Perception, Language, and the First Person

Mark Lance and Rebecca Kukla[1]

Pragmatism has enjoyed a major resurgence in Anglo-American philosophy over the course of the last decade or two, and Robert Brandom’s work – particularly his 1994 tome Making it Explicit (MIE) – has been at the vanguard of this resurgence (Brandom 1994).[2] But pragmatism comes in several surprisingly distinct flavours. Authors such as Hubert Dreyfus find their roots in certain parts of Heidegger and in phenomenologists such as Merleau-Ponty, and they privilege embodied, preconceptual skills as opposed to discursive practices as the basic sites of meaning and agency (Dreyfus 1991; Dreyfus 1992; Todes 2001). With strong inheritances from Dewey and Wittgenstein, Richard Rorty has championed a pragmatism whose core emphasis is on the rejection of transcendental truth and high metaphysical theorizing (Rorty 1982), and this anti-theoretical banner has been taken up by several prominent ethicists, among others. For his part, Brandom, who purports to offer a systematic theory of language and meaning grounded on a foundation of pragmatic normative relationships between speakers, looks back instead to Sellars and Quine for his stripe of pragmatism. Near the start of MIE, he writes:

The explanatory strategy pursued here is to begin with an account of social practices, identify the particular structure they must exhibit in order to qualify as specifically linguistic practices, and then consider what different sorts of semantic contents those practices can confer on states, performances, and expressions caught up in them in suitable ways. (Brandom 1994, xiii)

Despite his professed pragmatism, Brandom is no foe of high theory or metanarratives, and he is vastly more interested in language and theoretical reason than in the rest of human bodily activity. For Brandom, inferentially articulated discourse forms an autonomous domain of normativity, while our bodily encounters with the world in perception and in action serve as language entry and exit points respectively. MIEdoes not offer a systematic theory of how the domain of bodily interaction and the discursive domain fit together; indeed, Brandom makes the remarkable claim that it is merely a contingent matter that discourse is bounded by perception and action, and that in principle it could exist without them (MIE 234). Hence even though he identifies as a pragmatist, the fact that it is agents with bodies and points of view who are the necessary enactors of practices seems to play no essential role in his systematic vision. There is irony in this version of pragmatism: although Brandom wants to ground meaning in practices rather than syntax or autonomous semantics, his practices are individuated merely formally – almost syntactically – as shifts in abstract scorecards of commitments and entitlements. In this paper, we aim, from a position sympathetic to Brandom’s commitment to the foundational role of normative pragmatics in meaning and language, to plant discursive practices firmly within the embodied, material domain. We maintain that one cannot really understand meaning, content, inference, commitment or entitlement without seeing these as growing out of and being constituted within a systematic network of concrete, materially and socially incarnated actions, including transactions between agents, and between agents and the world about which they speak and inquire.

We are not the first to claim that a Brandomian story of normative transactions that commit and entitle must acknowledge that these transactions are carried out by material bodies in a concrete environment.[3] But our goals are more specific than this. We argue that it is only concretely located agents that can have a first person perspective or voice. Furthermore, we claim that there must be room for such first person perspectives within a theoretical account of discursive practice, in order for it to successfully make sense of meaning, discourse, or normative statuses. Brandom’s abstract normative pragmatics, we claim, does not have the theoretical resources for identifying such perspectives, and hence it needs substantial enrichment before it can accomplish what it seeks to accomplish.

1. The Primacy of Assertions and the Problem of Voice

We will work towards this claim by first arguing that Brandom’s focus on the speech act ofasserting as the privileged building-block of language causes a problematic narrowing of his theoretical vision and resources. Analytic philosophers of all stripes act as though the most fundamental, important, and typical thing we do with language is use it to make propositionally structured declarative truth claims. Traditional philosophy of language begins its inquiry into meaning by way of an inquiry into semantics, and its inquiry into semantics by way of an analysis of truth conditions. Since only declarative truth claims have truth conditions, it has thus seemed ‘natural’ to philosophers to take such utterances as philosophically privileged and foundational, and to treat all other kinds of discourse as marginal and derivative, if they come up at all. In How to Do Things with Words, Austin demonstrated the panoply of pragmatic functions that language may have (Austin 1962). Austin’s influence, however – and indeed Austin’s own rhetoric – has done little to unseat the prevailing commitment to the idea that traditional declarative assertions form the ‘core’ of language, while utterances with other pragmatic forms – interrogatives, imperatives, performatives, etc. – are derivative monkey-business. (Although for a rare but pointed exception, see (Belnap 1990)).

