiral of understanding as corresponding to the walls of our cosmos.’ I am tempted, perversely, to pretend that I do

294

Roger F. Gibson

understand. But let us be fair, if he claimed not to understand me, I would not for a moment suspect him of pretending.”3 Needless to add, Quine had a lively sense of humor. Harvard had a policy (now illegal) of mandatory retirement at age seventy. Quine retired in 1978, just two years after I sat-in on Word and Object. Retirement for Quine meant retirement from teaching; he still had twenty years of writing and lecturing before him. Indeed, he published five books and over eighty articles during his retirement years. During those same years I was privileged to be in his presence at the University of Missouri, the University of Nebraska, Washington University, Boston, San Marino, Spain, Sweden, Norway, and the Czech Republic. On one of his visits to Washington University, where at the time I was a Visiting Assistant Professor who had applied for a tenure position there to begin the following year, Quine weighed in on my behalf and I got the job. Two years later, 1988, Washington University hosted a five-day international conference on Quine’s philosophy. Nearly 300 people attended the conference. In 1990 another fiveday international conference on Quine was held in San Marino. Also in 1990 Quine gave ten lectures in Girona, Spain. In Stockholm in 1993, the King of Sweden presented Quine with the first Rolf Schock Prize, followed by a twoday conference on Quine in Oslo. In 1995 there was a four-day conference on Quine in the Czech Republic. This list is by no means a complete list of Quine’s activities during his retirement years; rather, it is limited to those of his activities that I witnessed. *** During the last two years of his life Quine suffered from memory loss. My close friend Ernie Lepore (of Rutgers University) knew of this and suggested that we visit Quine on the occasion of his 91st birthday, June 25, 1999. On June 24 I flew from St. Louis to Boston and, just as I had done twenty-three years earlier, I took a taxi from Logan International to the Sommerville Holiday Inn. The 1976 room rate of $34 had inflated to $199. Ernie drove from New Jersey to the Holiday Inn. He had never seen Harvard so we took a taxi from the Holiday Inn to Harvard Square. Ernie’s first impression of the campus couldn’t have been more different than my first impression twenty-three years earlier. June 24, 1999 was a beautiful day in 3 W. V. Quine. “Reply to Henryk Skolimowski,” in The Philosophy of W. V. Quine, edited by Lewis E. Hahn (LaSalle, IL: Open Court, 1986), p. 493.

How I Came to Know Quine: A Reminiscence

295

Cambridge. The flowers and flowering trees were in full boom, the temperature was in the 70s, and there were students everywhere. The campus was alive and well. The next morning, Quine’s 91st birthday, Ernie and I drove to Quine’s home, 38 Chestnut Street. Sadi, Quine’s wonderful live-in housekeeper, was expecting us and had told Quine we were coming. However, when we arrived it wasn’t clear that Quine knew who we were, though he seemed to recognize my voice. Ernie soon had Quine buckled up in the front seat of Ernie’s Toyota with me in the back seat. We were heading north on a scenic route to picturesque Marblehead. Quine seemed delighted to be out and about, even if it were with two “strangers.” We arrived at our destination around noon. Ernie parked the car and set out on foot to find a restaurant. He soon returned with good news: there was a restaurant just a few blocks away. We drove to the restaurant where we enjoyed a splendid lunch. Quine seemed to be enjoying himself, but Sadi had cautioned us that “The Professor” needs several rest periods daily. So after lunch we headed back to 38 Chestnut Street, arriving in plenty of time for Quine to catch a nap before we were to dine with Quine’s closest philosophical ally, Burt Dreben, and his lovely philosopher wife, Juliet Floyd, at their flat near Boston University. We arrived at Burt and Juliet’s at around 6:15, fifteen minutes early. Juliet served us drinks while she continued working in the kitchen. Burt was in his bathrobe when we arrived but when he joined us fifteen minutes later he had on a three-piece suit and tie. Seeing Burt was bitter sweet, for he had cancer, and he knew that he had only months to live. We had a wonderful evening, but sadly it was to be the last time I saw Burt. In 1990 Burt had confided to me that he was going to have a difficult time when Quine died, after all he had been by Quine’s side at Harvard since the 1940s. However, in a twist of fate befitting Somerset Maugham, it was Burt’s death that was to pose a difficult time for Quine. We left the dinner party around 11:00. Ernie dropped Quine and me off at the front door of 38 Chestnut Street. I spent the night at Quine’s, but Ernie couldn’t find a parking space and so he returned to the Holiday Inn. The next morning Quine called upstairs, “Roger, breakfast.” Breakfast was light, however, for we were going to the Harvard Faculty Club for Sunday brunch later. Once Ernie arrived we three piled into his Toyota and headed to the Faculty Club. Following Quine’s verbal instructions, Ernie had no difficulty getting us

296

Roger F. Gibson

to campus and locating the Faculty Club. Quine and I waited while Ernie parked the car. Since we made our reservation in Quine’s name, the maitre d’ had reserved for us “The Professor’s favorite table.” He may have been retired for twenty-one years, but he was still a celebrity on campus. I half-wondered if there might still be Quine graffiti in the men’s room in Emerson Hall! After a fine brunch, Ernie dropped Quine and me off at 38 Chestnut Street. Ernie was meeting some friends at M.I.T. Quine showed me his office on the third floor. This is where he did most of his work, not in his office on campus. There were a number of daggers hanging on one wall of his third-floor office, but when I asked about them he couldn’t remember where they had come from. Nor at this point in his life could he recall the titles of his books and articles. Quine retired to his room for a nap. Soon I would have to leave for the airport to catch my flight to St. Louis. Much to my regret Quine was still resting when I left so I didn’t get to say good-bye. That gnawed on me later, for I didn’t know if I would ever see him again. *** However, I did see him one more time; the occasion was his 92nd birthday. I flew into Boston; Ernie came by train. This time we stayed at the Parker House. By an uncanny coincidence, the next morning June 25, 2000 was both Quine’s birthday and the unveiling of Burt Dreben’s headstone. Ernie and I rented a car and drove to Burt’s ceremony. Of course Juliet, Burt’s widow, was there. Doug Quine, Van’s son, informed us a week earlier that his dad was hospitalized, but that there was a slim chance that he might be home in time for his birthday. However, we then learned from Sadi that Quine would not be released that soon, but he could have visitors at the hospital. On the afternoon of his birthday I called Juliet to see if she wanted to accompany Ernie and me to visit Quine in the hospital. She did, so the three of us went together. We found Quine sitting up in bed, taking a little oxygen. He was very alert, enough so as to correct Juliet’s pronunciation of the names of some of the places in Scandinavia she told him she had visited recently, and enough so as to talk for ten minutes on Ernie’s cell phone with Donald Davidson in Berkeley about Davidson’s upcoming trip to Turkey. Sadi had told us that one of Quine’s daughters, Norma, had visited him on the morning of his birthday. However, when I asked Quine if Norma visited him this morning, he said, “No, and she should have; its my birthday.” Well, geography is one thing, visi-

How I Came to Know Quine: A Reminiscence

297

tors are another. On Quine’s mental map visitors came and visitors went, but the continents remained conveniently fixed. His passion for geography and for maps truly was life-long. We stayed about an hour. Juliet gave Quine a kiss on the cheek and said good-bye; Ernie did the same. That left me alone face to face with the Great Man. I shook his hand and thanked him for all he had done for me. As I turned to join Juliet and Ernie in the hallway, Quine said in a loud clear voice, “Roger, thank you for coming.” I thought, “and thank you, Van, for allowing me ‘to migrate for such slender faire’ twenty-three years ago.” Van Quine, teacher and friend, died on Christmas Day, 2000 – exactly six months after his 92nd birthday. When he died there was a spate of Quine obituaries published around the globe. Many of them followed the lead of The New York Times by quoting the following sentence from my 1982 book: “Quine’s philosophy is a systematic attempt to answer, from a uniquely empiricistic point of view, what he takes to be the central question of epistemology, namely, ‘How do we acquire our theory of the world?’.” When that sentence was first penned (probably at one of the large tables in Widener in 1976) it expressed something fairly new, but today it comes close to being the received view. Quine nudged things in that direction when he wrote in the foreword of the book, “In reading Gibson I gain a welcome perspective on my own work.” Did I mention that Quine was a generous person? To have known him is an honor and a privilege; I shall miss him.

298

Hans Lenk, Matthias Maring

Responsibility and Globalization Hans Lenk, Matthias Maring

I. Globalization: Ideology or Fact? Everybody talks about “globalization” these days, but there are very different interpretations of that term. Basically, it seems to be a concept covering the economic internationalization of the markets and of the worldwide interconnected economic activity of powerful economic agents – mainly socalled multinational corporations (or for short “multis”) which seem to evade some political restrictions in terms of tax paying, legal conditions of hiring for jobs, participating in and contributing to conditions set by national trade unions and/or social welfare programs. Globalization, first of all seems to be an economic concept. The question is whether or not it has something to do with philosophy and ethics at all. Philosophy has an analytic and also an ethical task to deal with. Both are relevant in the case of the concept of globalization. The first seems to come under the topic of rhetoric and ideology, arguments of social excuses and interested strategies usually adapted by the aforementioned multis in a local or national setting and politicians as well as supporters of pure deregulation and neoliberal market economists, stock analysts etc. in particular. It is well known that there is a new rhetoric of globalization of markets, competition for jobs, production and world wide deliverance as well as supply of goods and services, of information and particularly of the economy pertaining to the markets and international finances, e. g. shareholder values. International capital flows as well as power centers in terms of economic concentration of the multis etc. seem to emphasize a trend in some kind of revitalization of pure liberal market economics or even international capitalism with some intended as well as some unintended political and social secondary consequences interfering with social welfare and retirement systems, social security as well as national and local politics of economics, infrastructure, organized trade union systems as well as the tax revenue systems of the respective national and federal states, including in particular problems of justice with respect to distributing and collecting taxes from corporate (especially international) agents. The German sociologist Ulrich Beck thinks that this international syndrom amounts to a new ideology which gains some kind of “seman-

Responsibility and Globalization

299

tic hegemony, a publicly incited ideology of globalization” being a “decisive source of power augmenting strategic potential on the side of the corporations” (Frankfurter Rundschau 9.1.1997). He deems this a kind of “side effect of industrial modernization of society” calling for study and endeavors of taking care of the secondary consequences of the side effects – so to speak the indirect metaconsequences for society, solidarity, justice and – let us include – practical humanity in different parts of the world. Many representatives of big business and the respective employers’ pressure groups use the argument of extant necessities of globalization for demanding dramatic changes with regard to wages (usually allegedly “too high”), social and welfare services (dito), daily working times and shift schedules and – in particular in a social welfare state like Germany – for rescuing the socalled (economic) “site Germany” (“Standort Deutschland”). The more or less direct consequences of globalization in economic terms would be a danger of losing the export status of the nation, particularly dependent on an outstanding in international export. The imbalance of respective burdens and profits of the corporations on a national scale and particularly of the subdivisions of international corporations would allegedly lead to a certain kind of economic drain in the sense that enterprises would look for cheap job countries to invest and produce there and to produce a special internal high figure of unemployment with all the conceivable respective social reverberations. Also, international competition in terms of prices and production costs would lessen or even damage the national capability of exports etc. Economist Dolata (Frankfurter Rundschau 30.1.1997) thinks that the argument of globalization in national economies (again he thinks of the German case) would be used to justify wage limitation, destabilization of social welfare and a kind of dumping of social security of workers in general in order to minimize costs of their economic output in the international competition. But he emphasizes that this would not be true in the first place because the globalization or rather internationalization of German economy does indeed particularly concentrate on European neighbour countries. Therefore, export is “shaped by a pertinent Europeization rather than by progressive globalization”. Industrial production in foreign countries is naturally organized in high wage countries of Western Europe and North America serving mainly for opening of new markets, maintaining and extending of economic market positions there etc. and does not yet have a remarkable impact on production and jobs at home. There seems to be only a narrow sector of German economy which can be called “globalized” in the true sense, e. g. chemical industry.

300

Hans Lenk, Matthias Maring

Otherwise, it seems to be much more a European internal market problem calling for a European social union with a certain kind of obligatory social charta instead of as yet calling for a truly planetary and international globalization. (Indeed the export figures seem to substantiate or justify these assessments.) Another author (Reuters ibd. 3.12.1996) emphasizes that the true causes of massive unemployment are not to be seen in the deficient export, but seem to be a consequence of the fact that individual jobs and workforce seems to be required less and less for generating an ever climbing economic social product. (The GNP today is at least (and really was already in 1995 in West Germany) three times as high as 1960 and required only 80 percent of the respective working hours. The main cause of joblessness and unemployment would be “the fact that one did not take into consideration adequately the decrease of the (necessary) amount of work. It would have been more meaningful, consequently to foster” the social conditions for decreasing working hours: instead of globalization the problem would be rather “the belief that it – the globalization itself, H. L., M. M. – would be the cause of joblessness”. All this amounts to an allegedly wrong politics, a certain kind of downward spiral only augmenting the problems instead of solving or mitigating them. “The ghost of globalization” just would divert or deflect the focus from the basic problems like concentration of economic power of firms, income and capital. A Christian philosopher of economics, Friedhelm Hengsbach, thinks that “the universal market fever” tends to concentrate just on “Olympic champions” in economics rebuilding society “only according to the rules for winners.” This would “eventually lead to a spiral of structural irresponsibility” in which corporations shift the burdens by firing elderly employees thereby passing the cost and buck to the community of solidarity of the social security system. The federal finance minister would do the same with respect to early retirement of state clerks, teachers etc. This would lead to a social trap situation which can not be solved by individual groups alone but only by a new form of cooperation – as for instance with the paragon example taken by Volkswagen. This firm introduced a new plan of cooperation and profit sharing for employees including the four working-day week as well as social security. Hengsbach deems “the economically speaking reasonable (das wirtschaftlich Vernünftige) would possibly be reconcilable with the morally required (dem moralisch Gebotenen)” (Zeit, 12.1.1996). We, however, think that this would be too optimistic a stance as regards the social traps which indeed seem to shift off the requirement of justification to

Responsibility and Globalization

301

arguments about international competition at the cost of national and social programs as well as wage structures. This seems to be true also on a regional basis whenever cities and communities have to put on strict saving strategies whereas the federal state or the states impose new responsibilities (e. g. the social right to a kindergarden acceptance for every child and the exploding climb of public social help and health systems) amounting to a real “in vain trap” (Kuban, Duisburg) considering the “cost instigating blurring of responsibilities” (kostentreibende Verantwortungsverwischung) (Frankfurter Rundschau 21.9.1998). It is true, social thinking should not give up hope, but indeed the auspices are rather black or at least grey. And there seems to be no means available effectively to counter the spirals and economic as well as social trap dynamics. For instance, there certainly is a tax-dumping amounting to an erosion of the tax sovereignty of the national states – not only in the European Union. Only global answers seem to be able to remedy this problem. Many multinational corporations shift their tax paying obligations to countries and regions of minimal tax burdens: for instance a big electrical multinational of Germany did not pay any profit tax in Germany in 1995; the biggest automobile corporation (prior to its fusion with Chrysler) did apparently not pay any corporation taxes since 1995 by “intelligently” using tricky manipulations and nationalization of economic losses and buying up other firms as well as transferring profits to foreign branches in low-tax countries etc. Whereas corporation taxes decrease (they are but pretty high in Germany still), wage taxes tended to rise in the last decades so that somebody (Münster ibd. 12.8.1998) talked about a “wage-tax state” because the wage taxes are reaching over one third of the whole wage income of the state whereas profit and capital income tax amounts to less than 15 percent (1998). There seems on the one hand to be a necessity for an international harmonization of tax developments particularly for corporations and multis in order to avoid the trap situations. But on the other hand it is even an advantage for competing with other states to have big firms invest in one’s own country. So, particularly some little or even a few developing countries figuring as a kind of “tax oases” are certainly not interested in having a global or even continental harmonization of tax arrangements on an international scale. The same is true for international fusions of multicorporations with many side effects or direct consequences for employees, jobs, consumers, states and governments (in terms of tax income) etc. The trap situation is certainly increasing with the international dynamics and the power of transnational corporations which is still rising all the time: indeed, in 1998 they contrib-

302

Hans Lenk, Matthias Maring

uted already a third of the world’s Gross National Product. For instance, export business of the US and UK almost exclusively is in the hands of these multis already (Hamm 1998). What may a philosopher say with regards to the terminology of globalization under these aspects? 1. First of all, there is a terminological consideration: economic globalization should not be confounded with or mistaken for societal globalization in terms of, say, information exchange, cultural exchange and mixtures (multicultural societies) etc. 2. Although most of the economic exchanges of the export states in Europe are relating to the European Union and North America this certainly is not the whole story. The same is not true for East and South Asian states, we think: they are basically even oriented much more in a world-wide export strategy than the Europeans. Criticizing economic globalization in that respect is in a way a rather European perspective. Indeed, there is an extension of global markets in every field which can not possibly be denied with respect to the future. 3. With respect to the polarization between economic globalization on the one hand and structural unemployment on the other, we think, that even though everything said by the quotations at the beginning would turn out to be true, the rather interested argumentations on both sides of the socalled traditional social divide between employers and employees (which, by the way, is being blurred already – in a few cases – nowadays by shareholdings and stockholdings in the hands of employees) is too dichotomic: interested and ideologically argumenting representatives on both sides look for arguments of the type “not that is the cause but on the contrary that other factor” – which does not indeed really live up to scientific differentiation. For instance, in the case of globalization and unemployment both factors (economic globalization and competition on an international scale as well as a decreased requirement for workforce) seem to be influential. One has to dispense with one-sided thinking and using the term of “globalization” as an ideological strategy – at least from a scientific and social-philosophic point of view. (Certainly the interested sides will use these arguments in public debates also in the future.) 4. Generally speaking, one-factor theories and dichotomic arguments and one-sided perspectives seem to be almost always wrong – particularly in complex sociopolitical and socioeconomic settings on a planetary scale. And if everybody just points to the other partner on the other side to gain an advantage in economic or political or social terms then a certain kind of downward spiral

