BRAIN

AND

LANGUAGE

38, 555-575 (19%)

A Kinematic Analysis of Anticipatory Coarticulation in the
Speech of Anterior Aphasic Subjects Using
Electromagnetic Articulography
WILLIAM
University

KATZ

of Texas

at Dallas

J~CHEN MACHETANZ
The Technical

University

of Munich

AND
UTE ORTH AND PAUL SCHGNLE
University

of Giittingen

An investigation was made into the extent and time course of anticipatory
coarticulation in the speech of two normal and two anterior aphasic, Germanspeaking subjects. Both labial and velar coarticulation gestures were investigated.
Subjects produced sentences containing target words contrasting in postconsonantal vowel rounding (e.g., [gali:ga]/[galy:ga]) and in nasality (e.g., [ti:da]/
[ti:na]). Speech kinematics were monitored by means of electromagnetic articulography. The data revealed that for correct productions, aphasic speakers’
coarticulatory patterns were more highly variable than those of control subjects.
These differences, however, were found chiefly for spatial displacement characteristics, while the temporal aspects of articulator movement necessary for
anticipatory coarticulation appeared largely intact. Articulator mistiming did not
appear to explain a small corpus of stop/nasal substitution errors produced by
one of the aphasic speakers. 8 1990 Academic PXSS. hc.

The authors thank John Bobela and Erich von Berger for their assistance in data analysis.
This research was sponsored by Grant NS 08176-OlAl to William Katz, University of
Goettingen/DFG grant to Jochen Machetanz, and BMFT Grant 0706839A to Paul-Walter
Schonle. Send reprint requests to William F. Katz, The University of Texas at Dallas,
Callier Center for Communication Disorders, 1966 Inwood Rd., Dallas, TX 75235.

0093-934)</90 $3 .OO
Copyright
Q 1990 by Academic
Press. Inc.
All rights of reproduction
in any form reserved.

556

KATZ ET AL.

BACKGROUND
An important goal of neurolinguistic
research has been to provide an
accurate account of the speech production capabilities of aphasic subjects. For this reason, a number of studies have been conducted investigating articulator timing and coordination in aphasic speech. This research has generally analyzed speech output at the level of individual
sounds or “phonemes.”
The data have been used to develop theories
concerning probable levels of speech-processing impairment in aphasia.
One theory which has been highly influential considers the errors in
speech produced by posterior aphasics to originate at the “phoneme
planning level” whereas “phonetic”
or articulatory errors are thought
to be more typical of anterior, nonfluent aphasia. This notion, first described explicitly by Alajouanine,
Ombredane, and Durand (1939), has
been supported by a number of recent acoustic and perceptual studies,
including investigations
of the voice onset time (VOT) of consonants
(Blumstein,
Cooper, Zurif, & Caramazza, 1977; Blumstein,
Cooper,
Goodglass, Statlender, & Gottlieb,
1980; Shewan, Leeper, & Booth,
1984; Gandour & Dardarananda, 1984) and the duration (Tikofsky, 1965;
Peterson & Lehiste, 1960; Ryalls, 1981, 1986), formant frequencies, and
fundamental frequencies (Ryalls, 1986) of vowels.
The general picture which has emerged is now thought to be more
complicated than originally conceived, yet still indicates that for anterior
aphasic subjects there are impairments
in the timing and integration of
movements of the articulatory
system. Interestingly,
anterior aphasic
difficulties do not seem to involve “across-the-board”
speech motor
impairment,
but instead affect particular speech gestures while leaving
other gestures relatively intact. For example, anterior aphasics seem to
have particular difficulty in gestures requiring precise laryngeal control
and the concurrent movement of independent articulators, as is required
in producing voice onset time (VOT) distinctions.
On the other hand,
the work of Shinn and Blumstein (1983), Katz (1987, 1988a), and others
have shown that anterior aphasics are capable of reaching the spatial
endpoints of articulatory positions, e.g., in positioning the vocal tract to
the appropriate place of articulation
for stop consonants. These data
suggest that spatial aspects of articulatory movement are relatively preserved in the speech of anterior aphasic patients. In contrast to anterior
aphasics, posterior aphasics appear to produce fewer errors of articulator
timing and coordination,
but instead make errors which are generally
“selectional”
or “phonemic”
in nature.
It is interesting to note that the majority of studies in this area have
been concerned with analyzing discrete consonant or vowel segments.
In recent years, however, investigators have also begun to examine
“sound-transitional”
or coarticulatory
phenomena in aphasic speech.
Coarticulation
refers to the fact that speech sounds are not strung to-

