Mind the consequences of inferentialism and normativism: conceptual mental episodes ain’t in the head (at all)

In this short paper, we briefly expose a position dubbed Vehicleless Externalism for Conceptual Mental Episodes (VECME). According to it, the constitutive relations there are between the production of conceptual mental episodes by an individual and the inclusion of this individual in social discursive practices make it impossible to equate, even partially, conceptual mental episodes with the occurrence of intracranial events. Conceptual mental episodes do not have subpersonal vehicles; they have owners: persons in interpretational practices. In Section 1, the context and the scope of this position, notably in Brandom’s philosophical system, is clarified. In Section 2, we situate the roots of VECME in a marriage between inferentialism and normativism. Section 3 presents some aspects of VECME. 1 Externalism(s) and the dependences between mental activities and linguistic environments Let us begin our argument with a quote from Wilfrid Sellars, a quote itself often used by Brandom: “Clearly human beings could dispense with all discourse, though only at the expense of having nothing to say” (Sellars 1980, p.152). In the Brandomian framework, the point here is not merely to hold that the framing of new ideas, intentions or beliefs by an individual depends on the fact this individual enjoys discursive abilities. Indeed, discursive or linguistic abilities only exist as special kinds of practices, nested in social practices. It is rather the fact some individual is a member of these social practices that enables him to acquire linguistic capacities, and thus a conceptually-articulated form of mental life. Whereas numerous contemporary philosophers also remind us that language is a constitutive tool for the development of mental life (Carruthers, Clark, Dennett), Brandom reminds us that the use of that tool is not free. If language is a tool, it is a tool that enables us to do and think about many things because its use is based on 1 For instance, for abilities such as memorizing, categorizing, introspecting, reasoning, metarepresenting, formalizing, acquisition and formation of new concepts. constraints we freely abide by. These normative constraints (inheritance of normative statuses for instance) define or constitute the meaningfulness of the new abilities and performances we gain by using these tools. These normative constraints are related to social practices. These constraints we bind ourselves by define the space in which meaning and expressive freedom may appear. This is a basic difference between normative and natural constraints (Brandom 1979). Besides these socially-mediated causal relations between linguistic tools and mental activity, one generally considers that there is another essential relation between our conceptual mental life and the linguistic social practices we are embedded in: classical semantic externalism (Putnam, Burge) considers that there are important taxonomic relations between the referential dimension of the contents of one’s beliefs (at least the ones involving some kinds of concepts) and the linguistic environment “in” which these thoughts occur. Still, this externalist framework leaves intact the possibility that mind and mental episodes are still in the head. Even if the referential dimension or even all the semantic dimensions of content can only be individuated if one considers the relations an agent entertains with its environment, the bearers or the vehicles of this content have an intracranial localization (they are physically specifiable entities). We here have what McDowell and McGinn called a Duplex Conception of the Mental : mental episodes are in the head (that is: their vehicles are in the head); but they should be conceived, attributed or individuated in terms of their relations to what is outside the head. We want to argue here, following amongst others Brandom, that the fact individuals are members of social practices is not only a cause or an enabling condition of their having a meaningful cognitive life, and a descriptive constraint one should respect in the individuation of their mental contents: it entails the fact that the conceptual mental lifes of these individuals are not in their head. Or, at least: conceptual mental episodes (CME) such as judging, thinking, considering or intending that p do not have vehicles, at least in the authoritative sense in philosophy of mind: internal material entities or events that, by intrinsically providing or encoding some kind of content would, modulo some functional and causal roles, be the realizers of these CME. To put it otherwise: CME ain’t in the head at all – even in a layered model of what happens in the intracranial world. They are not realized in, supervenient on or constituted by the occurrence of inner events. This thesis we dub Vehicleless Externalism on Conceptual Mental Episodes (VECME). In an abstract way, one might figure a vehicle as what remains in one event of thinking that p if one descriptively strips this event of all its intentional and phenomenal properties: a nonintentionnally specifiable phenomenon is supposed to remain; this meaningless bundle of subpersonal material properties can be described and individuated with reference to its formal, syntactical, or neurophysiological properties. Still, it is a content-providing material entity (Clark 2006); it is an entity that comes to bear some content (Rowlands 2003, p.156). We here understand the 'providing relation' in a twofold sense: first it might mean the vehicle does encode the 2 Beliefs are not here included in CME: beliefs are mental states (whose occurrences are judgments). Still, if one wants to consider beliefs as mental episodes, then one should see them, in a Brandomian framework, as commitments to the truth of a proposition. very same content as the content of the CME, as it has the same formal structure as this content. Second it might mean the vehicle encodes some content (at t) that causally explains the simultaneous presence of the personal-level content without being a formally-conform translation of it. The first case applies to a symbolic conception of the vehicles of CME (the vehicles of CME are Mentalese sentences). The second case is proper to subsymbolic conceptions of the vehicles of