- Email: [email protected]
Effect of vowel environment on fricative consonant duration in speech produced during simultaneous communication
EFFECT OF VOWEL ENVIRONMENT ON FRICATIVE CONSONANT DURATION IN SPEECH PRODUCED DURING SIMULTANEOUS COMMUNICATION ROBERT L. WHITEHEAD and BRENDA H. WHITEHEAD National Technical Institute for the Deaf, Rochester, New York
NICHOLAS SCHIAVETTI, DALE EVAN METZ, and KIMBERLY FARINELLA State University of New York, Geneseo, New York
This study investigated the effect of vowel environment on fricative consonant duration in contextual speech produced during simultaneous communication (SC). Previous studies (Schwartz, 1969) of vowel influences on consonant duration supported the notion of anticipatory scanning, in which final vowel targets influence the duration of preceding fricative consonants. Ten normal-hearing, experienced sign language users recorded palatal and alveolar fricatives produced in four vowel environments in contextual sentences under SC and speech-only (SO) conditions. Results indicated longer sentence durations for SC than for SO, and significant effects of vowel context on fricative consonant duration in contextual speech in both SC and SO conditions that revealed similar anticipatory scanning effects as seen in previous studies. These data confirm previous research indicating that the temporal alterations produced by simultaneous communication do not involve violations of the temporal rules of English speech. © 1999 Elsevier Science Inc. Educational Objectives: The reader will (1) acquire knowledge and understanding of simultaneous communication and its role in communication with children who are deaf; and (2) understand the relationships among speech rate, fricative duration, and temporal changes in speech during simultaneous communication. KEY WORDS: Simultaneous communication; Fricatives; Anticipatory scanning
Address correspondence to Robert L. Whitehead, National Technical Institute for the Deaf at RIT, Carey 2408, 96 Lomb Memorial Drive, Rochester, NY 14623-5604. Tel: 716-475-6457; Fax: 716-4757101; E-mail: ,[email protected]..
J. COMMUN. DISORD. 32 (1999), 423–434 © 1999 by Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
0021-9924/99/$–see front matter PII S0021-9924(99)00019-2
424
WHITEHEAD et al.
Previous research considered temporal changes in speech produced during simultaneous communication (SC), when speech is combined with manually coded English (sign and finger-spelling) for the production of each word of an utterance using both spoken and manual modalities (Whitehead, Schiavetti, Whitehead, & Metz, 1995). Reasons for the use of SC with children who are deaf include the enhancement of language development and exposure to segmental and suprasegmental features of normal speech, although some researchers have criticized SC as presenting an unnatural speech model (Huntington & Watton, 1984; Marmor & Petitto, 1979; Strong & Charlson, 1987; Vernon & Andrews, 1990). Tye-Murray, Spencer, and Woodworth (1995) reported that children who used SC before a cochlear implant did not discontinue the use of sign language after implantation. Their results stress the importance of research on the effects of SC on speech presented to deaf children to determine the potential influences of SC use on their speech and language development. Research on the temporal characteristics of speech produced during SC is necessary to determine the degree to which SC influences the speech models presented by persons with normal hearing to those with hearing impairment. For example, Hyde, Power, and Leigh (1998, p. 117) have raised questions about the disturbed prosodic quality of speech during SC “that may influence the acceptability of such speech as an auditory-oral model for deaf students.” Research with experienced and inexperienced signers has demonstrated that speech produced during SC is slower in rate than speech produced alone and has longer interword intervals and vowel durations (Whitehead et al., 1995; Whitehead, Schiavetti, Metz, & Farrell, 1999), and longer voice onset times (VOT) (Schiavetti, Whitehead, Metz, & Moore, 1999; Schiavetti, Whitehead, Metz, Whitehead, & Mignerey, 1996). These temporal modifications are the result of attempting to maintain simultaneity between the spoken and manual modes of communication (Windsor & Fristoe, 1989; 1991). Research has demonstrated that even though there are temporal elongations in speech produced during SC, the temporal rules of spoken English that have been studied so far are maintained, thereby providing an accurate speech model to deaf children for these specific parameters. These temporal rules include the effect of vowel height and voicing characteristics of the following consonants on vowel duration (Whitehead et al., 1995; 1999) and VOT differences between voiced and voiceless members of cognate pairs of plosives (Schiavetti, et al., 1996; 1999). The segmental measures investigated in these studies mainly concern vowels or the time intervals between consonants and vowels, but none of these studies investigated the effect of SC on specific consonant characteristics, such as duration or spectrum. The importance of studying consonantal effects of SC is highlighted by the classic statement of Hirsh (1952, p. 125) that “The consonants of English seem to be responsible for most of the intelligibility
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
425
carried by individual English words.” If SC disrupts the temporal or spectral patterns of consonants, then serious intelligibility repercussions could result from using this mode to communicate with deaf children who are attempting to develop oral language. A temporal rule of spoken English that has not been studied in the speech of persons using SC is the influence of vowels on consonant duration. This effect was labeled “anticipatory scanning” by Schwartz (1969) as a result of his study of consonant durational differences in repeated nonsense syllables, and has recently been shown to occur in contextual speech produced by adult male and female speakers (Weglarski, Sewall, Whitehead, Schiavetti, & Metz, 1999). This temporal rule suggests that alveolar and palatal fricatives are longer in duration when the following vowel is a high front vowel rather than a low back vowel. Schwartz (1969) theorized that the difference is the result of the smaller distance the tongue must travel from the consonant place of articulation to the position for the following high front vowel in the CV syllable /si/, compared with the larger distance the tongue must travel to the position for the following low back vowel in the CV syllable /sa/. The purpose of this study was to examine the effect of vowel environment on the duration of fricative consonants produced by adults in contextual speech when producing speech only, versus SC. The durations of alveolar and palatal fricatives preceding four different vowels were examined to study the degree to which anticipatory scanning effects seen in previous studies (Schwartz, 1969; Weglarski et al., 1999) that used speech-only conditions would be observed in speech produced during SC as well.
