한국어 고급 학습자의 발음 숙달도와 발음 학습 전략 사용 간의 관계 연구

임우열,김영주 국제한국어교육학회 2014 한국어 교육 Vol.25 No.4

This study aimed to investigate the use of pronunciation learning strategies of advanced Korean learners and the differences in the use of the strategies depending on their pronunciation proficiency. This study evaluated the pronunciation proficiency of 50 advanced Korean learners, by dividing it into fluency, comprehensibility, accuracy, and global foreign accent areas, based on recorded materials. The study also conducted the survey on the pronunciation learning strategies. The results showed that in the evaluation of the pronunciation proficiency of the whole groups, the score of global foreign accent was the lowest, while the score of comprehensibility was the highest. The strategy use of the top 30% and the lower 30% groups in the test, there appeared a difference in the use of the hypothesis testing strategies. For the whole groups, the use of the hypothesis testing strategies showed positive correlations with the detailed areas of the proficiency, indicating that the use of a specific strategy had correlation with the improvement of the pronunciation proficiency.

시각적 단어 재인에 미치는 단어 빈도와 의미적 연관성의 영향: 한국어 고급 학습자의 예

임우열,이선진,이선영,전현애,남기춘,김영주 국어국문학회 2020 국어국문학 Vol.0 No.192

The present study aims at investigating how semantic context and word frequency have effects on visual word recognition in Korean, and how these two factors interact with each other. In this experiment, 72 participants who are advanced Korean learners took part in the masked lexical decision priming task at SOA 43ms, the visual lexical decision priming task at SOA 230ms, and the cross-modal lexical decision priming task. The results of experiments indicated that the word frequency effect was significant in all tasks. Specifically, it had an impact on the cross-modal task, SOA 230ms task, and SOA 43ms task in turn. However, there were no significant semantic priming effects statistically in all three conditions, including SOA 230ms task, SOA 43ms task and the cross-modal task. Moreover, the experiment's result did not show an interactive effect between the word frequency and the semantic association. Consequently, the arguments put forward in this result showed that Korean word processing adopted different processing model between native Korean speakers and advanced Korean learners. These results were contrary to the study that the word frequency and the semantic context have the interactive effect at the initial stage of automatic semantic priming targeting native Korean participants. 본 연구는 고급 학습자의 한국어 시각 단어 재인 시 단어 빈도와 점화-표적 간 의미적 연관성이 어떤 영향을 미치며, 두 요인이 어떠한 관계를 맺고 있는지를 확인하는 데 그 목적이 있다. 이를 위해 한국어 고급 학습자 72명을 대상으로 하여 SOA 43ms 시각점화 어휘판단 과제, SOA 230ms 시각점화 어휘판단 과제, 교차양상 점화 어휘판단 과제를 시행하였다. 연구 결과를 보면, 단어의 빈도효과는 모든 과제에서 유의미하게 나타났으며 그 효과의 크기는 교차양상 점화 어휘판단 과제, SOA 230ms 시각점화 어휘판단 과제, SOA 43ms 시각점화 어휘판단 과제 순이었다. 의미적 연관성에 따른 의미점화 효과는 SOA 43ms, SOA 23oms, 교차양상 과제에서 모두 통계적으로 유의미하지 않았다. 또한 단어 빈도와 의미적 연관성 간의 상호작용 효과도 발견되지 않았다. 이러한 결과는 자동적 활성화 확산 이론을 증명한 모어 화자를 대상으로 한 연구와는 상반되는 것으로 한국어 모어 화자와 고급 학습자들의 어휘처리 모형에 차이가 있다는 주장을 뒷받침하는 결과이다.

중국인 한국어 학습자의 개인적 변인과 발음 숙달도 간의 상관성

임우열,김영주 한국외국어교육학회 2015 Foreign languages education Vol.22 No.3

This study examined whether learning context and language aptitude affected the use of pronunciation learning strategies and pronunciation proficiency. One hundred thirty-eight Chinese learners (88 KFL Ss & 50 KSL Ss) learners were investigated for the correlation between learning context(KSL, KFL), language aptitude (high/low), pronunciation learning strategy and pronunciation proficiency. Participants' sound discrimination ability was measured by PLAB and then they were divided into high group (upper 30%) and low group (lower 30%) by scores. SPLS and personal information survey were also conducted. The results showed that (i) a significant difference was found in pronunciation proficiency pursuant to learning context and language aptitude, (ii) use of pronunciation learning strategy differed significantly pursuant to learning context and language aptitude, and (ⅲ) use of pronunciation learning strategy could predict pronunciation proficiency. The findings suggest that effective pronunciation learning strategy should be investigated and taught in L2 Korean classroom.

