A Grammar of Hunzib (with texts and lexicon) By Helma van den Berg (review)

884 LANGUAGE, VOLUME 73, NUMBER 4 (1997) that eliminated one symbolic entry (No. 265)—and then lists items in the previously discussed two categories . A measure of B's careful and conservative approach is that he now discusses and lists 'Nilo-Saharan fragments', which would have been included in the main listings compiled by many other comparativists but which B only provisionally includes because they 'are those forms found in the data in at least two families whose representations and/or distributions are inadequate for status as a probable isogloss ' (142). These are conveniently detailed alphabetically (143-57). Thoroughness is B's aim, and he follows the 'fragments' with a 'Fragment index' in which the items of the preceding list are re-organized into semantic categories, and citations of the previous list are repeated in full (158-70). Just before his summary, B discusses at length the different ideas of Christopher Ehret on Nilo-Saharan classification, phonology, and root identification. The discussion is nonpolemic. Bender contends that Ehret's method of taking 'the segmental system of the proto-language ... to be very close to that of the present-day language having the most complicated consonantal inventory' is a wrong approach and adds (and I agree) 'The attractive feature of this ' 'maximalist ' ' alternative is that it makes it possible to account for nearly all correspondences as reductions (mergers) from a rich proto-system. But it flies in the face of common sense and experience in other language phyla in which splits are at least as common as mergers' (182). After a concise summary, this excellent essay concludes with 'Chapter notes', an appendix of 'NiloSaharan linguistic/cultural information' ; and indices of 'Highest-level reconstructed forms by semantic sets', 'English glosses' keyed to item numbers, an index of the reconstructed forms arranged alphabetically keyed to item numbers so that the reader will not have to research each list to find an entry; a table ofreferences (abbreviations used in the text); a 'General references' section (228-46); and a languagespecific reference section, in which the bibliography is organized by language family. The book concludes with two rather informal maps of 'African language phyla and Nilo-Saharan families' and a 'Nilo-Saharan language locator'. Although B has himself characterized this opus as an essay, it is an essay only in the sense that the author is aware that all currently available data may not suffice to give final answers; but, essay or not, I am sure that this work will be an important standard reference for all interested Africanists for some years to come. [Patrick C. Ryan] A Grammar of Hunzib (with texts and lexicon). By Helma van den Berg. Lincom studies in Caucasian linguistics 1. Munich & Newcastle: Lincom Europa , 1995. Pp. xvi, 366. Hunzib is an East Caucasian language (Daghestanian -Tsezic) spoken by approximately 2000 people of several small villages in Daghestan, Russia, and two villages in Georgia. Although there are short Hunzib texts of anthropological interest, there is no significant literary tradition in Hunzib; and the rainy mountainous climate of the Hunzib homeland is not likely to encourage tourism. This record of Hunzib should be viewed as archival in the sense that Hunzib probably will not be a living language within the next few years. It will be of interest primarily to Caucasianists for the reconstruction of (Proto-)Tsezic and secondarily to people like myself, who are broad comparativists. Van den Berg describes the linguistic affinities ofHunzib (and Tsezic) and recent historical conditions and proceeds to a very thorough description of Hunzib phonology and word and phrase stress. Assimilation characteristics are detailed, making the following discussions of morphology—which are organized logically under the headings of nominal; adjectival, pronominal , preverbal, postpositional, numeral; and verbal —easy to follow. After a discussion of Hunzib syntax, a number of simple Hunzib texts are included (139-266) m which, B thoughtfully provides the Hunzib, an analysis of the forms, and the translation, enabling anyone who is reading for comparative purposes to easily grasp the morphology and syntax. How many included texts have we all seen that involved phenomena that were buried in a footnote, making understanding not impossible but unnecessarily laborious ? An example of her excellent presentation...