Responsibility and Globalization

303

of shuffling off of alleged responsibilities or social “burdens” as well as costshifting would result. Methodologically speaking, these strategies and dynamics would mostly be including circular arguments and even naturalistic fallacies. There are some international groups like the group of Lisbon which try to propose solutions for the outlined globalization problems. For instance Riccardo Petrella, president of the mentioned group of Lisbon, proposes “a contract for another globalization” (1998), he points to the fact that the real basic problem calling for efficient measures would be the fulfillment of basic needs of eight millions (English), billions (American) of people in the world within the next 25 years with respect to water, food, energy, health, lodging, education, information, transport, communication, personal and political as well as social participation in arts and also in the shaping of commons and the communities. He thinks that first of all some traditional prejudices have to be changed: At first the priorities of economical proposals like the privatization of commons and societal as well as social services like the mentioned ones. Judging only on an economic basis would deprive our societies of all principles of social citizenship, solidarity, equality and “brotherhood” – as well as – let us add – practical humanity. Secondly, the influence of capital markets should be restricted by political controls of central banks, international parliaments etc., e. g. by a 0.5 percent Tobin tax on worldwide financial transactions of stock, elimination of tax oases, bank controls and abolition of bankers’ discretion, establishment of a world council for economic and financial security and measures fostering the wellbeing of the total planetary population. Thirdly, the role played by science and technology should be changed with respect to supporting and fostering the common weal and not only the competitive interest of corporations. Fourthly, a reorganization and redistribution of wealth by redistributing profits between capital and employees in the interest of the citizens etc. Many of these proposals seem to be utopian – particularly because of economic constraints and political power interests as well as international competition between states, corporations etc. Indeed, these proposals have to be worked out further – in particular with respect to a potential operationalization and institutionalization in practice. Also, we fifthly would like to add, the environmental perspective has to be emphasized much more than by the club of Lisbon. Indeed, as everywhere in complex societal areas and particularly in political and interareal interconnections the devil is so to speak sitting in the detail, as a

304

Hans Lenk, Matthias Maring

proverb reads. Economically speaking, a way towards a world wide “economic internal market” (can we say world wide “domestic market”?) is on the way by GATT/ WTO treaties – despite all the extant deficiencies and backlashes – particularly with respect to environmental and regional poverty problems. However, as Vallender wrote (1998, 576f) the “globalization of economic freedom” has to lead towards “the globalization of responsibility” with regard to the “protection of the Global Commons”. He proposes a new “world sustainability organization” (WSO) and a world environmental politics strictly oriented at the principle of subsidiarity. However, this organization WSO should be a rather influential and powerful organization. This is all the more true regarding the as yet thwarted international endeavors in terms of environmental global treaties like those of Rio de Janeiro, Kyoto and recently The Hague. To be sure, environmental initiatives are gaining weight – in particular in highly industrialized countries like continental Europe. Basically, the motto “think global, act local, but remember you do not act alone, there are to be considered synergetic consequences” is getting already some kind of economic impact – at least regionally speaking. Local initiatives like in our place the socalled “Environment Charta Agenda 21” (Umwelt Charta Agenda 21) try to foster and promote sustainable consumption, fair products and prices, with respect to the minimizing of environmental burdens on resources, packaging and transport as well as consumer information or transparency, public transportation, working and job conditions etc. This is locally an influential measure multiplied by positive resonance in the local and regional press. However, on a planetary perspective these initiatives taken by concerned citizen groups are rather touching and kind endeavors regarding the international economic, political and social trap situations and the mentioned spirals of economic power concentration, globalization of poverty and the ever growing clash between the North and the South or highly industrialized and socalled developing or underdeveloped countries – notably in Southern continents. Indeed a globalization of responsibility has to follow suit. There is a certain responsibility of rich countries for the losers of globalization processes – particularly in social – and moral – terms. There is a certain necessity to not only to raise consciousness regarding the internationalization and challenges of responsibility with respect to social justice, sustainability, practical humanity and the fulfillment of basic needs everywhere. Even the World Bank summarizing a questionnaire of 60000 affected persons in 60 countries highlights the

Responsibility and Globalization

305

three basic needs as a strategic ‘kernel or core of initiatives’: “equality of opportunity, participation and empowerment, security”. (What about environmental sustainability?) Vice president Stern underlines: “If we want that the poor would profit as quick as possible from growth, we have to directly work towards an extension of their empowerment (and competences) and the amelioration of their security situation” (September 2000). To practically materialize some of these goals would require definite measures on the national, regional and international levels including growing efforts for international remissions of debts, more efficient development aid and easier access to the markets of the industrialized countries. However, a liberalization of export and import should be taken up rather cautiously in order not to turn out to be detrimental or counterproductive to the poor ones. The developing countries should themselves fight ethnic, social and sexual inequalities, concentrate on infrastructural investments to poorer regions, decentralize administration and open up the public institutions making them more transparent. Any coherent and efficient program of developing countries fighting poverty should be thoroughly supported from the world economy outside. The most important agents however would be the poor ones themselves – thus spoke the World Bank indeed. What about environmental problems? Indeed, these problems are getting a rather prominent, if not highest priority impact – even with backlashes to the developed regions of the world. The US World Watch Institute asked for a comprehensive tax reform for rescuing the earth. However, the latest experiences with international treaties for and in environmental protection seem to be rather pessimistic as we saw. Usually, the value of nature and the “services” of nature for humans are not taken into account in any economic report of balances whatsoever. To be sure, there are some economists and biologists who tried to assess the services of nature for mankind in economic terms amounting to 28 (American) trillion, (English) billion U.S. Dollars – almost double as much as the whole Gross Social Product of all economies of the world (Constanza 1997). It is a question though whether or not economic figures can be apt to cover ecological impacts and, mostly, damages in adequate terms. However, at least one would get by that a certain guideline of the general vast importance of the respective influences – which might make some impression or sense in a basically economically thinking society. Generally speaking we would like to forward the thesis that the dynamics of globalization – as yet rather outstanding in ideological argumentation – should

306

Hans Lenk, Matthias Maring

be modified and differentiated as regards empirical evidence (most international markets are not really global yet – maybe with the exception of the Far and Southern East!), but we think that now and in the future there will occur special cases of secondary social traps. Alleged disadvantages regarding the economic situation of national production sites and exports render a certain kind of trap situation with respect to a quite biased reception and action of participating agents (notable corporate ones and in particular the socalled Global Players). First of all, the talk of “globalization” seems to be an ideological instrument leading towards a pressure on the social welfare state, on employees, unemployed people and generally towards a certain kind of pressure on lower social strata. The use of such rather ideological simplifications and arguments seems to imply or even incur examples of self-fulfilling prophecies (Merton 1976) or self-destroying prophecies, respectively. To be sure, there seem to be primary social traps with respect to economic situations of environmental protection programs, some middle sized or small corporations as well as for many affected people, including employees of lower social strata. However, the ideological usage of the globalization talk seems to lead to rather secondary social traps of shuffling off responsibility to other alleged bearers of responsibility. This trap of globalization conceptualization combines with the usual dogma of economic supply policies (profits of today are investments and jobs for tomorrow and after tomorrow) together with the ideology of automatic “self-healing” dynamics of markets in traditional liberal market economy, dominating in practical economics and large parts of politics. Domestic demand is usually not considered at all. And by the end of the millennium growing profits admittedly have not yet lead to more investments and more jobs. Generally speaking globalization traps are but special cases – though rather globally comprehensive ones – of social traps and social trap situations. They can certainly not be dealt with or sought by individualistic measures of one (corporate or state) agent alone but only by institutional, structural and legal approaches on a world wide and planetary scale. An essential question to avoid social traps including globalization traps is how is it possible to avoid or at least minimize defections by corporate agents, not only via universally supported sanctions, but preferably by economic and social incentives? All these dilemma situations are but examples of prisoners’, assurance etc. dilemmas problems well known in economic theorizing but difficult to deal with in practice. There is also a further social psychological problem with social traps consisting in the fact that a single defection would disturb trust: and it is very easy to

Responsibility and Globalization

307

destroy trust much easier than to build up trust. Again, building up trust and an attitude towards responsibility and readiness to bear it is a very difficult and lengthy endeavor (like the the “boring of thick boards” the famous sociologist Max Weber alluded to with respect to politics). This is certainly true also for respective social trap complexes of the socalled globalization trap (“Globalisierungsfalle” is the title of a book by H.-M. Martin and H. Schumann (1996)). Indeed, all general theses regarding processes of globalization and ideology of globalization arguments lead to the postulate that the respective dynamics of globalization in economics should lead to an emphasis of the globalization of responsibility with respect to protection of the global commons! (Vallender 1998). The same is true for the respective argumentations. This is certainly a task for philosophical analysis and research: How can we in the face of primary and secondary globalization processes attribute collective, corporate and personal responsibilities in the vast dynamic net of ever interacting and more and more globally acting corporate and institutional agents? (This is but a small, however, important analysis which should be addressed by philosophers, notably social philosophers and analytic philosophers – for firstly, but not only on a descriptive-conceptual level.) Certainly, we can not deal with all the studies and approaches the both of us have worked on with respect to the conceptualisation and differentiation as well as practical overlap with respect to different types of responsibilities we have worked on for the two last decades and publicized widely. However short, rather abstract, summary with respect to collective, corporate and the differentiation as well as distribution problems of responsibility may be outlined here. Even a list of priority rules for the cases of responsibility conflicts practically found everywhere may be added for thinking over and developing normative evaluations with respect to social practice on the local and the global level. Though these analytical differentiations seem to be rather abstract and academic, they should provide a basis for the perception of complex and conflicting perspectives in practice enabling the practitioner to see clearer through the mist of rather complex globalization and responsibility phenomena, talks, traps and problems! As a provisional thesis regarding responsibility problems in the face of extant and future globalization traps – including the ideological and rhetorical effects connected therewith – the following insight should be emphasized: Located at centre stage of the problem of collective, cooperative, and corporative responsibility lies the question of the operationally effective and analytically clear distribution of normative and descriptive responsibility – according to a theory of action – and the problem of the possibility of (equivalent) indi-

308

Hans Lenk, Matthias Maring

vidualistic reduction of or the secondary handling and coping with the collective responsibility of agents – whether individual or corporate.

II. Collective Responsibility and Distribution Problems What about collective responsibility as such? We cannot summarize here the elaborate typology of forms of responsibility distinguished and analyzed elsewhere (cf., e. g. Lenk 1991 in Durbin (1991) and just depicted by diagrams in the appendix, below.) Instead, let us just relate few nice anecdotes regarding different sorts of responsibilities and the relevant questions of collective bearing and distribution. Let us begin with a nice story by somebody called anonymous who could have been anybody, but it is certainly not a story by nobody, since nobody is a no-body: This is a story about four people: Everybody, Somebody, Anybody, and Nobody. There was an important job to be done and Everybody was asked to do it. Everybody was sure Somebody would do it. Anybody could have done it, but Nobody did it. Somebody got angry about that because it was Everybody’s job. Everybody thought Anybody could do it, but Nobody realized that Everybody wouldn’t do it. It ended up that way that Everybody blamed Somebody when actually Nobody asked Anybody. (quoted from nowhere in: Aus Politik und Zeitgeschichte, Beilagen zu Das Parlament (1996), B 7/96, 3) Our additional version goes like this: Somebody deemed Anybody responsible – instead of him- or herself. But since everybody is anybody and vice versa, Anybody shoved the mentioned responsibility off to Everybody. Everybody in turn argued that (s)he won’t be perfect, and perfect indeed is nobody. Therefore nobody is the culprit. Nobody is the real culprit. This is the current state of responsibility problems in complex social systems. Everybody passes the buck: Somebody to Anybody, that one in turn to Everybody; again, Everybody passes the buck to Nobody. And that’s where the buck stays, rotates, vacillates or whatever. Responsibilism vacillates between somebody-“elsism”, shove-offism (for short, “shovism” instead of chauvinism) and intraceability as well as, consequently, nobodyism. To repeat:

Responsibility and Globalization

309

Nobody is the real culprit. Nobody is responsible (as, e. g., allegedly in the Bhopal accident!). What can wo do in that situation? In order to handle the problems of Anybody’s and Everybody’s responsibilities we need to understand and treat, hence analyse, joint responsibilities. Is there a joint or participatory responsibility? Or is this leading nowhere? Ambrose Bierce held in his famous The Devil’s Dictionary as a definition of “RESPONSIBILITY: A detachable burden easily shifted to the shoulders of God, Fate, Fortune, Luck or one’s neighbor. In the days of astrology it was customary to unload it upon a star.” (The latter word now has a double meaning probably both applying.) But what really is that beast called “joint responsibility”? How to dissed bisect and/or reassemble it? We can understand human beings as the normative and responsible beings which means that they are morally distinguished from other creatures by the capacity to bear, acknowledge, consciously identify and accept responsibilities for the outcome of their individual and group or collective actions and role fulfilments and, beyond that, for the lives, limits and well-being of other humans and creatures as well as – increasingly – for the state and development of ecosystems. Humans are so to speak the moral beings. Yet, moral responsibility is but one sort of responsibility which might be located within a rather complex realm of different responsibilities, e.g. engendered by contracts or some other mutual agreements which might not necessarily be moral in the narrower sense, i.e., that they might not affect the life, limbs, psyche, andwellbeing of other people or living beings in general. These ethically speaking not morally relevant responsibilities might be called ethically neutral. But they are still normative and prima facie to be abided by the respective persons who have taken over these non-moral responsibilities. In addition, these ethically neutral responsibilities might get into conflict with moral duties and ethically relevant obligations, i.e. moral duties in the narrower sense. Here, differentiation of different sorts, types and levels of responsibility seems required for decision making – especially in an ever-growing complex interconnection and globalization of future societies. Addressing problems of distributing responsibility and the involved social trap constellation there is another nice illustrating case in the literature on collective responsibility.

310

Hans Lenk, Matthias Maring

A vicar had accumulated a remarkable amount of accomplishments for a hamlet of wine-growers. The wine-growers decided therefore to donate a barrel of wine to him celebrating the thankfulness on a special occasion. Each wine-grower should contribute two litres of his best wine from his cellar. Consequently, everyone of these poured the two litres they had agreed on into the open barrel.On the occasion of the respective celebration and after solemn speeches the barrel was opened and the first glass presented to the vicar. But the glass contained but pure water, and the festive mood changed to general shamefulness (Jöhr 1976, 127). It is not known whether the event mentioned in this example did really take place, but it is a very nice illustration of the problem of distributing responsibility. The model example shows immediately how intriguing the problems of responsibility and its distribution are: Who is responsible in this case? Everybody? Not one individually? Each to a certain degree? Problems of distributing responsibility are to be found today in particular in highly developed industrial societies shaped by technology and advanced economies. Personal acting seems to disappear behind collective and group actions. Group and collective action is on the one hand the acting of and the acting within organizations (corporate acting) and on the other hand the action of many actors under strategic and competitive conditions; sometimes the actors are rather independent of one another. With respect to collective actions there are at least two classes of distribution problems or rather distributability problems (which may however overlap): 1. the problem of attributing and distributing responsibility within the organization with respect to internal corporative segregation of work and role assignment as well as with respect to the corporate division of labour and production and 2. the problem of attributing responsibility in the case of collective actions of many actors (be they corporations or individuals) and with respect to dynamic processes and development of the market and technologies. Today and in the near future these problems are becoming extremely relevant and pressing. This is true, to be sure, not only because of the impact of new systems-technological phenomena and processes on their own, but due also to the high social interconnectiveness of action systems, markets and the evergrowing worldwide interrelatedness of societies in general. Questions of distributing collective and corporate responsibility can be distinguished and analysed according to the following approaches: • relations and mutual dependence of agents and legal rules,

Responsibility and Globalization

311

• models of moral responsibility distribution, • responsibility and non-corporate actions, • responsibility and corporate (institutional) actions, • responsibility and (information) systems and • priority rules to address and help to resolve responsibility conflicts. Cases in which somebody fully and exclusively has to take the responsibility are examined in philosophy as a rule. But are there not also other cases of cooperative responsibility, collective/co-operative decisions, and collective action in general, that are gaining much more importance today, in which someone carries full responsibility by sharing responsibility, according to the degree of the individual co-operation or accountability? In other words, does the extent of the distribution of responsibility generally reduce the degree of moral responsibility? As mentioned above, to address and handle problems of collective or corporate responsibilities and the related distribution to individual actors is dependent on the form of collective actions and causes; the respective form of collective action is also decisive and should, in the following, constitute a criterion for the distinction of various ways of attributing responsibility. For instance, a further point of emphasis is the distribution in terms of the responsibility type. If one makes a distinction between a duty to compensate and moral responsibility, then a division as a solution is more likely in the former case than in the latter. Particularly relevant to the distribution of responsibility are negative formulations of preventative and preservation responsibility as well as the responsibility to avoid omissions and failures, which seems to be more suited to be open to a regulation of the contributional and participatory form of responsibility distribution. One should also differentiate in regard to necessary and sufficient conditions of the onset of consequences and damages depending on the failed or omitted or unintentionally neglected actions of several actors. Basic problems of responsibility distribution do not only arise from the noncorporate collective actions of many actors (be they corporations or individuals), but also out of specific strategic conditions, particularly in division-oflabour processes, that is, in labour segregation in the market external to corporations. The effects, results and side-effects of such actions have – and not just nowadays – an increasing impact. Maybe the difficulty can be clarified with the help of examples and model hypotheses regarding of social traps, which until now were discussed mostly within the realm of individual rationality vs. collective irrationality (e.g., Prisoners’ Dilemma, the “Tragedy of the Commons” etc, cf. Hardin 1968, Lenk/Maring 1990).