Eye-Tracking Analysis of Advanced Korean Language Learners' Reading of Multi-modal Texts

임우열,김영주 한국응용언어학회 2024 응용 언어학 Vol.40 No.1

This study explores the cognitive mechanisms of second-language learners of Korean while navigating through multi-modal texts. Employing eye-tracking methodologies, we thoroughly examined the simultaneous processing of linguistic and visual elements by both native Korean speakers and advanced learners. The study cohort comprised 34 participants, comprising 18 native speakers and 16 advanced learners, all holding TOPIK Level 6 proficiency credentials from esteemed metropolitan universities. The findings revealed distinct reading behaviors in multi-modal texts among these cohorts. Notably, both groups allocated less time to multi-modal texts compared to their unimodal counterparts, aligning with the principles of dual-coding theory. Moreover, we discovered that factors such as phonological working memory, as proven by previous research, and spatial working memory also influence the reading process. However, in the case of advanced learners in multi-modal reading, the significant impact of working memory on the reading process was not observed. This implies that advanced Korean learners manifest distinct textual engagement patterns from their native counterparts. Recognizing these intricacies provides valuable insights for refining reading pedagogies and curriculum designs, potentially enhancing Korean learners' reading competencies.

한국어 고급 학습자의 용언 활용형 처리에 대한 점화 실험 연구

임우열(Wooyeol Lim),이선진(Sunjin Lee),김영주(Youngjoo Kim) 한국응용언어학회 2020 응용 언어학 Vol.36 No.2

This study investigated how L2 Korean advanced learners process regularly and irregularly inflected forms of Korean verbs, and whether this processing mechanism follows a single mechanism account or a dual mechanism account. In addition, it examined the effect of the learner’s working memory capacity on the reaction time or/and the accuracy rate in the chosen priming tasks. The patterns of facilitation on present tense verb targets were examined in three experiments. In the study, there were 72 participants with TOPIK level 6, who performed three lexical decision tasks with either a masked visual prime presented during 43ms (Experiment 1) or 230ms (Experiment 2) or an auditory prime (Experiment 3). Next, the priming for six different types of verbal inflections were compared: regular verbs with no orthographical or phonological changes, regular verbs with only phonological changes to their stem, regular verbs with only orthographical changes, irregular verbs with their stem omitted or alternated, irregular verbs with changes in their endings and irregular verbs with changes in both their stem and endings. The study found determined that (i) no priming effects in either regular or irregular inflections revealed in most conditions, which supports a single mechanism account, and (ii) the influence of working memory capacity of L2 learners showed as higher the working memory was, faster the reaction time was completed in those cases.

한국어 고급 학습자의 과거시제선어말어미 결합 용언 처리에 대한 점화 실험 연구

이선진,임우열,김영주 한국국어교육학회 2020 새국어교육 Vol.0 No.122

Purpose:The purpose of present study was to examine whether the past tense pre-final endings are stored in the mental lexicon as a full-listing model or decomposition model and to investigate the role of working memory in language processing. Methods:In the experiment, 72 advanced learners took lexical decision priming tasks with 43ms and 230ms SOA and cross-modal priming task. The working memory task was conducted to check the response time and the accuracy. Results:The analyzed data showed no significant difference between the stem frequency of the target words and the response time in the 43ms and 230ms SOA conditions, which supports the full-listing model. Whereas, in the cross-modal priming task, the full-listing model and the decomposition model were supported by the low-frequency words and high-frequency words, respectively. The learners’ working memory did not influence the response time of the target words; however, for the lexical decision priming task at SOA 230ms, the higher the working memory, the higher the accuracy. Conclusion:The processing mechanism of inflected forms of Korean verbs in the advanced Korean language learners(level 6 in TOPIK) mainly supports full list hypothesis and working memory positively influences accuracy rates only at SOA 230ms. 목적:본 연구는 한국어 고급 학습자의 과거시제선어말어미 결합 용언 처리기제가 전체목록가설, 완전분해가설, 혼합가설 중 어떤 가설을 지지하는지 살펴보는 데 목적이 있다. 방법:고급 학습자 72명을 대상으로, SOA 43ms와 230ms 점화 어휘판단 과제와 교차양상점화 과제를 실시하여 표적 단어에 대한 반응 시간 및 정답률을 살펴보았으며, 작업 기억 과제를 실시하여 작업 기억이 반응 시간 및 정답률에 미치는 영향을 검토하였다. 결과:SOA 43ms와 230ms 과제의 고빈도 및 저빈도 어간에서, 교차양상점화 과제의 저빈도 어간에서 전체목록가설을 지지하는 결과를 보였으나, 교차양상점화 과제의 고빈도 어간에서 완전분해가설을 지지하는 결과를 보였다. 이어서 학습자의 작업 기억 영향은 SOA 230ms 점화 어휘판단 과제에서의 정답률에서만 나타났다. 결론:한국어 고급 학습자의 과거시제선어말어미 결합 용언 처리 기제는 주로는 전체목록가설을 따르며 학습자의 작업 기억은 의식적인 처리가 가능한 환경(SOA 230ms)에서 작업 기억이 높을수록 정답률이 높아지는 양상을 보였다.