METHOD Speakers Speakers for this study were ten normal-hearing adults (5 females and 5 males), all of whom were faculty members at the National Technical Institute for the Deaf. The speakers had taught young deaf adults for at least eight years, using simultaneous communication in all classes. Each speaker’s sign language performance had been evaluated on the Sign Communication Proficiency Index (Caccamise, Updegraff, & Newell, 1990), and all speakers were classified at the advanced level or higher on this instrument. Thus, the speakers were considered to be fluent in the use of speech combined with signed English and finger-spelling.
SPEECH MATERIALS The speech samples investigated consisted of the carrier sentence, “I can say ___________ again clearly” and eight experimental words that varied in the
426
WHITEHEAD et al.
initial fricative and following vowel embedded in the blank slot in the carrier sentence. Two fricatives were studied: alveolar /s/ and palatal /ʃ/ preceding each of four vowels: high front /i/, low front /{/, low back /a/, and high back / u/. The experimental words included seat, sheet, sat, shack, sock, shock, suit, and shoot.
Recording Procedures The experimental sentences were spoken at a comfortable conversational loudness level as they were presented on flashcards. Audio recordings were made in a sound-treated booth using an Audio-Technica AT-816 microphone that was placed 15 cm from each speaker’s mouth and was connected to a Tascam 202MKII tape deck. The speakers produced each sentence, with its embedded experimental word, under two conditions: (a) speech only; and (b) speech combined with signed English for all words in the sentence except the experimental word, which was finger-spelled. The sentences containing the experimental words were presented to the speakers on flashcards in two different random orders, one for each experimental condition; the order of experimental condition (speech-only versus SC) was counterbalanced across speakers. The speakers were shown the experimental words before the recording, so they could familiarize themselves with the signs and finger-spellings to be used in SC.
Acoustic Analysis Procedures For each speech sample, duration measures in milliseconds were determined for the total sentence and the initial fricative consonant in the experimental word. The acoustic signal from the audio recording of each sentence was digitized with 16-bit precision at 20 kHz using Kay Elemetrics Computerized Speech Lab (CSL; Model 4300B). When the digitizing process is initiated, CSL applies an appropriate internal low-pass anti-aliasing filter to the raw acoustic signal (at a digitizing rate of 20 kHz, the upper frequency cut-off is 8 kHz), stores the digital results in memory, and displays the resultant waveform on a VGA graphics monitor. We measured total sentence duration in the two experimental conditions to verify that speech rate was slower in SC than in the speech-only condition. Total sentence duration was measured by visually isolating the first positively going portion of the waveform associated with the initiation of the diphthong /a/ and marking the location with a cursor. The cursor was then moved to the end of the sentence, and the last positive going portion of the waveform associated with the /i/ segment in the word “clearly” was visually isolated and marked with a cursor. The temporal interval between the two cursors was taken as the value for total sentence duration.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
427
Fricative consonant duration was measured by visually isolating the initiation of the aperiodic waveform following the diphthong /e/ in the word “say” and marking the location with a cursor. The cursor was then moved to the end of the aperiodic waveform, and the temporal interval between the two cursors was taken as the value for fricative consonant duration. As a measure of intraobserver reliability of the measurement procedure, the recordings of one male and one female speaker were selected at random, and all the fricative durations were measured a second time by the same person who made the first measurement. The mean difference between the first and second measure for the male speaker was was 1.89 msec (SD 5 0.58 msec) and the mean difference between the first and second measure for the female speaker was was 2.12 msec (SD 5 0.86 msec). These low mean differences generally indicate sufficient intraobserver reliability for these measures. As a measure of intraspeaker reliability, one male and one female speaker were selected at random to record all speech materials twice. The second recordings were digitized on the CSL, and all fricative durations were measured for comparison with the original recordings. For each speaker, the repeated measures for each token were subtracted from the original measure of each token to examine the absolute value of the difference between the two recordings. For the female speaker, the mean fricative duration difference for the speech-only condition was 3.84 msec (SD 5 1.76 msec) and for the SC condition was 6.93 msec (SD 5 4.93 msec). For the male speaker, the mean fricative duration difference for the speech-only condition was 4.31 msec (SD 5 2.11 msec) and for the SC condition was 8.62 msec (SD 5 5.88 msec). These low mean differences in fricative duration generally indicated sufficient intraspeaker reliability for these measures.
Statistical Analysis of Duration Data Statistical analyses of the sentence duration and fricative data were accomplished with a three-way (2 3 2 3 4) analysis of variance (ANOVA) with repeated measures on all independent variables: (a) experimental condition (two levels: speech-only versus SC); (b) fricative place of articulation (two levels: alveolar versus palatal); and (c) vowel (four levels: high front /i/ versus low front /{/ versus low back /a/ versus high back /u/).
