Clinical Linguistics & Phonetics, March 2015; 29(3): 236–245� 2015 Informa UK Ltd.ISSN: 0269-9206 print / 1464-5076 onlineDOI: 10.3109/02699206.2014.991449
Lateral bracing of the tongue during the onset phase of alveolarstops: An EPG study
ALICE LEE, FIONA E. GIBBON, & JUDITH OEBELS
Department of Speech and Hearing Sciences, University College Cork, Cork, Ireland
(Received 31 July 2014; revised 18 November 2014; accepted 20 November 2014)
AbstractAlthough raising the sides of the tongue to form a seal with the palate and upper teeth – lateral bracing –plays a key role in controlling airflow direction, providing overall tongue stability and building up oralpressure during alveolar consonant production, details of this articulatory gesture remain poorly understood.This study examined the dynamics of lateral bracing during the onset of alveolar stops /t/, /d/, /n/ producedby15 typical English-speaking adults using electropalatography. Percent tongue palate contact in the lateralregions over a 150-ms period from the preceding schwa to stop closure was measured. Rapid rising of thesides of the tongue from the back towards the front during the 50-ms period before closure was observed,with oral stops showing significantly more contact than nasal stops. This feature corresponds to well-documented formant transitions detectable from acoustic analysis. Possible explanations for increasedcontact for oral stops and clinical implications are discussed.
Keywords: Alveolar stops, electropalatography, lateral tongue activity
Introduction
Theories of phonetics have always involved some notion of place of articulation (Ladefoged &
Maddieson, 1996) because placement identifies the location within the oral cavity at which major
articulatory events occur. Place of articulation retains a pre-eminent status in phonetic description;
this is probably because early phonetic studies were based on X-rays, which recorded midsagittal
data. Whilst not lessening the importance of place of articulation, Fletcher (1992) highlighted the
importance of contact between the lateral borders of the tongue and the hard palate, especially for
production of lingual stops, lingual fricatives and high vowels. Fletcher (1992) went on to state
that, in order to produce normal alveolar stops, the speaker must ‘‘tense the lateral borders of the
tongue to produce a spoon-shaped configuration’’ (p. 99).
Stone, Faber, Raphael, and Shawker (1992) also discussed the importance of the lateral margins
of the tongue in providing anchorage, which gives stability to the whole tongue during stop
production. Fuchs, Perrier, Geng, and Mooshammer (2006) postulated that the sides of the tongue
play an important role in deceleration of the tongue during the fast ballistic stop movement
Correspondence: Alice Lee, PhD, Department of Speech and Hearing Sciences, Brookfield Health Sciences Complex, University College
Cork, College Road, Cork, Republic of Ireland. Tel: +353 (0)21 490 1540. Fax: +353 (0)21 490 1542. E-mail: [email protected]
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
involved in stop production, and in so doing, these regions of the tongue provide overall tongue
stability in typical speakers. Fuchs et al. (2006) also highlighted that the sides of the tongue allow
a tight intraoral seal to develop for the purpose of building up the necessary oral pressure during
alveolar obstruents and that the action of lateral bracing reduces the degree of freedom of tongue
movement and supports the formation of tongue shapes for articulation. Furthermore, bracing of
the tongue provides somatosensory feedback regarding its position to the speakers (Gick, Wilson,
& Derrick, 2013).
Phonetically, alveolar stops are described as consisting of a sequence of three events: onset
(also called the approach or shutting phase); closure; and release. The approach phase of stops is
of particular importance to normal stop production because changes in vocal tract shape, which
are affected by the actions of the sides of the tongue, are reflected in the formant transitions
contained in the acoustic signal. Kent and Read (2002) stated that a reliable transition period from
the vowel to the stop provides the necessary acoustic transition cues for accurate identification of
sounds in listeners’ perception of stops. Accurate movement control of the sides of the tongue
during the onset period is therefore necessary in order to achieve normal alveolar stops, and also
alveolar fricatives, during speech production.