COARTICULATION

IN APHASIA

557

gether like beads on a string, but rather are overlapped in a time-compressed manner. The study of anticipatory (or right-to-left) coarticulation
is of special interest because it is considered to be a measure of the
preplanning of upcoming speech segments. By examining speakers’ ability to anticipate articulatory configurations, it is possible to gain insight
into the size and nature of motor sequencing units, as well as their
control.
Studies of coarticulation
in aphasic speech have investigated whether
damage to specific areas of the brain results in delays or deficits in the
execution of local anticipatory
speech gestures. The existence of this
type of speech timing difficulty is potentially important, because coarticulatory-type
difficulties could constitute an underlying cause of the
phonological speech errors observed in anterior aphasic speech. Claims
have been made, for example, that speech errors produced by anterior
aphasic subjects which may on the surface appear to be substitution
errors could, in fact, be due to discrete interarticulatory
timing difficulties
(e.g., Ziegler & von Cramon, 1985).
These studies have largely used acoustic or perceptual methodologies.
In acoustic studies of coarticulation,
research has focused upon labialization and nasalization.
In the labialization
studies, anticipatory
liprounding has been assessed by measuring the amount of formant shift
in prevocalic consonants as a function of the rounding feature of the
following vowel. For American English-speaking
aphasics, there have
been two acoustic studies of coarticulation
completed to date, which
have yielded somewhat mixed results. These mixed findings are likely
due to slightly differing methodologies.
Katz (1987, 1988a) examined two
regions of the consonantal portion of CV stimuli produced by normal
and anterior aphasic subjects. The degree of vowel-anticipatory
spectral
information
present in these consonantal regions was measured using
linear predictive coding (LPC) analysis. The results demonstrated that
both normal and anterior aphasic subjects produced robust acoustic correlates of anticipatory labialization,
with no discernable differences as a
function of Group.
Tuller and Story (1987) conducted a similar investigation
of normal
and aphasic subjects’ fricative-vowel productions, focussing upon slightly
different regions of the fricative waveform. For acoustic regions analogous to those examined by Katz, the results showed coarticulatory
patterns with few differences as a function of Group (similar to the
findings of Katz). However, at a slightly earlier region of the fricative
waveform (i.e., fricative midpoint),
acoustic correlates of anticipatory
coarticulation
were noted for only one out of five nonfluent speakers
studied. These findings were interpreted as indicating possible Group
differences in anticipatory coarticulation,
at a relatively early point in
fricative production.

558

KATZ ET AL.

The finding of significant coarticulatory
differences between anterior
aphasic and normal control subjects has also been reported in an acoustic
study of German. Ziegler and von Cramon (1986b), using a discriminant
analysis of LPC reflection coefficients, found less acoustic information
signalling labial anticipatory
coarticulation
present in the speech of
apraxia of speech’ subjects than in the speech of normal, control subjects.
In sum, although current acoustical techniques have promise for analyzing coarticulatory
patterns, it is difficult to arrive at a uniform picture
of the anticipatory labialization
capabilities of aphasic subjects from the
data currently available.
The perceptual studies of anticipatory labialization
conducted to date
have shown somewhat better concordance in their results, yet still do
not completely agree. In these studies, gated speech segments containing
coarticulatory information relating to the following vowel are presented
to listeners for a vowel-identification
task. Data for both American English (Katz, 1987, 1988a) and German (Ziegler & von Cramon, 1985) have
suggested that the speech of anterior aphasics contains a diminished
amount of perceptual information
for anticipatory coarticulation.
From
their results, Ziegler and von Cramon (1985) have concluded that apraxia
of speech patients show a “consistent delay in the initiation of vowel
gestures.” On the other hand, the data of Katz (1987, 1988a) showed
differences between listeners’ correct identification
of aphasic speakers’
fricative and stop-consonant stimuli, with listeners having a far more
difficult time detecting coarticulatory
information
in [ti/tu] stimuli as
compared with [si/su] stimuli. This suggests that instead of a “uniform
delay” in coarticulatory gestures, a more complicated interaction of deficits (based upon the level of complexity of each speech gesture) may
be involved.
With respect to velar movement and nasality in aphasic speech, relatively little acoustic or perceptual work has been undertaken. Ziegler
and von Cramon (1986b) measured the sound pressure level (SPL) of
words containing nasals, spoken by apraxia of speech patients and normal
control subjects. It was reasoned that the SPL would reflect the “smoothness in the diversion of the airstream from nasal to oral.” Specifically,
it was argued that an adequate synchronization
between nasal and oral
gestures would not cause a marked change in SPL levels, whereas there
would be detectable damping of the SPL when premature velar elevation
’ “Apraxia of speech” is a syndrome involving erroneous sound productions consisting
of mainly distortions, substitutions, and additions. Whether apraxia of speech is a uniquely
specified syndrome separate from aphasia, or whether it is a subset of Broca’s aphasia
remains a matter of debate (see Blumstein, 1981). For the present discussion, it will be
sufficient to note that speech errors made by subjects presenting with apraxia of speech
symptoms are relevant to investigations into anticipatory coarticulation breakdown.

COARTICULATION

IN APHASIA

559

occurred while the oral passage was still occluded by the tongue.* It was
found that certain apraxia of speech patients showed “marked damping”
in the transition towards the final vowel of [anal and [ina] utterances.
In addition, the apraxia of speech subjects (as a group) showed a broader
distribution of SPL changes than did the normal control subjects. From
these data, it was inferred that the speech of apraxia of speech patients
may show premature occlusion of the velar port.
One way to address these issues more directly is to conduct kinematic
studies of articulator movement. Sussman, Marquardt, MacNeilage,
and
Hutchison (1988) have recently reported some important kinematic findings concerning labial coarticulation.
The authors used EMG and straingauge monitoring systems to analyze the anticipatory labialization
patterns of a group of Broca’s aphasic subjects. It was found that for correct
productions of CV, CCV, and VC CCV stimuli, Broca’s aphasic subjects
exhibited lip and jaw coarticulatory behavior similar to normal subjects.
This finding replicated the acoustic data of Katz (1987, 1988a). In addition, a number of error-prone productions were observed, e.g., stuttered initial consonant segments, unusually slow and prolonged progression of the lip rounding gesture, and epenthetic vowels inserted into
consonant clusters. The authors stressed that study of these error-prone
productions are difficult, but that such analyses are necessary in order
to fully understand how brain lesions interrupt normal motor speech
processing.
With respect to velar coarticulation,
a series of kinematic experiments
have been completed by Itoh and colleagues. Itoh, Sasanuma, and Ushijima (1979) conducted a fibroscopic investigation of velar movements
during speech by a Japanese-speaking
apraxia of speech patient. The
main findings were that this subject’s general successional pattern of
velar gestures was similar to those of normal subjects, although the
apraxic subject showed much more variability.
Anticipatory
coarticulation was considered to be essentially intact, although some deviations
from normal patterns were observed. Specifically, there were cases in
which the verbal apraxic subject began coarticulatory
velar lowering
earlier than did normal subjects.
These findings were replicated in a subsequent study using an X-ray
microbeam technique, which allowed simultaneous observation of several
articulators (Itoh, Sasanuma, Hirose, Yoshioka, & Ushijima,
1980). It
was found that temporal organization between the lip and velum of the
apraxia of speech patient was generally intact, but was occasionally
disturbed. The authors concluded that “motor programming for the temporal organization of different articulators becomes difficult in apraxia
* Note, however, that without actual airflow measurements of the oral and nasal cavities,
it may be problematic to impute velar synchronization difficulties based upon acoustic
amplitude shifts.

