Pragmatics, Conceptual Content and Inference

PRAGMATICS, CONCEPTUAL CONTENT AND INFERENCE*

Eduardo de Bustos

Dpt. of Logic, History and Philosophy of science

UNED – Spain

Abstract

The general idea behind the semantics of R. Brandom (1994, 2000) is that the conceptual content (the meaning) of expressions is the outcome of its function in inference. In others words, the conceptual content is not previous, in the causal (or methodological) order to its application in discourse. The relation between the normative concepts and the action is nothing but a special case of this relation between the conceptual and the practice. The general philosophical stance of R. Brandom in this respect is a variety of functionalism: the meaning of the expressions is a function of their (inferential) use in language.

This stance is congruent not only with the most popular varieties of pragmatism, but it fits as well with various general contentions that may be considered as characteristic or part of the definition of pragmatics. In particular, it is congruent with the so named pragmatic perspective (Verschueren and Bertucelli-Papi, 1987). Such a perspective takes the meaning not as an immanent property of linguistic expressions, but a property that expressions receive (is ascribed to) when they are used in communication.

But the issue I intend to discuss is not about the congruence relations between R. Brandom´s inferentialism and the pragmatic theory of meaning, but about theoretical relations of articulation, explicitness and complementation. The semantic theory of R. Brandom can be interpreted as an attempt to make explicit the way in which the assertive use of language gives rise to the cognitive content of linguistic expressions. The pragmatic theory of meaning has been in this point very general or indeterminate. In fact, it sticks to the emblematic statement that the meaning is the use in communication. The speech act theory has been considered as the theoretical means that concretes this functional dimension of meaning. Relevance theory is supposed to be a significant progress in advancing a number of theses that tie the functional dimension of language to its cognitive content, but the appeal of R. Brandom´s inferentialism is that he proposes a general philosophical position that it is coherent, makes sense and attempts to defend detailed contentions about this link. Concretely the notion of discursive commitment, or discursive responsibility, could be used not only to delimitate the notion of cognitive or conceptual content, as R. Brandom proposes, but also to reach an understanding of the general normative dimension that percolates the linguistic communication. The thesis of R. Brandom about this issue, if conveniently generalized, can be an important contribution to understand in a proper way the normative character of the maxims of conversation and the principle of cooperation. The paper critically analyses some of the R. Brandom particular proposals about the mechanism of conceptual, cognitive or propositional adscription to linguistic expressions. Particularly, it disputes the viability of his conception about the delimitation of the inferential role (and a fortiori of discursive commitment). While analysing some concretes examples concerning the semantics of coordination and metaphoric statements, it criticizes the sufficient nature of material implication for the adscription of propositional content. It contends that a more generous and real (cognitively speaking) notion of inference is needed if the inferences that conform the conceptual content as the meaning of an expression are to be accounted for.

1. Introduction

The main semantic idea around which the philosophy of language of R. Brandom is turning is that the cognitive content of a proposition is a function of its inferential role in discourse, i.e., of its capacity to fulfil the function of a premise or conclusion in an inferential process. The cognitive content of the elements of the proposition is set through a substitutional process (salva content) with a methodology essentially fregean.

On the other side, Brandom´s theory sticks on a paradigmatic way to the speech act of assertion or, better, to the content of the assertion, the proposition. R. Brandom recalls the Kantian intuition: the proposition is the minimal unit of which the speaker is discursively (rationally) responsible. In R. Brandom’s theory, the notion of proposition equals to the notion of propositional or cognitive content. In this concern, it is not detached from the traditional or orthodox theories in semantics, as the truth conditional semantics. Its novelty in this point is methodological over all, referring to the explanatory strategy to be used. What in traditional semantics are prior and primitive theoretical notions, as the reference and truth, come to be derived from other notions. His way to characterise the traditional notions make use of more basic concepts, of a discursive nature, trying to specify the very concept of linguistic use or discursive practice.

In the following, we will not explore in detail the whole semantic project of R. Brandom nor its articulation with the more general project of the rationalist pragmatism. We will attempt to analyse in critical way the nuclear theses of his semantics, that the propositional content is ascribed to assertions or other kinds of speech acts trough an inferential process conceived in terms of material inference. This attempt, that actually pursues an extended application of the R. Brandom’s inferentialism, takes for granted that such an inferentialism has unsustainable limitations as a theory of meaning. In rather rough terms, to be précised later, such restrictions can be detailed in two points:

1) a limitation to consider just the speech act of assertion in a way almost exclusive (and excluding), as the source of discursive commitments and entitlements.

While R. Brandom is a conspicuous adversary of representationalism, he share with this dominant philosophy of language and mind the assumption that the discursive practices are centred in representation or expression of facts through the assertion or its content, the proposition. In fact, his theory of meaning is essentially a theory of theory of judgment in an Aristotelian or Kantian way.

However, there are no reasons to think that such a limitation should be a methodological or epistemological need. A proper pragmatic perspective should stress the variety and heterogeneity of linguistic uses or discursive practices that give form to our communicative linguistic behaviour. It should contribute as well to discard the idea that discursive commitments and entitlements have as exclusive source the propositional or cognitive content of the propositions. As a practice constitutively normative, linguistic communication is ruled by a heterogeneous and manifold set of sources of discursive commitments and entitlements.

