基于语料库的《道德经》英译本翻译风格研究

论文总字数：29242字

摘 要

《道德经》作为诞生于2500多年前的中国经典文化作品，在西方世界受到广泛关注。自古以来，已有无数研究者对《道德经》的释义和翻译做出贡献。本论文用语料库研究的方法，选取刘殿爵译本、冯家福amp;简·英格里希合译本、亚瑟·韦利译本作为研究对象，对这些经典翻译的翻译风格的进行研究。

本论文使用自建小型单语语料库，以WordSmith作为工具，从词汇、句法、衔接等角度进行研究。词汇方面，本论文从词汇复杂度、词汇密度等方面进行研究。句法方面，本论文研究了各译本的平均句长和句长的标准差。衔接方面，本论文主要考察了三个译本的两种衔接手段。

研究发现，目标读者在翻译风格的一些方面具有一定影响。对于不同的目标读者，译者可能采取归化的方法，使用较为简单的用词和符合英语习惯的句式，或选择保留《道德经》原始风格，使用较为正式和简洁的句式，从而导致译文具有明显不同的风格。

关键词：《道德经》；翻译风格；语料库研究

Table of Contents

Chapter One Introduction 6

Chapter Two Literature Review 7

2.1 Previous Researches on Tao Te Ching 7

2.2 Corpus-based Translation Style Study 8

Chapter Three Methodology 9

3.1 Research Questions 9

3.2 Data Collection 9

Chapter Four Corpus-based Analysis 10

4.1 Lexical Analysis 10

4.1.1 Lexical Variability 10

4.1.2 Lexical Density 11

4.1.3 Summary 12

4.2 Syntactical Analysis 13

4.2.1 Sentence Length 13

4.2.2 Standard Deviation of Sentence Length 13

4.2.3 Summary 14

4.3 Cohesion Analysis 14

4.3.1 Personal References 14

4.3.2 Conjunctions 16

4.3.3 Summary 18

Chapter Five Conclusion 19

5.1 Major Findings of This Research 19

5.2 Limitations and Prospects for Future Research 19

References 20

1. Introduction

Tao Te Ching has long been regarded as a book full of wisdom of the past. Over the past centuries, countless attempts have been made to introduce this Tao Te Ching to other countries. Yet, the arcane and ambiguous wording of Tao Te Ching has set up a barrier for readers and translators to deduce its true meaning from its words. No people can guarantee that they have completely uncovered the whole wisdom of Tao Te Ching. Thus, the translation of Tao Te Ching has become an endless endeavor that can only be bettered, but can hardly become perfect.

In the past, a considerable number of analyses has been made to compare different translations of Tao Te Ching, and to find out the features of each translations, which serves as a guide for future translations. Some of these studies look into the handling of cultural images, while many others concentrate on the translation of certain key words. Corpus-based studies have also gained popularity in recent years, which should apparently be attributed to recent development of computer technology. This approach allows a more measurable result, and thus is less biased and more credible than the traditional approach. In this paper, a corpus-based approach is taken to inspect the samples from the dimensions of words, sentences, and overall styles. This paper also includes a more specific view on particular representative translations. It is believed that in this way, a more thorough view of translations can be produced, which makes it more helpful for future translators.

In the first part of this paper, the use of words is discussed. For each translation, lexical variability and lexical density are assessed. A case study is made to compare the translation that gets the highest score in the assessment with one that gets an average score.

The second part analyzes the sentence uses in the translations. This part computes the sentence length and complexity. A case study is similarly conducted to find out how different styles of sentences affect the outcome of translation.

The third and last part discusses the overall styles of the translations in terms of a specific language use in the translation.

1. Literature Review
3. 1. Previous Researches on Tao Te Ching

Tao Te Ching has been an attractive researching topic in China as well as in other countries for many years. In other countries, studies tend to view Tao Te Ching from the perspective of philosophy and Asia cultural studies. While in China, a linguistic approach is often taken.

In China, the study of variety in translation has been found and studied no later than 1992. Wang (1992) studies four translations of Tao Te Ching, and notes their vast difference. He attributes the cause of this variety to the varying viewpoints of the translator. He further discusses the translation of the key word 道 (Tao), which he claims to be very difficult to translate. He points out that with the rich connotation of 道, it is impossible to find a matching word in English, and then he discusses possible ways of translating 德 (Te). In the last part of his paper, he reviews the history background of Tao Te Ching, and observes that this work was already hard to interpret two thousand years ago, and there’s no wonder that the translators’ understanding is necessarily involved during the translation.