312

Hans Lenk, Matthias Maring

We all know that negative external synergetic and/or cumulative effects may occur when a large number of actors act along the lines of individual need calculations (only directly responsible for their own interests and acts). Particular components, that as such are relatively, i.e., subliminally, harmless, can lead as a whole to damages or even to the loss of highly appreciated “commons” or public property. It is characteristic of these damages that property rights, i.e., individual rights to use resources, e.g. public ones, are poorly or not at all defined or that they are not observed at all. Externalities are characterized by an incongruity between that outcome for which one is actually responsible and that for which one is made responsible (liable). To avoid the external social costs, these results must, for example, be internalized – incorporated into the ‘production functions’ of a business. Distribution of responsibility comes in at least two forms - In regard to this problem two subproblems emerge: firstly, the question of distribution of responsibility for or in view of cumulative and synergetic damages and, secondly, the question of responsibility for unforeseen or even enforeseeable consequences. With regard to moral judgement, it follows from both of the two subproblems that a personal action responsibility in many a situation and case cannot in general be attributed to an individual agent alone nor, under many a circumstance, can the cause be attributed to a single domain. Not only in the sense of task and role responsibility, but also in the moral and legal sense do the concerned individuals bear a co-responsibility corresponding to their active, potential or formal participation, to their constituting or influential shares (to be determined in each individual case). An extension of the responsibility of operationally manageable models of the distribution of (co-)responsibility are imperative, at least considering the consequences of collective action. Just appeals to the avoidance of social trap situations alone are not very useful. One should also introduce operationally available and efficient measures such as legal sanctions (product liability, collective responsibility, etc.), financial incentives to change production, determination of property rights for public goods, etc. The following rule could serve as a guideline: as many laws, regulations and prohibitions as necessary; as many incentives, individual initiatives and individual responsibility as possible. A different trait with respect to the problems of responsibility distribution includes the external responsibility of corporations, i.e., the corporation and some or all of its members (representative responsibility vs. participatory responsibility), and corporation members alone (reducible to the specific types of responsibility); internal responsibility in differently structured corpora-

Responsibility and Globalization

313

tions (hierarchies etc.) as individual responsibility and co-responsibility; the delegation of responsibility; and varying types of responsibility. Moral responsibility – this is the main and leading hypothesis – is usually, in regard to (at least ideal) corporate action, differentiable and ramified: it is corporations as such, corporation members, or the corporation and its members among others that can be held morally responsible. Individual personal responsibility, though, seems still to be the prototype case in point. The attribution of individual moral responsibility must be separately justified in each case. In general, one should make a distinction between the external (moral, legal, role) responsibility of the corporations and the (corresponding) internal responsibility distribution. Indeed, beside the role or task-specific, the legal and the corporative action responsibility, corporations and institutions do have a personal moral responsibility or an accountability analogous to moral responsibility. This moral responsibility can also be understood as a higher level responsibility; it would exist in addition to and at times independent of the specific individual responsibilities of the individual corporation member – at least analytically speaking. Individual responsibility and corporate responsibility do not as such have the same meaning; they cannot simply be mutually reduced to one another. The responsibility of one kind or type does not replace the other form, although in general, these analytically distinguished model concepts may overlap in social reality. Corporations can act “intentionally”, though in a manner non-reducible to individual action (i.e., they act in the secondary sense, on a higher level of social fiction, on a symbolic or semantically structured and interpreted plane; their actions because of this and the social consequences are no less real than a person’s actions). Such a corporate responsibility, that is not equivalent to the immediately bearable, direct, personal responsibility applies to businesses, the state and corporations as well as to technical and scientific organizations. Until now, the traditional a priori combination of the attribution of moral responsibility only to natural persons, i.e., the concepts of responsibility linked to individuals, appeared to be insurmountable barriers regarding the attribution of moral responsibility to corporations and institutions. Must that be so? We think, no. Rather, the exclusive limitation to the individualistic model thwarted. Should one not rather develop a hierarchical model that adequately and differentially puts the responsibilities on the various levels? Making or holding corporations responsible may also constitute a first step of attributing responsibility in corporate action; the (corporation internal) dis-

314

Hans Lenk, Matthias Maring

tribution problem can be dealt with in a second step. The latter is difficult to deal with according to responsibility types. The division of labour in corporations and large-scale projects on the one hand and the coordination of actions through markets on the other, but especially unwanted and yet unseen combinations of unfavourable factors, which are inclined to result in catastrophes, or even subliminal negligence or carelessness as with the so-called “normal catastrophes” analysed by Perrow will complicate the attribution of (unwanted) consequences of actions, and the attribution of responsibility in all its kinds. The individualistic concepts of ethics in philosophy, technology and economics do not suffice to tackle these problems, they are obviously not adequate any more, since they usually focus almost exclusively on individual actions and not on interactional, collective and corporate forms of actions or structural and systematic contexts. Thus far ethical approaches have indeed been too much oriented toward individual persons, they have not paid enough attention to social aspects, they are not yet adequately adjusted to socioethics and sociophilosophy. This has been mentioned quite a while ago (Lenk, 1979, pp. 69ff.). It is true, so far the problems of complex constellations of causes and the problems of responsibility have been discussed only in a very generalized way in philosophical literature, while jurisprudence is considering them in a much more detailed manner and has indeed come up with some very interesting approaches for solutions – which are of interest for philosophy as well. Yet, as a resumé – and in a modifying and restricting sense – one has to mention that the actually convincing principle of causation faces difficult problems if taken as the only precondition for the attribution of responsibility; these difficulties result primarily from the fact that there are various forms of collective action. Moreover, there obtains a far reaching impossibility to individualize causal factors within synergetic and cumulative processes largely due to the formation of groups and to the effects of adverse combinations of many intruding factors. To a large extent legal regulations (de lege lata) fail as ways for obtaining adequate precautions or even solutions when faced with ecological damages and damages that occur far from the sources of emission. Often a need for legal regulations is acknowledged indeed, yet for strategic environmental planning and policy-making it is thus far only insufficiently considered, nor has it until now been rendered controllable or sanctionable. In the literature one finds discussion and propositions about joint (and several) liabilities including a mutual right to compensation through the formation of sensitive areas and feedback loops of dangerous products and strict liability, the turn-about of the

Responsibility and Globalization

315

burden of proof in cases of high probabilities of causation, compensation through special public funds, (structural) incentives for the internalization of external effects, environmental emission bonuses to be bought by corporations, etc. The main difficulties for legal solutions are amongst others the extant non-liability connected with permitted subliminal (damaging) individual actions, and the definition of limiting and threshold values. Today and in the near future these problems will figure as extremely relevant and pressing not only because of the impact of new systems-technological phenomena and processes on its own. This is not – it seems to me – the only characteristic of the new situation in the superindustrial world of today generating a new kind of ethical problem related to technological progress. The most decisive new perspective for a new interpretation or a new application of ethics is, beyond any doubt, the immense growth in the technological power of man. Due to some specific facts, it requires new ethical perspectives – with wider scopes. Such risk-increasing facts are, e.g.: The number of people affected by technological measures or their sideeffects has increased tremendously. Those affected frequently do not directly interact, in the same context of action, with the other – sometimes intervening – agents. Natural systems literally became an object of human action – at least in a negative way. The human being can permanently disturb or destroy them by his technological measures. Without doubt, this is an absolutely new ethical situation: Humans never before possessed the power globally, or even regionally, to destroy all life in an ecological system, to debase it decisively by technological manipulations. Since some of these encroachments are not controllable and may be irreversible, nature, both as an ecological system and in terms of particular species, should gain a new ethical relevance within the new dimensions of technological power. In the past, ethics was essentially anthropocentrically oriented around actions, interactions, and their consequences among men; however, ethics must now gain an ecological relevance, giving significance to other forms of life (as for example in Schweitzer’s “ethics of reverence for life”). Taking into account possible irreversible harmful effects (e.g., changes of climate, injuries by radiation, technological erosion, overfertilizing and the respective phosphate- and nitratetriggered impacts on rivers, lakes and even seas – cf. the algae problem in recent summers occurring off the coast of Norway and Denmark –, the ozone hole in the Antarctic), the fate of man is also at stake – but not only his fate.

316

Hans Lenk, Matthias Maring

It is particularly the responsibility for usage of limited land and diminished resources of nature, of their respective systems, ecosystems and species, the responsibility for maintaining ecological balances, the co-responsibility for survival conditions of future generations, for maintaining, regenerating and moderately as well as reasonably utilizing resources and the responsibility for side-effects and remote consequences of actions which are dramatically gaining increasing moral and political relevance. With respect to the first class of distributability problems we would like to mention only two theses here (cf. Lenk 1985) and point out to some actual discussions 1. Responsibility is a function of power, impact and knowledge. The more strategically central one’s position is in terms of power, influence and knowledge, the higher his responsibility would be. This is an idea which can be worked out in more detail by using distributive models of graph theory and predistribution assignments of rights and duties according to different levels. The Analytic Hierarchy Process by Thomas Saaty, e.g., uses such a differentiation with respect to levels to employ social values in public decisionmaking. In cooperation with the former Science Advisor of North Carolina, Earl MacCormac, we have set out to apply such concepts to the problem of distributing responsibilities according to strategic centrality in social responsibility systems. A person is co-responsible or assigned/attributed coresponsibility in as much as he or she is actively participating in an action system or can possibly destroy or disturb the system. His or her responsibilities grow with centrality and the respective degree of influence. The basic idea is that every potential actor in the system is co-responsible in proportion to his practical influence. It is not only one leading position holder who would be responsible for everything in the action system. On the other hand, nobody who has no potential influence on the system is to be held responsible. With respect to moral co-responsibility though, there should be no vanishing effect amounting to the well-known ‘Large Committee Irresponsibility’: The more people are on the committee, the less responsibility every member has. Moral co-responsibility cannot be a case of group distribution, trade-off or bargaining. It does and should not vanish with growing numbers of participants. There is still much work to be done to combine the influence-oriented differentiation of co-responsibilities and this non-vanishing effect. (Another problem of adequate distributing responsibility with respect to corporate action should be mentioned here: frequently, in complex organizations the decider who makes up his mind and determines an action in that res-

Responsibility and Globalization

317

pect would not be identical with the literal actor who carries out the action itself.) 2. As came out rather recently, besides legal responsibilities, corporations seem also to bear moral responsibilities (particularly if they do not want to induce or aggravate dangerous conditions as, e.g., the management of Convair in the DC-10 case (Eddy/Potter/Page 1976) or the Air New Zealand in the case of the crash on the Antarctic Mount Erebus (French 1984, ch. XI)). This is certainly a type of moral responsibility different from an individual’s moral accountability. The corporate moral responsibility frequently coincides, but need not be identical with the moral co-responsibility of members of a decision-making board. Corporate moral responsibility therefore is analytically not to be confounded with moral co-responsibility of individual group members partaking in a collective action or decision-making process. Therefore, moral responsibility is analytically to be separated from corporate acting: responsible may be either the corporation as an institution, or members of the corporation – the latter ones not only as role-takers, but particularly as individual persons – or both. In addition, generally speaking one should distinguish the external responsibility for the organization and corporate action from the internal problem of responsibility distribution. The same is true with respect to institutional moral versus personal moral responsibility, although certainly an indirect connection obtains to be traced and analysed case by case which however does not amount to a general definitional reduction of corporate responsibility to solely individualistic terms or factors. Besides role or task-specific responsibility and the legal responsibility and the responsibility for actions, corporations and institutions often also harbour a moral responsibility or a responsibility that is analog to a moral one. This moral responsibility must be understood as a “secondary” responsibility (on a secondary level of action and effect, so to speak), it may exist besides and partially independent from the specific responsibilities of individual members of the corporation, although relationships as between these different types do usually exist, as mentioned above. Individual and corporate responsibilities do not refer to the same moral phenomena; they cannot simply be reduced to the same meaning, yet they do relate to each other. One of these responsibilities does not substitute the other. The second problem of distributing responsibility does not result from collective corporate action by itself, but only if many individuals act under strategic (competitive) conditions, if negative external, synergistic and/or cumulative effects occur. Indeed, “strategic conditions” means that the final result is

318

Hans Lenk, Matthias Maring

dependent on the (relatively independent) acting of many individual actors. Synergistic and cumulative effects would only result, if different components have a joint and mutually escalating impact; the individual components might by themselves be (relatively) harmless, i.e. remain under a certain thresholdvalue, but result in the deterioration or even loss of a highly valued common good (think of the example of the continental European forest “dying” from pollution by acid rain and erosion). The distributability problem of responsibility in these cases consists in the fact that side-effects cannot be attributed to a single originator and that they usually were/are or even could not be foreseen or predicted. We have two partial problems: first the question of participatory responsibility with respect to cumulative and synergistic harmful effects and second the question how to responsibly deal with unforeseen or even unpredictable facts or side-effects. The first problem can be called the problem of distributing responsibility under strategic conditions. For instance, is the legal principle of attributing “causality” and responsibility valid in Japan since the case of the Minamata disease according to which the statistically assessed contribution to the common harm by relevant polluters in the vicinity is ascertained as the pertaining causality by law, indeed satisfactory? The burden of proof here lies so to speak on the side of the potential originator, the polluter, who has to prove the unharmfulness of his emissions. This reversal of the burden of proof seems to be at least a controllable and operational measure to allow for attributability wherever environmental damages are in question. In these detriments usually land, water and air use or misuse are combined. They can at least be tendentiously forestalled or diminished in a controllable way by assigning sanctions. In that respect the Japanese legal principle of attributing causality might foster environmental protection. But there are methodological and legal as well as moral problems connected with such a regulation. First of all, adjacency and the guessing of causality can never be a proof of a causal origin. In addition, an easier part of the problem is how to attribute distributively the responsibility in the cases of synergistic and cumulative damages, particularly those with below-threshold-contributions of individual actors. Another problem is how reliably to distinguish between a descriptive assessment of causal origination and the normative attribution of responsibility, between causal responsibility and liability after Hart (1968). How could one possibly distinguish between the causal impact, the descriptive responsibility (Ladd 1975), i.e. the descriptive attribution of responsibility, and the respective normative attribution of responsibility for contributions – the amount of which is individually ineffec-

Responsibility and Globalization

319

tive, below the threshold of harmfulness? And how is one to distribute this kind of responsibility in general? Would it not be meaningful to postulate a normative collective responsibility of all pertinent corporations within the respective region in the sense of a joint liability? This would, however, mean a liability of all relevant corporations for the total damages. The impaired parties could sue for damages, claim in court for compensation and/or indemnification from any presumably participating corporation. Does this make sense, if connected with an overall generalization? This regulation, however, would have the advantage of dispensing with the proof of damage in respect of each singular damaging or aggrieving party – as, e.g., a respective norm in German Civil law would prescribe (cf. BGB §§ 830 I, 840 I, 421, 426). This kind of regulation would, in some way independent of individual case argumentation, interpret all (non-collective) agents as quasi one corporative agent being liable in total. The internal distribution and compensation within this quasi-group of corporate agents would then be a problem of mutual bargaining of all aggrieving parties. Social traps (see Lenk-Maring 1990) can arise within the contexts of joint (and severable) liabilities, especially with respect to liability funds. Such traps could and should be prevented through specific regulations, e.g. regulations of premia or (positive as well as negative) incentives in correspondence with the potential hazards and damages, rights to compensation against single damagers etc. There is also a sort of social traps involved in abiding by or profiting from ethical codes: People who follow the rules must often deal with disadvantages, while those who transgress them can benefit from advantages (especially when the infringement can be hidden). Problems of control, sanction, trust and security also arise; these cannot be solved through codes alone. Additional institutional measures are indeed necessary. Regarding responsibility conflicts in practice, there are no isolated solutions or suggestions for such cases; instead, applicability rules or practical guidelines on an intermediate level should be developed. These rules should differentiate, for example, between moral ideals (covering virtues as mentioned above) and moral (obligatory) rules (Hennessey/Gert). A combination of individual and institutional measures is necessary: To further and strengthen individual ethical competence is a necessary, but by no means a sufficient step for the efficient solution of responsibility problems and conflicts. An implementation of ethical considerations in law and politics would supplement and enhance this step. In particular, the codes should explicitly set priorities and decision criteria,

320

Hans Lenk, Matthias Maring

which would aid in the solution of the respective responsibility conflicts. Notwithstanding these arguments another kind of total liability with respect to product safety and hazards in terms of environmental damages of public goods should be established. It should be noted that there is a European Community agreement (1985) with regard to product liability laws. Causal originators of damages would then/now be liable in the sense of a strict liability in tort, whether or not they are really guilty in terms of intent or only negligent. Causal origination would already ascertain descriptive causal action responsibility and with respect to the damage of a good to be protected also normative responsibility for the respective action and its consequences. This form of liability would hopefully be deterrent enough to prevent infringements. If, however, damages would nevertheless occur it would at least not be necessary to prove fault or guiltiness as a presupposition of any claim for compensation. These arguments are equally relevant for the second partial problem with regard to non-corporate action. Are the systems responsible? The second important partial problem already mentioned is the question of the possible distribution of responsibility for unforeseen or even unforeseeable system impacts and side effects occurring by synergistic and cumulative effects. Also here single agents cannot be held responsible alone, but the whole system would seem to be quasi-responsible. A system however cannot bear responsibility. This is true even if computer scientists and some social philosophers of computer systems now tend to attribute moral and legal responsibility to information systems and locate the respective responsibility only within the system itself. (Klaus Haefner (1984), a German computer scientist, literally wanted to locate the responsibility “within the systems” since no one can survey and pursue all the ramifications of processes of decision within the system, no single man or woman would be able to be held responsible for the system decision. Is this true?) However, if responsibility would reside merely within the systems and no longer with persons, persons might all too easily claim for excuses. This is a further problem of the responsibility distribution emerging from the use of expert and information systems (cf. Lenk 1989). Can these be responsible? Resides the “responsibility [...] in the systems” (Haefner); can we make complex informational decision-making systems and expert systems responsible? Is that not an unnoticed introduction of irresponsibility with no

Responsibility and Globalization

321

one to be appointed guilty, an infringement or transgression of a taboo or even a categorical mistake on the side of the analyser? It is indeed meaningful and important to make computer systems more reliable, but it is not meaningful to attribute moral trustworthiness and responsibility to them. Indeed, that would be absurd and sound odd! Computers are not moral beings, just as information systems are not social beings. In spite of their far-reaching social implications, human beings must carry the full responsibility for the use or misuse of technical systems – but which human beings? The programmer? The director of the computing centre? The entrepreneur? The politician? ... The responsibility may, in view of the possible farreaching implications of responsible decisions for humans, especially individuals, hardly seem bearable, but morally it still exists. Human beings cannot morally deprive themselves of their power of decision and their accountability, they cannot and should not cede their moral responsibility to computers and information systems. (This thesis must, however, still be established and worked out). In view of the factual expansion of automated conditioned decisions, this responsibility dilemma, which the participants and the higher decision-makers cannot avoid, will become more and more pressing in the future. Responsibility cannot be allowed to be diluted, either in anonymity or under the protection of committees, or in the information and decision-making systems. Possibilities of a counter-reaction exist in a higher sensitization of the responsibility awareness, in the development of a Code of Ethics for computer experts, in the interdisciplinary research and in an alliance in teaching and training of all knowledge-oriented disciplines, etc. Are humans because of their immense power of technical encroachment and feasibility beyond any beforehand imagination and control responsible for much more, so to speak, than they could possibly foresee and literally (intentionally) be normatively responsible for? Should they not take over responsibility for unforeseen or even unforeseeable side-effects of his actions with respect to technological and scientific big science projects? But how could they possibly do that? There is no way of a really morally being held responsible for something it or one does not know or even could not know. In the sense of causal responsibility (descriptive origination) one can be held responsible in some sense, if an unintended damage occurs; the question however is, whether one could be held responsible also in a normative-moral sense. The so-called principle of causation if interpreted in a moral and legal sense, would – at least in tendency – adequately design normative responsibility too. One would have to answer for, to make good and to be liable for consequences in the sense

