Formal issues in lexical-functional grammar Ed. by Mary Dalrymple, Ronald M. Kaplan, John T. Maxwell, III, and Annie Zaenen (review)

BOOK NOTICES 231 domly selected series of entries will be compared: coarticulation to coherence (66-8, 1991 :59-60) and mellow to metalanguage (238-39, 1991:216-7). In the former, entries for coarticulation, cocktailparty, coda, cognate, cognitive meaning, cognize (cognise), and coherence are found in both editions. The only differences in these entries in the fourth edition were the deletion of references to other publications, the addition of a cross-reference anticipatory to coarticulation , the expansion of the coda entry to mention the distinction between simple and complex codas and to add a sentence on coda constraints, and the change of the British spelling cognise to the American cognize. Added were entries for codification (codify), cognitive grammar, cognitive metaphor, and cognitive semantics. In the second comparison, both editions had entries for mellow, mentalism, merger (merge), mesolect (al), metagrammar, and metalanguage (metalinguistic (s)) that were unchanged except for the deletion of references. The fourth edition adds new entries for melodic tier, mentalese, merge (as used in the minimalist program and different from merger [merge] used in historical linguistics), and meronym (y). As the discipline of linguistics grows and specialization becomes narrower, frequent updates of the dictionary have become necessary. This new edition is to be welcomed as a means to cope with the constantly changing terminology. Although the deletion of references to useful sources of further information is regrettable, first time purchasers of the dictionary will have acquired the best available reference for linguistic terms. Owners of previous editions may find the update a worthwhile investment. [Frances Ingemann, University of Kansas.] Formal issues in lexical-functional grammar. Ed. by Mary Dalrymple, Ronald M. Kaplan, John T. Maxwell , III, and Annie Zaenen. Stanford: Center for the Study of Language and Information, 1996. Pp. 449. $47.50. The goal of this collection of articles is to outline and make available to a wide readership developments in lexical-functional grammar (LFG) theory from 1982 to 1994. The editors believe that LFG's advantages he in 'its use of formally different representations for linguistically different kinds of information ' (ix) and have chosen papers to illustrate that this feature has allowed LFG researchers to reveal new aspects of linguistic structures. In addition to surveying research over a twelveyear period, the book provides an excellent sense of context for the papers through sectional introductions . The introductions describe the articles' relationships to work from LFG and other approaches both at the time of their original publication and at present. The volume also contains name and subject indices. 'Part I: Formal architecture' consists of two papers : 'The formal architecture of lexical-functional grammar' (7-27) and 'Lexical-functional grammar: A formal system for grammatical representation' (29-130) both by Ronald M. Kaplan, the second with Joan Bresnan. Kaplan and Bresnan's seminal paper provides an overview of LFG. The authors introduce the structures, show their relationships, and define their notation and application. The article summarizes functional requirements, discusses reasons for separating the three representational structures , and provides tools for describing long-distance dependencies. 'Part II: Nonlocal dependencies' consists of four articles which address developments in characterizing f-structure relationships. The original LFG framework did not discuss coordination, so Kaplan and Maxwell extend the theory to include an approach to coordination in 'Constituent coordination in lexical-functional grammar' (199-210). The authors explain coordination and discuss problems representing conjunctions. Further, Kaplan and Maxwell propose generalizing the expression of coordination with distribution, usually described using Kaplan and Bresnan's equation (fa) = v, by modifying the expression to deal with sets of functions. 'Part III: Word order' consists of two articles. In 'Formal devices for linguistic generalizations: West Germanic word order in LFG' (215-39). Zaenen and Kaplan examine some syntactically significant dependencies that c-structure does not capture but which word order affects. In 'Linear order, syntactic rank, and empty categories: On weak crossover' (241-74), Bresnan argues against applying weak crossover to syntactic phrase structure. Bresnan suggests that syntactic rank explains syntactic asymmetries without referring to c-command relations and can be separated from linear order. 'Part IV: Semantics and translation' consists of three articles which discuss projection architecture and information representation. In 'Translation by structural correspondences' (311-29...