Flege, J. (1995). Two methods for training a novel second-language phonetic contrast. Applied
Psycholinguistics, 16, 425-442.
Flege, J., Munro, M., & MacKay, I. (1995). Factors affecting degree of perceived foreign accent in a second
language. Journal of the Acoustical Society of America, 97, 3125-3134.
Flege, J., Munro, M. & MacKay, I. (1995). The effect of age of second language learning on the production of
English consonants. Speech Communication, 16, 1-26.
Flege, J. & Schmidt, A. (1995). Native speakers of Spanish show rate-dependent
processing of English stop consonants. Phonetica, 52, 90-111.
Schmidt, A. & Flege, J. (1995). Effects of speaking rate changes on native and non-native production.
Phonetica, 52, 41-54.
|Articles published in the period 1990-1995
|Flege, J., Takagi, N. & Mann, V. (1995). Japanese adults can learn to produce English
/r/ and /l/ accurately. Language and Speech, 38, 25-56.
Note: a companion piece published in 1996 examined the perception of English /r/ and
/l/ by the same experienced and inexperienced native Japanese participants
Abstract. This study examined word-initial productions of English /r/ and /l/. It was the
first study to compare native Japanese (NJ) adults having extensive conversational
experience in English (M = 20.8 years of residence in the U.S., range = 12-29) to NJ Ss
with relatively little such experience (M = 1.6 years, range = 0.7-3). The production of
English liquids by the NJ Ss and a native English (NE) comparison was elicited using a
"Definition" task which obviated the unwanted influence of orthography, by having Ss
read a word lis the "Reading" task), and by having they extemporaneously produce
sentences containing target words ("Spontaneous"). In four experiments the elicited
productions of /r/ and /l/ were evaluated by NE speaking listeners. To ensure
generalizabity to other NE speaking listeners, between group differences were
considered significant only if that held true in talker-based and listener-based ANOVAs.
Also, a subset of /r/ and /l/ tokens were examined acoustically.
The results indicated that many of the 12 experienced NJ Ss produced English liquids
as well as NE speakers did (scores falling within +/- 2 SDs of the NE mean), and a few
of the 12 inexperienced NJ Ss did so as well. The acoustic analyses revealed that when
production by the NJ Ss differed from those of NE speakers, it was due to an F2
Exp. 1. Examined production of 38 words (e.g., ride, lied, room, loom) that were read
from a list. The word endings were removed to enable listeners to focus exclusively on
the word-initial liquids. The resulting 1,368 CV stimuli were presented to 10 NE listeners
who classified the randomly presented CVs in a 6 alternative forced choice test. The
response alternatives were: 1-definitely L, 2-probably L, 3-possibly L, 4-possibly R, 5-
probably R, 6-defintely R. Analysis focused on simple intelligibility (responses 1-3 for /l/,
4-6 for /r/, and “weighted intelligibilty”, which reflected listener’s degree of certainty in
the classification of tokens as /r/ or /l/. Scores were significantly higher for the NE and
experienced Japanese (EJ) Ss than for the relatively inexperienced Japanese (IJ) Ss
indicating that additional experience in English augments performance by NJ learners of
Exp. 2. Ten words from the word-list reading experiment were examined: read-lead, root-
loot, rate-late, road-load, rot-lot. The /r/ and /l/ tokens were presented in separate
blocks for ratings on an EAI scale ranging from 1-strong foreign accent to 7-no foreign
accent. All three groups differed significantly, indicating that the EJ Ss as well as the IJ
Ss differed from the NE Ss. Unlike the results of Exp 1, this suggested that as a group,
the EJ Ss did not produce English liquids in a completely English like way. However, an
examination of data for individual Ss revealed that some NJ Ss obtained ratings that
falling within the NE range. For /r/ this held true for 7 of the 12 EJ Ss and for 1 of the 12
IJ Ss. For /l/ this held true for 8 EJ and 3 IJ Ss.
Abstract. Four experiments, all focussing on vowel duration, assessed Chinese subjects' production and
perception of the contrast between /t/ and /d/ in the final position of English words. Vowel duration was
measured in minimal pairs in Exp. 1. In Exps. 2-4, the stimuli in natural-edited beat-bead and bat-bad continua
(in which vowel duration varied in 20-ms steps) were presented to native English and Chinese subjects in a
forced-choice test, in an experiment using the method of adjustment, and in an imitation task.
The non-natives who learned English in childhood closely resembled native speakers in all four experiments.
