Primary Stress And Intelligibility Research To Motivate The Teaching Of Suprasegmentals
Primary Stress and Intelligibility:
Research to Motivate the
Teaching of Suprasegmentals
LAURA D. HAHN
University of Illinois at Urbana-Champaign
Urbana-Champaign, Illinois, United States
This study examined native English speakers’ reactions to nonnative
primary stress in English discourse. I measured North American under-
graduate students’ processing, comprehension, and evaluations of
three versions of an international teaching assistant’s speech: with
primary stress correctly placed, incorrectly placed, or missing entirely.
Results indicated that when listening to speech with correct primary
stress, the participants recalled signiﬁcantly more content and evalu-
ated the speaker signiﬁcantly more favorably than when primary stress
was aberrant or missing. Listeners also tended to process discourse
more easily when primary stress was correct, but the result was not
signiﬁcant. These ﬁndings provide insights into how using primary
stress affects international TAs’ intelligibility. They also provide empiri-
cal support and suggest new ideas for current pedagogical practices that
emphasize suprasegmentals in teaching pronunciation.
N umerous pedagogical resources on ESL/EFL pronunciation advo-
cate teaching nonnative speakers (NNSs) suprasegmentals to im-
prove the intelligibility of their speech. However, little empirical support
exists for such claims. Knowing how the various prosodic features
actually affect the way native speakers (NSs) process nonnative speech
would substantially strengthen the rationale for current pronunciation
pedagogy. Such information would also affect the training of interna-
tional teaching assistants (ITAs), who interact with students in extended
discourse. Because the relationship between suprasegmentals and intelli-
gibility is so complex, it is helpful to isolate particular suprasegmental
features for analysis. One feature that warrants attention is primary
stress, also known as sentence stress (Bardovi-Harlig, 1986; Schmerling,
1976), accent (Bolinger, 1972; Gunter, 1974), nucleus (Cruttenden, 1997;
Gussenhoven, 1983, 1985), and tonic (Crystal, 1969; Halliday, 1967a,
1967b). Primary stress is realized in speech by combining a detectable
TESOL QUARTERLY Vol. 38, No. 2, Summer 2004 201
change in pitch with increased vowel duration and increased intensity
(e.g., Bolinger, 1986; Cruttenden, 1997; Crystal, 1969; Ladd & Cutler,
1983; Lehiste, 1969; Lieberman, 1967).
In English discourse, primary stress signals new and contrastive
information (Bardovi-Harlig, 1986; Brown, 1983; Chafe, 1976; Clark &
Haviland, 1977; Halliday, 1967a, 1967b; Halliday & Hasan, 1976; Sperber
& Wilson, 1986). For example, in the following dialogue between Person
A and Person B, when B introduces new information to the discourse
context it receives primary stress (indicated by ´), while lexical items that
are given or that represent old information are destressed (indicated by
a wavy underline):
A: Are you ready?
B: I’m álways ready.
Similarly, primary stress marks contrasts, as in “I prefer réd wine to whíte
Thus in English, new and contrastive information is presented in
stressed elements, and old or given information is expressed in un-
stressed elements. For convenience, these relationships can be called the
given-new stress connection (GNSC). This article presents results of a study
examining the GNSC’s effects on the intelligibility and evaluation of
nonnative speech, and it discusses the pedagogical implications of
ﬁndings for ESL pronunciation teachers.
PEDAGOGICAL CLAIMS ABOUT SUPRASEGMENTALS
In her 1991 article summarizing the state of the art in pronunciation
pedagogy, Joan Morley identiﬁes suprasegmentals as one guiding prin-
ciple for pronunciation teaching. Current practice “[redirects] priorities
within the sound system to a focus on the critical importance of
suprasegmentals (i.e., stress, rhythm, intonation, etc.) and how they are
used to communicate meaning in the context of discourse” (1991, p.
493). This statement reﬂects a commonly accepted claim in pronuncia-
tion pedagogy today: that suprasegmentals play a crucial role in commu-
nication. However, claims about suprasegmentals’ relevance to intelligi-
bility can be traced back at least to structural linguists such as Nida
How often we have had the experience of hearing some foreigner speak
English with perfectly intelligible consonants and vowels and with standard
grammatical forms; and yet we have had the greatest of difﬁculty in under-
standing because the intonational patterns were entirely unnatural and
202 TESOL QUARTERLY
strange to us. Moreover, we may completely misinterpret a person speaking
English. . . . Proper intonation contributes a high percentage to the total
intelligibility of speech. (pp. 117–118)
Such assertions continued in ESL/EFL pedagogy, reaching a peak in
the 1980s and 1990s with the advent of communicative language teaching
and research on discourse. This era brought a proliferation of claims about
the value of teaching suprasegmentals and reﬂected a deeper understand-
ing of the relationship between suprasegmentals and intelligibility. For
example, McNerney and Mendelsohn (1992) claim that “a short term
pronunciation course should focus ﬁrst and foremost on suprasegmentals
as they have the greatest impact on the comprehensibility of the learner’s
English. We have found that giving priority to the suprasegmental aspects
of English not only improves learners’ comprehensibility but is also less
frustrating for students because greater change can be effected” (p. 132).
Brazil, Coulthard, and Johns (1980), Pennington and Richards (1986),
Morley (1991), Brown (1995), Clennell (1996), Celce-Murcia, Brinton,
and Goodwin (1996), and Chela-Flores (1998) have presented similar
arguments about pronunciation instruction. Although these authors base
their assertions on a theoretical understanding of prosody in discourse,
they offer little, if any, empirical evidence to support their claims about
how suprasegmentals affect the intelligibility of nonnative speech.
NNSs’ USE OF PRIMARY STRESS
In general, NNSs ﬁnd the rhythm and stress patterns of English
challenging. Studies such as Adams (1979), Wenk (1985), Mochizuki-
Sudo and Kiritani (1991), and Anderson (1993) have investigated NNS’
production of English rhythm and concluded that transfer from the
native language interfered with learners’ ability to appropriately pro-
duce English-like stress alternations across a phrase.
Other studies have more speciﬁcally examined NNSs’ production and
perception of English primary stress. Wennerstrom (1994) found that NSs
of Thai, Japanese, and Spanish failed to use pitch movement to highlight
new or contrastive information to the same degree that native English
speakers do. She also observed that the NNSs used “less reduction of pitch
and volume” on the utterances’ given information (p. 409). Juffs (1990)
found that Chinese learners of English used pitch movement on every
word in a message unit, “using tonic [i.e. primary] stress on virtually every
lexical item, whether it be semantically important or a function word” (p.
107). Other studies documenting NNSs’ difﬁculties with primary stress
include de Bot (1986), Watanabe (1988), Hahn (1999), and Pennington
and Ellis (2000).