322

Hans Lenk, Matthias Maring

of being liable to pay compensation etc. The range and power of action seems to have multiplied to such a degree that anticipation cannot follow quickly enough or pursue all the complex ramifications of impacts, consequences and side-effects. This seems to be an intriguing dilemma of responsibility in our systems technological age impregnated by complex systems interactions and dynamic changes easily transgressing linear thinking and traditional causal disciplinary knowledge. However, we have to be venturous, to enter new frontiers, in order to be able to discover, invent and test new scientific and technological solutions and to uphold social progress in a mass society, but we have generally speaking and with respect to big science, big technology and their hazardous potential with regard to man and nature to be rather precautious and proceed very cautiously, taking into account systems interrelations (notably also non-linear ones usually amounting even in cases and phases of chaotic, yet deterministic systems to unpredictability) and systems interactions and even potential occurrences or probabilities and risks of negative side-effects. Perhaps here Karl Popper’s idea of a piece-meal social engineering and piecemeal technology avoiding probably reversible unwanted consequences should be relevant not only with respect to public goods. Technology, technological progress and economic-industrial development in combination with the respective damages for land, clean air and water turn out to be multidimensional phenomena asking for interdisciplinary and complex approaches. The multi-perspectivity is the result of a mutual on-going interaction between diverse realms and actions of many corporate and individual agents. This is leading to a rather great complexity of individual, collective and corporate contributions, different areas and social background factors. The exponential structure of technological development in terms of range, energy, acceleration, interaction feedback phenomena etc. is a familiar insight of traditional sociology of science, technology and economic development. This insight is generally true for any multi-ramified and interdisciplinarily interlocked social phenomenon of development. With respect to moral judgment it seems clear that here causal origination also, (e.g., of some principally man-made environmental effects) usually cannot be attributed to one single individual agent nor can causation be solely assigned to one single realm, if development and acceleration are dependent on a multiplicity of mutually interacting and escalating impacts nor can moral responsibility simply be attributed to the whole systems as such. Not only with respect to different sorts of role and task responsibility, but also in moral and legal terms participating and contributing individual agents have to bear a

Responsibility and Globalization

323

certain co-responsibility according to their active, potential or actual or formal participation. For instance, in gene biology the exposure of gene-technologically produced bacteria cannot remain only within the moral responsibility of individual scientists who could not even possibly foresee or predict some harmful side-effects. This is certainly an intriguing example drastically increasing in prominence today and in the near future. Undoubtedly, all responsibility of researchers in applied science and technology has to be preventionoriented and take into account possibly affected individuals, species and ecosystems, whenever and wherever adverse developments and damaging effects can be pre-assessed and avoided. This is especially true with directly applied technological projects. Preventive responsibility is particularly important in these critical realms, situations and constellations. With regard to responsibility in general, it is not only corporations and institutions in economics and industry which have to bear responsibility, but also the state and its representative decision makers. Corporate responsibility has to be connected with individual responsibilities of the respective representative decision makers. This is true also for big technology projects, particularly if they are run by the state itself. There should be not only a legal, but also a moral balance of powers in terms of checks and controls similar to the traditional distribution of power between legislature, government and jurisdiction. The central point with respect to the problem of the distribution of responsibility is the question concerning the normative and descriptive distributability in terms of the theory of action and the possibility of adequately reducing the collective responsibility to individual actors in relation to the form of collective actions and causations. Thus, the respective form of a collective action is of determining value and figures as a criterion for the distinction of various attributions of responsibility. Another important point is that the distribution of responsibility is dependent on the kind of responsibility: If one differentiates between a legal liability for compensation and moral responsibility, a distribution is (more) easily attainable with the first kind, while it might be not (so) easy with moral responsibility. In particular, the negative formulations of a responsibility for prevention of damages and the preservation of states of well-being etc. are relevant for the distribution of responsibilities – as is the responsibility to prevent omissions, which is more easily accessible for a regulation of responsibility. One should also differentiate between the sufficient and the necessary conditions of a consequence or damage in relation to several involved persons’ failures to act. So, the individual failure to

324

Hans Lenk, Matthias Maring

act is causally sufficient for the occurrence of the consequence or damage, if non-ommission of the act would have prevented this occurrence. Again, the upshot of this in terms of moral responsibility might be formulated as this: The extension of individualistic responsibility is to be combined with the development of a socially proportionate co-responsibility, and with the establishment and analytic as well as institutional elabouration of corporate responsibility and a new sensitivity of and for moral conscience. Types of responsibility have to be analysed in a more differentiated way than hitherto (see, e.g., Lenk 1992, Lenk in Durbin 1991, 187ff.). Only in this way we may be able to cope with the most complex structures of causal networks and far-ranging consequences of actions, be they individual, strategic, collective, or corporate. Concepts for a more social orientation of responsibility and conscience should be given most attention. Ethics and moral philosophy have to take seriously these new systemic challenges by technically multiplied possibilities and impacts of action and systems networks. This does not mean that our traditional basic intuitions of ethics and morality have to be denied. On the contrary, in so far as morality has to deal with the well-being of other men and creatures (maybe even in the sense of Schweitzers’ “reverence for life”), analytic and normative ethics have to pay attention to the ramifications of systemic impact and action patterns. The accrued “incongruence of technical and ethical competency” should be diminished in the future (ibid. 167). Therefore, an extension of our horizon of moral judgments with respect to future generations, quasi-rights of some non-human natural beings, e. g., higher animal species, but also of eco-techno-systems etc. is important, if not paramount for any ethical agenda. Ecotechnofacts and such systems are going to pose real challenges for applied ethics in the future. Thus, there exists an ethical obligation for humans to take care that especially humankind as well as other natural kinds dependent on the human power for intervention do not get extinguished. It is true that individual beings, which have not yet been conceived, have no individual moral or legal right to be born, and one cannot impose an individual obligation on particular human couples to procreate, but it seems to be a sensible extrapolation from the constitutional rights of humankind, which are else often only constructed as rights of repulse and protection, to develop a collective responsibility of today’s living humans that they must not let their species be extinguished or destroyed. Humans have not only the – negative – responsibility to leave behind wholesome conditions of environment and life for future generations, which means they should not totally exploit non-regenerable raw- materials

Responsibility and Globalization

325

and that should refrain from lethal poisoning, depletion and destruction of the environment. They collectively also have an obligation and responsibility to actively prevent this from happening and to work for a future existence of humankind in life conditions worthy of human beings. This is at least a moral demand which originates in the integrity and continued existence of humankind, which are considered the highest desirable values by various ethical systems. Even a version of Kant’s (AA IV, p. 341) formal Categorical Imperative refers to the actual content of the ‘principle of humankind and of any reasoning nature’ as things in themselves. Judged morally, then, future generations’ relative rights or quasi-rights to existence do exist, even though no singular existence of a non-conceived individual can be sued for on a moral or legal basis. Thus, certain general human and moral obligations transcend those which are individualistically and juristically concretized. Moral value commitments are more comprehensive and determining than moral or legal individual responsibilities. Morality is more than a singular individual responsibility or obligation. Whether a summarily conducted extension of responsibility with respect to side-effects of collective action by corporations is enough, will be certainly a matter of debate. Peter French (1984, 132ff.) proposed an “Extended Principle of Accountability”, according to which somebody is morally responsible not only for intentional acts, but also “for some of the unintended effects of ... actions”, which these persons “should have (known) or did know would occur”. However, “being willing to do something ... does not entail intending to do it”. Even if an operational model in terms of legal compensation etc. were introduced, it remains debatable whether this criterion might be really treated as a moral criterion of responsibility. French presents a second extension of the principle of responsibility implying that a person “may be held accountable for those non-original or second effects that involve the actions of other persons that he obliquely or collaterally intended or was willing to have occur” or occurred. Paul Thompson even goes further than that: “Another strategy is to describe the act as one which risks the harmful consequence, and to assume that, in the absence of exculpatory circumstances, this is sufficient to serve as the basis for an ascription of moral responsibility” (Thompson 1985, 301). The decisive new approach here is “that in risking, one is doing something for which one may be held responsible” (ibid. 308, our italics). Imposing risk on somebody is now an action for which one is responsible; it is no longer as in traditional economics for instance just a probability distribution of action consequences or events and results which do not contain any objective evil (ibid.

326

Hans Lenk, Matthias Maring

307, 309). “Causal agency requirements are parts and parcel of risking” (ibid. 310). In this important causal connection of results and impacts and even unforeseeable side-effects of actions to the action of risk imposing, Thompson sees the advantage of his own approach compared to French’s book. For “some cases of morally untoward risking, however, will not be cases in which the agent was in any sense willing to countenance the harm that actually does materialize” (ibid. 309). If we would like to follow this extension of risk assignment and responsibility attribution we have to develop operational models of distribution of and participation in responsibility with respect to collective actions. Just ideal postulates without practical chances of realization by institutional measures would remain rather ineffective. Just pleading to avoid social traps does not seem very useful as such. Pleas are necessary, but per se they are not sufficient, not effective. One has to introduce operational procedures, measures and consequences as e.g. legal sanctions, financial stimuli and incentives to change hazardous productions and products, public and moral institutional means of pillorying as well as guaranteeing of property rights and protection of public goods combined with mechanisms of control and sanctioning in order to be effective. As mentioned above, a general guide-line pertaining to institutional measures to be established could be: As many laws, precepts and decrees and prohibitions as necessary, as many incentives and stimuli for self-initiative and individual responsibility as possible. Even if this might sound rather global and vague, it might generate the direction of a necessary moral reorientation in public matters of social trap constellations.

III. Priority Rules In considering different types and levels of responsibility, we will also have to develop priority rules, as for example, that moral responsibility takes precedence over role-responsibility. We would like to propose the following 20 rules of preference and priority which are not systematically arranged in a successive order and valid under prima-facie-conditions (that but except the first four ones presented just is they may be over-ruled by higher and more binding moral obligations). (The first four rules are adapted from Werhane 1985, pp. 72.) 1. To weight moral rights of the respective individual; these moral rights are un-alienable predistributive or primordial rights overriding utility considerations.

Responsibility and Globalization

2.

3.

4.

5.

6. 7. 8.

9. 10. 11.

12.

327

To seek a compromise taking into consideration interests of everyone on an equal basis; in case of an unsolvable or seemingly unsolvable conflict between equally relevant basic rights the condition mentioned in the clause is especially important. Only after considering the moral rights of each party one should vote for the solution which causes the least damage or maximizes utility for all involved parties. Only after application of rule 1., 2. and 3. utility considerations are to be weighted against potential harm. That means in general: Non-alienable (predistributive) moral rights are prior to considerations of avoiding harm and damage and these latter are prior to utility considerations. In practically unsolvable conflicts one should look for fair compromises (that is for compromises which involve proximatly equally distributed or proportionally justified distributions of disadvantages and utilities respectively.) General (higher level) moral responsibility is to obtain a preference over restricted nonmoral prima-facie-obligations. Universal moral responsibility generally takes preference over role and task responsibility. Direct or primary moral responsibility is usually but not always to be considered prior to indirect responsibility for remote consequences. (This is true because of urgency but should at times and in cases be modified according to importance and impact of consequences and long range effectiveness.) (See also rule 20 below.) Primary and personal moral responsibility precedes the secondary or second-level corporate responsibility. The public weal as well as “the common good” surpasses all other specific, special, specious and particular interests. In technical rules and regulations for applied science important principles of priority are formulated regarding safety regulations, e.g. Rule DIN 31.000 of the German Technical Regulation DIN explicitly states: “With respect to safe design (it is imperative that) that solution has to be preferred for which the safety goals will be reached in a technologically meaningful way and the best economical manner. In case of doubt safety requirements take precedence over economical consideration.” Safety goes on top of technical-functional and economic considerations (as for instance DIN 31.000 would postulate).

328

Hans Lenk, Matthias Maring

13. Global or continental as well as regional and local environmental compatibility are to be differentiated and have to be taken into account: system-relevant or decisive environmental compatibility and usually the comparatively speaking more comprehensive compatibility stay in front. Sustainable development of ecosystems is particularly urgent on each of these levels. 14. Urgency of ecocompatibility and sustainability (especially a system-decisive one) are to top economic utility. 15. Social and human compatibility would in the case of conflict precede above environmental and nature or species compatibilities which are however still to be considered by the way of reaching for meaningful compromises. 16. Human and social acceptability surpass functional efficiency and utility. 17. Anticipated acceptability of and compatibility with the requirements for the survival and quality of life of future human generations should take very high priority. 18. Social and political planning at large should take into account endeabours to achieve a relative maximum of liberty and freedom of decisionmaking (openness and flexibility of large-scope planning) and largely equal opportunities for future generations. 19. In the same vein, a relative (potential) multiplicity of options for the generations to come should have a high priority (“multioptions society”), e.g., no important options should be blocked for them (avoiding total resource depletion and environmental pollution by favoring sustainable development). 20. Practical (or practice-oriented) humanity and humaneness should go in front of abstract requirements and formal universal principles. Such rules of priorities are conducive to tracing and probably solving conflicts between different types of responsibilities obtaining in a particular actual situation of conflict. Whereas differentiating between the levels and types of responsibilities is necessary for the discovery and identification of conflicts, the rules of priorities could helpfully be applied in solving or at least regulating and assessing the respective conflict-situations and in tracing their special sources. Yet, in this realm much work has still to be done in the future.

Responsibility and Globalization

329

IV. Conclusion and Abstract The paper tries to delineate a moral and in some sense pragmatic-political approach to typical globalization problems and social, corporate, political as well as personal responsibility confronting the new challenges of globalization or the ideology of that in today’s and future societies. Some analytical tools of analyzing the intricacies of the ideology of globalization and practical social and political consequences of that are outlined and compared with economic and social as well as political trap situations as well as spirals or exacerbation or degeneration processes, notably regarding environmental degradation and resource exploiting. The leading idea is that globalization of liberal economics should be accompanied by a sort of globalization of responsibility in order to protect the global future of mankind and global commons and possibly to avoid particularly secondary social traps stemming from the ideology of globalization nowadays publicly utilized by corporations and some hardcore-liberal market economists. It is necessary to avoid not only traps and spirals of self-destroying prophecies, but also tendencies of a structural irresponsibility flowing from the talk and ideology as well as some processes of world wide globalization in economics. Environmental risks and consequences of non-corporate strategic collective actions of many actors (notably corporations) leading to what can be called a globalization trap are discussed and related to the problem of collective and corporate as well as institutional moral responsibility. Problems of distributing collective corporate responsibilities are certainly questions of a secondary philosophic analysis as well as a problem of outlining some priority rules and – last but not least – of introducing institutional measures for resolving responsibility conflicts and practices. There are no isolated solutions for individuals, corporations and states respectively, but there is a growing necessity of combining negative sanctions and codes with legal sanctions, financial stimuli and incentives to change hazardous productions and products, to protect the environments and eco-systems as well as public and moral means of pillorying infringements as well as protecting public goods in order to be effective. Above all, international treaties like the as yet rather ineffective environmental covenants (like Rio, Kyoto and, disappointingly, The Hague lately) are necessary and to be rendered more effective than hitherto. – For philosophers it is an interesting task to delineate the problems of distributing responsibilities, notably collective and corporate ones, and to develop priority rules applying to cases of responsibility conflicts.

330

Hans Lenk, Matthias Maring

References Birnbacher, D. 1988. Verantwortung für künftige Generationen. Stuttgart. Buchanan, A. 1985. Ethics, Efficiency, and the Market. Totowa, NJ. Constanza, R. et al. 1997. “The value of the world’s ecosystem services and natural capital.” Pp. 253-260. In Nature Nr. 6630, Vol. 387. Demsetz, H. 1967. “Toward a Theory of Property Rights.” Pp. 347-359. In American Economic Review 57. Diederichsen, U., Scholz, A. 1984. “Kausalitäts- und Beweisprobleme im zivilrechtlichen Umweltschutz.” Pp. 23-46. In Wirtschaft und Verwaltung. Donaldson, Th. 1982. Corporations and Morality. Englewood Cliffs, NJ. Durbin, P. T. (Ed.) 1991. Critical Perspectives on Nonacademic Science and Engineering. Bethlehem (Lehigh), London – Toronto (Assoc. Press). Eddy, R., Potter, E., Page, B. 1976. Destination Disaster. New York. Frankena, W.K. 1981. Analytische Ethik. München. French, P.A. 1984. Collective and Corporate Responsibility. New York 1984. Haefner, K. 1984. Mensch und Computer im Jahre 2000. Basel, Boston, Stuttgart 1984. Hamburger, H. 1979. Games as Models of Social Phenomena. San Francisco, CA. Hamm, B. 1998. “Internationale Verflechtung und Globalisierung.” Pp. 339-351. In Schäfers, B., Zapf, W. (Hg.). Handwörterbuch zur Gesellschaft Deutschlands. (Red.: S. Misosch.) Opladen. Hardin, G. 1968. “The Tragedy of the Commons.” Pp. 1243-1248. In Science 162. Hart, H.L.A. 1968. Punishment and Responsibility. Oxford 1968. Hennessey, J.W., Gert, B. 1985. “Moral Rules and Moral Ideals: A Useful Distinction in Business and Professional Practice.” Pp. 105-115. In Journal of Business Ethics 4. Hesse, H. (Ed.) 1988. Wirtschaftswissenschaft und Ethik. Berlin. Homann, K. 1989. “Entstehung, Befolgung und Wandel moralischer Normen: Neuere Erklärungsansätze.” Pp. 47-64. In Pappi 1989. Jöhr, W.A. 1976. “Die kollektive Selbstschädigung durch Verfolgung des eigenen Vorteils.” Pp. 127-159. In Neumark, F., Thalheim, K., Hölzler, H. (Eds.). Wettbewerb, Konzentration und wirtschaftliche Macht. Berlin. Kerber, W. 1988. “Sittlich handeln unter dem Zwang ökonomischer Sachzwänge.” Pp. 241-258. In Hesse 1988. Ladd, J. 1975. “The Ethics of Participation.” Pp. 98-125. In Nomos 16 (1975). Lenk, H. 1975. Pragmatische Philosophie. Hamburg. – 1979. Pragmatische Vernunft. Stuttgart. – 1981. “Herausforderung der Ethik durch technologische Macht. Zur moralischen Problematik des technischen Fortschritts.” Pp. 5-38. In Gesellschaft für Rechtspolitik. Trier (Ed.). Bitburger Gespräche. Jahrbuch 1981. München. – 1982. Zur Sozialphilosophie der Technik. Frankfurt 1982. – 1983. “Verantwortung für die Natur.” Pp. 1-18. In Allgemeine Zeitschrift für Philosophie 8. – 1985. “Mitverantwortung ist anteilig zu tragen – auch in der Wissenschaft.” Pp. 102-

Responsibility and Globalization

331

109. In H.M. Baumgartner and H. Staudinger (Eds.). Entmoralisierung der Wissenschaften. Ethik der Wissenschaften vol. II. München – Paderborn. – 1987a. “Über Verantwortungsbegriffe und das Verantwortungsproblem in der Technik.” Pp. 112-148. In Lenk/Ropohl 1987. – 1987b. Zwischen Sozialpsychologie und Sozialphilosophie. Frankfurt a.M. – 1988a. “Verantwortung zwischen Individualismus und Institutionalismus.” Pp. 6593. In Ethische Aspekte des Leistungssports. dvs-protokolle Nr. 33. Red. E. Franke. Clausthal-Zellerfeld 1988. – 1988b. “Welterfassung als Interpretationskonstrukt.” Pp. 69-78. In Allgemeine Zeitschrift für Philosophie 13. – 1989. “Können Informationssysteme moralisch verantwortlich sein?” Pp. 248-255. In Informatik-Spektrum 12. – 1989. “Types of Responsibility.” Pp. 1-5. In: RSA 2000: Dialogue with the Future11 (2). – (Ed.) 1991. Wissenschaft und Ethik. Stuttgart. – 1991. “Real-world Contexts and Types of Responsibility.” Pp. 183-192. In Durbin (Ed.) (1991). – 1991. “Ethikkodizes – zwischen schönem Schein und ‘harter’ Alltagsrealität.”Pp. 327-353. In Lenk, H., Maring, M. (Eds.). Technikverantwortung. Güterabwägung – Riskiobewertung – Verhaltenskodizes. Frankfurt a.M. – 1992. Zwischen Wissenschaft und Ethik. Frankfurt a.M. – 1994. Macht und Machbarkeit der Technik. Stuttgart. – 1994a. Von Deutungen zu Wertungen. Eine Einführung in aktuelles Philosophieren. Frankfurt a.M. – 1995. “Towards a Systematic Schema Interpretationism.” Pp. 57-69. In: Imamichi, T. (Ed.). Entretiens sur les enjeux actuels de l’ éthique. The Present Commitments of Ethics. Entretiens de Kyoto 1994 de l’ Institut International de Philosophie. Tokyo (Centre International pour Etude Comparée de Philosophie et d’ Esthétique). – 1992. Zwischen Wissenschaft und Ethik. Frankfurt a.M. – 1994. Von Deutungen zu Wertungen. Eine Einführung in aktuelles Philosophieren. Frankfurt a.M. – 1997. Einführung in die angewandte Ethik: Verantwortlichkeit und Gewissen. Stuttgart, Berlin, Köln. – 1998. Konkrete Humanität. Vorlesungen über Verantwortung und Menschlichkeit. Frankfurt a.M. – 1999. Praxisnahes Philosophieren. Eine Einführung. Stuttgart. –, Maring, M. 1984. “Gemeinwohl als Aufgabe für Ethik und Recht.” Pp. 35-44. In Straube, M., Weimar, R. (Eds.). Jurist und Technik zwischen Wissenschaft und Technik. Festschrift für Josef Kühne zum 60. Geburtstag. Wien. –, Maring, M. 1989. “Der Ingenieur als Experte eines Freien Berufs und seine Verantwortung in der technischen Gesellschaft.” Pp. 333-347. In Steinmann/Löhr 1989. – , Maring, M. 1990. “Responsibility for Land Use and the Problem of Social Traps. Soziale und ökonomische Aspekte der Bodennutzung.” Pp. 31-50. In Fitch, D.B.S.

