Earticle

This study investigated domain-initial strengthening in the prosodic hierarchy for Korean male and female speakers and examined whether there is any phonological transfer in the phonetic realization of English and Korean stops produced by Korean speakers. Prosodic structure proved to play a significant role in phonetically- strengthened realization of articulatory properties such as VOT, F0, and [H1-H2]. Correlation patterns among these parameters indicate that Korean speakers rely on VOT and F0 for both Korean and English stops. [H1-H2] turned out to be proportionate to VOT in English and Korean stops produced by Korean speakers. The [H1-H2] values overlapped between lenis and aspirated stops. The VOT values of voiced stops were similar to those of lenis stops, but the [H1-H2] values of the former were lower than those of the latter. The voicing and phonetic type of the stop category were phonetically-implemented across the three prosodic domains with the interaction of VOT, F0 and [H1-H2] for Korean speakers.

This study investigates speech rate development of Korean-speaking young children in a relation to VOT distinction. Speech rate is a duration-based acoustic measure and crucial parameter to determine language development and language processing. To examine children’s speech rate, a production experiment was conducted with three- and four-year-old Korean-speaking children. The two child groups showed insignificant speech rate differences in mono- and bi-syllabic words, while the difference from adult speakers apparently increased in bi-syllabic words. Even though children showed adult-like VOT distinction patterns across three stop categories, the ratio of VOT to word duration was different from that of adult speech. This finding indicates that adult-like VOT distinction development in a segment level would emerge ahead of adult-like speech rate development. The results suggest that children in those ages did not show adult-like speech rate control, and for the speech rate control development, multiple linguistic features and cognitive factors are related. The 3- or 4-year-olds need to further develop speech rate.

The purpose of this paper is to examine the effects of age and gender on VOT, F0, and H1-H2 in Korean voiceless stops produced by Daegu speakers. The results revealed that VOT, F0, and H1-H2 are primary acoustic cues for the distinction between male and female speakers. For the effect of age, F0 plays a role in differentiating three age groups while VOT distinguishes between the child and the two adult groups. However, the H1-H2 is not affected by age. With regard to phonation type, Daegu speakers use VOT for a three-way laryngeal contrast, but both F0 and H1-H2 make a two-way distinction, which indicates that VOT is a primary cue for the phonation type. Furthermore, the young female speakers show a unique characteristic revealing that the VOT difference between the lenis and the aspirated stop categories was smaller than other groups, which implies that they are sensitive to a sound change and prefer to use standard language. Regarding the place of articulation, only VOT distinguishes the place of articulation on the stops.

This study analyzed the values of VOT of Korean stops in two groups and surveyed the characteristics of Vietnamese migrant females’ pronunciation of Korean stops. Thirteen Vietnamese migrant females and thirteen Korean females participated to this study. The speech samples were composed by monosyllabic Korean stops following vowels (/a/, /i/ and /u/). All speech samples were collected and analyzed by Praat program. As for the average of the VOT values according to the manner of articulation, the aspirated stops were the longest in both groups. However, VOT of aspirated and lenis stops for Korean females was longer than those for Vietnamese migrant females. VOT of glottalized stops for Korean females was shorter than that for Vietnamese migrant females. As for errors in pronunciation by Vietnamese migrant females’ group, there were the most significant errors of pronouncing the lenis as the glottalized, the aspirated as the lenis and the glottalized as the lenis. It was found that there were statistically significant differences in the average of the values of VOT between the two groups in the entire lenis stops.

The study explores the acoustic properties of stops produced by Korean speakers of Jeju and examines how they differ by comparing VOT, F0, and H1-H2 in terms of speaker gender and age. The findings indicated that as for the VOT, the lax stops increase while the aspirated stops decrease as speakers get younger. In particular, a significant change in the aspirated stops was observed for the middle-aged female speakers. For the F0, the middle-aged adult groups showed a significant difference between the lax and the aspirated stops than in any other groups, especially for the female speakers. This implies that middle-aged female speakers complement VOT with F0 when discriminating the lax from the aspirated stops. The unique acoustic characteristics that appear for the middle-aged female speakers can be explained as a temporary phenomenon, not a diachronic change. In addition to VOT and F0, it is worth noting that H1-H2 could be employed as an acoustic cue for the classification of stops. The current study suggests that the acoustic properties observed in Jeju Korean should be interpreted through sociophonetic perspectives, and in particular, the acoustic changes in the aspirated stops are worthy of notice.