PRIMARY STRESS AND INTELLIGIBILITY 203
Clearly, many nonproﬁcient NNSs from many linguistic backgrounds
have difﬁculty mastering the primary stress system in English. They
exhibit two major problems: misplacing primary stress (often stressing
given information instead of new) and stressing all words in an utterance
more or less equally, without one prominent stress. In both cases, these
learners of English violate the GNSC: They fail to stress new information
by increasing pitch and volume and to destress old information by
reducing pitch and volume.
INTELLIGIBILITY OF NONNATIVE
How, then, does NNSs’ difﬁculty with primary stress in discourse affect
intelligibility? Intelligibility is a complex construct, and as such it has
been deﬁned and measured in numerous ways (Albrechtsen, Henriksen,
& Faerch, 1980; Fayer & Krasinski, 1987; Jenkins, 2000, 2002; Ludwig,
1982; Munro & Derwing, 1995a, 1999; Nelson, 1983; Smith & Raﬁqzad,
1985; Varonis & Gass, 1982).
Some studies of ITAs’ speech have proposed correlations between
using suprasegmentals accurately and speaker intelligibility. Tyler, Jeffries,
and Davies (1988) studied the discourse of ITAs whom undergraduates
perceived as disorganized and unfocused. They found that these ITAs
used too many pauses, too many primary stresses per message unit, and
inappropriate falling intonation. Gallego (1990) had undergraduates
view videotapes of ITAs and stop the tape when they felt that communi-
cation broke down. When ESL experts analyzed these breakdowns, they
discovered that most of them occurred when ITAs made pronunciation
errors, usually word stress. However, these studies do not show that the
errors caused the undergraduates’ perceptions.
Outside of the ITA context, other research has suggested that
suprasegmental features such as wave duration and wave peak amplitude
(Constantinou, 1993), word stress (Bansal, 1969), and the interstress
interval (Anderson, 1993) affect intelligibility. Jenkins (2002) examined
communication breakdowns of nonnative to nonnative speech, and
found “a combination of phonological errors which caused the most
serious problems in [her] data: misplaced tonic (nuclear) stress along
with a consonant substitution within the wrongly stressed word” (p. 89).
Psycholinguistic studies of speech processing suggest that even when
NSs violate the GNSC in ways that NNSs do, sentence processing for the
listener becomes more difﬁcult (e.g., Birch & Clifton, 1995; Bock &
Mazella, 1983; Cutler, 1984; Eefting, 1992; Fowler & Housum 1987;
Terken & Hirschberg, 1994; Wells, 1986). However, the stimuli in these
studies involved only very brief contexts, not longer pieces of discourse
204 TESOL QUARTERLY
where the cumulative effect of aberrant primary stress might be more
salient, as Nash (1971) hypothesized.
Sociolinguistic perspectives also elucidate intelligibility because par-
ticipants in a communicative event often do more than process speech in
the linguistic area of their brains; they have other, more subjective
experiences as well. Consciously or not, they may judge the speaker’s
personality and linguistic abilities and the quality of the message (Eisen-
stein, 1983; Ludwig, 1982; Ryan, 1983). Listeners will usually downgrade
the personalities of those who speak accented English (Anisfeld, Bogo, &
Lambert, 1962; Mulac, Hanley, & Prigge, 1974; Raisler, 1976; Webster &
Kramer, 1968). However, these types of studies often do not specify the
nature of “accentedness” or distinguish among the effects of deviances in
speciﬁc segmental and suprasegmental features in nonnative discourse.
One exception is Pickering (2001), who argues that nonnative tonal
structure (overuse of falling tones) led listeners to characterize her
subjects as “unsympathetic and uninvolved” (p. 233).
Evidence is contradictory concerning the extent to which NSs evaluate
NNSs’ linguistic ability according to other features of the communicative
event (Eisenstein, 1983). Some studies of ITA teaching have suggested
that undergraduate students may evaluate ITAs’ speech according to
such factors as the ITA’s perceived country of origin (Brown, 1988) or
the general “social mythology” associated with ITAs (Orth, 1982) rather
than an ITA’s actual linguistic ability. Other studies, however, conﬁrm
that untrained NSs can accurately evaluate linguistic skills (e.g.,
Albrechtsen, Henriksen, & Faerch, 1980; Flege, 1984; Palmer, 1973).
THE CURRENT STUDY
The literature reviewed shows that NNSs frequently violate the GNSC.
A practical question then emerges: How do such violations affect native-
speaking listeners? The psycholinguistic literature indicates that NSs may
process and comprehend such discourse with more difﬁculty and that
nonnativelike stress may elicit negative evaluations from NSs. The
question also has pedagogical ramiﬁcations, as reﬂected in the summons
from Pennington (1994): “As a basis for deciding which features to
teach, research is needed to identify those phonological features to
which native listeners attend most and react most strongly” (p. 8).
The current study took up this challenge to determine the impact that
NNSs’ misplaced or missing primary stress could have on native-speaking
listeners. To control for other linguistic variables, three versions of a text
were adapted from an authentic academic lecture (Triandis, 1990) and
recorded by a NNS. The versions were identical except that Versions B
and C reﬂected nonnative usage and violated the GNSC as follows:
PRIMARY STRESS AND INTELLIGIBILITY 205
Version A: GNSC was maintained.
Example: The úrban environment is more individualistic than the rúral
Version B: GNSC was violated through misplaced primary stress.
Example: The urban environment is more individualistic than the rural
Version C: GNSC was violated through absence of primary stress.
Example: The urban environment is more individualistic than the rural
The following research questions were investigated: Would NSs pro-
cess the discourse more easily when the speaker maintained the GNSC?
Would they comprehend more of the content? Would they evaluate the
speaker and the speech more favorably?
An experimental study was designed using oral texts constructed to
systematically vary the GNSC, and naïve listeners were asked to respond
to three tasks, each corresponding to one of the research questions.
The study had 90 subjects, all of them ﬁrst-semester freshman students
at a large Midwestern public university and NSs from homes where only
North American English was spoken; they were randomly assigned (30
each) to one of three experimental groups. Though exposure to
nonnative speech can affect the listeners’ ability to comprehend native
speech (Smith & Bisazza, 1982), a postexperiment questionnaire (Hahn,
1999) revealed that participants had no signiﬁcant differences in their
exposure to nonnative speech [χ2(2) .643, p .73]. The participants
also showed no signiﬁcant difference in prior familiarity with the content
of the lecture [χ2(2) 2.1, p .35].
The stimuli for the experiment consisted of three versions of a text
(see Appendix). The language and style of the original lecture were
maintained wherever possible. For the sake of cohesion, the text
included sentences that were stressed on the last content word (the most
206 TESOL QUARTERLY
neutral pattern); these yielded only one third of the primary stresses in
the text. Recordings of the text by seven NSs and two NNSs conﬁrmed
the researcher’s normal placement, and thus misplacement and
nonplacement, of the primary stresses.