332

Hans Lenk, Matthias Maring

– Pikalo, A. (Eds.). Socio-Economic Aspects of Land Use Plannung. Frankfurt a.M., Bern, New York. –, Maring, M. 1990. “Verantwortung und soziale Fallen.” Pp. 49-57. In Ethik und Sozialwissenschaft 1 (1990). –, Maring, M. 1995b. “Wer soll Verantwortung tragen? Probleme der Verantwortungsverteilung in komplexen (soziotechnischen-sozioökonomischen) Systemen.” Pp. 241-286. In Bayertz, K. (Ed.). Verantwortung. Prinzip oder Problem? Darmstadt. –, Maring, M. (Eds.) 1992. Wirtschaft und Ethik. Stuttgart. –, Maring, M. 1993. “Verantwortung – Normatives Interpretationskonstrukt und empirische Beschreibung.” Pp. 222-243. In Eckensberger, L., Gähde, U. (Hg.). Ethische Norm und empirische Hypothese. Frankfurt a.M. –, Maring, M. 1995a. “Wer soll Verantwortung tragen? Probleme der Verantwortungsverteilung in komplexen (soziotechnischen-sozioökonomischen) Systemen.” Pp. 241-286. In Bayertz, K. (Hg.). Verantwortung. Prinzip oder Problem? Darmstadt. –, Maring, M. 1995b. “Zum Verhältnis von Technikethik und Wirtschaftsethik.” Pp. 3340. In Geschichte und Gegenwart 14. –, Maring, M. 1996. “Technikbewertung – Ernstfall der vorsorgenden Verantwortungsethik.” Pp. 78-95. In Jahrbuch für Christliche Sozialwissenschaften 37. –, Maring, M. 1997. “Technik zwischen Können und Sollen: Wer verantwortet die Technik?” Pp. 92-118. In Putlitz, G. Frhr. zu, Schade, D. (Hg.). Wechselbeziehungen Mensch – Umwelt – Technik. Stuttgart. –, Maring, M. 1998a. “Das moralphilosophische Fundament einer Ethik für Organisationen – korporative und individuelle Verantwortung.” Pp. 19-35. In Blickle, G. (Hg.). Ethik in Organisationen. Göttingen. –, Maring, M. 1998b. “Wirtschaftsethik und Technikethik aus interdisziplinärer Sicht. Themen und Erfahrungen.” Pp. 72-103. In Lohmann, K.R., Schmidt, T. (Hg.). Akademische Philosophie zwischen Anspruch und Wirklichkeit. Frankfurt a.M. –, Ropohl, G. (Eds.). 1987. Ethik und Technik. Stuttgart 1987 (2. Auflage 1993). Maring, M. 1989. “Modelle korporativer Verantwortung.” Pp. 25-41. In Conceptus 23. Martin, H.-P., Schumann, H. 1996. Die Globalisierungsfalle. Reinbek. Mellema, G. 1985a. “Groups, Responsibility, and the Failure to Act.” Pp. 57-66. In International Journal of Applied Philosophy 1985. Mellema, G. 1985b. “Shared Responsibility and Ethical Dilutionism.” Pp. 177-187. In Australasian Journal of Philosophy 63. Mellema, G. 1988a. “Groups, Responsibility, and Risk Taking in Business Organizations.” Pp. 593-603. In Journal of Business Ethics 7. Mellema, G. 1988b. “Causation, Foresight and Collective Responsiblitity” P p. 44-50. In Analysis 48. Merton, R.K. 1976. “Die Eigendynamik gesellschaftlicher Voraussagen.” Pp. 144-161. In Topitsch, E. (Hg.). Logik der Sozialwissenschaften. Köln 1976. Olson, M. 1968. Die Logik des kollektiven Handelns. Tübingen. Pappi, F. U. (Ed.) 1989. Wirtschaftsethik. Gesellschaftliche Perspektiven. Special Issue, Christiana Albertina University. Kiel 1989.

Responsibility and Globalization

333

Perrow, C. 1987. Normale Katastrophen. Frankfurt a.M., New York. Popper, K. 1974. Das Elend des Historizismus. Tübingen. Rapp, F., Durbin, P.T. (Eds.) 1982. Technikphilosophie in der Diskussion. Braunschweig, Wiesbaden. Rawls, J. 1975. Eine Theorie der Gerechtigkeit. Frankfurt a. M. Ropohl, G. 1985. “Die Idee des technischen Fortschritts.” Pp. 153-174. In Wissenschaft Sei(n) Lust. Festschrift für Hans Lenk. Karlsruhe (xeroc.) Ropohl, G. 1989. An den Grenzen der Ingenieurethik. Pp. 560-568. In Universitas 44. Saaty, Th.L. 1980. “The Analytic Hierachy Process: Planning, Priority Setting, Ressource Allocation.” New York u.a. Sen, A. 1987. On Ethics and Economics. Oxford 1987. Stehling, F. 1989. “Ökonomische Aspekte des Umweltschutzes – Ökonomie und Ökologie im Konflikt?” Discussion Paper (Nr. 360). Institut für Wirtschaftstheorie und Operations Research. Universität Karlsruhe. Steinmann, H., Löhr, A. (Eds.) 1989. Unternehmensethik. Stuttgart. Taylor, M. 1987. The Possibility of Cooperation. Cambridge. Thompson, P. B. 1985. “Risking or Being Willing: Hamlet and the DC-10.” Pp. 301310. In The Journal of Value Inquiry 19. Tsuru, S., Weidner, H. (Eds.) 1985. Ein Modell für uns: Die Erfolge der japanischen Umweltpolitik. Köln. Unger, S.H. 1982. Controlling Technology: Ethics and the Responsible Engineer. New York. Vallender, K.A. (1998): “‘Weltbinnenmarkt’ und Umweltschutz.” Pp. 553-577, 371. In Ruch, A. u.a. (Hg.). Das Recht in Raum und Zeit. Festschrift Lendi. Zürich. Vanberg, V. 1982. Markt und Organisation. Tübingen. Weidner, H. 1985. “Bahnbrechende Urteile gegen Umweltverschmutzer.” Pp. 92-108. In Tsuru/Weidner 1985. Werhane, P.H. 1985. Persons, Rights, and Corporations. Englewood Cliffs, N.J.

334

Published Volumes

CONTRIBUTORS Dr. ir. Vincent Buskens, Utrecht University, Department of Sociology / ICS, Heidelberglaan 1, 3584 CS Utrecht, The Netherlands. Prof. Alvin I. Goldman, University of Arizona, Department of Philosophy, Tucson AZ 85 85721, USA. Prof. Roger Gibson, Washington University, Campus Box 1073, One Brookings Drive, St. Louis, Missouri 63130-4899, USA. Prof. Margaret Gilbert, University of Connecticut, Philosophy Department U-54, Storrs, Ct 06269, USA. Dr. Frank Hindriks, Erasmus University Rotterdam, Faculty of Philosophy, P.O. Box 1738, 3000 Dr Rotterdam, The Netherlands. Prof. Hans Lenk, Universität Karlsruhe, Fakultät für Geistes und Sozialwissenschaften, Institut für Philosophie, 76128 Karlsruhe, Germany. Dr. Mathias Maring, Universität Karlsruhe, Fakultät für Geistes und Sozialwissenschaften, Institut für Philosophie, 76128 Karlsruhe, Germany. Prof. Anthonie Meijers, Eindhoven University of Technology, Department of Philosophy and Social Sciences, Den Dolech 2, P.O. Box 513, 5600 MB Eindhoven, The Netherlands. Dr. Pekka Mökelä, Academy of Finland, University of Helsinki, Department of Philosophy, PO Box 9, Siltavuorenpenger 20, 00014 Helsinki, Finland. Prof. Steve Miller, Loyola University. Department LFCD, School of Education, Mallinckroth Campus, 1041 N. Ridge Road, Wilmetta IL. 60091, USA. Privatdozent Dr. habil. Gerhard Preyer, J. W.-Goethe-Universität, 60054 Frankfurt am Main, Germany. Prof. Lambér Royakkerss, Eindhoven University of Technology, Faculty of Technology Management, P.O. Box 513, 5600 MB Eindhoven, The Netherlands. Prof. Fred Schmitt, Department of Philosophy, Sycamore Hall 026, Indiana University, 1033E Third Street, Bloomington, Indiana 47405-7005, USA. Prof. Deborah Tollefsen, The Ohio State University, 350 University Hall, 230 North Oval Mall Columbus, OH 43210, USA. Prof. Raimo Tuomela, Academy of Finland, University of Helsinki, Department of Philosophy, PO Box 9, Siltavuorenpenger 20, 00014 Helsinki, Finland.

Published Volumes

335

IMPRINT ProtoSociology: An International Journal of Interdisciplinary Research ISSN 14344319 Editor: Gerhard Preyer Johann Wolfgang Goethe-Universität Frankfurt am Main, Department of Social Sciences Editorial Staff: Georg Peter Editorial of the Vol. 16 2002: Georg Peter Layout and digital publication: Georg Peter Editorial Office: ProtoSociology, Stephan-Heise-Str. 56, 60488 Frankfurt am Main, RFA, Phone: 069-769461, Email: [email protected], [email protected] Bank: Dresdner Bank AG, Frankfurt am Main, account: 44 121 168 01, BLZ: 500 800 00 SWIFT-BIC: DRES DE FF IBAN DE60 5008 0000 4412 1168 01 The Vol. 16 is sponsored by GASS INTERNATIONAL, Vorlandstr. 4, D-77756-Hausach, www.gass-logistics.com )))))))))))))))))))))))))))))))))))))))))) Die Zeitschrift soll 1/2jährlich erscheinen. Die Anzahl der jährlich erscheinenden Hefte und Sonderhefte bleibt jedoch vorbehalten. Copyright: Die in dieser Zeitschrift veröffentlichten Beiträge sind urheberrechtlich geschützt. Alle Rechte sind vorbehalten. Übersetzungen, Nachdruck, Vervielfältigung auf fotomechanischem oder ähnlichem Weg oder im Magnettonverfahren, Wiedergabe durch Vortrag, Funk- und Fernsehsendungen sowie Speicherung in Datenverarbeitungsanlagen, auch auszugsweise, sind nur mit Genehmigung des Herausgebers möglich. Für eingereichte Beiträge wird keine Haftung übernommen. Weitere Publikationsrechte von Artikeln bleiben vorbehalten. Zweitpublikationen seitens des Autors werden eingeräumt. Bei einer Zweitpublikation ist das Heft (Nummer, Titel, Erscheinungsjahr) der PROTOSOCIOLOGY zu zitieren. Für unaufgefordert eingesandte Manuskripte wird keine Haftung übernommen. Gerichtsstand ist Frankfurt am Main. Copyright: All rights reserved. This publication may not be reproduced, stored or transmitted in any form or by any means without the prior permission in writing of the publisher. Additional publications of the articles are reserved. The authors retain the personal right to re-use their own articles. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients is garanted by PROTOSOCIOLOGY, provided that the base fee is paid directly to VG Wort, Goethestr. 49, 80336 München RFA.The publisher accepts no responsibility for submitted manuscripts. ProtoSociology im Internet: www.protosociology.de

336

Published Volumes

ON PROTOSOCIOLOGY Protosociology occupies an important position in the European intellectual scene, bridging philosophy, economics, sociology and related disciplines. Its volumes on rationality bring together concerns in all these topics, and present an important challenge to the cognitive sciences. Donald Davidson, Berkeley (USA) Protosociology publishes original papers of great interest that deal with fundamental issues in the human and social science. No academic library is complete without it. Nicholas Rescher, Pittsburgh (USA) Protosociology has been remarkably successful in publishing interesting work from different tradition and different disciplines and, as the title signals, in giving that work a new, eye-catching slant. Philipp Pettit, Canberra, Australia Protosociology is a truly premier interdisciplinary journal that publishes articles and reviews on timely topics written by and for a wide range of international scholars. The recent volumes on rationality are remarkable for their breadth and depth. Protosociology would be a great addition to any library. Roger Gibson, St. Louis (USA

Published Volumes

337

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research Vol. 15, 2001 On a Sociology of Borderlines: Social Process in Time of Globalization Gerhard Preyer, Mathias Bös (eds.) CONTENTS Introduction: Borderlines in Time of Globalization (Gerhard Preyer, Mathias Boes) I RECONCEPTIONALIZATIONS OF THE GLOBAL: BORDERLINES IN WORLD SOCIETY Shmuel Noah Eisenstadt The Continual Reconstruction of Multiple Modern Civilizations and Collective Identities Christopher Chase-Dunn: Globalization: A World-Systems Perspective Thomas D. Hall: World-Systems, Frontiers, and Ethnogenesis: Incorporation and Resistance to State Expansion Richard E. Lee: After History? The Last Frontier of Historical Capitalism II DEFINING BORDERLINES IN WORLD SOCIETY: THE EMERGENCE OF NEW MEMBERSHIPS Gerhard Preyer: Globalization and the Evolution of Membership Barrie Axford: Enacting Globalization: Transnational Networks and the Deterritorialization of Social Relationships in the Global System Mathias Bös Immigration and the Open Society: The Normative Patterns of Membership in the Nation State Uta Gerhardt, Birgitta Hohenester: A Transformation of National Identity? Refugees and German Society After World War II Marja Keränen: Citizenship, UniversalSpeak, and Local-Speak

III THE GLOBAL AND THE LOCAL: THE COLLAPSE AND RECONSTRUCTION OF BORDERLINES Christie Davies, Eugene Trivizas The Collapse of the National Morality and National Moral Boundaries of Small Peripheral Countries: not Globalisation but the Imposition of Liberty Walter L. Bühl: Former GDR between “Transformation” and “Social Evolution” F. Peter Wagner: Beyond “East” and “West”: On the European and Global Dimensions of the Fall of Communism Ramón Grosfoguel ‘Cultural Racism’ and ‘Borders of Exclusion’ in the Capitalist World-Economy: Colonial Caribbean Migrants in Core Zones Francisco Entrena: Socio-Economic Restructurings of the Local Settings in the Era of Globalization ON A SOCIOLOGY OF VIOLENCE Konrad Thomas Ein anderes Verständnis von Gewalt: Der gesellschaftsanalytische Beitrag des Literaturwissenschaftlers René Girard CRITICAL REVIEW: ProtoSciology, Vol. 12, 1998-Special Edition: After the Received View. Developments in the Theory of Science George N. Schlesinger

338

Published Volumes

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research

Vol. 14, 2000 Folk Psychology, Mental Concepts and the Ascription of Attitudes: On Contemporary Philosophy of Mind CONTENTS Alvin I. Goldman Folk Psychology and Mental Concepts Philip Pettit How the Folk Understand Folk Psychology Jane Heal Understanding Other Minds from the Inside Christopher S. Hill From Assertion to Belief: The Role of Linguistic Data in the Practice of Belief-Ascription David Rosenthal Content, Interpretation, and Consciousness Jay L. Garfield Thought as Language: A Metaphor Too Far Robert M. Gordon Sellars’s Ryleans Revisited Louise Röska-Hardy Self-Ascription and Simulation Theory Brian P. McLaughlin Why Intentional Systems Theory Cannot Reconcile Physicalism With Realism about Belief and Desire Gerhard Preyer Primary Reasons: From Radical Interpretation to a Pure Anomalism of the Mental

Rebecca Kukla How to Get an Interpretivist Committed David Pitt Nativism and the Theory of Content Raffaella De Rosa On Fodor’s Claim that Classical Empiricists and Rationalists Agree on the Innateness of Ideas Consuelo Preti Belief and Desire Under The Elms Erwin Rogler On David Lewis’ Philosophy of Mind Barbara Von Eckardt, Jeffrey S. Poland In Defense of the Standard View

ON CONTEMPORARY PHILOSOPHY OF LANGUAGE, ONTOLOGY AND MORAL THEORY Thomas Baldwin Nearly Logic James E. Tomberlin Logical Form, Actualism, and Ontology Arend Kulenkampff, Frank Siebelt What a Noncognitivist might tell a Moral Realist

www.protosociology.de

Published Volumes

339

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research Vol. 13 1999: Reasoning and Argumentation D. Mans, G. Preyer (eds.) CONTENTS Ralph H. Johnson: Reasoning, Argumentation and The Network Problem Leo Groarke: The Fox and the Hedgehog: On Logic, Argument, and Argumentation Theory J. Anthony Blair: Presumptive Reasoning/Argument: An Overlooked Class Robert C. Pinto: Argument Schemes and the Evaluation of Presumptive Reasoning: some Reflections on Blair’s Account Douglas Walton: The New Dialectic: A Method of Evaluating an Argument Used for Some Purpose in a Given Case Manfred Kienpointner: Comments on Douglas Walton’s Paper Christopher W. Tindale: The Authority of Testimony John Woods: Peirce’s Abductive Enthusiasms Henry W. Johnstone, Jr.: “‘Any,’ ‘Every,’ and the Philosophical Argumentum ad Hominem” ON CONTEMPORARY PHILOSOPHY Hans Lenk Interdisziplinarität und Interpretation Ellery Eells Causal Decision Theory

www.protosociology.de

340

Published Volumes

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research VOL. 12, 1998 – Special Edition AFTER THE RECEIVED VIEW – Developments in the Theory of Science Gerhard Preyer, Georg Peter, Alexander Ulfig (Eds.)