CME. According to VECME, if CME are not encoded or realized in inner vehicles, they are not encoded or realized in external vehicles as well, that is, in external material entities or substances (books, diaries, sentences, pictures,...). CME rather have owners: persons (and not subpersonal parts of them, or impersonal items they would manipulate), actively standing in interpretational and institutional practices. The basic idea behind VECME is not new. Brandom devoted a paper to it in 2004 anyway. It can even be seen a consequence of an idea he already defended in 1994: neither linguistic intentionality nor mental intentionality have foundational priority in the explanation of mental and linguistic activities (1994, p.151). Both of them are inherited from the intentionality of social practices. Still, I believe the originality of VECME has not been sufficiently appreciated, and especially its importance in Brandom’s system. Actually, it might even somehow be strange of trying to understand Brandom’s originality in recent (analytic) philosophy of language by still (possible) endorsing a basic assumption of (analytic) contemporary philosophy of mind and philosophy of language that Brandom’s work also attempts to undermine. This main assumption is vehicle-internalism: the idea that the conceptual mental episodes agents produce in the game of giving and asking for reasons ultimately have an intracranial localization (although they would be individuated and attributed from social practices). On this vehicle-internalist picture, as public productions, assertions are expressions of mental phenomena; they manifest the presence of intracraniallylocated phenomena (beliefs, intentions, meaning,..) that, at least partially (for the asserter), endow these assertions with meaning (before inferential, environmental, indexical and contextual components are considered). VECME can be seen as a logical consequence of both Brandom’s sophisticated inferentialism and normativism (the thesis that meaning is constitutively normative). If it is correct, VECME makes it very difficult to endorse both an inferentialist and normativist model of linguistic activities and a traditional vehicle-internalist model of mental episodes, as Mark Lance for instance does by holding that, in his inferentialist story, “to attribute a belief is, roughly, to attribute an internal state which disposes one to make a public move of asserting a propositional content within the social game of giving and asking for reasons.” (Lance 1998, p.220). 3 Philosophers such as Wittgenstein, Arthur Collins, Lynne Ruder-Baker or Vincent Descombes endorsed or endorse it. Note that some of Brandom’s main philosophical influences (Sellars, Rorty) do not. Ditto for Davidson. 4 Anomalous monism, functionalism, reductionism, supervenience, or dual aspect theories all endorse vehicle-internalism. 2 The normative and inferentialist grounds of VECME A controversial version of the “meaning is normative” idea is the following (and we will adopt it here): there is no meaning prior to the norms that define this meaning. That is: meaning is constitutively normative. It is not only the case that means implies some oughts: more fundamentally, “for every means, there is an ought that implies it” (Gibbard 1994, p.101). Still, we will see, these oughts are more constraints than obligations. For this version of the “meaning is normative” slogan, it is not only the case that applying a concept can sometimes oblige one to apply another concept, often preclude one from applying a different concept, or permit one to apply other concepts: these normative constraints constitute the content of the initial concept (and the correct conditions of its mastery as well). This essential normative character of meaning is here best understood in an inferentialist conception of meaning (but note that inferentialism does not entail normativism, and conversely). As it is well known, according to inferentialism, conceptual content should not be (primarily) analysed in terms of reference or representation, but in terms of inference, more precisely in terms of inferential relations between conceptual content and other contents, but also perception and action (which are non-inferential entries and exits of a system). For the inferentialism we consider here, the meaning of a sentential expression (and not only of a logical concept) is its use, role, or function within its language, including perceptual entries and behavioural exits. Inferential relations constitute the inferential role of the content, which is to be equated with its meaning. But meaning is not just inferential role. This inferential role is indeed the product of inferential rules. The inferential uses, roles, or functions are thus framed and constrained by inferential rules. The meaning of a linguistic concept or expression therefore consists in its normatively framed inferential role in some language. This normative foundation of inferences makes inferential rules, and not inferrings or inferential relations, the building stone of semantics. From this point of view, the contentfulness of CME consists in their inferential role, being governed by inferential rules that are socially instituted and used (and not being governed by mere causal-functional relations within one’s cognitive economy, as some functional role semantics for mental content hold). These inferential rules typically expressed by conditionals, material or formal do not primarily consist in obligations for speakers or believers; they rather constraint our linguistic practices by delimitating what, on an inferential point of view, we may and may not do by entertaining conceptual contents. They should better be seen as normative uniformities characterizing the pattern-governed behaviours of speakers. Negative uniformities play the most important role; the rules that govern them are therefore to be construed as constraints, rather than obligations or incentives (Sellars 1974; Peregrin 2006). Inferential rules do not prescribe what ought to be done, but what is allowed and what is not allowed to do when one asserts that p. For Brandom’s inferentialism, the inferences that articulate conceptual contents are mainly material inferences (and not primarily and necessarily formal inferences); they are cashed out in terms of normative statuses (entitlements and