Plenary: Linguistic data types and documentary linguistics

Abstract

This talk presents a basic typology of the kinds of data relevant for linguistic theorizing, based on the degree to which they are analyzed and hence abstracted from a specific act of linguistic performance. The three basic stages in data processing distinguished here are: raw data, primary data, and structural (or secondary) data. Although these stages are clearly interrelated and interdependent, it is argued that for each type and processing stage there is a different set of useful practices and evaluation metrics, some of which are discussed in more detail in the remainder of the talk. While strictly speaking, only structural data are relevant for linguistic theorizing, the discipline also needs to pay close attention to the first two types of data, at least if it aspires to be an empirical science. Which provides the (disciplinary) raison d’être for documentary linguistics.