This paper is concerned with how accent types can contribute to the perceived prominence of the accented words, both when they are presented in a sentence devoid of context and when a context is used to build expectations of more or less prosodic prominence. In a prominence rating experiment in German, target words were presented with four different levels of prosodic prominence (L+H*, H*, H+L*, deaccented) in isolated sentences. In a second prominence rating experiment, words with two of these levels of prominence (L+H* and H+L*) were presented with a prior context evoking the expectation that the upcoming information was either exciting or neutral. In a third experiment, this reduced set of stimuli was also assessed for appropriateness, i.e. as to whether the target stimulus matched expectations evoked by the context provided. Results indicate that, both with and without prior context, pronounced accentual rises (L+H*) involve a particularly high level of perceptual prominence. Contextually induced expectations had no effect on the perception of prominence. However, the extra prominence for L+H* was only appropriate in the exciting context. Falling accents (H+L*) were lower in prominence, and were appropriate in both contexts, indicating that they did not need to be contextually licensed.
Cite as: Röhr, C.T., Baumann, S., Schumacher, P.B., Grice, M. (2020) Perceptual Prominence of Accent Types and the Role of Expectations. Proc. Speech Prosody 2020, 366-370, doi: 10.21437/SpeechProsody.2020-75
@inproceedings{rohr20_speechprosody,
author={Christine T. Röhr and Stefan Baumann and Petra B. Schumacher and Martine Grice},
title={{Perceptual Prominence of Accent Types and the Role of Expectations}},
year=2020,
booktitle={Proc. Speech Prosody 2020},
pages={366--370},
doi={10.21437/SpeechProsody.2020-75}
}