You are on page 1of 9

Original Paper

Folia Phoniatr Logop 2005;57:125–133


DOI: 10.1159/000084133

Influence of Gender and Environmental


Setting on Voice Onset Time
Michael Robba Harvey Gilbertb Jay Lermanb
a
University of Canterbury, Christchurch, New Zealand, and
b
University of Connecticut, Storrs, Conn., USA

Key Words
Voice onset time  Speech production  Speech perception

Abstract
The influence of gender and environmental setting on voice onset time
(VOT) was examined. Participants produced six stop consonants paired with
three vowels in a consonant + vowel (CV) context. Recordings were made when
participants were seated inside a sound booth (laboratory setting) and outside
a sound booth (non-laboratory setting). Results of the analysis for VOT and CV
duration indicated that females produced voiceless stops with significantly lon-
ger VOT durations than males in both settings. Both gender groups produced
CV tokens with significantly longer duration in the laboratory setting; however,
females continued to produce CV tokens with longer duration in the non-labora-
tory setting. Findings suggest that differences in the vocal anatomy between
genders may account for some but not all differences in temporal measures of
speech production. Sociophonetic factors, such as speaking style, may also con-
tribute to gender differences in speaking behavior.
Copyright © 2005 S. Karger AG, Basel

Introduction

The importance of voice onset time (VOT) in speech perception and speech
production is well known [1, 2]. VOT is a reliable cue to distinguish between voiced
and voiceless stops in English, as well as a number of other languages [3–5]. As an
acoustic cue, VOT also serves as an inferential estimate of speech motor control, re-
quiring fine motor coordination of the respiratory, phonatory, and articulatory struc-
tures [6].
198.143.32.65 - 1/27/2016 12:19:26 AM

© 2005 S. Karger AG, Basel Prof. Michael Robb, Department of Communication


1021–7762/05/0573–0125$22.00/0 Disorders, University of Canterbury, Private Bag 4800
University of Pittsburgh

Fax +41 61 306 12 34 Christchurch 8020 (New Zealand)


Downloaded by:

E-Mail karger@karger.ch Accessible online at: Tel. +64 3 364 2401, Fax +64 3 364 2760
www.karger.com www.karger.com/fpl E-Mail michael.robb@canterbury.ac.nz
VOT has been measured in children [7], the elderly [8], across racial groups [9],
non-English-speaking groups [10, 11], and speech-disordered groups [12, 13]. The
effects of gender on VOT have also been explored. A number of these studies indi-
cate that females and males differ in their production of word-initial stop consonants
[14–16], with these differences exhibited as early as 11-years of age [17]. The excep-
tions to these results include a study by Sweeting and Baken [8] who found no gender
differences among adult speakers, and Smith [18] who reported longer VOTs for
initial /d,g/ in adult males. In general, the overall trend is that females produce stops
with longer VOT durations compared to males.
Gender differences in respiratory function [19], laryngeal behavior [20–23], and
supralaryngeal behavior [24] for speech production have been documented. Presum-
ably, the gender differences found for VOT production reflect fundamental biologi-
cal differences in the various speech subsystems. For example, female production of
prevocalic stops is associated with greater oral pressure compared to male produc-
tions [25]. Bless and Abbs [26] noted that a characteristic feature of female phona-
tion is a posterior glottal opening (‘chink’) which occurs during vocal fold closing.
This opening results in an air leak into the supralaryngeal space during moments of
vocal fold closing. In the context of female production of VOT, it is possible that
more air escapes into the supralaryngeal space prior to vowel production thus result-
ing in greater oral pressure for stops compared to male speakers. The greater oral
pressure results in a longer interval between stop burst and voicing onset.
Although gender differences in VOT production may result from biological dif-
ferences, there are also methodological and sociophonetic factors that might contrib-
ute to VOT differences. For example, VOT is negatively correlated with increases in
speech rate or word duration [27]. Speaking slowly and producing words with a long
duration result in longer VOT values, especially for voiceless stops [28–30]. Presum-
ably if gender differences exist in speaking rate, a gender difference would also be
apparent in VOT production. However, there are conflicting data concerning the
speaking rate of men versus women. Some researchers have found no group differ-
ences in speaking rate [31–33]; while there are studies indicating that women speak
faster than men [34], and men speak faster than women [35]. Previous studies ex-
amining gender differences in VOT have not considered the influence of speaking
rate. Therefore, it is possible that speaking rate differences (such as word duration)
rather than biological differences are responsible for the sex-linked VOT pattern.
Gender differences in VOT may also arise as a result of the method in which
VOT values are summarized. A commonly used procedure when calculating and re-
porting VOT is to average voicing-lead (i.e., negative VOTs) and short-lag values
together. This procedure dates back to Lisker and Abramson [36] and has been ad-
opted by several researchers since then [9, 30]. However, computing mean VOT
values on both negative and positive values may obscure possible patterns in VOT
production, particularly voiced stop production. Examining short-lag VOT values
apart from negative VOTs may shed new information on gender differences in
VOT.
Finally, there are suggestions that sex differences in VOT may be due to a so-
ciophonetic influence, whereby the differences reflect conscious manipulation of
speech patterns to overtly or tacitly convey gender identity. Cheshire [37] and others
[17, 35, 38] report that females and males differ in their general pattern of pronun-
ciation, with females tending to use more carefully articulated speech and adopt this
198.143.32.65 - 1/27/2016 12:19:26 AM

