Intelligibility is considered as a critical component of second language acquisition. Understanding the effects of foreign accented speech is important for both theoretical inquiries and evaluation on the nature of intelligibility. Much of previous re...
Intelligibility is considered as a critical component of second language acquisition. Understanding the effects of foreign accented speech is important for both theoretical inquiries and evaluation on the nature of intelligibility. Much of previous research, investigating Korean accented speech has been limited to measuring gross intelligibility or accuracy by English speaking listeners. This study examines the assessment of intelligibility, comprehensibility and accentedness of Korean speakers by different L1 background listeners. It also figures out the effect of pronunciation errors on intelligibility and compares the findings with Jenkins' Lingua Franca Core(LFC).
This study elicited speech samples from 10 Korean undergraduate and graduate students of varying proficiency and 1 English speaker. These speakers were recorded speaking freely about 4 different easy topics. 44 sentences were extracted and the stimuli were played to 20 listeners from 4 different L1 backgrounds(5 each Korean, Mandarin, Filipino and English listeners). They were asked to transcribe orthographically sentence level stimuli(Intelligibility) and also asked to judge each stimulus on 9 subjective Likert-type rating scales regarding how easy the stimulus was to understand(Comprehensibility) and how heavily foreign accented it was(Accentedness). Additionally participants checked the meanings of some key words which were considered to influence the intelligibility rating scores. They performed the lexical decision test(LDT) to assess each listener's lexical and orthographic knowledge.
The effects of listener, speaker, LDT, and familiarity factors were compared by means of MANOVA and two-way ANOVA analysis. On intelligibility measures, the Korean listeners had significantly less difficulty in understanding speakers with Korean accented speech than Mandarin and Filipino listeners. However, English listeners had significantly higher scores than any other groups. All nonnative listeners who found the native speaker's stimuli less intelligible showed higher intelligibility scores on Korean speakers' stimuli. Familarity with Korean accented speech was found not to affect intelligibility scores. No significant differences were noted on measures of comprehensibility and accentedness among listener groups. Also, no significant correlations were found on measures of intelligibility and accentedness, suggesting that heavy foreign accents are not strongly correlated with listeners' intelligibility.
The researcher and an English instructor identified pronunciation and morphosyntactic errors. Rhythm and intonation were rated by two English instructors. Many pronunciation errors, frequent in Korean speakers, were found. The result shows that segmental errors had correlation to intelligibility scores of Korean and American listeners only. 8 out of 40 Korean speakers' stimuli were found to be more intelligible to English listeners than native speaker's stimuli. Good rhythm and intonation scores do not correspond to the intelligibility measures, which corresponds to Jenkins' study. The segmental errors and word-level intelligibility scores were also compared to investigate the influence of each segmental errors. Substitution and deletion had more detrimental effect on intelligibility than addition of a phome. Some vowels differing vowel quality and application of Korean phonological rules were found to be problematic, which differs from Jenkins' LFC.
In conclusion, Korean speakers' stimuli had many errors and were less intelligible to Mandarin and Filipino listeners than Korean and American listeners. This suggests that in classrooms, pronunciation teaching needs to be enhanced both by focusing on accuracy and by giving students more opportunity to be exposed to various accents of English.