In this paper, I wish to revisit a topic that I addressed many years ago (cf. Jacob 1997) from a novel perspective. Much philosophy of mind of the latter part of the twentieth century has been devoted to naturalizing intentionality or the contents of mental representations. One of the landmarks of naturalistic philosophy of mind of the past thirty years is unquestionably Ruth Millikan’s teleosemantic framework. Teleosemantic theories are teleological theories that seek to explain content by appealing to the functions of representations. Like most teleosemantic approaches, Millikan (1984, 2004) embraces an etiological conception of function, according to which functions are selected effects (Millikan 1984, 1989b; Neander 1991, 1995, 2004; Wright 1973): the function of a trait is the effect caused by the trait that explains the continued reproduction (survival or proliferation) of past tokens of this trait.
Millikan’s teleosemantic approach is particularly impressive for two related reasons. First, it applies in a single stroke to the contents of intentional mental representations, whose function is to mediate between pairs of cognitive mechanisms located within single brains, and also to the meanings of intentional conventional linguistic signs, whose function is to mediate between pairs of cognitive mechanisms located in the brains of distinct individuals. Second, her overall teleological (or teleofunctional) approach, based on the etiological theory of functions, is meant to offer an account of the proliferation or continued reproduction of both biological entities and non-biological cultural things, such as linguistic and non-linguistic conventions.
Following Mayr (1961), evolutionary biologists and philosophers have long argued that the distinction between so-called ultimate and proximate explanations of biological traits (e.g., behaviors) is central to evolutionary theorizing. Roughly speaking, ultimate explanations address why-questions: for example, why do birds sing? Why does singing confer a selectional advantage (or greater fitness) to birds? Proximate explanations address how-questions: for example, what are the particular external circumstances which trigger singing in birds? What are the internal brain mechanisms that allow birds to sing?
The distinction between ultimate and proximate biological explanations raises some deep scientific and philosophical questions. One such question is whether ultimate explanations should be construed as non-causal answers to why-questions. Some philosophers have argued that ultimate explanations are selectional explanations based on natural selection. Natural selection can account for the prevalence of some trait in a population of individuals, but it cannot track the causal process whereby the trait is generated in each individual in the first place (Sober 1984, pp. 147–152; Dretske 1988, pp. 92–93; Dretske 1990, pp. 827–830). Other philosophers have replied that selectional explanations are causal explanations, on the grounds that no token of a trait whose type has been selected for fulfilling its (etiological) function could proliferate unless it was linked by a causal chain to the earlier production of the selected effect by ancestor tokens of the same type of trait (Millikan 1990, p. 808).[1]
In this paper, I will not address such perplexing issues. I will simply accept the validity of the distinction and assume that (whether ultimate explanations are causal explanations or not) ultimate and proximate explanations are complementary, not competing, explanations. Given that why-questions are fundamentally different from how-questions, it is likely that ultimate explanations offer few (if any) constraints on proximate explanations, and vice versa. I will further assume that the distinction carries over from biological to cultural evolution and applies to the evolution of human communication (cf. Scott-Philipps et al. 2011). In particular, Millikan’s basic teleosemantic account of the proliferation of intentional conventional linguistic signs can usefully be construed as a kind of ultimate explanation of human (verbal and non-verbal) communication. Its main task is to address questions such as: what is the evolutionary or cultural function of human communication? Why do humans engage in communication at all? As with other kinds of ultimate explanations, it needs to be supplemented by specific proximate explanations whose role is to disclose the particular human cognitive capacities and mental processes whereby humans produce and understand intentional conventional signs.
The goal of this paper is to assess the balance between Millikan’s broad teleosemantic approach to the cooperative function of human communication and the choice of particular proximate psychological mechanisms that she endorses. In particular, I will focus on her anti-mentalistic view, namely that verbal understanding of another’s utterance is a kind of direct perception of whatever the utterance is about, and her correlative rejection of the basic Gricean pragmatic assumption that verbal understanding is an exercise in mindreading. One of the distinctive features of the human mindreading capacity is that it enables individuals to make sense of two kinds of agency: instrumental and communicative agency. In order to make sense of an agent’s instrumental action, one must represent the contents of both her motivations and epistemic states. In order to make sense of an agent’s communicative action, as Grice has basically argued, the addressee must infer what the agent is trying to convey, i.e., her communicative intention, whose very fulfilment requires that it is recognized by the addressee. What is distinctive of human intentional communication is that it enables the communicative agent to cause her addressee to acquire new psychological states, and thereby to manipulate his mind.
Thus, I shall examine the contrast between the particular proximate mechanisms favored by Millikan and the Gricean pragmatic tradition. In the first section, I shall spell out the basic Gricean mentalistic framework. In the second section, I will spell out Millikan’s teleosemantic machinery. In the third section, I will examine Millikan’s view that verbal understanding is an extended form of perception. In the fourth section, I will examine the extent to which Millikan’s account of conventions can support her rejection of the Gricean assumption that verbal understanding is an exercise in mindreading. Finally, in the last section, I will show that recent developmental findings in the investigation of early human social cognition are relevant to the controversy between Millikan and the Gricean tradition over the choice of proximate mechanisms underlying human communication.