Prominence across languages - production differences and their implications
Speaker:Bill Barry
Institution:Saarland University
Abstract:
Prominence is the perceptual consequence of marking certain parts of an utterance as more important than others. The marking can stem from different structural and functional levels: Citing individual words of more than one syllable requires the lexically stressed syllable to be marked. Informational structuring of a sense-group requires certain words to be marked as more important to the message, i.e., to be accented, others to be negatively "marked" as less important, i.e. de-accented. This informational marking can be further differentiated so that several words or word-groups share the prominence (e.g. in broad focus conditions), or one word is singled out (narrow focus). The phonetic means at a speaker's disposal for accenting and de-accenting are well known, and are considered to be universal: duration, intensity, fundamental frequency and spectral structure. An initial hypothesis might therefore be that all speakers in all languages use these four properties in the same way to induce the necessary prominence percepts. On the other hand, psycho-acoustics tells us that duration and intensity form a trading relation within the durational range that is typical for spoken syllables. A second hypothesis might be that each speaker (irrespective of the language) develops a strategy for achieving the necessary prominence by favouring either intensity or duration. A third position could be that the four dimensions are used differently in different languages because they are required to a differing degree for other phonological purposes and are therefore less manipulable for prominence purposes. These questions will be addressed through data from a number of the world's languages (How many depends on how the project has progressed!)