The Effect of Stress and Speech Rate on Vowel Coarticulation in

Comments

Transcription

The Effect of Stress and Speech Rate on Vowel Coarticulation in
Complimentary Author PDF: Not for Broad Dissemination
JSLHR
Research Article
The Effect of Stress and Speech Rate on
Vowel Coarticulation in Catalan
Vowel–Consonant–Vowel Sequences
Daniel Recasensa
Purpose: The goal of this study was to ascertain the effect of
changes in stress and speech rate on vowel coarticulation in
vowel–consonant–vowel sequences.
Method: Data on second formant coarticulatory effects as
a function of changing /i/ versus /a/ were collected for
five Catalan speakers’ productions of vowel–consonant–
vowel sequences with the fixed vowels /i/ and /a/ and
consonants: the approximant /δ/, the alveolopalatal nasal
/ɲ /, and /l/, which in the Catalan language differs in darkness
degree according to speaker.
Results: In agreement with predictions formulated by the
degree-of-articulation-constraint model of coarticulation,
the size of the vowel coarticulatory effects was inversely
related to the degree of articulatory constraint for the
consonant, and the direction of those effects was mostly
carryover or anticipatory in vowel–consonant–vowel sequences
with highly constrained consonants (/ɲ/, dark /l/) and more
variable whenever the intervocalic consonant was less
constrained (/δ/, clear /l/). Stress and speech-rate variations
had an effect on overall vowel duration, second formant
frequency, and coarticulation size but not on the consonantspecific patterns of degree and direction of vowel
coarticulation.
Conclusion: These results indicate that prosodically
induced coarticulatory changes conform to the basic
principles of segmental coarticulatory organization.
C
the vowel effects, namely, whether the V2-dependent anticipatory effects extending to the preceding consonant and
transconsonantal V1 are more or less prominent than the V1dependent carryover effects extending forward to the following consonant and transconsonantal V2. Thus, for example,
although /a/ may cause the tongue dorsum to lower during
preceding /t/ and V1 = /i/ in the sequence /ita/ and during following /t/ and V2 = /i/ in the sequence /ati/, the prominence
of the former (anticipatory) and latter (carryover) tonguedorsum lowering effects may not be the same. The study of
the direction of vowel coarticulation is of particular importance because it provides some insight into the cognitive and
peripheral mechanisms involved in segmental production in
running speech (Farnetani & Recasens, 2010; Grosvald,
2009; Guenther, 1995). In particular, anticipatory effects have
been taken to reflect segmental planning and carryover effects
to depend mostly on inertia and the biomechanical characteristics of the vocal-tract articulators (Whalen, 1990).
The present study analyzed these research issues
using second formant second formant (F2) data on coarticulatory effects exerted by /i/ or /a/ (also referred to throughout as the changing vowel) in Catalan VCV sequences with
three consonants—the alveolopalatal nasal /ɲ/, the alveolar
lateral /l/ (which may vary from strongly dark to less dark or
moderately clear according to speaker in present-day Catalan;
onsecutive phonetic segments influence each other
at the articulatory and acoustic level in the speech
chain. These coarticulation effects occur, for example, between vowels and consonants in vowel–consonant–
vowel (VCV) sequences. The present study deals with the
role that the segmental composition of a VCV sequence
plays in the extent to which the first or second vowel (V1
or V2) affects the other two phonetic segments. Two major
issues will be subject to investigation. First is the size of the
vowel coarticulatory effects during the intervocalic consonant (V-to-C coarticulation) and the size and temporal
extent of those effects during the transconsonantal vowel
(V-to-V coarticulation). Thus, for example, coarticulation
data show that V2 = /a/ may cause the tongue dorsum
to lower during the preceding consonant and V1 = /i/ in
the sequences /ipa/ and /ita/ and that these effects are
larger and last longer in the former sequence than in the
latter because /p/ involves no lingual activity, whereas /t/
does. The second major research topic is the direction of
a
Universitat Autònoma de Barcelona, Spain
Correspondence to Daniel Recasens: [email protected]
Editor: Jody Kreiman
Associate Editor: Ewa Jacewicz
Received July 15, 2014
Revision received February 2, 2015
Accepted June 18, 2015
DOI: 10.1044/2015_JSLHR-S-14-0196
Disclosure: The author has declared that no competing interests existed at the time
of publication.
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015 • Copyright © 2015 American Speech-Language-Hearing Association
1407
Complimentary Author PDF: Not for Broad Dissemination
Simonet, 2010), and the dental approximant /δ/—and the
transconsonantal vowels /i/ and /a/ (also referred to as the
fixed vowel). In principle, F2 coarticulation should be positively correlated with tongue-dorsum coarticulation—that
is, the higher and more anterior the tongue body, the higher
the F2 frequency (Fant, 1960). Anticipatory effects from V2 =
/i/ versus /a/ will be measured in the sequence pairs /aCi/–/aCa/
and /iCa/–/iCi/, and carryover effects from V1 = /i/ versus /a/
will be analyzed in the sequence pairs /iCa/–/aCa/ and /aCi/–
/iCi/. These sequence pairings allow an evaluation of essentially how much changing /i/ affects the consonant and fixed
/a/ and how much changing /a/ affects the consonant and
fixed /i/ (e.g., in pairs such as /aCi/–/aCa/ and /iCi/–/iCa/,
respectively) along the two coarticulatory dimensions. Analysis results will be evaluated within the framework of the
degree-of-articulatory-constraint (DAC) model of coarticulation (see Recasens, Pallarès, & Fontdevila, 1997).
A novelty of the present investigation with respect
to previous lingual coarticulation studies (for a summary,
see Recasens, 1999) is elicitation of the extent to which
changes in stress and speech rate affect the degree and
direction of vowel coarticulation in VCV sequences that
differ regarding segmental composition. It has been known
for some time that mostly vowels, but also consonants,
undergo articulatory undershoot and acoustic reduction in
unstressed versus stressed position and in fast versus normal (also in normal vs. overarticulated) speech (Agwuele,
Sussman, & Lindblom, 2008; Farnetani, 1990; Gay, 1978;
Kent & Netsell, 1971; Moon & Lindblom, 1994; but see
Mok, 2011, for studies failing to report more vowel undershoot as speech rate increases). The general prediction is
that, given that stressed vowels are produced with more
prominent articulatory gestures than unstressed vowels, the
former ought to exert more coarticulation on the latter in
VCV sequences than vice versa. There is indeed supporting
evidence in the literature that unstressed vowels are more
sensitive than stressed vowels to V-to-V effects (Fowler,
1981; Magen, 1997; Yun, 2007; but see Cho, 2004). Also
in line with previous experimental evidence (Krull, 1989;
Matthies, Perrier, Perkell, & Zandipour, 2001), phonetic
segments should be more reduced and more prone to coarticulate with other segments at faster versus slower speech
rates and in less versus more formal speech.
The remainder of the introduction first describes how
the articulatory constraints for the consonants and vowels
subject to investigation in the present study affect the degree of vowel coarticulation and render vowel coarticulation more prominent at the anticipatory or carryover level
in VCV sequences. It also deals with the effect of changes
in stress and speech rate on the degree and direction of
vowel coarticulation in the segmental conditions subject to
analysis. A set of testing hypotheses about these research
goals is summarized at the end of the introduction.
Degree of Coarticulation
The DAC model predicts that the extent to which
vowels affect consonants in VCV sequences should vary
1408
inversely with the degree of tongue-dorsum constraint for
the intervocalic consonant and, more specifically, with the
extent to which the consonantal primary articulator is
involved in the formation of a closure or a constriction.
The model was formulated by Recasens et al. (1997) after
several studies showing that the extent to which consonants resist coarticulation from adjacent vowels is positively conditioned by closure size or degree of constriction
narrowing.
Table 1 (top, second and third columns) indicates the
predicted degrees of coarticulatory resistance to the vowel
effects for the consonants under investigation in the present
study and whether vowel effects should be exerted by /i/
rather than by /a/ or vice versa. The consonants /ɲ/ and
dark /l/ are expected to exhibit the highest degree of tonguedorsum coarticulatory resistance and therefore to block
vowel coarticulatory effects in F2 frequency to a large extent. This should be so because the two consonants impose
strict demands on the tongue dorsum: active tongue-body
raising and fronting toward the alveolopalatal zone in the
case of /ɲ/ and some active tongue-predorsum lowering
and postdorsum retraction, in addition to the formation of
an apicoalveolar contact, in the case of dark /l/ (Proctor,
2009). Moreover, vowel coarticulation is expected to be
most prominent when the vowel is antagonistic to the consonant and thus makes the lingual constriction for the
latter especially hard to achieve, namely in the case of the
effects from /a/ on /ɲ/ (the tongue-dorsum position is low
and back for /a/ and high and front for /ɲ/) and from /i/
on dark /l/ (the tongue dorsum occupies a high and front
position for /i/ and a low and back position for dark /l/).