126 Folia Phoniatr Logop Robb/Gilbert/Lerman


University of Pittsburgh

2005;57:125–133
Downloaded by:
speaking style in experimental settings. There have been no direct attempts to deter-
mine whether gender differences in VOT may be socially motivated.
In summary, it remains to be determined whether gender differences in VOT
result from biological differences in the speech production mechanism, or whether
methodological and sociophonetic factors are involved. The purpose of this study
was to consider the influence of word duration, short-lag VOT and environmental
setting on sex-linked differences in VOT.

Method

Participants
Twenty young Caucasian adults consisting of 10 females (mean age = 20 years) and 10 males
(mean age = 20 years) participated in the study. All participants were native monolingual speak-
ers of American English and were selected from the undergraduate population at the University
of Connecticut (Storrs, Conn., USA). None of the participants reported ever having a speech,
language or hearing disorder.

Recording Procedures
Each participant was audio recorded in two environmental settings. One setting, termed the
laboratory setting, was a sound-treated booth housed in the Department of Communication Sci-
ences at the University of Connecticut. In the laboratory setting, participants were seated at a
table. A non-functioning (‘dead’) microphone was mounted on a stand and positioned approxi-
mately 12 in (30 cm) from the participant’s mouth. The lead from the microphone was connect-
ed to one channel of a portable audiocassette recorder (Marantz PMD-360); however, the micro-
phone was turned off throughout the recording session. The actual audio recordings were obtained
using a second microphone worn by the researcher who was situated approximately 24 in (60 cm)
from the participant. The researcher wore a wireless frequency-modulated (FM) lavaliere micro-
phone (NADY E-701) clipped to her lapel. The microphone was attached to a body-pack trans-
mitter (NADY LT-10) that was placed in a shirt pocket. The transmitter delivered an FM signal
to a receiver that was attached to a second channel of the audiocassette recorder. Both micro-
phones and the audiocassette recorder were visible to the participant. The use of a sound booth,
visible microphones, and visible audiocassette recorder in the laboratory setting was a deliberate
attempt to create an environment whereby each participant was fully cognizant of their speaking
behavior. The second setting, termed the non-laboratory setting, consisted of a carpeted enclosed
room in which the sound booth, visible microphones, and visible audiocassette recorder were not
used. In this non-laboratory setting, each participant sat at a table across from the researcher.
Audio recordings were obtained using the same wireless FM microphone system used in the
laboratory setting; however, the microphone was concealed in a stack of books situated on the
table approximately 24 in from the participant’s mouth. All speech samples were recorded using
high quality audiocassette tape (Maxell PI, low noise/high output).
Although it is likely laboratory influences could not be entirely removed from the non-labo-
ratory setting, the intent was to de-emphasize or shift attention away from a formal speaking
situation. By placing participants in a setting that was less contrived, we assumed this would al-
low for a more natural style of speaking compared to the sound booth. Participants were ran-
domly assigned to either the laboratory or non-laboratory setting for their first recording and
returned within 2-weeks time for audio recording in the second setting. In both settings partici-
pants were instructed to speak naturally.