The purpose of this paper is to explore what acoustic cues are employed for the distinction of three phonation types of stops and affricates in Jeju Korean, and it examines whether two different manners of consonants with the same three-laryngeal contrasts are undergoing a sound change. The findings of the study revealed that for stops and affricates, a main effect on phonation type was found for VOT, F0, and H1-H2. As for the age effect, a significant difference was found in VOT, but not in H1-H2. A main age effect in F0 was observed for stops but not for affricates. With respect to the interactional effects, the lax and the aspirated consonants showed shorter duration as speaker age decreased, causing these two phonation categories to merge. In terms of F0, the aspirated consonants for these two consonant manners had higher F0 as speaker age decreased. Concerning H1-H2, there was no significant difference in the lax consonants for stops and affricates as a function of age, but H1-H2 in the aspirated consonants appeared higher as speaker age decreased. The current study suggested that Jeju Korean is undergoing a consistent sound change between the stops and the affricates.

This study investigates the acoustic properties of Korean stops in word-initial position produced by Jeju speakers and examines the effects of phonation type, place of articulation, and speaker’s gender on VOT, F0, H1-H2, and the spectral properties of the stop release bursts. The findings show that VOT, F0, the amplitude difference in the first harmonic and the second harmonic(H1-H2), and the amplitude intensity of a release burst for Korean stops contribute to the phonation type, whereas VOT and the four spectral moment analysis contribute to the place of articulation. That is, Jeju speakers employ the distinctive acoustic cues for the stop consonants in terms of phonation type and place of articulation. Aside from the fact that acoustic parameters of speech encode linguistic information, the present study suggests that F0, H1-H2, skewness, kurtosis, and the amplitude intensity of a release burst are employed to convey social information about a speaker’s gender. In particular, the present study suggests that the release burst of stops can convey social information as well as linguistic information.

본 연구는 중국 내 한국어 학습자에 의해 산출된 한국어 어두 파열음의 음향적 특성을 분 석하였다. 이를 위해 산출실험을 실행하였으며 실험 환경은 두 가지로 하였다. 하나는 주어진 자료를 보고 읽는 것이고, 다른 하나는 주어진 실험 자극을 듣고 따라 읽는 것이다. 나아가 두 실험에서 산출된 어두 파열음이 어떠한 차이를 갖는지 VOT와 후행 모음의 F0를 통해 관 찰하였다. 실험 결과는 다음과 같다. 보고읽기 실험에서 산출된 VOT는 평음-경음-격음의 삼 중 대립을 잘 나타내고, 후행 모음의 F0도 평음-경음/격음의 대립을 잘 보였다. 그러나 따라 읽기 실험에서 산출된 VOT와 후행 모음 F0는 평음과 격음 사이에서 통계적인 차이가 관찰 되지 않았다. 이는 한국어 평음의 지각적 특성이 한국어 산출에 오류를 유발하기 쉽다는 것을 의미하며, 또한 조음 방법을 중심으로 한 발음 교육의 중요성이 부각되기도 한다.

The purpose of this study is to analyze the acoustic characteristics of Korean word-initial stops produced by students majoring in Korean in China. To address the goal, pronunciation assessments were performed, and the assessments consisted of two parts. One was to look at and read the material given, and the other was to listen to and repeat the experiment stimulation given. Moreover, how the word-initial stops produced at the two experiments differed was observed with the VOT and following vowel F0. The results of the experiments are as follows. The VOT produced at the experiment of looking at and reading the material given, the three-way (lenis-fortis-aspirated) phoneme distinction was well observed, and the following vowel F0, too, revealed lenis-fortis/aspirated phoneme distinction well. In the VOT produced at the experiment of listening and repeating, difference between the lenis and aspirated phonemes was not observed. This implies that the auditory perceptual features of Korean word-initial lenis phonemes do not influence Korean pronunciation negatively, and also, it stresses the importance of pronunciation education focused on the manner of articulation.