It is perhaps unsurprising to find that electropalatographic studies have shown that some groups
of individuals with speech disorders are not able to ‘‘brace’’ the tongue in the normal way to form
accurate stop closures (Gibbon, 1999; Timmins, Hardcastle, Wood, & Cleland, 2011). Gibbon and
Wood (2010) described electropalatography (EPG) data, including lateral bracing, from children
with developmental speech disorders. Compared with children with typical speech, these authors
illustrated different types of EPG patterns, all of which involved abnormal lateral contact. For
example, one child made extensive contact across most of the palate during productions, whereas
another child produced a pattern similar to an alveolar stop but with some asymmetry and
incomplete lateral seal. A third child had contact predominantly in the posterior regions of the
palate so had minimal lateral contact for alveolar target sounds. In a summary of current EPG
studies on cleft palate speech, Gibbon and Lee (2010) reported similar difficulties with the control
of the lateral regions of the tongue in the production of stops and fricatives in individuals in this
clinical group. In this group also, some speakers are unable to use the sides of the tongue to form a
complete seal. Where there is an incomplete seal, air is likely to escape and result in sound
distortions, such as ‘‘spirantisation’’, where friction occurs during a stop target, or lateral
fricatives for sibilant targets.
Many EPG studies have reported tongue palate contact for typical articulatory stops.
Traditionally, /t/, /d/ and /n/ in typical English speakers are described as having alveolar
placement, which indicates that they have similar place of articulation, and EPG studies have
confirmed that /t/, /d/ and /n/ have similar shaped tongue palate configurations (Dagenais,
Lorendo, & McCutcheon, 1994; Goozee, Murdoch, & Theodoros, 1999; Hardcastle & Gibbon,
1997; McAuliffe, Ward, & Murdoch, 2001; McLeod, 2006; McLeod & Roberts, 2005). An EPG
study by Gibbon, Yuen, Lee, and Adams (2007) found that /t/ and /d/ were similar, but had more
contact and were more likely to have complete lateral contact than /n/.
Compared with the closure and release phases of stops, which carry the primary acoustic cues
for phoneme identification, the articulatory characteristics of the onset phase are relatively under
researched and have generally received scant attention in the literature. McLeod (2011) surveyed
practising speech and language therapists and found that they did not fully understand the role of
the lateral regions of the tongue during the articulation of speech sounds. McLeod found that
while the therapists demonstrated good knowledge of tongue palate contact along the midline
(sagittal plane), they had poor knowledge about lateral bracing for alveolar consonants. McLeod
concludes that therapists’ knowledge of tongue palate contact ‘‘may be influenced by the sagittal
images of consonant production provided in textbooks’’ (p. 1010).
Lateral bracing in alveolar stops 237
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
The lack of knowledge described above means that it is problematic for researchers or speech
and language therapists to identify the nature and extent of abnormal EPG patterns for stops in
individuals with speech disorders. Given the important role of the lateral tongue regions in the
onset of alveolar stops, the lack of research and understanding is a surprising knowledge gap. One
likely explanation for the relative neglect is the lack of suitable methodologies to investigate
tongue activity in the peripheral areas of the oral cavity. One technique –EPG – is well suited for
the purpose of recording lateral tongue activity because it can detect contact with the lateral
margins of the hard palate as the tongue apex forms a closure against the alveolar ridge. The aim
of this study was to investigate the dynamic characteristics of lateral contact during onset in the
time period leading up to the formation of the alveolar closure in /t/, /d/ and /n/ in typical adult
speakers.
Methods
Participants
The participants were 15 typical adults who were native speakers of English and aged from 24 to
61 years with a mean age of 44 years. Five of them were male (M1–M5) and 10 were female
(F1–F10). None of the participants had a self-reported history of speech, language or hearing
difficulties. The participants were either faculty members at Queen Margaret University in
Edinburgh or speech and language therapists in the UK who used EPG in their clinical work, and
each of them had a custom-made Reading EPG plate (Gibbon et al., 2007). The research project
was approved by the Research Ethics Committee of Queen Margaret University, Edinburgh.
