Grammatical roles and relations By F. R. Palmer (review)

Abstract

BOOK NOTICES 217 meaning' (90-115) that the distinction between utterance -type meaning and utterance-token meaning in pragmatics is important and offers a three-tiered theory of communication in which utterance-type meaning has a special place. In 'Does spoken language have sentences' (116-35), Jim Miller adduces that the evidence from spontaneous spoken language indicates that the clause and not the sentence is the primary unit of linguistic analysis. 'Grammaticalisation and social structure: Non-standard conjunction-formation in East Anglian English' (136-47) by Peter Trudgill discusses a specific form of grammaticalization process and its relation to the social structure of a small, isolated community of speakers. Bernard Comrie's 'German Perfekt and Präteritum: Speculations on meaning and interpretation ' (148-61) distinguishes semantic and pragmatic contributions to the overall interpretation process by discussing the differences between the two German forms, the simple past Präteritum (ich liebte ? loved') and the compound past Perfekt (ich hatte geliebt ? had loved'). In 'The possessed' (162-74), John Anderson looks at the two morphological processes of concord and rection, copying processes which are traditionally differentiated by the role ofthe trigger (the former involving the copying of a category value, the latter the copying of a noncategory value), especially with regard to possessives . R. M. W. Dixon's 'Complement clauses and complementation strategies' (175-220) discusses the semantic tasks which all languages have to perform and their manners of grammatically achieving them by looking specifically at complement strategies in Dyirbal (and complement clauses in English and Fijian). In the concluding 'Grammar and meaning' (221-49), John Lyons comments on issues which remain important to him and which link the volume's papers—the relationship between the grammatical structure of language and its determination by meaning ; whether or not semantics is a separate level of analysis; and whether the tripartite division (phonology , grammar, semantics) posited by linguists is reality or the result of theoretical and methodological decisions. [Leo Obrst, Boeing, Philadelphia.] Grammatical roles and relations. By F. R. Palmer. Cambridge: Cambridge University Press, 1994. Pp. xv, 359. This excellent textbook explains in a clear and rigorous way some of the recent theoretical advances in our understanding ofgrammatical roles and relations, transitivity, valency-changing derivations, and the like. Palmer summarizes the literature and adds his own insights, which are often most instructive. The discussion is illustrated by a treasure trove of wellchosen examples (from over 130 languages), some from fairly obscure sources. There are nine chapters. After 'Introduction' (1-21), we get 'Roles and relations' (22-52), discussing agent and patient, different kinds of object, instrument , and the like. Then there is 'Accusative, ergative and agentive systems' (53-87), with discussion ofsplit-ergative systems, followed by 'Syntactic relations' (88-1 16), on the role of pivots in coordinate and subordinate constructions and such matters as 'dative subjects' . We then have an excellent discussion ofthe function and use of 'Passive' (117-41) (but lacking any mention of 'get'-type passives), and 'Passive: Related and problem issues' (142-75), discussing the use of passive markers as signals of a reflexive construction, and the role of middle (in the traditional sense). Applicatives are, oddly, discussed here, in relation to which argument in an applicative construction can become passive subject. There are then well-informed chapters on 'Antipassive' (176-200), including the role of antipassive in accusative languages, on 'Topic and inverse systems' (201-13), where by 'topic' P means Philippine-type focus systems, andon 'Causatives' (214-41). A glossary , a list of references together with the pages on which each is cited, and language and subject indexes round out the book. This book is particularly accessible since it steers clear of any of the current formalisms. P begins by saying that the volume 'will not contain a great deal of theoretical discussion' (1). In fact the whole book is theoretical; it is written in terms of what has recently come to be called basic linguistic theory, the theoretical paradigm which has been built up over the past 2,000 years and which is the basis for almost all grammatical descriptions. (The here-today-gonetomorrow formalists generally take one part of basic theory, blow it up a bit...