Knowledge of Meaning and Linguistic Communication

Abstract

The central question guiding this thesis is, how do we determine what information is semantic? I argue that the amount of information semantically-encoded is proportional to the role semantic competence plays in linguistic communication. My reasons for this are meta-theoretical. Natural language semantics is the theory of the information encoded by linguistic expressions. As such, it should proceed in accordance with normal naturalistic inquiry on the model of the core natural sciences. This includes the practice of hypothesizing entities with the goal of explaining otherwise inexplicable phenomena and theoretical virtues like parsimony. These same considerations lead me to reject a standard claim in natural language semantics, that proper names semantically refer to their bearers. My argument proceeds as follows. Ordinary assumptions about natural language semantics include the assumption that semantics is essentially a theory of truth conditions; and that its tasks, including the analysis of logical properties like validity and consistency, are to be determined a priori. In Chapter 1, I argue that these assumptions must be earned, not stipulated, for a natural language semantics understood as a scientific theory of linguistic meaning. Instead, the nature of the semantic properties we ascribe to language should be determined by the explanatory niche they fill, which in turn is determined by the broader theoretical context. I argue that this broader context should be understood as the theory of linguistic communication and the cognitive mechanisms underlying it. In Chapter 2, I examine one strategy for delimiting the domain of semantic properties, the argument from competence. Arguments from competence infer from the limitations of the cognitive mechanisms devoted to linguistic interpretation to negative claims about what semantic properties cannot be. I take a detailed look at an instance of this strategy, Kent Bach’s argument that reference is not a semantic property of proper names. In Chapter 3, I argue that the viability of Bach’s argument, and arguments from competence in general, depends on the role played by semantic knowledge. We can infer to expressions’ semantic contents from speakers’ knowledge of semantic properties only if that knowledge is necessary for linguistic communication. In Chapter 4, I argue that, because linguistic communication is ostensive-inferential, not code-based, knowledge of semantic properties is not necessary for linguistic communication. Arguments from competence, including Bach’s argument that reference is not a semantic property of name, fail. However, many of the same considerations which lead to this failure point the way to a different argument against the referential theory of names, the argument from methodological semantic minimalism. Methodological minimalism requires us to posit only those semantic properties necessary to explain the phenomena of linguistic communication. Because these phenomena are ostensive-inferential, they depend primarily on pragmatic cognitive mechanisms, and we can explain many phenomena pragmatically, rendering semantic explanation redundant. In Chapter 5, I apply methodological semantic minimalism to the thesis that reference is a semantic property of proper names.