NEW EDITION

The Nilo-Saharan Phylum is the most controversial outcome of Greenberg's genetic classification of African languages dating from 1963. It combines a number of previous "isolates" and its coherence as against the other phyla (Afrasian, Niger-Congo, and Khoisan) has not been satisfactorily demonstrated. In this volume, the author undertakes the demonstration, taking as a starting point major groupings he established in previous morphological comparisons and then examining a lexical data base of over 600 items from all the documented languages. The large Central Sudanic and East Sudanic families are represented by reconstructed forms from previous and forthcoming publications respectively.

The outcome is a set of 359 reconstructed items divided among Nilo-Saharan isoglosses (174 in number), isoglosses for subfamilies (83), symbolic forms (21), areal items (35), items linking Nilo-Saharan and Niger-Congo (30), and items linking Nilo-Saharan and Afrasian (26). The last four groups are not evidence for establishment of Nilo-Saharan as a genetic family, but are equally important in delimiting the phylum and setting it off from other phyla, a task which has not been given proper weight in the past.

The book begins with a lengthy and valuable Introduction covering purpose and method, an annotated list of languages genetically arranged, a demographic and cultural overview summary of speakers, and typological and areal overviews. A history of Nilo-Saharan studies is included and in a separate chapter a detailed comparison with the only competing classification (CH. EHRET, forthcoming).

The main body of the text consists of the establishment of proto-Nilo-Saharan segments and modern reflexes, consideration of the role of morphology, and representation of the 359 reconstructed forms. Four indices allow the reader to locate items by semantic set, English gloss, reconstructed head-forms in alphabetical order, and selected secondary reconstructions under head-forms.