CONTENTS

in memoriam Wolfgang Stegmüller

Introduction: Developments in the Theory of Science (Gerhard Preyer, Georg Peter, Alexander Ulfig) LOGICAL OPERATIONALISM – SIGNIFICANCE AND MEANING Wilhelm K. Essler: Truth and Knowledge. Some Considerations concerning the Task of Philosophy of Science Gerhard Preyer: The Received View, Incommensurability and Comparison of Theories – Beliefs as the Basis of Theorizing Robert Schwartz: Reflections on Projection Jeffrey E. Foss: The Logical and Sociological Structure of Science STRUCTURALISM – MEANINGFUL MEASUREMENT – THE CONCEPTION OF PHYSICAL LAW C. Ulises Moulines: Structuralism vs. Operationalism Nicholas Rescher: Meaningless Numbers R. I. G. Hughes: Laws of Nature, Laws of Physics, and the Representational Account of Theories James R. Brown: Einstein’s Principle Theory INDUCTIVE INFERENCES – INTERPRETATION OF PROBABILITY – GAME THEORY Kevin T. Kelly, Cory Juhl Transcendental Deductions and Universal Architectures for Inductive Inferences

Howard H. Harriott: R.A. Fisher and the Interpretation of Probability Brian Skyrms: Evolution of an Anomaly PROPERTIES – UNDERDETERMINATION – SCIENTIFIC REALISM George N. Schlesinger: Degrees of Characterizations Carl A. Matheson: Observational Adequacy as distinct from the Truth about Observables Thomas R. Grimes: Scientific Realism and the Problem of Underdetermination Paul C. L. Tang: On Paul Churchland’s Treatment of the Argument from Introspection and Scientific Realism RATIONALITY – METAPHORS – VALUES IN SCIENCE David Resnik: Scientific Rationality and Epistemic Goals Aldo Montesano: Rationality in Economics: A General Framework Joseph Agassi: Science Real and Ideal: Popper and the Dogmatic Scientist Michael Bradie: Models and Metaphors in Science David Gruender: Values and the Philosophy of Science

341

Published Volumes

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research

Vol. 11, 1998 Cognitive Semantics II – Externalism in Debate CONTENTS RADICAL INTERPRETATION, ONTOLOGY AND THE STRUCTURE OF LANGUAGE Gerhard Preyer, Michael Roth: On Donald Davidson’s Philosophy: An Outline Richard Manning: All Facts Great and Small Barbara Fultner: Of Parts and Wholes: The Molecularist Critique of Semantic Holism Louis Goble: Re-Evaluating Supervaluations David Simpson: Interpretation and Skill: On Passing Theory Wulf Kellerwessel: Katz on Semantics and Pragmatics

EXTERNALISM

AND THE INDIVIDUATION OF

CONTENT Ron Wilburn: Knowledge, Content, and the Wellstrings of Objectivity Anthony Brueckner: Content Externalism and A Priori Knowledge Consuelo Preti: The Irrelevance of Supervenience Michael Liston: Externalist Determinants of Reference Arnold Silverberg: Semantic Externalism. A Response to Chomsky Gerhard Preyer: Interpretation and Rationality: Steps from Radical Interpretation to the Externalism of Triangulation

www.protosociology.de

342

Published Volumes

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research

Vol. 10, 1997 Cognitive Semantics I – Conceptions of Meaning

CONTENTS MEANING, TRUTH AND THE USE OF LANGUAGE Herman Cappelen, Ernie Lepore: Semantic Theory and Indirect Speech Kirk Ludwig: The Truth about Moods Louise Röska-Hardy: Language, Use and Action Jeffrey King: The Source(s) of Necessity Filip Buekens: The Genesis of Meaning (a Myth) REFERENCE, INDEXICALS AND SPEAKER MEANING Robert Hanna: Extending Direct Reference Peter Ludlow: Semantics, Tense, and Time: a Note on Tenseless Truth-Conditions for Token-Reflexive Tensed Sentences Gerhard Preyer: Verstehen, Referenz, Wahrheit. Zur Philosophie Hilary Putnams

Reinaldo Elugardo: Descriptions, Indexicals and Speaker Meaning RATIONAL DIALETISM OF ANTINOMIES Mark Sainsbury: Can Rational Dialetheism Be Refuted By Considerations about Negation and Denial? CONTEMPORARY DEBATES IN THE PHILOSOPHY OF MIND Klaus Sachs-Hombach: Philosophy of Mind: Die Simulationstheorie IN RETROSPECT Joseph Agassi: Wittgenstein – The End of a Myth Kent A. Peacock, Richard Feist: The Einstein-DeSitter Controversy

www.protosociology.de

Published Volumes

343

Forthcoming

PROTOSOCIOLOGY An International Journal of Interdisciplinary Research PARSONS-SPECIAL EDITION CLASSICAL SOURCES

AND

CURRENT USES

OF

TALCOTT PARSONS’S THOUGHT

edited and with an introduction by Bernard Barber and Uta Gerhard This volume wishes to cut across the vast array of misunderstandings, misreadings, and misinterpretations which have marred the proper recognition of Parsons’s work for the last thirty years. If the question is asked why such recognition would be important, the answer is twofold. For one, Parsons’s work is a classic that reaches from the l930s into the l970s of the twentieth century while sociology’s now arguably uncontroversial classics like Max Weber, or Emile Durkheim - were only able to analyze society until about the end of World War I. Therefore, Parsons’s oeuvre has a special modernity in that his work takes into account the New Deal, National Socialism, McCarthyism, and the Student Rebellion, to name but a few timely events which are part and parcel of Parsons’s sociological analysis. Furthermore, his conception of social order and anomie, especially since they seriously discuss the problematic viability of utilitarian economics, are of particular importance today. Introductionary Essay: The Parsons Agenda Bernard Barber, Uta Gerhardt PART 1: Revisiting The Structure of Social Action Mark Gould: Neoclassical Economics, Social Order, and “The Structure of Social Action” Uta Gerhardt: National Socialism and the Politics of “The Structure of Social Action” Richard Münch: The Problem of Order: Sixty Years After “The Structure of Social Action” PART 2: Understanding Parsons’s Thought Bryan S. Turner: Sacricity and Solidarity in Talcott Parsons’s Economic Sociology Edward A.Tiryakian: An Emergent French Connection: Revisiting Parsons’s Durkheim Bernard Barber: The Nature and Functions of Religion and Beyond PART 3: Evaluating the Scope of Sociology Daniel Bell: The Limits of General Theory in the Social Sciences Hans Jürgen Puhle: (Title to be announced)

344

Bookpublications

LOGICAL FORM

AND

LANGUAGE

Gerhard Preyer, Georg Peter (eds.) To Donald Davison This volume brings together some of the world’s most distinguished linguistically-minded philosophers and philosophically-minded linguists in an outstanding collection of new papers on what logical form is supposed to be and why it matters for the study of language. - Zoltan Szabo, Cornell University CONTENTS Gerhard Preyer, Georg Peter: Introduction PART I: THE NATURE

LOGICALFORM

Stephen Neale Logically Unformed

Norbert Hornstein A Grammatical Argument for A Davidsonian Semantics

Ernie Lepore, Kirk Ludwig What is Logical Form

Jason Stanley Nominal Restriction

OF

Paul M. Pietroski Function and Concatenation Jeffrey King Two Sorts of Claim about “Logical Form” Peter Ludlow LF and Natural Logic. The Syntax of Directional Entailing Environments Robert May/R.F. Fiengo Identity Statements PART II: INTENSIONALITY, EVENTS SEMANTIC CONTENT

AND

James Higginbotham Why is Sequence of Tens Obligatory? Richard Larson The Grammar of Intensionality Barry Schein Events and the Semantic Content of Thematic Relations

PART III: LOGICAL FORM, BELIEF ASCRIPTION PROPER NAMES

AND

Bernard Linsky Russell’s Logical Form, LF and Truth Conditions Lenny Clapp, Robert Stainton “Obviously Propositions are Nothing.” Russell and the Logical Form of Belief Reports Robert Matthews Logical Form and the Relational Conception of Belief Marga Reimer Ordinary Proper Names Reinaldo Elugardo The Predicate View of Proper Names

Oxford University Press, Oxford 2002 www.protosociology.de/Books.htm

345

Bookpublications

The Contextualization of Rationality Gerhard Preyer, Georg Peter (eds.)

Introduction: Problems, Concepts and Theories of Rationality RADICAL INTERPRETATION, NATURALIZED EPISTEMOLOGY

AND

NORMATIVITY

John Heil The Propositional Attitudes

Roger F. Gibson Stich on Intentionality and Rationality

David K. Henderson Epistemic Rationality, Epistemic Motivation, and Interpretive Charity

Paul K. Moser and David Yandell Against Naturalizing Rationality Harvey Siegel Naturalism, Instrumental Rationality, and the Normativity of Epistemology

Gerhard Preyer Interpretation and Rationality: Steps from Radical Interpretation to the Externalism of Triangulation INTENTIONS

AND THE

SOCIAL ASPECT

OF

RATIONALITY

Alfred R. Mele Rational Intentions and the Toxin Puzzle

Raimo Tuomela Rational Cooperation and Collective Goals

Peter Gärdenfors The Social Stance

Peter French Rationality and Ethics

Timo Airaksinen and Katri Kaalikoski Instrumental Rationality

Julian Nida-Rümelin The Plurality of Good Reasons and the Theory of Practical Rationality

CONCEPTS

OF

EXPLANATION, JUSTIFICATION

AND

REALITY

Philip Pettit Three Aspects of Rational Explanation

Alexander Ulfig Validity, Justification and Rationality

Keith Lehrer Rationality and Trustworthiness

Nicholas Rescher Reason and Reality

250 Seiten mentis Verlag, Paderborn 2000

346

Bookpublications

Analytische Ästhetik Eine Untersuchung zu Nelson Goodman und zur literarischen Parodie Georg Peter I. ANALYTISCHE ÄSTHETIK:

DIE

ZEICHENTHEORIE NELSON GOODMANS

1. Eine ‚andere‘ Fragestellung / 2. Zu Nelson Goodman in allgemeiner Absicht / 3. Zeichenbedeutung: Sprachliche und nichtsprachliche Denotation / 4. Von der Probe zum passenden Etikett: Exemplifikation und Ausdruck / 5. Die Nebenbeschäftigung der Symbole: Ein bedeutungstheoretischer Exkurs zu Metapher und Wahrheit / 6. Ästhetische Bedeutung und kognitive Funktion: Das Kunstwerk als Probe / 7. ‚Wann‘ oder doch ‚Was ist Kunst?‘ II. SCHEMAINTERPRETATION

UND DIE IDENTIFIZIERBARKEIT ÄSTHETISCHER

ZEICHEN

1. Schemainterpretationismus und Symboltheorie: Vom Symbol zum Schema / 2. Der konstruktive Schemainterpretationismus von Hans Lenk / 3. Hermeneutik und Schemainterpretation / 4. Die soziale Verankerung der Interpretation: Ästhetische Kompetenz und Kultursoziologie / 5. Thesenhaftes zu einer schemaorientierten Symboltheorie: Wann ist (es) Parodie III. ROBERT NEUMANNS PARODIEN: DIE PROBE

AUF DIE

EXEMPLIFIKATION

1. Literaturwissenschaft und Parodie: eine vorbelastete Beziehung / 2. Die Parodie als eine Form des Komischen / 3. Das Problem der Wertung: Die kritische und die (bloß) komische Parodie / 4. Nachahmung, Bezugnahme und Robert Neumanns Theorie der Parodie / 5. Von der Funktion der Zeichen zur Funktion der Parodie: Goethe und Brecht parodiert / 6. Komplexität der Parodie und die komplexen Parodien Robert Neumanns / 7. Zeigen und Sagen: Funktion und Qualität der Parodie anhand einer Benn-Parodie / 8. Die Grenzen der Parodie und der Parodierbarkeit / 9. Parodien und andere Formen der Nachahmung IV. ZUR METHODIK

DES

ANSATZES

UND SEINER

WEITERFÜHRUNG

ca. 330 Seiten, Erschienen in der Reihe: PHILOSOPHISCHE ANALYSE / PHILOSOPHICAL ANALYSIS, Herausgegeben von / Edited by Herbert Hochberg · Rafael Hüntelmann et al., Deutsche Bibliothek der Wissenschaften, Dr. Hänsel-Hohenhausen AG, Frankfurt a.M., München, New York. www.protosociology.de/Books.htm

347

Bookpublications

Borderlines in a Globalized World New Perspectives in a Sociology of the World System Gerhard Preyer, Mathias Bös (eds.) To Walter L. Bühl Introduction Gerhard Preyer, Mathias Bös Borderlines in Time of Globalization: New Theoretical Perspectives

I RECONCEPTIONALIZATIONS

OF THE

GLOBAL: BORDERLINES

IN

WORLD SOCIETY

Shmuel N. Eisenstadt: The Continual Reconstruction of Multiple Modern Civilizations and Collective Identities Christopher Chase Dunn: Globalization: A World-Systems Perspective Thomas D. Hall: World-Systems, Frontiers, and Ethnogenesis: Incorporation and Resistance to State Expansion. Richard Lee: After History? The Last Frontier of Historical Capitalism

II DEFINING BORDERLINES

IN

WORLD SOCIETY: THE EMERGENCE

OF

NEW MEMBERSHIPS

Gerhard Preyer: Globalization and the Evolution of Membership Barrie Axford: Enacting Globalization - Transnational Networks and the Deterritorialisation of Social Relationship Mathias Bös: Immigration and the Open Society: The Normative Patterns of Membership in the Nation State Uta Gerhardt, Birgitta Hohenester: A Transformation of National Identity? Refugees and German Society after World War II

III. THE GLOBAL

AND THE

LOCAL : THE COLLAPSE

AND

RECONSTRUCTION

OF

BORDERLINES

Christie Davies, Evgenios Trivizas: The Collapse of the National Morality and National Moral Boundaries of Small Peripherial Countries not Globalization but the Imposition of Liberty F. Peter Wagner: Beyond “East” and “West”: On the European and Global Dimensions of the Fall of Communism Francisco Entrena:Socioeconomic Restructurings of the Local Settings in the Era of Globalization Kluwer Academic Publishers Social Indicators Research Series Vol. 9 Dordrecht, Netherlands

348

Bookpublications

PROTOSOZIOLOGIE IM KONTEXT “Lebenswelt” und “System” in Philosophie und Soziologie Gerhard Preyer, Georg Peter, Alexander Ulfig (Hrsg.) in memoriam Paul Lorenzen Einleitung: “Lebenswelt” und “System” in Philosophie und Soziologie, Gerhard Preyer, Georg Peter, Alexander Ulfig

ZUM BEGRIFF

DER

LEBENSWELT

Ernst W. Orth: Lebenswelt als unvermeidliche Illusion? Husserls Lebensweltbegriff und seine kulturpolitischen Weiterungen, Walter Biemel: Gedanken zur Genesis der Lebenswelt, Alexander Ulfig: Lebenswelt und Reflexion. Anhang: Lebenswelt als Fundament der Wissenschaft, Gerhard Preyer: Hintergrundwissen: Kritik eines Begriffs, Hubert A. Knoblauch: Soziologie als strenge Wissenschaft? Phänomenologie, kommunikative Lebenswelt und soziologische Methodologie

LEBENSWELT – BEGRÜNDUNG – WISSENSCHAFT Jürgen Mittelstraß: Das lebensweltliche Apriori, Peter Janich: Die Rationalität der Naturwissenschaften, Jürgen Mittelstraß: Rationalität und Reproduzierbarkeit, Elisabeth Ströker:Lebenswelt durch Wissenschaft, Paul Janssen: Lebenswelt, Wissen und Wissenschaft, Richard T. Murphy: E. Husserl’s Phenomenology of Reason LEBENSWELT/LEBENSFORM – SPRACHE Pierre Kerszberg: Lifeworld and Language, John F.M. Hunter: The Motley Forms of Life in the Later Wittgenstein, Peter A. French: Why did Wittgenstein read Tagore to the Vienna Circle? Georg Peter: Die Nebenbeschäftigung der Symbole. Zu Wahrheit und Funktion der Metapher

SYSTEM – SOZIALSYSTEM – GESELLSCHAFT Niklas Luhmann: Die Lebenswelt nach Rücksprache mit Phänomenologen, Niklas Luhmann: Observing Re-entries, Gerhard Preyer: System-, Medien- und Evolutionstheorie. Zu Niklas Luhmanns Ansatz, Richard Münch: Autopoesis per Definition, Hans Zitko: Codierungen der Kunst: Zur Kunstsoziologie Niklas Luhmanns, James Bohman: The Completeness of Macro-Sociological Explanations: System and Lifeworld, Göran Ahrne: Outline of an Organisational Theory of Society, Anhang: Karl Otto Hondrich: Zu Göran Ahrnes Ansatz

392 Seiten Verlag Königshausen & Neumann, Würzburg 1996 Digitale Neuauflage bei: Humanities Online, Frankfurt am Main2001 www.humanities-online.de

349

Bookpublications

INTENTION – BEDEUTUNG – KOMMUNIKATION Kognitive und handlungstheoretische Grundlagen der Sprachtheorie Gerhard Preyer, Maria Ulkan, Alexander Ulfig (Hrsg.) Einleitung: Zu kognitiven und handlungstheoretischen Grundlagen der Sprachtheorie, Gerhard Preyer, Maria Ulkan, Alexander Ulfig I INTENTIONEN

UND KOMMUNIKATIVE

HANDLUNGEN

Maria Ulkan: Kommunikative und illokutionäre Akte; Georg Meggle/Maria Ulkan: Grices Doppelfehler. Ein Nachtrag zum Griceschen Grundmodell; Jan Nuyts: Intentionalität und Sprachfunktionen II INTERPRETATION UND BEDEUTUNG Gerhard Preyer: Kognitive Semantik, Anhang Sprechaktsemantik: J.L. Austin, J.R. Searle, H.P. Grice, P.F. Strawson; Louise Röska-Hardy: Sprechen, Sprache, Handeln; Frank Siebelt: Zweierlei Holismus. Überlegungen zur Interpretationstheorie Donald Davidsons; Peter Rothermel: Semantische Implikaturen; Volkmar Taube: Referenz und Interpretation. Zur Theorie nichtsprachlicher Symbolisierung; Georg Peter: Zu Richtigkeit und Interpretation der Metapher: Kognitive Funktion und rekonstruktive Schemainterpretation III KLASSIFIKATION

