ABSTRACT The project is to arrive at a partial reconstruction and a sub- classification of the Nilo-Saharan Phylum of African languages by examining a lexicon of over 600 items divided into about 65 sets on semantic bases. An already completed classification based on grammatical isomorphs serves as a basis for starting work. The methods are the standard comparative-historical ones of looking for regular phonological correspondences and thereby establishing cognate sets. A sub-project to be done near the end is to run Multi-Dimensional Scaling programs on the cognate sets to see what goodness of fit may tell us about the usefulness of this quantitative method and also for possible alternative subgroupings. Results of the entire project will be published in the form of a monograph on the Nilo-Saharan Phylum containing (1) a long introduction to the phylum, its languages and its speakers, (2) the methods used, (3) tables and charts of the results, and (4) a table of all significant cognate sets.