Liao (2004) provides a more structural explanation for the variety of translation. In particular, he points out that the open-endedness of Tao Te Ching enables different ways of interpretation. He then notes that Vorurteil, or prejudgement of the translator plays an indispensable role in the production of the translation. Vorurteil may include the translator’s knowledge of ancient Chinese culture, as well as their own cultural background. He also points out that the subjectivity of the translator helps realize the variety.

Zhang amp; Bao (2010) perform a classification and review of the various English versions of Tao Te Ching. They categorize each version based on the philosophy behind each version, and notes that the variety of these versions is based on the variety of their understanding of God. They point out that Tao Te Ching is both translatable and untranslatable, as its basic ideas are generic enough to be understood by people around the world, but the complete implication cannot be replicated in translated versions.

Together, these works are good sources for recognizing the variety of the translations, and to find out the cause of the variety.

• 1. Corpus-based Translation Style Study

Mona Baker is regarded as a pioneer in Corpus-based translation studies (Han amp; Jiang amp; Yuan 2019). In Towards a Methodology for Investigating the Style of a Literary Translator, Baker (2000) defines "style of translation" as "a kind of thumb-print that is expressed in a range of linguistic — as well as non-linguistic — features", that is, "recurring patterns" in the translator's use of language. He further notes that it is important to access the "forensic stylistics" that lie in the translators works. Further, Baker selects a specific set of factors to assess the styles of the selected works. Some of these factors, such as type/token ratio and average sentence length, are still commonly used in corpus-based translation style studies today.

Works have been done to apply corpus-based translation style study in the research about Tao Te Ching. Notably, Zhao (2015) compares the style of translations produced by John C. H. Wu and by Arthur Waley from three levels: word, sentence and discourse. In word and sentence levels, Zhao uses a similar set of factors as used by Baker, while in discourse level, Zhao assesses the use of personal pronouns of the translations. Zhao attributes the differences in translation styles to the translators' understandings of Tao Te Ching, the purposes of translation, and the strategies used in translation.

The above works have a deep influence in the methodology used in this paper.

1. Methodology
2. 1. Research Questions
3. Is there any significant difference in translation styles, in different translations of Tao Te Ching?
4. If the differences exist, what are the differences in translation styles in translations of Tao Te Ching?
5. If the differences exist, what are the effects of these differences in translation styles?
   1. Data Collection

Data are collected using WordSmith 4.0 and homemade programs. WordSmith is an integrated set of programs created and maintained by Scott M. at Oxford University. It features three main programs: Concord, KeyWords, and WordList. Concord is a program that makes concordance based on the given text. KeyWords provides functionality for locating a predetermined set of key words in the text. WordList is a tool that outputs a frequency list of all words in the text files. In this study WordSmith is used to compute the type-token ratio of the texts.

This study also uses TreeTagger to tag the part-of-speech for each word, in order to study the lexical density of each text. TreeTagger is a tool developed by Helmut Schmid for annotating text with part-of-speech and lemma information.

1. Corpus-based Analysis
2. 1. Lexical Analysis

In this section, several lexis-related features are examined to uncover the style of each translation.

• • 1. Lexical Variability

In corpus linguistics, lexical variability, also known as vocabulary richness, is usually quantified by the type-token ratio (TTR), where type is the number of distinct words in the text, and token is the total number of words. A high type-token ratio indicates that the writer uses a wider range of words, while a low type-token ratio represents a more restricted vocabulary. Thus, type-token ratio is proportional to the diversity of words in a text. Baker (2000) points out that texts addressed to non-native English speakers tend to have a lower type-token ratio than a native speaker.

However, when the lengths of texts are different, type-token ratio may become less reliable, as longer texts naturally have more repetition of words, and thus are characterized with a lower TTR than shorter texts. In this case, standard type-token ratio (STTR) may be a more reliable figure. Standard type-token ratio is obtained by first computing the type-token ratio for every N words in the text, then taking the average of all computed TTR, where N is usually set to 1000.

The following table presents the statistics of basic lexical data of the translations in question.

Table 1

As demonstrated in the table, while Lau’s translation has over a thousand more words (types) than Feng’s work, their tokens are much closer in quantity. Waley’s version has the highest number of types as well as tokens among all three versions. Thus, there’s no wonder that Lau’s work is much lower than Feng’s in terms of both TTR and STTR. But it’s probably more surprising to see that while Waley’s work has a much lower TTR than Feng’s, its STTR is not much lower than Feng’s. This probably is an indication that the difference in TTR between Waley’s and Feng’s works is primarily attributed to the difference in length, and not so much the difference of their inherent lexical variability.