Speech Sampling
In each setting, participants read a list of 18 monosyllables formed from the six stop conso-
nants combined with the vowels /i,u,a/. Stimuli were presented in CV (consonant + vowel) forms
and presented in randomized order within each repetition set. Each CV was embedded in the
198.143.32.65 - 1/27/2016 12:19:26 AM

Gender and VOT Folia Phoniatr Logop 127


University of Pittsburgh

2005;57:125–133
Downloaded by:
carrier phrase, ‘It’s a ____.’ The carrier phrase was used to encourage naturalistic productions and
to minimize intrasubject variability in stress patterns and fundamental frequency contours [8].
Five repetitions were recorded for each participant; however, the first and fifth repetitions were
excluded from the acoustic analyses. The first repetition served as a task rehearsal for each par-
ticipant. If one of the original three tokens needed to be discarded (e.g., background noise), the
token was replaced with the fifth exemplar.

Acoustic Analysis
Each audio-recorded CV token was digitized at 10 kHz using a speech analysis system (Kay
CSL 4300) and simultaneously displayed on a computer monitor as an amplitude ! frequency
display and wideband (500 Hz) sound spectrogram (Blackman window weighting). A pair of ver-
tical cursors was overlaid on the displays. The left cursor was positioned at the burst release and
the right cursor was placed at the first instance of vocal fold vibration at the level of the second
formant [36]. The VOT was defined as the time interval between cursors and reported in milli-
seconds. Instances in which voicing occurred before the burst release were measured by placing
the left cursor at the onset of the pre-burst voicing and the right cursor was positioned at the burst
release. The VOT for this time interval was reported as a negative value. In addition to VOT, the
overall duration of each CV token was measured. The CV duration was defined as the time in-
terval between cursors at the burst release and the last instance of vocal fold vibration associated
with the vowel. The CV duration was reported in milliseconds.

Reliability Analysis
To ensure consistency and reliability in the VOT measurements, a test of intra-judge and
inter-judge reliability was carried out. Intra-judge reliability was performed by randomly select-
ing one subject from each of the recording settings and re-measuring the VOT for each of the 18
monosyllables. A second individual experienced in the acoustic analysis of speech was used for
determining inter-judge reliability. This individual measured the VOT for one subject from each
of the recording settings and the values were compared to the original measurement values. The
average intra-judge re-measurement difference was 3 ms (range = 0–7 ms). The average inter-
judge measurement difference was 4 ms (range = 0–10 ms). Pearson product-moment correlation
tests were used to calculate the intra-judge and inter-judge reliabilities. A significant correlation
(p ! 0.05) was found for both analyses indicating a high level of measurement reliability.

Statistical Analysis
Data were analyzed using a series of two-way repeated measures analysis of variance (ANO-
VA) tests [39] to assess the effects of environmental setting (laboratory versus non-laboratory) on
the features of VOT and CV duration produced by the men and women. Gender and experimen-
tal setting were the between-groups factors and features of VOT and CV were the within-groups
factors.