과거시제선어말어미 결합 용언의 처리 기제-점화실험 연구-

이선진,이은하,임우열,김영주 국어국문학회 2018 국어국문학 Vol.0 No.185

This present study aims at examining the way inflected forms of Korean verbs are processed: whether the mental representation of inflected forms can be accounted for by a full-list model or a decomposition model. To meet the goal, eighty-four Korean native speakers took a masked priming task with SOA of 43ms, SOA of 230ms and cross-modal priming task. To examine the effects of response time on the target word according to the condition of frequency of the stem(high, low) and priming word(identical condition, experimental, high frequency final ending, low frequency final ending condition, control condition). As a result, the frequency of the stem affected the response time in the masked priming task with SOA 43ms(high frequency of stem), supporting the decomposition model while the frequency of the stem did not affect the Eojeol recognition in the masked priming task with SOA of 43ms(low frequency of stem), supporting the full-list model. Overall the results of SOA 43ms supported the hybrid model. The results of SOA 230ms and cross-modal model(both high and low frequency of stem) showed to support the decomposition model. 본 연구는 과거시제선어말어미가 결합한 용언의 활용형이 심성어휘집에서 전체 어절 단위로 인식되는지, 형태소 단위로 분해되어 인식되는지를 살펴보는 것을 목적으로 한다. 이를 위해 모어 화자를 대상으로 SOA 43ms와 230ms 점화 어휘판단 과제와 교차양상점화 어휘판단 과제를 실시하여 표적 어간 빈도와 점화 조건에 따라 표적 단어에 대한 반응 시간에 미치는 영향을 살펴보았다. 그 결과, SOA 230ms 점화 어휘판단 과제와 교차양상점화 어휘판단 과제의 고빈도 및 저빈도 어간에서 그리고 SOA 43ms 점화 어휘판단 과제의 고빈도 어간에서 완전분해가설을 지지하였다. 반면, SOA 43ms 점화 어휘판단 과제의 저빈도 어간에서는 전체목록가설을 지지하는 결과를 보였다.

의미적 맥락과 단어빈도가 한국어 시각단어 재인에 미치는 영향: 점화연구

이은하,이선진,임우열,김영주 국어국문학회 2019 국어국문학 Vol.0 No.186

The present study aims at investigating how semantic context and word frequency have effects on visual word recognition in Korean, and how these two factors interact with each other. A primed lexical decision task was implemented on 144 Korean native speakers with three conditions: visual target verbs paired with visual prime nouns with SOA (stimulus onset asynchrony) 43ms inbetween, visual targets paired with visual primes with SOA 230ms inbetween, and visual targets paired with auditory primes with no delay inbetween. The results revealed that, (i) high-frequency targets were recognized faster than low-frequency targets; (ii) targets preceded by semantically associated primes were recognized faster than those preceded by not associatively related primes; (iii) greater priming for low-frequency targets was found only at the SOA 43ms condition without statistically significant interactive effects of semantic context and word frequency; finally, (iv) no statistically signigicant facilitatory semantic priming effects were shown for both high- and low- frequency words at the cross-modal condition despite signigicant main effects for semantic congruency. The findings suggest that semantic context and word frequency have significant influence on visual word recognition in Korean although the interaction between semantic context and word frequency may occur only at the initial period of automatic semantic priming, and that auditorily presented primes are likely to be converted to the other modality during reading visual targets in Korean. These results can be considered as a supportive evidence for multistage activation model and conversion hypotheses. 본 연구에서는 시각단어 재인 시 점화-표적 간 의미적 연관성과 단어빈도(표적용언 어간의 빈도)가 어떤 관계를 맺고 있으며, 의미적 맥락과 단어빈도 그리고 두 요인의 관계가 어떤 영향을 미치는가를 고찰했다. 이를 위해 한국어 모어화자 144명을 대상으로 SOA 43ms 차폐점화 어휘판단 과제, SOA 230ms 시각점화 어휘판단 과제 그리고 교차양상 점화 어휘판단 과제를 시행했다. 연구결과를 보면. (i) 모든 과제에서 고빈도 표적이 저빈도 표적보다 더 빨리 처리되었으며, (ii) 모든 과제에서 점화와 의미적으로 관련된 표적을 처리할 때 무관한 표적을 처리할 때보다 반응시간이 짧았고. (iii) 의미적 맥락과 단어빈도 간 상호작용은 SOA 43ms 차폐점화 과제의 저빈도 표적에서만 유의미했다. 끝으로, (iv) 교차양상 과제에서 유의미한 의미점화 효과를 보였으나, 빈도를 나누어 고빈도와 저빈도 표적을 살펴보았을 때는 유의미하지 않았다. 본 연구의 결과는 단어빈도와 의미적 관련성이 한국어 시각 어휘처리에 영향을 주나 자동적 의미점화의 초기시점에서만 의미적 맥락과 단어빈도가 상호작용적 관계를 나타내며, 점화와 표적이 서로 다른 감각경로로 제시될 경우 점화가 표적의 감각경로로 변환됨을 시사한다. 이러한 결과는 어휘처리의 다단계 활성화 모형과 변환 가설을 지지하는 것으로 해석할 수 있다.