| old_uid | 16771 |
|---|
| title | Tongue positioning differences between stops and nasals: Evidence from cross-language EPG and ultrasound corpora |
|---|
| start_date | 2018/11/29 |
|---|
| schedule | 14h-15h30 |
|---|
| online | no |
|---|
| details | Séminaire du département Parole et Cognition |
|---|
| summary | Alveolar/dental stops and nasals (/t, d/ and /n/) are produced with the same lingual gesture (a tongue tip closure at the alveolar ridge or upper teeth). Yet this gesture for the nasal /n/ is known to be phonetically less stable/more variable, and highly susceptible to phonological changes, such as assimilation and deletion (e.g. Cho & Keating, 2009; Jun, 1995). Some of this behaviour of nasals has been traditionally attributed to acoustic/perceptual factors, such as less reliable acoustic cues to place of articulation for nasals (Ohala, & Ohala, 1993). Yet, previous articulatory work has suggested that nasals can be produced with weaker, less complete closures than the corresponding stops (Kohler, 1990; Gibbon et al., 2007on German and English), thus suggesting an articulatory basis of the variation. Given the somewhat limited empirical basis of previous observations, it is not clear how general the noted stop-nasal differences are, that is, whether they hold for different languages and larger samples of speakers. To explore this question, this talk presents results of a cross-language study of /t/ vs. /n/ differences based on corpora recently developed at the University of Toronto Phonetics Lab.
An examination of electropalatographic (EPG) data from 30 participants (native speakers of English, French, Japanese, Korean, Serbian, and Spanish) reveals that the syllable-initial nasal /n/ is consistently produced with a weaker linguopalatal contact than the corresponding stop /t/, regardless of the speakers’ language background. Ultrasound midsagittal tongue imaging data from 85 participants (L1 and L2 English, Gujarati, Kannada, and Punjabi) show a strong tendency towards the nasal being produced with a more front or lower position of the tongue body compared to the stop (as well as overall greater variability than in the EPG data). These, relatively subtle tongue positioning differences can have predictable acoustic consequences (as shown for Kannada) and can contribute to the merger of phonological contrasts (/n/ vs. /n/ in Punjabi).
Taken together, the results suggest that tongue configurations for stops and nasals of the same place are not necessarily identical, despite the consonants being commonly classified as having the same place. Moreover, these differences appear to be language-independent, while also showing some individual variation. The tongue positioning differences are likely to result from distinct aerodynamic and acoustic requirements for the two types of consonants (such as the absence or presence of the air pressure build-up leading to the burst release). Overall, the findings have implications for phonological (gestural or segmental) representations for stops and nasals, as well as for our understanding the sources of synchronic variability and diachronic sound change involving these consonants. |
|---|
| responsibles | Meyer, Ito |
|---|
| |