Speech material
The minimal pair sets, a tab/a dab/a nab and a tip/a dip/a nip, were used in this study. Each word
was preceded by a schwa /@/, which elicited target consonants from a neutral tongue position. The
target stimuli were embedded among distracters, for example, ‘‘a Dee, a nab, a tomb, a sip’’, and
these phrases were presented in a randomised order. Each participant was instructed to repeat the
entire list of phrases 10 times at a natural, conversational rate, yielding 60 tokens of target stimuli
per participant. Due to technical problems, six tokens were discarded, leaving 894 tokens for
analysis.
Instrumentation and annotation
The WinEPGTM system (Articulate Instruments, Edinburgh, UK) was used for data recording
(Wrench, Gibbon, McNeill, & Wood, 2002). The EPG data were sampled at 100 Hz
simultaneously with the acoustic signal at 22 050 Hz. The Reading EPG plate used in this study
contained 62 electrodes arranged in eight horizontal rows, with six electrodes in the most anterior
row (row 1) and eight in rows 2–8 (Hardcastle & Gibbon, 1997). Annotation of the EPG data was
carried out using the Articulate Assistant software, version 1.18 (Wrench et al., 2002). The
annotation used in Gibbon et al.’s (2007) study – the closure phase of the /t, d, n/ segments defined
from the first EPG frame to the last that showed complete constriction – was used as reference for
the new annotation in this study. As illustrated in Figure 1, using the first EPG frame of complete
constriction as the landmark (frame 234), the annotated segment started from the 10th EPG frame
preceding the landmark to the fourth EPG frame following this frame. That is, each annotated
238 A. Lee et al.
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
segment contained 15 EPG frames (total period of 150 ms) consisting of three phases each of five
EPG frames (50 ms): the first phase capturing data during the preceding schwa vowel; the second
phase, capturing the transition from the vowel to the following alveolar stop; and the third phase
capturing the start of the closure phase of the stop. All preceding schwas were approximately
100 ms in duration.
Data analysis
The amount of contact between the tongue and the lateral area of the palate (expressed as a
percentage) was calculated for each of the 15 EPG frames within the annotated segment for the
894 tokens. The lateral area was defined as the two outermost columns on the left side and the two
outermost columns on the right side of the EPG plate, extending from row 2 to row 8 of electrodes
(Gibbon & Nicolaidis, 1999; Nicolaidis & Baltazani, 2011). The measurement was made, and the
data exported using the ‘‘data export’’ function of the Articulate Assistant software.
For the statistical analysis, first, the percent lateral contact data was aggregated over the five
EPG frames within each phase for each of the productions for each speaker, yielding a total of
2682 data points. The mean percent lateral contact (predictive variable) was then analysed in a
linear mixed effect (LME) model (with maximum likelihood estimation) with consonant type
(three levels), vowel context (two levels) and phase (three levels) as fixed effects; and speaker
(n¼ 15) as random effect using IBM SPSS Statistics 21. The two-way interactions, conson-
ant� vowel, consonant� phase and vowel� phase, and the three-way interaction, conson-
ant� vowel� phase were included as fixed effects as well.
Results
The mean percent lateral contact of all speakers over the 15 EPG frames for the three target
consonants /t/, /d/ and /n/ is plotted in Figure 2. In terms of a general description of the results, the
three target consonants showed a similar trend in terms of the change in the amount of lateral
contact overtime – the mean percent lateral contact was low during the preceding schwa (phase
1 – EPG frames 1–5), followed by a rapid increase in lateral contact in phase 2 (frames 6–10)
before forming complete constriction for the stops in phase 3 (frames 11–15). The speakers had
lower mean percent lateral contact for the nasal stop /n/ compared with the oral stops /t, d/, and
this difference appeared to be greater during phases 2 and 3.
Figure 1. Illustration of annotation for ‘‘a tab’’ produced by one participant (F10). The line marking ‘‘Start’’ in
the waveform indicates the beginning of the annotated segment and it corresponds to EPG frame 224 at the bottom.