A male NS of Korean with a high proﬁciency level in oral English,
experience as an ITA, and graduate-level training in phonetics recorded
each version of the text. Multiple recordings were made to ensure that
the versions did not contain signiﬁcant segmental distortions and were
prosodically identical to each other except for primary stress. Digital
editing techniques were employed (using Macromedia SoundEdit 16) to
make the versions equal in other ways, including volume and length
(approximately 4.5 minutes). For example, some of the speaker’s pause
times were either shortened or lengthened to ensure consistency across
versions. Auditory analyses by two native-English-speaking pronunciation
teachers conﬁrmed that Versions A, B, and C satisfactorily represented
their intended stress patterns. Conﬁrmation was corroborated by acous-
tic analyses of the pitch, energy, and duration of relevant syllables in the
text using a computerized speech lab (model 4300, Kay Elemetrics).
Measure of Difﬁculty Processing Discourse
The subjects’ difﬁculty processing the discourse in each text version
was evaluated using the dual-task paradigm prevalent in psycholinguistic
studies of spoken language processing (e.g., Hecker, Stevens, & Williams,
1966; Munro & Derwing, 1995b; Schmidt-Nielsen, Kallman, & Meijer,
1990; Tun, Wingﬁeld, & Stine, 1991). Tun, Wingﬁeld, and Stine summa-
rize this approach:
The logic of this method is that if attentional capacity is constrained by a
limited pool of processing resources, then changes in the resources allocated
to a primary task such as listening to a speech should be reﬂected by
reciprocal changes in performance on a secondary task that is carried out
concurrently. If an individual devotes more resources to the primary task,
performance should be poorer on the secondary task. This will, of course, be
true to the extent that task demands overlap. (pp. 3–4)
The participants’ primary task was to understand and remember the
lecture’s content. Their secondary task was to monitor for a tone
presented intermittently in the background of the speech; subjects
clicked a computer mouse when they heard the tone, and a computer
measured the reaction time between the tone and the mouse click. The
tones were generated randomly at 3- to 7-second intervals (Tun, Wingﬁeld,
& Stine, 1991) throughout the audio lecture. The same random pattern
was applied to all three versions such that the tones occurred at very
PRIMARY STRESS AND INTELLIGIBILITY 207
nearly the same places in the text. A computer program presented the
stimulus (the oral text and the tones) and recorded the reaction time (in
milliseconds). A pilot study indicated that subjects listening to Version A
without the tones did not score signiﬁcantly higher on a recall task [F(3,
50) .107, p .96] or a comprehension quiz [F(3, 50) 1.126, p
.35] than subjects in any of the other groups. The tone detection task
itself therefore did not interfere signiﬁcantly with subjects’ comprehen-
sion of the discourse.
Measures of Comprehension
The study measured overall comprehension of the lecture in two ways.
In the ﬁrst measure, subjects were asked to write down as much as they
could recall from the lecture so that the researcher could determine the
number of main ideas and details each subject had understood and
retained. The second measure was a short-answer comprehension quiz,
designed to test how well the subjects understood and remembered
particular information in the lecture. The questions, developed through
a series of pilot tests, elicited each main idea in the lecture. Answers to
the quiz were assigned a maximum value of 2 points each. The scoring
scheme consisted of 0 points for “no answer” or “completely incorrect,”
1 point for “partially correct,” and 2 points for “completely correct.”
Measure of Evaluative Reactions to the Speaker
To elicit NSs’ subjective reactions to the speaker and the lecture,
attitudinal questions were selected from the Instructor and Course
Evaluation System (ICES) Item Catalog, which is a pool of Likert-type
items that instructors at the University of Illinois at Urbana-Champaign
regularly use to collect summative feedback from students on their
classroom instruction. ICES items are presented in rating-scale form, and
students respond by selecting the one option on a numerical ﬁve-point
scale that best describes their opinion (Braskamp, Ory, & Brandenburg,
1984). The items for the present study (see Table 4) were selected from
a subset of items on communication skills. They focused on the speaker’s
voice and English-speaking ability (Items 1, 4, 8, 9, and 10), the content
of the lecture (Items 2, 3, and 11), and the general delivery of the
message (Items 5, 6, and 7). The ﬁnal item on this questionnaire was an
open-ended question inviting short-answer comments about the TA.
208 TESOL QUARTERLY
The 90 subjects were randomly assigned to one of three experimental
groups and tested individually. The researcher explained the instruc-
tions and the format of the experiment. Speciﬁc instructions and
practice for the reaction time task were provided on the computer. The
subjects were then informed that they would be listening to a TA and
were then asked to complete the reaction time task, the recall, the quiz,
and the ICES ratings in that order.
The ﬁrst dependent variable under consideration was reaction time. A
histogram of the reaction time data revealed a negatively skewed
distribution, indicating that regardless of experimental conditions, most
subjects’ reaction times were rather long, with just a few very small values.
Because sample statistics are most meaningful when the value distribu-
tion is normal, the reaction times were reciprocally transformed to push
the extreme values toward the center, which distributed the values more
symmetrically. To use whole numbers rather than fractions in the
analyses, the scores were then multiplied by 10,000. The raw data were
also examined to determine whether subjects experienced any particular
difﬁculties completing the clicking task. The number of outliers and
errors in responses (e.g., double-clicking) was so low that no consistency
or pattern could be detected; therefore, the aberrant values were
replaced with the mean of the two neighboring values. Consequently, the
values obtained are best described in terms of speed not reaction time;
that is, a shorter reaction time has a higher speed. Descriptive statistics
and effect sizes for reaction time speed appear in Table 1.
The subjects listening to the nativelike Version A reacted more quickly
(i.e., with higher speed) to the tones than subjects listening to either of
the nonnativelike versions. The difference between mean reaction times
for Versions A and B was greater than that for Versions A and C.
A one-factor analysis of variance (ANOVA) showed no statistically
signiﬁcant differences in reaction time speed among the three groups
[F(2, 87) 1.707, p .19]. (A 0.05 signiﬁcance level was selected for
this two-tailed test and each subsequent comparison.) Although the
differences in the scores were too small to be attributed to the indepen-
dent variable, the large discrepancy in effect size between Versions A and
B suggests that the difference may have some practical signiﬁcance when
considered with the other results.
PRIMARY STRESS AND INTELLIGIBILITY 209
Descriptive Statistics and Effect Sizes: Reaction Time Speeds for Versions A, B, and C
Version N Mean Standard deviation Effect size
A 30 133.54 16.46 0.50 (A-B)
B 30 125.01 18.86 0.15 (B-C)
C 30 127.81 19.21 0.39 (A-C)
Note. Effect sizes describe the discrepancy between two means in terms of the standard deviation
and were computed using Cohen’s d.