VON

SPRECHAKTEN

Maria Ulkan: Informations- und Aufforderungshandlungen; Dirk Hartmann: Konstruktive Sprechakttheorie; Volkmar Taube: Bildliche Sprechakte IV KOMMUNIKATIVES HANDELN

UND INTERSUBJEKTIVE

GÜLTIGKEIT

Jürgen Habermas: Sprechakttheoretische Erläuterungen zum Begriff der kommunikativen Rationalität; Karl-Otto Apel: Illokutionäre Bedeutung und normative Gültigkeit. Die transzendentalpragmatische Begründung der uneingeschränkten kommunikativen Verständigung; Peter-Paul König: Kommunikatives und strategisches Handeln. Kritische Bemerkungen zu zwei zentralen Begriffen der “Theorie des kommunikativen Handelns” von Jürgen Habermas; Alexander Ulfig: Präsuppositionen und Hintergrundwissen. Eine Kritik am formalpragmatischen Präsuppositionsbegriff V DIALOGSTRUKTUR

UND

ARGUMENTATION

Wilhelm Franke: Konzepte linguistischer Dialogforschung; Franz Hundsnurscher: Streitspezifische Sprechakte: Vorwerfen, Insistieren, Beschimpfen; Dieter Mans: Argumentation im Kontext, Exkurs: Zu Christoph Lumers “Praktische Argumentationstheorie” 408 Seiten Westdeutscher Verlag, Opladen 1997 Digitale Version: Humanities Online 2001

350

Bookpublications

STRUKTURELLE EVOLUTION

UND DAS WELTSYSTEM

Theorien, Sozialstruktur und evolutionäre Entwicklungen Gerhard Preyer (Hrsg.) Einleitung: Gerhard Preyer: Strukturelle Evolution und das Weltsystem: Theorien, Sozialstruktur und evolutionäre Entwicklungen ZENTRUM

UND

PERIPHERIE –

INSTITUTIONELLE

ENTWICKLUNG –

ASKRIPTIVE

SOLIDARITÄT

Shmuel Noah Eisenstadt: Social Division of Labor, Construction of Centers and Institutional Dynamics: A Reassessment of the Structural–Evolutionary Perspective, Bernhard Giesen, Kay Junge: Strukturelle Evolution, Gerhard Preyer: Mitgliedschaftsbedingungen. Zur soziologischen Kerntheorie einer Protosoziologie, Anhang: Die modernen Gesellschaften verstehen. Richard Münchs Entwicklungstheorie moderner Gesellschaften, Erwin Rogler, Gerhard Preyer: Relationslogische Darstellung der sozialen Gesetze, Gerhard Preyer: Die modernen Gesellschaften “verstehen”. Zu Richard Münchs Entwicklungstheorie moderner Gesellschaften, Dieter Claessens: Bemerkungen zur Entstehung der modernen Ökonomie: Das Organistionsproblem, Richard Pieper: Strukturelle Emotionen, elementare Strukturbildung und strukturelle Evolution DIE EVOLUTION

POLITISCHER

ORDNUNGEN

Rainer C. Baum: Parsons on Evolution of Democracy, Mathias Bös: Zur Evolution nationalstaatlich verfaßter Gesellschaften, Konrad Thomas: Das Ethnische und das Staatliche, Volker Bornschier: Die westeuropäische Integration als Gesellschaftsmodell im Zentrumswettbewerb ZUR SOZIOLOGIE

DES

WELTSYSTEMS

Immanuel Wallerstein: Evolution of the Modern World-System, Christopher ChaseDunn, Thomas D. Hall: The Historical Evolution of World-Systems: Iterations and Transformations, Albert Bergesen: Postmodernism: A World System Explanation, Richard Münch: Modernity and Irrationality: Paradoxes of Moral Modernization, Walter L. Bühl: Transformation oder strukturelle Evolution? Zum Problem der Steuerbarkeit von sozialen Systemen STATE

OF THE

ART: Michael Schmid: Soziologische Evolutionstheorien

Suhrkamp Verlag, Frankfurt am Main 1998, Suhrkamp Taschenbuch Wissenschaft

Bookpublications

LEBENSWELT – SYSTEM – GESELLSCHAFT Konstruktionsprobleme der “Theorie des kommunikativen Handelns” von Jürgen Habermas Gerhard Preyer I DIE ENTWICKLUNGSLOGIK VON WELTBILDERN 1. Die Rationalisierung von Weltbildern 2. Strategien der Analyse von Weltbildern 3. Folgeprobleme und Kritik II GESELLSCHAFT ALS LEBENSWELT UND SYSTEM 1. Die Strukturen der Lebenswelt 2. Allgemeine Bezugsprobleme der soziologischen Evolutionstheorie 3. Die Verständigungsformen 1. Weltbilder und soziale Integration 2. Zur Durkheim-Interpretation 4. Kommunikationsmedien und generalisierte Kommunikationsweisen 5. Folgeprobleme und Kritik III DIE FORMAL-PRAGMATISCHE BEDEUTUNGSTHEORIE 1. Die sprechakttheoretische Grundlegung 2. Interpersonal geltende Bedeutungskonventionen 3. Regelbewußtsein und Handlungskompetenz 4. Die Expansion des semantischen Gehalts 5. Verständigung und die Herstellung interpersonaler Beziehungen IV DER ERWERB DES MORALISCHEN BEWUßTSEINS 1. Die sozial-kognitive Grundausstattung 2. Diskurs und moralisches Bewußtsein 3. Folgeprobleme V KONSTRUKTIONSPROBLEME UND KRITIK 1. Rekonstruktionshypothesen 2. Zu den Konstruktionsproblemen ANHANG: Max Webers Religionssoziologie als eine Typologie des Rationalismus 290 Seiten digital bookpublication: Verlag Humanities Online, Frankfurt am Main www.humanities-online.de

351

352

Bookpublications

Die globale Herausforderung Wie Deutschland an die Weltspitze zurückkehren kann Gerhard Preyer Der Begriff Globalisierung hat bereits Schlagwortcharakter bekommen und ist in aller Munde. Globale Märkte haben die Grundsituation wirtschaftlichen und politischen Handelns grundsätzlich verändert. Das heute entstehende globale Weltsystem und der Medienverbund auf dem es beruht, fordert uns alle heraus. Was bedeutet Globalisierung? Worin bestehen die globalen Herausforderungen, die zu bewältigen sind? Darauf wird eine Antwort gegeben. NEUE MEDIEN, WISSENSCHAFT

UND

TECHNIK

IM

ZEITALTER

DER

GLOBALISIERUNG

Was heißt Globalisierung? – Die neuen Medien: Eine Kopernikanische Wende – Der Überlebensimperativ: Technologieentwicklung – Die Zukunft der deutsche Universitäten VERÄNDERTE KONSTELLATIONEN Globale Wirtschaft und globale Ordnung – Die Globalisierung der Finanzmärkte – Auf dem Weg zur virtuellen Organisation – Das Ende des klassischen Arbeitsmarktes – Zu einer neuen Wirtschaftspolitik EUROPA

IM

ZEITALTER

DER

GLOBALISIERUNG

Zur Ausgangssituation im Zentrumswettbewerb – Frankreichs Zentralismus: Grenzen der Marktwirtschaft – Das Überlebenssystem Italien – Großbritannien zwischen Tradition und Modernisierung – Deutsche Stärken im Umbruch – Neu gemischte Karten – Die Wiedergewinnung des Standorts Deutschland DIE EVOLUTION

DES

MITGLIEDSCHAFTSCODES

Zur Soziologie der Grenzziehungen – Gesellschaft, Organisation und Interaktion – Gesellschaftsinterne Globalisierung und Entwicklungstrends – Die globalisierte Gesellschaft

290 Seiten Frankfurter Allgemeine Zeitung/Gabler Edition, Frankfurt am Main 1998

Bookpublications

353

INTEGRIERTES MANAGEMENT Was kommt nach der Lean-Production? Gerhard Preyer/Jakob Schissler Es gibt heute im Weltmaßstab kein Unternehmen, das seine Umstrukturierung nicht an den Bestandteilen des japanischen Modells der schlanken Produktion und des schlanken Managements, der Lean-Modelle, orientiert. Die Lean-Modelle sind zu einem integrierten Management fortzuentwickeln: dem über Vernetzung integrierten segmentierten Unternehmen, mit durchlässigen, ständig wechselnden Grenzen zwischen Unternehmen, Lieferanten und Kunden. Integriertes Management ist das Unternehmensmodell, das eine erhöhte Qualifizierung der Mitarbeiter und die Fähigkeit, in sich selbst steuernden Teams mitzuarbeiten, erfordert. Das bedeutet aber auch die Entwicklung eines neuen Führungsstils. Mit integriertem Management kann jedes Unternehmen sofort anfangen. Dazu wird in diesem Buch ein Weg gezeigt. Es wird eine wichtige Funktion als Impuls für weitere Entwicklungen haben. Inhalt DIE NEUEN HERAUSFORDERUNGEN 1. Vom japanischen und amerikanischen Management lernen? 2. Information, Wissen und Entscheidungen KREATIVITÄT – ORGANISATION – FÜHRUNG 1. Managementfähigkeiten 2. Innovative Organisation 3. Management und Führungsstile STRATEGIE – MARKT – KULTUR 1. Unternehmensstrategie 2. Controlling 3. Marketing 4. Zur Unternehmenskultur STICHWORT Was kommt nach der Lean Production? Frankfurter Allgemeine Zeitung, Verlagsbereich Wirtschaftsbücher, BlickbuchWirtschaft, Frankfurt am Main 1996, 164 Seiten, 32,- DM

354

Bookpublications

ÜBER DAS KÄMPFEN Zum Phänomen des Kampfes in Sport und Gesellschaft Axel Binhack Kämpfen ist stets als ein Beziehungsphänomen, gleichsam als “Fortsetzung der Kommunikation mit anderen Mitteln” zu verstehen. Anhand von sechs Strukturmerkmalen wird eine “Prototypik des Kampfes” entwickelt. Ihre zusammenhängende Dynamik ist gesellschafts- und kulturunabhängig in vielfältigen Konfliktformen zu erkennen. In dem vorgelegten interdisziplinären Anschnitt wird die Struktur der Kampfbeziehung von vergleichbaren Sozialphänomenen abgegrenzt und am Beispiel des Kampfsports in Form des “Zweikampfes” analysiert. 1. Merkmale einer Formalstruktur des Kampfes Der duale Beziehungsaspekt: Antagonismus und Ambivalenz – Der duale Trägeraspekt: Entscheidungsorientiertheit und tendenzielle energetische Totalität – Der duale Inhaltsaspekt: Zweckgerichtetheit und riskante Offenheit 2. Die Strukturmerkmale des Kampfes im Vergleich mit Arbeit, Spiel und den Beziehungsphänomenen Tausch, Flirt und Macht Der “antagonistische” Unterschied – Freiheit und Zweckgerichtetheit – “Zeitenthobene” innere Unendlichkeit und “zeitbeschleunigende” Entscheidungsorientiertheit – Die formalstrukturelle “Folgenhaftigkeit” des Kampfes 3. Zum Kampfsport Kampfsport als Radikalisierung des sportlichen Kampfes – Der Kampfsport als kulturelles Phänomen – Der Kampfdiskurs als Symbol männlichen Agierens – Ästhetische Perspektiven 4. Die Kampfsportart Karate-Do als Beispiel Historische Entwicklungslinien des Karate-Do – Der Schwertkampf und die “Leere Hand”: Zum karatespezifischen Aspekt der Waffenlosigkeit – Die Kultivierung des Kampfphänomens im Karate-Do – Der Kampf als Lehrmeister im Sinne des Zen Campus Verlag, Campus Forschung Bd. 768, Frankfurt am Main 1998

355

Bookpublications

System der Rechte, demokratischer Rechtsstaat und Diskurstheorie des Rechts nach Jürgen Habermas hrsg. von Werner Krawietz / Gerhard Preyer Editorial: Zivilgesellschaftliche Assoziationen und spätmoderner Rechtssstaat (Werner Krawietz)

I JURISTISCHE ENTSCHEIDUNG, BEGRÜNDUNG

UND

DISKURSIVER

RECHTFERTIGUNG

DES

RECHTS

IN DER

PERSPEKTIVE

REFLEXIONSTHEORIE

Enrique P. Huba: Standortbestimmung in der zeitgenössischen Rechtstheorie - Rawls, Dworkin, Habermas, Alexy und andere Mitglieder der modernen ‘Heiligen (Rede-) Familie’, Thomas McCarthy: Legitimacy and Diversity: Dialectical Reflexions on Analytical Distinctions, Gerhard Preyer: Rechtsgeltung – Argumentation – Entscheidung II ORDNUNG, RECHT

UND

RATIONALITÄT

IM JURISTISCHEN

DISKURS

Karl-Heinz Ladeur: Rechtliche Ordnungsbildung unter Ungewißheitsbedingungen und intersubjektive Rationalität, Ulfrid Neumann: Zur Interpretation des forensischen Diskurses in der Rechtsphilosophie von Jürgen Habermas, Ota Weinberger: Diskursive Demokratie ohne Diskurstheorie III KONSENSUSTHEORIE

SUBJEKTPHILOSOPHIE? VORAUSSETZUNGEN UND FOLGEN DISKURSES IM DEMOKRATISCHEN RECHTSTAAT

VERSUS

JURISTISCHEN

DES

Michael Pawlik: Die Verdrängung des Subjekts und ihre Folgen. Begründungsdefizite in Habermas’ “System der Rechte”, Uwe Steinhoff: Probleme der Legitimation des demokratischen Rechtsstaats, William Rehg: The Place of Consensus in Democratic Legitimation: A Recommendation

Rechtstheorie Zeitschrift für Logik, Methodenlehre, Normentheorie und Soziologie des Rechts 3 1996, Habermas-Sonderheft (erschienen 1998) Duncker und Humblot, Berlin

356

Bookpublications

Wittgensteins Spätphilosophie Analysen und Probleme

Wulf Kellerwessel, Thomas Peuker (Hrsg.) Die Spätphilosophie Ludwig Wittgensteins entfaltet nicht nur in der gegenwärtigen Sprachphilosophie ihre Wirkung, sondern auch zunehmend über den Bereich der Sprachphilosophie hinaus. Zwar macht dieser einen Schwerpunkt des Wittgensteinschen Denkens aus, aber dennoch verdanken die Philosophie des Geistes, die Erkenntnistheorie und die Philosophie der Religion der Wittgensteinschen Sprachphilosophie wichtige Impulse. Diese Schwerpunktsetzung zugunsten der Sprachphilosophie wie auch ihre Zusammenhänge mit den genanten anderen Disziplinen der Philosophie zeichnet der vorliegende Band aus.

INHALT W. Kellerwessel, T. Peuker: Einleitung, A. Ofsti: Methodischer Solipsismus, Metasprachen(problem) und Dexis. Einige Überlegungen zum Privatsprachenproblem, T. Peuker: Die faktische Öffentlichkeit der Sprache. Zu Wittgensteins Privatsprachenargument, P. Niesen: Gemeinschaft, Normativität, Praxis: Zur Debatte über Wittgensteins Regelbegriff, A. Berndzen: Einer Regel entsprechen – einer Regel folgen. Zu einer Kontroverse in der Interpretation von Wittgensteins ”Philosophischen Untersuchungen”, G. Preyer: Sprachbedeutung ohne Regelbefolgung, C. Stetter: Sprache und Schrift bei Wittgenstein, H.J. Schneider: Mentale Zustände als metaphorische Schöpfungen, W. Kellerwessel: Zum Begriff der Gewißheit in Wittgensteins ”Über Gewißheit”. Ein Kommentar, M. Kroß: “Glaube Du! Es schadet nicht”. Ludwig Wittgensteins Vermischte Schriften zur Religion - Anhang. 302 Seiten

Verlag Königshausen & Neumann, Würzburg

Bookpublications

Donald Davidsons Philosophie Von der radikalen Interpretation zum radikalen Kontextualismus Gerhard Preyer Für Bruce Aune EINLEITUNG I RADIKALE INTERPRETATION, LOGISCHE FORM UND EREIGNISSE 1. Donald Davidsons Philosophie: Ein Überblick 2. Wahrheit, Bedeutung und radikale Interpretation 2.1 Radikale Interpretation als ein radikaler Externalismus (1) Der Zirkel zwischen Überzeugung und Bedeutung: Die Asymmetrie von RI (2) Die Strukur von Inhaltssätzen: sagen, daß ... (3) Der Grundsatz der Nachsicht (4) Die Autonomie der Bedeutung (5) Radikaler Externalismus: Der innertheoretische Schritt (6) Die Ontologie von RI 2.2 Von der Idiolekttheorie zum dritten Dogma des Empirismus (1) Ausgangs- und Übergangstheorien der Interpretation (2) Die Demontage eines Mythos (3) Die epistemischen Beschränkungen von Verstehen 3. Die logische Form von Handlungssätzen und die singuläre Kausalaussage 3.1 Logische Form und adverbiale Modifikation 3.2 Kausale Beziehungen 4. Wahrheit und Überzeugung 4.1 Zum Haupteinwand 4.2 Zur empirischen Offensichtlichkeit von Einstellungszuschreibung 4.3 Rationalität als normativer Begriff II KÖRPERBEWEGUNGEN UND HANDLUNGEN 1. Das logische Verknüpfungsargument 2. Primäre Gründe und die Identitätsthese: Die synkategormatische Fassung von Absichten 3. Basisakte 4. Flucht vor den Körperbewegungen 4.1 Thalbergs Handlungstheorie und der “Akkordeon-Effekt” 4.2 A.I. Goldmans Kritik an der Identitätsthese

357

358

Bookpublications

4.3 H.L.A. Hart: Zuschreibungen 4.4 Körperbewegungen als Bestandteil von Handlungen III EINE RADIKALE THEORIE DES HANDELNS 1. Der Begriff der Einstellungsrationalität 1.1 Die Homogenität der Interpretation 1.2 Bewertende Einstellungen 2. Handlungsbeschreibungen und Handlungsverursachung 3. Eine Handlungserklärung 3.1 Überzeugungen, Absichten, Situationen 3.2 Praktische Gedanken 3.3 Handlungsgründe 3.4 Zu Hempels Ansatz 4. Praktische Schlüsse 4.1 Zur Gültigkeit praktischen Schließens 4.2 Entscheidungen und die Ausführung von Absichten 4.3 Überzeugungen und willentliche Handlungen 5. Radikaler Kontextualismus Literatur 290 Seiten

Verlag Humanties Online, Frankfurt am Main www.humanities-online.de

359

Bookpublications

LANGUAGE, MIND

AND

EPISTEMOLOGY

On Donald Davidson’s Philosophy Gerhard Preyer, Frank Siebelt, Alexander Ulfig (eds.)

