摘要
本文基于自建的《水浒传》汉英平行语料库，综合利用语料库语言学、对比语言学、功能语言学及文体学、信息论等理论，从宏观和微观两个层面，对《水浒传》四个英译本的翻译特征进行多维度的对比分析。试图回答如下五个问题：
（1）《水浒传》四个译本作为翻译语言，在语言维度上各体现出什么特征？具体讲，在词汇、词块使用上各有什么特征？形成这些语言特征的原因是什么？
（2）不同译者在翻译时，如何在句子层次进行操作的？其操作规范与其翻译策略之间的关系又是什么？
（3）在语篇层次，译者是如何安排译文的？对原文是如何进行重组的？这样的重组与其翻译策略又有什么样的关联？
（4）四个译文主题词与主题语义域分析能否揭示翻译文本的特征？
（5）不同译本间人物语言行为差异所体现出译本翻译特征的同时，它与译者的价值取向有什么样的关系？
为探究这些问题的答案，挖掘四译本的翻译特征，全文分成6章。
第一章是绪论，介绍了本文研究背景及意义，研究内容和研究方法及全文的框架。
第二章系《水浒传》英译的研究现状。在这部分中，笔者对近几十年来关于《水浒传》英译研究作了回顾，按研究内容所涉及的版本的多少，将其划分为两个层次，即基于单译本的研究和基于多译本的研究，在分类的基础上，厘清每一研究的特点和存在的问题，最后统筹所有研究，指出基于平行语料库研究的必要性。这是因为基于平行语料库的范式，能够从宏观、微观层面对译本进行定量、定性分析，得出的结论更加科学、客观；同时将四个译本纳入研究范围，一方面扩大了研究视野，另一方面也使对比更加深入，更能揭示各译本的特征。
第三章着重介绍了《水浒传》汉英平行语料库的建库过程和方法，以及语料库加工的方法，详细描述了语料的选择、语料间的段落对齐和句子对齐、语料的标注；简要介绍了本文主要使用的语料库及语料库建设工具和分析工具，阐述了基于平行语料库的研究范式等。
第四章是对四译本宏观翻译特征的对比研究。结合平行语料库与可比语料库及参照语料库，运用语料库语言学的方法，主要在宏观层次对比研究各译文词频、词汇、词块、句对、语篇及主题词、主题语义域的特征和四译文的相似度，从而把握各译文词汇及词块使用特征、译者在翻译时对句子理解和组合的特征及在语篇层面上的译文推进机制、译文的主题词及主题语义域的特征及译文间的关系。研究发现：
（1）词频上，赛译与杰译、沙译及登译这三个译本的词表首位词不同，赛译的词表首位词是连词and，而其它三个译本的词表首位词均是定冠词the；统计发现，赛译本中的连词and与其它三个译本存在显著性差异，再以赛氏原创作品《大地三步曲》为参照语料库，考察发现该作品的词表首位词也同样是定冠词and，从而可断定赛氏的有关中国题材的无论是译文还是作品呈一致的特征。此外，and的过量使用，是句子变长的主要因素之一，下章的研究印证了这一推断。最后，将四译本与TEC语料库相比，杰译更接近于翻译语言，而赛译、沙译及登译这三个译文具有相似性，稍微远离翻译语言。
（2）词汇特征包括词汇密度、频率参数、类符/形符比与平均句长、独特词四个方面。词汇密度的分析发现沙译的词汇密度最高，其次分别为登译、杰译和赛译。四个译本词汇密度都小于英语原创文本的词汇密度，除沙译外，其它三个译文的词汇密度均小于文献中译语的密度，显然这三个译文作为翻译语言的特征都较为明显。由于词汇密度是衡量文本信息量大小的一个标准，可认为沙译相对于另外三个译文本信息量最大，而赛译信息量最小；频率参数的调查发现赛译的高频词的数目、累计比例、高频词重复率及高频词与低频词之比都是最大的，这说明赛译中使用高频词最多，而高频词一般为虚词，这也印证了前面研究中赛译词汇密度最小的结论。同理，沙译词汇密度最大也得以印证。此外，四个译本最显著的特点是高频词所占的比例极高，高于文献中其它语言或题材译文的高频词比例，而且高频词的重复率极高，这些充分表明四位译者在词汇使用方面有“简化”的特征，印证了翻译普遍性之“简化”假说；在标准类/形符比上，沙译最高，登译、杰译和赛译分别依次降低，说明沙译的词汇使用最为多样和丰富，而赛译则最缺乏变化，最为单调。这一点也支持上述核心词汇模式分析的结论，同时也支持了前面研究得出的沙译词汇密度最大和赛译词汇密度最小的结论。另外，四个译文的标准类符/形符比均小于原创小说，表明四译文整体用词变化性低于原创小说，这与翻译文本中和词语“简化”假设相一致，即翻译小说倾向于使用较少的词汇，而英语原创文本倾向于使用较多的词汇，故作为译语的特征较为明显；至于平均句长，四译文从大到小分别为赛译、登译、沙译和杰译，其中赛译是杰译的一倍多。且四个译文句长要明显大于原创文本，说明都接近译语的特点，其中赛译句子冗长的原因之一就是使用过多的and句，这和上面的词频结论又得到相互印证。