RESULTS Sentence Duration The ANOVA revealed a significant effect of communication mode on sentence duration (F 5 33.66; df 5 1,9; p , .001) . No other main effects or interaction effects on sentence duration were significant. Figure 1 presents the
Figure 1. Means and standard deviations of durations of sentences containing experimental words with initial alveolar and palatal fricatives preceding each of four vowels in speech-only vs. SC conditions.
428 WHITEHEAD et al.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
429
means and standard deviations of the sentence duration in speech-only and SC for both fricative places of articulation preceding each of the four vowels. Inspection of Figure 1 reveals that sentence duration was always longer in SC than in SO, indicating a slowing of speech while speakers attempted to sign and speak at the same time.
Fricative Duration The ANOVA revealed significant effects of communication mode (F 5 64.27; df 5 1,9; p , .0001) and vowel (F 5 94.71; df 5 3,27; p , .0001) on fricative duration. No other main effects or interaction effects on sentence duration were significant. Direct means contrast tests comparing the main effect of vowel environments (collapsed across experimental condition and place of articulation) indicated that the durations of the fricatives preceding each vowel were significantly different from the durations of the fricatives preceding each of the other three vowels. Figure 2 illustrates the means and standard deviations of the durations of the two initial fricatives in each of the four vowel environments for both the speech-only and SC conditions. Inspection of Figure 2 reveals longer fricative durations in SC than in the speech only condition. In addition, the fricative duration was greater preceding high and front vowels when compared with durations preceding low and back vowels.
DISCUSSION The results of the present investigation are consistent with previous research demonstrating that speech is elongated in duration when it is combined with signed English and finger-spelling in simultaneous communication (Whitehead, et al., 1995). Specifically, for the dependent variables of sentence and fricative duration, there was a consistent pattern of significant durational differences between the speech-only and SC conditions; that is, sentences and fricatives were always significantly longer in duration in the SC condition than the speech-only condition. The combination of two communication systems that operate at different rates results in a retardation of the more rapid articulatory act of speech in order to maintain a high degree of simultaneity (Windsor & Fristoe, 1989; 1991). However, despite the increase in duration of fricatives during SC, the speakers maintained the temporal rule that alveolar and palatal fricatives are longer in duration when the following vowel is a high front vowel compared with a low back vowel. Thus, during SC these speakers demonstrated anticipatory scanning of subsequent vowel anatomical position, as was reported by Schwartz (1969) for nonsense syllables produced in a speech-only condition. In addition, they showed the same pattern that was recently reported by Weglarski et al. (1999) for contextual speech, in which both subsequent vowel
Figure 2. Means and standard deviations of durations of alveolar and palatal fricatives preceding each of four vowels in speech-only vs. SC conditions.
430 WHITEHEAD et al.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
431
height and front-back advancement influenced the duration of preceding fricatives. These results are consistent with previous studies that reported temporal elongation during SC but maintenance of the temporal rules of spoken English in the relative durations of various vowels in different consonant context (Whitehead et al., 1995; 1999), voice-onset times of voiced and voiceless cognates (Schiavetti et al., 1996; 1999), and relative vowel duration before voiced and voiceless consonants (D’Avanzo, Graziano, Metz, Schiavetti, & Whitehead, 1998; Metz, Schiavetti, Lessler, Lawe, Whitehead, & Whitehead, 1997;). Thus, it can be concluded that although SC results in an altered temporal pattern of speech, it does not result in a pattern that violates these specific temporal rules of spoken English. These results are interesting in relation to the research of Hyde, Power, and Leigh (1998, p.124) who “found the slowed speech rate under the simultaneous communication condition to be the most understandable” and concluded that “the slowed rate of many teachers using simultaneous communication improves the intelligibility of that speech for deaf students.” Hyde et al. (1998) reached this conclusion despite their finding that speech produced during SC was rated as less natural-sounding than speech produced under two other speech-only conditions. They suggested further research to determine what speech characteristics during SC contributed to this intelligibility increase. Acoustic parameters such as increased fricative duration might well be responsible for some of the increase in intelligibility along with other factors, such as the increased vowel duration reported by Whitehead et al. (1995; 1999) or the increased voice onset time reported by Schiavetti et al. (1996; 1999). Future research is important to determine the relative contribution of each of these temporal changes to any improvements found in the intelligibility of speech during SC. In addition to examining the influence of temporal changes on the intelligibility of speech produced during SC, it is important to consider the possible influence of SC on spectral characteristics of speech. Kent and Read (1992) have summarized many of the spectral characteristics of fricative consonants that are importantly related to their perception, and future research should consider the degree to which fricative consonant spectral characteristics may be affected as speech is slowed during SC. An important line of inquiry would concern whether any consonant spectral changes that might be found would affect the intelligibility of speech produced during SC. The research of Picheny, Durlach, & Braida (1986) concerning speaking clearly for the hearing-impaired presents a good research design model for such studies of the influence of acoustic changes on intelligibility in different speaking conditions. Future research on SC should focus on empirical analysis of its relative benefits in improving communication between hearing and hearing-impaired persons, and analysis of the intelligibility characteristics of speech produced during SC is an important step toward this goal.
