Grammatical determinants for concepts of entities: The importance of being (a) kind

This talk presents empirical and conceptual arguments for thinking that language substantially constrains the range of what can thought as the content of a word (here, a noun). I will outline a theory of nouns as grammatical constructs, based on the idea that nouns are fundamentally names for abstract categories (like 'the bear' as a kind). Grammar turns these into expressions that refer over a domain with a part structure, creating the linguistic mass-count distinction, and allowing reference to concrete particulars. This template for possible meanings makes for a predictive theory of possible nouns disallowing, for instance, nouns meaning "round square" or "non-red thing". Evidence that conceptualization is rooted in perception and action (embodied cognition) is not alternative to this mentalistic approach, but provides the necessary bridge between mind-internal content and external reality.

The first part of the presentation will make the case for giving the linguistic ability a role in shaping and constraining what can be "thought-as-a-word. The second part will detail the proposed structure, and the new data it can explain; for instance, the fact that 'bears' can refer to a kind ('bears are extinct') but the seemingly synonymous 'bear specimens' cannot; while when a mass noun like 'water' allows a plural 'waters' with mass reading, it is only the singular that can refer to the kind.