Taking the measure of phonetic structure

Abstract

From his earliest work, Peter Ladefoged has insisted that phonetic dimensions and categories (e.g., vowel charts, IPA symbols, features) must be measurable if they are to provide an adequate basis for a universal phonetic representation. This view has led him to two questions, the pursuit of which has been enormously informative and has set the agenda for much contemporary phonetic research. (1) What are the appropriate reference frames in which to make these measurements? The debate triggered by Ladefoged as to the correct choice of reference frame is a key component of current theoretical discussion in phonetics. While the empirical data yield tantalizing hints, their interpretation is far from unambiguous, as can be seen even in some of Ladefoged’s earliest work. (2) Are there universal fixed phonetic categories within these reference frames from which individual languages choose (and if so, what are they)? Here, Ladefoged’s research has been at the leading edge of a growing consensus that there are, in fact, no fixed categories and that the division of continua into categories in particular languages is largely a random process, constrained by the nature of speech production and perception. [Work supported by NIH.]