Hanyong Park

Perceptual category mapping between English and Korean prevocalic obstruents: Evidence from mapping effects in second language identification skills

Abstract The current study develops an approach to quantify the extent to which native language (L1) categories are used in second language (L2) category identification, and uses this approach to examine the identification of a set of English obstruents by Korean learners of English as a foreign language. Forty native Koreans listened to nonsense English CV words consisting of /p b t d f v θ ð/ and /a/, and were asked to identify the consonant with both Korean and English labeling. They also gave gradient evaluations of the goodness of the Korean labels to the stimuli. The results of the Korean labeling task were analyzed to predict what confusion patterns would be expected if listeners used L1 categories and probabilistically mapped them onto L1 category responses. Results show the perceptual patterns of L2 stops can be successfully predicted by use of L1 categories alone if the listeners’ goodness rating scores were used to weight the probabilistic mapping from L1 to L2 in the predictions. Accuracy for other segments, such as /p/ and /f/, was higher than predicted. In general, this increase in accuracy over what is predicted from the L1 mapping data was negatively correlated with the average goodness-of-fit to the Korean. These results provide quantitative corroboration of acquisition models claiming that some L2 categories can function by using existing L1 categories alone while others must be indicative of the addition of a new linguistic category.

Evidence for featural units in the acquisition of speech production skills: Linguistic structure in foreign accent

Abstract This study examines correlations in accuracy of the production of one set of segments with accuracy in segments that share a featural contrast in Korean EFL (English as a Foreign Language) learners. Results indicate that accuracy rates for segment sets that share gestures in production tend to correlate, while segments that contrast in the same feature, but require the acquisition of different gestures do not correlate. Data here are from two tasks, a reading task and a mimicry task. Correlation results are similar across the two tasks, though a larger range of inter-subject differences in overall accuracy is evident in the mimicry task. Comparison of correlation patterns with previously published correlation patterns in perceptual identification indicates that patterns differ for perception and production, indicating that the structure of the skill sets, and hence, the acquisitional units for production and perception are different.

VOWEL EPENTHESIS AND SEGMENT IDENTITY IN KOREAN LEARNERS OF ENGLISH

Recent literature has sought to understand the presence of epenthetic vowels after the productions of postvocalic word-final consonants by second language (L2) learners whose first languages (L1s) restrict the presence of obstruents in coda position. Previous models include those in which epenthesis is seen as a strategy to mitigate the effects of coda licensing restrictions in the L1; others see epenthesis as a result of misperception of consonant releases in the L2 as indicating the presence of an additional syllable nucleus. The current study examines segmental identification and syllable counting in inexperienced Korean learners of English as a foreign language. Across stimuli, two effects were found. Increased perceptual epenthesis correlated with increased identification of a segment as voiceless, indicating a joint perception of voicelessness and an epenthetic vowel. Also, rate of epenthesis increased with accuracy of detecting sibilants and coronal segments, sounds characterized by salient consonant noise. Taken together, these results indicate a perceptual mechanism that evaluates segmental content and prosodic structure from an integrated percept. Analyses across individual listeners do not support a model in which perceptual epenthesis is a strategy to increase segmental accuracy; rather, they support models in which both syllable counting and segmental identification are skills that are reflections of increased proficiency in the L2. These results are discussed with respect to models of L2 phonology that rely on the robust encoding of lexical items.

Cross‐language perceptual category mapping: Korean perception of English obstruents

Models, such as SLM and PAM, predict that performance on second language sounds is determined by the perceptual relationship of the sounds to the original language categories. To measure this relationship, Schmidt [J. Acoust. Soc. Am. 99, 3201–3211] had native Korean speakers classify English consonant productions into Korean orthographic categories, and assess the similarity of the consonants to the chosen categories. The current experiment further examines how Korean labeling relates to accuracy in using English orthographic categories. Results show Koreans poorly identify sounds rated as dissimilar from Korean categories. Similarly, sounds that are inconsistently labeled with Korean labels are less accurately identified. These results suggest that accuracy varies with the nearness of the English and a Korean category, and thus English categories are developed from original Korean categories. However, other results indicate that sounds straddling Korean categories can be very accurately labeled, suggest...

Neutralization in the perception and production of English coda obstruents by Korean learners of English

Phonological neutralization rules require the suspension of differences between segments in perception and production. This paper examines the role of neutralization in the production and perception of Korean learners of English. In Korean, laryngeal and manner contrasts found in initial obstruents are systematically neutralized in coda position into stops that sound like voiceless stops in English. The current paper pursues two questions: (1) Does neutralization have the same effect on perceptual and production abilities? (2) are neutralization effects found with all English segments, or are they restricted to stops, which are transferred from Korean? /p, b, t, d, f, v, θ, δ/ were placed after /a/. Two tasks are compared: (A) Identification. Forty Korean learners identified coda consonants produced by four native speakers of English, using English labels. (B) Reading. Ten native English listeners identified coda consonants produced by four Korean learners. Identification errors were largely unidirectiona...