• • 1. Lexical Density

Lexical density is defined as the proportion of notional words among all words. In this study, the part-of-speech of each word is tagged using TreeTagger with English parameter file (BNC tagset) downloaded from http://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/data/english-bnc.par.gz, then the number of notional words are counted using regular expression.

The following table summarizes the statistics related to lexical density.

Table 2

As is illustrated in the table, Feng’s version has an apparently higher level of lexical density than Lau’s, which in turn has a higher level of lexical density than Waley’s. That is, Feng’s translation has the highest proportion of notional words and the lowest proportion of function words among all three, while Waley’s has the lowest proportion of notional words and the highest proportion of function words.

A lower lexical density indicates that the text is more accessible to the reader, while a higher lexical density indicates than the text is more difficult to comprehend. Thus, the data signifies that Waley’s version is the most reader-friendly. However, as Tao Te Ching is an arcane work from ancient time, one can also conclude from the data that Feng’s version is better in reproducing the original style of Tao Te Ching when compared with the other two translations.

• • 1. Summary

The above data show that Feng’s translation is higher in both vocabulary richness and density. While Waley’s translation has a higher lexical variability than Lau’s, it has a lower lexical density. That is to say, Feng’s translation has a richer set of words and a denser use of notional words, which preserves a mysterious aura inherited from the original work of Tao Te Ching, while Lau’s and Waley’s translations are easier to read.

The difference in style is probably caused by two factors: the translators’ background and the target readers. As noted by Baker (2000), texts from a non-native speaker tend to have lower STTR than those from a native speaker. Thus, although D. C. Lau is a professional in English translation, his work still demonstrates characteristics of a work from a non-native English speaker. On the other hand, both of the other versions in question have passed though the hands of native speakers (Waley is a native speaker, while Feng’s is a cooperative work by him and his wife, a native speaker), and therefore has a similar level of STTR. The difference in target readers probably contributes to the difference in lexical density. Although all translations are aimed at introducing the work to foreign readers, their exact targets are not the same. Waley aims at making Chinese literacy more accessible to Western readers, who are not familiar with this topic. Feng, on the other hand, aims at introducing the work to people who are interested in ancient Chinese philosophy. Thus, Waley’s work is simpler so it can be read by a wider range of people, while Feng’s work is more formal and complex, to keep the aura of Chinese mystery which appeals its readers. Both styles turn out to be successful, as both are widely accepted by people at their time.

• 1. Syntactical Analysis
     1. Sentence Length

Sentence length has traditionally been used as an indicator of the sentence’s complexity. In this study, WordSmith is used to compute the mean sentence length of the analyzed texts. The results are shown below.

Table 3

While the mean sentence length of Lau’s and Waley’s translations is very close (the difference is as low as 0.1), the mean sentence length of Feng’s translation is vastly different. This indicates that Lau’s and Waley’s translations have a much higher tendency to use complex sentence structure than Feng’s, while the latter features much more concise sentences.

• • 1. Standard Deviation of Sentence Length

Standard deviation is a measurement for the variation of a set of data values. A high standard deviation signals that the data have a high degree of dispersion, while a low standard deviation signals that the data tend to be close to the mean.

As can be seen from the above statistics, Feng’s translation has the lowest value of standard deviation, while Lau’s has the highest value. That is to say, Feng’s translation is more monotonic with less variation in sentence length, while Lau’s translation features a diversified use of long and short sentences.

• • 1. Summary

From the analysis, it can be found that Lau’s and Waley’s translations both have higher sentence length on average, with more diversified sentence structure, while Feng’s translation is characterized with shorter sentences and monotonic sentence structure. Further inspection shows that Lau’s and Waley’s translations are much richer in the use of subordinary conjunctions. This indicates that Feng chooses a much different translation style from the other two. While the other two more or less attempt to explicitate the relationship between sentences by adding conjunctions, Feng decides to keep the sentences separated. By using shorter sentences, and keeping the relations between sentences implicit, Feng effectively produce a style that is more commonly seen in Chinese texts, especially ancient Chinese texts, than English texts. This, again, reflects the difference between the intended targets of each translator as elaborated in the previous section.

• 1. Cohesion Analysis

Cohesion is a noteworthy concept in discourse study. It signifies the relations between sentences in a discourse. In this section, two major forms of cohesion are analyzed to find out the translation styles of the works as expressed in discourse level.

• • 1. Personal References

The use of personal pronoun is one of the significant differences between Chinese and English. In Chinese, personal pronouns tend to be omitted more frequently than in English. Because of this, it is common for translators to add pronouns when translating Chinese texts to English. The following table summarizes the frequency of personal pronouns in each assessed translation.

