New Tools and Speech Databases

Abstract

Before moving on to the actual empirical pursuits of this monograph, it is worth backing up a bit to the origins of this project, to its foundations in work on data management in sociolinguistics and in consideration of the nature of sociolinguistic data. The research discussed here was initially inspired by my ongoing work on methodologies for databasing and archiving speech recordings (Kendall, 2008a, 2009, 2010a, 2011, forthcoming a, b, Kendall and Bradlow 2011). Over the past half-dozen years, I have been involved in the development of two projects in particular, SLAAP and OSCAAR (described, and acronyms expanded, below), which center on the creation of web-based digital archives built around time-aligned annotation frameworks. Ultimately, it was the development of these time-aligned frameworks and an exploration of theories of transcription (Kendall 2005, 2006–2007), which led to my interest in speech timing phenomena. In a sense, I stumbled into questions about pause timing (Kendall 2006) as I explored various ways users might interact with the time-aligned transcription model implemented in SLAAP.