A Corpus—based Stylistic Study of Two English Version of The Three Character Classic

饒瓊

Abstract：This paper use Antconc， along with statistical methods， to make a quantitative and qualitative analysis of the type/token ratio， subject personal pronouns， and mean sentence length in two English version of The Character Classic by Herbert Giles and Guo zhuzhang from a corpus-based approach. The five aspects show the translators style.

Key words：corpus； the English versions of San Zi Jing； translators style； constrastive analysis

1. Introduction

The Three Character Classic is one of the Chinese classic texts. It was probably written in the 13th century to introduce children to the history of China， the great works of philosophy， and the basics of Confucian ethics. The most well-known English translation of the text was completed by Herbert Giles in 1900 and revised in 1910. The latest translation version from A Chinese-English Collection of the Best Chinese Traditional Primers was compiled by Guo Zhuzhang， who translated this book in modern Chinese and English respectively detailed notes and comments. This book has been well recognized by readers and attracted peoples attention to elementary education. San Zi Jing， as one of the passages is representative. Therefore， this paper chooses two translation versions by Giles and Guo zhuzhang to analyze translators style.

2. Literature review

Translator Style separates the broad sense and the narrow sense. In a broad way， the translator style refers to the translators personality in terms of language use and non-linguistic features， including selection of source texts， translation strategies application， prefaces and postscripts and translation of the analects etc. Narrowly speaking， translator style refers to translators preferences in terms of language application or language expression， or recurring language expressing in translation versions （Hu， 2011， P.19）.

Applying corpus and corpus retrieval tools to stylistic studies is a new research tool of translation style. On the platform of TEC， Mona Baker （2000） once described and analyzed the difference of translation style of the British translators Peter Bush and Peter Clark from three aspects， including type/token ratio， mean sentence length and narrative.

Recently， some domestic scholars do some researches on San Zi Jing. However， almost no one uses corpus to research translators style. So， this paper will use the corpus and compare some statistics to analyze the two translators style.

3. Research Method

In this study， the author references Bakers research methods and builds a comparable corpus to do empirical research. In order to ensure objectivity and comprehensiveness， self-built comparable corpus selects San ZI Jings translation versions of Guo Zhuzhang and Giles. The corpus retrieval and analysis software used in the research is Antconc 3.2.1 which compares and analyze the corpora in the corpus from type/token ratio， personal pronouns， mean sentence length and dominant connected components in order to understand different translators style.

4. Results and Discussion

（1） Type/token ratio

Type refers to the number of different words in a specific text and token refers to the total numbers of specific text. Type/Token Ratio （TTR） refers to the ratio between the type and token in a specific text. The separate number of type and token does not reflect the essential feature of the text， but their ratio reflects the nature of the text to a certain extent， that is the variability of words （Yang， 2002）. When comparing the translation by means of TTR， the ratio of the difference texts shows the degree of translators influenced by source texts. When the number of the corpus surpass 1000， standard type/token ratio （STTR） is used to measure the changeability of words. The table 1 is TTR statistics.

As is shown in the table 1， the Giles token is slightly larger than Guos， but TTR is lower than Guos. This shows that the length of the translation from Giles version is greater than Guos translation， but vocabulary diversity is not as good as Guos translation. In contrast， Guos Translation carries more information， vocabularies use is richer， and is more readable.

（2） Subject personal pronouns

Sinclair said that comparing high frequency words is one of the important methods to research language based on corpus. It can provide researchers with the basic information of the text. In general， the high frequency words of each text remain steady. Therefore， the subtle difference of order may lead to greatly different style （Sinclair， 1991， P.30）. The results are shown in table 2.

According to the table 2， in the two versions of Three Character Classic， The raw frequency of personal pronouns as subject is 30 and38 respectively. Nevertheless， the ratio between the frequency and the total numbers can reflect the differences of personal pronouns as subject in two translation versions （1.33<1.56）. The figures shows that Guo uses personal pronouns as subject much less compared with Giles in the translation version of San Zi Jing. This might be that Guo breaks the habits of English grammars and imitates Chinese grammar structure directly. In concrete terms， ancient Chinese has higher hypotaxis and often makes the subjects in the grammar structure more implicit. The agent is deduced from context and logic relation （王力， 2002）. In contrast， English has higher parataxis and the agent is generally expressed by definite subject. Therefore， when translating ancient Chinese into English， the explicitation of personal pronouns as subjects should increase. However， in Guos translation， the phenomena of explicitation of the personal pronouns as subjects is not clear. He replaces the subject personal pronouns with content nouns.

（3） Mean sentence length

The mean sentence length refers to the average length of sentence in specific text. In general， shorter the mean sentence length is， the easier the syntactic structure of the text is， and sentences are easy to comprehend. Otherwise， syntactic structure of the text is more complicated， and sentences are more difficult to understand. According to Antconc， the sentence information statistics of two translation versions is shown as follow.

We can see from the table， Guos mean sentence length is 14.37 words， but Giless mean sentence length is 15.12 words. The two translation version use more moderate sentence length and are helpful for readers to understand. In concrete term， Giless sentence uses are more complicated and the change of sentence length is more significant. The main cause of this difference is that Giles is very faithful to convey the original meaning and increases the number of modified and explanatory language， resulting in increase in the sentences number of its translation version and complicated syntax. And Guo is pursuing to reproduce the original texts.

5. Conclusion

Based on the above comparison and analysis of the corpus， although Guos translation length is shorter， the type/token ratio is higher. Therefore， Guos translation version carries a greater amount of information. Giless translation texts is longer， but the type/token ratio is lower. Therefore， the information carried by Giles translation is lower. In contrast， Giless mean sentence length is longer， and the sentence structure is more complicated， which improve the difficulty of reading. However， Guos mean sentence length is shorter， and the simple sentence， make it a lower reading difficulty. The reason is mainly because Giles takes the translation strategy of explicitation during the process of translation.

By using corpus retrieval software to retrieve and analyze relevant data， we can research different translation style of translators. This empirical research combines quantitative analysis with qualitative analysis， which will be helpful to grasp the original style and translation style objectively. Corpus-based translation style research avoids the subjectivity and commentary research model in traditional translation criticism， and it has a strong objectivity and persuasiveness. Its application provides a new method for the translation study.

References

[1] Giles， H. （1900）. Elementary Chinese San Tzu Ching. Three Character Classic.

[2] 郭著章. 漢英對照蒙學(xué)經(jīng)典[M]. 武漢： 武漢大學(xué)出版社， 2004.

[3] 胡開寶. 語料庫翻譯學(xué)概論[M]. 上海： 上海交通大學(xué)出版社， 2011.

[4] 楊惠中. 語料庫語言學(xué)導(dǎo)論[M]. 上海： 上海外語教育出版社， 2002.

[5] 趙彥春. 《三字經(jīng)》英譯詰難與譯理發(fā)凡[J]. 天津外國語學(xué)院學(xué)報， 2014.