560

KATZ

ET AL.

of speech under certain conditions resulting in inconsistent articulatory
errors” (p. 73).
In the present experiment, we have sought to provide additional data
regarding labial and nasal anticipatory
coarticulation
in the speech of
anterior aphasic subjects. Using the recent technology of electromagnetic
articulography,
both labialization and nasalization were investigated, using stimuli produced in three places of articulation.
High quality, synchronized acoustic data were also obtained in order to track the acoustic
and perceptual correlates of anticipatory coarticulation.
These acoustic
data are described in a separate article (Katz, Machetanz,
Orth, &
Schonle, in press).
METHOD
There has been some recent controversy concerning methodology used to study anticipatory coarticulation in aphasia (Katz, 1988b; Sussman et al., 1988; Ziegler, 1989). Sussman
et al. (1988) have discussed the fact that analyzing coarticulatory movement patterns of
anterior aphasic speech may be problematic because these subjects’ deviant output makes
coarticulatory patterns difficult to determine. For example, the authors, describing their
aphasic speaker database, state:
There were different kinds and degrees of articulatory behavior encountered that
could be classified as deviant. Some caused only slight problems in the measurements, others required adjustments to the normal measurement procedures,
and others precluded any measurements from being taken. (p. 375)
Having initially encountered problems similar to those described by Sussman et al. (1988),
we would like to emphasize a number of methodological steps taken in the present research
which assisted in the measurement of coarticulation, and provided greater generahzability
across subjects. First of all, the use of a magnetometer system permitted precise, relatively
noninvasive examination of articulator movement. Second, kinematic waveform-averaging
techniques were employed in order to elucidate regularities across speakers’ individual
utterances. Third, RMS distance-averaging techniques allowed precise quantification of
spatial variation within the speech of individual speakers. Finally, by defining coarticulation
as the first point in time at which context-dependent, horizontal-, or vertical-direction
differences arise for a given articulator, it was possible to obtain a definition of “gestural
anticipation” useful for aphasia research.

Subjects
Subjects included two anterior aphasic and two normal, control speakers. All subjects
were right-handed native speakers of German, from similar dialect regions. Aphasic subject
classifications were based upon clinical exam and speech pathology assessment, including
the Aachen Aphasia Exam (Huber, Weniger, Poeck, & Willmes, 1980).
Anterior aphasic subject EG was a Sl-year-old female librarian, who had a stroke in
1982. CT scans showed a large lesion in the region of the left middle cerebral artery (Fig.
1). The lesion extended deep into white matter, but spared the angular gyrus and supramarginal gyrus. The patient was initially classified as Global aphasic, but was subsequently
classified as a Broca’s aphasic.
Anterior aphasic subject AW was a 62-year-old engineer, who had a stroke in 1987. CT
scans revealed a large ischemic insult of the left middle cerebral artery, extending deep

COARTICULATION

FIG. 1. CT scan of patient EG demonstrating
left middle cerebral artery.

IN APHASIA

561

an extensive lesion in the region of the

into the insula, but sparing subcortical structures (Fig. 2). Patient AW was classified as a
Broca’s aphasic.
The two control subjects were matched for the sex and the approximate age of each
aphasic subject. Control subjects had no history of neurological disease, and no speech
or language difficulties.

Procedure
Speech kinematics were measured using the electromagnetic articulography system developed at the University of Gottingen Department of Clinical Neurophysiology (Schonle
et al., 1987). This device allows for the simultaneous recording of multiple points in and
outside of the vocal tract. Subjects were seated in a quiet testing room and were fitted
with a helmet containing three magnetic transmitter coils. Minute receiver coils were
attached to the upper lip (UL) and lower lip (LL) (for the labial study) and to the tongue
tip and velum (for the nasal study). The UL and LL coils were placed on the midsagittal,
exterior surface of each lip, at the most anterior point of the red margin. Velar coils were
placed on the midsagittal, uvular surface of the velum, at a position equidistant between
the uvular apex and base. Tongue coils were placed on the midsagittal, dorsal surface,
approximately l/3 the length of the tongue body from the tip. A computer was used to
sample positional data for two receiver coils, and to record acoustic data.

Speech Material
Speakers were asked to produce real word stimuli designed to probe the timing of
anticipatory labialization during consonant production and anticipatory nasalization during
vowel production (Table 1). There was a total of 10 stimulus items (4 labial, 6 nasal). The

562

KATZ

ET AL.

FIG. 2. CT scan of patient AW demonstrating
middle cerebral artery.

a large lesion in the region of the left

labial stimuli ([gali:ga]/[galy:ga]; [gale:za]/[gaL:za])
consist of word pairs contrasting minimally in the rounding feature of the vowel following the consonant [l]. The nasal stimuli
([ti:ga]/[tingl]; [ti:da]/[ti:na]; [ti:ba]/[ti:ma]) consist of word pairs differing in the nasality
features of each word’s initial vowel and middle consonant. The nasal stimuli were also
selected to represent velar, alveolar, and labial place of articulation for middle (wordinternal) consonants. All stimuli were embedded in the carrier phrase: “Ich sagte _
zweimal” (“I said twice”).
Each stimulus was repeated 15 times (in succession), with brief pauses between sentences. Intertrial intervals were cued by one of the researchers, and were approximately
3 set in duration. Labial and nasal productions were obtained in two separate testing
sessions, each of which lasted approximately 45 min. The order in which labial and nasal
stimulus sets were presented was counterbalanced between the four subjects. Stimulus
production within each testing session was fixed according to a uniform schedule.