본 연구는 한국어 3중대립파열음의 발성 유형 체계에 있어 구형태 사용 지역인 부산과 신형태 사용 지역인 서울의 중국인 학습자가 구형태와 신형태를 인식하는 양상을 조사하여, 파열음의 지역 변이형이 파열음 습득에 미치는 영향을 분석하는 것을 연구의 목적으로 삼았다. 연구를 위해, 부산 지역의 중국인 학습자 40명(초급과 고급 각 20명)과 서울의 중국인 학습자 40명(초급과 고급 각 20명)을 대상으로, 신형태와 구형태 파열음에 대한 청취 실험이 실시되었다. 조사 결과, 구형태에 대해서는 부산 학습자와 서울 학습자 모두 학습 기간이 길어지면서 평음과 격음 간, 평음과 경음 간 변별 능력이 향상이 되어, 결과적으로 평음을 경음, 격음과 유사한 수준으로 습득해 냈다. 신형태의 경우는 부산 학습자는 학습 기간이 길어져도 평음과 격음 간 변별 능력이 서울 학습자 정도로 향상되지 않았다. 그리고 그 결과, 평음과 격음을 경음만큼 정확도가 높게 습득하지 못하는 것으로 나타났다. 그에 반해 서울 학습자는 신형태의 경우도 평음과 격음 간 변별 능력이 상당 수준으로 습득되어. 그 결과, 평음을 경, 격음과 유사한 정확도 수준으로 습득해 내는 것으로 나타났다.

The study examined Chinese learners of Korean in Busan, where the old form of phonation type is used, and in Seoul, where the new form is used, to perceive old and new forms of the phonation types of Korean three-way plosives, and analyzed how the regional variants of plosives influenced Chinese learners acquiring them. Based on the results of the analysis, pedagogical implications were suggested. For the study, 40 beginning and 40 advanced Chinese learners of Korean in Busan and Seoul (40 in each region) participated in a perception test. The perception test was designed with identification tasks that consisted of 90 questions developed with nine plosives, pa, p’a, pʰa, ta, t’a, tʰa, ka, k’a, kʰa. As a result of the analysis, in terms of the old form, both Busan learners and Seoul learners improved in discriminating lenis from aspirated or tense so that their lenis displayed a similar degree of accuracy with aspirated and tense. In the case of the new form, the ability of Busan learners to discriminate between lenis and aspirated did not improve, and as a result, they showed a lower degree of accuracy in lenis compared to Seoul learners. On the other hand, Seoul learners acquired a considerable level of discrimination ability between lenis and aspirated and they acquired lenis with a similar level of accuracy as aspirated and tense.

The purpose of this study was to explore the Spectral times of the Broca's Aphasic patients' stops caused by a stroke. A survey was conducted on 5 Broca's Aphasic patients and 7 people in control group. Spectral times were analized using Praat (version 4.4.22) and a microphone (shure PG48) was directly connected to a notebook PC for recording. The results were as follows: 1) The Broca's Aphasic patients' VOT, post-vowel length and utterance length of whole word were longer than those of the control group. 2) The VOT of the control group and the Broca's Aphasic patient group showed a significant difference in all of bilabial consonants, alveolar consonants, velar consonants and fortis consonants. The VOT lengths of the control group were all shorter than those of the Broca's Aphasic patient group.

The purpose of this study is to investigate the acoustic characteristics of Korean velar tensification between normal and fast speech rate. This study focuses on four acoustic characteristics of two-syllable words with tensed velar stops for 34 female speakers: VOT, CoG, Closure duration and H1-H2. According to the study, these four acoustic features become lower or shorter as speech rate increases. 1) VOT is affected by the articulation place of the coda preceding velar tensification. VOT is the shortest when the coda is a velar stop. 2) CoG shows the same value at the same speech rate, regardless of the coda preceding velar tensification. However, CoG depends on the articulation place of consonants that precede velar tensification. 3) Closure duration becomes shorter at fast speech rate than at normal speech rate without exception. Closure duration is a lot longer in the word with a stop coda than with a sonorant coda. 4) As speech rate increases, H1-H2 moves in the negative direction. 5) The closer the articulation place is moved toward the velar, the lower the correlation between fast and normal speech rate becomes in VOT, CoG, and Closure duration. H1-H2 shows the high correlation between fast and normal speech rate in all words except of ‘palgil’ with the lateral coda.