Two native-English-speaking graduate students scored the written
recalls using Johnson’s (1970) protocol for scoring and distinguishing
between main ideas and details, and the mean scores were used for
analysis. A Pearson product-moment correlation coefﬁcient was calcu-
lated to determine interrater reliability, which was high (.92). Table 2
contains descriptive statistics and effect sizes comparing the ideas
recalled by subjects in each group. These results indicate that subjects
listening to Version A recalled more main ideas and more details about
the lecture than subjects listening to either of the other two versions.
An ANOVA yielded statistically signiﬁcant differences for total amount
of information recalled [F(2, 87) 7.128; p .001] and for number of
main ideas recalled [F(2, 87) 7.891; p .001] but not for number of
details recalled [F(2, 87) 1.062; p .35]. Homogeneity of variance was
established between groups (Levene’s statistic for total ideas .007, p
.99, and for main ideas .011, p .99). Tukey’s honestly signiﬁcant
difference test was chosen as a posthoc test because it uses the Studentized
range statistic to make all pairwise comparisons between groups; thus, in
effect, it ﬁxes the familywise error rate at alpha against all possible null
hypotheses. Subjects listening to Version A remembered a signiﬁcantly
greater number of ideas than subjects listening to Version B (p .001)
or to Version C (p .02). Subjects listening to Version A remembered
signiﬁcantly more main ideas than subjects listening to Version B (p
.001) or to Version C (p .05).
The comprehension quizzes were scored by two native-English-speaking
graduate students; interrater reliability was again high (Pearson’s r
.94). Table 3 contains descriptive statistics and effect sizes for the quiz
results. Although the mean score of subjects listening to Version A was
approximately one point higher than the mean scores of subjects
listening to either of the other two versions, the effect sizes were small.
210 TESOL QUARTERLY
Descriptive Statistics and Effect Sizes: Main Ideas, Details, and Total Number of Ideas
Recalled from the Text for Versions A, B, and C
Version N Mean Standard deviation Effect size
A 30 13.27 3.78 0.95 (A-B)
B 30 9.45 3.65 0.37 (B-C)
C 30 10.93 3.83 0.58 (A-C)
Version N Mean Standard deviation Effect size
A 30 3.78 1.77 0.29 (A-B)
B 30 3.22 2.16 0.06 (B-C)
C 30 3.37 1.94 0.35 (A-C)
Total number of ideas
Version N Mean Standard deviation Effect size
A 30 17.05 4.58 0.88 (A-B)
B 30 12.63 4.54 0.24 (B-C)
C 30 13.82 4.94 0.65 (A-C)
A one-factor ANOVA of the quiz scores detected no signiﬁcant
difference among groups of subjects [F(2, 87) 1.071, p .35]. Thus it
appears that the version of the lecture did not affect any subject’s
performance on this quiz. However, the groups’ relative performance on
this dependent variable has stayed the same as for the other measures:
Scores were highest for subjects listening to Version A, second-highest
for subjects listening to Version C, and lowest for subjects listening to
Descriptive Statistics and Effect Sizes: Quiz Scores for Versions A, B, and C
Version N Mean Standard deviation Effect size
A 30 10.333 2.79 0.32 (A-B)
B 30 9.317 3.09 0.01 (B-C)
C 30 9.644 3.55 0.33 (A-C)
PRIMARY STRESS AND INTELLIGIBILITY 211
For convenience and consistency in interpretation, the scale for Items
6, 8, and 10 was inverted so that 5 represented the most positive score
and 1 the most negative. Descriptive statistics and effect sizes for the 11
ICES items are reported in Table 4. The subjects who listened to Version
A rated the speaker higher on every item than did the subjects who
listened to Version B or C. Effect sizes indicate that the differences
between Version A and the other two are moderate to substantial. In
comparing the means, Version B received higher ratings than Version C
on eight items (Items 1, 3, 4, 5, 7, 8, 10, and 11), while Version C received
higher ratings than Version B on Items 2 and 6. These two groups had
equal mean ratings on Item 9.
A multivariate ANOVA was conducted with the 11 ICES items as the
dependent variables and the three conditions under which the subjects
heard the text as the factor variables. The purpose was to determine
whether the three groups differed signiﬁcantly in their responses to the
multiple dependent variables. Results (see Table 5) indicated a signiﬁ-
cant overall effect for condition (version of the text) [F(2, 87) 2.442,
Tests of between-subjects effects (see Table 6) examined each item
individually and indicated that all ICES items except one (Item 9) were
signiﬁcantly different depending on the version of the text; Item 9 had a
medium effect size for A compared to B and C. Pairwise comparisons
were then calculated for each item; statistically signiﬁcant comparisons
appear in the ﬁnal column of Table 6. Version A was rated signiﬁcantly
higher than Version B on 10 out of 11 items (all except Item 4), and
signiﬁcantly higher than Version C on 10 out of 11 items (all except Item
6). Furthermore, for the items rated higher in Version B than in Version
C, only Item 8 showed a statistically signiﬁcant difference. In addition,
for Item 6 the mean rating was signiﬁcantly higher for Version C than for
Overall, subjects’ written comments about the speaker in Version A
were the most positive, responses to Version B were more mixed, and
responses to Version C were also mixed but more critical of the TA’s
delivery than either Version A or B. The most salient ﬁnding in the
written comments that was not reﬂected in the ICES items was that 9 of
30 subjects (30%) who listened to the stressless Version C remarked that
the TA spoke too fast. The TA for Version A received no such comment,
and the TA for Version B received only one. Because the pauses, overall
pace, and total speaking time for each version were virtually equal, oral
text without primary stresses apparently conveys the impression that the
speaker is speaking rapidly.