For Brandom there is no route to meaningful content – much less to mental content – that is independent of pragmatic transactions. Semantic notions are to be constructed on the basis of an independent account of social normative pragmatics. Given this, one might thus assume that Brandom, who aims for a systematic and comprehensive account of discursive practice, would be interested in language in all of its pragmatic forms. It would be an odd confidence, after all, that assumes that no part of the structure of pragmatic space other than the specifically assertional would be relevant to Brandom’s larger project. Within the space of inferential relations, Brandom is a holist about content. So why should he rule out the possibility that essential aspects of assertion must themselves be understood in terms of a role within a broader practical space? Nonetheless, Brandom echoes the methodological priorities of traditional philosophy of language: he is interested in only one kind of discursive practice, namely asserting, or declarative truth-claiming.[4] On the basis of a detailed account of asserting alone he attempts to earn a semantically significant notion of inference and sub-sentential semantics. We argue that this focus on assertions to the exclusion of other pragmatic acts is problematic – it distorts our understanding of discourse as a normative phenomenon, including our understanding of assertion itself.

A crucial feature of declarative truth claims is that they are not essentially indexed to any particular speaker or audience – they are inherently “impersonal” rather than structurally bound to a first-, second-, or third-person voice. A declarative such as “Ottawa is the capital of Canada” has no personal voice. Many declaratives do have a voice: “I am sick of crappy Mexican food”; “You have schmutz on your face”; “Louise thinks that Toronto is the capital of Canada”. However, to the extent that what we are interested in when it comes to the pragmatic force of declaratives is their status as assertions or truth claims – and remember, that is the only kind of pragmatic force that interests Brandom – any declarative can be translated from one personal voice to another without changing its force in the least. Thus, qua truth-conditional assertion, “I am sick of crappy Mexican food” (spoken by Mark) is just the same as “Mark is sick of crappy Mexican food”, and “You have schmutz on your face” (directed at Rebecca) is just the same as “Rebecca has schmutz on her face”, or (uttered by Rebecca) “I have schmutz on my face”. In this sense, and relative to these concerns, the pragmatic voice (as opposed to the surface-level grammatical voice) of an assertion, qua assertion, is always impersonal – it works the same way regardless of who says it and to whom.

Now of course, one can note all kinds of pragmatic differences between, for instance, announcing one’s own schmutz and pointing it out in someone else – these utterances are governed by different rules of etiquette, have different impacts on people’s motivational structure, etc. But none of these differences show up if, like Brandom, the only pragmatic dimension we are interested in is the way that truth claims circulate and grant commitments and entitlements. The practice of asserting is, as we shall put it, agent-neutral with respect to both speaker and audience. An assertion makes the same claim regardless of who is speaking and who is listening.

In contrast, consider two types of utterance to which pragmatic voice is structurally essential. Imperatives are always spoken in the second person. An imperative must be issued to someone in order for it to count as an imperative at all. The idea of ‘translating’ an imperative into the third or first person while retaining its meaning or force does not even get off the ground conceptually. “Close the door!” makes a specific demand upon someone in particular (or some particular group of people). A ‘translation’ into the third person, such as “Mark ought to close the door”, is a declarative with a totally different pragmatic structure; it does not constitute an order at all. Nor is there any clear sense in which it ‘means’ the same thing as the original imperative. Similarly, consider vocatives, such as “Hey, Eli!” or “Yo, Emma!”, which serve to recognize someone in the second person. There is no even roughly equivalent third-personal correlate of such speech acts; “I see Eli sitting there” does not have anything like the same meaning or force, on any reasonable account. Imperative and vocative speech acts are essentially second-person transactions. At the same time, these acts are not transferable at the level of their speaker. For example, only someone who actually encounters and recognizes Eli and Emma can properly use a vocative to hail them. Imperatives and vocatives are thus – as we shall put it – agent-relative with respect to both speaker and audience.

To the extent that Brandom builds his philosophical account of language entirely out of assertions, he makes no room for any discovery that such agent-relative features of discourse play an important role in constituting or enabling meaningful discursive practices within a linguistic community. Because they are just tallies of commitments and entitlements to truth-claims, all of the normative statuses – the scorecards of commitments and entitlements – in Brandom’s theoretical universe are, of necessity, impersonal statuses that can retain their identity regardless of which subject holds them.

Why should we think that this restriction is important? That Brandom’s normative statuses are inherently impersonal should bother us only if agent-relative statuses are essential to the normative functioning of language. But as Joseph Rouse (2002) has argued, we cannot engage in a Brandomian practice of imputing and assuming entitlements and commitments unless (at a bare minimum) we are able to responsively recognize other speakers, their claims, and their normative position in the game of giving and asking of reasons, and to actively take up and accord normative statuses ourselves. These are skills that are materially incarnated, and they rely on perception and action as much as do any other normatively governed skills. That is to say: to commit and to entitle is to do and to see. Now in order for the normative force of anyone’s scores of commitments and entitlements to actually grip practice in this way, it is not enough that the commitments and entitlements be embodied in material activity.[5] For my own commitments and entitlements to govern my practices, I have to recognize that I have them, and this means recognizing that they are mine.[6] I must understand and recognize not only the shape of the normative web of commitments and entitlements, and not only how new speech acts change this web, but where I am located within this web: which scores are mine and which changes in scores affect me. This perspectival character is a logical condition for any of these scores making any difference to me at all.[7]