commitments), objects of normative attitudes (undertakings (or acknowledgement), attributions (or ascriptions)). The content of what we say and judge is inferentially articulated by being pragmatically determined in normative practices of scorekeeping. These practices confer contents to performances; they institute and exercise the linguistic norms concerning how it is correct to use expressions, under what circumstances it is appropriate to produce some conceptual performances and what the appropriate consequences of these performances are. Inferential relations are thought of in terms of preservation, inheritance and exclusion relations among commitments and entitlements. The crucial point here is that these statuses do not supervene on physical inner properties of their owners: even when they are undertaken by an individual, their significance is always related to their attribution by other scorekeepers and to the attribution (by other scorekeepers as well) of other statuses that follow from them and that define their content. This point is not only crucial when we consider the content of CME: it is central when it comes to define what judging some content is. Judging that p (in fore interio) is endorsing p; it is undertaking a propositionnaly contentful commitment to p (Brandom 2004, p.244). The conceptual content (inferential role) of our judgings is mainly determined by the commitments and entitlements that are (or would be) attributed to us from the commitments we undertake in this judging. Brandom’s inferentialism comes with a fine-grained interpretationism or phenomenalism. For interpretationism, any performance can be contentful only in virtue of the fact that it is or can be treated as such in contentconferring practices. What one what is committed or entitled to, depends on what one is taken to be committed or entitled to. It is misleading to say that for this interpretationism, S is judging that p if and only if someone interprets (or would be ready to interpret) him as judging that p. It is better to say that S is judging that p if and only if he is able to acknowledge the various deontic statuses that some member(s) of the linguistic community would be ready to attribute to him and that inferentially define the content of its judging. Attributions and ascriptions need not be actual; they might just be potential. Even actual, they are more implicit than explicit: implicit attribution can equate with the tracking of S’s inferential moves and with dispositions to correct them. The members of the community do not primarily attribute mental episodes: they rather attribute the deontic statuses that define the inferential scope of the contents of these episodes (but also what these episodes are: intendings, judgings,...), and are ready to correct deviant uses of concepts. For this interpretationism, the nature of the propositional content of a mental episode (thinking, fearing, intending, judging that p) is defined by its pragmatically and socially-mediated inferential articulation. In this interpretationism, nobody can judge that p if he is not recognizable as mastering the inferential articulation of p in attributed commitments and entitlements. Understanding content is knowing the circumstances in which one becomes committed or entitled towards this content, but also the consequential commitments and entitlements one inherits by endorsing this content. Judging that p is not necessarily judging that one ought to judge various things entailed by p; it is being disposed to move to and to accept (or to make accept) these entailed contents. It is not actual interpretation that matters here, but interpretability. Only interpretability is required for someone genuinely to be binding herself by conceptual norms. This interpretability requires the mastery of the pragmatically-determined inferential proprieties that articulate the contents of CME. A good criterion enabling us to establish the presence of this practical conceptual knowledge by an agent is the fact he is disposed to correct deviant inferential uses of the content made by other agents. Let us now turn to the consequences of inferentialism and normativism for VECME. 3 Conceptual mental episodes, discursive practices, and intracranial processes Judging something is undertaking a commitment to some content that is defined in terms of other commitments and entitlements. Judging constitutively includes undertaking a commitment to the truth of a proposition. This normative stand can only exist in normative practices (Brandom 2004, p.244), where one is able to justify this commitment and to exhibit inferential mastery of this content (especially its relations with actions). The normative statuses (commitments, entitlements) that make up the contents of judgings and that define what it is for some agent to judge something are the products of social practical attitudes (Brandom 2004, p.249). Their gain or loss is therefore not reflected at all in some intrinsic properties of their owner. Undertakings of commitments, endorsing claims are social proprieties, not natural or material properties that could be instantiated in subpersonal inner events. No subpersonal item can engage in these practices; these practices have as objects the abilities of subjects, not what happens inside of their crane. If the brain can exhibit responses to various states of affairs, these responses cannot constitute our responsibilities or undertakings of commitments to things being thus-and-so. The status of being committed to the truth of p or to take a propositionally contentful stand is not vehicled by inner processes, for this status can only be realized by persons in social-interpretational (scorekeeping) practices. From this point of view, judging that p is like borrowing a book at the library or being a mayor: these social statuses do not supervene on the intrinsic properties of their owner. Two individuals might be internally physically indiscernible, one and not the other might yet be a believer, a mayor or a book borrower only in virtue of the environmental social practices he is actively embedded in. It is not only content that is determined by inferential norms. More basically, judging some content is itself determined by practices that confer the having of contents by attributing contentful commitments and entitlements. It is not only