212 TESOL QUARTERLY
Descriptive Statistics and Effect Sizes: Results from ICES Ratings
Version A Version B Version C Effect Sizes
Item Mean S.D. Mean S.D. Mean S.D. A-B B-C A-C
Is the instructor a good
speaker? 3.67 0.76 2.97 1.13 2.80 0.76 0.72 0.18 0.90
How would you characterize
the instructor’s ability
to explain? 4.20 0.55 3.53 0.97 3.57 0.97 0.74 0.04 0.70
How interesting was the
instructor’s presentation? 3.43 0.97 2.70 1.06 2.47 1.25 0.63 0.20 0.83
The instructor emphasized
important points by raising
voice, repeating, etc. 3.13 1.25 2.80 1.13 2.23 1.30 0.26 0.45 0.71
It was easy to hear and
understand the instructor. 3.80 1.10 2.90 1.12 2.60 0.93 0.82 0.28 1.09
The instructor’s lecture
seemed to ramble. 3.93 0.98 3.07 1.41 3.67 1.03 0.72 0.50 0.58
presentation would allow
for easy note-taking. 4.20 1.03 3.57 1.25 3.53 1.07 0.55 0.03 0.58
The instructor spoke in a
monotone, rarely showing
expression in voice. 3.03 1.40 2.40 1.16 1.63 0.89 0.49 0.60 1.08
Was the instructor’s voice
pleasant or irritating to
listen to? 3.43 0.68 2.97 1.07 2.97 0.93 0.50 0.00 0.50
The instructor’s lack of
facility with English
of ideas. 4.03 0.76 3.60 1.19 3.27 1.28 0.65 0.28 0.92
Was the lecture easy to
understand? 4.37 0.76 3.60 1.19 3.27 1.28 0.65 0.28 0.92
In this study, subjects generally responded more positively to NNS
discourse when the speaker maintained rather than violated the GNSC;
these ﬁndings are congruent with the related literature. First, subjects
listening to Version A with its correct primary stresses responded
somewhat more quickly to the randomized tones than subjects listening
to either Version B or C, which did not have correctly placed primary
PRIMARY STRESS AND INTELLIGIBILITY 213
Multivariate Test Comparing ICES Item Ratings by Condition (Version of Text)
Value F Hypoth. df Error dif. Sig. Eta squared Noncent. parameter Observed power
Wilks’s lambda .541 2.250 22 156 .001 .256 55.442 .998
Tests of Between-Subjects Effects and Pairwise Results for ICES Items
Eta Observed Significant group
Source Item # SS df MS F Sig. squared power differences
Corrected model 1 12.689 2 6.344 7.837 .001 .153 .946 A>B, A>C
2 8.467 2 4.233 5.788 .004 .117 .859 A>B, A>C
3 15.267 2 7.633 6.317 .003 .127 .889 A>B, A>C
4 12.422 2 6.211 4.105 .020 .086 .714 A>C
5 23.400 2 11.700 10.526 .000 .195 .987 A>B, A>C
6 11.822 2 5.911 4.418 .015 .092 .747 A>B, C>B
7 8.467 2 4.233 3.359 .039 .072 .620 A>B, B>C
8 29.489 2 14.744 10.767 .000 .198 .988 A>B>C
9 4.356 2 2.178 2.657 .076 .058 .515 A>B, A>C
10 39.200 2 19.600 13.830 .000 .241 .998 A>B, A>C
11 19.089 2 9.544 7.831 .001 .153 .946 A>B, A>C
stress. Although this ﬁnding was not statistically signiﬁcant, the effect size
suggests some practical evidence that subjects processed Version A with
slightly less effort. Second, the mean scores for each experimental group
on the recall task and the comprehension quiz followed the same pattern
as the mean scores for reaction time speed: Group A scored higher than
Group C, which in turn scored higher than Group B. This result suggests
that correctly placed primary stress enhances comprehension. Misplaced
primary stress may also impede comprehension more than no primary
stress, but the data are not conclusive. Additional corroborative evi-
dence, however, might show that processing Version B—(subconsciously)
recognizing stress anomalies and backtracking to ﬁnd the real new
information—is psycholinguistically more challenging than processing
Version C—simply looking for new information without stress clues.
Third, analyses of the attitudinal ICES data indicate that NSs tend to rate
NNSs who use primary stress more positively than NNSs who either
misplace primary stress or do not use it at all. Thus, this pronunciation
feature has a clear impact on listeners’ perceptions of a speaker’s ability
The differences among the three groups of subjects were statistically
signiﬁcant for the recall (comprehension) data and nearly all of the ICES
items. Although the differences in reaction time speed and the compre-
hension quiz scores were not statistically signiﬁcant, the results of those
measurements follow the pattern found in the other data. This consistent
pattern in the results supports the general proposition that correct
primary stress in extended nonnative discourse facilitates communication.
Implications for ITA Programs and Curricula
The results provide evidence that undergraduate students appear to
understand and respond more positively to an ITA when primary stress is
“doing its job” than when it is not. It therefore seems logical that the ITA
curriculum should include instruction in primary stress. However, opin-
ions regarding the role of pronunciation instruction in ITA programs
have diverged considerably (e.g., Briggs, Hyon, Aldridge, & Swales, 1990;
Hinofotis & Bailey, 1981; Hoekje & Williams, 1992; Smith, Byrd, Nelson,
Barrett, & Constantinides, 1992). For example, Hoekje and Williams (1992)
argue that ITA programs should focus on sociolinguistics, discourse, and
“compensatory” strategies, and they suggest that pronunciation instruc-
tion may not be effective beyond providing a basic level of proﬁciency and
correct pronunciation of academic terms. Yet this view neglects the
contributions of pronunciation, and suprasegmentals in particular, to
different aspects of communicative competence.
The GNSC, as an important source of cohesion in spoken English
PRIMARY STRESS AND INTELLIGIBILITY 215
discourse, is essential to discourse competence—and perhaps to an
overall perception of communicative competence. The following com-
ments from subjects in this study reveal some of the connections they
made in this regard:
Version A: “It was an enjoyable speech! Pleasant voice, used good inﬂection
and tone, and I had no trouble understanding what was said.”
Version B: “The TA didn’t change tone. He was difﬁcult and confusing. I was
Version C: “The TA needs to change the tone of his voice to have a more
lasting impact. . . . It was one big ﬂow.”
Other connections between the sound system and communicative com-
petence can be identiﬁed. For example, Chun (1988) acknowledges the
role of intonation in discourse competence:
Intonation is fundamental to genuine communication because communica-
tive competence is the ability not only to formulate grammatically correct
utterances, but also to signal interactional strategies, such as interrupting,
asking for clariﬁcation, taking the ﬂoor, changing the subject, concluding an
argument, or constraining a hearer to reply. (p. 295)
Suprasegmentals also contribute to sociolinguistic competence. Chun
suggests that intonation contributes to our interpretations of utterances
as polite or impolite and that it marks the role of the participants’
relationships in a conversation. In addition, the use of linking and
blending devices in connected speech depends on the register as well as
the situation’s degree of formality (Anderson-Hsieh, Riney, & Koehler,
1994; Lass, 1984; Temperley, 1987).
Pronunciation is not by any means the only key to ITA effectiveness.
However, a broad perspective that acknowledges the various roles that
pronunciation features play in communicating meaning in discourse
would enhance ITA programs and materials. Of course, these implica-
tions would also be relevant for other ESL learners such as health-care
workers, business professionals, and other NNSs who participate in
extended oral discourse.
Implications for Pronunciation Pedagogy and Materials
Empirical evidence strengthens pedagogical claims about the impor-
tance of teaching suprasegmentals. The results of this study complement
current literature advocating the inclusion of primary stress in an ESL
216 TESOL QUARTERLY
pronunciation curriculum. For example, Jenkins (2002) includes pri-
mary stress in her Lingua Franca Core—a set of phonological features
that “seem to be crucial as safeguards of mutual intelligibility in
interlanguage talk” (p. 96), indicating that primary stress is an important
feature for NNSs who interact with other NNSs as well as with NSs.