Results

Voice Onset Time


A total of 1,080 VOT measurements were made in the present study (54 mea-
sures per participant). Each participant’s VOT values corresponding to each stop
token were averaged. The individual means were than averaged for each gender
group. The mean and standard deviation values obtained for each stop token accord-
ing to gender group and recording location are listed in table 1. For both groups, and
across both settings, the VOT of voiceless and voiced stops increased as a function
of place of articulation. Bilabials yielded the shortest VOT values, followed by al-
veolar stops, then velar stops. Negative VOT values were found only for /b/ tokens
(37 in total), occurring for both male and female speakers in both settings. In gen-
198.143.32.65 - 1/27/2016 12:19:26 AM

128 Folia Phoniatr Logop Robb/Gilbert/Lerman


University of Pittsburgh

2005;57:125–133
Downloaded by:
Table 1. Mean VOT (ms) for male and female stop + vowel productions collected in laboratory
and non-laboratory settings

Group /p/ /t/ /k/ /b/ /d/ /g/ /–b/

Laboratory setting
Males 58 (13) 76 (15) 78 (12) 6 (4) 15 (5) 22 (6) –3 (2)
Females 73 (15) 83 (18) 88 (16) 4 (4) 12 (7) 18 (7) –4 (1)

Non-laboratory setting
Males 58 (15) 78 (16) 82 (15) 7 (6) 14 (5) 25 (6) –5 (2)
Females 73 (17) 86 (26) 99 (20) 5 (5) 11 (6) 25 (8) –3 (2)

Negative VOT values are reported separately. Note that negative VOTs were only associated
with /b/ tokens. Standard deviations are shown in parentheses

eral, these values compare favorably to previous studies of VOT production in young
adults [2, 8, 13, 29, 40].
Separate 2-way ANOVAs were performed according to stop consonant voicing
to examine the influence of gender and environmental setting on VOT production.
The data obtained for each stop consonant were collapsed according to voiced and
voiceless stop categories. Negative VOT values were removed from the analysis. Re-
sults of the analysis for voiced stops indicated no significant main effects for gender
or environmental setting. Both gender groups produced voiced stops similarly in the
two settings. There was no significant gender by environmental setting interaction.
Results of the 2-way ANOVA for voiceless stops identified a significant gender by
environmental setting interaction [F(1, 19) = 8.61, p ! 0.01, h2 = 0.31]. A post-hoc
Scheffé t test indicated that females produced significantly longer voiceless stop
VOTs compared to males in the non-laboratory setting (p ! 0.01). The results ac-
cording to VOT voicing categories and setting are displayed in figure 1.

CV Duration
The duration values for each CV token produced by each subject were collapsed
according to stop voicing categories. The data were further averaged according to
gender group. The mean CV duration for male and female productions according to
voicing category and setting are listed in table 2. Results of a 2-way ANOVA for
voiced CV duration identified a significant gender by environmental setting interac-
tion [F(1, 19) = 17.78, p ! 0.01, h2 = 0.47]. Both gender groups produced voiced CVs
with shorter duration in the non-laboratory setting compared to the laboratory set-
ting. However, results of a post-hoc t test indicated that males demonstrated a sig-
nificantly larger decrease in CV duration in the non-laboratory setting compared to
females (p ! 0.01). Results of a 2-way ANOVA for voiceless CV durations yielded
results similar to those for voiced CV durations. A significant gender by environmen-
tal setting interaction was found [F (1, 19) = 20.82, p ! 0.01, h2 = 0.53]. Both groups
produced voiceless CVs with shorter duration in the non-laboratory setting compared
to the laboratory setting. However, males demonstrated a significantly larger decrease
in CV duration in the non-laboratory setting compared to females (p ! 0.01).
198.143.32.65 - 1/27/2016 12:19:26 AM

Gender and VOT Folia Phoniatr Logop 129


University of Pittsburgh

2005;57:125–133
Downloaded by:
Fig. 1. Mean VOT values grouped according to voice/voiceless stop categories as produced in
laboratory and non-laboratory settings. ( = Male participants; $ = female participants. The data
obtained for voiced stops exclude negative VOT values.

