Anna Marie Schmidt

Cross‐language identification of consonants. Part 1. Korean perception of English

Twenty native Korean-speaking subjects heard 22 English word-initial consonants in three vowel contexts produced by three native English talkers. The subjects orthographically labeled each English consonant as the closest Korean consonant. They then judged how similar the English consonant was to the Korean consonant on a scale of 1 to 5. Some English consonants were labeled consistently as a single Korean consonant and judged to be very similar. Other English consonants were labeled consistently as a single Korean consonant but judged to be less similar. Still other English consonants were inconsistently labeled. Korean acoustic cues, vowel context, and token differences appeared to influence labeling choices.

Perceptual prothesis in native Spanish speakers

Previous research suggests a perceptual bias exists for native phonotactics [D. Massaro and M. Cohen, Percept. Psychophys. 34, 338–348 (1983)] such that listeners report nonexistent segments when listening to stimuli that violate native phonotactics [E. Dupoux, K. Kakehi, Y. Hirose, C. Pallier, and J. Mehler, J. Exp. Psychol.: Human Percept. Perform. 25, 1568–1578 (1999)]. This study investigated how native‐language experience affects second language processing, focusing on how native Spanish speakers perceive the English clusters /st/, /sp/, and /sk/, which represent phonotactically illegal forms in Spanish. To preserve native phonotactics, Spanish speakers often produce prothetic vowels before English words beginning with /s/ clusters. Is the influence of native phonotactics also present in the perception of illegal clusters? A stimuli continuum ranging from no vowel (e.g., ‘‘sku’’) to a full vowel (e.g., ‘‘esku’’) before the cluster was used. Four final vowel contexts were used for each cluster, result...

Perception of English /n l/ with amplified consonants or transitions by native speakers of Chinese

Previous work has shown that speakers of some southern Chinese dialects have difficulty perceptually differentiating English /n l/. In this study, an attempt was made to increase the perceptual salience of some possible acoustic cues. Seven native Chinese speakers labeled the first consonant of three combined sets of ten English /n l/ minimal pairs produced by a male and a female native English speaker. One set consisted of unaltered words from the earlier study. The consonant portion of the second set had been normalized for peak amplitude. The peak amplitude of the consonant was then equal to the peak amplitude of the vowel. In the third set, a portion of the CV transition had been normalized. Native English speakers labeled the words with high accuracy. Results and implications will be discussed.

Developing new pre-lexical processing skills in adults

Studies of adults learning to decode words may provide a window on understanding of the pre-lexical processing of words as learning progresses. Second language (L2) learners (for example) have been found to assimilate new phones to first language (L1) phonemes when asked for the closest sound in their languages (e.g., Schmidt, 1996). Ratings of fit with L1 phonemes was related to details of acoustics of stimuli, L1 acoustics, and vowel context suggesting that L1 phonetic details rather than phoneme categories were compared. What happens after attention has been focused on acoustic/articulatory characteristics of these new sounds (as in training studies) and learning progresses? Do adults establish new phoneme categories to aid in word processing or is it more likely that native categories are expanded based upon developing within L1 category perceptual skills and orthographic word prediction skills. Evidence will be reviewed indicating that learned production of L2 sounds can be become native-like but L2 perception outcomes remain non-native like suggesting that new phoneme categories are not formed. Thus, new learned words can be accurately produced but success in pre-lexical processing of new heard words will depend upon linguistic factors such as context. Studies of adults learning to decode words may provide a window on understanding of the pre-lexical processing of words as learning progresses. Second language (L2) learners (for example) have been found to assimilate new phones to first language (L1) phonemes when asked for the closest sound in their languages (e.g., Schmidt, 1996). Ratings of fit with L1 phonemes was related to details of acoustics of stimuli, L1 acoustics, and vowel context suggesting that L1 phonetic details rather than phoneme categories were compared. What happens after attention has been focused on acoustic/articulatory characteristics of these new sounds (as in training studies) and learning progresses? Do adults establish new phoneme categories to aid in word processing or is it more likely that native categories are expanded based upon developing within L1 category perceptual skills and orthographic word prediction skills. Evidence will be reviewed indicating that learned production of L2 sounds can be become native-like but L2...