Fortunately, primary stress can be learned. Pennington and Ellis (2000)
found that learners who received explicit instruction signiﬁcantly im-
proved primary stress production, reinforcing Dalton and Seidlhofer’s
(1994) argument that primary stress is both teachable and valuable for
communicating. Furthermore, Derwing, Munro, and Wiebe (1998)
found that ESL students who received 20 minutes per day of instruction
in suprasegmentals (including primary stress) signiﬁcantly improved
comprehension and ﬂuency in a narrative speaking task.
Many pronunciation textbooks (e.g., Dauer, 1993; Dickerson, 1989;
Gilbert, 1984; Grant, 2000; Hagen & Grogan, 1992; Miller, 2000) do
teach at least the basics of primary stress, including its role in signaling
old and new information. Results from this study support the rationale
for using such materials. Speciﬁc strategies for teaching primary stress in
the ESL/EFL classroom that follow from this study include
• Integrating relevant suprasegmental topics with other course mate-
rial focusing on sociolinguistic and discourse competence. A general
oral skills class could, for example, address contrastive stress when
practicing functions such as disagreeing or contradicting, or con-
tracting and blending when practicing informal conversations such
as small talk. Additional ideas for such integration can be found in
Levis and Grant (2003).
• Telling students that maintaining the GNSC facilitates communica-
tion. A rationale that points to improved intelligibility may motivate
students’ learning. Examples of both high-stakes and low-stakes
miscommunications could be used to emphasize primary stress’s
• Including perception exercises for primary stress. For example,
students could practice identifying the primary stress in recordings
of spoken discourse. Comparing speech samples with correct, mis-
placed, and missing primary stress may also help learners perceive
primary stress and its meaning.
• Using longer pieces of discourse to allow students to practice
primary stress. Most pronunciation textbooks provide short, limited
contexts for practicing primary stress. Practicing the GNSC requires
dialogues and passages beyond two or three sentences. Focused
practice in correctly producing and placing primary stress in stu-
dents’ own discourse (conversations and oral presentations) would
also be valuable.
PRIMARY STRESS AND INTELLIGIBILITY 217
• Teaching students how to destress old information. In this study,
stressing old information appearing at the end of a message unit
made the discourse more difﬁcult to understand. As indicated
earlier, NNSs often misplace the primary stress in this way (as in the
sentence “So I’m an indivídualist. But I’m not always an indivídual-
ist.”). Although many textbooks explain how and why the word
always would receive primary stress, they do not provide guidelines
on how and why the second individualist is destressed. NNSs need to
learn that keeping old information at a lower pitch provides an
important contrast between the stressed and unstressed words that
facilitates communicating the discourse’s meaning.
This research provides evidence that primary stress contributes signiﬁ-
cantly to the intelligibility of nonnative discourse, and it strengthens the
broadly stated claims in the pedagogical literature on ESL pronunciation
that teaching suprasegmentals is important. However, further studies are
needed to corroborate or contradict these ﬁndings. For example, par-
allel studies could investigate NSs’ reactions to nonnative primary stress
with other speaker accents or with different lengths and styles of oral
texts. Other measures of intelligibility should also be explored and
validated. Similar studies of other linguistic variables such as intonation,
word stress, and various segmental features would also help us to more
fully understand intelligibility and NS reactions to nonnative pronuncia-
tion. Such insights can enhance our understanding of prosodic features
in English, help us set pedagogical priorities, and provide guidelines for
helping ITAs and other ESL learners achieve communicative compe-
tence in English.
Laura D. Hahn teaches courses in applied English phonology and ESL/EFL
methodology at the University of Illinois at Urbana-Champaign and at Wheaton
College. She has taught numerous pronunciation courses and is chair-elect of
TESOL’s Speech, Pronunciation, and Listening Interest Section. Her research
interests include suprasegmentals, teacher preparation, and intelligibility.
Adams, C. (1979). English speech rhythm and the foreign learner. The Hague: Mouton.
Albrechtsen, D., Henriksen, B., & Faerch, C. (1980). Native speaker reactions to
learners’ spoken interlanguage. Language Learning, 30, 365–396.
Anderson, P. (1993). The interstress interval as an indicator of perceived intelligibil-
218 TESOL QUARTERLY
ity among nonnative speakers of English. (Doctoral dissertation, Wichita State
University, 1993.) ProQuest Dissertation Abstracts, AAC 9325699.
Anderson-Hsieh, J., Riney, T., & Koehler, K. (1994). Connected speech modiﬁcations
in the English of Japanese learners. Issues and Developments in English and Applied
Linguistics, 7, 31–52.
Anisfeld, M., Bogo, N., & Lambert, W. (1962). Evaluational reactions to accented
English speech. Journal of Abnormal and Social Psychology, 65(4), 223–231.
Bansal, R. (1969). The intelligibility of Indian English. Hyderabad, India: Central
Institute of English.
Bardovi-Harlig, K. (1986). Pragmatic determinants of English sentence stress. Bloomington,
IN: Indiana University Linguistics Club.
Birch, S., & Clifton, C. (1995). Focus, accent, and argument structure: Effects on
language comprehension. Language and Speech, 38(4), 365–391.
Bock, J. K., & Mazzella, J. (1983). Intonational marking of given and new informa-
tion: Some consequences for comprehension. Memory & Cognition, 11(1), 64–76.
Bolinger, D. (1972). Accent is predictable—if you’re a mind-reader. Language, 48,
Bolinger, D. (1986). Intonation and its parts. Stanford, CA: Stanford University Press.
Braskamp, L., Ory, J., & Brandenburg, D. (1984). Evaluating teaching effectiveness: A
practical guide. Newbury Park, CA: Sage Publications.
Brazil, D., Coulthard, M., & Johns, C. (1980). Discourse intonation and language
teaching. London: Longman Group.
Briggs, S., Hyon, S., Aldridge, P., & Swales, J. (1990). The international teaching
assistant: An annotated critical bibliography. Ann Arbor: The English Language
Institute, The University of Michigan.
Brown, A. (1995). Minimal pairs: Minimal importance? ELT Journal, 49(2), 169–175.
Brown, G. (1983). Prosodic structure and the given/new distinction. In A. Cutler &
D. R. Ladd (Eds.), Prosody: Models and measurements (pp. 67–77). Berlin: Springer-
Brown, K. (1988). Effects of perceived country of origin, educational status, and native
speakerness on American college student attitudes toward non-native instructors. Unpub-
lished doctoral dissertation, University of Minnesota (UMI No. AAT 8815268).
Celce-Murcia, M., Brinton, D., & Goodwin, J. (1996). Teaching pronunciation: A
reference for teachers of English to speakers of other languages. Great Britain: Oxford
Chafe, W. (1976). Givenness, contrastiveness, deﬁniteness, subjects, topics, and point
of view. In C. Li (Ed.), Subject and topic (pp. 27–55). New York: Academic.
