2018 Fiscal Year Annual Research Report
Laryngeal mechanisms in speech production and perception: A cross-language study on F0 perturbation
| Project/Area Number |
17F17006
|
|---|
| Research Institution | Sophia University |
|---|
Principal Investigator |
荒井 隆行 上智大学, 理工学部, 教授 (80266072)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
GAO JIAYIN 上智大学, 理工学部, 外国人特別研究員
|
|---|
| Project Period (FY) |
2017-10-13 – 2020-03-31
|
| Keywords | Tokyo Japanese / voicing / pitch (f0) / voice onset time / electroglottography / laryngeal mechanism / speech production / speech perception |
|---|
| Outline of Annual Research Achievements |
[Voicing in Japanese: production] We conducted comprehensive analyses on the acoustic data collected in 2017-2018. Results showed the following trends: (1) strong overlap between the VOT (Voice Onset Time) distributions of word-initial voiced and voiceless plosives; (2) in word-initial position, a large voiceless-voiced f0 (fundamental frequency) distinction, and (3) in non-initial position, a robust distinction in phonetic voicing with a negligible distinction in f0.
These results have been submitted to Journal of Phonetics.
[Voicing in Japanese: perception by native and nonnative listeners] We conducted several perception experiments. We tested the discrimination and the identification of voiced vs. voiceless plosives, using synthesized stimuli. We tested the identification of the two plosive series by Japanese vs. French listeners, using modified natural stimuli. These results showed (1) a highly categorical perception of VOT (categorical boundary around 10 ms); (2) a secondary use of pitch in perception, and (3) a higher sensitivity to raised pitch signalling voicelessness than to lowered pitch signalling voicedness, for both Japanese and French listeners. These results have been/will be presented at several conferences.
[EGG (electroglottography)] We tested a native speaker of Korean on the production of three plosive series plus a nasal series. Preliminary results show phonation differences related to the four consonant types. In collaboration with Prof. Donna Erickson, we did a recording on singing voice and are currently analyzing the data.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
We investigated the link between production and perception of the plosive voicing in Japanese. Our new findings about the relation between voicing and pitch in Japanese are crucial for comparative studies with other languages and will give insights into how the voicing contrast is replaced with a tonal contrast in a great number of languages.
Our data show that although voiced plosives are not robustly produced with voice lead, listeners are very sensitive to voice lead in perception. This result challenges a widely accepted model of sound change (Ohala, 1981), according to which perception leads production in a sound change. Another important result concerns the perception of pitch. We find that listeners are more sensitive to high pitch than low pitch. This is in line with recent studies on production (Hanson, 2009; Kirby & Ladd, 2016) as well as our production data from Japanese. Interestingly, the same pattern is found with French listeners.
A paper about these results will be published the ICPhS conference proceedings. It has received a Goesta Bruce Scholarship offered by the International Phonetic Association.
|
| Strategy for Future Research Activity |
[Other perception experiments] We intend to conduct at least another perception experiment, in order to dissect some factors that were mixed in the previous experiment (tonal context and pitch range).
[EGG study] We will soon set up a production experiment with more speakers of Korean, as well as speakers of Japanese.
[Conclusion of the research] Based on the results from the series of experiments we conducted/will conduct, we aim to propose a model of synchronic variation and diachronic change with regards to pitch and voicing. We aim for three journal paper publications to present results and our model. Some unanalyzed data will be left for future collaboration between the JSPS fellow and the host lab.
|
