Neural networks and the linguistics of speech

Kretzschmar, W.A. (2008) Neural networks and the linguistics of speech. Interdisciplinary Science Reviews, 33(4), pp. 336-356. (doi: 10.1179/174327908X392898)

Abstract

Traditional ideas of dialect represent an essentially aesthetic view, in which evidence from language variation is perceived according to social conventions about language and geography. Objective perception (aesthetics in the philosophical sense) of the same speech evidence can lead to quite different generalisations about language in use. The linguistics of speech (i.e., what people actually say and write to and for each other) must be distinguished from 'the linguistics of linguistic structure' that characterises many modern academic ideas about language. Taken together, the basic elements of speech correspond to what has been called a 'complex system' in sciences ranging from physics to ecology to economics. Computational modeling of neural networks appears to be a good match for analysis of language data, and can offer clues about neural function. This essay will show, using results from experiments with an implementation of a self-organising map algorithm, that application of the model from the linguistics of speech to computer neural network analysis of geographical language data can explain anomalies in traditional ideas of dialects. A clear understanding of the aesthetics of perception, combined with developments from neuroscience, can lead us to new objective generalisations about language in use, and potentially about other topics subject to aesthetic judgments.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Kretzschmar, Professor William
• Authors: Kretzschmar, W.A.
• College/School: College of Arts & Humanities > School of Critical Studies > English Language and Linguistics
• Journal Name: Interdisciplinary Science Reviews
• ISSN: 0308-0188
• ISSN (Online): 1743-2790