Training new second language category formation

The Speech Learning Model (Flege, 1995) hypothesizes that, in second language (L2) learning, the greater the difference the learner perceives between in a sound from a first language (L1) category and a sound from an L2 category, the more likely it is that a new L2 category will be formed. Finding an L1 category that is not similar in some way to an L2 category is difficult. However, a L2 sound, that is similar to an L1 sound, could be consciously taught as a new allophone of that sound in a way that focuses attention on phonetic differences. In this study, Chinese L1 speakers, from a region of China where /n/ and /l/ in word initial position in their dialect (and in Mandarin) are produced as /n/, will be taught a dark Arabic /l/ as a new position specific allophone to be used in word initial position in English and Mandarin Chinese. Results from perception and production of /l/ and /n/ before and after training will be reported.

Production and perception training of /r l/ with native Japanese speakers

Visual feedback with electropalatometry was used to teach accurate /r/ and /l/ to a native Japanese speaker. Perceptual differentiation of the phonemes did not improve. A new perceptual training protocol was developed and tested.

Production and perception of two vowels in Northeast Ohio.

In Northeast Ohio, three dialect regions converge: the Inland North, North Midland, and Pittsburg regions (Labov et al. 2005). It is believed that the vowel in “bat” is raising and fronting while the vowel in “bet” is lowering and backing as part of the Northern cities vowel shift in the Inland North region. Earlier data from 50 talkers (aged 18–25) from these regions suggest vowel movement in each region with a full merger or reversal of the two vowels only for Inland North talkers. Talkers from all regions produced longer bat vowels than bet vowels. The current study examined perception as well as production of the vowels by talkers from these regions. Participants heard two synthetic vowel continua and identified the vowels: a duration continuum, in which the formant pattern was constant but the vowel duration varied, and a formant continuum, in which the duration was constant but the formant pattern varied from a low backed vowel to a high fronted vowel. Perception patterns were compared to production...

Developmental processes in production of Korean /s/ and /s*/

The purpose of this study was to investigate development of Korean children’s productions of tense and lax Korean s. Korean alveolar fricatives are acquired late; /s*/ is generally later than /s/. However, little is known about developmental processes in acquiring the contrast between tense and lax Korean alveolar fricatives. Twenty typically developing monolingual Korean children between 4 and 8 and ten monolingual Korean adults were recruited in Korea. All children fell within normal limits on both a Korean articulation test and a standardized Korean vocabulary test. A picture‐naming picture task elicited productions of initial fricatives, alveolar stops, and affricates followed by /a, i, u/ in monosyllabic words. Production error patterns will be analyzed to describe general developmental trends in production. More importantly, the contrast between productions of /s/ and /s*/ will be presented based on acoustic parameters including frication durations, centroid frequency, and harmonic‐related amplitude...

Phonological awareness of English by Chinese and Korean bilinguals

This study examined non‐native speakers phonological awareness of spoken English. Chinese speaking adults, Korean speaking adults, and English speaking adults were tested. The L2 speakers had been in the US for less than 6 months. Chinese and Korean allow no consonant clusters and have limited numbers of consonants allowable in syllable final position, whereas English allows a variety of clusters and various consonants in syllable final position. Subjects participated in eight phonological awareness tasks (4 replacement tasks and 4 deletion tasks) based on English phonology. In addition, digit span was measured. Preliminary analysis indicates that Chinese and Korean speaker errors appear to reflect L1 influences (such as orthography, phonotactic constraints, and phonology). All three groups of speakers showed more difficulty with manipulation of rime than onset, especially with postvocalic nasals. Results will be discussed in terms of syllable structure, L1 influence, and association with short term memory.

Cross‐language perception of fricative/vowel syllables by speakers of English and Mandarin

A recent study [Radhakrishnan and Schmidt (2001)] on the perception of fricative/vowel syllables by English and Mandarin speakers demonstrated that English speakers use the fricative noise while Mandarin speakers use both the fricative noise and vowel transitions in categorizing syllable initial fricatives in their native languages. In this study, cross‐language mappings for fricative/vowel syllables were obtained from native English and Mandarin speakers. Native English and Mandarin speakers labeled and rated three sets of tokens from L2. These tokens were (1) fricative noises from four different vowel environments, (2) natural fricative vowel syllables, and (3) F+V combinations in which the fricative noise from one language was combined with vowel transition from the other language. Vowel length was manipulated in the third set of stimuli. Results indicate differences in the cross‐language mappings by the two language groups.