Chela-Flores, B. (1998). Teaching English rhythm: From theory to practice. Caracas,
Venezuela: Fondo Editorial Tropykos.
Chun, D. (1988). The neglected role of intonation in communicative competence
and proﬁciency. The Modern Language Journal, 72(3), 295–302.
Clark, H., & Haviland, S. (1977). Comprehension and the given-new contract. In
R. Freedle (Ed.), Discourse production and comprehension: Vol. 1 (pp. 1–40). Norwood,
Clennell, C. (1996). Promoting the role of English prosody in a discourse-based
approach to oral interaction. Prospect, 11(3), 17–28.
Constantinou, A. (1993). Intelligibility of Mandarin speakers of English: Correlation
of acoustic and perceptual measures. (Masters thesis, California State University,
Long Beach). ProQuest Dissertation Abstracts, AAC 1351387.
Cruttenden, A. (1997). Intonation. Cambridge: Cambridge University Press.
Crystal, D. (1969). Prosodic systems and intonation in English. Cambridge: Cambridge
Cutler, A. (1984). Stress and accent in language production and understanding. In
PRIMARY STRESS AND INTELLIGIBILITY 219
D. Gibbon & H. Richter (Eds.), Intonation, accent, and rhythm: Studies in discourse
phonology. Berlin: Walter de Gruyter.
Dalton, C., & Seidlhofer, B. (1994). Pronunciation. Oxford, UK: Oxford University
Dauer, R. (1993). Accurate English. Englewood Cliffs, NJ: Regents/Prentice-Hall.
de Bot, K. (1986). The transfer of intonation and the missing data base. In E. Keller-
man & M. Sharwood Smith (Eds.), Crosslinguistic inﬂuence in second language
acquisition (pp. 111–119). New York: Pergamon Institute of English.
Derwing, T., Munro, M., & Wiebe, G. (1998). Evidence in factor of a broad
framework for pronunciation instruction. Language Learning, 48(3), 393–410.
Dickerson, W. (1989). Stress in the speech stream: The rhythm of spoken English. Urbana,
IL: University of Illinois Press.
Eefting, W. (1992). The effect of accentuation and word duration on the naturalness
of speech. The Journal of the Acoustical Society of America, 91(1), 411–420.
Eisenstein, M. (1983). Native reactions to non-native speech: A review of empirical
research. Studies in Second Language Acquisition, 5(2), 160–176.
Fayer, J., & Krasinski, E. (1987). Native and nonnative judgments of intelligibility and
irritation. Language Learning, 37(3), 313–326.
Flege, J. (1984). The detection of French accent by American listeners. The Journal of
the Acoustical Society of America, 76(3), 692–707.
Fowler, C., & Housum, J. (1987). Talkers’ signaling of “new” and “old” words in
speech and listeners’ perception and use of the distinction. Journal of Memory and
Language, 26, 489–504.
Gallego, J. C. (1990). The intelligibility of three nonnative English speaking teaching
assistants: An analysis of student-reported communication breakdowns. Issues in
Applied Linguistics, 1(2), 219–237.
Gilbert, J. (1984). Clear speech. Cambridge: Cambridge University Press.
Grant, L. (2000). Well said: Advanced English pronunciation. Boston: Heinle & Heinle.
Gunter, R. (1974). Sentences in dialog. Columbia, SC: Hornbeam.
Gussenhoven, C. (1983). Focus, mode and the nucleus. Journal of Linguistics, 19, 377–
Gussenhoven, C. (1985). Two views of accent: A reply. Journal of Linguistics, 21, 125–
Hagen, S., & Grogan, P. (1992). Sound advantage: A pronunciation book. Englewood
Cliffs, NJ: Prentice Hall/Regents.
Hahn, L. (1999). Native speakers’ reactions to non-native speech in English discourse.
Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign
(UMI No. AAT 9944870).
Halliday, M. A. K. (1967a). Intonation and grammar in British English. Janua
Linguarum. The Hague: Mouton.
Halliday, M. A. K. (1967b). Notes on transitivity and theme in English: Part 2. Journal
of Linguistics, 3, 177–274.
Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. New York: Longman.
Hecker, M., Stevens, K., & Williams, C. (1966). Measurements of reaction time in
intelligibility tests. The Journal of the Acoustical Society of America, 39(6), 1188–1189.
Hinofotis, F., & Bailey, K. (1981). American undergraduates’ reactions to the
communication skills of foreign teaching assistants. In J. Fisher, M. Clarke, &
J. Schacter (Eds.), On TESOL ’80: Building bridges: Research and practice in teaching
English as a second language (pp. 120–136). Washington, DC: TESOL.
Hoekje, B., & Williams, J. (1992). Communicative competence and the dilemma of
international teaching assistant education. TESOL Quarterly, 26(2), 243–270.
220 TESOL QUARTERLY
Jenkins, J. (2000). The phonology of English as an international language. Oxford: Oxford
Jenkins, J. (2002). A sociolinguistically based, empirically researched pronunciation
syllabus for English as an international language. Applied Linguistics, 23(1), 83–
Johnson, R. (1970). Recall of prose as a function of the structural importance of the
linguistic units. Journal of Verbal Learning and Behavior, 9, 12–20.
Juffs, A. (1990). Tone, syllable structure and interlanguage phonology: Chinese
learners’ stress errors. International Review of Applied Linguistics, 28(2), 99–117.
Ladd, D. R., & Cutler, A. (1983). Models and measurements in the study of prosody.
In A. Cutler & D. R. Ladd (Eds.), Prosody: Models and measurements (pp. 77–85).
Lass, R. (1984). Phonology. Cambridge: Cambridge University Press.
Lehiste, L. (1969). Suprasegmentals. Cambridge, MA: M.I.T. Press.
Levis, J., & Grant, L. (2003). Integrating pronunciation into ESL/EFL classrooms.
TESOL Journal, 12(2), 13–19.
Lieberman, P. (1967). Intonation, perception, and language. MIT Research Monograph
No. 38. Cambridge, MA: M.I.T. Press.
Ludwig, J. (1982). Native-speaker judgments of second-language learners’ effort at
communication: A review. The Modern Language Journal, 66(3), 274–283.
McNerney, M., & Mendelsohn, D. (1992). Suprasegmentals in the pronunciation
class: Setting priorities. In P. Avery & S. Ehrlich (Eds.), Teaching American English
pronunciation (pp. 185–196). Oxford: Oxford University Press.
Miller, S. (2000). Targeting pronunciation. Boston: Houghton Mifﬂin.
Mochizuki-Sudo, M., & Kiritani, S. (1991). Production and perception of stress-
related durational patterns in Japanese learners of English. Journal of Phonetics, 19,
Morley, J. (1991). The pronunciation component in teaching English to speakers of
other languages. TESOL Quarterly, 25(3), 481–520.
Mulac, A., Hanley, T., & Prigge, D. (1974). Effects of phonological speech foreignness
upon three dimensions of attitude of selected American listeners. The Quarterly
Journal of Speech, 60(4), 411–420.
Munro, M., & Derwing, T. (1995a). Foreign accent, comprehensibility, and intelligi-
bility in the speech of second language learners. Language Learning, 45(1), 73–97.
Munro, M., & Derwing, T. (1995b). Processing time, accent, and comprehensibility
in the perception of native and foreign-accented speech. Language and Speech,
Munro, M., & Derwing, T. (1999). Foreign accent, comprehensibility, and intelligibil-
ity in the speech of second language learners. Language Learning, 49(supp. 1),
Nash, R. (1971). Phonemic and prosodic interference and intelligibility. In A. Rigault
(Ed.), Proceedings of the seventh international congress of phonetic sciences (pp. 570–
573). The Hague: Mouton.
Nelson, C. (1983). Intelligibility and non-native varieties of English. In B. Kachru
(Ed.), The other tongue: English across cultures (pp. 58–73). New York: Pergamon.
Nida, E. (1957). Learning a foreign language. Ann Arbor: Friendship.
Orth, J. (1982). University undergraduate evaluational reactions to the speech of foreign
teaching assistants. Unpublished doctoral dissertation, University of Texas at Austin
(UMI No. AAT 8309183).
Palmer, L. (1973). A preliminary report on a study of the linguistic correlates of
raters’ subjective judgments of non-native speech. In R. Shuy & R. Fasold (Eds.),
PRIMARY STRESS AND INTELLIGIBILITY 221
Language attitudes: Current trends and prospects (pp. 41–59). Washington, DC:
Georgetown University Press.
Pennington, M. (1994). Recent research in L2 phonology: Implications for practice.
In J. Morley (Ed.), Pronunciation pedagogy and theory: New views, new directions (pp.
64–91). Washington, DC: TESOL.
Pennington, M., & Ellis, N. (2000). Cantonese speakers’ memory for English
sentences with prosodic cues. The Modern Language Journal, 84(3), 372–389.
Pennington, M., & Richards, J. (1986). Pronunciation revisited. TESOL Quarterly,
Pickering, L. (2001). The role of tone choice in improving ITA communication in
the classroom. TESOL Quarterly, 35(2), 233–253.
Raisler, I. (1976). Differential response to the same message delivered by native and
foreign speakers. Foreign Language Annals, 9, 256–259.
Ryan, E. (1983). Social psychological mechanisms underlying native-speaker evalua-
tions of non-native speech. Studies in Second Language Acquisition, 5(2), 48–159.
Schmerling, S. (1976). Aspects of English sentence stress. Austin, TX: University of Texas
Schmidt-Nielsen, A., Kallman, H., & Meijer, C. (1990). Dual-task performance using
degraded speech in a sentence-veriﬁcation task. Bulletin of the Psychonomic Society,
Smith, L., & Bisazza, J. (1982). The comprehensibility of three varieties of English for
college students in seven countries. Language Learning, 32, 259–269.
Smith, L., & Raﬁqzad, K. (1979). English for cross-cultural communication: The
question of intelligibility. TESOL Quarterly, 13, 371–380.
Smith, R., Byrd, P., Nelson, G., Barrett, R., & Constantinides, J. (1992). Crossing
pedagogical oceans: International teaching assistants in U.S. undergraduate education.
ASHE-ERIC Higher Education Report No. 8. Washington, DC: The George
Washington University, School of Education and Human Development.
Sperber, D., & Wilson, D. (1986). Relevance: Communication and cognition. Cambridge,
MA: Harvard University Press.
Temperley, M. (1987). Linking and deletion in ﬁnal consonant clusters. In J. Morely
(Ed.), Current perspectives on pronunciation: Practices anchored in theory (pp. 59–82).
Washington, DC: TESOL.
Terken, J., & Hirschberg, J. (1994). Deaccentuation of words representing ‘given’
information: Effects of persistence of grammatical function and surface position.
Language and Speech, 37(2), 125–145.
Triandis, H. (1990, April). Individualism and collectivism: Implications for communication.
Paper presented at the colloquium of the Division of English as an International
Language, University of Illinois at Urbana-Champaign.
Tun, P., Wingﬁeld, A., & Stine, E. (1991). Speech-processing capacity in young and
older adults: A dual-task study. Psychology and Aging, 6(1), 3–9.
Tyler, A., Jeffries, A., & Davies, C. (1988). The effect of discourse structuring devices
on listener perceptions of coherence in non-native university teachers’ spoken
discourse. World Englishes, 7(2), 101–110.
Varonis, E., & Gass, S. (1982). The comprehensibility of nonnative speech. Studies in
Second Language Acquisition, 4, 114–126.
Watanabe, K. (1988). Sentence stress perception by Japanese students. Journal of
Phonetics, 16, 181–186.
Webster, W., & Kramer, E. (1968). Attitudes and evaluational reactions to accented
English speech. The Journal of Social Psychology, 75, 231–240.
Wells, W. H. G. (1986). An experimental approach to the interpretation of focus in
spoken English. In C. Johns-Lewis (Ed.), Intonation in discourse (pp. 53–75).
London: Croom Helm.
222 TESOL QUARTERLY
Wenk, B. (1985). Speech rhythms in second language acquisition. Language and
Speech, 28(2), 157–175.
Wennerstrom, A. (1994). Intonational meaning in English discourse: A study of non-
native speakers. Applied Linguistics, 15(4), 399–420.
Sample of ITA Lecture
I will start by deﬁning the topic for today, | which is individualism and collectivism.
Individualism concerns the placing of personal goals ahead of group goals. And collectivism
concerns placing group goals ahead of personal goals. So let’s suppose you have a conﬂict at
work about break time. Let’s say your co-workers want longer breaks, | but you want shorter
breaks. If you’re a collectivist, you’ll give in to the group. But if you’re an individualist, you’ll go
against the group.
First of all, there are many determinants of individualism and collectivism. Culture is a
determinant, | but it’s only one of the determinants. But let me start with culture. Basically, the
European cultures, | particularly those in northwestern Europe, | are highly individualistic.
England is in northwest Europe, | and it’s typical of the individualist pattern. And the East Asian
cultures, | such as China and Japan, | are much more typical of the collectivist pattern. But in
between, you have different combinations of the patterns. And I’ll discuss that in a minute. But
let me mention some other determinants | of individualism and collectivism.
One determinant that’s very important is social class. There’s a tendency for the upper
classes to be more individualistic than the lower classes. In other words, people at the top of a
social structure | are more likely to think and behave like an individualist | than those near the
bottom of the structure. For example, if you look at the history of China, | the emperor of China
is more individualistic than the working class in China. So certain classes in a culture | may be
more individualistic than the entire population in a culture.
PRIMARY STRESS AND INTELLIGIBILITY 223