Advancing the Utility of the Transcript: A Computer-Enhanced Methodology

Abstract

The transcript is often the primary mediating apparatus between theory and data in language research. Researchers from a wide array of linguistic disciplines and across the social sciences rely on transcripts for the analysis and presentation of their data, yet despite some important contributions to the literature (for example, Edwards 200 I, Edwards and Lampert ] 993, Ochs 1979) most transcripts remain text-based documents, varying in their conventions from researcher to researcher, and limited in their utility to the project-at-hand. While we know, as Jane Edwards writes, that "transcripts are invaluable [since 1they provide a distillation of the fleeting events of an interaction, frozen in time, freed from extraneous detail, and expressed in categories of interest to the researcher" (2001 :321), we also know that the form of and information in a given transcript will influence our interpretations of the data (Edwards 2001; Ochs 1979). Decisions as seemingly straightforward as how to layout the text to those more nuanced - like how much non-verbal information to include and how to encode minutiae such as pause-length and utterance overlap - have far reaching effects on the utility of a transcript.
This paper presents the approach to the transcript undertaken by the North Carolina Sociolinguistic Archive and Analysis Project. This approach, I argue, helps combat the confusions that arise from text-based transcripts and moves the transcript in new directions, with results that are of benefit to language researchers.