Are transcripts reproducible?

Abstract

The research reported here is part of a larger psycholinguistic project on transcribing and the use of transcripts. It is hypothesized that reproducing transcripts originally prepared on the basis of current transcription systems overloads the capability of those who carry out transcript reproduction and therefore occasions an excessive error rate. Ten reproduced transcripts were taken from (a) three current textbooks (Duranti 1997; Garman 1990; Whitney 1998), and from (b) an earlier textbook (Levinson 1983); and (c) six versions were taken from a German transcript (Keppler 1987). Additions, deletions, substitutions, and relocations of notations were identified according to five categories: Verbal, prosodie, paralinguistic, extralinguistic, and format components. The hypothesis is supported: The overall rate of change is 6.6 syllables per change (2032/308) across all 41 comparisons. Factors underlying this excessive amount of change are discussed. The proposal is made that only those notations be made which are to be used for analyses in keeping with the purposes of the research in question.