Toward an Analytic-Pragmatist Account of Folk Psychology? Derek W. Strijbos Leon C. de Bruin Dept. of Philosophy Radboud University Nijmegen Erasmusplein 1 P.O. Box 9103 6500 HD Nijmegen Institute for Philosophy Leiden University Matthias de Vrieshof 4 P.O. Box 9515 2300 RA Leiden In this paper we want to establish two points of interest to Brandom’s project of analytic pragmatism. First, we aim to show how, in light of this project, Brandom’s deontic scorekeeping model can be used as a valuable descriptive tool for characterizing folk psychological interpretation, and how this reveals certain problematic assumptions underlying dominant positions in the debate on folk psychology. Second, we will consider the relevance of empirical study of folk-psychological practice to Brandom’s project of analytic pragmatism: to what extent should pragmatic analyses of the sort Brandom proposes be subject to empirical research on what people actually say in the practice of giving and asking for reasons? 1. Introduction In this paper we want to establish two points of interest to Brandom’s project of analytic pragmatism (2008). First, we will show how the project can be relevant for a philosophical debate that is not directly associated with Brandom’s work: the debate on so-called ‘folk psychology’ in the philosophy of mind. The positions that still dominate this debate are inspired by mentalistic and formalist assumptions regarding folk-psychological interpretation (section 2). Getting these assumptions into clear view and assessing their propriety requires an overarching account of interpretation that is at the same time informative in its characterization of what people are doing during interpretation and neutral regarding rivaling explanations of this practice. Brandom’s pragmatic analysis of deontic normative vocabulary (in terms of ‘commitments’ and ‘entitlements’) explains how his deontic scorekeeping model (1994) can meet this demand and why it should thus be taken seriously as a starting point in the debate on folk-psychology. Doing just that, we shall briefly argue why mentioned assumptions are problematic and sketch the outlines of an alternative account (section 3). Second, we shall consider the relevance of empirical study of folk-psychological practice to Brandom’s project of analytic pragmatism. On the one hand, analytic pragmatism appears to have an unprecedented potential of opening up to empirical research. On the other hand, it is not clear in what ways empirical results are to contribute to the formulation of the kind of pragmatically mediated semantic relations that Brandom aims at (section 4). 2. Mentalistic and Formalist Assumptions Regarding Folk Psychology There has been a tendency among philosophers of mind to use the term ‘folk psychology’ to refer to the practical lore and know how that ‘the folk’ puts to use in making sense each other in everyday life. Our focus in this paper is on a fairly ‘high-level’, sophisticated folk-psychological capacity: the capacity to interpret each other’s behavior in terms of reasons for action. It is important to be clear from the start about what this capacity comprises: in recent years, a growing number of philosophers and psychologists have convincingly argued that much of our daily social interaction doesn’t involve the attribution of full-blown reasons to fellow agents. The lion’s share of our embodied engagements with each other doesn’t seem to require interpretation at a propositional level (Gallagher 2001). Furthermore, our social interactions take place in socially structured, normalized environments in which the need for interpretation in terms of the individual agent’s reasons is often obviated. As long as people do what they are supposed to do, according to the rules of social practice, they can often get along fine without focus on each other’s reasons (Bruner 1990, Hutto 2004, 2008). Still, there are occasions on which it is important to find out the reason in light of which the agent performed a particular action. While it is true that traditional accounts of folk psychology have significantly overestimated the scope of this particular folk psychological competence, this by itself says nothing about the theories they put forward concerning its nature. It is the latter issue that is our prime target here. There are two dominant positions in the debate on folk psychology: theory theory (TT) and simulation theory (ST). According to TT, interpretation proceeds by calling upon an acquired or innate folk psychological theory - a body of generalizations that functionally structures