（3）词块维度是以5词词块为代表，发现赛译中基本是以组织性词块为主，旨在组织篇章结构，实现语篇行为，表达语义—逻辑关系，揭示信息流动的方向，组织性词块构筑和强化着文本的衔接和连贯，使文本成为一个有机的整体，因而译文逻辑清楚，易于阅读；杰译基本是以立场性词块为主，立场性词块主要是发表评价或表明态度，提出建议或忠告，主要出现在对话中，促进情节的发展；沙译的词块中与赛译相仿，大部分是组织性词块；登译基本是以命题性词块为主，实现命题意义，主要以时间的交替和空间的转移来推动故事情节的发展，体现出过程性。
（4）句对涉及"翻译单位"的选择和实施，研究发现，登译1:1句对最多，表明其最尊崇原文，下来依次是沙译、赛译和杰译。赛译中2:1类别的句对比其他三译文都多，因其经常使用and句，过多的and将句子进行合并，从而导致这类句对类型增多，同时也印证了前面分析平均句长得出的结论：赛译的平均句长最长。此外，四个译本中，都有一定数量的句子没有翻译，其中杰译最多，赛译最少，沙译和登译分别居第二、三位，这体现出译者翻译的主观性及能动性。
（5）语篇维度研究发现：四译文主要是通过对话来组织发展和推进篇章的。另外，时间在四个译文中所占的比例均为第二，其中赛译最高，其余依次为杰译、登译和沙译，表明时间的变化和推移也是译文发展的一个重要手段，尤其是赛译，其时间所占的比例与对话相比差别不是太大。地点类相对来说在译文中所占的比例较少，但赛译中仍占较大的比例，这说明赛译中以对话为主，以时间为次，辅之于地点的变化来推进和发展语篇的；而杰译、沙译和登译则以对话为主，以时间为次，这两方面交织共同组织和发展语篇的。
（6）主题性分析发现四译本中的主题名词基本是《水浒传》中的关键人物，像宋江、吴用、李逵等。其中，沙译及登译尤其是后者广泛使用人物的绰号（而不直接使用人名）是其一大特征。主题动词包括运动型这一共性外，区别在于报道性动词的使用；从主题词云图发现杰译、沙译及登译最鲜明的一个特征是：Song（梁山英雄首领宋江的姓）居于主题词的中心，但赛译位于主题词的中心的却是and，而Sung即“宋”，却位居次席。这从另一个角度印证了前面讨论赛译多and句的句式特点。另外，四译文主题词均含有一定量的人称代词，这也是小说特征之一，基本用于对话当中，体现出人物间的交互性及认同感。负主题词，即译文中某些词汇的过度少用，体现翻译文本的共同特征。
（7）四译本相似度历时聚类研究表明沙译与登译相似度最高，而赛译与登译相似度最低。居于其间相似度按大小依次是：杰译与沙译，杰译与登译，杰译与赛译，赛译与沙译。译本间的相似度可从不同译者的翻译策略或翻译方法中得到解释。
第五章是四译本微观翻译特征的对比研究，是对第四章中宏观研究的补充和深化，着重探讨微观层面上的翻译特征，包括动词高频动词take的类联接及搭配词的特征、四译本回目翻译、话语标记翻译、典型动词翻译、詈词翻译，研究发现：
（1）动词take的类联接及搭配研究发现，四位译者均使用此词最常用的类联接及对应的最常用的搭配。
（2）登译的回目翻译，采用意译法，以英雄双行体诗歌的形式，较好地对应了原文对仗的特点。其它三个译文，赛译基本运用的是逐字逐句、亦步亦趋的直译方法，和原文形成了物理对应；杰译则主要运用了意译的方法，译文比较简捷，基本能够清楚地传递关键意思；沙译大多采用了直译法，能够较好地保留了原文对仗的形式，且意义的传达也非常到位。此外，对回目的词汇进行定量分析揭示了四位译者的不同特征。
（3）话语标记翻译方面，开始标记语“话说”的翻译对比研究发现赛译均为IT IS SAID，杰译与沙译可以说是全部去除了这些词语，不予翻译。登译除有少量未译外，别的都做了巧妙处理，没有影响英语行文。且常常是有变化的，有浓重的文学色彩，实现了引出话题的语用功能；结束标记语"且听下回分解"翻译对比研究表明登译能够根据上下文具体情况选择合适的译文。与之相比较，赛译和沙译均仅有一种译法，而杰译大多数情况下不予翻译。
（4）报道动词“道”的翻译考察发现赛译较为单调；杰译则经常将原文中的直接引语译为间接引语，且使用对应报道动词的种类少于沙译和登译，沙译译法丰富。动词“掿”、“攧”的翻译对比发现：赛译往往选择一些高频常用词；杰译略译的情况较多，即使译出也选择常用词；沙译则会考虑上下文的语境，选择合适的词汇。结合语义韵理论，可发现沙译是在汉语语篇语义韵的考量下翻译的，因而是在语篇层次上的翻译，译文合适、得当，能够将形式和功能及意义紧密地联系在一体。相对于另外三个，登译体现出多样性和丰富性。
（5）詈词“厮”的翻译对比研究显示赛译基本上呈中性；杰译也基本呈中性，部分未译，因而赛译与杰译有“净化”的趋向；与赛译及杰译相反的是，沙译与登译往往使用含有感情色彩较浓、有明显谩骂的词汇来表达“厮”的内涵意义。相比于沙译，登译的数量更加丰富，词汇的贬义或蔑视的程度更加强烈。“鸟”的翻译对比研究发现赛译单一，杰译大多略去不译，沙译和登译丰富多样，在大多数场合中主要用来表示说话者辱骂、愤怒、怨恨和鄙视等感情或是表示说话者诅咒、厌烦、惊讶和强调等语气。
