1. An ultrasound study of the trough effect in VhV sequences Natalia Zharkova Queen Margaret University College, Speech and Hearing Sciences nzharkova@qmuc.ac.uk Ultrafest III, Arizona 16 April 2005

2. Triggered by:  1. Recent research on troughs in VCV sequences like /aba/  2. What we know about /h/:

3. /h/  /h/ is unspecified for tongue position (e.g. Keating 1988, Pierrehumbert & Talkin 1992, Ladefoged 2001, Karbownicki 2004)  /h/ is also unspecified for lip and jaw position…   /h/ has a lot of freedom for coarticulation

4. Troughs  A “trough”, or a lowering of the tongue, has been found in bilabial consonants surrounded by identical vowels (e.g. Houde 1967, Gay 1974, Gay & Ushijima 1974, Bell-Berti & Harris 1974, Engstrand 1988, Svirsky et al. 1997, Lindblom et al. 2002, Fuchs et al. 2004, Vazquez Alvarez, Hewlett & Zharkova 2004)  Bilabial consonants, like /h/, are considered unspecified for a particular tongue position

5.  So: interesting to see what happens when /h/, which is even more unspecified than bilabials, is between two identical vowels

6. Questions:  What would be the pattern of tongue behaviour during VhVs? Specifically:  Does the tongue maintain the same position throughout the VhV sequence?  If not – what differences occur?  E.g. is there a trough on /h/? Is the V1 position different from the V2 position?

7. Data collection  QMUC ultrasound system  three native British English speakers  data = /ihi/, /uhu/, /aha/  carrier phrase “I said … too” (“eehee”, “oohoo”, “aha”)  sixteen times each

8. /aha/

9. /ihi/

10. /uhu/

11.  Creating three annotations: mid /h/, V1, V2 Analysis mid V1mid /h/same distance /h/ - V2

12.  Creating three splines – V1, /h/, V2: Analysis

16.  splines superimposed on each other: Analysis

17. Typical tongue contours during /uhu/ mid V1 mid V2 mid C

18.  1. Comparing occurrence of different tongue shape patterns Analysis

19. Distances along vertical measure bar: V1 – CC – V2 V1 V2 /h/  2. Measuring tongue movements throughout VhVs

20.  extracting xy spline coordinates from US analysis software  importing xy values into Matlab  3. Comparing whole contour shapes

21. Black solid line – V1 Red solid line – /h/ Blue dashed line – V2 Typical tongue shape pattern during /uhu/

22.  Calculating the distance from each point on the C curve to its nearest neighbour on the V1 curve and separately on the V2 curve  Plotting these distances Distances between V1, C and V2 curves

23. Black solid line –V1Red solid line – /h/ Blue dashed line – V2

24. Results 1. Comparing occurrence of different tongue shape patterns  Trough (highest point of C below both VV)  Antitrough (highest point of C above both VV)  Neutral (highest point of C between two VV)

25. Tongue shape patterns distribution by vowel

26. Results  2. Distances of tongue movement throughout VhVs

27. Distances of tongue movement Very small distances !!!!!!!!!!!!

28. Significant differences in tongue displacement sizes – no significant differences in tongue displacement sizes – /i/ vs /u/ /a/ vs /i/, /a/ vs /u/

29.  3. Comparison of whole contour shapes tongue contours, by vowel… Results

30. Average tongue shape pattern during /ihi/ Black solid line – V1 Red solid line – /h/ Blue dashed line – V2 Average tongue shape pattern during /ihi/

31. Distances between V1, C and V2 curves, /ihi/ V1 /h/ V2 Middle part of the tongue typically lowers during the consonant !“Trough”!

32. Distances between V1, C and V2 curves, /ihi/ B ack part of the tongue on average moves slightly backwards during the consonant V1 /h/ V2 !Relaxation of Advanced Tongue Root!

33. Average tongue shape pattern during /uhu/ Black solid line – V1 Red solid line – /h/ Blue dashed line – V2 Average tongue shape pattern during /uhu/

34. Distances between V1, C and V2 curves, /uhu/ V1 /h/ V2 Middle part of the tongue typically lowers during the consonant !“Trough”!

35. Distances between V1, C and V2 curves, /uhu/ Tongue typically goes backwards from V1 into C, and forwards again for V2 V1 /h/ V2 !Relaxation of Advanced Tongue Root!

36. Average tongue shape pattern during /aha/ Black solid line – V1 Red solid line – /h/ Blue dashed line – V2 Average tongue shape pattern during /aha/

37. /aha/  One obvious and rather consistent pattern: /aha/

38. Distances between V1, C and V2 curves, /aha/ Front part of the tongue is on average lower for V2 than for V1 V1 /h/ V2 !Second syllable stressed!