432
WHITEHEAD et al.
A portion of this research was conducted at the National Technical Institute for the Deaf in the course of an agreement between the Rochester Institute of Technology and the United States Department of Education. Part of this research was supported by funds from the Geneseo Foundation provided through the Research Council.
REFERENCES Caccamise, F., Updegraff, D., & Newell, W. (1990). Staff sign skills assessment-development at Michigan School for the Deaf: Achieving an important need. Journal of the Academy of Rehabilitative Audiology, 23, 27–41. D’Avanzo, S.B., Graziano, T., Metz, D.E., Schiavetti, N., & Whitehead, R.L. (1998). Production and perception of final consonant voicing in speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 31, 337–346. Hirsh, I.J. (1952). The measurement of hearing. New York: McGraw-Hill. Huntington, A., & Watton, F. (1984). Language and interaction in the education of hearing–impaired children (Part 2). Journal of the British Association of Teachers of the Deaf, (8)5, 137–144. Hyde, M., Power, D., & Leigh, G. (1998). Oral-only and simultaneous communication speech characteristics of teachers of the deaf. In A. Weisel (Ed.), Issues unresolved: New perspectives on language and deaf education (pp. 117–125). Washington, DC: Gallaudet University Press. Kent, R.D., & Read, C. (1992). The acoustic analysis of speech. San Diego: Singular. Marmor, G.S., & Petitto, L. (1979). Simultaneous communication in the classroom: How well is English grammar represented? Sign Language Studies, 23, 99–136. Metz, D.E., Schiavetti, N., Lessler, A., Lawe, Y., Whitehead, R.L., & Whitehead, B.H. (1997). Production and perception of final consonant voicing in speech produced during simultaneous communication. Journal of Communication Disorders, 30, 495–505. Picheny, M.A., Durlach, N.I., & Braida, L.D. (1986). Speaking clearly for the hard of hearing II: Acoustic characteristics of clear and conversational speech. Journal of Speech and Hearing Research, 29, 434–446. Schiavetti, N., Whitehead, R.L., Metz, D.E., & Moore, N. (1999). Voice onset time in speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 32, 37–49.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
433
Schiavetti, N., Whitehead, R.L., Metz, D.E., Whitehead, B.H., & Mignerey, M. (1996). Voice onset time in speech produced during simultaneous communication. Journal of Speech and Hearing Research, 38, 565–572. Strong, M., & Charlson, E.S. (1987). Simultaneous communication: Are teachers attempting an impossible task? American Annals of the Deaf, 132, 376–382. Schwartz, M.F. (1969). Influence of vowel environment upon the duration of /s/ and /e/. Journal of the Acoustical Society of America, 46, 480–481. Tye-Murray, N., Spencer, L., & Woodworth, G.G. (1995). Acquisition of speech by children who have prolonged cochlear implant experience. Journal of Speech and Hearing Research, 38, 327–337. Vernon, M., & Andrews, J.F. (1990). The Psychology of Deafness. New York: Longman. Weglarski, A., Sewall, A., Whitehead, R.L., Schiavetti, N., & Metz, D.E. (1999). Effect of vowel environment on consonant duration: An extension of normative data to contextual speech produced by adults. Manuscript submitted for publication. Whitehead, R.L., Schiavetti, N., Whitehead, B.H., & Metz, D.E. (1995). Temporal characteristics of speech produced during simultaneous communication. Journal of Speech and Hearing Research, 38, 1014–1024 . Whitehead, R.L., Schiavetti, N., Metz, D.E., & Farrell, T. (1999). Temporal characteristics of speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 32, 79–95. Windsor, J., & Fristoe, M. (1989). Key word signing: Listeners’ classification of signed and spoken narratives. Journal of Speech and Hearing Disorders, 54, 374–382. Windsor, J., & Fristoe, M. (1991). Key word signing: Perceived and acoustic differences between signed and spoken narratives. Journal of Speech and Hearing Research, 34, 260–268. Manuscript received 1 March 1999; revised 6 July 1999; accepted 15 July 1999.
434
WHITEHEAD et al.
