Data and language documentation

The topic of this chapter is the relationship between data and language documentation. Unlike many fields of study, concerns regarding data collection and manipulation play a central role in our understanding of, and theorizing about, language documentation. The field to a large extent, in fact, owes its existence to a shift in focus in the goals of linguistic field work from concerns regarding outputs derived from primary data, like grammars and dictionaries, to the collection of the primary data itself. When trying to understand the role of data in language documentation, the first question we must consider is what precisely do we mean by data? Beginning with the work of Himmelmann (1998), it has become customary in language documentation to distinguish between primary data—constituting recordings, notes on recordings, and transcriptions—and analytical resources—like descriptive grammars and dictionaries—constructed on the basis of, and via generalization over, primary data. While making this conceptual distinction is essential to the practice and theorizing of language documentation, most individuals or teams working on language documentation projects are ultimately interested in both collecting primary data and producing the kinds of analytical resources associated with traditional language description, most prominently grammars, dictionaries, and texts (whether oriented for community or academic use). Therefore, each will be considered here. That is, the discussion will cover topics both regarding the collection, storage, and manipulation of primary data as well as the mobilization (see Holton (this volume)) of that data to create analytical resources. While it is also important to keep in mind that data is not synonymous with digital data, for the most part, in this chapter, only digital data will be discussed. Generally, digital, rather than analog, data has been the focus of work in language documentation both because new data is typically captured solely in digital form at present and because analog data is increasingly being digitized so that it can be manipulated and disseminated with digital tools. Discussion of important aspects of digitization—i.e., the process through which a digital representation of a non-digital object is created—can be found in the E-MELD School of Best Practices in Digital Language Documentation (Boynton et al. (2006)), and an exemplary case study of the digitization process can be found in Simons et al. (2007). This chapter will focus on conceptual issues rather than specific technical recommendations, though such recommendations may be discussed to provide illustrative examples. This is because our understanding of the conceptual issues evolves at a much slower rate than the technical recommendations, which change as the technologies we use for capturing and analyzing data themselves change and, therefore, largely outpace the speed through which works like this one make their way into publication. At least for the time being, the best way to find answers to questions like What audio recording device should I use? or What