We can note that while Strict* tokens for each member of the set, e.g. [yes or yeah] and [no] are quite acoustically distinct, one from the other, a range of 'lax* pronunciations of each can result in objects which are almost identical, one to the other, e.g. 'nuh' and 'yiih', 'neh* and 'yeh', etc. Also we can find sounds like 'mnyeh,' 'neuh,' 'nyem', etc., which are used for both affirmative and negative. The result is that the token itself may not be usable to decide which member of the set is being produced. It appears that for the most part such tokens are no problem. On the occurrence of one of them, a hearer can decide which member of the set has been produced, and, as in the following fragment (line 5), can produce a sequentially appropriate next utterance.
No takes yet. Share an insight, caveat, or question.
Gail Jefferson (1978) studied this question.