The approximant /δ/ and clear /l/, however, are
less constrained because the tongue body is not actively
controlled during their production and therefore ought
to allow more vowel coarticulation than /ɲ/ and dark /l/.
In addition, some more predorsum lowering for the passage
of airflow through the sides of the oral cavity for clear /l/
versus /δ/ should cause the alveolar lateral to be somewhat
more constrained and thus more resistant to vowel coarticulation than the dental approximant (on dental consonants involving possibly some front dorsum lowering as
well, see Dart, 1991). In this case, tongue-dorsum and F2
raising effects from /i/ are expected to be more prominent
than tongue-dorsum and F2 lowering effects from /a/ because, as pointed out in the following, the former vowel
exhibits a higher degree of tongue-body constraint than the
latter.
The prominence of the vowel-dependent effects in
VCV sequences ought to vary as a function of the transconsonantal vowel as well (see Table 1, bottom, second and
third columns). As reported for other languages (Cho, 2004;
Mok, 2011), these effects should be smaller when the fixed
vowel is /i/ (most resistant) than when it is /a/ (least resistant). The rationale for this outcome is that, as predicted
for /ɲ/, tongue-dorsum raising and fronting ought to render
/i/ more constrained than /a/. In addition, effects on vowels
ought to be more prominent if they are exerted by highly
constrained consonants than by consonants that are less
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
Table 1. Predicted vowel-consonant-vowel coarticulation scenarios for different conditions of intervocalic consonant and fixed vowel.
Segmental condition
Intervocalic consonant
/ /
dark /l/
clear /l/
/ɲ/
Fixed vowel
/i/
/a/
Coarticulatory effects
Coarticulatory
resistance
Vowel
Minimal
Maximal
Intermediate
Maximal
/i/ > /a/
/i/ > /a/
/i/ > /a/
/a/ > /i/
Maximal
Minimal
Coarticulatory direction
Consonant
Consonant-to-vowel
Vowel-to-consonant/vowel
—
Ant > carr
—
Carr > ant
Carr > ant
Ant > carr
Ant ≥ carr
Carr > ant
dark /l/ > / /, clear /l/, and /ɲ/
/ɲ/ > / / and clear and dark /l/
Note. Carr = carryover; ant = anticipation; — = unavailable data.
is higher at closure offset and during V2 than at closure onset and during V1 (Recasens, 2007). As for
the vowel effects, tongue-dorsum and F2 lowering
exerted primarily by /a/ ought to be more prominent
at the carryover than at the anticipatory level, given
that the (more prominent) C-to-V carryover effects
should interfere with vowel anticipation to a larger
extent than the (less prominent) C-to-V anticipatory
effects interfere with vowel carryover.
However, dark /l/ is expected to exert more C-to-V
anticipation than C-to-V carryover (mostly on
antagonistic /i/) because the lowering and retraction
of the tongue body occurs in anticipation of the
tongue front raising gesture for this consonant (Sproat
& Fujimura, 1993). Regarding vowel coarticulation,
the tongue-dorsum and F2 raising effects exerted
mostly by /i/ ought to operate at the anticipatory rather
than the carryover level because now the C-to-V
coarticulatory effects are mostly anticipatory and
therefore should interfere with V1 rather than with V2.
constrained, in particular if the vowel and consonant do
not share the same or a similar articulatory configuration.
Consequently, effects on /i/ ought to be exerted mostly by
dark /l/ and those on /a/ by /ɲ/, whereas clear /l/ and /δ/ ought
to cause less C-to-V coarticulation to occur.
In sum, I predict that vowel coarticulation ought to
vary inversely with the degree of tongue-body constraint
for the contextual phonetic segments as follows: for the
intervocalic consonant in the progression /δ/ > clear /l/ >
dark /l/ and /ɲ/ and for the transconsonantal vowel in the
progression /a/ > /i/.
Coarticulatory Direction
The DAC model of coarticulation also makes specific predictions about whether vowel effects in VCV sequences should be anticipatory rather than carryover
or vice versa (Recasens, 2007; Recasens et al., 1997). According to the model, the extent to which vowel effects
occur along one direction rather than the other depends on
the directionality pattern of the C-to-V effects. Thus, for
example, if the consonant exerts more carryover coarticulation
on V2 than anticipation on V1, one would expect the vowel
effects on the consonant and the transconsonantal vowel
to be more prominent at the carryover than at the anticipatory level as well. This should be so because, in this particular case, there will be more blocking of the V2-dependent
anticipatory effects (by C-to-V2 carryover) than of the
V1-dependent carryover effects (by C-to-V1 anticipation).
As a consequence, one must know whether effects from
consonants on vowels favor the anticipatory or the carryover
direction in order to predict which one of the two directions
of vowel coarticulation will be favored in a given VCV
sequence. As summarized in Table 1 (top, fourth and fifth
columns), the following predictions may be made regarding
the directionality of the consonant and vowel coarticulatory effects in VCV sequences with (a) highly constrained
/ɲ/ and dark /l/ and (b) less constrained /δ/ and clear /l/:
1.
Data on C-to-V coarticulation reveal that /ɲ/ should
exert carryover rather than anticipatory effects on
vowels (mostly on antagonistic /a/) in line with the
fact that the tongue-dorsum activity for the consonant
2.
It is hard to make predictions about the direction of
the vowel coarticulatory effects on the basis of the
direction of the consonant-dependent effects in the
case of VCV sequences with /δ/ and clear /l/, given
that the production of the two dentoalveolars involves
no active tongue-dorsum motion associated with the
formation of a closure or constriction. In principle,
and in parallel to effects exerted by other alveolopalatal segments such as /ɲ/, effects from /i/ rather than
from /a/ should be greater at the carryover than at
the anticipatory level. Moreover, these vowel effects
ought to take place in VCV sequences with /δ/ rather
than in those with clear /l/ because the need to lower
the tongue predorsum somewhat in order to form
lateral channels for the passage of airflow during the
preceding vowel should already render the consonant
and vowel anticipatory components especially salient
(similarly to dark /l/).
In sum, VCV sequences ought to show less vowel
coarticulation for /ɲ/ and dark /l/ than for /δ/ and clear /l/
in view of differences in degree of tongue-body involvement
during consonant articulation. There ought to be clear-cut
Recasens: Effect of Stress and Rate on Coarticulation
1409
Complimentary Author PDF: Not for Broad Dissemination
differences in vowel coarticulatory direction for the highly
constrained consonants /ɲ/ and dark /l/—that is, vowel carryover should exceed vowel anticipation in VCV sequences
with the nasal consonant, and the reverse should occur in
VCV sequences with the alveolar lateral. Regarding VCV
sequences with the less constrained consonants /δ/ and clear
/l/, vowel carryover effects as a function of /i/ are expected
to prevail over vowel anticipatory effects for the former
consonant but not for the latter.
Stress and Speech Rate
In the present investigation, the predictions about
degree and direction of vowel coarticulation in VCV sequences formulated by the DAC model were evaluated
in Catalan VCV sequences bearing stress on V1 (′VCV)
and on V2 (V′CV) and produced at normal and fast speaking rates.
Regarding coarticulation degree, vowel coarticulation effects induced by the presence versus absence of stress
and by different speech rates were expected to be smaller
for highly constrained consonants than for less constrained
ones, the rationale being that the former ought to be
more resistant to variations in closure or constriction
degree than the latter. Thus, in comparison to /δ/ and to a
large extent clear /l/, consonants subject to stricter articulatory requirements such as /ɲ/ and dark /l/ should be
more resistant to different degrees of coarticulation that
may arise when stress position and speech rate are subject
to change.
