The Poland of linguistics

The editorial introduction to this book quotes a description of morphology (the study of word formation) as "the Poland of linguistics". American descriptive linguists of the 1950s tended to treat morphology as an annexe to the phonemic analyses that were commonly the centrepiece of their research. Later, when generative linguists turned syntax into the "hot" branch of linguistics, morphology was often subsumed uneasily under that heading. Over the past 20 years or so, according to Andrew Spencer and Arnold Zwicky, linguists have come to see morphology as a topic comprising substantial facts and principles of its own, which cannot adequately be subordinated to other areas.

One contributor (Robert Beard, in a chapter on "Derivation") continues to embrace the axiom that morphology should be reduced to syntax, but on the whole the authors of the book's 32 chapters (who are predominantly American, though they include representatives of other parts of the English-speaking world and Europe) reflect the editors' claim that morphology is now enfranchised as an independent branch of modern linguistics.

The book is divided into five parts. Part one, "The phenomena", surveys the word-formation processes found in the world's languages - individual chapters cover topics such as inflexion, compounding, clitics, etc. Part two, "Morphology and grammar", examines interactions between word-structure and various other linguistic levels. Part three, "Theoretical issues", considers the implications of morphological data for theoretical linguistic controversies. Part four, "Morphology in a wider setting", contains chapters relating morphology to topics such as child language acquisition, aphasia and word recognition. Part five contains sketches of the morphological systems of ten languages chosen from diverse, often little-known language families, focusing particularly (but not exclusively) on phenomena that are distinctively characteristic of those respective languages.

Many readers will find part five the most valuable section. Some of its chapters draw attention to possibilities of which one would never become aware through experience of the handful of languages that are widely studied. In Archi, a language spoken by 1,000 people in one village 7,500 feet above sea level in Daghestan, Aleksandr Kibrik tells us that verb inflexion is so complex that one stem can have more than 1,500,000 inflected forms.

Speakers of European languages usually know without hesitation whether some inflected form is a valid word or not, even if they happen never to have encountered it before. Archi speakers apparently often begin by denying that a complex inflected form is possible and then radically change their judgement on further consideration.

The discipline imposed by the need to make connexions with bodies of knowledge independent of linguistics makes these chapters commendably free of abstruse theorising for theory's sake. However, the core of the book is its first three parts. These probably mention, somewhere, virtually every morphological phenomenon likely to engage linguists' attention. But it is debatable how far they can be seen as succeeding in covering their topics in the clear, definitive style required for a handbook.

Furthermore, when the book quotes various data from familiar languages, worryingly often the facts are wrong. I was surprised by a claim, used to support the conclusion of Mark Aronoff and Frank Anshen's chapter, that the word "specialism", unlike "specialist", has come into use quite recently in British English. In fact, the earliest Oxford English Dictionary citations for these words, from different sources, are both dated 1856 (and the precise sense of "specialism" described by Aronoff and Anshen is cited from 1868). Or consider a displayed example, in a chapter by an author who will be nameless here, that translates "John sees it" into French as Jean le vois. If the combined efforts of the author, two editors and the publisher's copy editor cannot avoid this sort of error with a language as universally known as French, it is difficult to trust the book when it makes odd-sounding claims (as it often does) about less familiar languages. When the data come from languages such as Greenlandic or Chichewa, their accuracy is anyone's guess.

Editors and publisher have done less than they might have to turn this collection of research articles into a systematic, coherent survey of its subject. Languages using non-Roman alphabets (Russian, Hebrew), for example, seem to be transliterated on different systems by individual authors.

The book does achieve comprehensiveness, but in other respects it does not really fulfil its task satisfactorily.

Geoffrey Sampson is reader in computer science and artificial intelligence, University of Sussex.