Prosodic Interaction Between Speakers of American and British English
Jelena Krivokapic Linguistics Department
Yale University [email protected]
2pSC2
Overview
Experimental study of speaker convergence between speakers of different dialects Examining prosodic and segmental characteristics
Synchronous Speech Paradigm (Cummins 2002, Cummins 2003)
Krivokapic, ASA Baltimore 2010 2
Background: Speaker convergence
Speakers’ production converges in the course of a conversation (Pardo 2006) Speakers’ productions of lexical items are more similar to their co-
speaker’s when produced during a conversation than when produced before or after the task.
Speakers’ production changes depending on the language environment (Sancier & Fowler 1997) A bilingual speaker’s VOT shifted in both of her languages
depending on the country she was staying in. The speaker’s Portuguese VOT shifted towards the VOT of English when in the US, and her English VOT production shifted towards Portuguese when in Brazil (Sancier & Fowler 1997).
In a similar study examining Spanish speakers living in the US, convergence was found for VOT, but not for F2 in vowels (Tobin 2009).
Krivokapic, ASA Baltimore 2010 3
Convergence across dialects
Ní Chiosáin (2007) examines interaction between two dialects of Irish, using the synchronous speech paradigm. Investigated variables: lexical stress, vowel duration,
lenition Only small effects of convergence found, mainly for
lexical stress
Krivokapic, ASA Baltimore 2010 4
Convergence as Gestural Drift
Convergence can be explained by speakers’ inclination towards imitation (Sancier & Fowler 1997, Kuhl & Meltzoff 1996).
Perception of articulatory gestures drives production, leading to automatic gestural attunement (Fowler, 1986, 1996, Fowler et al. 2003, Goldstein & Fowler 2003, Sancier & Fowler 1997) Gestural drift: “perceptually guided changes in speech
production” (Sancier & Fowler 1997)
Can form the basis for continuous language learning and accent change
Krivokapic, ASA Baltimore 2010 5
Synchronous speech paradigm (Cummins 2002, Zvonik & Cummins 2002, 2003)
Minimizes individual, non-linguistic variation without inducing artificial temporal properties
Dyads read a text simultaneously seated facing each other and recorded on stereo channels
with head-mounted microphones
Captures in a unique way speakers’ shared knowledge of linguistic timing (Cummins 2002)
Reduces variability in F0 contour (Kim & Nam 2009)
Krivokapic, ASA Baltimore 2010 6
Outstanding research questions
Nature of speaker convergence What is the effect of interaction between speakers of
different dialects (British and American English)? Examining whether convergence occurs across dialects
How are prosodic characteristics affected? Examining whether convergence occurs in
suprasegmental properties and whether it differs from convergence in phonemes
Synchronous Speech Paradigm (Cummins 2002, Cummins 2003) as a means to investigate speaker interaction
Krivokapic, ASA Baltimore 2010 7
Acoustic experiment
Part of a larger study
Here: A short story containing 10 test words differing between the dialects in vowel quality and 4 words differing in stress placement Test words from Wells 1982, Berg 1999, story
adapted from Honorof, McCullough & Somerville 2000.
Subjects read 6 repetitions of the story, with 84 filler sentences between each repetition.
Recordings on separate days for solo and synchronous condition solo condition always first 8
Stimuli The nurse took a bath when she woke up. Then she put on a plain yellow dress and a fleece jacket, picked up her goose Nico, and headed north to work. She worked with her father in a company that produced cloth for cleaning crystal. But a strange thing happened when she walked out of the door: She saw a goat standing near the big garage where she kept her coffee and tools. That reminded her of a story where a dictator had a square hat and always talked about caffeine at press conferences. She was a young adult when she first heard the story, and her favorite meal was tuna with parsley.
9 Adapted from Honorof, McCullough & Somerville 2000
Subjects
8 speakers (4 dyads) 4 native speakers of American English 4 native speakers of British English
Length of stay in the US: recently arrived: 3 months (subject C), 1 year (subject F)
long time residents: 5 years (subject A), 19 years (subject H)
Krivokapic, ASA Baltimore 2010 10
Measurements
Segmental convergence: vowel quality
Prosodic convergence: stress placement, intonation contour
IPA transcription for vowel quality and stress placement
Vowel quality: F1, F2 (for ‘bath’, ‘dress’, ‘cloth’, ‘goat’) F3 to examine r-coloring
F0: intonation contour comparisons 11
Statistical analysis
Two-factor ANOVA on F1, F2, and F3 data, testing for each speaker separately the effects of 1 Speaking condition (solo and synchronous)
2 Test word
Criterion for significant difference p < .05
Krivokapic, ASA Baltimore 2010 12
Results: Observations
One dyad was excluded from the analysis The speakers did not synchronize well enough, with
one subject speaking much faster than the other, and the other attempting to match the speed.