Analysis
Subjects’ productions were perceptually screened by the researchers, and speech errors
were identified. For the normal control subjects, no speech errors were noted. For the
aphasic subjects, a total of 22 speech errors were detected (=7.3% of the 300 total repetitions by aphasic subjects). Errors were separately classified for further analysis. From
a total of 15 repetitions recorded for each stimulus type (per speaker), the first 10 correct
repetitions were selected for analysis.
Kinematic data were analyzed from graphic representations of articulator position and
tangential velocity. A time-synchronized display showing two channels of articulatory
movement and the acoustic waveform were displayed on a CRT. The direction of articulator
movement could be determined from separate plots showing anteroposterior (x-axis) and

COARTICULATION
TABLE

IN APHASIA

563

1

STIMULI

Labial

Nasal

“Geliege”
[galiga]
(gerund/lying

“Tiger”
[ti:ga]
(tiger)

down)

(ger./prevarication)

“Tingei”
[tingl]
(performance on a country stage)

“Gelese”
[gale:za]
(ger./reading)

“Tide”
[ti:da]
(tide)

“GelBse”
[gals:za]
(get./solving)

“Tine”
[ti:na]
(woman’s name)

“Geliige”

bhvl

“Tiber”
[ti:ba]
(Tiber river)
“Thieme”
[ti:ma]
(company name)

vertical (y-axis) motion. By comparing the extent and direction of articulator movement
during the production
of speech segments contrasting in labialization
(e.g.,
[gali:ga]/[galy:ga]) or in nasality (e.g., [ti:da]/[ti:na]),
it was possible to directly view
anticipatory coarticulation effects.
In addition, quantitative records of speech timing were obtained using interactive software designed for speech movement analysis. This software allowed researchers to place
cursors at specific points in the CRT display of speech kinematics, and automatically write
time and displacement values at these positions to files for statistical analysis.

RESULTS
I.
A.

AND CONCLUSIONS
Correct Productions

Labial Stimuli
i. Articulator displacement.
Figure 3 shows, for each speaker, plots
of lip movement
for stimuli contrasting
in labial rounding
(i.e.,
[gale:za]/[gals:za]
and [gali:ga]/[galy:ga]).
Each panel shows articulator
movement during the portion of the utterances beginning at the midpoint
of the vowel [a:] in the word “sagte” and ending at the beginning of the
diphthong [a’] in the word “zweimal.”
This portion was selected because
it could be precisely located in the kinematic and acoustic displays, and
because it contains the speech segments critical for analyzing anticipatory
coarticulation.
The kinematic plots shown in these figures are averaged
over 10 correctly-produced
stimuli. The uppermost graphs show move-

564

KATZ ET AL.

(Aphasic) AW

(Control) EO

(Aphasic) EG
___.-.-._._
-_---.-..

FIG. 3. Averaged articulator displacement for labial stimuli produced by normal and
aphasic speakers. Each panel plots horizontal (x-axis) and vertical (y-axis) movement for
10 repetitions of stimuli containing the vowel pairs [i:J/[y:] (left), and [e:l/[s:] (right).

ment for UL, and the bottom graphs show movement for LL (and jaw).
For each articulator, the x-axis represents displacement in the anteroposterior direction, while the y-axis represents displacement in the vertical direction. Thus, positive deflection (“rise”) along the x-axis indicates anterior motion of an articulator, while negative deflection (“fall”)
indicates posterior motion. Similarly, positive deflection along the y-axis
indicates upward motion, while negative deflection indicates downward
motion.
Lip rounding may be characterized in terms of both extension and
vertical movement (raising or lowering, depending upon the region of
the lip examined, and upon individual subject characteristics). The two
control speakers produced robust UL protrusion (extension and lowering)
for rounded (as compared with unrounded) stimuli. These lip protrusion

COARTICULATION

IN APHASIA

565

gestures were rapid, concise, and were clearly related to production of
the rounded vowel in the utterance. In addition, speaker PS showed
substantial extension and lowering of LL (and jaw) for both stimuli series,
while speaker EO demonstrated only slight LL (and jaw) protrusion for
[galy:ga] as compared with [gali:ga].
Considering next the anterior aphasic data, UL protrusion was also
found for rounded (as compared with unrounded) stimuli. The overall
time course of these gestures differed somewhat between the two aphasic
subjects. Speaker AW showed rapid UL movements corresponding with
production of rounded vowel segments (i.e., resembling the data of the
control subjects). In contrast, speaker EG showed more gradual UL
movements, with a much lesser amount of net displacement.
In terms
of LL (and jaw) movement, both aphasic speakers showed articulator
protrusion (extension and raising) for rounded (as compared with unrounded) stimuli.
In order to determine the extent to which speakers differed in the
regularity of articulator movement, a measure of item-to-item
displacement variation was computed. This involved determining the average
displacement
for 10 repetitions, calculating the Euclidean (root mean
square) distance between each individual repetition and the average, then
dividing this sum by the number of points. This may be represented by
the formula:
Root mean square
(RMS) displacement deviation = v I/n * 2:‘: , (a, - 6,)’
where n = number of points
a, = individual trial displacement values
6; = average displacement values

A graph of these data is shown in Fig. 4, which displays item-to-item
deviation for labialization
gestures by the four speakers. Data for the xand y-axes are displayed separately. The aphasic speakers showed
greater overall variation than the control speakers, with particularly high
variation for LL (and jaw) displacement. In order to quantify these data,
two-way (Group x Articulator [UL, LL]) analyses of variance (ANOVA)
were conducted. The results showed a significant main effect for Group
[F(l, 748) = 187.4, p < .OOl], and Articulator [F(l, 748) = 6.67, p <
.Ol]. These data indicate that aphasic speakers are more variable than
control speakers, and that LL (and jaw) shows greater displacement
variability than UL.
ii. Articulator timing. Because it was considered important to examine the extent to which anticipatory coarticulation
varied with speaking rate, the overall duration of speakers’ utterances were analyzed. Fig.
5 displays these data for labialized (left) and nasalized (right) stimuli.