Table 2. Mean CV duration (ms) of male and female productions according to stop consonant
voicing collected in a laboratory and non-laboratory setting

Group Laboratory setting Non-laboratory setting

voiceless CVs voiced CVs voiceless CVs voiced CVs

Males 385 (72) 350 (72) 297 (77) 251 (75)


Females 380 (41) 337 (53) 344 (44) 313 (31)

Standard deviations are shown in parentheses.

Discussion

The influence of gender and environmental setting on the production of VOT


and CV duration was studied. Across laboratory and non-laboratory settings, wom-
en were found to produce voiceless stops with significantly longer VOT values com-
pared to men. The common finding across both settings lends support to the sugges-
tion of a gender difference accounting for the VOT differences [17, 41]. However,
this difference was only apparent for voiceless stop production. A gender difference
in voiced stop production was not found in either speaking environment. This find-
ing is in contrast to Ryalls et al. [9] and Swartz [15] who found that female speakers
198.143.32.65 - 1/27/2016 12:19:26 AM

130 Folia Phoniatr Logop Robb/Gilbert/Lerman


University of Pittsburgh

2005;57:125–133
Downloaded by:
produced significantly longer VOT values for voiced stops than did male speakers.
Past studies identifying a gender difference in the VOT of voiced stops have aver-
aged short lag and negative VOT values together [9, 15]. Examination of the results
from these studies indicates that negative VOT values tend to occur more often
among male speakers compared to females. This finding is supported in a videoen-
doscopic study by Cooke et al. [42] who found a tendency (although statistically
nonsignificant) for males to produce voicing onset sooner than females, indicating
that the distance between the vocal processes is less for males than females. In the
present study, it was suggested that the practice of including negative VOT values in
the voiced stop ‘category’ may distort the pattern of voiced VOT production among
gender groups. By examining only voiced stops (with a voicing lag), as in the present
study, the gender groups appear more similar.
The influence of speaking rate on VOT production was inferred by examining
the duration of CV tokens. It was assumed that VOT values would be directly in-
fluenced by CV duration, with longer durations contributing to longer VOT values
[30]. There were no durational differences between groups in their production of
either voiced or voiceless CVs in the laboratory setting; yet women produced voice-
less stops with longer VOT values. Differences existed between genders in the pro-
duction of CV tokens in the non-laboratory setting, with men producing voiced
and voiceless CV tokens with a significantly shorter duration compared with wom-
en. However in light of these CV durational differences, the same pattern of VOT
production found in the laboratory setting was observed in the non-laboratory set-
ting. The finding of a consistent pattern of VOT production among men and wom-
en across both environmental settings, in spite of variable production of CV dura-
tions would suggest that speaking rate had no clear effect on VOT production, thus
lending additional support for a biological influence on VOT. This finding is sup-
ported by the recent work of Allen et al. [43] who found that VOT productions
among men and women tend to vary even when differences in speaking rate are
controlled.
The present results provide insight regarding a sociophonetic influence on speech
production behavior. Both men and women were found to produce CV tokens with
significantly longer durations in the laboratory compared to non-laboratory setting.
Once outside the laboratory setting, both genders reduced the duration of CV pro-
ductions, with men demonstrating a considerably reduced CV duration compared
to women. Even though both groups were found to produce CV tokens more slowly
in the laboratory setting, female speakers tended to persist in producing the CV to-
kens slowly in the non-laboratory setting. Two observations can be made from the
findings for CV duration. First, it appears that speaking environment may indeed
have an influence on speech production behavior, as evidenced by both gender groups
speaking more slowly in the laboratory setting. Secondly, the female group’s persis-
tence in producing CV tokens slowly in the non-laboratory setting would indicate
that the group treated both settings similarly (i.e., as laboratory settings). This find-
ing would lend support to the suggestion that females tend to use more carefully ar-
ticulated speech in experimental settings compared with males [35, 44]. However,
environmental setting may tend to have a greater influence on rather long speech
segments (such as CV syllables) as opposed to short segments (such as VOT). Future
research examining the influence of environmental setting on speech segments of
varying length seems warranted.
198.143.32.65 - 1/27/2016 12:19:26 AM