As for the effect of stress on coarticulatory direction,
American English data for VCV sequences with /b/, /d/,
/g/, and several vowels indicate that V-to-V effects from
stressed vowels onto unstressed vowels in VCV sequences
occur at the carryover rather than the anticipatory level
(Agwuele, 2004; Hertrich & Ackermann, 1995). Within the
DAC model framework, patterns of vowel coarticulatory
direction in VCV sequences with highly constrained consonants were expected to be enhanced when stress falls on
the vowel (V1 or V2) that triggers more coarticulation. If
so, there ought to be an increase in vowel carryover coarticulation relative to vowel anticipation in VCV sequences
with /ɲ/ when stress falls on V1 (/′VɲV/) compared with
when it falls on the CV2 string (/V′ɲV/), and in vowel anticipation relative to vowel carryover in VCV sequences
with dark /l/ when stress falls on CV2 (/V′lV/) compared
with when it falls on V1 (/′VlV/). As for the more unconstrained dentoalveolars /δ/ and clear /l/, changes in stress
position could yield more vowel carryover in /′VδV/ versus
/V′δV/ sequences and more vowel anticipation or no preferred direction in /V′lV/ versus /′VlV/ sequences. Regarding
the role of speech rate, the articulatory gesture for vowels
and consonants could be anticipated to a larger extent as
speakers speak faster, which should result in an increase of
anticipation relative to carryover for both the consonantand the vowel-dependent effects in VCV sequences in general. This should be so because articulatory gestures for
speech segments are typically anticipated in time, and this
1410
anticipatory activity is expected to become more prominent
as undershoot causes the mechanico-inertial characteristics
of the articulators to diminish at faster speech rates.
In sum, if coarticulatory direction is affected by
changes in stress and speaking rate, the presence of stress
in the crucial changing vowel ought to cause an increase in
the degree of vowel anticipation vis-à-vis vowel carryover
in VCV sequences with dark /l/ and the reverse for /ɲ/ and
an increase in the preferred vowel coarticulatory direction
in VCV sequences with /δ/ and clear /l/ (presumably in
vowel carryover for the former consonant but not for the
latter). However, there should be more consonant and
vowel anticipation at faster versus slower speaking rates
as a general rule.
Summary of Testing Hypotheses
The present study predicts that the degree and direction of coarticulation effects as a function of /i/ versus /a/
in VCV sequences ought to vary with the degree of tonguebody constraint for the intervocalic consonant and the
fixed vowel. Vowel coarticulation should decrease in
the progression /δ/ > /l/ > /ɲ/ (and for dark vs. clearer
varieties of /l/) and for fixed /a/ more than for fixed /i/.
Coarticulatory direction is expected to favor vowel carryover in /VɲV/ sequences and vowel anticipation in /VlV/
sequences with dark /l/ and perhaps carryover effects from
/i/ in /VδV/ sequences, whereas effects for /VlV/ sequences
with clear /l/ could favor vowel anticipation or neither
direction. There should also be more vowel coarticulation
from stressed versus unstressed vowels and in VCV sequences
produced at fast versus normal speaking rates, and this
increase ought to occur mostly when the intervocalic consonant is relatively unconstrained. Moreover, the preferred
vowel coarticulatory direction could be enhanced when
stress falls on the crucial changing vowel (e.g., on V1 in
/VɲV/ sequences), and vowel anticipation ought to gain
weight at fast versus normal speech rates irrespective of the
segmental conditions involved.
Method
Speech Materials
Acoustic recordings were made of 10 tokens of the
symmetrical and asymmetrical sequences /aCa/, /iCi/, /iCa/,
and /aCi/, with the consonants /δ/, /ɲ/, and /l/ and the
two stress conditions /′VCV/ and /V′CV/, where stress falls
either on the first syllable and thus on V1 or else on the
second syllable and thus on the CV2 string. These VCV sequences allow the study of vowel coarticulation over time
in all symmetrical and asymmetrical combinations with
changing and fixed /i/ and /a/.
VCV sequences were embedded in an [əβ__βə] string
and thus were flanked by a voiced labial consonant (which
happens to be an approximant if occurring intervocalically
in Catalan), in order to maximize possible coarticulatory
effects in lingual activity and thus in F2 frequency that
may take place during the transconsonantal vowel. The
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
nonsense [əβVCVβə] sequences were uttered in isolation in
order to make sure that differences in vowel duration and
intensity between the stressed and unstressed syllables in
the /′VCV/ and /V′CV/ conditions would not be canceled,
at least in part, in the case that one or more words with
specific stress levels were placed before and after the target
nonsense word.
Speakers and Recording Procedure
Five 40- to 60-year-old Catalan speakers (DR, AL,
DP, CE, and LU) read the speech material 10 times at
their normal speech rate and subsequently another 10 times
at a fast speech rate. All five speakers are university or
high school teachers and speak Catalan on a daily basis.
Overall, 2,400 sequence tokens were recorded and submitted to analysis (3 Cs × 4 VCV combinations × 2 stress
conditions × 2 speech rates × 10 tokens × 5 speakers). The
small number of speakers renders this study a preliminary
investigation about the effect of stress and speech rate on
vowel coarticulation in VCV sequences, though this limitation is offset by the large number of data points that were
subject to measurement.
Audio recordings were carried out at a 22050-Hz
sampling rate in a soundproof room with a Rode NT1-A
microphone and the CSL4500 (Computerized Speech Lab)
system by Kay Pentax (Montvale, NJ). The acoustic signal
was downsampled to 11025 Hz and analyzed with the CSL
acoustic analysis software.
Acoustic Analysis
Because all the consonants (/δ/, /ɲ/, and /l/) under
study have formant structure, VC and CV segmental
boundaries were determined visually on spectrographic displays on the basis of a change in formant intensity level—
that is, formants are more intense for vowels than for
approximants, laterals, and nasals—and on spectral discontinuity whenever available, as in the case of the /VɲV/
sequences. For each speech file, F2 frequency values were
collected at C midpoint and at several temporal points during the vowel segments separately for V1 and V2: at vowel
onset and offset; at vowel midpoint; at the midpoint between vowel onset and vowel midpoint; and at the midpoint
between vowel midpoint and vowel offset. Frequency measurements were taken visually at these 11 time points on
spectrographic displays by placing a cursor in the middle
of F2. Figure 1 shows a spectrogram in which these points
are coded with numbers 1 through 11. Overall, 26,400 measurements were taken (2,400 sequence tokens × 11 time
points). This analysis method was preferred to the linear
predictive coding peak-picking method, which, according
to results from some pilot work, often yielded unsatisfactory frequency values whenever F2 was not clearly visible
or exhibited a rapid spectral change. A reliability check
was not carried out on the present data sample but was
performed on F2 vowel measurements taken by the same
experimenter in previous studies (Recasens & Espinosa,
2009). Some notion of the measurement precision can be
gathered from the F2 data at the consonant midpoint
reported in Table 2, where coefficients of variation do not
exceed 5% for about 90% of the F2 values and range between 5% and 10% for the remaining 10%. Special care
was taken to place the cursor at the center of the formant
in relatively fast vowel transitions and whenever F2 was
hardly visible.
The sizes of the V2-dependent anticipatory effects
in the sequence pairs /aCa/–/aCi/ and /iCi/–/iCa/ and of
the V1-dependent carryover effects in the sequence pairs
/aCa/–/iCa/ and /iCi/–/aCi/ were obtained by subtracting
F2 for changing /a/ from F2 for changing /i/ at seven points
in time: at the edge of the changing vowel (i.e., at V2 onset
for vowel anticipation and at V1 offset for vowel carryover), at C midpoint, and at five equidistant points during
the transconsonantal vowel (see Figure 1). The reason why
V2- and V1-dependent effects were measured at V2 onset
and V1 offset, respectively, is to be found in the fact that
the F2 frequency at those temporal points is strongly dependent not only on the vowel but on the intervocalic consonant as well. The sign of the difference had to be positive
because F2 is at about 2000 Hz for /i/ and at about 1200–
1500 Hz for /a/; as in previous studies (Recasens, 2002;
Recasens et al., 1997), negative values that occurred occasionally when F2 for /a/ exceeded F2 for /i/ during the consonant and/or the transconsonantal vowel were converted
into 0 for statistical analysis.
Statistical Analysis
F2 differences between /i/ and /a/ were tested statistically with a repeated measures design using linear mixed
models and five independent factors: Fixed Vowel (/a/, /i/),
Consonant (/δ/, /ɲ/, /l/), Stress Condition for the changing
vowel (stressed, unstressed), Coarticulatory Direction
(anticipatory, carryover), and Speech Rate (normal, fast).
Separate tests were run on the F2 differences between
changing /i/ and /a/ at seven equivalent temporal points for
evaluating anticipatory vowel coarticulation (e.g., the difference between V2 = /i/ and V2 = /a/ at V1 offset at Point 5)
and carryover vowel coarticulation (e.g., the difference
between V1 = /i/ and V1 = /a/ at V2 onset at Point 7). The
temporal comparisons subject to statistical testing were as
follows (see Figure 1):
1.
Between V2 anticipatory effects at V2 onset (Point 7)
and V1 carryover effects at V1 offset (Point 5)
2.
