Prosodic Marking of Playful Teasing Exchanges

Document Sample
Prosodic Marking of Playful Teasing Exchanges Powered By Docstoc
					Prosodic marking of playful teasing exchanges
Debra L. Burnett and Linda M. Milosky Department of Communication Sciences and Disorders, Syracuse University
Abstract The present study seeks to contrast selected acoustic characteristics in mothers’ utterances constituting playful mislabeling teases with those that were sincere (i.e., not teasing). Playful teasing was observed between 15 children, ages 18-33 months, and their mothers and mislabel tease episodes were identified. Acoustic analyses of fundamental frequency (f0) were conducted using Praat. Overall analysis of these acoustic measures as well as analysis of the measures by sentence type (question vs. assertion) are reported.

Poster 1520 Board 196

Example of an assertion: (While reading the book “Goodnight Moon”) Mislabel tease: “Goodnight sun.” Sincere utterance: “Goodnight moon.” Findings for Assertions (see Figures 1&2): 11/15 mothers demonstrated teases in the form of an assertion. Paired t-tests (α=.05) were used to verify that the sentences were of equal length. Mislabel tease utterances averaged 1.37 seconds and sincere utterances sentences averaged 1.13 seconds (p=.25). Paired t-tests (α=.05) of the mothers’ mean f0 and range f0 during teasing and sincere utterances were not significant (p=.88, p=.74 respectively).
Figure 1. Mean fundamental frequency (f0)
325.00 300.00 275.00 250.00 225.00 200.00 175.00 150.00 125.00 100.00 75.00 50.00 25.00 0.00

Findings for Questions (see Figures 1&2): Each of the mothers demonstrated a tease in the form of a question (n=15). Paired t-tests (α=.05) were used to verify that the sentences were of equal length for both sentence types. Teases averaged 1.06 seconds and sincere utterances averaged 1.07 seconds (p=.93). A paired t-test (α=.05) of the mothers’ range f0 was significant (p=.005) indicating that mothers used a wider range during teases than during sincere utterances but mean f0 was not significantly different (p=.13).
Figure 2. Mean range of fundamental frequency (maximum f0-minimum f0)
325.00 300.00 275.00 250.00 225.00 200.00 175.00 150.00 125.00 100.00 75.00 50.00 25.00 0.00 Assertions (n=11 participants)
Mislabel tease utterances Sincere utterances
Fundamental frequency (Hz)

Playful teasing is defined as exchanges in which the speaker intentionally provokes the target in a playful, nonthreatening way with the use of verbal and/or nonverbal off-record markers such as facial expression or use of a playful tone of voice (Keltner, Capps, Kring, Young, & Heerey, 2001). One type of playful tease is object/attribute mislabeling. For example, a mother might pick up a toy dog and call it a “Kitty.” Such teases may constitute early opportunities for children to monitor comprehension (Skarakis-Doyle, 2002) and are clear exposures to early nonliteral utterances. Previous research has examined the use of prosodic markers during play and ordinary activities (Reissland & Snow, 1996) but not prosody during playful mislabeling utterances and sincere utterances. In an interview regarding teases practices, 60% of mothers indicated that they used a “special tone of voice” during teases to indicate playful intent (Burnett & Milosky, 2005). The present study used a seminaturalistic methodology to answer the following question to determine if mothers were correct in their perception that they use a special tone of voice for playful teasing: Do the prosodic characteristics of fundamental frequency (f0) differ between selected mislabeling tease utterances and sincere (i.e., not teasing) utterances?

Example of a question:
Mislabel tease: “Is that a bird?” Sincere utterance: “Is that a flower?”
Findings for Entire Corpus: The entire corpus consisted of 60 mislabel tease utterances and 60 matched sincere utterances for a total of 120 utterances. Paired t-tests (α=.05) were used to verify that the sentences were of equal length. For teases, sentences averaged 1.14 seconds and for sincere utterances, sentences averaged 1.10 seconds (p=.58). A paired t-test (α=.05) of the mothers’ mean f0 during teasing and sincere utterances approached significance (p=.047). A paired t-test (α=.05) of the mothers’ range f0 (maximum-minimum) was significant (p=.02) indicating that mothers used a wider pitch range during teases than during sincere utterances.