the relation between perceptions, mental states and actions. ST claims that our default procedure in making sense of behavior in terms of reasons for action consists in simulating the mental process responsible for the action by, in some way or other, ‘placing ourselves in the shoes’ of the agent under consideration. A remarkable feature of both approaches is that they tend to conceive of reason interpretation as a process of mental state attribution. Inspired by the metaphysical thesis of functionalism in the philosophy of mind, TT maps folk psychological understanding on the alleged functional or causal roles of mental states. On all construals (‘modular TT’ (Fodor 1992, Leslie et al. 2005), ‘scientific TT’ (Gopnik and Meltzoff 1997), ‘modal TT’ (Maibom 2003), ‘external TT’ (Braddon-Mitchell and Jackson 2007) the folk’s understanding of other people’s actions in some way or other depends on innate or acquired (tacit) knowledge that specifies the functional roles of mental states. ST’s original aim was to fend off such functionalist rendering of folk psychological interpretation (Gordon 1986, Heal 1986). While some ST-ists (most notably Gordon 1996, 2000, 2002) have also resisted the heavily mentalized picture of interpretation that went with it, the dominant cognitivist strand in ST simply replaced the ‘theory-driven’ interpretation process postulated by TT with a ‘process-driven’ mindreading procedure (Goldman 1989). On this ‘pretense-driven off-line simulation’ account (Stich and Nichols 1997), interpretation proceeds by feeding pretend beliefs and pretend desires into one’s own offline practical reasoning system and attributing the resulting pretend decisions (and corresponding actions) to the agent who is being simulated (Goldman 1989, 2006). These mentalistic views are often accompanied by a strong commitment to the belief-desire model of action interpretation. Accordingly, interpreting others in terms of reasons for action requires the (tacit) (re)construction of the action under consideration in terms a constellation of beliefs and desires, primarily a desire toward some goal and a belief regarding the means. The interpretation process thereby obeys the ‘central action principles’ of belief-desire psychology, perhaps the most salient one being ‘if A wants P and believes that doing q will bring about p, then ceteris paribus, A will 1 q.’ (Borg 2007, p. 6) On TT-accounts, such principles are simply part of people’s theory of mind. As such, they must be (tacitly) represented in some way or other. On cognitivist ST-approaches, the process-driven mindreading procedure must at least mirror these principles: a pretend-decision can only be reached by feeding a pretend belief and a pretend desire into one’s own offline practical 2 reasoning mechanism. As a consequence of all this, reason explanations in ordinary discursive practice in terms of facts, values, beliefs or desires are treated as essentially truncated versions of a kind of practical syllogism that specifies the underlying interpretation process. On these mentalistic and formalist assumptions, such explanations are only regarded as appropriate in virtue of a mindreading procedure filling in the omitted premises. 3. Interpretation as Deontic Scorekeeping These assumptions, we argue elsewhere, stem from a deep-rooted bias of mentioned accounts in favor of third-person interpretation practices (Strijbos and De Bruin, forthcoming). Here the caveat regarding the scope of folk psychology, mentioned at the beginning of the previous section, becomes extremely relevant. For what is at issue, and what has always been at issue in the debate, is the folk’s capacity to interpret each other in terms of their reasons for action. And what appears to have been 1 Frith and Happé (1999, p. 2), for example, say that ‘in everyday life we make sense of each other’s behaviour by appeal to a belief-desire psychology.’ 2 Thus, Goldman thinks that ‘a decision-making mechanism normally takes genuine (nonpretend) desires and beliefs as inputs and then outputs a genuine (nonpretend) decision. In simulation exercises, the decision-making mechanism is applied to pretend desires and beliefs and outputs pretend decisions.’ (2006, p. 29) systematically overlooked is that when and where it really matters to know the agent’s side of the story, that is: to know his reasons, people don’t trust their own theories and simulations; they tend to ask someone, preferably the agent himself. Speculating about the reasons of others is something people normally engage in only when asking is inappropriate, inconvenient or impossible. Reason interpretation, as Hutto (2004, p. 565) puts it, is not primarily a ‘spectator sport’. Compared to the answers we get ‘from the horse’s mouth’, the explanations and predictions drawn from reason speculation are generally far less accurate and successful. In short, there are good reasons to think that when and where it is important to get the agent’s perspective into clear view, the primary way to achieve this is by participating in reason discourse. Reason discourse being its primary context of application, participants in the debate should reconsider their mentalistic and formalist assumptions regarding action interpretation in terms of reasons. Shifting focus to the second-person practice of giving and asking for reasons, it seems reasonable to start by taking reason discourse at face value and closely study what people are actually saying when they are trying to make themselves understood. Consider the following brief conversations: A: “Why are you wearing a tie?” B: “Because it is important to make a good impression.” C: “Why aren’t you coming to the party tonight?” D: “I want to finish my paper.” E: “Why are you getting up so early?” F: “It is Monday.” G: “Why did you get up so early?” H: “I thought it was Monday.” In all four examples the interpreter asks the agent for his reasons for action. Now suppose that the interpreter accept the agent’s answers in these different scenarios. Taking these examples at face value, we ought to at least consider the agents’ responses as their reasons. That is: to consider their reasons to be 1) facts (that it is Monday), 2) values (that making a good impression is important), 3) desires (towards finishing the paper) or 4) beliefs (that it was Monday). To many this may seem an outrageous proposal. For how could interpretation of actions, even in the context of second-person reason discourse, get of the ground without at least implicit reference to the agent’s motivation? This rhetorical question is well taken, but hard to pin down. It all hinges on what is meant by ‘implicit’ and ‘motivation’. On TT and (cognitivist) ST proposals, ‘implicit’ equals ‘tacit’ and ‘motivation’ comes down to ‘well-formed belief-desire pair’. Taking it this way is surely begging the question at this point: arguments need to be provided as to why this is the interpretation of ‘implicit’ and ‘motivation’ that we should endorse. We think there is clearly another sense in which it is indeed nigh inconceivable to understand an agent’s reasons without a grasp of what it is that moves him. But making this sense explicit is a complicated affair as long as the terminology available is dominated by a mentalistic, formalist reading. The challenge is to give a non-question begging characterization of what it is that people are doing when they are participating in the practice of giving and asking for reasons. It is against this background that we should look at Brandom’s deontic scorekeeping model (1994), especially in light of the pragmatically mediated semantic relation he proposes for deontic normative vocabulary in chapter 4 of his 2008. On the scorekeeping model, interpreters are withholding entitlement to (having non-inferentially responded to) a practical commitment (to act) when they ask an agent for his reasons for action. In giving his reason, the agent is attempting to vindicate entitlement for his practical commitment by acknowledging (and undertaking) a commitment (and claiming entitlement) to his answer, to the effect that he takes responsibility for vindicating entitlement to it (and its committive consequences) when challenged. Accepting the agent’s answer, the interpreter 1) grants authority to the agent in attributing entitlement to the commitment expressed in the answer and 2) endorses that answer as providing a reason for action by attributing entitlement to the agent’s original practical commitment (to act). According to this model, what people do in participating in reason discourse is keeping score of what each participant is committed and entitled to say and do, given the features of the world in which the game unfolds, their past claims and actions, social standards and the authority of others. In saying or doing something, participants change their deontic scores by altering the constellation of moves they are committed and entitled to make and the moves they are prohibited from making. Thus, actions (and speech acts) may stand in need of reasons in case they constitute a move to which fellow scorekeepers do not attribute entitlement. Under such circumstances, performing a certain speech act will count as giving a reason if it can vindicate one’s entitlement. Now why should this model provide the neutral description of interpretation practices we are looking for? By characterizing reason discourse in terms of the attribution of commitments and entitlements, it is left wide open as to whether such attribution further requires the attribution of mental states and whether this be done according to the belief-desire model. It seems to us that all contenders in the debate on folk psychology could agree on this: that withholding entitlement to a commitment to act is indeed a way of saying what one does in asking someone for a reason for action. Mutatis mutandis for the other characterizations given above. But looked at in light of the project of analytic pragmatism, there is more to it than this. For why should we be confident with this particular vocabulary? With Brandom, we could endorse what he calls the ‘normative Kant-Sellars thesis’ (2008, pp. 109-116) and say that the normative vocabulary of ‘commitment’ and ‘entitlement’ stands in an ‘Elaborated-Explicated’ (LX) relation to the practice of giving and asking for reasons. What this means is that the linguistic practice that suffices to use this normative vocabulary can be algorithmically elaborated from the practice of giving and asking for reasons (which on Brandom’s account is moreover a necessary condition for utterances to count as linguistic utterances) and in turn suffices to make explicit this very same practice by enabling people to say what they are doing when they are saying things in the course of reason conversation (see figure 3 1). For Brandom, this means that mentioned normative vocabulary is a species of expressive (logical) vocabulary that enables one to codify the know-how present in the practice of giving and asking for reasons in explicit know-that. Normative Kant-Sellars Thesis: Normative Vocabulary is Elaborated-Explicating (LX) Res1 :VV 1-5 VNormative VEmpirical 1: PV-suff 4: PV-suff 5: VP-suff 2: PV-nec PAlgEl 3: PP-suff PGiving and PNormative Asking for Reasons PADP Figure 1 Importantly, in having this specifically expressive function, the normative vocabulary does not alter the reason discourse it is directed at: it does not add anything substantial to what people are already saying to each other in giving and asking each other for reasons. Rather it makes explicit what was already implicit in reason discourse. What is implicit is a particular kind of doing, not some piece of tacitly represented propositional content. Consider the above example again. A: “Why are you wearing a tie?” B: “Because it is important to make a good impression.” 3 Figure adopted from http://www.pitt.edu/~brandom/locke/locke-w4.html In accepting B’s answer that it is important to make a good impression, A attributes entitlement to B’s acknowledged commitment that it is important to make a good impression and thereby attributes entitlement to his wearing a tie. Nothing substantial is added to what is actually being said during the conversation. Compare this to a mindreading account, according to which by accepting B’s answer, A must be tacitly attributing to B something like a desire (felt duty?) to make a good impression and a belief that wearing a tie is a (the?) way to make a good impression. Characterizing the interpretation process this way significantly expands the content of interpretation beyond what is actually being said, without proper argument. Using Brandom’s deontic scorekeeping model as a neutral descriptive tool to characterize what people are doing in interpreting each other is of course only a first step to providing an explanation as to how people are able to do these things. Here we think all options should prima facie be taken seriously, including the versions of TT and cognitivist ST criticized above. Considerations of parsimony strongly speak against these formalist mindreading accounts, however (Strijbos and De Bruin, forthcoming). Taking the second-person practice of reason discourse as the primary mode of acquiring and deploying reason interpretation skills, our alternative account uses Sellars’ (1953) (and Brandom’s) notion of material inference to explain how people withhold and attribute commitments and entitlements to each other (ibid.). Consider the examples above once more. Our suggestion, in short, is that in responding to the question why he is wearing a tie by saying that it is important to make a good impression, B is constructing the practical material inference “It is important to make a good impression, therefore I