Between V2 anticipatory effects and V1 carryover
effects at C midpoint (Point 6)
3.
Between V2 anticipatory effects at V1 offset (Point 5)
and V1 carryover effects at V2 onset (Point 7)
4.
Between V2 anticipatory effects at Point 4 and V1
carryover effects at Point 8
5.
Between V2 anticipatory effects at V1 midpoint
(Point 3) and V1 carryover effects at V2 midpoint
(Point 9)
Recasens: Effect of Stress and Rate on Coarticulation
1411
Complimentary Author PDF: Not for Broad Dissemination
Figure 1. Spectrographic representation of the sequence /ala/, with temporal points at which the F2 measurements were
taken. V1 = first vowel position; C = consonant; V2 = second vowel position.
6.
Between V2 anticipatory effects at Point 2 and V1
carryover effects at Point 10
7.
Between V2 anticipatory effects at V1 onset (Point 1)
and V1 carryover effects at V2 offset (Point 11)
The statistical model included all main effects and
two- and three-way interactions. Bonferroni post hoc tests
were performed on those significant variables that had
more than two levels, and simple effects tests were carried
out on the significant factor interactions. A low threshold
for significance was chosen in order to account for the
application of multiple statistical tests on the temporal
coarticulation data. Because seven comparisons were carried out, the Bonferroni correction requires that the p value
be .007. Although significant effects are reported for
p < .05, only those differences approaching the p < .001
level of significance are considered to be robust. Partial
eta-squared (ηp2) values and thus a measure of the strength
of the main effects and significant interactions will also
be provided.
Additional repeated measures analyses were run on
the data for the symmetrical sequences /iCi/ and /aCa/ in
order to uncover differences in segmental duration and
in F2 frequency triggered by changes in stress and speech
rate. Two tests were performed on vowel duration and on
the F2 frequency values at the vowel midpoint, with Stress
(stressed, unstressed), Position of the stressed vowel (V1,
V2), Vowel Quality (/i/, /a/), and Speaking Rate (normal,
fast) as factors. Another analysis of variance was carried
out on sentence duration, with Speaking Rate as the independent variable.
Results
Coarticulatory Resistance
A first goal of this study was to analyze the extent
to which /δ/, /ɲ/, and /l/ exhibit the predicted degrees of
coarticulatory resistance to vowel effects. In order to investigate this issue, Figure 2 plots the F2 frequency values at
1412
consonant midpoint for the three consonants in the symmetrical sequences /iCi/ and /aCa/ across stress and rate
conditions, speakers, and tokens. According to the figure
(see also Table 2) and in agreement with the initial predictions, coarticulatory resistance is minimal for /δ/, with
F2 occurring at 1900–2000 Hz in the /i/ context and at
1200–1350 in the /a/ context. Except for speaker CE, the
alveolopalatal /ɲ/, whose F2 frequency is about 2000 Hz
or higher, is slightly more resistant than /l/—that is, the
F2 difference between /i/ and /a/ is somewhat smaller for
the nasal than for the lateral.
Differences in vowel coarticulation between /l/ and
/ɲ/ may be related to variations in /l/ darkness degree
among speakers. Darkness degree may be evaluated in the
sequence /ili/, where the lingual gestures for the vowel and
the consonant are more antagonistic if /l/ is dark than if
it is clear. Thus, although /i/ is articulated with a high front
tongue-body position, dark /l/—and much less so clear /l/
—is produced with a low predorsal and back postdorsal
configuration. These articulatory characteristics ought to
cause F2 to be about 2000 Hz for /i/, 1000–1300 Hz for
dark /l/, and 1400–2000 Hz for clear /l/ and cause dark /l/
to be more resistant than clear /l/ to coarticulatory effects
in tongue-body position and F2 frequency associated
with the high front vowel. Data for /l/ in the sequence /ili/
plotted in Figure 2 show that darkness degree is higher
for speakers DR, AL, and DP than for CE and LU (F2
for /l/ is at about 1250–1350 Hz for the three former subjects
and about 1550 Hz for the two latter ones). Moreover, the
expected positive relationship between coarticulatory resistance and darkness degree—that is, darker variants of /l/
should be more resistant to F2 differences from /i/ to /a/ than
clearer variants of the consonant—holds only to some
extent. As expected, coarticulatory resistance is higher for
speakers DR, AL, and DP than for speaker LU (the F2
difference between /ili/ and /ala/ amounts to 200–350 Hz for
the three former subjects and to 433 Hz for the latter). In
disagreement with the expected positive relationship between
darkness degree and coarticulatory resistance, however, /l/
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Table 2. Mean F2 values (standard deviations) at consonant midpoint in symmetrical /VðV/, /VɲV/, and /VlV/ sequences.
Speech
rate
Normal
Stressed
vowel
position Vowel
V1
/i/
/a/
V2
/i/
/a/
Fast
V1
/i/
Recasens: Effect of Stress and Rate on Coarticulation
/a/
V2
/i/
/a/
Speaker DR
/ð/
/ɲ/
Speaker AL
/l/
/ð/
/ɲ/
2011.7
2192.4
1271.2
1812.1
2310.8
(60.37)
(15.23)
(51.10) (202.74)a (30.62)
1284.9
2034.6
1066.6
1325.7
2162.8
(32.46)
(83.32)
(12.80)
(62.86)
(60.16)b
2222.2
1324.2
1797.8
2327.2
1956.5
(34.78) (101.57) (96.94)a
(28.99)
(66.57)b
1286.3
1963.5
1074.9
1294.5
2030.3
(25.97)
(27.75)
(28.44)
(40.65) (142.48)d
2207.9
1316.1
1915.6
2321.8
2012.2
(27.24) (119.61) (163.81)a (44.42)
(61.80)a
1069.8
1333.4
2129.2
1320.5
2049.1
(20.49)
(49.12)
(54.92)a
(34.03)
(69.15)a
2227.3
1342.9
2044.8
2300.1
2066.9
(114.16)c (47.72) (104.64) (102.20)a (30.13)
1072.5
1343.2
2171.2
1286.3
1960.6
(36.33) (167.24)e
(30.90) (103.31)b (16.01)
Speaker DP
Speaker CE
/l/
/ð/
/ɲ/
/l/
/ð/
/ɲ/
/l/
1250.3
(35.98)
1070.1
(26.56)
1252.0
(45.97)
1038.0
(37.40)
1276.7
(44.85)
1074.2
(38.21)
1254.8
(46.73)
1061.9
(49.12)
1973.1
(31.70)a
1254.2
(32.11)
1913.4
(92.17)a
1198.2
(17.04)
1954.7
(84.89)d
1307.4
(35.93)
1943.6
(71.77)
1232.3
(39.51)
2039.3
(17.37)
1947.1
(28.15)
1980.4
(32.51)
1897.5
(57.86)
2076.5
(45.10)
1988.0
(42.76)
2005.3
(29.95)
1881.2
(68.84)
1299.6
(40.57)
1007.9
(25.11)
1278.7
(46.39)
976.0
(35.55)
1362.2
(35.20)
1000.1
(17.90)
1350.4
(46.99)
1007.9
(32.15)
1917.3
(65.05)
1335.4
(30.59)
1941.7
(66.24)
1391.3
(51.71)
1921.0
(49.39)
1312.8
(74.16)
1962.3
(47.39)
1379.7
(57.25)
2173.4
(31.69)
1834.5
(93.56)
2120.0
(44.79)
1828.0
(90.63)
2170.6
(21.93)
1821.9
(77.72)
2088.8
(55.20)
1828.6
(64.39)
1568.9
(36.04)
1355.3
(53.16)
1542.2
(25.31)
1368.8
(58.85)
1554.8
(43.17)
1374.6
(51.35)
1549.0
(24.20)
1387.8
(66.87)
Note. Number of repetitions for each sequence is 10 unless otherwise noted. V1 = first vowel position; V2 = second vowel position.
a
Speaker LU
Number of repetitions = 9. bNumber of repetitions = 8. cNumber of repetitions = 7. dNumber of repetitions = 6. eNumber of repetitions = 5.