The line ‘‘Complete constriction’’ corresponds to EPG frame 234 (the beginning of the closure phase) and the line ‘‘End’’
to frame 238.
Lateral bracing in alveolar stops 239
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
The results of the statistical analysis showed that, after the non-significant effects were
removed, the reduced LME model consisted of the fixed effects, consonant type, vowel context
and phase, and the two-way interaction, consonant� phase. The full model did not provide a
better fit than the reduced model, as indicated by the absence of significant difference in the –2 log
likelihood value between the two models (�2¼ 7.74, df¼ 8, p¼ 0.459). There were significant
main effects for consonant type and phase, and a significant interaction effect for conson-
ant� phase; however, the main effect for vowel context was not significant (see Table 1 for
results). Analyses of simple main effects of consonant within phase were conducted –
differences between consonants were found at phase 1 [F(2, 2667)¼ 18.39, p50.001], phase 2
[F(2, 2667)¼ 99.87, p50.001] and phase 3 [F(2, 2667)¼ 144.40, p50.001]. Pairwise
comparisons revealed that there were significant differences in the percent lateral contact
between the three consonants. In phases 1 and 2, /t/ had higher mean percent lateral contact than
/d/ and /n/; and /d/ had higher mean percent lateral contact than /n/. For phase 3, both /t/ and /d/
showed higher mean percent lateral contact than /n/; however, the difference between /t/ and /d/
was not significant (Table 2).
Figure 2. Mean percent lateral contact of the 15 typical adult speakers over the 15 EPG frames, with the first five frames
capturing the amount of contact during the preceding schwa (phase 1), the following five frames capturing the transition
from the vowel to the following alveolar stop (/t, d, n/) (phase 2) and the last five frames showing the start of the closure
phase of the alveolar stops (phase 3).
Table 1. Results of the tests of fixed effects.
Predictor variable F df p
Consonant type 223.94 2 50.001
Vowel context 0.13 1 0.723
Phase 8903.93 2 50.001
Consonant� phase 19.35 4 50.001
240 A. Lee et al.
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
Figure 3 shows an example of EPG contact patterns for the target consonants /t, d, n/ produced
in the minimal seta tab/a dab/a nab by one participant (M4). This figure shows that the contact
pattern trajectories for all the target consonants were similar – the contact started in the posterior
lateral region of the hard palate and moved rapidly forward towards the alveolar region over a
50-ms period. The rapid increase of lateral contact during phase 2 is further detailed in Figure 4.
This figure shows the mean percent lateral contact of all speakers for each row (row 2–8) from
EPG frame 6 through 11. There was contact mainly in the back (rows 7 and 8) at frame 6 and the
contact increased rapidly forwards. The darker shades evident for /t/ and /d/ in Figure 4 indicate
that the mean percent contact for these stops was higher compared to /n/. In other words, the
dynamics of the sides of the tongue during onset phase show that oral stops are produced with
greater amounts of contact for /t/ and /d/ compared to the nasal stop.
Discussion
A consistent finding of the present study is that speakers showed a rapid increase in lateral contact
during the 50-ms period prior to full alveolar closure for oral stops /t, d, n/ (i.e. phase 2). This
rapid increase in lateral contact is interpreted as reflecting a rapid upward movement of the sides
Table 2. Results of pairwise comparisons regarding the percent lateral contact between /t/ and /d/,
/t/ and /n/ and /d/ and /n/ for phases 1, 2 and 3.
95% CI
Phase
Pairwise
comparison
Mean difference in
percent lateral contact p Lower limit Upper limit
1 /t/ vs. /d/ 2.1% 0.004 0.5% 3.8%
/t/ vs. /n/ 4.1% 50.001 2.5% 5.7%
/d/ vs. /n/ 1.9% 0.012 0.3% 3.5%
2 /t/ vs. /d/ 3.7% 50.001 2.0% 5.3%
/t/ vs. /n/ 9.4% 50.001 7.8% 11.1%
/d/ vs. /n/ 5.8% 50.001 4.2% 7.4%
3 /t/ vs. /d/ 1.4% 0.134 �0.3% 3.0%
/t/ vs. /n/ 10.5% 50.001 8.9% 12.1%
/d/ vs. /n/ 9.2% 50.001 7.5% 10.8%
Figure 3. An illustration of EPG contact patterns over the 15 EPG frames for /t, d, n/ produced in the minimal pair seta
tab/a dab/a nab by one participant (M4).