CONTINUING EDUCATION Effect of Vowel Environment on Consonant Duration in Speech Produced during Simultaneous Communication QUESTIONS 1. As used in this study, simultaneous communication (SC) involved: a. Speech-only produced by normal hearing speakers b. American Sign Language produced by deaf signers c. Speech combined with signed English and finger-spelling d. Speech-only produced by deaf speakers e. American Sign Language produced by normal hearing signers 2. Fricative duration is: a. Longer before high vowels than low vowels b. Longer before front vowels than back vowels c. Longer before back vowels than front vowels d. Longer before low vowels than high vowels e. Both a and b 3. In the present article, sentence duration was: a. Longer in speech-only than in SC b. Longer in SC than in speech-only c. Longer in ASL than in SC d. Shorter in ASL than in SC e. Both a and c 4. In the present article, fricative duration was a. Longer in speech-only than in SC b. Longer in SC than in speech-only c. Longer in ASL than in SC d. Shorter in ASL than in SC e. Both a and c 5. The results of this study indicate that speech during SC: a. Is slower, has longer fricative duration, and violates temporal rules of English speech b. Is faster, has shorter fricative duration, and violates temporal rules of English speech c. Is slower, has longer fricative duration, and does not violate temporal rules of English speech d. Is faster, has shorter fricative duration, and does not violate temporal rules of English speech e. Is unchanged in sentence or fricative duration
NICHOLAS SCHIAVETTI, DALE EVAN METZ, and KIMBERLY FARINELLA State University of New York, Geneseo, New York
This study investigated the effect of vowel environment on fricative consonant duration in contextual speech produced during simultaneous communication (SC). Previous studies (Schwartz, 1969) of vowel influences on consonant duration supported the notion of anticipatory scanning, in which final vowel targets influence the duration of preceding fricative consonants. Ten normal-hearing, experienced sign language users recorded palatal and alveolar fricatives produced in four vowel environments in contextual sentences under SC and speech-only (SO) conditions. Results indicated longer sentence durations for SC than for SO, and significant effects of vowel context on fricative consonant duration in contextual speech in both SC and SO conditions that revealed similar anticipatory scanning effects as seen in previous studies. These data confirm previous research indicating that the temporal alterations produced by simultaneous communication do not involve violations of the temporal rules of English speech. © 1999 Elsevier Science Inc. Educational Objectives: The reader will (1) acquire knowledge and understanding of simultaneous communication and its role in communication with children who are deaf; and (2) understand the relationships among speech rate, fricative duration, and temporal changes in speech during simultaneous communication. KEY WORDS: Simultaneous communication; Fricatives; Anticipatory scanning
Address correspondence to Robert L. Whitehead, National Technical Institute for the Deaf at RIT, Carey 2408, 96 Lomb Memorial Drive, Rochester, NY 14623-5604. Tel: 716-475-6457; Fax: 716-4757101; E-mail: ,[email protected]..
J. COMMUN. DISORD. 32 (1999), 423–434 © 1999 by Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010
0021-9924/99/$–see front matter PII S0021-9924(99)00019-2
424
WHITEHEAD et al.
Previous research considered temporal changes in speech produced during simultaneous communication (SC), when speech is combined with manually coded English (sign and finger-spelling) for the production of each word of an utterance using both spoken and manual modalities (Whitehead, Schiavetti, Whitehead, & Metz, 1995). Reasons for the use of SC with children who are deaf include the enhancement of language development and exposure to segmental and suprasegmental features of normal speech, although some researchers have criticized SC as presenting an unnatural speech model (Huntington & Watton, 1984; Marmor & Petitto, 1979; Strong & Charlson, 1987; Vernon & Andrews, 1990). Tye-Murray, Spencer, and Woodworth (1995) reported that children who used SC before a cochlear implant did not discontinue the use of sign language after implantation. Their results stress the importance of research on the effects of SC on speech presented to deaf children to determine the potential influences of SC use on their speech and language development. Research on the temporal characteristics of speech produced during SC is necessary to determine the degree to which SC influences the speech models presented by persons with normal hearing to those with hearing impairment. For example, Hyde, Power, and Leigh (1998, p. 117) have raised questions about the disturbed prosodic quality of speech during SC “that may influence the acceptability of such speech as an auditory-oral model for deaf students.” Research with experienced and inexperienced signers has demonstrated that speech produced during SC is slower in rate than speech produced alone and has longer interword intervals and vowel durations (Whitehead et al., 1995; Whitehead, Schiavetti, Metz, & Farrell, 1999), and longer voice onset times (VOT) (Schiavetti, Whitehead, Metz, & Moore, 1999; Schiavetti, Whitehead, Metz, Whitehead, & Mignerey, 1996). These temporal modifications are the result of attempting to maintain simultaneity between the spoken and manual modes of communication (Windsor & Fristoe, 1989; 1991). Research has demonstrated that even though there are temporal elongations in speech produced during SC, the temporal rules of spoken English that have been studied so far are maintained, thereby providing an accurate speech model to deaf children for these specific parameters. These temporal rules include the effect of vowel height and voicing characteristics of the following consonants on vowel duration (Whitehead et al., 1995; 1999) and VOT differences between voiced and voiceless members of cognate pairs of plosives (Schiavetti, et al., 1996; 1999). The segmental measures investigated in these studies mainly concern vowels or the time intervals between consonants and vowels, but none of these studies investigated the effect of SC on specific consonant characteristics, such as duration or spectrum. The importance of studying consonantal effects of SC is highlighted by the classic statement of Hirsh (1952, p. 125) that “The consonants of English seem to be responsible for most of the intelligibility
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
425
carried by individual English words.” If SC disrupts the temporal or spectral patterns of consonants, then serious intelligibility repercussions could result from using this mode to communicate with deaf children who are attempting to develop oral language. A temporal rule of spoken English that has not been studied in the speech of persons using SC is the influence of vowels on consonant duration. This effect was labeled “anticipatory scanning” by Schwartz (1969) as a result of his study of consonant durational differences in repeated nonsense syllables, and has recently been shown to occur in contextual speech produced by adult male and female speakers (Weglarski, Sewall, Whitehead, Schiavetti, & Metz, 1999). This temporal rule suggests that alveolar and palatal fricatives are longer in duration when the following vowel is a high front vowel rather than a low back vowel. Schwartz (1969) theorized that the difference is the result of the smaller distance the tongue must travel from the consonant place of articulation to the position for the following high front vowel in the CV syllable /si/, compared with the larger distance the tongue must travel to the position for the following low back vowel in the CV syllable /sa/. The purpose of this study was to examine the effect of vowel environment on the duration of fricative consonants produced by adults in contextual speech when producing speech only, versus SC. The durations of alveolar and palatal fricatives preceding four different vowels were examined to study the degree to which anticipatory scanning effects seen in previous studies (Schwartz, 1969; Weglarski et al., 1999) that used speech-only conditions would be observed in speech produced during SC as well.