第六章为结论，总结本研究的主要发现，根据四个译本翻译的共性特征，归纳对翻译研究与翻译实践的四点启示：
（1）四译本在词汇层次都表现出“简化”的共性，即表头高频词覆盖率特别高，因此对高频词的掌握特别重要，翻译要优先使用高频词，尤其是它们的常用、典型的类联接及对应的典型搭配。
（2）四译本翻译句对中一对一比例最高，即译者的翻译单位基本上是以句子为单位进行的，在此基础上适当的进行句子组合，因此在注重语篇的基础上，以句子为单位进行翻译不失为一个可靠的选择。
（3）小说翻译的组织及推进方面，在对话的基础上，选择时间和空间的变换组织发展基本上是译者遵循的原则，因此对于汉语小说英译有极大的参考价值。译者在破段时，可考虑时间及空间变换因素，将此作为依据及线索。
（4）平行语料库可对词语、短语等语言因素所在的句子进行链接，观察上下文可发现对应的翻译，由于这些翻译有足够的上下文语境，因此对编写高质量的翻译词典能够发挥十分重要的作用。
最后，在本章末指出本研究存在的不足之处，并就后续研究提出建议。
In light of theories of corpus linguistics, contrastive linguistics, functionallinguistics, stylistics and information theory, the dissertation probes into thetranslation features of four versions of Shuihu Zhuan from both the macro and themicro levels with a multi-dimensional comparative approach. Based on self-builtShuihu Zhuan Chinese-English parallel corpus, it tries to answer the following fivequestions:
(1) What features of translational languages do the four versions reflect?Specifically, what are the features on the use of words and the chunks? What are thereasons for the formation of such language features?
(2) How do the four translators cope with the source text in the sentence level?What translation strategies have they adopted doing their exertion?
(3) How do the four translators handle the source text at the textual level? Howdo they restructure the original and what translation strategies does the reconstructionimply？
(4) To what extent do key words analysis and key domains analysis of the fourversions reveal their translation features？
(5) As the actual language behavior distinction between the characters in differentversions, what are the relationships between the features and their value orientation?
The paper is divided into six chapters aiming to find the answers to thesequestions and explore the translation features of the four versions.
The first chapter is the introduction, and it describes the background andsignificance of this study, the research contents and methods as well as the frameworkof the paper.
