Tonal alignment patterns in Russian under changes in pitch range

Yosuke Igarashi 1 2

1 The Graduate School of Area and Culture Studies, Tokyo University of Foreign Studies
2 Japan Society for the Promotion of Science

In the present study, an experiment was performed to explore the invariant features (alignment of the F0 valley and peak, rise duration or slope) of Russian rising pitch accents under changes in pitch range, and to test general plausibility of segmental anchoring (Arvaniti et al. 1998) in this language. Pitch range was manipulated by asking speakers to modify loudness. The results revealed that 1) except for one of four speakers, the F0 valley and peak were moved backwards when loudness increased, 2) rise duration was not constant but was rather influenced by segmental duration of the accented syllable, and 3) slope was not the invariant feature of the accents but highly dependent on changes in pitch range. The results did not support a view that a given type of F0 movement has a constant duration and/ or constant slope, which should underlie a theory that treats F0 movements per se as primitive. Nor did they fully confirm segmental anchoring, a phenomenon which can be regarded as a strong support for a theory in which intonational contours consist of primitive level tones.
The factor that causes retraction of the valley and peak was also explored. It was found that speakers who retracted the peak produced longer syllable duration as loudness increased. It was suggested that increasing of loudness had an effect similar to factors such as the upcoming word or phrase boundaries or the following stressed syllable, which have been reported to cause syllable lengthening and retraction of the peak in the previous studies (e.g. Silverman and Pierrehumbert 1990, Prieto et al. 1995). The factors which cause the retraction of the valley could not be found. The variable alignment of valley, found in the experiment against general agreement that it should be consistently anchored with the onset of the accented syllable, should be an issue for further research.
The results also revealed that the valley and peak were somehow independent from each other: retraction of the valley and that of peak seems to be caused by different factors, and they are independently aligned with some segmental points not keeping constant rise duration. It was suggested that this independency is fairly consistent with a theory that treats not movements but level tones as intonational primitive.