Lateral bracing in alveolar stops 241
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
of the tongue immediately prior to alveolar closure. The contact in all speakers began in the
posterior lateral region of the hard palate and moved rapidly forwards until closure in the alveolar
region was attained. This finding provides physiological evidence in support of Kent and Read’s
(2002) view that a reliable temporal constant of stop articulations is a 50-ms transition period
from the vowel to the stop and that ‘‘within this 50-ms interval, all formant frequencies shift from
their values from the vowel to the stop’’ (p. 155).
The findings from this study in relation to typical speakers’ onset phase productions will
provide valuable normative data for researchers and speech and language therapists when
interpreting EPG data from individuals with speech disorders. Although the onset phase contains
fewer of the major acoustic cues to identification compared to the closure and release phases of
stop production, nevertheless any abnormal actions of the sides of the tongue have clinical
implications. Abnormal tongue dynamics or positioning during the onset phase would hinder, or
even prevent, the formation of a normal stop, which in turn would have deleterious consequences
on speech intelligibility. For example, one consequence of abnormal lateral activity during the
onset phase would be diminished tongue apex control, with the result that the closure and release
phases would become unstable and would likely involve abnormal place of articulation. To
illustrate this point, Gibbon (1999, p. 389) reported (in Figure 4 of the paper) the EPG data of
word-final /d/ in the word ‘‘shed’’ produced by an eight-year-old girl with a speech sound
disorder. During the onset phase of /d/, the tongue palate contact can be seen to start in the
posterior lateral region of the hard palate; however, an atypical contact pattern develops thereafter.
The contact moves forward but it is confined to the palatal and velar regions of the hard palate.
There is no contact between the tongue apex and the alveolar region. Moreover, during the 50 ms
before /d/ closure, increased contact towards the middle of the hard palate in the palatal and velar
regions can be observed. Although the production in this particular example was perceived as a
‘‘correct’’ [d] by listeners, probably due to the release occurring in the post-alveolar region.
Abnormal tongue palate contact during closure and release phases, such as the one described
above, could have direct consequences for the acoustic cues necessary for sound identification.
Mayo, Gibbon, and Clark (2013) investigated how listeners’ perception of conflicting acoustic
cues, similar to those present in disordered articulations, affected listeners’ categorisation of place
of articulation for lingual stops. These authors found that listeners did not assign consistent
categorical placement labels (such as /d/ or /g/) when conflicting transitional cues were present in
the speech stimuli. Their study suggested that listeners do not interpret the presence of conflicting
acoustic cues in a consistent, categorical way, and this inconsistency in listener perceptions would
negatively impact on speech intelligibility.
The importance of a lateral bracing during the onset phase of alveolar stops and fricatives
production for individuals with speech disorders also has an important developmental perspective.
Figure 4. Mean percent lateral contact of the 15 typical adult speakers for each row (rows 2–8) from EPG frame 6 through
11 for the target consonants /t, d, n/. Frame 11 was the point where complete constriction (closure) in the alveolar region
was formed.
242 A. Lee et al.
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
Normal motor control of the sides of the tongue is thought to underpin the production and
subsequent development of other sounds (Fletcher, 1992; Gibbon, 1999). Fletcher (1992)
maintains that the ability to produce normal anterior stops is a prerequisite for the development of
velar and sibilant sounds, stating that ‘‘the anterior stop gesture is postulated to serve as a
fountainhead for the remaining lingual consonant gestures’’ (Fletcher, 1992, p. 99). A corollary of
Fletcher’s view is that any difficulties controlling the movement of the lateral regions of the
tongue are likely to have far reaching negative consequences not just for the intelligibility of
speech but also for the acquisition of other speech sounds that require similar lingual gestures.