The second chapter summarizes the actuality of Shuihu Zhuan C-E translationresearch with the existing problems. The author first reviews the researches of Englishtranslations of Shuihu Zhuan in recent decades. According to the versions involved, itis divided into two categories: one version and one-odd version. Based on thefollowing two categories, the advantages and disadvantages of all the previous studies are put forward. Therefore, it is necessary and important to conduct the parallelcorpus-based work. This is because parallel corpus-based paradigm enablesresearchers to conduct quantitative and qualitative analysis from both the macro andmicro levels. Thus, its finding will be scientific and objective. In addition, putting thefour versions into the consideration, on the one hand, will expand the research horizon;on the other hand, the deeper contrasts can be better to reveal the features of eachtranslation.
Chapter three focuses on the building of Shuihu Zhuan Chinese-English parallelcorpora, including the working procedure, the choices and processing of the texts, thealignment of the paragraphs and sentences and the annotation of the corpora. It gives abrief introduction of the corpora and the concordancing tools which are used in thepaper. At the end of the chapter, the parallel corpus-based research paradigm ispresented.
The fourth chapter is devoted to the macro study of the translation features of thefour versions. With the help of the parallel corpora, comparable corpora and referencecorpus and adoption of the approach of corpus linguistics, the paper conducts acomparative study of the four translations in the dimensions of word frequency, words,chunk, sentence pairs, discourse, key words, key domains and similarity, so as tograsp the translation features at the corresponding levels. For example, the features ofthe words and chunks, the features of the translators' comprehension and restructureof sentences, the underlying mechanism for progressing the texts, the relationshipbetween the four translations and its keywords and key domains.The findings of this chapter are as follows:
(1) Word frequencies shows that the top words in Jackson's version, Shapiro'sversion and Dent-Young's version are the same word the, but the top word of Pearl'sversion is and. Statistics find that there is a significant difference between Pearl's andthe other three's in the use of the word and. The same finding is conformed by usingthe reference corpus of Pearl's China-related theme novel, Good Earth Trilogy, inwhich the top word is also and. Therefore, it is concluded that whether her novels or translation are characterized by the use of word and. In addition, the overuse of and isone of the factors which leads to long sentence. The next chapter confirms thishypothesis. Finally, compared with the TEC corpus, Jackson's translation is closer tothe translation language, the other three translations are similar, slightly different fromthe translation language.
(2) Lexical Features include four aspects, namely, lexical density, frequencyparameters, the type/token ratio, the mean sentence length, the unique words. Lexicaldensity analysis finds that Shapiro's translation has the highest one, followed byDent-Young's, Jackson's and Pearl's. Four versions' lexical densities are all less thanthat of the original English text. Except Shapiro's, the other three translations' lexicaldensity are all lower than that in the previous study, which shows that the threeversions take on the features of the translational language. As lexical density isregarded as the symbol of the amount of information, it can be concluded thatShapiro's enjoys the highest information, while Pearl's have the lowest information;the survey of frequency parameters shows that the number, the cumulative proportion,the repetition rate of high frequency words and the ratio of high frequency wordsversus low frequency words are all the highest in her translation. This, on the onehand, clearly proves that Pearl tends to use high frequency words, which are mainlynominal ones; on the other hand, it also confirms the above conclusion that Pearl' shas the lowest lexical density. Similarly, Shapiro's highest lexical density is alsoconfirmed. In addition, the four versions share the same feature that high frequencywords enjoy the high proportion, even higher than those of the previous studies.Besides, the repetition rate of high frequency words is high too. All lexical featureslend support to the hypothesis of simplification of the translated texts.
In standard type/token ratio (STTR), the highest one is Shapiro's, followed byDent-Young's, Jackson's and Pearl's. This shows that Shapiro's tends to be rich inword choice while Pearl's is fixed and monotonous. In this point, the finding supportsboth the core patterns of lexical use and the conclusion of the lexical density analysis.Furthermore, the four versions' STTRs are all less than the original ones. The fact is consistent with the simplification hypothesis in word use. That's to say, translatednovels tend to use less words while original ones tends to use more words.Consequently, the four translations share the distinct features of translated texts.
As for the mean sentence length (MSL), the longest one is Pearl's, followed byDent-Young's, Shapiro's and Jackson's, of which the longest one is nearly as twice asthe shortest. In addition, the MSL of each version is significantly greater than that ofthe original text, which shows that they are all close to the translated texts. One reasonfor the long sentences in Pearl's lies to her use too many and in sentences. Thisconfirms the findings of the frequency analysis.