Three groups of non-natives who had learned English in adulthood, on the other hand, differed from the native
English (NE) comparison group. These late learners produced significantly longer vowels in words ending in /d/
than /t/. However, their vowel duration differences were much smaller than the native English speakers', and
were correlated significantly with degree of foreign accent in English. The late learners differed from the NE
speakers in several ways in the two perception experiments and imitation task.
The pattern of significant and non-significant between-group differences were consistent with the hypothesis
that second language speech production accuracy is limited by the adequacy of perceptual representations for
L2 sounds. However, the data for individual subjects did not always conform to the expected pattern.
retroflex /t/. Journal of Phonetics, 19, 213-229.
Moeini, M., Flege, J. & McCutcheon, M. (1990). The design of a microcomputer controlled voice onset time
analyzer. Biomedical Instrumentation and Technology, 24, 357-362.
Bohn, O.-S. & Flege, J. (1990). Interlingual identification and the role of foreign language experience in L2
vowel perception. Applied Psycholinguistics, 11, 303-328.
Quarterly Journal of Experimental Psychology, 43A, 701-731.
Abstract. When someone who is learning a second language (L2) produces a sound in the L2 using a familiar,
native-language (L1) category, the L2 sound is said to have been "identified with" an L1 sound. Although
interlingual identification exerts a powerful influence on L2 pronunciation, it is still poorly understgood.
Orthographic classification was used here to assess the interlingual identification of Spanish and English
vowels. Sixty native speakers of Spanish in three experiments judged the vowels found in the English words
beat, bit, bet, bat (spoken by 10 native speakers of American English). The Ss labelled each English vowel by
circling one of the 5 letters used to spell the vowel in Spanish (viz. <i,e,a,o,u>) or by circling "none" if they
thought they had heard a vowel not found in Spanish.
Ss who spoke English as an L2 used the "none" label more often than Spanish monolinguals did, suggesting
that learning an L2 heightens bilinguals' awareness of cross-language phonetic differences. Exerienced
Spanish speakers of English used "none" more often than inexperienced Ss did (M = 42% vs. 18%). A few Ss
used the "none" label consistently for the vowels in bat and bit, suggesting that they may have regarded these
vowels as "new" (i.e., non-Spanish). However, the group data provided little support for the hypothesis that
adult Spanish learners of English treat either vowel as new. The great majority of Ss, even those who were
experienced in English, identified English the vowel in bat with their Spanish /a/
produced in a second language. Journal of the. Acoustical Society of America, 89, 395-411.
study was motivated by an unexpected earlier finding obtained for early native Spanish learners of English.
Flege & Eefting (1987) examined two groups of native Spanish adults who learned English as young children
(i.e., early learners). The members of both groups were living in Puerto Rico when tested. Both groups of
participants produced English /p t k/ with significantly shorter VOT values than did age-matched monolingual
native speakers of English.
The aim of this this 1991 study was to determine if indeed "pre-critical period" learners are unable to accurately
produce L2 stop consonants that differ from those found in the L1.
input from native speakers of English and used English frequently. As expected, the early learners were fully
Special care was taken in selecting participants to ensure that the early learners in this study had received
native-like in their production. Also as expected, the average VOT values obtained for the late learner group
was about mid-way between the average VOT values observed in the production of monolingual native Special
care was taken in selecting participants to ensure that the early learners in this study had received speakers of
Spanish and English. It is important to noted, however, variation existed among individual late learners, some
being more "Englishlike" in their production of English stops than others.
Flege, J., Munro, M., & Fox, R. (1994). Auditory and categorical affects on cross-language vowel perception.
Journal of the Acoustical Society of America, 95, 3623-3641.
Flege, J. & Munro, M. (1994). The word unit in L2 speech production and perception.
Studies in Second Language Acquisition, 16, 381-411.
contrast. Journal of the Acoustical Society of America, 93, 1589-1608.
Bohn, O.-S. & Flege, J. (1993). Perceptual switching in Spanish/English bilinguals:
Evidence for universal factors in stop voicing judgments. Journal of Phonetics, 21, 267-290.
Fox, R., Flege, J., & Munro, M. (1995). The perception of English and Spanish vowels by native English and
Spanish listeners: A multidimensional scaling analysis. Journal of the Acoustical Society of America, 97, 2540-
Abstract. Groups of native Spanish- and English-speaking listeners participated in a multidimensional scaling
(MDS) study examining vowel tokens drawn from Spanish (/i e a/) and English (/i ɪ eɪ ɛ ˄ æ ɑ/). The naturally
produced vowel stimuli were spoken by three monolinguals (Spanish or English). All possible pairing of the
vowels (n = 405 after excluding pairs consisting of two tokens of the same vowel category) were presented to
30 English monolinguals and 30 native Spanish speakers of English L2. The adult listeners rated the members
of each vowel pair using a 9-point dissimilarity scale. The perceptual distances between the ten vowels
examined were then analyzed using the individual-differences version of ALSCAL. Results demonstrated that
the English monolinguals used three underlying dimensions in rating vowels while the Spanish-English
bilinguals used just two.