shall wear a tie.” In attributing entitlement to B’s wearing a tie, A herself endorses the proposed practical material inference against the background of B’s deontic score. Similarly, by answering that he wants to finish his paper, B is constructing the material inference “I want to finish my paper, therefore I shall not go to the party tonight.” Etc. An attractive feature of this account, we argue, is that it puts much less cognitive burden on the shoulders of the interpreter. The interpreter can simply follow the lead of the agent by endorsing (or rejecting) the material inference proposed by the agent. Moreover, the content of the inference is entirely explicit in what is actually said, nothing needs to be tacitly added in order to understand the agent’s answer. Of course the material inference needs to be assessed against the background of the agent’s deontic score. But the interpreter doesn’t have to guess at these scores (by means of theory or simulation). In normal intra-cultural practice, the deontic scores of agent and interpreter significantly overlap. And to the extent that they don’t, the additional ‘personalized’ scores will normally have been provided by the agent himself, by his sayings and doings in the past. 4. Empirical Issues In our view, making progress on mentioned how-question regarding the attribution and withholding of commitments and entitlements requires both philosophical and empirical effort. There is simply no way to decide between rivalling accounts without paying proper attention to what people are actually saying and doing in reason discourse. The question we would like to raise in this last section is to what extent this is also true of the pragmatically mediated semantic analyses that form the centrepieces of Brandom’s project of analytic pragmatism. At first sight, Brandom’s project seems to have real potential to reach out to the empirical sciences. By considering the semantic relations between different vocabularies as being born out of ordinary discursive practice, study of uses of these vocabularies in discursive practice can be expected to yield important insights into the logical structure of concepts that are of philosophical importance. Taking the semantic relations between different vocabularies to a pragmatic level, the project of philosophical analysis frees itself from its often-criticized commitment to a priori methods of inquiry and opens up possibilities for an empirically informed and essentially hermeneutical way of investigation. Endorsing such a hermeneutical approach would of course make our own adoption of Brandom’s deontic scorekeeping model vulnerable to empirical considerations. It is open to philosophers, psychologists and other researchers to refute our claim of neutrality with regard to the normative vocabulary of ‘commitment’ and ‘entitlement’ by collecting evidence that speaks against the LX-analysis that backs it up. What in our view really speaks for Brandom’s analytic pragmatism is that such considerations can be allotted a proper place within the project itself. It is not clear to us to what extent Brandom himself would endorse this empirical- hermeneutical approach. In particular, it would be interesting to find out if and to what extent he would want to see LX-relations between different vocabularies reflected in 1) ontogenetic development and 2) the structure of actual discursive practice. With regard to the first: to what extent should the developmental order of infants’ mastery of different vocabularies be a criterion for the accuracy of proposed LX-relations? Here it is interesting to note that Brandom elsewhere suggests to treat the vocabulary of desire as a species of expressive, logical vocabulary (1994, 2000), making explicit material inferences from beliefs (‘doxastic commitments’) to intentions (‘practical commitments’) and actions. Yet developmental evidence shows that infants use conative vocabulary before they acquire the capacity to wield doxastic vocabulary (Wellman 1990, Gopnik and Wellman 1994, Harris 1996). What should the consequences of such empirical findings be for the pragmatic analysis proposed? As to the second: how should the different kinds of expressive, logical vocabulary be assessed in light of the interpersonal dynamics of discursive practices? It seems plausible to think that people normally try to make themselves explicit only when conversational partners demand it, when the success of discursive interaction depends on it. This suggests that we could, or perhaps even should, identify logical vocabulary by paying close attention