/ð/
/ɲ/
1948.3
2074.6
(72.52)b (83.28)
1269.0
1969.0
(107.86) (66.45)
1997.6
2171.4
(117.35)b (50.61)
1186.8
1993.5
(64.42)
(50.09)
1875.9
2109.2
(122.26) (41.74)
1157.5
1707.2
(37.12) (74.27)a
1900.2
2140.7
(96.17)a (27.79)
1211.9
1812.7
(38.16)
(57.09)
/l/
1534.5
(32.05)
1101.2
(59.43)
1524.9
(43.25)
1102.4
(45.96)
1560.2
(49.07)
1075.0
(52.88)
1492.1
(90.17)
1100.0
(39.79)
1413
Complimentary Author PDF: Not for Broad Dissemination
Figure 2. Mean F2 frequency values at the midpoint of /δ/, /ɲ/, and
/l/ in symmetrical vowel–consonant–vowel sequences with /i/ and
/a / across stress and rate conditions, plotted as a function of
consonant, vowel context, and speaker.
exhibits comparable degrees of vowel coarticulation, or
even less, in the case of speaker CE (182 Hz) than of speakers
DR, AL, and DP (200–350 Hz). As the figure shows, this
unexpected result appears to be related to the presence of
a higher F2 for /l/ next to /a/ for this subject than for the
other four speakers, presumably meaning that the tongue
body is not pulled down much by the low vowel during the
production of the alveolar lateral.
As a complement to these data on consonant coarticulatory resistance, Figure 3 presents the F2 trajectories
for the symmetrical sequences /iCi/ and /aCa/ with all three
consonants and stress falling on V1 and on V2 (normal
speaking rate only). Less V-to-C coarticulation from /i/ on
/l/ than on /δ/ at the midpoint of the consonant is consistent
with the alveolar lateral exhibiting a much lower F2 target
than the dental approximant in the sequence /iCi/; moreover, as pointed out already, a shorter-than-expected F2 distance between /ili/ and /ala/ for speaker CE is related to
the entire F2 trajectory for /ala/ occurring at a higher frequency for this speaker than for the other four. As for the
VCV sequences with /ɲ/, the alveolopalatal nasal exhibits
an F2 peak when flanked by /a/ (/aɲa/), which comes close
to the F2 frequency for the consonant in the context of
/i/ (/iɲi/); the peak in question may occur at about closure
midpoint (for speakers AL and DP) or toward closure
offset (for speakers DR, CE, and LU), depending presumably on the point in time when maximal linguopalatal contact takes place. An aspect worth noting is the extremely
high F2 frequency for the consonant and the two vowels
in speaker AL’s productions of the sequence /iɲi/, which
suggests that articulatory overshoot—that is, a considerable F2 increase whenever the VCV sequence is composed
of several alveolopalatal segments in succession—is at
work.
Regarding the effect of consonants on vowels, Figure 3 also shows that F2 frequency changes associated with
the consonant have a relatively restricted temporal span
in the sequence /ili/ (i.e., F2 lowering starts at about V1
1414
midpoint and F2 raising ends at V2 midpoint) and extend
over a longer period of time in the sequence /aɲa/ (i.e., F2
rising starts at V1 onset and F2 lowering ends at V2 offset). Moreover, regarding the latter VCV sequence, the F2
lowering trajectory during V2 = /a/ often proceeds more
slowly than the F2 rising trajectory during V1 = /a/, which
results in higher F2 frequency values at comparable temporal points during V2 versus V1 and thus more prominent
C-to-V carryover than anticipatory effects (compare, e.g.,
the F2 frequency value at V2 midpoint with that at V1
midpoint in the sequence /aɲa/ for all speakers).
Differences in the temporal extent of vowel coarticulation among VCV sequences with /δ/, /ɲ/, and /l/ are
consistent with the scenario at the midpoint of the consonant just described. As revealed by the statistical results
for the main consonant effect presented in Table 3 (left
column), differences in vowel coarticulation among the
three intervocalic consonants were found to occur at consonant midpoint (Temporal Comparison 2) and extended
until the midpoint of the fixed transconsonantal vowel
(Temporal Comparison 5) or even further. Moreover, the
F2 difference between changing /i/ and /a/ over the time
domain turned out to be greater when the intervocalic
consonant was /δ/ than when it was /l/ or /ɲ/ and generally for the lateral than for the nasal. This consonantspecific behavior may be inferred from the height of the
bars in Figures 4–6 displaying the size of the vowel effects in VCV sequences with each of the three consonants
at all temporal points.
As for differences in coarticulatory resistance between the two fixed vowels, Table 3 (middle and right
columns) also reports a main fixed vowel effect and a significant Consonant × Fixed Vowel interaction lasting until
about the vowel midpoint (Temporal Comparison 5). As
revealed by Figures 4–6, this finding is related to the existence of more vowel coarticulation during fixed /a/ in VCV
sequences with /δ/ and /ɲ/ and the reverse (i.e., more coarticulation during fixed /i/ than during fixed /a/) in VCV
sequences with /l/. Indeed, for the temporal comparisons
yielding a significant effect, bars are higher for the fixed /a/
condition than for the fixed /i/ condition in VCV sequences
with /δ/ (see Figure 4) and to a lesser extent in VCV sequences with /ɲ/ as well (see Figure 5) and for fixed /i/
than for fixed /a/ in VCV sequences with /l/ (see Figure 6).
These results are only in partial agreement with the initial
hypothesis that /a/ should be more sensitive to coarticulation
effects than /i/ and are interpreted in terms of gestural affinity and antagonism for consonants and vowels in the
Discussion section.
Coarticulatory Direction
As shown in Table 4, differences in coarticulatory
direction—that is, whether a higher F2 for changing /i/ versus /a/ extends more toward the right (carryover) or toward
the left (anticipatory)—may also reach significance during
the consonant and the transconsonantal vowel. In particular,
Consonant and Fixed Vowel interact significantly with
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
Figure 3. F2 trajectories for symmetrical /′VCV/ sequences (solid lines) and /V′CV/ sequences (dashed lines) with /δ/, /ɲ/, and /l/ and /i/ and /a/, for
all speakers. Data correspond to the normal speaking-rate condition. V1 = first vowel position; C = consonant; V2 = second vowel position.
Recasens: Effect of Stress and Rate on Coarticulation
1415
Complimentary Author PDF: Not for Broad Dissemination
Table 3. Significant Consonant and Fixed Vowel effects for F2 differences between changing /i/ and /a/ over time.
Consonant
(df = 2, 199)
Temporal Comparison
1. V2 on/V1 off
2. C midpoint
3. V1 off/V2 on
4.
5. V1 midpoint/V2 midpoint
6.
7. V1 on/V2 off
Fixed Vowel
(df = 1, 199)
Consonant × Fixed Vowel
(df = 2, 199)
F
ηp2
F
ηp2
F
ηp2
137.4***
193.8***
86.3***
77.3***
15.9***
.57
.66
.46
.43
.13
14.2***
84.5***
25.5***
52.9***
16.2***
.07
.29
.11
.20
.07
15.9***
7.2**
40.4***
61.8***
10.7***
.14
.07
.28
.37
.09
.04
6.4*
.03
4.6**
Note. df = degrees of freedom; V1 = first vowel position; V2 = second vowel position; C = consonant.
*p < .05. **p < .01. ***p < .001.
Direction in ways that are interpreted next with reference
to the coarticulation data plotted in Figures 4–6.
As shown in Figure 4, vowel effects in VCV sequences
with /δ/ are generally larger at the carryover level in the
fixed /a/ context (effects from /i/ on /δ/ and on fixed /a/)
and favor no specific direction in the fixed /i/ condition (effects from /a/ on /δ/ and on fixed /i/). Indeed, the four right
bars in each graph are higher than the four left bars for
most temporal comparisons when the fixed vowel is /a/,
whereas there is no clear prevalence of the four right bars
over the four left bars or vice versa in the fixed /i/ context.
As shown by the F2 trajectories for /iδa/ and /aδi/ for a
representative group of speakers in Figure 7, the presence
of more carryover than anticipatory effects from /i/ is consistent with F2 being located at a higher frequency at C
offset in the sequence /iδa/ than at C onset in the sequence
/aδi/, and also during the first portion of V2 in the former
sequence than during the last portion of V1 in the latter,
for all three speakers. There may be two (perhaps concomitant) reasons for this finding: The mechanico-inertial constraints involved in the tongue raising and fronting gesture
for the high front vowel may explain why effects from /i/
favor the carryover direction, and V1 = /a/ may cause the
tongue dorsum to stay relatively low during /δ/, thus blocking to a greater or lesser extent the tongue-dorsum raising
and fronting anticipatory effects exerted by V2 = /i/.