566

KATZ
I
W

q
0

n

ET AL.
I

I

I

l W

.Q

vertical LL
horizontal
LL
vertical UL
horizontal
UL

00

SUBJECT
FIG. 4. Root mean square displacement variation for labial stimuli produced by normal
(PS, EO) and aphasic (AW, EG) speakers.

Each panel shows the median and quartile distribution
of durations for
the portion of each utterance begining with the midpoint of the [a:] in
the word “sagte” and ending with the midpoint of the diphthong [a’] in
the word “zweimal.”
Outiier data points (i.e., exceeding the interquartile
range) are plotted with asterisks. These graphs show that the aphasic
speakers produced slower speech, with greater variation in segment timWar

Labial

s

2000

s
!I
a

IWO

I

0
D.

00
.w
SUBJECT

.a

p.

.O
.w
SUBJECT

l e

FIG. 5. Duration of normal (PS, EO) and aphasic (AW, EG) speakers’ productions.
Each plot shows the median and quartile distribution for utterance durations beginning
Values exceeding the
with the [a:] of “sagte” and ending with the [a’] of “zweimal.”
interquartile range are represented with asterisks.

COARTICULATION

567

IN APHASIA

ing, than is found in the speech of normal, control subjects. In addition
to these overall durational characteristics,
further analysis of aphasic
speakers’ temporal patterns indicated that these subjects showed a prevalence of intersyllabic pauses, as well as occasional segment prolongations (particularly for vowels). These findings are in accord with previous
descriptions of verbal apraxic speech (Kent & Rosenbek, 1982, 1983;
Rosenbek, Kent, & LaPointe, 1984; Ziegler & von Cramon, 1986a).
In order to analyze the point of anticipatory coarticulation
onset, individual repetitions were inspected with attention paid to the exact point
in the acoustic waveform at which kinematic, coarticulatory effects could
be noted. That is, it was asked at which point in the speech waveform
could anticipatory movement adjustments pertaining to labialization first
be determined. Examples of these kinematic/acoustic
correspondences
may be observed in Fig. 6, which plots sample lip movement, speech

,._-_-,- -I...-. “PPERL,p

2. (Control) EO
..
- ....._ .-

3a. (Aphasic)

AW

4a. (Aphasic)

EG

3b. (Aphasic)

AW

4b. (Aphasic)

EG

1. (Control)

PS

,I 5 a: kta gxly-&

[I F

za:k

13

g2 I”

_ -..

?2,

FIG. 6. Examples of kinematic and acoustic data for speakers’ labial stimuli productions. Shown are horizontal (x-axis) and vertical (y-axis) movement plots for UL and LL
during production of sentences containing the target word [galy:ga] (underlined). Each
panel plots 2.56 set of speech.

568

KATZ ET AL.

waveforms, and phonetic transcriptions for sentence portions containing
[galy:ga] produced by the four speakers.
The results demonstrated
that lip protrusion was, for all subjects,
confined to a region proximate to the rounded vowel. The beginning of
labial protrusion generally began either shortly before, or during production of the syllable [go], i.e., the syllable preceding that containing
the rounded vowel.
Although speakers were quite consistent across repetitions (with control subjects showing greater consistency than aphasics), there were
notable Speaker- and Group-dependent
differences in coarticulation
onset. Examples of these differences may be seen in Fig. 6. Control speaker
PS consistently began protrusion at a point in time corresponding with
the end of the burst of [g], and the beginning of voicing for [a]. Control
speaker EO began slightly earlier, at a point approximately
10 msec
before the beginning of the burst for [g]. Aphasic speaker AW showed
more variable patterns. In five repetitions, this subject began anticipatory
labialization at approximately
the same point in the speech waveform as
the control subjects, i.e., within the first third of the segment corresponding to the vowel [a] of [go]. An example of this pattern is shown
in panel 3a. However, in the other five repetitions analyzed, rounding
was begun slightly earlier than in the speech of the control subjects. In
these repetitions, an example of which is shown in panel 3b, AW began
rounding anywhere from 40 msec preceding the burst up until the exact
onset of the burst. Aphasic subject EG produced speech with an UL
movement onset point resembling that of control subject EO, i.e., at the
very beginning of the burst (panel 4a). Yet, in one repetition rounding
was initiated approximately
40 msec before the burst for [g] (panel 4b).
As with the data for control subjects, the timing of coarticulatory
onset
for the LL coil showed local patterns closely associated in time with the
onset positions of the UL coil.
In summary, these data suggest that anterior aphasic subjects initiated
coarticulatory
labialization
at approximately
the same point in time as
normal, control subjects. Aphasic subjects showed more variation in
their onset positions, although this did not fit the pattern of a uniform
delay. Rather, aphasic speakers’ variation generally involved unusually
early lip protrusion in comparison to that found for control subjects.
B.

