PublisherUniversity of Arizona Linguistics Circle
JournalCoyote Papers: Working Papers in Linguistics, Linguistic Theory at the University of Arizona
AbstractThis paper seeks to evaluate and quantify the acoustic cues utilized in the production and perception of intonational meanings in English. Although much previous work has started from observed contours and looked for their meaning based on the context where they were found (Ward and Hirschberg 1985, Gussenhoven 1984, Crystal 1969), the opposite approach is taken here. Arbitrary contours generated in a systematic way (detailed below) were presented to study participants, who were then asked to rate these items on a series of semantic scales (scales with antonyms on either end). The items were word-contour pairs consisting of one of four words in combination with one of twenty-seven contours. Contours were created on a grid which had three time points and three pitch levels. The words were chosen to have a variety of types of semantic/dictionary meanings. Participants’ ratings were used to determine the emotive meaning of the contours, and thus the degree to which the contours interrelate. As defined by Osgood (1957), the emotive meaning of an utterance or word does not relate directly to its dictionary meaning, but instead relates to the location of that utterance or word in the semantic space. By looking at what items cluster together in semantic space and comparing the acoustic and lexical/semantic characteristics of these items it is possible to understand which dimensions play a role in the assignment of meaning to intonational contours. Results show that the judgments on the semantic scales are influenced by both the intonational contours and the lexical items. Multivariate Analysis of Variance tests were performed to determine what acoustic characteristics contributed the most to the formation of clusters of items. It was found that the presence versus absence of High pitch points in the contour as well as the word used to convey the contour were the most influential factors for the participants. The approach taken here is able to uncover new shades of intonational meaning, as well as pinpoint the acoustic cues used to assess these meanings.