A grammar of Kissi: A Southern Atlantic language By G. Tucker Childs (review)

878LANGUAGE, VOLUME 73, NUMBER 4 (1997) A grammar of Kissi: A Southern Atlantic language. By G. Tucker Childs. Berlin & New York: Mouton de Gruyter, 1995. Pp xv, 370. Reviewed by David J. Dwyer, Michigan State University Spoken by almost 500,000 people in the Republic of Guinea and in adjacent areas of Sierra Leone and Liberia, Kissi belongs to the Atlantic branch of Niger-Congo, along with such betterknown languages as Fula, Wolof, Biafada, Temne, and Limba. While there has been little disagreement over what languages belong in the Atlantic branch, the degree of linguistic diversity within this modestly documented linguistic group has proven challenging for internal classification . Nevertheless, following Dalby (1965) we find general agreement that Kissi belongs to the Mel subgroup of Atlantic's Southern branch. One might imagine that a grammar of a language like Kissi is the sort of thing that someone just sits down and writes, beginning with the phonology and ending with the syntax in an orderly sort of way. But after reviewing A grammar of Kissi, one becomes aware of several crucial decisions that the author had to confront. The most obvious question is that of obtaining data. The only other major work (Mukarovsky 1948) relied primarily on the unpublished notes of Dora Earthy (n.d.) In contrast, Childs drew heavily from his own field investigations gathered 'intermittently over a period extending from 1981-84' (1 1), and supplemented with subsequent interviews with native speakers as the writing progressed. But before embarking on a project to produce this book a number of more crucial concerns confronted the author, beginning with the awareness that the grammar will likely stand as the only major grammatical analysis of the language for the foreseeable future, if previous literature on this language is any indication. This means that in addition to the presumed audience of general linguists, this work will be read by Kissi speakers, second language learners of Kissi who have no other reference grammar at their disposal, comparative linguists seeking to test their notions of universale or historical development, and even future generations of linguists who have moved beyond the current modes of language theory. How then does one write a grammar that meets the needs of such a diverse audience? In browsing through the modest set of grammars written for other Atlantic languages, one is struck by the obsolescence of grammars that work within a specific linguistic paradigm now considered outmoded and by the durability of grammars that transcend the confines of specific grammatical theories. This, then, poses a third question of how to write a durable grammar of Kissi. Finally, in providing an in-depth description of a language, how does one organize a readable, coherent account, particularly when syntactic realities mingle with its morphology and phonology? Following an introductory chapter, the book begins with a typological overview of Kissi. In addition to situating Kissi typologically, this chapter provides an overview of the workings of the Kissi grammar so that when these topics are presented in more detail later on, the reader has a sense of how the detail fits into the language. Subsequent chapters move from phonology to morphology to syntax. The chapters on phonology address segmental and prosodie systems, with an entire chapter devoted to the latter, and phonological rules with subdivisions dealing with syllable structure, segmental mies, and tone. A sizable portion of the book deals with aspects of word-level phenomena necessitated by the complexities imposed on Kissi by its noun-class system, a common, but not universal feature of Niger-Congo languages. Later chapters illustrate how this feature plays itself out in Kissi's extensive nominal and verbal morphology. The final two chapters address clause-level and sentence-level syntax. There is no chapter dealing with higher levels of discourse though Appendix E contains several texts. With respect to the question ofrepresentation, C has refrained from adopting or even advocating a particular linguistic model and uses terminology common to current general linguistics. In fact, he has avoided using highly specialized terminology, preferring to use citations and end notes to refer the reader to technical discussions rather than include such discussions in the text. More REVIEWS879 frequently, however, he uses these...