Table 4

The data shows that the three translations have a roughly similar proportion of personal pronouns, which helps to clarify the structure of the text, instead of requiring the readers to infer from the context. Of these personal pronouns, "it" and "he" are most commonly used, while "she" is least common. This is probably due to the nature of the source text, as its topics are largely concerning abstract concepts and lifeless objects, while gender topic is out of scope.

It can also be observed from the data that the assessed translations have some difference in their use of personal pronouns. Overall, Feng’s translation has a less frequent use of personal pronouns than the other two translations, while Waley’s translation uses personal pronouns slightly more frequently than Lau’s. The choice of words is also different, Lau’s translation tends to use "it" more often, while Waley seems to prefer the word "he". This can again be attributed to the cooperative effect of translators’ own styles and the target readers. While the diversity in the choice of specific words can largely be attributed to the personal preference of the translators, the relative scantiness of personal pronouns in Feng’s translation might be a characteristic feature that makes the translated text have a style closest to that of the original text.

Example 1: Source text: 是以圣人处无为之事，行不言之教，万物作焉而不辞，生而不有，为而不恃，功成而不居。夫唯不居，是以不去。 (Chapter 2)

Lau’s translation: Therefore the sage keeps to the deed that consists in taking no action and practises the teaching that uses no words. The myriad creatures rise from it yet it claims no authority; It gives them life yet claims no possession; It benefits them yet exacts no gratitude; It accomplishes its task yet lays claim to no merit. It is because it lays claim to no merit that its merit never deserts it.

Feng’s translation: Therefore the sage goes about doing nothing, teaching no-talking. The ten thousand things rise and fall without cease, Creating, yet not. Working, yet not taking credit. Work is done, then forgotten. Therefore it lasts forever.

Waley’s translation: Therefore the Sage relies on actionless activity, Carries on wordless teaching, But the myriad creatures are worked upon by him; He does not disown them. He rears them, but does not lay claim to them, Controls them, but does not lean upon them, Achieves his aim, but does not call attention to what he does; And for the very reason that he does not call attention to what he does He is not ejected from fruition of what he has done.

From the example, it can be seen that while each translation has a distinct way of using personal pronouns, the end density of personal pronouns in Lau’s and Waley’s translations are similar. Feng’s translation, on the other hand, uses nearly no personal pronouns, which ensures that the translation has a similar level of concision (and ambiguity) to the original text.

• • 1. Conjunctions

Compared to English, ancient Chinese tend to be much more paratactic. In the original text of Tao Te Ching, conjunctions other than gu (故) and er (而) are more often omitted than not when forming sentences. In English, however, conjunctions are usually a necessity to connect words, phrases, or clauses. Thus, translations tend to add conjunctions such as "and", "though", "if", "therefore" to reveal the implicit relations between clauses. The following table summarizes ten most frequently used conjunctions in each assessed version.

Table 5

Overall, all three translations have a similar proportion of conjunctions. This suggests that all these translations use the strategy of adding conjunctions to explicate the underlying relations of the original texts, so that their translations flow more naturally.

Example 2: Source Text: 化而欲作，吾将镇之以无名之朴，镇之以无名之朴，夫将不欲。不欲以静，天下将自定。(Chapter 37)

Lau’s translation: After they are transformed, should desire raise its head, I shall press it down with the weight of the nameless uncarved block. The nameless uncarved block Is but freedom from desire, and if I cease to desire and remain still, the empire will be at peace of its own accord.

Feng’s translation: If they still desired to act, they would return to the simplicity of formless substance. Without form there is no desire. Without desire there is tranquility. And in this way all things would be at peace.

Waley’s translation: And if having been transformed they should desire to act, we must restrain them by the blankness of the Unnamed. The blankness of the Unnamed Brings dispassion; To be dispassionate is to be still. And so, of itself, the whole empire will be at rest.

It can be observed from this example that while the source text does not use any conjunctions, all three translations choose to add conjunctions to improve the fluency and cohesion of the translated texts.

• • 1. Summary

The study of personal references and conjunctions confirms some of above conclusions. In the study of personal references, Feng’s translation exhibits a different characteristic from Lau’s and Waley’s translations. As with its sentential characteristics, the low frequency of personal pronouns helps Feng’s translation retain a style that directly resembles the style of original text. However, all three translations have a similar proportion of conjunctions, which suggest that all of them take care to improve the clarity and coherence of the translation by adding conjunctions.

1. Conclusion
2. 1. Major Findings of This Research

At lexical level, statistics of lexical variability and lexical diversity are collected. The major findings are: (1) the translator’s background affects the lexical variability of the translated text. Thus, Lau’s translation has a lower STTR than the other translations. This factor, however, does not manifests in other parameters. (2) Target readership affects the lexical diversity of the text. To arise the interest of the readers, a simpler text is more effective. However, to readers who are already interested in Chinese philosophy, a version with higher lexical density might be more appealing.