(3) Chunks, with the five-word as the representative, finds that Pearl's are mainlyorganizational ones, aiming to organize the discourse structure, to realize thediscourse acts, to express the relationship between the meaning and the logic, toreveal the direction of information flow. By strengthening the cohesion and coherenceof the text, this contributes a lot to the formation of a organic and complete one. Thesechunks help readers to read. Jackson's is mainly made up of stance chunks. Thesekinds of chunks convey people's evaluation, attitude, suggestion or warning. Most ofthem appear in the dialogues with the purpose to advance the development of the plot.Shapiro's is identical to Pearl's, with propositional chunks as the main type.Dent-Young's, however, is mainly of the propositional ones. They consist of temporaland spatial chunks, with the joint purpose to make the story develop.
(4) Sentence pair involves the choice and operation of translation unit. Theanalysis of the Chinese-English sentence alignment type reveals that Dent-Young'sversion respects the original while the other tends to regroup or paraphrase theoriginal, with Jackson's as the most free one. This conclusion is based on the statisticsof the Chinese-English1:1sentence pair type. In comparison with the other three, thetype2:1in Pearl's is far more. She frequently uses and as a way to form longsentences, which, in turn, confirms that her version possesses long sentences. Besides,no translators translated all the original, of which, Pearl finished the most, followedby Shapiro, Dent-Young and Jackson. This indicates the translators' subjective choice.
(5) The discourse dimensional study finds that four translations develop throughdialogues to advance the texts. In addition, the factor time ranks the second in all theversions, of which Pearl's is the most, followed by Jackson', Dent-Young's andShapiro's. This indicates that the change of time and the passage of time is also animportant means to develop the text. For Pearl's, both dialogue and time account foralmost the same proportion. Though the place factor is less in all the four versions, itstill accounts for a large proportion in Pearl's. The conclusion can be made that Pearl'sdevelops with dialogues, time and place, with the former as the main one and thelatter the subordinate one. In contrast, the other three develop with primary dialoguesand secondary time.
(6) Key words analysis reveals that the key noun words in four translationsbasically are the names of key figures in the novel, like Song Jiang, Wu Yong, Li Kui.In Dent-Young's and Shapiro's translations these names are replaced with thecorresponding nicknames. The widespread uses of nicknames rather than the namesare their major features. Key verb words include the common displacement verbs, thereporting verbs, with the latter as their differences. Moreover, the key words cloudscan find the word Song lies in the center while word and is located in the center ofPearl's and Sung(Song) beside it. This again confirms the lexical feature of Pearl's. Inaddition, there are many personal pronouns in the key word list, which is a feature ofnovel. In dialogue, these words reflect the interaction and identity between thecharacters. Negative keywords, or the underuse words show a common feature of thefour translated texts.
(7) Cluster Analysis of the four translations shows that Shapiro's and Dent-Youngshare the highest similarity while Pearl's and Dent-Young's have the least similarity,sandwiched with Jackson's and Shapiro's, Jackson's and Dent-Young's, Jackson's andPearl's, Pearl's and Shapiro's. The similarities or differences among the four versioncan be explained from the translators ' strategies or their translation methods.
The fifth chapter examines micro translation features of the four translations,which can be viewed as a way to supplement and deepen Chapter four's macro research. It includes the survey of the high frequent word take, the comparative studyof the chapter title translation, the discourse marker translation, typical verbtranslation and swearword translation.
The findings of this chapter are as follows:
(1) Four translators all use the most common colligation and the correspondingtypical collocations of the verb take.
(2) In chapter title translation, Dent-Young uses the free translation method withthe heroic couplet poetry style, which better corresponds to the characteristics of theoriginal antithesis. Pearl basically uses word for word translation method, formingphysical mapping with the original; Jackson mainly paraphrases the original, and histranslation is relatively simple, but the key meaning is clearly conveyed; Shapiro'stranslations are mostly the literal ones, and they are able to retain the originalantithesis form and the meaning is fully conveyed.
(3) The beginning discourse marker huà shuō in the original all are translated intoIT IS SAID by Pearl; however, Jackson and Shapiro nearly remove the markers andneglect them. Dent-Young almost translates all the markers with various translationsand literary flavor without affecting the English wording. These various expressionsserve the pragmatic function of introducing the topic; end Marker qiě tīng xià huí fēnjiě in the original text has one fixed translation in Pearl's and Shapiro's respectively.Jackson, in most cases, turns a blind eye to them and does not translate them. Incomparison, Dent-Young always translates them according to their different contexts.