METHOD Speakers Speakers for this study were ten normal-hearing adults (5 females and 5 males), all of whom were faculty members at the National Technical Institute for the Deaf. The speakers had taught young deaf adults for at least eight years, using simultaneous communication in all classes. Each speaker’s sign language performance had been evaluated on the Sign Communication Proficiency Index (Caccamise, Updegraff, & Newell, 1990), and all speakers were classified at the advanced level or higher on this instrument. Thus, the speakers were considered to be fluent in the use of speech combined with signed English and finger-spelling.
SPEECH MATERIALS The speech samples investigated consisted of the carrier sentence, “I can say ___________ again clearly” and eight experimental words that varied in the
426
WHITEHEAD et al.
initial fricative and following vowel embedded in the blank slot in the carrier sentence. Two fricatives were studied: alveolar /s/ and palatal /ʃ/ preceding each of four vowels: high front /i/, low front /{/, low back /a/, and high back / u/. The experimental words included seat, sheet, sat, shack, sock, shock, suit, and shoot.
Recording Procedures The experimental sentences were spoken at a comfortable conversational loudness level as they were presented on flashcards. Audio recordings were made in a sound-treated booth using an Audio-Technica AT-816 microphone that was placed 15 cm from each speaker’s mouth and was connected to a Tascam 202MKII tape deck. The speakers produced each sentence, with its embedded experimental word, under two conditions: (a) speech only; and (b) speech combined with signed English for all words in the sentence except the experimental word, which was finger-spelled. The sentences containing the experimental words were presented to the speakers on flashcards in two different random orders, one for each experimental condition; the order of experimental condition (speech-only versus SC) was counterbalanced across speakers. The speakers were shown the experimental words before the recording, so they could familiarize themselves with the signs and finger-spellings to be used in SC.
Acoustic Analysis Procedures For each speech sample, duration measures in milliseconds were determined for the total sentence and the initial fricative consonant in the experimental word. The acoustic signal from the audio recording of each sentence was digitized with 16-bit precision at 20 kHz using Kay Elemetrics Computerized Speech Lab (CSL; Model 4300B). When the digitizing process is initiated, CSL applies an appropriate internal low-pass anti-aliasing filter to the raw acoustic signal (at a digitizing rate of 20 kHz, the upper frequency cut-off is 8 kHz), stores the digital results in memory, and displays the resultant waveform on a VGA graphics monitor. We measured total sentence duration in the two experimental conditions to verify that speech rate was slower in SC than in the speech-only condition. Total sentence duration was measured by visually isolating the first positively going portion of the waveform associated with the initiation of the diphthong /a/ and marking the location with a cursor. The cursor was then moved to the end of the sentence, and the last positive going portion of the waveform associated with the /i/ segment in the word “clearly” was visually isolated and marked with a cursor. The temporal interval between the two cursors was taken as the value for total sentence duration.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
427
Fricative consonant duration was measured by visually isolating the initiation of the aperiodic waveform following the diphthong /e/ in the word “say” and marking the location with a cursor. The cursor was then moved to the end of the aperiodic waveform, and the temporal interval between the two cursors was taken as the value for fricative consonant duration. As a measure of intraobserver reliability of the measurement procedure, the recordings of one male and one female speaker were selected at random, and all the fricative durations were measured a second time by the same person who made the first measurement. The mean difference between the first and second measure for the male speaker was was 1.89 msec (SD 5 0.58 msec) and the mean difference between the first and second measure for the female speaker was was 2.12 msec (SD 5 0.86 msec). These low mean differences generally indicate sufficient intraobserver reliability for these measures. As a measure of intraspeaker reliability, one male and one female speaker were selected at random to record all speech materials twice. The second recordings were digitized on the CSL, and all fricative durations were measured for comparison with the original recordings. For each speaker, the repeated measures for each token were subtracted from the original measure of each token to examine the absolute value of the difference between the two recordings. For the female speaker, the mean fricative duration difference for the speech-only condition was 3.84 msec (SD 5 1.76 msec) and for the SC condition was 6.93 msec (SD 5 4.93 msec). For the male speaker, the mean fricative duration difference for the speech-only condition was 4.31 msec (SD 5 2.11 msec) and for the SC condition was 8.62 msec (SD 5 5.88 msec). These low mean differences in fricative duration generally indicated sufficient intraspeaker reliability for these measures.
Statistical Analysis of Duration Data Statistical analyses of the sentence duration and fricative data were accomplished with a three-way (2 3 2 3 4) analysis of variance (ANOVA) with repeated measures on all independent variables: (a) experimental condition (two levels: speech-only versus SC); (b) fricative place of articulation (two levels: alveolar versus palatal); and (c) vowel (four levels: high front /i/ versus low front /{/ versus low back /a/ versus high back /u/).