As revealed by the graphs in Figure 5, F2 differences
between changing /i/ and /a/ are greater at the carryover
than the anticipatory level in VCV sequences with /ɲ/ and
the two fixed vowels /i/ and /a/ mostly in Temporal Comparisons 1, 2, and 3. The vowel carryover effects also exceeded
the vowel anticipatory effects in Comparisons 4–7 during
fixed /i/, though this directionality difference appears to be
associated with speaker AL only, presumably because he
overshot the vowel (/i/) when it was preceded by another
segment involving much dorsopalatal contact (/ɲ/). Inspection of the F2 trajectories for the two asymmetrical VCV
sequences /iɲa/ and /aɲi/ plotted in Figure 8 may shed
some light on why effects from /a/ on the consonant and
fixed /i/ are greater at the carryover versus the anticipatory
level. Analogous to the scenario for /δ/, data for all speakers
indicate the presence of a much higher F2 frequency at C
offset and during the first half of V2 = /a/ in the sequence
/iɲa/ than at C onset and during the second half of V1 = /a/
Figure 4. Cross-speaker F2 differences for changing /i/ versus /a/ at seven temporal points in /VδV/ sequences with fixed /i/ (top) and fixed
/a / (bottom). In each graph, the four left bars correspond to the V2 anticipatory effects (Ant) and the four right ones to the V1 carryover effects
(Carr). Within each bundle of four bars, four stress/rate conditions may be identified: STn (stressed changing vowel, normal speech rate),
USTn (unstressed, normal), STf (stressed, fast), and USTf (unstressed, fast). V1 = first vowel position; V2 = second vowel position.
1416
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
Figure 5. Cross-speaker F2 differences for changing /i/ versus /a/ at seven temporal points in /VɲV/ sequences with fixed /i/ (top) and fixed
/a/ (bottom). See Figure 4 caption for details.
in the sequence /aɲi/, which follows from the presence of
more tongue-dorsum raising toward the palate at closure
offset than at closure onset. This may explain why tonguebody lowering effects exerted by /a/ are blocked at the
offset of the consonant rather than at its onset and, consequently, why carryover effects from V1 = /a/ are larger
than anticipatory effects from V2 = /a/.
Patterns of coarticulatory direction for sequences
with /l/ are to a large extent determined by whether the
consonant is darker or clearer. Speakers AL and DP, with
strongly dark variants of the consonant, favored vowel
anticipation in the fixed /i/ condition up to temporal comparison 4 (see Figure 9, top graphs) and in the fixed /a/
context mostly for Comparisons 1 and 2 (see Figure 9,
bottom graphs). Coarticulation data across speakers presented in Figure 6 also reveal the prevalence of vowel anticipation for Comparisons 2–4 in the context of /i/ and
for Comparisons 1 and 2—but not at temporal points situated further away from the changing vowel—in the context
of /a/. F2 trajectories for the sequences /ila/ and /ali/ produced by speakers AL and DP allow a determination in
some detail of why effects from /i/ on dark /l/ and fixed /a/
are greater at the anticipatory versus the carryover level (see
Figure 10). According to the figure, the two speakers show
a lower F2 for /ila/ than for /ali/ at C midpoint and at
equivalent temporal points during the vowels—for example, at C onset and during V1 for /ila/ compared with at C
offset and during V2 for /ali/. This finding should be attributed to a considerable blocking of the carryover effects
from V1 = /i/ by the tongue lowering/backing anticipatory
activity for dark /l/; this becomes especially salient whenever the consonant is followed by a low vowel because both
dark /l/ and /a/ share an analogous low/back tongue-dorsum
configuration. Speakers CE and LU, with clearer variants
of /l/, as well as speaker DR, favor no major pattern of
vowel coarticulatory direction, perhaps because, at least for
the former subjects, anticipatory effects for the consonant
counteract the carryover action of V1 = /i/ to some extent,
though less so than when the alveolar lateral is dark.
Stress and Speech Rate
Sentence duration was affected significantly by Speech
Rate, F(1, 114) = 185.29, p < .001, ηp2 = .67. As shown in
Figure 6. Cross-speaker F2 differences for changing /i/ versus /a/ at seven temporal points in /VlV/ sequences with fixed /i/ (top) and fixed /a/
(bottom). See Figure 4 caption for details.
Recasens: Effect of Stress and Rate on Coarticulation
1417
Complimentary Author PDF: Not for Broad Dissemination
Table 4. Significant Coarticulatory Direction effects for F2 differences between changing /i/ and /a/ over time.
Direction
(df = 1, 199)
Temporal Comparison
1. V2 on/V1 off
2. C midpoint
3. V1 off/V2 on
4.
5. V1 midpoint/V2 midpoint
6.
7. V1 on/V2 off
Direction ×
Consonant
(df = 2, 199)
Direction ×
Fixed Vowel
(df = 1, 199)
F
ηp2
F
ηp2
F
ηp2
5.3*
11.9**
.03
.06
13.5***
8.7***
6.1**
4.8**
4.3*
.12
.08
.06
.04
.04
9.8**
.05
20.3***
6.8**
.09
.03
6.3*
.03
5.9*
.03
Direction × Consonant ×
Fixed Vowel
(df = 2, 199)
F
ηp2
4.2*
5.7**
11.1***
4.4*
.04
.05
.10
.04
Note. df = degrees of freedom; V1 = first vowel position; V2 = second vowel position; C = consonant.
*p < .05. **p < .01. ***p < .001.
Figure 11 (top), sentences were shorter when produced at
fast versus normal speech rates, and shortening degree was
speaker dependent—that is, maximal for speakers LU and
DR and minimal for speaker DP.
Vowel duration yielded a main effect of Stress,
F(1, 221) = 94.75, p < .001, ηp2 = .30; Speech Rate, F(1, 221) =
21.25, p < .001, ηp2 = .09; and Vowel Quality, F(1, 221) =
24.49, p < .001, ηp2 = .10, but not of Vowel Position, and
Figure 7. F2 trajectories for the sequences /iδa/ and /aδi/ for speakers DR, DP, and CE (stressed V1, normal speaking-rate condition only).
V1 = first vowel position; C = consonant; V2 = second vowel position.
1418
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
Figure 8. F2 trajectories for the sequences /iɲa/ and /aɲi/ for speakers DR, DP, and CE (stressed V1, normal speaking-rate condition only).
V1 = first vowel position; C = consonant; V2 = second vowel position.
no significant interactions. It was greater for /a/ than for /i/
and, as shown by Figure 11 (bottom graph), it varied with
stress and speech rate in the progression stressed/normal >
stressed/fast > unstressed/normal > unstressed/fast for
most speakers and therefore as a function of changes in
stress position rather than of changes in speech rate.
The vowel F2 frequency also varied significantly as
a function of Stress, F(1, 221) = 4.06, p < .05, ηp2 = .02;
Vowel Position, F(1, 221) = 21.91, p < .001, ηp2 = .09; and
Vowel Quality, F(1, 221) = 2,696.38, p < .001, ηp2 = .92,
but not Speech Rate. As shown by the F2 trajectories for
/iCi/ and /aCa/ displayed in Figure 3, F2 was slightly higher
for stressed vowels than for unstressed ones for essentially
all speakers (in the figure, continuous lines correspond to
the stressed V1 condition and discontinuous lines to the
stressed V2 condition). F2 frequencies were also higher for
vowels placed in the V2 versus the V1 position and for /i/
versus /a/. Moreover, vowel duration and F2 frequency
turned out to be moderately correlated across Stress,
Speech Rate, Consonant, and Vowel Position and speaker
conditions—that is, F2 increased with vowel duration
mostly for /i/, r = .420, p < .001, but also for /a/, r = .252,
p < .01.
As shown in Table 5, Stress also had an effect on
the degree of vowel coarticulation according to Temporal
Comparisons 4–7, whereas the effect of Speech Rate was
barely significant and held for Temporal Comparison 4,
for the most part. Figures 4–6 illustrate graphically the
stress- and rate-dependent differences in vowel coarticulation occurring during VCV sequences composed of /δ/, /ɲ/,
and /l/ and fixed /i/ and /a/. Regarding stress, the size of
the vowel coarticulation effects appears in dark bars for
triggering stressed vowels and in gray bars for triggering unstressed vowels. For VCV sequences with all three consonants, results for Temporal Comparisons 4–7 reveal a
trend for the degree of coarticulation to increase when
the triggering vowel is stressed (and thus the target vowel
is unstressed) than when the triggering vowel is unstressed
(and the target vowel is stressed). A relevant finding is that
the effect of Stress on vowel coarticulation operates during the transconsonantal vowel but not for Comparison 2
and thus during the consonant (see also the Discussion).