Nasal Stimuli
i. Articulator displacement.
Examples of velar movement patterns,
averaged across 10 repetitions, are given in Fig. 7. In this figure, the
alveolar consonant stimulus pairs [ti:da]/[ti:na]
are shown. The patterns
noted for these stimuli were found to be similar to stimuli differing in
consonant place-of-articulation
(i.e., [ti:ba]/[ti:ma]
and [ti:ga]/[tingl]).
The two control speakers showed no movement of the velum during

COARTICULATION

IN APHASIA

569

(Control)

PS

(Aphasic)

AW

(Control)

EO

(Aphasic)

EC

.-----

(Cl
(C]

= In1
= [d]

FIG. 7. Averaged articulator displacement for velar stimuli produced by normal and
aphasic speakers. Each panel plots horizontal (x-axis) and vertical (y-axis) movement for
10 repetitions of stimuli containing the stimulus words [ti:nal and [ti:da].

production of the non-nasalized stimulus [ti:da], whereas robust velar
port opening (i.e., velar extension and lowering) was observed for productions of the nasalized stimulus [ti:na]. In contrast, the aphasic speakers showed quite different movement patterns. Examination
of whole
sentence data for speaker AW showed that this subject tended to lower
his velum towards the end of words, then abruptly raise the velum to
achieve a closed velar port just prior to the beginning of a new word or
syllable. In his production of the non-nasalized stimulus [ti:da], there
was no velar lowering during the syllable [ti:], indicating a correct closure
of the velar port for non-nasalized, stop consonants (as found for control
subjects). There was, however, an unusual drop in velar height with the
production of the second syllable [da]. Superimposed upon this general
movement pattern, AW nevertheless appeared to produce appropriate
velar opening gestures for the nasalized stimulus [ti:na]. During nasalized
productions, there was a rapid lowering of the velum occurring at approximately the burst for [t], followed by a gradual rise in velar height
for the production of the subsequent word (“zweimal”).
This velar lowering had a more rapid trajectory than its non-nasalized counterpart.
Little anteroposterior motion was recorded for the velum of this subject.

570

KATZ

ET AL.

Aphasic subject EG showed a similar overall movement pattern as
subject AW, however EG showed even greater degrees of impairment.
Examination
of whole sentence patterns revealed a tendency for inappropriate velar lowering. Whereas in the data of AW this was noticeable
toward the end of words and syllables, in the speech of EG velar lowering
began shortly after the beginning of each new segment. Similar to AW,
EG produced rapid, abrupt velar closing gestures before new words and
syllables. For example, during production of the non-nasalized stimulus
[ti:da], there was inappropriate velar fronting, and slight velar lowering.
Nevertheless, EG’s velar movement for the nasalized stimulus [ti:na]
were in a direction appropriate for nasalization. As with the data of AW,
the velum began a more rapid descent for [ti:na] than for [ti:da], and
showed greater anterior displacement. This context-dependent
difference
appeared at approximately
the start of the burst for [t].
As with the labialization
stimuli, variation in velum displacement was
quantified by calculating the RMS distance between individual and averaged displacement waveforms. The results are shown in Fig. 8, which
plots, for each speaker, tongue and velum displacement variation. It was
found that the aphasic speakers produced greater overall variation in
velum and tongue displacement
than the control speakers. This was
particularly true for the velar movement of aphasic subject EG. In order
to quantify these data, two-way, Group x Articulator (tongue, velum)
analyses of variance (ANOVA) were conducted. The results showed a
significant main effect for Group [F(l, 1368) = 480.7, p < .OOl], and

El vertical velum
fB horizontal velum
0 vertical tongue
n horizontal tongue

00

aw

SUBJECT

FIG. 8. Root mean square displacement variation for velar stimuli produced by normal
(PS, EO) and aphasic (AW, EG) speakers.

COARTICULATION

571

IN APHASIA

Articulator [F(l, 1368) = 137.2, p < .OOl]. There was also a significant
Group x Articulator interaction [F(l) 1362) = 10.1, p < .002], reflecting
the fact that control speakers produced significantly more lingual than
velar variation, whereas this was not the case for the aphasic speakers.
In particular, aphasic subject EG showed the opposite pattern (i.e.,
greater velar than lingual variation). Taken together, these data indicate
that the two aphasic speakers, considered as a “group,”
were more
variable than controls, and that aphasic subject EG showed a particularly
high degree of velar movement variation.
ii. Articulator timing. The time course of correct velar stimuli productions was investigated by conducting comparisons of the acoustic and
kinetic data. As with the labialization results, subjects were found to be
quite consistent from utterance to utterance, though aphasics showed
greater variability
in onset position. Examples of speakers’ kinematic
and acoustic data for sentence portions containing the word [ti:na] are
given in Fig. 9. Each panel displays data for a subject’s single production,
showing tongue and velum displacement, the acoustic waveform, and a
phonetic transcription. Both control speakers showed context-dependent
differences at a region of the speech waveform located between the end
of the aspiration following the [t] segment, and the first 40 msec of vowel
pulsing. Similarly,
the first noticeable context-dependent
difference in
the movement traces of the aphasic subjects (i.e., visible changes in the
rates of velar lowering) was noted to be within this same temporal region.
(Control) PS

(Control) EO

(Aphasic) AW

(Aphasic) EG

[i$sa:kra

w

D Y.i Ink-l

1

I$

22:

lrts

LLi_lLa

cr..1

FIG. 9. Examples of kinematic and acoustic data for speakers’ single labial stimuli
productions. Shown are horizontal (x-axis) and vertical (y-axis) movement plots for the
tongue tip and velum during production of sentences containing the target word [ti:na]
(underlined). Each panel plots 2.56 set of speech.

572

KATZ

ET AL.