Introduction: On Donald Davidson’s Philosophy Alexander Ulfig)

(Gerhard Preyer, Frank Siebelt,

PHILOSOPHY OF LANGUAGE Jerry Fodor, Ernie Lepore (New Brunswick, USA): Meaning, Holism, and the Problem of Extensionality; Olav Gjelsvik (Oslo, Norway): Davidson’s Use of Truth in Accounting for Meaning; Wilhelm K. Essler (Frankfurt/Main, Germany): Was ist Wahrheit? Arend Kulenkampff (Frankfurt/Main, Germany): Eigennamen und Kennzeichnungen EPISTEMOLOGY Roger F. Gibson (St. Louis, USA): Quine and Davidson: Two Naturalized Epistemologists; Eva Picardi (Bolongna, Italy): Davidson and Quine on Observation Sentences; Ralf Naumann (Düsseldorf, Germany): Events and Externalism; Dorit Bar-On (Chapel Hill, USA): Conceptual Relativism and Translation; David K. Henderson (Memphis, USA): Conceptual Schemes after Davidson; Frank Siebelt (Frankfurt/Main, Germany): Singular Causal Sentences and two Relational Views PHILOSOPHY OF MIND AND THEORY OF ACTION Louise M. Antony (Raleigh, USA): The Inadequacy of Anomalous Monism as a Realist Theory of Mind, Louise Röska-Hardy (Darmstadt, Germany): Internalism, Externalism and Davidsons Conception of the Mental, Marcia Cavell (Berkeley, USA): Dividing the Self, Ralf Stoecker (Bielefeld, Germany): Willensschwäche - Wie ist das nur möglich? Klaus Puhl (Graz, Austria): Davidson on intentional Content and Self-Knowledge, Johannes Brandl (Salzburg, Austria): Sharing Beliefs and the Myth of the Subjective, Kirk Ludwig (Gainesville, USA): First Person Knowledge and Authority, Gerhard Preyer (Frankfurt/Main, Germany): Rationalität: Absichten, Primärgründe und praktisches Denken. Donald Davidson: Dialectic and Dialogue 445 Seiten Kluwer Academic Publishers, Synthese Library, Dordrecht

360

Bookpublications

Reality and Humean Supervenience Essays on the Philosophy of David Lewis Gerhard Preyer, Frank Siebelt (eds.) “This is a very fine collection of essays about David Lewis’s work. Centered around such central Lewisian doctrines as Humean supervenience, modal realism, the counterfactual analysis of causation, and physicalism, the collection is a fitting tribute to his outstanding work and to its influence.” – Jeremy Butterfield, University of Cambridge “This collection gathers papers on David Lewis’s metaphysics. Since the editors have assembled a superb group of authors, the result is at the cutting edge of the best work in contemporary metaphysics. Together with Lewis’s writings it would make for an outstanding seminar in metaphysics.” – Ernest Sosa, Brown University “David Lewis’s work occupies a central place in contemporary metaphysics, and in this volume some of our leading metaphysicians comment, interpret, and critically appraise Lewis’s important and influential claims and arguments on issues like modality and possible worlds, Humean supervenience, endurance and perdurance, counterfactuals and time, causation, and the mind-body problem. This is a timely and highly useful collection indispensable to students of Lewis’s work in metaphysics and philosophy of mind.” – Jaegwon Kim, Brown University Preface Gerhard Preyer, Frank Siebelt Reality and Humean Supervenience - Some Reflections on David Lewis’ Philosophy MODAL REALISM Phillip Bricker: Island Universes and the Analysis of Modality, John Bigelow: Time Travel Fictions, Peter Forrest: Counting the Cost of Modal Realism, Paul Teller: Lewis’s Defence of Counterpart Theory, Harold W. Noonan: The Case for Perdurance PHYSICALISM, CAUSATION AND CONDITIONALS Daniel Bonevac: Naturalism for the Faint of Heart, D. M. Armstrong: Going throug the Open Door again: Counterfactual vs. Singularists Theories of Causation, Jonathan Bennett: On Forward and Backward Counterfactual Conditionals REDUCTION

OF

MIND

Terence Horgan: Multiple Reference, Multiple Realization and the Reduction of Mind, Michael Tye: Knowing what it is like: The Ability Hypothesis and the Knowledge Argument

Rowman and Littlefield Publishers, Lanham, USA

Bookpublications

361

Wilhelm K. Essler „Unser die Welt“ Sprachphilosophische Grundlegungen der Erkenntnistheorie Ausgewählte Artikel Herausgegeben von Gerhard Preyer

Einleitung: Zur Struktur von Erfahrung 1. Fundamentals of a Semi-Kantian Metaphysics of Knowledge 2. Kant und kein Ende 3. Tarski on Language and Truth 4. Am Anfang war die Tat – Semantische und epistemologische Anmerkungen zur Sprachstufung 5. Was ist Wahrheit? 6. Gorgias hat recht! 7. Was ist und zu welchem Ende treibt man Metaphysik 8. Das logische Aufbauen von Welten 9. Unsere Welt – trotz alledem 10. Selbst das Selbst ist nicht selbst 11. „Erkenne dich selbst!“ Der Versuch von Hinweisen auf die Weisheit des Sokrates und auf die des Buddha Schakyamuni 12. Offenes Philosophieren

Digitale Publikation: Verlag Humanities Online, Frankfurt am Main www.humanities-online.de

362

Published Volumes

Subscription / Digital Version I subscribe PROTOSOCIOLOGY – pdf-file– 12,- Euro each issue, abroad: $ 25 (including courtage); payable by cheque or transfer. Date, signature ................... Separate order I order PROTOSOCIOLOGY volume Nr. ....... Single issue: 15 Euro, abroad: $ 27 (including courtage); payable by cheque or transfer. Email-posting with keyword. Date, signature ............ Your address: Name ....................

Address ....................

E-Mail ................... Send the order to: PROTOSOCIOLOGY ISSN 0940-4147 Editorial office: Stephan-Heise-Str. 56, D-60488 Frankfurt am Main 90, E-Mail: [email protected], Tel. 069/769461 PROTOSOCIOLOGY participates in Poiesis Online Service, Philosopical Dokumentation Center, Bowling Green State University, Bowling Green, OH 43403-0189, http:// www. bgsu.edu/pdc/

Editor: Gerhard Preyer, Johann Wolfgang Goethe-Universität Frankfurt am Main, Department of Social Sciences. Editorial staff: Georg Peter. www.protosociology.de

Published Volumes

363

ProtoSociology: Digital Volumes available Vol. 11 is a free download. Visit our homepage under “Special Service”. Vol. 16 – Understanding the Social: New Perspectives from Epistemology Vol. 15 – On a Sociology of Borderlines: Social Process in Time of Globalization Vol. 14 – Folk Psychology, Mental Concepts and the Ascription of Attitudes Vol. 13 – Reasoning and Argumentation Vol. 12 – After the Received View – Developments in the Theory of Science Vol. 11 – Cognitive Semantics II – Externalism in Debate (free download!) Vol. 10 – Cognitive Semantics I – Conceptions of Meaning Vol. 8/9 – Rationality II &III (double volume)

Send the order to: PROTOSOCIOLOGY, editorial office: Stephan-Heise-Str. 56, D-60488 Frankfurt am Main 90, Tel. 069/769461 Order via E-Mail: [email protected] or [email protected]

Editor: Gerhard Preyer, Johann Wolfgang Goethe-Universität Frankfurt am Main FB 3: Department of Social Sciences. Editorial staff: Georg Peter. www.protosociology.de

364

Published Volumes

Published Volumes

365

366

Published Volumes

Transactions of the Charles S. Peirce Society A Quarterly Journal in American Philosophy

Peter H. Hare Randall P. Dipert Editors Since its founding in 1965, this has been the premier journal specializing in the history of American philosophy. Although named after the founder of American pragmatism, all types of American thought are covered, from the colonial period until the recent past. All books published in the field are discussed in essay reviews.

Membership/Subscription Rates $20 Special* Individual $35 Institution $60 Sustaining $100 * Special Rate: Students/Retired/Unemployed

Postage Outside USA Surface Mail: $10 Air Mail: $25 (USA postage included in Rate above)

Back volumes available upon request.

Transactions of the C. S. Peirce Society SUNY-Buffalo, Philosophy Department 135 Park Hall Buffalo, New York 14260 fax: 716-645-6139 e-mail: [email protected] http://www.pragmatism.org/tps/

Published Volumes

Die Arbeit wurde mit folgenden Preisen ausgezeichnet:

l Carl-von-Rotteck-Preis 1992 der Albert-Ludwigs-Universita¨t Freiburg

l European Award for Legal Theory 1994 der European Academy for Legal Theory

Michel Paroussis

Theorie des juristischen Diskurses Eine institutionelle Epistemologie des Rechts Schriften zur Rechtstheorie, Heft 169 Abb.; 170 S. 1995 h3-428-08402-0i A 48,– / sFr 85,–

Einer Theorie des Diskurses ist die Aufgabe auferlegt, einesteils die juristische Rede als Aussagenensemble in ihrem genetischen Rahmen zu positionieren, anderenteils Sektoren juristischer Sprachformen auszudifferenzieren und theoretisch zu behandeln. Solche Sektoren juristischer Sprache bilden die von der Linguistik entwickelten Ebenen des Textes, des Satzes und des Wortes. Ihrer nimmt sich die Diskurstheorie an, ohne jedoch an analytischen Mustern der Sprechakttheorie, der generativen Grammatik, der Semiotik oder der Logik zu haften. Diskurse, hier nicht als einfache, normierbare, subjektbezogene Dialogprozeduren aufgefaßt, bilden den institutionell bedingten epistemologischen Rahmen einer Analytik der juristischen Sprache, die a) Recht (auf der Textebene) nicht als Fachsprache, sondern als epistemisch diskursive Einheit, b) den Rechtssatz (auf der Satzebene) nicht als Normtext, Proposition oder Sprechakt, sondern als diskursive Aussage, und schließlich c) Rechtsbegriffe (auf der Wortebene) nicht als prima¨r normtextbezogene, sondern als diskursive Begriffe konzipiert.

Duncker & Humblot GmbH · Berlin Postfach 41 03 29 · D-12113 Berlin · Telefax (0 30) 79 00 06 31 http://www.duncker-humblot.de

367

368

Published Volumes

Erkenntnis An International Journal of Analytic Philosophy Founding Editors: Hans Reichenbach, Rudolf Carnap Editors: Wilhelm K. Essler, Universita«t Frankfurt, Frankfurt am Main, Germany; Wolfgang Spohn, Fachgruppe Philosophie, Universita«t Konstanz, Germany; Patrick Suppes, Stanford University, CA, USA

Erkenntnis is a philosophical journal publishing papers which are committed in one way or another to the philosophical attitude which is signified by the label ‘analytic philosophy’. It concentrates on those philosophical fields which are particularly inspired by this attitude, though other topics are welcome as well. These fields are: y epistemology y philosophy of science, foundations and methodology of science in general and natural and human sciences such as physics, biology, psychology, economics, social sciences, etc. in particular y philosophy of mathematics y logic, philosophy of logic, and all kinds of philosophical logics y philosophy of language y Ontology, metaphysics, theory of truth, theory of modality y philosophical psychology, philosophy of mind y practical philosophy, i.e. ethics, philosophy of action, philosophy of law, etc. Erkenntnis has as one of its objectives the provision of a suitable platform for the discussion of controversial issues; another is the provision of timely, competent reviews of important publications in an ever-growing field of research. In recent years, philosophers standing quite outside the pale of analytic philosophy have also paid careful, and indeed most welcome, attention both to precision of concept and language, and to well-grounded foundations. Erkenntnis provides for them, and for philosophers of all persuasions, a place of meeting, of discussion, and of disputation.

Subscription Information

ISSN 0165-0106

2002, Volumes 56-57 (6 issues) Subscription rate: EUR 602.00yUSD 603.00yGBP 386.00 Private rate: EUR 208.00yUSD 294.00yGBP 133.00 Subscription Rate refers to either the Paper version or the Online version. To receive the Combined Paper & Online Version please add 20%. The private rate, if applicable, is available for the paper version only. Special membership rate: DEM 190.00

Visit . . .

www.wkap.nl

P.O. Box 322, 3300 AH Dordrecht, The Netherlands, E-mail [email protected] P.O. Box 358, Accord Station, Hingham, MA 02018-0358, U.S.A., E-mail [email protected]

Published Volumes

Erkenntnis An International Journal of Analytic Philosophy Founding Editors: Hans Reichenbach, Rudolf Carnap Editors: Wilhelm K. Essler, Universita«t Frankfurt, Frankfurt am Main, Germany; Wolfgang Spohn, Fachgruppe Philosophie, Universita«t Konstanz, Germany; Patrick Suppes, Stanford University, CA, USA

Erkenntnis is a philosophical journal publishing papers which are committed in one way or another to the philosophical attitude which is signified by the label ‘analytic philosophy’. It concentrates on those philosophical fields which are particularly inspired by this attitude, though other topics are welcome as well. These fields are: y epistemology y philosophy of science, foundations and methodology of science in general and natural and human sciences such as physics, biology, psychology, economics, social sciences, etc. in particular y philosophy of mathematics y logic, philosophy of logic, and all kinds of philosophical logics y philosophy of language y Ontology, metaphysics, theory of truth, theory of modality y philosophical psychology, philosophy of mind y practical philosophy, i.e. ethics, philosophy of action, philosophy of law, etc. Erkenntnis has as one of its objectives the provision of a suitable platform for the discussion of controversial issues; another is the provision of timely, competent reviews of important publications in an ever-growing field of research. In recent years, philosophers standing quite outside the pale of analytic philosophy have also paid careful, and indeed most welcome, attention both to precision of concept and language, and to well-grounded foundations. Erkenntnis provides for them, and for philosophers of all persuasions, a place of meeting, of discussion, and of disputation.

Subscription Information

ISSN 0165-0106

2002, Volumes 56-57 (6 issues) Subscription rate: EUR 602.00yUSD 603.00yGBP 386.00 Private rate: EUR 208.00yUSD 294.00yGBP 133.00 Subscription Rate refers to either the Paper version or the Online version. To receive the Combined Paper & Online Version please add 20%. The private rate, if applicable, is available for the paper version only. Special membership rate: DEM 190.00

Visit . . .

www.wkap.nl

P.O. Box 322, 3300 AH Dordrecht, The Netherlands, E-mail [email protected] P.O. Box 358, Accord Station, Hingham, MA 02018-0358, U.S.A., E-mail [email protected]

369

370

Published Volumes

Husserl Studies Editor: William R. McKenna, Miami University, Oxford, OH, USA Associate Editor: Dieter Lohmar, Universita«t zu Ko«ln, Husserl-Archiv, Germany Book Review Editor: Sebastian Luft, Katholieke Universiteit Leuven, Husserl-Archief, Belgium

Husserl Studies is an international journal which underlines the relevance of Husserl’s phenomenology, both for contemporary philosophy and for the wider academic field. The journal serves as a forum for Husserlian studies, both systematic and historical. The publication of important texts from Husserl’s Nachla
makes such a forum even more necessary. Intercultural and interdisciplinary contributions are particularly welcomed by the journal. Occasionally, material by Husserl himself, or connected with the historical background of his thought is published. Husserl Studies also publishes critical reviews of current Husserl literature, as well as reviews or other philosophical works which have a direct bearing on Husserl research. There is also a section -Chronicle- devoted to recent developments in Husserl research. This section also serves as a bulletin board for both conferences and papers as well as books, articles and dissertations devoted to Husserl and ^ to a lesser extent ^ phenomenology in general.

Subscription Information

ISSN 0167-9848

2002, Volume 18 (3 issues) Subscription rate: EUR 236.00yUSD 237.00yGBP151.00 Private rate: EUR 121.00yUSD 134.00yGBP 78.00 Subscription Rate refers to either the Paper version or the Online version. To receive the Combined Paper & Online Version please add 20%. The private rate, if applicable, is available for the paper version only.

Visit . . .

www.wkap.nl

P.O. Box 322, 3300 AH Dordrecht, The Netherlands, E-mail [email protected] P.O. Box 358, Accord Station, Hingham, MA 02018-0358, U.S.A., E-mail [email protected]

371

Published Volumes

INFORMAL LOGIC Reasoning and Argumentation in Theory and Practice Editors: J.A. Blair, R.H. Johnson

INFORMAL LOGIC publishes articles which advance the dialectic in reasoning and argumentation in theory and practice, including but not restricted to: theory of argument, fallacy analysis and fallacy theory, criteria of good argument, rationality and argument, psychology of argumentation, argument fields, theory of critical thinking, the teaching of argumentation, informal logic, and critical thinking, as well as related topics in cognate fields.

TEACHING SUPPLEMENT Each issue of INFORMAL L OGIC now contains a teaching supplement with materials for use in critical thinking, reasoning skills, and informal logic courses  an invaluable tool for instructors.

H.V. Hansen, C.W. Tindale

VOLUME 20 HIGHLIGHTS Jonathan Adler Belief and Negation Louise Cummings Petitio Principii: The Case for Non-Fallaciousness Alec Fisher Informal Logic and Its Implications for Philosophy Jean Goodwin Wigmore's Chart Method Bram van Heuveln A Preferred Treatment of Mill’s Methods. Nicholas Rescher Reification Fallacies and Inappropriate Totalities Douglas Walton New Dialectical Rules for Ambiguity John Woods How Philosophical is Informal Logic?

INFORMAL LOGIC

Prices per volume (3 issues)

Department of Philosophy, University of Windsor

Canada (in $CDN): Individuals $40.00, Institutions $65.00, and Students $25.00. United States (in $US): Individuals $40.00, Institutions $65.00, and Students $25.00. Elsewhere (in $US): Individuals $43.00, Institutions, $70.00, and Students $25.00.

Windsor, Ontario, Canada. N9B 3P4 E-mail: [email protected]

www.uwindsor.ca/faculty/arts/philosophy/IL All taxes and shipping charges are included. Phone: (519) 253-3000, x2332

Fax: (519) 971-3610

Make cheque or money order payable to INFORMAL LOGIC.

372

Published Volumes

WHEN TALK IS A SCIENCE…

Linguistics and Language Behavior Abstracts Comprehensive, cost-effective, timely coverage of current ideas in linguistics and language research Abstracts of articles, books, and conference papers from nearly 1,500 journals published in 35 countries; citations of relevant dissertations as well as books and other media. Available in print or electronically through the Internet Database Service from Cambridge Scientific Abstracts (www.csa.com). Contact [email protected] for trial Internet access or a sample issue. Linguistics

Language Behavior Abstracts

Published by CSA