39. /aha/  fewer number of troughs and their significantly smaller size in /a/ than in the other two vowels possible explanation: for the open vowel /a/ raising, rather than lowering, would be expected during tongue deactivation (Lindblom et al. 2002, Vazquez Alvarez, Hewlett & Zharkova 2004) /aha/

40. Distances between V1, C and V2 curves ihi aha V1 /h/ V2 uhu

41. Distances between V1, C and V2 curves ihi aha V1 /h/ V2 uhu

42. Distances between V1, C and V2 curves ihi aha V1 /h/ V2 uhu

43. Differences between V1 and V2  On average V1 is further away from C than V2, suggesting a syllable boundary influence and showing asymmetrical nature of VCV: ihi uhu aha V1-C:0,539 0,582 0,362 V2-C:0,387 0,519 0,310

44. Conclusions  Tongue is in a very similar position for both vowels and /h/  However, there is some evidence that /h/ is more like V2 than like V1: a syllable boundary effect

45.  Some evidence for troughs, but they are small  More troughs in /i/ and /u/ contexts than in /a/ context  Troughs/antitroughs mainly occur in mid and back parts of the tongue  Front of the tongue – continuous movement from V1 to V2 Conclusions

46. Implications for the future  Why these patterns?  May be some properties of /h/  May be due to syllable boundary within the VhV sequence  May be due to stress position and its physical characteristics … Future research…..

47. Bell-Berti, F. & Harris, K.S. (1974). More on the motor organization of speech gestures. Haskins Labs. Status Rep. Speech Res., SR-37/38, pp. 73-77. Engstrand, O. (1988). Articulatory correlates of stress and speaking rate in Swedish VCV utterances. Journal of the Acoustical Society of America, 83, pp. 1863-1875. Fuchs, S., Hoole, P., Brunner, J. & Inoue, M. (2004). The trough effect – an aerodynamic phenomenon? [Oral presentation, “From Sound to Sense”, 11-13 June 2004, MIT.] Gay, T. (1974). Some electromyographic measures of coarticulation in VCV utterances. Haskins Labs. Status Rep. Speech Res., SR-44, pp. 137-145. Gay, T. & Ushijima, T. (1974). Effect of speaking rate on stop consonant-vowel articulation. Speech Commun. Semin., Stockh., SCS-74, pp. 205-208. Houde, R.A. (1967). A study of tongue motion during selected speech sounds. PhD diss. Speech Commun. Res. Lab., Santa Barbara, Monogr. No. 2. Karbownicki, L. (2004). Investigation of the coarticulation effects on [h] when preceding a vowel. BSc, Honours project, Queen Margaret University College. Keating, P.A. (1988). Underspecification in phonetics. Phonology 5.2, pp. 275- 292. REFERENCESREFERENCES

48. Kozhevnikov, V.A. & Chistovich, L.A. (1965). Rech: Artikulyatsiya i vospriyatiye (Speech: Articulation and perception). Moscow-Leningrad. Translation: Kozhevnikov, V.A. & Chistovich, L.A. (1965). Speech: Articulation and perception, No. 30, p. 543 (Joint Pub. Res. Service, Washington). Ladefoged, P. (2001). A Course in Phonetics. 4th edn. Orlando, FL: Harcourt College Publishers. Lindblom, B., Sussman, H.M., Modaressi, G. & Burlingame, E. (2002). The trough effect: Implications for speech motor programming. Phonetica, 59, pp. 245-262. Perkell, J. (1986). Coarticulation strategies: preliminary implications of a detailed analysis of lower lip protrusion movements. Speech Communication, 5, pp. 47-68. Pierrehumbert, J. & Talkin, D. (1992). Lenition of [h] and glottal stop. In J. Docherty & D.R. Ladd (eds.), Papers in Laboratory Phonology II: Gesture, Segment, Prosody. Cambridge: Cambridge University Press. Pp. 90-117. Svirsky, M., Stevens, K., Matthies, M., Manzella, J., Perkell, J. & Wilhelms- Tricarico, R. (1997). Tongue surface displacement during bilabial stops. Journal of the Acoustical Society of America, 102, pp. 562-571. Vazquez Alvarez, Y., Hewlett, N., & Zharkova, N. (2004). An ultrasound study of the "Trough Effect". [Poster at the British Association of Academic Phoneticians Colloquium 2004, University of Cambridge, Cambridge, UK.] REFERENCESREFERENCES