2) In a way closely related with the constriction to the speech act of assertion, in R. Brandom´s theory of communication it appears a conception too much restricted of the concept of inference. His theory of meaning deserves a main place to the explanation of the logical vocabulary, which on one side articulates the propositional content and, on the other, can be considered as the way in which the semantic auto conscience expresses itself. One more time, in this respect, Brandom sticks to one of the prejudices of the mainstream semantics: that the unique relevant inference for semantic theory is the deductive inference, that one that expresses paradigmatically the locution “if...then”, the material conditional¹. His originality resides in getting back, as a primary notion, the materially correct inference, over the logically valid inference, which is to be defined from it. But, if a conception too much restricted of inference is adopted, then it is possible not to account of elementary “semantic” facts or facts that are in the edge between semantic and pragmatics. The traditional and so long gone presuppositions (or the rest of them, see Boer and Lycan, 1976; Bustos, 1986) would be in this case, but also well known and analysed phenomena in the modern manuals of semantics (Mey, 1993, 2001; Verschueren, 1999; Jaszczolt, 2002). A theory of meaning trying to make explicit the identity of this notion with a set of uses or discursive practices should make use of a wide notion of inference, effectively ranging over real cognitive processes that the speakers of a language label as inferential. To reduce the conceptual adscription mechanism to material inference among propositional or cognitive contents is not only a non-justified platonic or rationalist starting point, but also it could result in empowering outcomes for a theory of meaning, making it equivalent to a theory about the propositional content.

2. Material inference and propositional content

In 2.iv.2 from Making it Explicit (MIE, from now on), Brandom examines the way trough which inference, as a cognitive process, settle the conceptual content. Leaning on ideas from the Begriffsschrift from Frege, he states that inference is not only the mean that allows us discriminate (settle, delimit, distinguish) conceptual contents. Is not just an epistemic resort, but also the outcome of such a process, what remains when we have calibrated the inferential roles (premise, conclusion) a concrete judgement (proposition) can fulfil. As is usually characterised, two propositions have the same content if their inferential functions coincide: they gives place to the same conclusions, as premises, and they are the conclusions of identical inferences (deductions). So the identity of conceptual content is characterised then in a fregean way, calling on substituibility. Two propositions have the same content if they can substitute each other in any inferential role without altering the value of the inference itself. The value of the inference is a normative value, in principle: “the fundamental semantic assignment of conceptual content to judgements is derived from the ultimately pragmatic correctness of inference” (MIE, 96). Now, what it is this notion of correctness of an inference? And what makes it pragmatic?

The answer to the first question, in R. Brandom’s point of view, is essentially Kantian, based on the distinction between the formal and the material. According to the orthodox logic (`the dogma’), the correction of an inference must be understood in terms of its validity, and its validity depends exclusively in the logical articulation of the proposition. In this point the methodological order of the explanation is not that crucial: saying that a valid inference is the one which preserves the truth of the premises in the conclusion is equivalent to stating that truth is what remains in the move from the premises to the conclusion, in a valid inference. What is important is that correction is understood (reduced) to the logical validity. Consequently, to recognize the validity of a deduction consists in assuming the belief in (the truth of) the correspondent conditional, the one that has (the conjunction of) the premises as antecedent and the conclusion as a consequent.

With regard to this rigorous conception of inferential correction, R. Brandom maintains, following W. Sellars (1997), the legitimacy of material inference. In traditional terms, material inference is the inference of which whose correction is based on the conceptual content of the elements within it; it is a step that goes from the content to the conclusion (which goes from `single’ to `not married’, from `whale’ to `mammal’). To say it in other words, the conclusion of a materially correct inference makes the conceptual content of the premises explicit, it displays its content. In order to use the same examples as R. Brandom: “As examples, consider the inference from ‘Pittsburg is to the West of Philadelphia’ to ‘Philadelphia is to the East of Pittsburg’, the inference from ‘today is Wednesday’ to ‘Tomorrow will be Thursday’ and that from ‘Lightning is seen now’ to ‘Thunder will be heard soon’” (MIE, 97-8) The usual way of considering the correction of such inferences is to state that conclusions are valid by virtue of the content of the concepts which appear in the premises. As it is well known, R. Brandom prefers the inverse explanation. Conceptual contents are the ones that are by virtue of being supported by the correction of those inferences. The primary, to R. Brandom, rightly I believe, is the practical ability (discursive, knowing how to) to manage or apply concepts, that is, firstly to know how to discriminate good material inferences from the ones that are not. Then, that practical ability can be applied, if wanted, to a presumed nature or essence of a concept, in the Aristotelian way. To a certain point, it is a question of emphasis. For a nativist, like Ray Jackendoff (1995), concepts (most of them) are innate, so the dominion of their application is a mere question of putting them in practice through learning. Discursive abilities consist fundamentally in putting to work that innate gift within a social context. Social context and communicative interaction are only necessary as triggers of genetically codified knowledge. From a more constructivist point of view, the moment of the processes of acculturation and learning is more important: they constitute the necessary component in the elaboration of a conceptual system peculiar to a community or society. However, this does not imply that the genetic or innate aspects do not carry any roll in the process of conceptual construction (the blank slate is a straw man); but that those aspects establish the cognitive limits in which that construction is elaborated.

Anyway, the centre of R. Brandom’s argument in this point turns on the primacy or not of material inferences (inferences based on the conceptual content) over strictly logical inferences. Even though material inferences can be reduced to logical inferences by means of the usual technical recourses (for example, using the postulates of meaning à la Carnap or making implicit premises explicit in the correspondent enthymeme), R. Brandom supports here the strategy of inversion: that of considering logical inferences as an extreme case of material inferences, the case in which the conceptual content tends to zero (for example, in logical connectives). Material inferences are the ones that, in the last resort, make what is implicit explicit in the discursive practice: the function of the logical vocabulary is to express the nature of the discursive practices of the speakers, as these practices attribute conceptual content.