A subset of 15 dyads, children ages 18 through 33 months (M=24), from urban and suburban areas of the greater Syracuse area was used for the current analyses since those mothers demonstrated mislabel teases. These children were part of a larger study of 31 dyads examining playful teasing between mothers and young children (Burnett & Milosky,2005) Each mother-child dyad was seen for one two-hour session at home involving an examiner-generated interview regarding teasing practices (Heerey, Capps, Keltner, & Kring, 2005) and 20-45 minutes of play in which the mother demonstrated typical playful teasing exchanges. Sessions were videotaped and analyzed for the presence of playful teases. Of the different types identified (Interrater reliability was 80% for identifying tease types), object/attribute mislabel teases (e.g. calling a dog a “kitty”) were selected for the current analysis. Each mother wore an external lapel microphone to capture optimal audio recording of her voice. Four mislabel teases were chosen for each mother (n=60) and a sincere utterance was matched to each tease utterance (n=60) for approximate length in words, utterance type (i.e., assertion or question) and when possible, initial semantic content (e.g., “Is that a cow?” matched to “Is that a fish?”). All utterances chosen were free from excessive background noise (e.g., loud toys). Analyses using Praat (Boersma & Weenink, 2006; Wood, 2005):

Fundamental frequency (Hz)

Assertions (n=11 participants) Questions (n=15 participants) Utterance type Mislabel tease utterances Sincere utterances

Questions (n=15 participants)

Utterance type

 Mothers demonstrated more pitch variation (range f0) during mislabel tease utterances than during sincere utterances.  Pitch variability, as measured by f0 range, differed for questions but not for assertions.  While average pitch did not differ between mislabel tease utterances and sincere utterances, the use of a wider pitch range demonstrates that mothers are correct in their assumption that they use a “special tone of voice” when teasing their children.

Boersma, P. and Weenink, D. (2006). Praat: doing phonetics by computer (Version 4.4.07) [Computer program]. Retrieved from Burnett, D. and Milosky, L. (2005). Language-based playful teasing between mothers and young children. Poster presented at ASHA Convention, San Diego, CA.. Heerey, E., Capps, L., Keltner, D. and Kring, A. (2005). Understanding teases: lessons from children with autism. Journal of Abnormal Child Psychology, 33, 55-68. Keltner, D., Capps, L., Kring, A., Young, R., and Heerey, E. (2001). Just teasing: a conceptual analysis and empirical review. Psychological Bulletin, 127, 2, 229-248. Reissland, N. and Snow, D. (1996). Maternal pitch height in ordinary and play situations. Journal of Child Language, 23, 269- 278. Skarakis-Doyle, E. (2002). Young children’s detection of violations in familiar stories and emerging comprehension monitoring. Discourse Processes, 33, 2, 175-197. Wood, S. (2005). Praat for beginners [Manual]. Retrieved from /Sidney/praate/

o Audio files were digitized using Praat from Panasonic Mini-DV Camcorder recordings using a 32000 Hz sampling rate.
o Each utterance was saved as a WAV file and then analyzed for mean f0, minimum f0, maximum f0, and standard deviation of f0. o The pitch extraction algorithm reported fundamental frequency (f0) values for each 0.01 seconds of voiced speech.

The research was supported in part by an ASHA Special Interest Division 1: Language Learning and Education Student Research Grant to the first author. Special thanks to Stefanie Hayes and Rachel Neuman for their help with data collection and to Hannah Burke, Amy Brown, Emily DeSalvo and Lindsay Rapke for all their help with data analysis.

For further information contact: Debra Burnett at

Shared By: