Kenyan Sign Language (KSL) is a thriving national sign language used by tens of thousands of signers in Kenya, and which emerged out of two deaf schools in western Kenya in the early 1960s. In this thesis, I provide a thorough description and analysis of the basic phonological components of the KSL lexicon used in the southwestern region of Kenya (formerly south Nyanza Province).
This phonological grammar of (SoNy)KSL makes contributions in three domains. In the descriptive domain, it provides a thorough report of the basic units in the main phonological parameters; i.e., Handshape (Ch. 4), Location (Ch. 5), and Movement (Ch 6, 7), as well as the evidence for the distinctiveness of each unit. The description for Movement and Location are particularly noteworthy because those parameters have received less attention in sign linguistics in general compared to Handshape.
In the methodological domain, the grammar is based on a KSL Lexical Database built for this project, in which over 50 phonetic characteristics of 1,880 non-compound signs were coded. This database is currently one of only a few such richly coded lexical databases of sign languages. In addition, this grammar employs a rigorous approach to determining lexical contrast, which has yielded a separate dataset of 461 minimal pairs (Ch. 3). This dataset is unique in sign linguistics and reveals patterns of lexical contrast that were not previously known—and which have generated new hypotheses about how lexical contrast may be constrained by degrees of visual similarity.
Finally, this thesis makes a theoretical contribution by comparing how different models of sign phonology can account for sign types in KSL. By evaluating the explanatory power of the main theories of sign phonology on the basis of specific descriptive data, this thesis gives unique insights into the theoretical validity of these models. It also proposes modifications in some cases, especially with regard to how the Dependency Model (DPM) can account for the representation of movement features and their relationship to the timing tier. In addition, a new movement feature, [dispersed], is described and its implementation worked out in the DPM.