Towards a Fieldwork Methodology for Eliciting Distinctionsin Lexical Aspect in Bantu

While analyses of lexical aspect have traditionally relied on Vendler’s (1957) typology and expansions thereof, more recent work has shown Vendler’s classes to be insufficient for characterising event types in many languages.This is particularly true for Bantu languages, which typically have large classes of “change-of-state” (COS) verbs (or verbal predicates), and which are underrepresented in cross-linguistic studies of lexical aspect. Understanding the nuances of lexical aspect in isiNdebele requires a toolkit that goes beyond the traditional Vendlerian tests, many of which produce misleading results in isiNdebele. During our 2016 fieldwork, we developed and refined a battery of tests aimed at teasing apart key distinctions within COS and other lexical aspect classes in isiNdebele. We then adapted the tests to Northern Transvaal Ndebele (Sindebele), which allowed us to make a direct cross-linguistic comparison of the construal of various events, and to see potential effects of language contact on event construal. However, even a detailed and robust series of tests can easily miss crucial phenomena related to event construal (often available to native-speaker linguists through introspection, but less easily accessed by linguists working on languages other than their own). To mitigate this issue, much of our research involved long, ethnographic-style discussions with Ndebele speakers of the potential meaning(s) and use(s) of verbs in various linguistic frames. Although this style of research is less amenable to producing neat categorisations of verbs that can be directly compared between languages, we argue that it both reduces artificially straightforward analyses, and allows us to discover which distinctions are actually worth comparing. In this paper, we describe our research strategies, and suggest how they might be adapted for use with other languages and in cross-linguistic comparison.