ISCA Archive SpeechProsody 2008
ISCA Archive SpeechProsody 2008

Early and late nuclei in yes-no questions: tails or high rises?

Nancy Hedberg, Juan M. Sosa, Emrah Görgülü

In this paper we analyze the placement of the nucleus in yesno questions and its significance in American English. We show that the vast majority of positive yes-no questions are expressed through a low rise, often with at least one word as a tail. High rise is the second most common yes-no question nucleus; and so we are interested in the question of why a high-rise instead of a low rise contour is sometimes selected by the speaker. It can be difficult to tell whether a question should end in a late high rise, or whether an early low-rise should be postulated with a tail that is part of the rise. We bring phonetic criteria to bear on this question, and also show that post-nuclear tails tend to consist of function words or else of information that is in some sense given in the discourse. Finally, we present evidence that the discourse function of high rises overlaps with the function of tails to such an extent that it is economical to consider the high pitch accent of a high rise nuclear tune as simply an accented part of what otherwise could be analyzed as a tail.