Language, Usage and Cognition

by Joan L. Bybee

Language demonstrates structure while also showing considerable variation at all languages differ from one another while still being shaped by the same principles; utterances within a language differ from one another while exhibiting the same structural patterns; languages change over time, but in fairly regular ways. This book focuses on the dynamic processes that create languages and give them their structure and variance. It outlines a theory of language that addresses the nature of grammar, taking into account its variance and gradience, and seeks explanation in terms of the recurrent processes that operate in language use. The evidence is based on the study of large corpora of spoken and written language, what we know about how languages change, as well as the results of experiments with language users. The result is an integrated theory of language use and language change which has implications for cognitive processing and language evolution.

Genres: Linguistics

264 pages, Hardcover

First published March 1, 2008

Reviews

[Katja · Rating 2 out of 5]

Bybee's book is "more a synthesis of previous work than it is truly original". The main point made is that frequency (instances of use) impacts language representation and that an exemplar model can adequately explain language change and other linguistic phenomena which are problematic for structuralist approaches. Bybee's model is in line with those of other cognitive/functional linguists, so the book has the usual bit of criticism targeted at the generative linguists. Three chapters are Bybee's studies of the English auxiliary, the grammaticalization of "in spite of" and modern English "can". In general, I find the usage-based approach convincing but the problem I have with the book is that the presentation of what the model really is and how it works is at times annoyingly vague. The explanations all make sense intuitively but from a book about language representations I would expect more concrete definitions and more examples of such representations. What does the construction network look like? How is it that certain items become more readily accessible? When searching for similar exemplars, how is similarity measured and how is ambiguity resolved? What is the exact relation between chunking, prefabs, constructions, utterances? As it is, the theory is hardly falsifiable.