RESULTS Sentence Duration The ANOVA revealed a significant effect of communication mode on sentence duration (F 5 33.66; df 5 1,9; p , .001) . No other main effects or interaction effects on sentence duration were significant. Figure 1 presents the
Figure 1. Means and standard deviations of durations of sentences containing experimental words with initial alveolar and palatal fricatives preceding each of four vowels in speech-only vs. SC conditions.
428 WHITEHEAD et al.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
429
means and standard deviations of the sentence duration in speech-only and SC for both fricative places of articulation preceding each of the four vowels. Inspection of Figure 1 reveals that sentence duration was always longer in SC than in SO, indicating a slowing of speech while speakers attempted to sign and speak at the same time.
Fricative Duration The ANOVA revealed significant effects of communication mode (F 5 64.27; df 5 1,9; p , .0001) and vowel (F 5 94.71; df 5 3,27; p , .0001) on fricative duration. No other main effects or interaction effects on sentence duration were significant. Direct means contrast tests comparing the main effect of vowel environments (collapsed across experimental condition and place of articulation) indicated that the durations of the fricatives preceding each vowel were significantly different from the durations of the fricatives preceding each of the other three vowels. Figure 2 illustrates the means and standard deviations of the durations of the two initial fricatives in each of the four vowel environments for both the speech-only and SC conditions. Inspection of Figure 2 reveals longer fricative durations in SC than in the speech only condition. In addition, the fricative duration was greater preceding high and front vowels when compared with durations preceding low and back vowels.
DISCUSSION The results of the present investigation are consistent with previous research demonstrating that speech is elongated in duration when it is combined with signed English and finger-spelling in simultaneous communication (Whitehead, et al., 1995). Specifically, for the dependent variables of sentence and fricative duration, there was a consistent pattern of significant durational differences between the speech-only and SC conditions; that is, sentences and fricatives were always significantly longer in duration in the SC condition than the speech-only condition. The combination of two communication systems that operate at different rates results in a retardation of the more rapid articulatory act of speech in order to maintain a high degree of simultaneity (Windsor & Fristoe, 1989; 1991). However, despite the increase in duration of fricatives during SC, the speakers maintained the temporal rule that alveolar and palatal fricatives are longer in duration when the following vowel is a high front vowel compared with a low back vowel. Thus, during SC these speakers demonstrated anticipatory scanning of subsequent vowel anatomical position, as was reported by Schwartz (1969) for nonsense syllables produced in a speech-only condition. In addition, they showed the same pattern that was recently reported by Weglarski et al. (1999) for contextual speech, in which both subsequent vowel
Figure 2. Means and standard deviations of durations of alveolar and palatal fricatives preceding each of four vowels in speech-only vs. SC conditions.
430 WHITEHEAD et al.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
431
height and front-back advancement influenced the duration of preceding fricatives. These results are consistent with previous studies that reported temporal elongation during SC but maintenance of the temporal rules of spoken English in the relative durations of various vowels in different consonant context (Whitehead et al., 1995; 1999), voice-onset times of voiced and voiceless cognates (Schiavetti et al., 1996; 1999), and relative vowel duration before voiced and voiceless consonants (D’Avanzo, Graziano, Metz, Schiavetti, & Whitehead, 1998; Metz, Schiavetti, Lessler, Lawe, Whitehead, & Whitehead, 1997;). Thus, it can be concluded that although SC results in an altered temporal pattern of speech, it does not result in a pattern that violates these specific temporal rules of spoken English. These results are interesting in relation to the research of Hyde, Power, and Leigh (1998, p.124) who “found the slowed speech rate under the simultaneous communication condition to be the most understandable” and concluded that “the slowed rate of many teachers using simultaneous communication improves the intelligibility of that speech for deaf students.” Hyde et al. (1998) reached this conclusion despite their finding that speech produced during SC was rated as less natural-sounding than speech produced under two other speech-only conditions. They suggested further research to determine what speech characteristics during SC contributed to this intelligibility increase. Acoustic parameters such as increased fricative duration might well be responsible for some of the increase in intelligibility along with other factors, such as the increased vowel duration reported by Whitehead et al. (1995; 1999) or the increased voice onset time reported by Schiavetti et al. (1996; 1999). Future research is important to determine the relative contribution of each of these temporal changes to any improvements found in the intelligibility of speech during SC. In addition to examining the influence of temporal changes on the intelligibility of speech produced during SC, it is important to consider the possible influence of SC on spectral characteristics of speech. Kent and Read (1992) have summarized many of the spectral characteristics of fricative consonants that are importantly related to their perception, and future research should consider the degree to which fricative consonant spectral characteristics may be affected as speech is slowed during SC. An important line of inquiry would concern whether any consonant spectral changes that might be found would affect the intelligibility of speech produced during SC. The research of Picheny, Durlach, & Braida (1986) concerning speaking clearly for the hearing-impaired presents a good research design model for such studies of the influence of acoustic changes on intelligibility in different speaking conditions. Future research on SC should focus on empirical analysis of its relative benefits in improving communication between hearing and hearing-impaired persons, and analysis of the intelligibility characteristics of speech produced during SC is an important step toward this goal.
432
WHITEHEAD et al.
A portion of this research was conducted at the National Technical Institute for the Deaf in the course of an agreement between the Rochester Institute of Technology and the United States Department of Education. Part of this research was supported by funds from the Geneseo Foundation provided through the Research Council.