The most salient perceptual dimension for both groups distinguished vowel height. However, for the English
listeners, this dimension was most significantly correlated with duration and indicated a language-dependent
sensitivity to this phonetic feature. The second dimension for the English listeners represented a front-back
distinction, while the third reflected a central/non-central distinction. For the Spanish listeners, the second
dimension was less easily interpreted. However, the perceptual data for the Spanish listeners was more
interpretable in terms of the distribution of the vowels in the two-dimensional perceptual plane. The vowels were
distributed in terms of three separate vowel clusters, each cluster near the location of a Spanish vowel.
Separate MDS analyses were carried out for subgroups of Spanish listeners who were relatively proficient or
non-proficient in English. The vowel space of the proficient Spanish listeners was more English-like than that of
the non-proficient Spanish listeners, suggesting that the perceptual dimensions used by listeners in identifying
vowels may gradually change as proficiency in the second language increases.
Keywords: English, Spanish, vowel perception, dimensions, MDS, multi-dimensional scaling, L2 experience,
duration, spectral quality
perceive the word-final English /t/-/d/ contrast. Journal of Phonetics, 17, 299-315.
Abstract. A contrast between /t/ and Id/ exists in the initial but not the final position of Chinese words. This
study examined identification of English word-final /t/ and /d/ tokens from which closure voicing and release
burst cues had been removed. The performance of three Chinese groups was compared before, during, and
after feedback training. The Cantonese subjects were expected to perform best because their Ll permits
unreleased /p t k/ in word-final position, and the Mandarin subjects were expected to perform most poorly
because their Ll permits no word·final obstruents. An intermediate level of performance was expected from the
Shanghainese subjects, whose Ll permits words to end with a glottal stop.
subjects, whose Ll permits words to end with a glottal stop.
As predicted, the Cantonese subjects were significantly more sensitive to the English It/-/d/ contrast than the
Mandarin subjects were, with the Shanghainese subjects showing an intermediate level of performance. The
subjects in all three groups showed a significant increase in sensitivity as a result of the training. The increase
in correct identifications averaged 25%. The results were interpreted to mean that native-Ianguage phonotactic
constraints influence how syllables are processed. More specifically, we conclude that the Cantonese subjects
focused greater attention on the end of the consonant- vowel - consonant stimuli than the Mandarin subjects
did, which better enabled them to use remaining acoustic cues to the / t/-/d/ contrast such as FI offset
speakers of English, Mandarin and Spanish, Journal of the Acoustical Society of America, 92, 128-143.
Flege, J. & Fletcher, K. (1992). Talker and listener effects on the perception of degree of foreign accent.
Journal of the Acoustical Society of America, 91, 370-389.
Bohn, O.-S. Flege, J., Dagenais, P. & Fletcher, S. (1992) Differentiation and variability of tongue positioning in
the production of German vowels. Zeitschrift fur Dialektologie und Linguistik, 72, 1-26.
Bohn, O.-S., & Flege, J. (1992). The production of new and similar vowels by adult German learners of English.
Studies in Second Language Acquisition, 14, 131-158.
/r/ or /l/. Overall, liquids produced by the NE and EJ Ss were identified correctly more often (96%, 97%) than
those of the IJ Ss (79%). An effect of elicitation task was evident only for the IJ Ss because only they were
below ceiling. For them, the percent correct identification scores obtained in the Spontaneous task were
lower (68%) than those obtained in the Reading and Definition tasks (85%, 83%).
The aim of Exp. 4 was to obtain scores that were below ceiling. The listeners from Exp. 3 returned for a
second session in which they rated the stimuli elicited in the Reading, Definition and Spontaneous task using
a scale ranging from 1-strong foreign accent to 7-no foreign accent. The /r/ and /l/ stimuli were presented in
fairly similar. Averaged across elicitation tasks and the two liquids, /r/ and /l/, 10 of the 12 EJ Ss and 3 of the
12 IJ Ss obtained scores falling within the NE range.
Note: Recently, I had another look at the Definition task data, separating the scores for r and l. For both /r/
and /l/, all 12 EJ and NE Ss obtained scores falling within the NE range. For the EJ Ss, on the other hand, 11
Ss fell within the NE range for /r/, just 4 for /l/. It appears that the greater difficulty of producing /l/ than /r/
disappeared as the NJ Ss became more experienced in English.