At sentential level, statistics of sentence count, mean sentence length and standard deviation of sentence length are collected. It is found that target readership plays a similar role as in lexical level. In primary stage, a translation with more conjunction is welcomed, while in a later stage, a version with shorter sentence length on average that better reproduces the original style of the text is preferred by people.

• 1. Limitations and Prospects for Future Research

Due to the limitation of the author, the research embodies a lot of shortage that can be improved in the future.

First, the subject of this study is limited. Due to various constraint, this study is limited to only three translations. Future researches could investigate more versions to get a more complete and thorough view of this topic.

Second, the methods used in this study is lacking in breadth and depth. With modern computer technology, more sophisticated statistical methods and parameters can be introduced to facilitate more accurate and fine-grained analysis of translation styles.

Third, the author’s understanding of Tao Te Ching is limited, which severely hinders this research from becoming more in depth. If the research could be carried on by someone more knowledgeable about this ancient Chinese masterpiece, a more detailed analysis would be possible.

References

1. Baker, M. Corpus linguistics and translation studies: Implications and applications[J]. In M. Baker, G. Francis, E. Tognini-Bonelli (Eds), Text and technology: In honor of John Sinclair. Amsterdam: John Benjamins. 1993.
2. Baker, M. Towards a Methodology for Investigating the Style of a Literary Translator[J]. Target, 2000, 12(2): 241-266.
3. Laviosa S. Core patterns of lexical use in a comparable corpus of English narrative prose[J]. Meta: journal des traducteurs/Meta: Translators' Journal, 1998, 43(4): 557-570.
4. Bebell DJ, Fera SM. Comparison and Analysis of Selected English Interpretations of the Tao Te Ching[J]. Asian Philosophy. 2000,10(2):133-147. doi:10.1080/713650897.
5. Halliday M.A.K., Hasan R. Cohesion in English[M]. London: Longman Group Limited.
6. 韩红建,蒋跃,袁小陆.大数据时代的语料库译者风格研究[J].外语教学,2019,40(02):88-93.
7. 王芳,任翠翠.《道德经》英译译本研究[J].语言教育,2014,2(04):61-66.
8. 朱安博."译可译,非常译"——汪榕培教授访谈录[J].山东外语教学,2013,34(03):3-6 28.
9. 王越西.适应与选择——从生态翻译学视角研究亚瑟·威利之《道德经》英译[J].东北师大学报(哲学社会科学版),2012(02):127-130.
10. 文军,罗张.《道德经》英译研究在中国[J].上海翻译,2012(01):19-23.
11. 辛红娟.《道德经》英译文本的另类阐释[J].中国外语,2011,8(02):93-97.
12. 张小钢,包通法.《道德经》英译版本的归类及思考[J].江南大学学报(人文社会科学版),2010,9(02):115-120.
13. 辛红娟."文化软实力"与《道德经》英译[J].外语与外语教学,2009(11):50-52.
14. 郑海凌.老子思想的翻译与传播[J].外语与外语教学,2008(11):54-56.
15. 班荣学,梁婧.从英译《道德经》看典籍翻译中的文化传真[J].西北大学学报(哲学社会科学版),2008(04):162-166.
16. 辛红娟,高圣兵.追寻老子的踪迹——《道德经》英语译本的历时描述[J].南京农业大学学报(社会科学版),2008(01):79-84.
17. 廖敏.试析《道德经》翻译的多样性[J].西南民族大学学报(人文社科版),2004(09):333-336.
18. 陈国华,轩治峰.《老子》的版本与英译[J].外语教学与研究,2002(06):464-470 480.
19. 苗玲玲.译可译,无常译——谈《道德经》翻译中的译者主体性[J].学术研究,2002(08):134-137.
20. 费小平.《老子》六译本选评与中国传统文化名著重译探讨[J].贵州教育学院学报(社会科学版),2001(01):61-65 107.
21. 崔长青.谈《道德经》英译[J].读书,2000(12):106-110.
22. 崔长青.《道德经》英译本初探[J].国际关系学院学报,1997(03):51-56.
23. 王平.比较《老子》的三种英译[J].外语与外语教学,1996(S1):57-60.
24. 李贻荫,金百林.D．C．Lau妙译《道德经》[J].外语研究,1995(02):42-44 46.
25. 汪榕培.译可译,非常译——英译《老子》纵横谈[J].外语与外语教学,1992(01):25-30.

剩余内容已隐藏，请支付后下载全文，论文总字数：29242字