Although the trajectories of lateral contact were similar for oral and nasal stops, the results
revealed a significant difference in the amount of lateral contact displayed by stops. The voiceless
stop /t/ had significantly more contact than the voiced stop /d/ in phases 1 and 2, and both oral
stops had significantly more contact than the nasal stop /n/ in phases 1, 2 and 3. This was a robust
finding that occurred in all speakers, regardless of individual differences in the overall amount of
contact displayed. The finding of higher amounts of lateral contact for oral compared to nasal
stops is consistent with the results from the studies by Gibbon et al. (2007) and McLeod (2006).
A higher amount of lateral contact for oral stops is interpreted as indicating that the sides of the
tongue are in a higher position within the vocal tract for oral compared with nasal stops. This
finding supports a previous study by Yang and Stone (2002), who used ultrasound imaging to
describe the shape of the surface of the tongue during various speech sounds. These authors found
that the sides of the tongue were raised for /t/, and at the same time, the tongue was deeply grooved
in the centre behind the alveolar stop closure. In comparison, the tongue had a much flatter
configuration during /n/. Considering together the findings of this study and the study by Yang and
Stone (2002), the EPG data from this study suggest that the tongue configuration and its position
within the vocal tract are different for the oral and nasal alveolar stops, including the onset phase
of their production.
It has been speculated that the reason for increased contact for /t/ and /d/ compared with /n/
may relate to the increased intraoral pressure requirements for producing a spectral burst for oral
stops (Gibbon et al., 2007; Subtelny, Worth, & Sakuda, 1966). The higher amount of tongue–
palate contact for oral stops compared with the nasal stop may be due to increased lingual pressure
against the hard palate necessary to meet the intraoral air pressure requirements. For the nasal
stop, the overall flatter tongue configuration reported for /n/ (Yang & Stone, 2002) may be
important for creating a higher acoustic impedance to sound transmission into the oral cavity, in
order to facilitate the sound’s transmission into the nasal cavity (Jones, 1991). In addition,
previous studies have reported that the jaw position was higher during the production oral stops /t,
d/ compared with nasal stop /n/ (e.g. Mooshammer, Hoole, & Geumann, 2006; Tabain, 2012).
These authors explained that the higher jaw position for the alveolar oral stops was probably due
to the need to use the lower teeth as a secondary obstacle noise source for producing the spectral
burst (Mooshammer et al., 2006; Tabain, 2012). It is possible that the different tongue positions
for alveolar oral stops and nasal stop might have been influenced by the different jaw positions for
the two types of speech sounds. It is postulated that the production of alveolar oral stops and nasal
stop do not only just differ by the presence of nasal coupling for the nasal stop but also the use of
different tongue shapes and positions and jaw positions, for the purpose of controlling the balance
of oral–nasal resonance and the production of spectral burst that requires adequate intraoral
pressure and the presence of obstacle noise source for the oral stops.
The finding in this study of increased lateral contact during /t, d/ compared with /n/, which
occurred even during the preceding vowel, is interpreted as a feature of anticipatory coarticulation
for the sound-specific tongue configurations required for the upcoming stop closures. Compared
with the front and back vowels, the schwa /@/ has neutral articulatory setting; hence, it is less
resistant to the anticipatory coarticulatory effects brought by the motor planning or programming
Lateral bracing in alveolar stops 243
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
for the following consonant (Hixon, Weismer, & Hoit, 2008; Recasens, 1999). This study did not
find a significant main effect of vowel context (/æ/ and /i/) on the percent lateral contact of /t, d, n/
during the preceding schwa vowel, transition from the schwa to the following alveolar stop and the
first 50 ms of the closure period of the stop. However, a significant vowel effect was reported in
the study by Gibbon et al. (2007). The different findings were probably related to the fact that
slightly different EPG measures were used (percent contact of the entire palate in the Gibbon et al.
study) and that, more importantly, the measurements were taken at different temporal locations
(closure period of the alveolar stops in the study by Gibbon et al.). Hence, these findings showed
that there was no significant anticipatory coarticulatory effect brought by the second vowel to the
first vowel (schwa), the schwa-alveolar stop transition and the first 50 ms of the closure period of
the alveolar stop in a schwa-alveolar stop-vowel sequence.
