ABSTRACT

This chapter explores the multiple sources of diachronic found data, valuable as exemplars of types. Such data are more readily available than one might assume and are not always a poor alternative to collecting new corpora, though they may be the poor linguist's primary alternative. It shows that one can augment found data with annotations and metadata prior to analysis, option generally not available when only papers, and not the underlying data, are published. The chapter discusses how to compensate for some gaps in data, metadata or coding to create a viable panel study. The default method of acquiring panel data is to collect it by recruiting and recording interviews of speakers at two or more points in time. The chapter also discusses the challenges in tailoring data for panel studies and focuses on reuse. It explains the factors which have been shown to complicate the analysis of panel studies but which can inform and enrich future analyses.