Gender and VOT Folia Phoniatr Logop 131


University of Pittsburgh

2005;57:125–133
Downloaded by:
Creating a truly non-laboratory environment to examine speech production be-
havior may be problematic. As such, the results obtained in the present study should
be considered preliminary. While sex-linked differences in voiceless VOT produc-
tion appear to be biologically determined, a word of caution is offered for future
studies comparing the speech production behavior of men and women. Method-
ological and environmental variables should not be overlooked.

References

1 Denes P: Effect of duration on the perception of voicing. J Acoust Soc Am 1955; 27: 761–764.
2 Lisker L, Abramson A: Some effects of context on voice onset time in English stops. Lang Speech 1967;
10: 1–28.
3 Abramson A, Lisker L: Voice-timing perception in Spanish word-initial stops. J Phonet 1973; 1: 1–8.
4 Hutters B: Vocal fold adjustments in aspirated and unaspirated stops in Danish. Phonetica 1985; 42:
1–24.
5 Ito J, Mester A: The phonology of voicing in Japanese: Theoretical consequences of morphological ac-
cessibility. L1 1986; 17: 49–73.
6 Kent R, Read C: Acoustic Analysis of Speech, ed 2. Albany, Delmar, 2002.
7 Eguchi S, Hirsh I: Development of speech sounds in children. Acta Otolaryngol (suppl) 1969; 257: 1–
51.
8 Sweeting P, Baken R: Voice onset time in a normal-aged population. J Speech Hear Res 1982; 25: 129–
134.
9 Ryalls J, Zipprer A, Baldauff P: A preliminary investigation of the effects of gender and race on voice
onset time. J Speech Lang Hear Res 1997; 40: 642–645.
10 Davis K: Stop voicing in Hindi. J Phonet 1994; 22: 177–193.
11 Kagaya R: A fiberscopic and acoustic study of the Korean stops, affricates, and fricatives. J Phonet
1974; 2: 161–180.
12 Borden G, Baer T, Kenney M: Onset of voicing in stuttered and fluent utterances. J Speech Hear Res
1985; 28: 363–372.
13 Gilbert H, Campbell M: Voice onset time in the speech of hearing-impaired individuals. Folia Phoniatr
1978; 30: 67–81.
14 Koenig L: Laryngeal factors in voiceless consonant production in men, women, and 5-year-olds. J
Speech Lang Hear Res 2000; 43: 1211–1228.
15 Swartz B: Gender differences in voice onset time. Percept Motor Skills 1992; 75: 983–992.
16 Whiteside S, Irving C: Speakers’ sex differences in voice onset time: Some preliminary findings. Percept
Motor Skills 1997; 85: 459–463.
17 Whiteside S, Marshall J: Developmental trends in voice onset time: Some evidence for sex differences.
Phonetica 2001; 58: 196–210.
18 Smith B: Effects of place of articulation and vowel environment on voiced stop consonant production.
Glossa 1978; 12: 163–175.
19 Stathopoulos E, Sapienza C: Respiratory and laryngeal function of women and men during vocal in-
tensity variation. J Speech Hear Res 1993; 36: 64–75.
20 Hanson H: Glottal characteristics of female speakers: Acoustic correlates. J Acoust Soc Am 1997; 101:
466–481.
21 Hanson H, Chuang E: Glottal characteristics of male speakers: Acoustic correlates. J Acoust Soc Am
1999; 106: 1064–1077.
22 Higgins M, Netsell R, Schulte L: Vowel related differences in laryngeal, articulatory, and phonatory
function. J Speech Lang Hear Res 1998; 41: 712–724.
23 Titze I: Physiologic and acoustic differences between male and female voices. J Acoust Soc Am 1989;
85: 1699–1707.
24 Hillenbrand J, Getty L, Clark M, Wheeler K: Acoustic characteristics of American English vowels. J
Acoust Soc Am 1995; 97: 3099–3111.
25 Holmberg E, Hillman R, Perkell J: Glottal airflow and transglottal air pressure measurements for male
and female speakers in soft, normal and loud voice. J Acoust Soc Am 1988; 84: 511–529.
26 Bless D, Abbs J: Vocal fold physiology: Contemporary research and clinical issues. San Diego, College-
Hill Press, 1983.
27 Klatt D: Voice onset time, frication, and aspiration in word-initial consonant clusters. J Speech Hear
Res 1975; 18: 686–706.
28 Kessinger R, Blumstein S: Effects of speaking rate on voice-onset time and vowel production: Some
implications for perception studies. J Phonet 1988; 26: 117–128.
198.143.32.65 - 1/27/2016 12:19:26 AM

