SALS-SIG Research Seminar

Predictivity vs. Stipulativity in the Lexicon

Abstract:

The question of the structure of the lexicon has become the focus of much linguistic research in recent years. Much emphasis has been placed on capturing *generalizations* about the behaviour of words, in terms of e.g. semantic properties and syntactic patterning. However, I will argue in this talk that the capture of such generalizations about word use must be balanced with the need to accurately reflect individual word usage. That is, *idiosyncrasies* associated with word use must be recorded in the lexicon in addition to any generalizations, in order to explain particular interpretations of a word in specific contexts. I will show that the decision of what information to associate with a word does not solely depend on the identification of rules and exceptions to those rules, but is also influenced by pragmatic reasoning. An important implication of this for lexicon construction is that the interaction of distinct knowledge sources in language understanding can influence what information must be explicitly represented in the lexicon, and the level of generality of lexical regularities.

I will focus on the phenomenon of the resultative construction to illustrate why a balance must be struck between stipulative (word-specific) and predictive information in the lexicon, and how context can influence this balance. I will furthermore introduce an analysis of the resultative construction, combining elements of the Construction Grammar approach to resultatives (Goldberg 1995) with pragmatic reasoning, which addresses both the predictivity and the stipulativity of this phenomenon. The approach I follow suggests that an acknowledgement of the influence of lexical conventionalisation and pragmatic reasoning on interpretation can lead to models which provide a fuller account of the influence of lexical conventionalisation and pragmatic reasoning on interpretation can lead to models which provide a fuller account of the resultative data. .

Enquiries: sals@mri.mq.edu.au