For different dialects the synchronous speech task was difficult especially for recently arrived speakers and during the
short story
many errors, comments by the subjects not typical for synchronous speech task
Krivokapic, ASA Baltimore 2010 13
Results: Vowel quality I
Transcription showed no categorical changes in vowels
F1 and F2: 3 subjects showed changes on F1 on some test words 5 subjects showed changes on F2 on some test words
In all but one case, the change involves lowering of the formant. In terms of convergence: inconsistent in direction, with
same subjects showing both converging and non-converging direction of change.
Krivokapic, ASA Baltimore 2010 14
Results: Vowel quality II
R-coloring
Two subjects showed an effect of condition and word (but no interaction), and one an effect of word and a trend of condition effect (all in the same direction)
British subjects
One subject (F) showed F3 lowering and one (C) a trend to F3 lowering => convergence for recently arrived subjects
American subject
Subject D showed F3 raising => convergence Krivokapic, ASA Baltimore 2010 15
Results: Stress
Solo condition: speakers produced expected stress patterns except for one British subject (A), who produced ‘garage’ with stress on the second syllable.
Synchronous condition:
British speakers
Subject C: twice stress placement on ‘caffeine’ unclear
Subject A: three times restart on ‘dictator’, once stress on first syllable
American speakers
Subject D: once stress placement on ‘caffeine’ unclear, three times stress on first syllable 16
Time (s)30.6898 33.4547
0
500
Time (s)32.4749 35.248
0
500
Time (s)18.4012 21.1803
0
500
Time (s)16.8779 19.4852
0
500
Time (s)19.3601 22.0318
0
500
0
0.2
0.4
0.6
0.8
1
Results: Intonation contour I Dyad AB (one sentence mid story)
17
American solo (B)
American synchronous
British solo (A)
British synchronous
0
0.2
0.4
0.6
0.8
1
Time (s)31.1616 33.7502
0
500
Time (s)32.9442 35.4308
0
500
Time (s)87.5169 90.0073
0
500
Time (s)15.853 18.3642
0
500
Time (s)15.8908 18.229
0
500
Time 0
0.2
0.4
0.6
0.8
1
Time (s)30.5468 32.9827
0
500
Time (s)16.6424 18.8347
0
500
Time (s)17.8006 20.1288
0
500
Time (s)15.66 17.91230
500
Time (s)15.2365 17.5167
0
500
Time
0
0.2
0.4
0.6
0.8
1
Time (s)30.6603 33.5096
0
500
Time (s)32.4583 35.4151
0
500
Time (s)18.3596 21.239
0
500
Time (s)16.8605 19.5109
0
500
Time (s)19.3411 22.1125
0
500
500
400
300
200
100
0
500
400
300
200
100
0
500
400
300
200
100
0
500
400
300
200
100
0
Time (s)26.3357 28.6217
0
500
Time (s)30.287 32.5933
0
500
Time (s)15.2943 17.5765
0
500
Time (s)16.1474 18.2666
0
500
Time (s)14.6726 16.8209
0
500
Time (s)15.9933 18.5413
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Time Time (s)16.2593 18.6469
0
500
Time (s)16.1251 18.406
0
500
Time (s)15.3339 17.4403
0
500
Time (s)15.067 17.5512
0
500
Time (s)15.067 17.5512
0
500
Time (s)17.3727 19.7441
0
500
Time (s)19.7265 22.0872
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Time
Time (s)51.9916 54.4612
0
500
Time (s)17.4509 19.8005
0
500
Time (s)30.2112 32.7716
0
500
Time (s)14.7529 17.0592
0
500
Time (s)15.5555 17.8582
0
500
Time (s)15.4077 17.8154
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0 Time (s)
52.0231 54.43590
500
Time (s)17.4509 19.8005
0
500
Time (s)30.1885 32.7698
0
500
Time (s)14.8255 17.1004
0
500
Time (s)15.5659 17.8548
0
500
Time (s)15.5303 17.8063
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Results: Intonation contour II Dyad CD (one sentence mid story)
18
American solo (D)
American synchronous
British solo (C)
British synchronous
Time (s)14.705 18.6203
0
500
Time (s)17.2023 19.7739
0
500
Time (s)18.2978 20.8897
0
500
Time (s)20.4196 23.0736
0
500
Time (s)22.6834 25.3601
0
500
Time (s)14.705 18.6203
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0 Time (s)
14.5147 18.6380
500
Time (s)22.5087 25.3383
0
500