132 Folia Phoniatr Logop Robb/Gilbert/Lerman


University of Pittsburgh

2005;57:125–133
Downloaded by:
29 Port R, Rotunno R: Relation between voice-onset time and vowel duration. J Acoust Soc Am 1979; 66:
654–662.
30 Volaitis L, Miller J: Phonetic prototypes: Influence of place of articulation and speaking rate on the
internal structure of voicing categories. J Acoust Soc Am 1992; 92: 723–735.
31 Kowal S, O’Connell D, Sabin E: Development of temporal patterning and vocal hesitations in sponta-
neous narratives. J Psycholing Res 1975; 4: 195–207.
32 Robb M, Maclagan M, Chen Y: Speaking rates of American and New Zealand varieties of English. Clin
Ling Phonet 2004; 17: 1–16.
33 Walker V: Durational characteristics of youn adults during speaking and reading tasks. Folia Phoniatr
1988; 40: 13–20.
34 Lutz K, Mallard A: Disfluencies and rate of speech in young adult nonstutterers. J Fluency Dis 1986;
11: 307–316.
35 Byrd D: Relation of sex and dialect to reduction. Speech Comm 1994; 15: 39–54.
36 Lisker L, Abramson A: A cross-language study of voicing in initial stops: Acoustical measurements.
Word 1964; 20: 384–422.
37 Cheshire J: Sex and gender in variationist research; in Chambers J, Trudgill P, Chilling-Estes N (eds):
The Handbook of Language Variation and Change. Oxford, Blackwell, 2002.
38 Mattingly I: Speaker variation and vocal tract size. J Acoust Soc Am 1966; 39:S1219.
39 Statistica: ver 5.1 [computer program], Tulsa, OK, StatSoft, Inc, 1997.
40 Zlatin M: Voicing contrast: Perceptual and productive voice onset time characteristics of adults. J
Acoust Soc Am 1974; 56: 981–994.
41 Whiteside S, Henry L, Dobbin R: Sex differences in voice onset time: A developmental study of pho-
netic contexts effects in British English. J Acoust Soc Am 2004; 116: 1179–1183.
42 Cooke A, Ludlow C, Hallet N, Selbie W: Characteristics of vocal fold adduction related to voice onset.
J Voice 1997; 11: 12–22.
43 Allen J, Miller J, DeSteno D: Individual talker differences in voice-onset time. J Acoust Soc Am 2003;
113: 544–552.
44 Henton C, Bladon R: Breathiness in normal female speech: Inefficiency versus desirability. Lang Com-
mun 1985; 5: 221–227.

198.143.32.65 - 1/27/2016 12:19:26 AM

Gender and VOT Folia Phoniatr Logop 133


University of Pittsburgh

2005;57:125–133
Downloaded by:

You might also like