to the kinds of words people use when they have to further justify their claims and actions, after having failed to make themselves understood in ‘normal’ terms. Yet such conversational assessment of pragmatic analyses is easier suggested than carried out. For how exactly are we to distinguish between instances of ‘normal’ and ‘justificatory’ discourse and thus to determine which vocabularies count as (non-)logical vocabularies in the folk’s discursive practices? There is a chance, moreover, that the terminology people actually use to make themselves explicit in ordinary conversation significantly differs from, and cannot by fully captured by, technical, philosophical vocabulary (such as ‘entitlement’ and ‘commitment’). What would this tell us about the status of this technical vocabulary? What exactly is such vocabulary supposed to make explicit if it is not born out of ordinary conversational practice? An option would be to say that it serves to make philosophical discourse explicit. But then the question reappears at the level of discourse: for how then is the philosophical dialogue supposed to be related to ordinary conversation? On the empirical-hermeneutical interpretation suggested here, these scenarios reveal a potential tension between the ‘analytic’ and ‘pragmatic’ part of the project of analytic pragmatism. In our view, however, there is no way of settling on these issues in advance; we should simply carry out the project and discover where it leads us. References Borg, E. (2007) “If Mirror Neurons are the Answer, What was the Question?”, Journal of Consciousness Studies, 14, pp. 5-19. Braddon-Mitchell, D and Jackson, F. (2007) Philosophy of Mind and Cognition, Oxford, Blackwell. Brandom, R.B. (1994) Making it Explicit, Cambridge, MA, Harvard University Press. Brandom, R.B. (2000) Articulating Reasons: An Introduction to Inferentialism, Cambridge, MA, Harvard University Press. Brandom, R.B. (2008) Between saying and doing. Toward an Analytic Pragmatism, Oxford, Oxford University Press. Bruner, J.S. (1990) Acts of meaning, Cambridge, MA, Harvard University Press. Fodor, J. (1992) “A Theory of the Child’s Theory of Mind”, Cognition 44, pp. 283-296. Frith and Happé (1999) “Theory of mind and self consciousness: What is it like to be autistic?”, Mind and Language, 14, pp. 1-22. Gallagher, S. (2001) “The practice of mind: Theory, simulation, or interaction?” Journal of Consciousness Studies 8, 5–7, pp. 83–107. Goldman, A. (1989) “Interpretation Psychologized”, Mind and Language 4, pp. 161–185. Goldman, A. (2006) Simulating Minds: The Philosophy, Psychology and Neuroscience of Mindreading, New York, Oxford University Press. Gopnik, A. and A.N. Meltzoff (1997) Words, thoughts, and theories, Cambridge, MA, MIT Press. Gopnik, A. and H. Wellman (1994) “The 'Theory Theory’”, in L. Hirschfield and S. Gelman (eds.) Mapping the Mind: Domain Specificity in Culture and Cognition, New York, Cambridge University Press, pp. 257-293 Gordon, R. M. (1996) “‘Radical’ Simulationism”, in P. Carruthers and P. Smith (eds.) Theories of Theories of Mind, Cambridge, Cambridge University Press, pp. 11-21. Gordon, R. M. (2000) “Simulation and the Explanation of Action”, in B. Koegler and K. Stueber (eds.) Empathy and Agency: The Problem of Understanding in the Social Sciences, Oxford, Westview Press, pp. 62-82. Gordon, R. M. (2002) “Simulation and Reason Explanation: The Radical View”, Philosophical Topics, 29, pp. 175-192. Harris, P.L. (1996) “Desires, beliefs, and language”, in P. Carruthers and P.K. Smith (eds.) Theories of theories of mind, Cambridge, Cambridge University Press, pp. 200-220. Heal, J. (1986) “Replication and Functionalism”, in J. Butterfield (ed.) Language, Mind, and Logic, Cambridge, Cambridge University Press, pp. 135-150. Hutto, D.D. (2004) “The Limits of Spectatorial Folk Psychology”, Mind and Language, 19, pp. 548– 573. Hutto D.D. (2008) Folk Psychological Narratives: The Sociocultural Basis of Understanding Reasons. Cambridge, MA, MIT Press. Leslie, A.M., T. German and P. Polizzi (2005) “Belief-Desire Reasoning as a Process of Selection”, Cognitive Psychology, 50, pp. 45-85. Maibom, H. (2003) “The Mindreader and the Scientist”, Mind and Language, 18, pp. 296-315. Sellars, W. 1953. Inference and Meaning. Mind, 62, pp. 313-338. Stich, S. and S. Nichols (1997) “Cognitive Penetrability, Rationality and Restricted Simulation”, Mind and Language, 12, pp. 297–326. Strijbos, D.W., and L.C. De Bruin (forthcoming) “Folk Psychology without Principles”. Wellman, H. (1990) The Child's Theory of Mind. Cambridge, MA, Bradford Books MIT Press