The purpose of this study is to present problems of translating skaz narrative novels, which ignore their narrative genre. May(1994) contended that many English translations of Russian skaz novels showed the problems through the translation strategies such deleting and modifying some key words indicating that the novels are skaz. This study is to show that those are not only the problem of English translation of Russian skaz but also the one observed in Korean translations of English skaz. Furthermore, two reasons of those translation phenomena will be discussed.

To do this, the researcher chose The Adventures of Huckleberry Finn written by Twain in 1884 and its nine Korean translations as the source text and target texts respectively. In the source text, there are four linguistic elements that can be considered as the skaz narrative devices, which are present tense, second person narratee, pragmatic markers, and regional dialects. Those are all used in the narrator's text(NT) by the teenage narrator Huckleberry in the novel. In the other narrative novels however, the four linguistic elements are easily seen in the character's text(CT). Used in the NT, those are indicating that the narrator is the teller of the story and the narrator is a specific person. Considering specific oral features in the NT show orality of the narrative, those four linguistic devices are very important in the novel as skaz.

Through the comparative analysis on the translation, differences were observed between NTs and CTs. 'Present tense' and 'pragmatic markers' also exist in the target language as the same linguistic markers as those in the SL (source language). However, 21% of present tense verbs in NT were translated into past tense unlike those all translated into present tense verbs in CT. In case of pragmatic markers, 52.2% were deleted in the NT compared to those of which 72.2% are translated into the pragmatic markers in the CT. 'Second person narratee' caused each translator's different choice of sentence-endings of target language in the nine TTs. Six of them used '-da' in the NT which does not include the listener in the utterance and only three chose '-a/uh', '-yo' to show the speaker's inclusion of listeners. In the translation of CT, all the translations chose the listener-including sentence-endings such as '-yo'. '-a/uh', '-kuna', etc. In the translation of regional dialects, the strategies applied to each dialect were different. Translation of the dialects of noun words shows no difference between the NT and CT. But in the translation of the dialects of verbs, unlike the translation of NT of which 66% was translated into the written style of TL, 100% was translated into spoken style of TL. In case of eye dialects, differences were observed not between NT and CT but each translation.

Through the comparative analysis, the researcher could verified that there are differences in translation between NT and CT. Those are the same results with what May insisted from English translation of Russian skaz novels. She presented the problems by showing some examples translated from Russian to English. As she mentioned in her book, there are two reasons of the phenomena. May presented linguistic difference between the SL and TL as the first and the problem of translators' choice as the second. In the translations of The Adventures of Huckleberry Finn, the two reasons are observed as well. In case of 'present tense' and 'pragmatic markers', they are transformed into past tense verbs or deleted in NTs even though they are translated into present tense and pragmatic markers in CTs. Those are outcomes of translators' choice. Unlike the two devices, second person narratees are translated into selective sentence-endings, which are different grammatical devices from the ones in ST. Translators also should select the one which they think the most appropriate in terms of the function and role of second-person narratee 'you' in ST. So, the translation of second-person narratee 'you' was due to the linguistic difference from SL and TL. The last one, 'regional dialect' is also the one which needs translator's selective choice. In CTs, translators used TL's regional dialects and spoken style using various words to express the existence of substandard words in ST. But in NTs the trials to use regional dialects are not observed at all.

The translations of four skaz narrative devices in The Adventures of Huckleberry Finn show that the semantic translation of those devices in NTs is true as May claimed. However, the languages which are English and Korean it deals with are different from those in May's research. It is meaningful that the thesis is the first to deal with the translation problems of skaz written in English and translated into Korean.

The translations choosing listener-including sentence endings in NTs such as TT4 and TT9 expressed skaz's dialogicity, orality, and spontaneity more smoothly than the others. And TT5 translated pragmatic markers in the ST into the same grammatical markers in TT and created new eye-dialects in TT in translation of ST's eye dialects. Also, present tense, second-person narratee, pragmatic markers, and regional dialects used in Korean non-translated novels written by Kim Dong-in and Chae Man-sik show that the four narrative devices are all able to be expressed in Korean even in NT. In that sense, translation of The Adventures of Huckleberry Finn as the skaz narrative is possible and new and more creative translations are expected in the near future.