Recasens: Effect of Stress and Rate on Coarticulation
1419
Complimentary Author PDF: Not for Broad Dissemination
Figure 9. F2 differences for changing /i/ versus /a/ at temporal points 1–4 in /VlV/ sequences with fixed /i/ (top) and fixed /a/ (bottom) for
speakers AL and DP. See Figure 4 caption for details.
The effect of Speech Rate on vowel coarticulation may
be traced by comparing the two left bars and the two
right bars within each bundle of four bars in each graph
of Figures 4–6 (i.e., the two left bars correspond to the
normal speech rate condition and the two right bars to
the fast speech rate condition). The figures reveal the
presence of slightly more vowel coarticulation for fast
versus normal speech rates, mostly in the case of Temporal
Comparisons 2–5.
A relevant finding was the lack of significant Stress ×
Consonant, Speech Rate × Consonant, Stress × Coarticulatory Direction, and Speech Rate × Coarticulatory Direction interactions (Speech Rate and Stress did not interact
with each other either). Therefore, an increase in coarticulation associated with the presence of a stressed versus
1420
unstressed changing vowel and an increase in speech rate
appears to take place independent of the intervocalic consonant and coarticulatory direction—namely, of whether
the consonant is /δ/, /ɲ/, or /l/ and whether vowel effects
are mostly carryover or anticipatory, depending on the segmental composition of the VCV sequence. Indeed, data
presented in Figures 4–6 show a similar increase in coarticulation as a function of stressed versus unstressed vowels
in the anticipatory and carryover directions for VCV
sequences with all three consonants. Consistent with this
finding, F2 trajectories for the asymmetrical sequences
/aCi/ and /iCa/ such as those plotted in Figures 7, 8, and
10 exhibit essentially the same temporal patterns independent
of whether they are produced in the stressed/normal,
unstressed/normal, stressed/fast, or unstressed/fast conditions.
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
Figure 10. F2 trajectories for the sequences /ila/ and /ali/ for speakers AL and DP (stressed V1, normal speaking-rate condition only). V1 =
first vowel position; C = consonant; V2 = second vowel position.
Discussion
The first hypothesis tested in the present study was
whether coarticulatory resistance varied with the degree
of lingual constraint involved in the production of the
intervocalic consonant and the transconsonantal vowel
(Recasens et al., 1997). VCV coarticulation data reported
in the Results turned out to be in general agreement with
this prediction regarding consonants: Vowel coarticulatory
Figure 11. (Top) Effect of changes in speech rate on sentence
duration across consonant, vowel quality, and stress position for
each speaker. (Bottom) Effect of changes in stress position and
speech rate on vowel duration in symmetrical vowel–consonant–
vowel sequences across consonant and vowel quality for each
speaker. Error bars correspond to one standard deviation.
effects from /i/ versus /a/ were greater for /δ/ than for /l/
and /ɲ/, and these consonant-dependent differences in
vowel coarticulation extended essentially until about the
midpoint of the fixed vowel. Moreover, the alveolopalatal
nasal turned out to be somewhat more resistant than the
alveolar lateral during the consonant and the fixed vowel.
Coarticulatory resistance for /l/ was related to speakerdependent differences in darkness degree to a large extent:
As expected, /l/ was highly coarticulation resistant for
the three speakers showing a dark consonantal variety;
among subjects showing a clearer variety of the consonant,
coarticulatory resistance was relatively high for one speaker
and as low as for dark /l/ in the case of the other subject.
Also as predicted, vowel effects in VCV sequences
with /δ/ and /ɲ/ were greater when the fixed vowel was less
constrained (/a/) than when it was more constrained (/i/).
It thus appears that there is more room for /δ/ and /ɲ/ to
adapt to coarticulatory effects when flanked by a vocalic
segment that does not cause the tongue-body requirements
Table 5. Significant Stress and Speech Rate effects for F2 differences
between changing /i/ and /a/ over time.
Stress
(df = 2, 199)
Temporal comparison
1. V2 on/V1 off
2. C midpoint
3. V1 off/V2 on
4.
5. V1 midpoint/V2 midpoint
6.
7. V1 on/V2 off
F
ηp2
11.6**
13.5***
13.8***
10.5**
.05
.06
.06
.05
Rate
(df = 1, 199)
F
ηp2
4.5*
6.8**
5.5*
.02
.03
.03
Note. df = degrees of freedom; V1 = first vowel position; V2 = second
vowel position; C = consonant.
*p < .05. **p < .01. ***p < .001.
Recasens: Effect of Stress and Rate on Coarticulation
1421
Complimentary Author PDF: Not for Broad Dissemination
for the consonant to increase considerably. Vowel effects
in sequences with /l/ were, however, greater in the fixed /i/
versus the fixed /a/ condition—that is, there was more
anticipatory F2 lowering induced by V2 = /a/ on the consonant and preceding fixed /i/ than anticipatory F2 raising
exerted by V2 = /i/ on the consonant and preceding fixed
/a/. This finding may be related to the joint tongue-dorsum
lowering and retraction action of dark /l/ and following
/a/ during preceding /i/.
As shown earlier in the literature, V-to-V effects in
VCV sequences decrease as one gets away from the vowel
that exerts the coarticulatory effect. Moreover, as also
reported in some studies (Fowler & Brancazio, 2000;
Recasens, 2002), vowel effects may occur during the transconsonantal vowel but not during the intervocalic consonant, mostly when, as for alveolopalatal consonants, a
large portion of the tongue is involved in making a closure
and thus is left not too free to adapt to the contextual
phonetic segments. Several predictions on coarticulatory
direction were also confirmed by the data:
1.
Sequences with /ɲ/ showed more carryover than anticipatory vowel coarticulation (triggered essentially
by low back /a/), which should be related to the articulatory configuration for the alveolopalatal nasal
being more /j/-like at closure release than at closure
onset. The coarticulation data for VCV sequences
with the alveolopalatal nasal also provided evidence
for articulatory overshoot in the case of the sequence
/iɲi/ produced by speaker AL, who exhibited the
highest F2 frequency trajectory of all subjects; the
overshoot effect in question appears to result from
tongue-dorsum raising and fronting for both V1 = /i/
and following /ɲ/, causing dorsopalatal contact and
F2 during V2 = /i/ to increase above the expected
threshold.
2.
Dark /l/ favored vowel anticipation over vowel carryover. This finding was attributed to the fact that the
tongue-dorsum lowering and retraction movement
for this consonantal variety is anticipatory rather
than carryover and therefore allows coarticulatory
effects from V2 = /i/ rather than from V1 = /i/ to take
place. Following /a/ appears to assist the anticipatory
tongue-dorsum lowering and backing motion for
the consonant to become especially prominent during preceding /i/. However, VCV sequences with a
clearer variety of /l/ did not show consistent patterns
of vowel coarticulatory direction: Unlike sequences
with dark /l/, they did not favor vowel anticipation,
which appears to be in line with the fact that clear /l/
does not involve much anticipatory tongue-dorsum
lowering and backing; unlike sequences with /δ/
(see the next point), they did not favor vowel carryover effects from V1 = /i/ either, presumably because
some anticipatory predorsum lowering associated with
laterality conflicts with the vowel effects in question.
3.
Vowel coarticulatory effects in sequences with /δ/ favored the carryover direction in the fixed /a/ context
1422
and no specific direction in the fixed /i/ context. This
finding appears to be related primarily to the tonguedorsum raising/fronting carryover effects exerted
by V1 = /i/ on consonants such as /δ/ not involving
active tongue-dorsum activity. This vowel carryover
action may be assisted by V1 = /a/ causing the
tongue dorsum to occupy a relatively low position
during /δ/ and some delay of the anticipatory tongue
fronting/raising motion for V2 = /i/ during the
consonant.
A main goal of this investigation was to study the
effect of changes in stress and speech rate on vowel coarticulation in VCV sequences. Changes in stress turned out to
affect vowel duration and F2 frequency to a larger extent
than speech-rate variations: Vowels were longer and less
centralized when stressed versus unstressed for essentially
all speakers, and an increase in speech rate did not induce
the same degree of vowel shortening in all subjects and
yielded small or no F2 changes. Vowel coarticulation was
also greater and extended over a longer temporal period
when induced by stressed versus unstressed vowels rather
than by an increase in speech rate. These Catalan data are
in agreement with data for other languages showing that
vowels produced with different stress levels and at different
speech rates may be hypoarticulated or hyperarticulated
(de Jong, Beckman, & Edwards, 1993).
The initial prediction that stress and rate effects could
be more prominent in VCV sequences with less constrained
consonants than in those with more constrained consonants did not hold. Thus, although constrained consonants
turned out to be more resistant than unconstrained ones
to vowel effects in tongue-dorsum raising/lowering, this difference in vowel coarticulatory resistance applied largely
independent of whether the changing vowel was stressed
or unstressed and sentences were produced faster or more
slowly.
