A Theory of Granularity and Its Application to Problems of Polysemy and Underspecification of Meaning

Communication using natural language is remarkably efficient, by allowing reuse (through the use of generative devices) of a finite vocabulary to describe a potentially infinite set of situations. This vocabulary reuse contributes to words having many related senses (polysemy). Further, meanings can be relatively vague or precise; in other words, varying in their degree of specification of meaning. I suggest that these problems can be addressed by developing a knowledge representation which makes explicit the notion of granularity. As the grain size changes, we may fold certain distinctions, or split meanings more finely. In this paper, I formalize a theory of granularity and demonstrate how it can be applied to problems of meaning representation. Such a theory requires a world model which provides a rich sortal differentiation of entities based on the distinctions made by natural language, including the representation of meronymic structure and reification. Granularity will be represented in terms of structural operations defined as abstractions. I illustrate how this applies to problems of polysemy and vagueness in nominalizations, where splitting and folding of meanings are particularly evident.