This study aims to examine the different aspects and characteristics of pronunciation errors made by Korean learners from Uzbekistan in order to present the basic materials required for pronunciation training on the Korean phonological system and its features. To this end, the phonological system of Uzbek, the native language of the learners, was organized, and Korean pronunciation errors were predicted by comparing and analyzing the Uzbek and Korean languages. Subsequently, an utterance recording experiment was conducted involving these learners, and the patterns of pronunciation errors were identified by analyzing the collected data. The analysis results revealed many errors in obstruents, including aspirated-sound errors in stop consonants, tense-sound errors in fricative and affricative consonants, and incorrect vowel pronunciation of /ㅣ/ and /ㅡ/. There were gender differences in errors in different phonological phenomena, such as lateralization, nasalization, stem-final post nasal tensification, palatalization, and /ㄴ/ addition. Learners’ proficiency levels also influenced the rates of pronunciation errors. Korean learners committed numerous aspirated-sound errors in stop consonants and tense-sound errors in fricative and affricate consonants. Voice onset time (VOT) and fundamental frequency (F0) were used only to distinguish tense sounds; the F0 of aspirated sounds was low, making them indistinguishable from lax sounds.
Learners frequently replaced /ㅣ/ and /ㅡ/, which is related to the realization of the Uzbek /i/ through the variants [i] and [ɨ]. The study findings confirmed that this variant sound environment affected Korean language learning, and errors in pronunciation of /ㅣ/ and /ㅡ/ became prominent under the influence of preceding consonants. Vowel pronunciation featured the overlapping of vowel spaces between /ㅗ/ and /ㅓ/, which was particularly conspicuous in male learners. This is an error caused by rearranging the vowel spaces because Korean contains one more monophthong than Uzbek. This overlapping of vowel spaces also affected diphthongs, and pronunciation errors that confused or replaced /ㅛ/ and /ㅕ/ were prominent.
The pronunciation error rate of male learners was higher than that of female learners. Female learners showed no difference in pronunciation error rates between intermediate- and advanced-level learners in all phonological phenomena, excluding palatalization. Conversely, male learners, however, showed a significant difference in pronunciation error rates between intermediate learners and advanced learners in all phonological phenomena.
Regarding patterns of pronunciation errors by the phonological phenomenon, no difference in pronunciation error rates was observed between male and female learners in any phonological phenomenon that occurred in the early stages of learning, such as neutralization, consonant-cluster simplification, /ㅎ/ deletion, aspiration, and post-obstruent tensification. However, in phonological phenomena such as lateralization, nasalization, stem-final post nasal tensification, palatalization, and /ㄴ/ addition, male learners’ pronunciation error rates were higher than those of female learners.
The rate of pronunciation errors in neutralization was low at 9.7% among all learners, with no differences based on gender or proficiency level. Among male learners, the pronunciation error rate differed according to the type of final consonants, with fewer errors in bilabial stops and more errors in palatal affricates. Pronunciation errors included closure failure errors and vowel insertion errors, as well as a combination of the two.
The pronunciation error rate for palatalization was 43.1%, which was high, and there were clear differences based on gender and proficiency level. Learners overapplied palatalization and learned less about the morphological condition compared to the phonological condition. Although the number of syllables and the parts of speech had no effect, some errors were noted in words such as “같이.” Due to word-specific factors, male learners had a lower rate of pronunciation errors when the preceding element was /ㅌ/ as compared to /ㄷ/.
The rate of pronunciation errors was higher for the nasalization of liquid sounds than the nasalization of stop sounds, with more pronunciation errors in male learners than female learners. Regarding the nasalization of liquid sounds, female learners made pronunciation errors when the preceding element was /ㄴ/ but tended to apply regressive lateralization to /ㄴ-ㄹ/. There were more nasal pronunciation errors of stop sounds in nominals than in verbals. When the trailing element was /ㄴ/, nasalization was better applied, while there were more errors when it was /ㅁ/. When the consonant-cluster simplification was applied, male learners produced errors in the trailing elements even in an environment where consonant-cluster simplification occurred. However, in the case of female learners, failure in consonant group simplification had a greater impact on pronunciation errors. Female learners realized the simplification of consonant-cluster of conjugation and declension differently, which led to an increase in conjugation errors. Particularly in the case of “ㄺ,” it was pronounced as [ㄹ] or [ㄱ], depending on the grammatical category, but pronunciation errors increased because the exact environment was not familiar.