Furthermore, the knowledge that these are my scores – that I belong on the normative map in this location – can’t itself be just another piece of third personal theoretical knowledge. I could recognize all the commitments and entitlements that attach to subject position x, and recognize that Rebecca is the inhabitant of subject position x, without thereby recognizing that I am bound by these commitments and entitlements. Any new theoretical reasons why they should apply to me could just be added to the score without getting that score any closer to gripping me as practically compelling. Recognizing all the commitments and entitlements that attach to subject position x, and recognizing that Rebecca is the inhabitant of subject position x, that is, gets one no closer to genuine commitment. Hence I must have a practical and perspectival grasp, not only of how to recognize and negotiate my commitments and entitlements, but also of the fact that they are mine – that they commit and entitle me.

Nor is this practical grasp of our first-person relationship to our commitments something extra that is added to them. The point is that nothing counts as a genuine commitment or entitlement solely in virtue of the way it is distributed among scorecards, either synchronically or diachronically. What makes a series of ones and zeros on abstract scorecards into genuine commitments is the way they are taken up in practice, and this taking up is an essentially first personal activity, just as holding one to a commitment, or recognizing someone’s entitlement, is inherently second personal. But Brandom’s focus on assertions, and hence on impersonal normative statuses, forecloses his ability to accommodate these structurally first- and second-personal dimensions of normative transactions between agents in discourse. We need to understand shifts in our commitments and entitlements as events that involve embodied agents with first-person perspectives as they act and communicate within a social context, and not merely as shifts in an abstract, Platonic space of scorecards that could be instantiated by any Universal Turing Machine. Hence we have prima facie reason to think that an account of discursive practice that recognizes only agent-neutral normative statuses will be insufficient. Whether agent-relative statuses must be built into language itself remains to be seen. We will argue that they must, by arguing (1) that perceptual episodes are structurally agent-relative, and not analogous with agent-neutral ‘inner assertions’, and (2) that a functioning language must include speech acts that give expression to such perceptual episodes and mirror their agent-relativity.

2. Perception and Perspective

We mentioned earlier that Brandom treats language as an autonomous domain that is independent of perception, although perception provides inputs or ‘entrance points’ for language. We will argue that he is wrong to do so, and that acts of perceiving are inextricably normatively and pragmatically bound up with acts of speaking in any meaningful discursive practice. Before we get to this argument, we will first argue that perception is an activity that (like some speech acts, as we argued above) cannot be understood except as agent-relative; acts of perception could not be cut free of a particular perceiver for whom they count as first-personal episodes. In contrast, Brandom apparently understands perceptual episodes as funny kinds of assertions (for instance see MIE 236 and 243).[8] If perceptual episodes had a normative pragmatic structure analogous to acts of asserting, then they would likewise inherit the structural impersonality or agent neutrality of asserting. Perceptual episodes could then be ‘passed on’ or transferred between agents without loss of identity.

It is true that if Rebecca perceives that there is a bunny in the bush, then not only is she entitled to claim that there is a bunny in the bush, but she is also in a position to, as Brandom would put it, pass on an entitlement to this claim to everyone in her discursive community. So she can now tell Mark that there is a bunny in the bush, and in turn, on her say-so, Mark is now entitled to believe and to assert that there is a bunny in the bush. However, Rebecca’s perceptual act is not exhausted by her acquisition of such a fungible, agent-neutral entitlement. A perceptual event is not just the transferring of a propositional belief into the head of the perceiver, but a bodily event that, as McDowell has stressed (McDowell 1994; McDowell 1998), ineliminably involves receptivity. When we come to be entitled to a belief on the basis of perception it is because we encounter the world in a certain way. But notice that such events of receptivity are essentially individuating, in a specific sense: an event of perception cannot be shared among several agents, even though several agents may perceive the same thing as a result of similar interactions with the world. As Heidegger would put it, perceptions are in each case mine (Heidegger 1996, I:1). Perceptions, taken as concrete receptive encounters, by their nature build in first-person singular ownership of that encounter rather than adding it on as an afterthought; episodes of perception cannot even notionally be understood as floating free of being someone’s perceptual episode in particular, and this ownership is inherently first-personal. To perceive is inherently to be uniquely placed, indexically, with respect to what I see. Perceptual episodes are inherently particular and non-fungible in just the way that the inferential and assertional entitlements to which they give rise are not. The inference and assertion licenses I ‘pass on’ when I express what I perceive meaningfully maintain their identity through their different incarnations in different speakers, but the original perceptual episodes do not. Even if we insist that you see the very same thing as I do on a given occasion – you see the bunny also – we have two perceptual episodes grounding two different receptive entitlements, and not one.