It has been a common experience for those who begin to work with verbatim transcripts of spoken language (or what pass for verbatim transcripts) to find them unexpectedly difficult to read. Even the most careful and exacting paper transcript loses something that seems so crucial to comprehension that extracting the message from a written transcript can become a daunting challenge. What are the extra clues that help a listener to understand a spoken text but disappear in writing?
This question invites a host of different kinds of answers. A good transcript should reflect, with reasonable accuracy, the lexicon, the syntax, and the segmental phonology of the original speech. Beyond these areas of traditional linguistic concern, however, there is no limit to the variety of background assumptions, knowledge, skills, and situational cues that a listener can be expected to exploit, but that a reader may not have available. In this paper, I use a simple spoken text as a means of exploring a few of the many clues to meaning that the listener can use, and consider some of the ways in which speakers offer help that goes beyond the more formal linguistic markers that we record in our transcripts.
Robbins Burling is Professor Emeritus of Anthropology and Linguistics at the University of Michigan