Another relevant finding is that stress and speech
rate did not interact with consonant-specific patterns of
vowel coarticulatory direction and therefore affected the
two coarticulatory directions alike—that is, changes in
vowel coarticulation over time for specific VCV sequences
occurred more or less equally in the anticipatory and carryover directions rather than in the direction that was
mostly favored by the consonantal articulatory gesture.
Thus, for example, the prevalence of vowel carryover over
vowel anticipation in /VɲV/ sequences did not become
more evident when stress fell on V1 than when it fell on V2
and therefore in /′VɲV/ versus /V′ɲV/ sequences; instead,
stress caused a similar increase in vowel coarticulation
in the target unstressed vowel in both /′VɲV/ and /V′ɲV/
structures and therefore at both anticipatory and carryover
levels.
To summarize, changes in vowel coarticulation associated with stress and speaking rate were found not to alter
the patterns of degree and direction of vowel coarticulation in VCV sequences composed of phonetic segments specified for different degrees of articulatory constraint. This
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015
Complimentary Author PDF: Not for Broad Dissemination
finding allows the conclusion that the contribution of segmental and prosodic factors to VCV coarticulation is organized hierarchically: Coarticulatory changes induced by
prosodic variations conform to segment-dependent patterns
of coarticulatory degree and direction. It also implies that
coarticulatory fields for specific phonetic segments may be
resized depending on prosodic condition.
The VCV coarticulation data reported in the Results
also support the view that prosodic variations affect vowels
rather than consonants. Indeed, stress-dependent vowel
effects were found to occur during the transconsonantal
vowel rather than during the consonant—that is, for Temporal Comparisons 4–7 rather than for Comparison 2 in
Figures 4–6. This finding is consistent with data from the
literature showing that vowel articulation is affected to a
larger extent than consonant articulation by an increase in
speech rate and by a shift from more formal to spontaneous
speech styles (see van Son & Pols, 1999) and also with the
notion that consonants but not vowels require the formation
of invariant articulatory constrictions.
Future work needs to be carried out in order to explore the extent to which the patterns of coarticulation reported in this study hold for more speakers, for real words
with the same segmental and stress patterns, and for
other languages besides Catalan. The perceptual effectiveness of these coarticulatory patterns should also be tested
with real speech stimuli excised from VCV sequences.
Acknowledgments
This research was funded by project FFI2013-40579-P from
the Spanish Ministry of Economy and Competitiveness, by ICREA
(Catalan Institution for Research and Advanced Studies), and
by the research group 2014SGR61 from the Catalan government.
I would like to thank Meritxell Mira for assistance with the analysis of the acoustic data.
References
Agwuele, A. (2004). Effect of stress on coarticulation. In J. Slifka,
S. Manuel, & M. Matthies (Eds.), From sound to sense: 50+
years of discoveries in speech communication (pp. C109–C114).
Cambridge, MA: Massachusetts Institute of Technology. Retrieved from http://www.rle.mit.edu/soundtosense/conference/
pdfs/fulltext/Saturday%20Posters/SB-Agwuele-STS.pdf
Agwuele, A., Sussman, H. M., & Lindblom, B. (2008). The effect
of speaking rate on consonant vowel coarticulation. Phonetica,
65, 194–209.
Cho, T. (2004). Prosodically conditioned strengthening and vowelto-vowel coarticulation in English. Journal of Phonetics, 32,
141–176.
Dart, S. (1991). Articulatory and acoustic properties of apical and
laminal articulations. UCLA Working Papers in Phonetics, 79,
1–155.
de Jong, K., Beckman, M. E., & Edwards, J. (1993). The interplay
between prosodic structure and coarticulation. Language and
Speech, 36, 197–212.
Fant, G. (1960). Acoustic theory of speech production. The Hague,
the Netherlands: Mouton.
Farnetani, E. (1990). V-C-V lingual coarticulation and its spatiotemporal domain. In W. J. Hardcastle & A. Marchal (Eds.),
Speech production and speech modelling (pp. 93–130). Dordrecht,
the Netherlands: Kluwer.
Farnetani, E., & Recasens, D. (2010). Coarticulation and connected
speech processes. In W. J. Hardcastle, J. Laver, & F. E. Gibbon
(Eds.), The handbook of phonetic sciences (2nd ed., pp. 316–352).
Malden, MA: Wiley-Blackwell.
Fowler, C. A. (1981). A relationship between coarticulation and
compensatory shortening. Phonetica, 38, 35–50.
Fowler, C. A., & Brancazio, L. (2000). Coarticulation resistance
of American English consonants and its effects on transconsonantal vowel-to-vowel coarticulation. Language and Speech,
43, 1–41.
Gay, T. (1978). Effect of speaking rate on vowel formant movements. The Journal of the Acoustical Society of America, 63,
223–230.
Grosvald, M. A. (2009). Long-distance coarticulation: A production
and perception study of English and American Sign Language
(Unpublished doctoral dissertation). University of California,
Davis.
Guenther, F. H. (1995). Speech sound acquisition, coarticulation,
and rate effects in neural network model of speech production.
Psychological Review, 102, 594–621.
Hertrich, I., & Ackermann, H. (1995). Coarticulation in slow
speech: Durational and spectral analysis. Language and Speech,
38, 159–187.
Kent, R. D., & Netsell, R. (1971). Effects of stress contrast on certain articulatory parameters. Phonetica, 24, 23–44.
Krull, D. (1989). Second formant locus patterns and consonantvowel coarticulation in spontaneous speech. Phonetic Experimental Research at the Institute of Linguistics, University of
Stockholm, 10, 87–108.
Magen, H. S. (1997). The extent of vowel-to-vowel coarticulation
in English. Journal of Phonetics, 25, 187–205.
Matthies, M., Perrier, P., Perkell, J. S., & Zandipour, M. (2001).
Variation in anticipatory coarticulation with changes in clarity
and rate. Journal of Speech, Language, and Hearing Research,
44, 340–353.
Mok, P. P. K. (2011). Effects of vowel duration and vowel quality
on vowel-to-vowel coarticulation. Language and Speech, 54,
527–545.
Moon, S.-J., & Lindblom, B. (1994). Interaction between duration,
context, and speaking style in English stressed vowels. The
Journal of the Acoustical Society of America, 96, 40–55.
Proctor, M. I. (2009). Gestural characterization of a phonological
class: The liquids (Unpublished doctoral dissertation). Yale
University, New Haven, CT.
Recasens, D. (1999). Lingual coarticulation. In W. J. Hardcastle
& N. Hewlett (Eds.), Coarticulation: Theory, data and techniques (pp. 80–104). Cambridge, United Kingdom: Cambridge
University Press.
Recasens, D. (2002). An EMA study of VCV coarticulatory direction. The Journal of the Acoustical Society of America, 111,
2828–2841.
Recasens, D. (2007). Patterns of VCV coarticulatory direction
according to the DAC model. In P. Prieto, J. Mascaró, &
M.-J. Solé (Eds.), Segmental and prosodic issues in Romance
phonology (pp. 25–39). Amsterdam, the Netherlands: John
Benjamins.
Recasens, D., & Espinosa, A. (2009). Dispersion and variability in
Catalan five and six peripheral vowel systems. Speech Communication, 51, 240–258.
Recasens, D., Pallarès, M. D., & Fontdevila, J. (1997). A model of
lingual coarticulation based on articulatory constraints. The
Journal of the Acoustical Society of America, 102, 544–561.
Recasens: Effect of Stress and Rate on Coarticulation
1423
Complimentary Author PDF: Not for Broad Dissemination
Simonet, M. (2010). Dark and clear laterals in Catalan and
Spanish: Interaction of phonetic categories in early bilinguals.
Journal of Phonetics, 38, 663–678.
Sproat, R., & Fujimura, O. (1993). Allophonic variation in
English /l/ and its implications for phonetic implementation.
Journal of Phonetics, 21, 291–311.
1424
van Son, R. J. J. H., & Pols, L. C. W. (1999). An acoustic description of consonant reduction. Speech Communication, 28, 125–140.
Whalen, D. H. (1990). Coarticulation is largely planned. Journal
of Phonetics, 18, 3–35.
Yun, G. (2007). Word frequency, stress and coarticulation in English.
Studies in Phonetics, Phonology and Morphology, 13, 315–331.
Journal of Speech, Language, and Hearing Research • Vol. 58 • 1407–1424 • October 2015

Similar documents