Perceptual category mapping between English and Korean obstruents in non-CV positions: Prosodic location effects in second language identification skills

Abstract This study examines the degree to which mapping patterns between native language (L1) and second language (L2) categories for one prosodic context will generalize to other prosodic contexts, and how position-specific neutralization in the L1 influences the category mappings. Forty L1-Korean learners of English listened to English nonsense words consisting of /p b t d f v θ ð/ and /ɑ/, with the consonants appearing in pre-stressed intervocalic, post-stressed intervocalic, or coda context, and were asked to identify the consonant with both Korean and English labeling and to give gradient evaluations of the goodness of each label to the stimuli. Results show that the mapping patterns differ extensively from those found previously with the same subjects for consonants in initial, onset context. The mapping patterns for the intervocalic context also differed by position with respect to stress location. Coda consonants elicited poor goodness-of-fit and noisier mapping patterns for all segments, suggesting that an L1 coda neutralization process put all L2-English sounds in codas as “new” sounds under the Speech Learning Model (SLM) framework ( Flege, 1995 ). Taken together, the results indicate that consonant learning needs to be evaluated in terms of position-by-position variants, rather than just being a general property of the overall consonant systems.

Predicting second language (L2) identification rates from first language (L1) mapping data: Similarity patterns for English and Korean obstruents in pre‐ and poststressed intervocalic, and postvocalic positions.

Park and de Jong [J.Phon, 36, 704–723 (2008)] found that listeners’ identification rates of second language (L2) categories can be predicted from mapping data, provided the L2 category has a high degree of subjective correspondence to native language (L1) categories. The current study examines whether identification rates can be predicted for consonants in different prosodic locations, and hence, whether the reliance on L1 categories is the same, regardless of prosodic position. Forty native Koreans identified /p b t d f v θo/ in pre‐ and poststressed intervocalic, and postvocalic positions, in nonsense English words, using both Korean and English labels, and also gave gradient evaluations of their identifications. The mapping data were used to predict confusion rates for the L2 identification data, finding results for prestress intervocalic consonants similar to the previous study for wordinitial consonants; stop identification rates were well‐predicted, but fricatives are systematically much better than...

Segments and segmental properties in cross‐language perception: Korean perception of English obstruents in various prosodic locations

Experimental models of cross‐language perception and second‐language acquisition (such as PAM and SLM) typically treat language differences in terms of whether the two languages share phonological segmental categories. Linguistic models, by contrast, generally examine properties which cross classify segments, such as features, rules, or prosodic constraints. Such models predict that perceptual patterns found for one segment will generalize to other segments of the same class. This paper presents perceptual identifications of Korean listeners to a set of voiced and voiceless English stops and fricatives in various prosodic locations to determine the extent to which such generality occurs. Results show some class‐general effects; for example, voicing identification patterns generalize from stops, which occur in Korean, to nonsibilant fricatives, which are new to Korean listeners. However, when identification is poor, there are clear differences between segments within the same class. For example, in identif...

Speaker similarity, acoustic properties, and perceptual learning with non-native speech

Variability in speech signals plays an important role in perceptual learning of non-native speech. Between-talker differences are a major source of variability. The present work aims to quantify within- and between-talker variability, following the approach of Silbert & Park (2014, JASA, 136, 2714) to quantify variation within and across Korean stop categories. We analyze 20 (10 male and 10 female) native speakers’ repeated productions of the Korean CVC nonword /p’ap/, using statistical techniques for dimensionality reduction to map multidimensional acoustic space to lower-dimensional talker-similarity space. We then compare the modeled similarity space to speaker discrimination data. Data were collected in an oddball task, where the listeners choose the oddball after listening to a set of triplets consisting of /p’ap/s spoken by two speakers (two distinct tokens from one speaker, one token from the other). The modeled similarity space allows us to automatically select tokens based on talker similarity an...

Statistical relationships between phonological categories and acoustic‐phonetic properties of Korean consonants

The mapping between segmental contrasts and acoustic‐phonetic properties is complex and many‐to‐many. Contrasts are often cued by a multiple acoustic‐phonetic properties, and acoustic‐phonetic properties typically provide information about multiple contrasts. Following the approach of de Jong et al. (2011, JASA 129, 2455), we analyze multiple native speakers’ repeated productions of Korean obstruents using a hierarchical multivariate statistical model of the relationship between multidimensional acoustics and phonological categories. Specifically, we model the mapping between categories and multidimensional acoustic measurements from multiple repetitions of 14 Korean obstruent consonants produced by 20 native speakers (10 male, 10 female) in onset position in monosyllables. The statistical model allows us to analyze distinct within- and between-speaker sources of variability in consonant production, and model comparisons allow us to assess the utility of complexity in the assumed underlying phonological c...