Time (s)20.2668 23.0828
0
500
Time (s)18.1716 20.8611
0
500
Time (s)17.1325 19.7299
0
500
Time (s)18.2641 20.9017
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Time (s)21.5178 24.334
0
500
Time (s)20.0607 22.7106
0
500
Time (s)18.8607 21.4178
0
500
Time (s)22.2443 25.0453
0
500
Time (s)20.2971 22.8482
0
500
Time (s)26.6112 29.2063
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Time
Results: Intonation contour III Dyad FG (one sentence mid story)
19
American solo (G)
American synchronous
British solo (F)
British synchronous
Time (s)22.3666 24.8764
0
500
Time (s)20.5939 22.902
0
500
Time (s)17.5518 19.9165
0
500
Time (s)15.9109 18.1992
0
500
Time (s)15.9122 18.3213
0
500
Time (s)17.4181 19.7782
0
500
0
0.2
0.4
0.6
0.8
1500
400
300
200
100
0
Time
Summary & Conclusions Interaction leads to convergence for both segmental and
prosodic properties, although only for some subjects. Subjects C, F (recently arrived British subjects) and D (American)
show convergence for vowel quality F3 (r-coloring), but not F1 and F2, shows effects of interaction
Subjects A, C (British), and D (American) show convergence for stress
Intonational contours show indication of reduced variability
British subjects converged more than American subjects. Most affected were the subjects who had recently arrived from Britain.
Synchronous speech can be used for examining convergence, but it is a difficult task for speakers of different dialects and therefore might interfere with the process of convergence.
20
References Berg, T. 1999. Stress variation in British and American English. World Englishes, 18, 123–143. Cummins, F. 2002. On synchronous speech. Acoustic Research Letter Online, 3, 7-11. Cummins, F. 2003. Practice and performance in speech produced synchronously. Journal of Phonetics, 31, 139-148. Fowler, C. A. 1986. An event approach to the study of speech perception from a direct-realist perspective, Journal of
Phonetics, 14, 3-28 Fowler, C. A. 1996. Listeners do hear sounds not tongues, J. Acoustical Society of America, 99, 1730-1741. Fowler, C. A., Brown, J. M., Sabadini, L., & J. Weihing. 2003. Rapid access to speech gestures in perception: Evidence
from choice and simple response time tasks. Journal of Memory & Language, 49, 396-413. Goldstein, L. M., & C. Fowler. 2003. Articulatory phonology: a phonology for public language use. In Phonetics and
Phonology in Language Comprehension and Production: Differences and Similarities, (Eds.) Antje S. Meyer &Niels O. Schiller. Honorof, D., McCullough, J. & B. Somerville. 2000. Available at: http://web.ku.edu/~idea/readings/comma.htm Kim, M. & H. Nam 2009. Pitch accommodation in synchronous speech, J. Acoustical Society of America, 125:4, 2575. Kuhl, P. & A. Meltzoff. 1996. Infant vocalization in response to speech: vocal imitation and developmental change,
Journal of the Acoustical Society of America, 100, 2425-2438. Ní Chiosáin, M. 2007. Effects of synchronous speech task on length and prosody in interdialecal nonprestige varieties''.
Language variation and change, 19, 51-62. Pardo, J. 2006. On phonetic convergence during conversational interaction. J. Acoustical Society of America, 119, 2382-2393 Sancier, M. L. & C. A. Fowler. 1997. “Gestural drift in a bilingual speaker of Brazilian Portuguese and English,” Journal of
Phonetics, 25, 421–436. Tobin, S. 2009. Gestural drift in Spanish-English speakers. J. Acoustical Society of America, 125:4, 2757. Wells, J. C. 1982. Accents of English. New York: CUP. Zvonik, E., & F. Cummins (2002). Pause duration and variability in read texts. In: Proceedings of the 2002 International
Conference on Spoken Language Processing. Denver, Colorado, pp. 1109-1112. Zvonik, E., & F. Cummins (2003). The effect of surrounding phrase lengths on pause duration. In: Proceedings of
EUROSPEECH 2003. Geneva, Switzerland, pp. 777-780.
21