It is recognized that word frequency is a significant factor influencing voice onset time (VOT). Nevertheless, the effects of word frequency factor on VOT have scarcely been investigated. The purpose of this paper is to examine how the VOT of the English voiceless stops in word-initial position is differently realized in high frequency words and low frequency words in terms of gender, age, place of articulation (POA) and vowel height. The results of the current study showed that in high frequency words the factors of age, POA and vowel height had a significant influence on VOT values, with longer VOT for younger people, for the alveolar stop /t/ and for the high vowels. In low frequency words only POA significantly affected VOT values, with longer VOT for the velar stop /k/. The interactions between POA and vowel height, and between age and vowel height were statistically different, although no interaction was found in the high frequency words.

Previous literature proposed that the ongoing VOT merger between aspirated and lenis stops in Korean is prosodically conditioned. To assess this prosodic account across a variety of prosodic positions, twelve younger speakers were recorded using a novel method of eliciting the stops in four positions with different boundary strengths, distributed over single utterances under broad focus. Bolstering previous findings, the merger occurred phrase-initially (i.e., utterance-initially, IP-initially, AP-initially), while the contrast was maintained by clear separation of F0. Notably, the pivotal role of F0 was also observed phrase-medially (i.e., Wd-initially) in the presence of the VOT cues retaining the feature primacy, even influencing the pitch range of the entire phrase. Additionally, the cue-shifting from a VOT-based to an F0-based system was more advanced in females’ speech. The magnitude of cue-shifting is discussed in light of boundary strengths, focusing on how variant forms involved in ongoing sound change interact with speech prosody.

This study aimed to examine whether Korean natives make use of VOT and F0 cues differently between the perception and the production of phonation types of Korean stops. It has an significant meaning in pronunciation education to know on which acoustic cues Korean natives placed greater weight when distinguishing among Korean /ㄱ, ㄲ, ㅋ/. In this study, the experiment on production of /ㄱ, ㄲ, ㅋ/ by Korean and Chinese natives were conducted to analyze the patterns of VOT and F0 realization. In addition, the experiment on perception was also conducted to identify whether Korean natives use the same acoustic cues as that of the production when judging foreign learners pronunciation. The results from the production analysis of Korean and Chinese indicated that all groups of subjects placed greater weight on VOT when distinguishing /ㄲ/ from the others. while F0 had statistically significant differences in /ㄱ, ㄲ, ㅋ/ only by the group of text-reading. In case of the perception, acoustic cues of Korean natives took on a different features from that of the production on distinguishing phonation types. These results imply that the way to judge appropriateness of a learner's pronunciation by VOT and F0 features on production should be reconsiderd.

The objective of this study is to examine the vowel duration that follow word initial-plosives in Korean language produced by Chinese learners, to identify its characteristics and propose how to use the following vowel duration for teaching pronunciation of word initial-plosives. To this end, 20 native Korean speakers from Seoul and Gyeonggido Province in their 20s and 20 Chinese speakers from Shandong Province in their 20s participated in the production experiment. They were asked to pronounce nine word initial-plosives in each sentence unit, and the vowel duration was analyzed. According to the result, native Korean speakers produced the longest sound for vowels that followed fortis, while producing shorter sounds for the following vowels of lax and aspirated. Another observation was an inverse relationship between VOT of word-initial plosives and the vowel duration. In comparison, the vowel duration produced by Chinese speakers had no differences among lax, aspirated, and fortis. Also, an inverse relationship was not observed between VOT and the vowel duration. To examine the effect of education, we taught Chinese participants that the vowel duration should vary by the phonation type of word initial-plosives. Afterwards, we asked them to read the same sentences again, and measured the vowel duration. The vowel duration produced by Chinese participants after the instruction remained different from that produced by native Korean speakers. However, compared to pre-instruction test, Chinese participants produced shorter sounds for the following vowels of lax and aspirated consonants. This suggests that the characteristics of the vowel duration by different phonation type can be applied to teaching word initial-plosives in Korean language. Teaching Chinese learners of Korean language to pronounce the vowel duration of fortis longer than the following vowel of lax or aspirated would help them make a distinction between lax and aspirated, and correctly pronounce them. (Korea University)