In other words, even though the displacement of velar position was noted
to be qualitatively
different and more variable for aphasic speakers, the
aphasic subjects appeared to initiate nasalization gestures at approximately the same point in time during speech as normal controls. These
data suggest that, although the overall ability to control spatial positioning
of the velum was clearly compromised
for aphasic speakers, temporal
aspects of velar, coarticulatory
movement appeared to be relatively
preserved.
II.

Error-Prone

Productions

An investigation was made of the claim that speech production errors
perceptually resembling phone substitutions might in fact be the result
of discrete, interarticulatory
phasing difficulties (e.g., Mlcoch & Noll,
1980; Ziegler & von Cramon, 1985). There were, however, very few
speech errors in the database containing the target structures of interest.
Speaker AW (who exhibited a lesser degree of speech difficulty than
EG) produced 3 errors, while EG produced 19 errors. Of the 22 total
errors recorded, the majority involved intraword or intrasyllable
hesitations, along with occasional consonant or vowel distortion errors (e.g.,
excessive consonant aspiration, inappropriate vowel nasalization). There
were six cases involving perceptually apparent substitutions. Of these,
three involved segments of interest. These consisted of instances in which
subject EG produced substitutions of “[ti:da]”
for [ti:na] targets. For
these cases, there were found to be [d] stop consonant bursts in the
acoustic waveforms. The kinematic data showed that in two of the cases,
velar movement more closely resembled correctly-produced
[d] than [n]
gestures, while in the third case the displacement patterns fell midway
between those typical of nasal and non-nasal consonants. Taken together,
these data do not suggest that a slightly mistimed [n] production resulted
in a [d] percept, but rather indicate that these errors likely resulted from
phone selectional (“phonemic”)
difficulties.
DISCUSSION
In this study, an investigation was made of the extent and time course
of local speech events produced by normal and anterior aphasic subjects.
Although the data must be considered preliminary
because there were
only two aphasic and two control subjects investigated, the findings
nevertheless address a number of important issues concerning the neurological bases of speech production.
To briefly summarize the results, it was found that for both labial and
nasal (correct) productions,
aphasic speakers’ coarticulatory
patterns
were more highly variable than those of control subjects. These differences, however, were found chiefly for spatial displacement characteristics, while the temporal aspects of articulator movement necessary for

COARTICULATION

IN APHASIA

573

anticipatory coarticulation
appeared largely intact. It was also found that
velum mistiming did not appear to explain a small corpus of stop/nasal
substitution errors produced by one of the aphasic speakers.
The chief question which these data address concerns whether anterior
aphasic subjects show systematic delays in the production of anticipatory
coarticulation
information
during speech. With respect to labial anticipation, uniform coarticulatory
delays on the order of 20-30 msec have
been proposed as a possible speech characteristic of German (Ziegler &
von Cramon, 1985) and English-speaking
(Tuller & Story, 1987) anterior
aphasic subjects. In contrast, Katz (1987, 1988a) and Sussman et al.
(1988) have provided evidence suggesting that temporal control of anticipatory labial coarticulation
is largely intact in anterior aphasic speakers’ correct productions. The data from the present experiment replicate
the findings of Sussman et al. (1988), and support the acoustic findings
of Katz (1987, 1988a), in that anterior aphasic subjects were found to
produce coarticulatory
gestures as early as (or earlier than) matched
control speakers.
A possible explanation for previous findings of perceptually apparent
delays in coarticulation
might concern the fact that anterior aphasic
speech has been shown to be impoverished
in a number of ways (e.g.,
marked by excessive aspiration of consonants, mistimed VOT, and highly
variable vowel formant frequencies). It is possible that speech having
such deficiencies could tax the perceptual systems of listeners to the
point where rounding information is obscured. In other words, there may
be a threshold effect, where the same amount of rounding cues present
in the speech of a normal speaker would not be recognizable in the
speech of an anterior aphasic speaker. Such an explanation would also
be compatible with the data of Katz (1987, 1988a), who showed that
American English-speaking,
anterior aphasic subjects’ acoustic cues pertaining to rounding were indistinguishable
from those of normal controls,
whereas normal listeners showed significantly poorer perception of coarticulatory information
when presented with the speech of the aphasic
(as compared to control) speakers. With respect to anticipatory
velar
movement, the present study addresses the question of whether discoordinations in velar movement might correspond to coarticulatory impairments (e.g., Mlcoch & Noll, 1980) and whether such discoordinations
might also account for perceptually apparent “substitution”
errors in the
speech of anterior aphasic subjects (Itoh et al., 1979, 1980; Ziegler &
von Cramon, 1986b). Concerning anticipatory impairments,
the present
data replicate the findings of Itoh et al. (1979, 1980), in that they demonstrate essentially intact anticipatory
coarticulation
in the correct
speech of anterior aphasic subjects.
As for theories concerning aphasics’ error-prone productions, the present data do not rule out the possibility that interarticulatory
phasing

KATZ

574

ET AL.

difficulties may account for occasional substitution errors. In the present
database, however, there was very little evidence for this. Of the three
stop/nasal substitution errors examined, all contained clearly identifiable
stop consonant bursts, suggesting selectional (“phonemic”)
rather than
interarticulatory
discoordination
in motor output. Only one of the three
substitution errors showed kinematic patterns allowing for the possibility
of interarticulatory
discoordination
(i.e., velar displacement patterns midway between those typical of nasal and non-nasal consonants).
To the extent that the present data indicate little coarticulatory deficit
in adult patients having substantial damage to anterior brain structures,
the data suggest that speech sound transitioning
capabilities are either
not subserved by these adult brain structures, or that these capabilities
somehow remain intact in the face of this damage. If future experimentation confirms the finding that highly automatized
behavior such as
anticipatory coarticulation does not require the integrity of anterior structures in adult subjects, then models implicating these regions (e.g., Broca’s region) as general “speech programming”
centers may require substantial revision.
REFERENCES
Alajuoanine,

T., Ombredane,

phonetique

dans I’uphasie.

