Topic and discourse structure in West Greenlandic agreement constructions by Anna Berge (review)

This book is a reworked version of the author's Ph.D. dissertation completed in 1997. It won the Mary R. Haas Award presented by the Society for the Study of the Indigenous Languages of the Americas (SSILA), but unfortunately, it had been held up in review for an unusually long period of time. Its publication is, thus, especially welcome.

Anna Berge's objective is to provide readers with an analysis of West Greenlandic agreement constructions based on the structure of discourse rather than on a traditional syntactic approach, by stressing the primacy of topic over the roles of subject and object. Her motivation for so doing stems from two factors: (i) the inadequacy of a purely syntactic analysis for describing ergative languages such as West Greenlandic (the Inuit dialect spoken on the western coast of Greenland), and (ii) the fact that studying syntax through decontextualized sentences entails various problems, including a high number of exceptions that the theory cannot explain without having recourse to nonsyntactic elements (e.g. information flow, 'psychological subject').

The study uses spoken West Greenlandic texts as raw materials—totaling more than 600 clauses—that B collected in the field by interviewing two local consultants in their native language, and by asking two more to talk freely about their personal recollections. All of the data appear in the appendix at the end of the book, with their morphological analyses and English translations. Data analysis is generally correct, except for a small number of errors that do not impair the validity of the study.

The book is divided into five chapters. It opens with an introduction discussing West Greenlandic grammar, its role (and that of other Inuit dialects) in syntactic theory, approaches to the study of discourse, and their application to discourse structure in West Greenlandic. This is followed by three core chapters that deal respectively with topic and theme as discourse roles, ergativity as a reflection of topic status, and a discussion of the relevance of switch-reference vs. thematic coherence and topic continuity. Each of these core chapters starts by positioning the chapter's matter within general linguistic theory, before describing in detail what West Greenlandic data analysis has to tell us about the topic under examination. The book ends with a conclusion summarizing its findings and hinting at some more general extrapolations that stem from West Greenlandic discourse analysis.

The author's core argument is that in West Greenlandic, the chief factor governing agreement marking is not strictly syntactic, because studies of the language, whether dealing with ergativity, transitivity, or the difference between the ergative and antipassive verbal voices, must always involve nonsyntactic elements in their explanation. This suggests that discourse is particularly important in grammatical analysis. If discourse can predictably affect syntactic constructions, it must possess predictable structural components—as syntax and semantics do—that are necessary to an adequate description of linguistic structure. Several theories have been used in studying West Greenlandic, but without a comprehensive description of what is needed to speak the language, and this at all levels of linguistic complexity (including semantics, discourse, pragmatics, and sociolinguistics), a linguist's analysis risks being tributary to the theory to which he or she adheres.

According to B, in order to be satisfactory, the study of discourse must consider central issues such as its effect on syntax, the patterns of information flow, and the thematic development of a text, as well as the sociocultural context within which a text is produced. The central premise of her study is that certain constructions that mark textual cohesion in West Greenlandic can best be understood by assuming that discourse roles such as topic—rather than syntactic roles such as subject or object—are the most relevant structural items. This means that discourse must be viewed as a structure, rather than...