Like many previous studies, this study investigated the phenomenon of lateral bracing in a
specific phonetic context, namely the onset phase of alveolar stops. However, it is relevant to
consider bracing in a broader context. Gick, Allen, Stavness, and Wilson (2013) recently
investigated lateral bracing as a pervasive feature of continuous speech. These authors used EPG
in combination with ultrasound and electromagnetic articulometry to investigate lateral bracing
during speech and found that speakers ‘‘brace the tongue continuously, except during a small
percentage of low vowels, and during a larger percentage of instances of /l/’’ (p. 4204). They
concluded that ‘‘bracing is a constant and necessary aspect of tongue motor control’’ (p. 4204).
Acknowledgements
The authors would like to thank Kathleen O’Sullivan for her advice and help with the statistical
analysis.
Declaration of interest
The authors report no conflicts of interest.
References
Dagenais, P. A., Lorendo, L. C., & McCutcheon, M. J. (1994). A study of voicing and context effects upon consonant
linguapalatal contact patterns. Journal of Phonetics, 22, 225–238.
Fletcher, S. G. (1992). Articulation: A physiological approach. San Diego, CA: Singular.
Fuchs, S., Perrier, P., Geng, C., & Mooshammer, C. (2006). What role does the palate play in speech motor control?
Insights from tongue kinematics for German alveolar obstruents. In J. Harrington & M. Tabain (Eds.), Speech
production: Models, phonetics processes, and techniques (pp. 149–164). New York: Psychology Press.
Gibbon, F. E. (1999). Undifferentiated lingual gestures in children with articulation/phonological disorders. Journal of
Speech, Language, and Hearing Research, 42, 382–397.
Gibbon, F. E., & Lee, A. (2010). Producing turbulent speech sounds in the context of cleft palate. In S. Fuchs, M. Toda, &
M. Zygis (Eds.), Turbulent sounds: An interdisciplinary guide (pp. 303–341). Berlin: de Gruyter.
Gibbon, F., & Nicolaidis, K. (1999). Palatography. In W. J. Hardcastle & N. Hewlett (Eds.), Coarticulation: Theory, data
and techniques (pp. 229–245). Cambridge: Cambridge University Press.
Gibbon, F., & Wood, S. (2010). Visual feedback therapy with electropalatography (EPG) for speech sound disorders in
children. In L. Williams, S. McLeod, & R. McCauley (Eds). Interventions in speech sound disorders (pp. 509–536).
Baltimore: Brookes.
Gibbon, F. E., Yuen, I., Lee, A., & Adams, L. (2007). Normal adult speakers’ tongue palate contact patterns for oral and
nasal stops. Advances in Speech-Language Pathology, 9, 82–89.
Gick, B., Allen, B., Stavness, I., & Wilson, I. (2013). Speaking tongues are always braced. The Journal of the Acoustical
Society of America, 134, 4204.
244 A. Lee et al.
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
Gick, B., Wilson, I., & Derrick, D. (2013). Articulatory phonetics. Malden, MA: Wiley-Blackwell.
Goozee, J. V., Murdoch, B. E., & Theodoros, D. G. (1999). Electropalatographic assessment of articulatory timing
characteristics in dysarthria following traumatic brain injury. Journal of Medical Speech-Language Pathology, 7,
209–222.
Hardcastle, W. J., & Gibbon, F. E. (1997). Electropalatography and its clinical applications. In M. J. Ball, & C. Code
(Eds.), Instrumental clinical phonetics (pp. 149–193). London: Whurr Publishers Ltd.
Hixon, T. J., Weismer, G., & Hoit, J. D. (2008). Preclinical speech science: Anatomy physiology acoustics perception. San
Diego, CA: Plural Publishing.
Jones, D. L. (1991). Velopharyngeal function and dysfunction. Clinics in Communication Disorders, 1, 19–25.