impossible to refer to some things without the appropriate environment: it is impossible to produce meaningful thoughts without actively standing in these normative practices. It is not only the case that p is inferentially and socially determined: judging it is a social status, an attributed or attributable commitment to the truth of p (whose undertaking comes with other commitments and entitlements, practically mastered by the person). This status belongs to a person, nothing inside of this person helps to realize it or constitutes it. All depends on the nature of the social practices in which the person is, and on her abilities to inferentially use the content of what she judges and so of committing herself to other judgments. The social practices define the content of the judgment, but also the behavioural and inferential factors in virtue of which the person is interpretable as judging that p (these behavioural factors include the fact the person can herself be an interpreting person). Normative practices do not attribute judgements; yet it is only by actively being a member of them (notably by being able to inferentially use conceptual contents) that a person is interpretable as entertaining contents in fore interio – that is, commitments that are judgings (and not assertings). No inner event can give or amount to the endorsement and commitment to some conceptual content. Intracranial processes are not only non-sufficient for that, they are non-necessary as well. From this claim should follow the claim that CME, as social status (commitments), are not even partially realized in or by some physical substance. There is no first an inner event that would then become an intentional episode, or an event that could be both described in nonintentional and intentional terms. The non-intentional specification of what happens in my head when I judge that p is unable to shed any explanatory lights on the happening of judging that p (and not only on the reference or on the content of this judging). VECME can accept that cognitive operations and events could exist independently of their potential attribution by mental agents – but this does not apply to CME, which are necessarily attributable or attributed social status (commitments to the truth of a proposition). If it turned out that S was unable to justify or to inferentially articulate the content p of the CME he claims to have (notably by showing brain-scans proving that “there is something in there”), we might refuse to attribute him the endorsement and thus the judging of p (and thus the CME). Conversely, more than often, we attribute CME and conceptual commitments to agents without looking at what happens inside of their head. All that matters are the behavioural capacities of the agents, as long as we can understand them as being inferential capacities (that is, as abilities to engage in pattern-governed semantic behaviour). The point is not only to say that the subpersonal internalist specifications of the socalled vehicles of CME cannot make intelligible the interpersonal normative properties in virtue of which personal mental episodes are CME (that is, endorsements of conceptual contents). It is to say that these interpersonal normative properties are essential in the definition of CME, and that they do not have necessary relations at all with subpersonal physical operations (which are thus useless for the definition and identification of CME). What is in the head is neither sufficient nor necessary for the occurrence of the essential features of CME: their normative properties. Nothing inside the head can encode the content persons express in their judgments. For the latter one is necessarily undertaken by the person, in virtue of its active insertion in linguistic practices. Nothing subpersonal can do what persons as persons do. One might argue that the fact some agent does not have inferential mastery of the content of some CME is a matter of intracranial parameters. But inferential skills are inferential skills only in virtue of inferential norms instituted by the community. Their exercise depends on intracranial processes, but the latter ones do not constitute what are correct or incorrect inferential skills and knowledge. True, sometimes neuropsychological damages may cause behavioural deficiencies that can be taken as inferential behavioural deficiencies. But what causes the deficiency does not constitute the rightness or wrongness of the deficient behaviour. Behavioural deficiencies might put into question the fact the subject really entertains some judgings or assertings only depending on our interpretations and charity principles – and not from what the brain would show or tell us. Behavioral inferential capacities are pattern governed behaviour abilities only from the outside, from actual or potential interpretations. There are only in the head the material conditions and operations that are empirically required for the possession and exercise of the behavioural skills that serve as criteria for the attribution of social statuses and thus mental attitudes and contents. Behavioural skills and performances are criteria for the phenomenal presence of the CME not because they are causally produced by intracranial operations, but because of our conceptions of what must contextually be the behavioural criteria of the presence of a CME. When we tend to attribute mental episodes to other agents, we attribute them social statuses, as abilities to accomplish the pattern-governed behaviours defining the meaning of the contents they judge or think. Brain factors do not influence our attributions. Sure, they influence the behavioural capacities of agents. But not in a way that is constitutive of the normative character of these capacities (making them pattern governed behaviour similar to the undertaking of commitments and entitlements). Intracranial processes rather causally influence the production of the pattern governed behaviour on the basis of which persons are interpretable as the owners of these CME. Put otherwise, inner events do not cause the occurrence of CME; they rather contribute to cause the occurrence of behavioural skills (which are inferential, correct or appropriate outside – they are skills for pattern governed behaviour) on the basis of which commitments and entitlements are attributable to the agent.