Pronunciation error rates in lateralization were higher in progressive lateralization than in regressive lateralization, and in progressive lateralization, pronunciation error rates were higher in three-syllable words than in two-syllable words. Regressive lateralization was affected by neither the number of syllables nor the frequency of word usage. When applying the phonological phenomenon, rather than understanding the relationship between consonant and nasal /ㄴ/ influences, the following element was replaced such that the preceding element was affected. Hence, the rate of realization of nasalization was high in progressive lateralization, and nasalization was more prominent in the /ㅀ-ㄴ/ sequence.
The analysis of post-obstruent tensification and stem-final post nasal tensification showed that post-obstruent tensification was mostly applied well when there was no problem in the articulation of obstruents. In the pronunciation of tense sounds, fricatives and affricates were more difficult than stop consonants, but errors in tense sounds were also prominent. Stem-ending post-nasal tensification had a higher error rate than post-obstruent tensification, with male learners making more errors than female learners; there were also clear differences by proficiency level. Pronunciation errors in three-syllable words were higher than in two-syllable words, and tensification was better applied when the trailing element was /ㄴ/. Moreover, when the trailing element was affricate, the pronunciation error was conspicuous.
The pronunciation error rate in consonant-cluster simplification was 43.8%, with no difference by gender, although male learners showed differences by proficiency level. Pronunciation errors were more prominent for verbal than for nominal. However, when consonant-cluster simplification did not occur, as no initial consonant in the element followed the consonant cluster, the rate of pronunciation errors in female learners was found to be higher than that of male learners; the pronunciation errors in nominal were more than that in verbal.
The error rate of /ㄴ/ addition was 82.6%, which constituted a very high pronunciation disqualification rate. The error rate among male learners was higher than that of female learners; additionally, male learners showed differences according to their proficiency level. In /ㄴ/ addition, there were more pronunciation errors when the root character of the two morphemes to be combined was weaker, and the pronunciation error rate was higher when the preceding final consonant was an obstructed consonant or a resonant consonant without anteriority. The rate of realization of /ㄴ/ addition was much higher in the auxiliary verb-binding type.
The pronunciation error rate for aspiration was 32.2%. Aspiration errors were more common in female learners than male learners, but there were no gender differences in the error rates for aspiration. Differences by proficiency level appeared only in male learners. Contrary to other phonological phenomena, aspiration showed opposite patterns of pronunciation errors between male and female learners. Male learners had a high rate of regressive aspiration errors, whereas female learners had a high rate of progressive aspiration errors. The leading consonant type of progressive aspiration is a major factor influencing male learners’ pronunciation, and the pronunciation error rate for /ㅀ/ was higher than for /ㅎ/ or /ㄶ/. However, lagging factors greatly affected female learners, and the error rate was higher for stop consonants than affricates. Females had a high percentage of errors in regressive aspiration, especially of consonants. There was no difference in the error rate when the leading consonants were consonant clusters, but errors in alveolar consonants were prominent in the case of single consonants. Alveolar consonants were the consonants at the articulation position with the highest aspiration-pronunciation error rate in female learners, confirming aspiration as a major factor in pronunciation error. However, male learners’ errors were not influenced by such factors.
From these patterns of pronunciation errors, it can be concluded that, apart from articulation ability, the eligibility rate increases when knowledge education about the application environment is properly performed. Thus, the study has high utilization value in the field of Korean language education as it presents basic data for Korean pronunciation education for learners from Uzbekistan.