REFERENCES Caccamise, F., Updegraff, D., & Newell, W. (1990). Staff sign skills assessment-development at Michigan School for the Deaf: Achieving an important need. Journal of the Academy of Rehabilitative Audiology, 23, 27–41. D’Avanzo, S.B., Graziano, T., Metz, D.E., Schiavetti, N., & Whitehead, R.L. (1998). Production and perception of final consonant voicing in speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 31, 337–346. Hirsh, I.J. (1952). The measurement of hearing. New York: McGraw-Hill. Huntington, A., & Watton, F. (1984). Language and interaction in the education of hearing–impaired children (Part 2). Journal of the British Association of Teachers of the Deaf, (8)5, 137–144. Hyde, M., Power, D., & Leigh, G. (1998). Oral-only and simultaneous communication speech characteristics of teachers of the deaf. In A. Weisel (Ed.), Issues unresolved: New perspectives on language and deaf education (pp. 117–125). Washington, DC: Gallaudet University Press. Kent, R.D., & Read, C. (1992). The acoustic analysis of speech. San Diego: Singular. Marmor, G.S., & Petitto, L. (1979). Simultaneous communication in the classroom: How well is English grammar represented? Sign Language Studies, 23, 99–136. Metz, D.E., Schiavetti, N., Lessler, A., Lawe, Y., Whitehead, R.L., & Whitehead, B.H. (1997). Production and perception of final consonant voicing in speech produced during simultaneous communication. Journal of Communication Disorders, 30, 495–505. Picheny, M.A., Durlach, N.I., & Braida, L.D. (1986). Speaking clearly for the hard of hearing II: Acoustic characteristics of clear and conversational speech. Journal of Speech and Hearing Research, 29, 434–446. Schiavetti, N., Whitehead, R.L., Metz, D.E., & Moore, N. (1999). Voice onset time in speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 32, 37–49.
FRICATIVE DURATION IN SIMULTANEOUS COMMUNICATION
433
Schiavetti, N., Whitehead, R.L., Metz, D.E., Whitehead, B.H., & Mignerey, M. (1996). Voice onset time in speech produced during simultaneous communication. Journal of Speech and Hearing Research, 38, 565–572. Strong, M., & Charlson, E.S. (1987). Simultaneous communication: Are teachers attempting an impossible task? American Annals of the Deaf, 132, 376–382. Schwartz, M.F. (1969). Influence of vowel environment upon the duration of /s/ and /e/. Journal of the Acoustical Society of America, 46, 480–481. Tye-Murray, N., Spencer, L., & Woodworth, G.G. (1995). Acquisition of speech by children who have prolonged cochlear implant experience. Journal of Speech and Hearing Research, 38, 327–337. Vernon, M., & Andrews, J.F. (1990). The Psychology of Deafness. New York: Longman. Weglarski, A., Sewall, A., Whitehead, R.L., Schiavetti, N., & Metz, D.E. (1999). Effect of vowel environment on consonant duration: An extension of normative data to contextual speech produced by adults. Manuscript submitted for publication. Whitehead, R.L., Schiavetti, N., Whitehead, B.H., & Metz, D.E. (1995). Temporal characteristics of speech produced during simultaneous communication. Journal of Speech and Hearing Research, 38, 1014–1024 . Whitehead, R.L., Schiavetti, N., Metz, D.E., & Farrell, T. (1999). Temporal characteristics of speech produced by inexperienced signers during simultaneous communication. Journal of Communication Disorders, 32, 79–95. Windsor, J., & Fristoe, M. (1989). Key word signing: Listeners’ classification of signed and spoken narratives. Journal of Speech and Hearing Disorders, 54, 374–382. Windsor, J., & Fristoe, M. (1991). Key word signing: Perceived and acoustic differences between signed and spoken narratives. Journal of Speech and Hearing Research, 34, 260–268. Manuscript received 1 March 1999; revised 6 July 1999; accepted 15 July 1999.
434
WHITEHEAD et al.
CONTINUING EDUCATION Effect of Vowel Environment on Consonant Duration in Speech Produced during Simultaneous Communication QUESTIONS 1. As used in this study, simultaneous communication (SC) involved: a. Speech-only produced by normal hearing speakers b. American Sign Language produced by deaf signers c. Speech combined with signed English and finger-spelling d. Speech-only produced by deaf speakers e. American Sign Language produced by normal hearing signers 2. Fricative duration is: a. Longer before high vowels than low vowels b. Longer before front vowels than back vowels c. Longer before back vowels than front vowels d. Longer before low vowels than high vowels e. Both a and b 3. In the present article, sentence duration was: a. Longer in speech-only than in SC b. Longer in SC than in speech-only c. Longer in ASL than in SC d. Shorter in ASL than in SC e. Both a and c 4. In the present article, fricative duration was a. Longer in speech-only than in SC b. Longer in SC than in speech-only c. Longer in ASL than in SC d. Shorter in ASL than in SC e. Both a and c 5. The results of this study indicate that speech during SC: a. Is slower, has longer fricative duration, and violates temporal rules of English speech b. Is faster, has shorter fricative duration, and violates temporal rules of English speech c. Is slower, has longer fricative duration, and does not violate temporal rules of English speech d. Is faster, has shorter fricative duration, and does not violate temporal rules of English speech e. Is unchanged in sentence or fricative duration