A., & Durand, M. 1939. Le syndrome
de desintegration
Masson et Cie.
Blumstein, S. E. 1981. Phonological aspects of aphasia. In M. Sarno (Ed.), Acquired
aphasia.
New York: Academic Press.
Blumstein, S. E., Cooper, W. E., Zurif, E. B., & Caramazza, A. 1977. The perception
and production of voice-onset time in aphasia. Neuropsychologiu, 15, 371-383.
Blumstein, S. E., Cooper, W. E., Goodglass, H., Statlender, S., & Gottlieb, J. 1980.
Production deficits in aphasia: A voice onset time analysis. Brain and Language.
9,
153-170.
Gandour, J., & Dardarananda, R. 1984. Voice onset time in aphasia: Thai, II. Production.
Brain

and Language,

23, 206-224.

Huber, W., Weniger, D., Poeck, K., & Willmes, K. 1980. Der Aachener Aphasie Test.
Nervenarzt,
51, 475-482.
Itoh, M., Sasanuma, S., & Ushijima, T. 1979. Velar movements during speech in a patient
with apraxia of speech. Brain and Language.
7, 227-239.
Itoh, M., Sasanuma, S., Hirose, H., Yoshioka, H., & Ushijima, T. 1980. Abnormal articulatory dynamics in a patient with apraxia of speech: X-ray microbeam observation.
Brain and Language,
11, 66-75.
Katz, W. F. 1987. Anticipatory labial and lingual coarticulation in aphasia. In J. Ryalls
(Ed.), Phonetic approaches
to speech production
in aphasia
and related disorders.
San Diego: College-Hill Press.
Katz, W. F. 1988a. Anticipatory coarticulation in aphasia: Acoustic and perceptual data.
Brain

and Language,

35, 340-368.

Katz, W. F. 1988b. “Methodological
considerations” reconsidered: Reply to Sussman et
al. 1988. Brain and Language,
35, 380-385.
Katz, W. F., Machetanz, J., Orth, U., & Schoenle, P. in press. Anticipatory labial coarticulation in the speech of German-speaking anterior aphasic subjects: Acoustic analyses.

Journal

of Neurolinguistics.

COARTICULATION

IN APHASIA

575

Kent, R. D., & Rosenbek, J. C. 1982. Prosodic disturbance and neurological lesion. Bruin
and Language,
15, 259-291.
Kent, R. D., & Rosenbek, J. C. 1983. Acoustic patterns of apraxia of speech. Journal of
Speech

and Hearing

Research,

26, 231-249.

Mlcoch, A. G., & NOR, J. D. 1980. Speech production models as related to the concept
of apraxia of speech. In N. J. Lass (Ed.), Speech and language:
Advances
in basic
research
and practice
(Vol. 4, pp. 201-239). New York: Academic Press.
Peterson, G. E., & Lehiste, I. 1960. Duration of syllabic nuclei in English. Journnl of the
Acousfical

Society

of America,

32, 693-703.

Rosenbek, J. C., Kent, R. D., & LaPointe. L. L. 1984. Apraxia of speech: An overview
and some perspectives. In J. C. Rosenbek, M. R. McNeil, & A. E. Aronson (Eds.).
Apraxia
of speech: Physiology,
acoustics,
iingttisrics,
managemenf.
San Diego: College-Hill Press.
Ryalls, J. H. 1981. Motor aphasia: Acoustic correlates of phonetic disintegration in vowels.
Neuropsychologia,
19, 365-374.
Ryalls, .I. H. 1986. An acoustic study of vowel production in aphasia. Brain and Language,
29, 48-67.

Schonle, P. W., Grabe, K., Wenig, P., Hohne, J., Schrader, J., & Conrad, B. 1987.
Electromagnetic articulography: Use of alternating magnetic fields for tracking movements of multiple points inside and outside the vocal tract. Brain and Language,
31,
26-35.

Shewan, C.. Leeper, H., & Booth, J. C. 1984. An analysis of voice onset time (VOT) in
aphasic and normal subjects. In J. Rosenbek. M. R. McNeil. & A. E. Aronson (Eds.),
Apraxia
of speech. San Diego: College Hill Press.
Shinn, P., & Blumstein, S. B. 1983. Phonetic disintegration in aphasia: Acoustic analysis
of spectral characteristics for place of articulation. Bruin and Language,
20, 90-I 14.
Sussman. H.. Marquardt, T.. MacNeilage, P., & Hutchison, J. 1988. Anticipatory coarticulation in aphasia: Methodological considerations. Brain and Language.
35, 369379.

Tikofsky, R. 1965. Phonetic characteristics
ofdysarthria.
tion, The University of Michigan, Ann Arbor, MI.
Tuller, B., & Story. R. 1987. Anticipatory coarticulation
Phonetic

approaches

to speech

production

in aphasia

Office of Research Administrain aphasia. In J. Ryalls (Ed.),
disorders.
San Diego:

and related

College-Hill Press.
Ziegler, W. 1989. Anticipatory coarticulation in aphasia: More methodology: A reply to
Sussman et ul. (1988) and Katz (1988). B rain and Language,
37, 172-I 76.
Ziegler, W.. & von Cramon, D. 1985. Anticipatory coarticulation in a patient with apraxia
of speech. Brain and Langrrage. 26, 117-130.
Ziegler, W., & von Cramon, D. 1986a. Disturbed coarticulation in apraxia of speech:
Acoustic evidence. Brain and Language,
29, 34-47.
Ziegler, W., & von Cramon, D. 1986b. Timing deficits in apraxia of speech, European
Archives

of Psychiatry

and Neurological

Sciences,

236, 44-49.