씰테크(주)

윤관희 한국접착및계면학회 2022 접착ㆍ코팅기술 Vol.1 No.2

The effects of neighborhood density and word frequency on the production of English-learning Korean speakers

윤관희 한국음운론학회 2010 음성·음운·형태론 연구 Vol.16 No.2

This study investigates whether English-learning Korean speakers’ production is affected by two parameters in organizing mental lexicon such as neighborhood density and word frequency. Speakers produced four types of real English words classified by high- and low- neighborhood density and high- and low-frequency. In order to see these effects, F1s and F2s of vowels, vowel duration and pitch were measured. First, the results showed that F1 and F2 values of vowels in the dense neighborhood words were significantly different from those of sparse neighborhood words as consistent with previous results for native speakers of English. Second, F1s and F2s of vowels in high-frequency words were also different from those in low-frequency words. Put together, these results show that vowel space in the low-frequency words and dense neighborhood words was more expanded than that in the high-frequency and sparse neighborhood words. Furthermore, pitches of vowels in sparse neighborhood words were significantly higher than those in dense neighborhood words. In addition, vowels were produced with longer duration in low-density words and high-frequency words than in high-density words and low-frequency words. However, vowel pitch did not significantly differ by word frequency but by neighborhood density. Overall, our findings suggest that lexical factors like neighborhood density and word frequency influence the phonetic-fine details and those two factors serve the function of parameters associated with the organization of L2 speakers’ mental lexicon as well as of L1 speakers’.

Phonologically-driven Intergestural Timing Relation

윤관희 한국언어학회 2008 언어 Vol.33 No.1

This paper presents the results of the experimental study on the influences of phonological rules or constraints on intergestural timing relations realized as the degree of vowel-to-vowel coarticulation in English. First, both ultrasound imaging and acoustic results showed that palatalization rules enhanced the degree of vowel-to-vowel coarticulation compared with the case of nonpalatalized forms. Second, we obtained the result that postlexical palatalization rules in English contributed to a greater degree of coarticulation than lexical palatalization rules. These results provide an interesting implication that the on-line application of phonological rules and the lexical status of phonological rules directly affect the low-level phonetic fine details such as intergestural timing relations. On the basis of these results, I adopt a unified model of phonology and phonetics, formally accounting for these results through feature-and-gesture based OT frameworks. Further, my analysis shows that intergestural timing relations can be formally incorporated in phonology grammar.

The Production of German Final Devoicing by Korean Speakers

윤관희 한국언어학회 2011 언어 Vol.36 No.4

Yun, Gwanhi. 2011. The Production of German Final Devoicing by Korean Speakers. Korean Journal of Linguistics, 36-4, 979-1013. This study investigates how Korean speakers produce German final devoicing and whether they have acquired voicing contrast in German obstruents. Korean learners of German produced real German words with the underlying voicing distinction in the five phonological contexts. One is the non-neutralization environment and the others are potential neutralization contexts (syllable-final, word-final, and utterance-final positions). In order to see whether the underlying voicing contrast is retained on the surface, five acoustic correlates of voicing were measured (duration of the preceding consonant, duration of the preceding vowel, duration of stop closure or frication, duration of release portion, and F0 of the preceding vowel). First, the results showed that in syllable initial or intervocalic position, the underlying voiceless obstruents were produced with longer duration of the preceding consonant, shorter duration of the preceding vowel, longer closure or frication duration, and longer release duration than their counterpart voiceless obstruents. Next, in syllable final, voiceless condition, the underlying voicing contrast was retained with respect to closure duration of the preceding consonant, and duration of the preceding vowel. Third, in a word-final position as well, the underlying voiceless obstruents were produced with longer duration of the preceding consonant, shorter duration of the preceding vowel, and longer closure or frication duration than their voiced counterparts. Finally, in utterance-final position, the underlying voiceless obstruents were produced with longer duration of the preceding consonant, shorter duration of the preceding vowel, and longer closure or frication duration than their voiced counterparts. These findings suggest that even Korean L2 speakers show incomplete neutralization in German final devoicing like some of the native speakers of German and can acquire phonemic voicing contrast, which does not exist in their L1. (Daegu University)

