On the Methodology of Semantic Fieldwork1

This paper presents and motivates a methodology for conducting semantic fieldwork on languages of the Americas. I first argue that one cannot gather adequate information about meaning from spontaneous discourse alone. Thus, direct elicitation (including asking consultants for judgments) is an indispensable methodological tool. I then present more detailed methodological suggestions. I offer techniques for eliciting translations and discuss how one should interpret the results of a translation task. I discuss which types of judgment are legitimate, how to obtain judgments, how to interpret the results, and how discourse contexts should be presented. I make the (somewhat controversial) claim that consultants are unlikely to be influenced by the use of a meta-language (such as English) and that a meta-language is often the best option when presenting discourse contexts. [