Kent, R.D., & Read, C. (2002). The acoustic analysis of speech (2nd ed.). New York: Singular/Thomson Learning.
Ladefoged, P., & Maddieson, I. (1996). The sounds of the world’slanguages. Malden, MA: Blackwell Publishing.
Mayo, C., Gibbon, F., & Clark, R. (2013). Phonetically trained and untrained adults’ transcription of place of articulation
for intervocalic lingual stops with intermediate acoustic cues. Journal of Speech Language and Hearing Research, 56,
779–791.
McAuliffe, M. J., Ward, E. C., & Murdoch, B. E. (2001). Tongue-to-palate contact patterns and variability of four English
consonants in an /i/ vowel environment. Logopedics Phoniatrics Vocology, 26, 165–178.
McLeod, S. (2006). Australian adults’ production of /n/: An EPG investigation. Clinical Linguistics and Phonetics, 20,
99–107.
McLeod, S. (2011). Speech-language pathologists’ knowledge of tongue/palate contact for consonants. Clinical Linguistics
and Phonetics, 25, 1004–1013.
McLeod, S., & Roberts, A. (2005). Templates of tongue/palate contact for speech sound intervention. In C. Heine, & L.
Brown (Eds.), Proceedings of the 2005 Speech Pathology Australia National Conference (pp. 104–112). Melbourne:
Speech Pathology Australia.
Mooshammer, C., Hoole, P., & Geumann, A. (2006). Interarticulator cohesion within coronal consonant production.
Journal of the Acoustical Society of America, 120, 1028–1039.
Nicolaidis, K., & Baltazani, M. (2011). An electropalatographic and acoustic study of the Greek rhotic in /Cr/ cluster.
Proceedings of the 17th International Congress of Phonetic Sciences (ICPhS XVII), Hong Kong (pp. 1474–1477).
Retrieved December 13, 2012, from http://www.icphs2011.hk/resources/OnlineProceedings/RegularSession/
Nicolaidis/Nicolaidis.pdf.
Recasens, D. (1999). Lingual coarticulation. In W. J. Hardcastle, & N. Hewlett (Eds.), Coarticulation: Theory, data and
techniques (pp. 80–104). Cambridge: Cambridge University Press.
Stone, M., Faber, A., Raphael, L. J., & Shawker, T. H. (1992). Cross-sectional tongue shape and linguopalatal contact
patterns in [s], [R
], and [l]. Journal of Phonetics, 20, 253–270.
Subtelny, J. D., Worth, J. H., & Sakuda, M. (1966). Intraoral pressure and rate of flow during speech. Journal of Speech
and Hearing Research, 9, 498–518.
Tabain, M. (2012). Jaw movement and coronal stop spectra in Central Arrernte. Journal of Phonetics, 40, 551–567.
Timmins, C., Hardcastle, W. J., Wood, S., & Cleland, J. (2011). An EPG analysis of /t/ in young people with Down’s
syndrome. Clinical Linguistics & Phonetics, 25, 1022–1027.
Wrench, A. A., Gibbon, F. E., McNeill, A. M., & Wood, S. E. (2002). An EPG therapy protocol for remediation and
assessment of articulation disorders. In J. H. L. Hansen, & B. Pellom (Eds.), Proceedings of the 7th International
Conference on Spoken Language Processing (ICSLP’02), Denver, USA (pp. 965–968). Retrieved March 7, 2005, from
http://eresearch.qmu.ac.uk/2291/1/wrench02_icslp.pdf.
Yang, C., & Stone, M. (2002). Dynamic programming method for temporal registration of three-dimensional tongue
surface motion from multiple utterances. Speech Communication, 38, 201–209.
Lateral bracing in alveolar stops 245
Clin
Lin
guis
t Pho
n D
ownl
oade
d fr
om in
form
ahea
lthca
re.c
om b
y U
nive
rsity
Col
lege
Cor
k on
03/
20/1
5Fo
r pe
rson
al u
se o
nly.
Top Related