Production and Processing of Phonaesthemes in L2 English

윤관희 한국언어학회 2018 언어 Vol.43 No.4

Yun, Gwanhi. 2018. Production and Processing of Phonaesthemes in L2 English. Korean Journal of Linguistics, 43-4, 781-806. It has been demonstrated that phonaesthemes are common across languages and hold psychological reality although they are not distinct morphological units in formal morphological theory. This study examines whether L2 phonaesthemes hold a significant place as psychological reality in L2 mental lexicon. To investigate this issue, two psycholinguistic experiments were conducted with Korean L2 English speakers: Naming task and lexical decision priming task. In naming experiment, it was found that Korean speakers created novel words with English phonaesthemes and their core meanings, providing the indirect indication of their implicit knowledge of phonaesthemes. In lexical decision priming task, we found that phonaesthemically related words elicited faster and higher accuracy in word recognition than form-only related stimuli or baseline stimuli. Moreover, the degrees of phonaesthemic priming effects were less than those by native speakers of English, but greater than those found for morphologically related words in English. Furthermore, their lexical judgment also was affected by individual phonaesthemes, suggesting different strengths in the representations. The results from these two psycholinguistic behavioral experiments indicate that L2 phonaesthemes might be represented in L2 lexicon to some degree and their strengths might vary depending on individual phonaesthemes.

Against Intrusive r Strategy in English Vowel Hiatus: Evidence from the Buckeye Corpus and L2 Speech

윤관희,이민경 한국영어학회 2022 영어학 Vol.22 No.-

As phonologically-documented, in vowel-vowel (VV) sequences within words or across word boundaries, when the first word ends in one of the non-high vowels such as /ǝ, ɑ, / and the following vowel gets unstressed, intrusive r is added intervocalically. This paper attempts to examine and analyze whether such /r/-epenthesis takes place to avoid vowel hiatus in the speeches of Korean L2 English speakers (KS) collected from the production task compared to those of English native speakers (ES) extracted from the Buckeye Corpus of spontaneous conversational speech. As stimuli for both KS and ES, VV sequences across word boundaries are mainly targeted based on the possible anti-hiatus strategies, i.e. r-intrusion, glottal stop insertion and vowel deletion as well as canonical variant. Interestingly but strikingly, r-intrusion is hardly observed in both groups. For the tokens of ES, vowel hiatus predominantly arose and vowel elision was second-best. For the KS’s tokens, no r was embedded, either. However, Korean L2 English speakers behave differently given their English proficiency. For KS with low proficiency (LP), a pause between two vowels results in glottal stop insertion but a canonical form is the most favored with no pause. However, for KS with high proficiency (HP), a pause is hardly placed in hiatus contexts and canonical variants predominantly surface. Unlike LP KS, a pause does not play a key factor to apply any other anti-hiatal strategies and further the effect of vowel height is quite significant, i.e., hiatus is more tolerated in V1-non-high than in V1-high. Contrary to the phonological claims of intrusive r as a hiatus breaker, vowel hiatus remains intact with no r intruded.

English L2 Speakers' Processing of Morphologically Simplex and Suffixed Words

윤관희 한국언어학회 2015 언어 Vol.40 No.3

Yun, Gwanhi. 2015. English L2 Speakers’ Processing of Morphologically Simplex and Suffixed Words Korean Journal of Linguistics, 40-3, 40-3, 425-450. Recent studies on morphological processing have shown that morphologically complex words such as class II affixed words or regular inflected words are processed by decomposition whereas class I affixed words and irregular inflected words are processed as whole-word units. The present study aims at exploring whether such patterns found for native speakers of English emerge differently in accordance with L2 proficiency. First, the results attained for intermediate proficiency learners show that complex words benefit from more processing efficiency than simplex words, derivational words cost more difficulty processing than inflected words for high-frequency condition, and class II suffixed words suffer heavier processing burden than class I suffixed words. Second, the results for advanced learner group show that simplex words are processed more efficiently than complex words, inflectional words are processed with more difficulty than derivational words in low-frequency words, and class II suffixed words pay more processing cost than class I suffixed words in high-frequency condition. These findings suggest that L2 processing mode for complex words is contingent on proficiency level and surface frequency is mediated in the processing. Furthermore, it is implied that though the extent of dual-route processing is not as strong as that of target language users, L2 learners seem to be under progress towards dual-route processing as the growth of proficiency. (Daegu University)