Oh, Kwan Young. “A study of the distributions and realizations of variants in the English voiceless alveolar stop.” Studies in English Language & Literature. 42.2 (2016): 213-237. The purpose of this study is to analyze the sound distribution of Korean speakers of English in terms of their pronunciations of the voiceless alveolar stop /t/ in six different contexts. We want to identify what kinds of variant sounds of the /t/ in each context were realized and also compare those sounds with the variant sounds of Americans in the same contexts. As a consequence of analyses, we confirmed that Koreans still lacked the knowledge of phonological contexts such as the difference between stressed and unstressed syllables and the relation between variant types and their contexts in which /t/ occurred. Also, we recognized that they pronounced the /t/ as an aspirated sound whenever it preceded a vowel. In addition, an interesting fact observed from Koreans was that the /t/ was realized as several types as in [th], [ɾ], RV(release and vowel insertion), and [d] in the intervocalic(foot-internal) context ((ˈ)V__V), and also in #__(word-initial) context, as in [ts], [t], [t˺], RV, and [·]. Through statistical analyses of R, we recognized that the differences in the measured values of close phase, VOT(Voice Onset Time), and the aggregate of the two between Koreans and Americans were significant. Finally, it was shown that each of the values of close phase, VOT, and the aggregate of the two were affected by both the speakers of the two groups and the contexts of variants through Interaction Plot. (Chonnam National University)

Han, Jeong-Im & Kim, Joo-Yeon. 2014. Acquisition of Korean Stops by Native Speakers of Mandarin Chinese: A One-year Longitudinal Study. Korean Journal of Linguistics, 39-2, 377-403. This study examined the production of the Korean stops (lenis, fortis, and aspirated) by Chinese learners over a one-year period following their first exposure to Korean. The main goals of the study were to investigate the learning trajectory of adult learners of second-language (L2) segmental acquisition and to determine whether L2 learners could make use of the multiple acoustic cues associated with the L2 segments. In Experiment 1, the voice onset time (VOT) and vowel-onset fundamental frequency (onset F0) were measured for each type of stops in Korean produced by the Chinese learners. The results indicate that L2 phonetic learning progresses gradually instead of following a pattern characterized by rapid initial learning and a subsequent plateau within a few months. The Chinese learners also learned to control the VOT quite easily but not the onset F0; this result suggests that L2 learners have difficulty producing the multiple acoustic correlates for L2 phonetic contrasts. In Experiment 2, it was shown that the asymmetry in the development of controlling VOT and onset F0 had a direct impact on the intelligibility of Chinese learners’ utterances to native Korean speakers. (Konkuk University)

Han, Jeong-Im. 2013. Testing the Influence of Different Use of L1 Acoustic Cues between Two Age Group Speakers on the Acquisition of L2 Stops. Korean Journal of Linguistics, 38-2, 493-517. The present study investigated the production and perception of voiced vs. voiceless contrast of English stops by two age groups of Korean learners. The issue addressed in this study was whether young and older Korean speakers differ in the way they produce and perceive the English stops. Given recent diachronic change from VOT to onset F0 as a primary cue to the Korean initial stops, though VOT is employed primarily to distinguish the English stops, it was explored whether young speakers make greater use of the onset F0 than old speakers to signal the English stops in word-initial position. This was to test whether relative importance of specific acoustic properties in the native language affects the extent to which those acoustic properties are used in the production and perception of a target language contrast. The results showed that in both production and perception, both young and older speakers depended primarily on VOT to distinguish the English stops, and onset F0 was employed redundantly. Two possible explanations for this result were put forth based on the characteristics of the speakers and listeners: both age group speakers might acquire already how to use acoustic cues to the English laryngeal contrast; or the ongoing change in the use of acoustic cues for Korean laryngeal contrast might not be clearly shown in the two age group speakers in the present study. (Konkuk University)