Smile, blow out candle and, hey presto, a bilabial fricative

It is one of the many merits of Henry Rogers's The Sounds of Language that it keeps one foot securely on the ground in introducing what is often, to many students beginning to study linguistics, a difficult and rather abstract subject. How can we know what a "bilabial fricative" sounds like or, more crucially, how to produce one successfully? Rogers comes to the rescue: "Try smiling; pretend to blow out a candle while still smiling."

And for the more difficult "ejective" stop consonants, as found in Quechua, a South American language: "Pretend you have a hair on the tip of your tongue. Extend the tip of your tongue just beyond the lips and try and spit the hair off."

This kind of advice peppers a book that manages more successfully than many introductions to phonetics to convey the reality of what working with speech sounds is like. In addition to extensive exercises at the end of each chapter, there is a section on calligraphy showing students how to write phonetic symbols in a recognisable way, and a discussion of the problems in devising a transcription system. All of this is very welcome. Rarely do we find linguists agreeing on every detail in their transcription systems.

Sacrifices there must be in a book with such a wide-ranging agenda. The section on intonation, although useful, is very brief, and there is no chance of learning much about the history of English in the two pages devoted to it. But for sheer energy, realism and level of information, this introduction could hardly be bettered.

If all western philosophy consists of footnotes to Plato there is a sense in which the same could be said of generative, and post-generative linguistics in relation to Noam Chomsky. All the authors represented in John A. Goldsmith's Phonological Theory: The Essential Readings have been influenced to a greater or lesser extent by Chomsky and Halle's The Sound Patterns of English (1968). The advance in phonological theory enabled by this work came from seeing that the mind could handle the vast array of data needed to process sounds only if it could generate rules sufficiently powerful to operate semi-autonomously. From this has flowed the preoccupation with rule ordering and application that has dominated modern linguistics.

This is not to ignore the achievement of the essays represented in Goldsmith's book, however, many of which are classics of modern phonological theory. Goldsmith is surely right in his contention that they are much better read in their original form and we have to be grateful for their reproduction here. His own essay on "Autosegmental phonology" is a case in point, where the patient attention to detail makes it possible for readers with even modest linguistic skills to follow the argument.

For the most part, however, this is a book for the advanced guard of modern phonology - and one that enables us to see just how much the traditional boundary between phonetics and phonology has been consigned to history.

Geoffrey Finch is senior lecturer in English, Anglia Polytechnic University.