Korean Learners' Perceptual Cues for Word Boundaries in English

윤관희 한국언어학회 2009 언어 Vol.34 No.1

This study investigates whether/how English-learning Korean speakers utilize allophonic cues such as aspiration and glottal stops in order to identify word boundaries in English. Given that speech stream is inherently continuous and causes segmentation problems, it has currently been controversial issue whether and what types of segmentation strategies L1 and L2 speakers adopt in word recognition and perception. Previous perception studies (Altenberg 2005; Ito and Strange 2007) showed that Spanish and Japanese learners of English are not as highly sensitive to allophonic cues like aspiration and glottal stops as native speakers of English. The purpose of this current study is to directly compare previous two studies with the result with English-learning Korean listeners. Thirty intermediate or advanced learners participated in perception experiments to see whether/how they segment word boundaries in English. The result showed that they utilized both aspiration and glottal stops above chance level, but that unlike our expectation, glottal stop stimuli were easier to perceive than aspiration stimuli. Furthermore, such results sharply pattern with those with Japanese and Spanish L2 learners even though both allophonic cues differ in utility in their consonant systems and the status of glottal stop differ in all three languages. Based on the current study, two interesting implications are provided. First, L2 learners might acquire phonetic-fine details though their sensitivity is not as high as native speakers, supporting Kuhl and Iverson (1995). Second, more reference to glottal stops over aspiration indicates that the former might be less marked than the latter across languages at least in the field of L2 phonetic or phonological acquisition.

What factors influence VOT ? : a comparative study of English, Japanese, and Korean

윤관희 대구대학교 인문과학연구소 2008 人文科學硏究 Vol.31 No.-

본고에서는 영어와 한국어와 일본어 자료를 대상으로 폐쇄음의 VOT(성대 진동 시점)를 비교, 분석했다. 실험 결과, 첫째, 영어와 한국어와 일본어는 각 언어의 폐쇄음의 음운적 구별을 위해 다른 범위의 VOT를 사용함이 드러났다. 둘째, VOT는 폐쇄음의 조음위치에 영향을 받는다는 점이 밝혀졌다. 영어와 한국어의 경우, 연구개음의 VOT가 순음과 치경음의 VOT보다 긴 반면, 일본어의 경우에는 순음의 VOT가 치경음이나 심지어 연구개음의 VOT보다 더 길었다. 이는 개별 언어에 따라 VOT가 조음위치 외에 다른 요소들에 의해 영향을 받을 수도 있음을 시사한다. 셋째,VOT는 영어와 일본어에서 후속모음의 높이와 정확히 유의미한 상관관계를 보이지 않았다. 그러나, VOT는 다른 모음 앞에서보다 고모음 앞에서 더 길었다. 한국어 폐쇄음의 경우에도 VOT는 후속모음의 높이에 따른 유의미한 상관 관계를 보이지 않았다. 한국어의 경우, 폐쇄음의 VOT는 선행 연구 결과들과는 달리, 중모음 앞에서보다, 저모음 앞에서 더 길었다. 선행 연구 결과와는 다른 본고의 결과는 VOT가 모음의 높이와 개구도 같은 다른 요인들과 복합적으로 작용함을 시사한다. 마지막으로, 영어와 일본어의 폐쇄음의 VOT는 유성폐쇄음과 무성폐쇄음의 대조를 극대화하기 위해 양극적 가치를 보여 주었다. 3가지 음운론적 대조를 나타내는 한국어 폐쇄음의 경우에도, VOT의 경우는 2가지 범주로 나타났다. 본고의 실험에서 가장 두드러진 발견은 일본어의 치경음과 연구개음의 경우 VOT의 중간단계가 존재한다는 결과였는데, 이는 폐쇄음의 적절한 기술에 기여할 것이다.

Soda-lime 유리 박막을 이용한 CIGS 박막의 Na 도핑효과 및 유연기판 태양전지의 셀특성 분석

윤관희,윤주헌,김종근,김원목,박종극,백영준,김동환,정증현 한국진공학회 2011 한국진공학회 학술발표회초록집 Vol.40 No.0