(4) As the translation of reporting verb dào, Pearl renders few differenttranslations; Jackson usually changes the direct quotation in the original text intoindirect speech. Shapiro's translation is rich and varied. Applying the semanticprosody theory to evaluating the translations of powerful verb nuò and diān, theresearch finds that Shapiro's translations are usually text-oriented andmeaning-function-oriented. They fully express the hidden semantic prosody; therefore,they are appropriate. At the same time, Pearl's, Jackson's and Dent-Young's show theirindividual characteristics.
(5) The survey of the translations of swearword sī shows that Pearl's are basicallyneutral; Jackson's are also almost neutral. Both of the two translators sometimesneglect some. So they take on the tendency of sanitation of swearwords. Conversely,Shapiro and Dent-Young tend to express it with emotional and abusive words.Compared to Shapiro's, Dent-Young's translations are richer, even containing stronglyderogatory words. Comparative study of the translations of diǎo shows that Pearl's arefixed, and Jackson mostly omits them without translation. Shapiro's and Dent-Young'sare varied in most occasions, primarily representing the speaker's abuse, anger,resentment and contempt or speaker 's curse, boredom, surprise, and emphasis.
The sixth chapter is the conclusion of the dissertation. It summarizes the mainfindings of the present study. Importantly, four implications for translation study andpractice are drawn, and they are as follows:
(1) In word level, four versions share the feature of "simplification", whichmeans that the few high frequency words make up more than half of the text (here,even if closed words are excluded, the rest open words also have a great proportion,such as words come, go, etc.) Therefore, it is very important to grasp the highfrequency words and priority should be given to using them in translation, especiallytheir common colligations and typical collocations.
(2) Sentence pair study reveals that one to one for Chinese to English pattern topsin the pairs, which clearly shows that the translation is basically based on the unit ofsentence. So with the consideration of the discourse, it's an appropriate choice tomake the sentence as the translation unit.
(3) Dialogues, time and space constitute the clues for discourse progress, ofwhich time and space changes are a reliable reference for the translators to find. Sowhen dealing with the fiction translation from Chinese into English and breaking thelong paragraph into short ones, the translator should take time and space symbols intothe consideration as clues.
(4) Parallel corpus provides linking bilingual or multilingual words, phrases withauthentic and enough context. These information and materials may play a crucial role in compiling high-quality translation dictionary.
At the end, it points out the area where further research may be carried out.
引文
1参见谭汝谦“中日之间翻译事业的几个问题”，《日本研究》1985年第3期。
2参见王丽娜“《水浒传》外文论著简介”，《湖北大学学报》（哲学社会科学版）1985年第3期。
3参见孙建成《<水浒传>英译的语言与文化》，第6页，上海复旦大学出版社，2008。
4参见p xxii, All Men Are Brothers, The George Macy Companies, Inc.,1948.
5参见p xiv Water Margin:Outlaws OF The Marsh. Tuttle Publishing,2010.
6参见p xiv The Part One of The Marshes of Mount Liang, the Chinese University Press,1994.
Abercrombie, D.1965. Pseudo-procedures in linguistics[A]. in D. Abercrombie (ed.) Studies inPhonetics and Linguistics [C]. pp.114–19. Oxford University Press.
Aijmer, K., Altenberg B. and Johansson M. Languages in contrast. Papers from a Symposium onText-based Cross-linguistic Studies, Lund4-5March1994.[C]. Lund Studies in English88.Lund: Lund University Press,1996.
Aijmer, K.&B. Altenberg.(Eds.) English Corpus linguistics[C]. London and New York:Longman,1991.
Ardekani, M.A.M. The translation of reporting verbs in English and Persian [J]. Babel,2002,(2).
Aston, G. Learning with Corpora: an Overview[A]. in Aston, G.(ed.) Learning With Corpora[C],Huston: Athelstan,2001.
Baker, M. Corpus linguistics and translation studies: Implications and applications[A]. In M.Baker, G. Francis&E. Tognini-Bonelli (eds.). Text and Technology: In Honour of JohnSinclair [C]. Amsterdam: John Benjamins,1993:233-250.
Baker, M. Corpus-based Translation Studies: The Challenges That Lie Ahead[A]. SOMERS H.Terminology, LSP and Translation: Studies in Language Engineering, in Honour of Juan C.Sager[C]. Amsterdam: John Benjamins,1996.
Baker, P. Using Corpora in Discourse Analysis[M]. London: Continuum,2006.
Biber, D. Stig Johansson, Geoffrey Leech, Susan Conrad, Edward Finegan. Longman Grammar ofSpoken and Written English.[M]. London: Pearson Education Limited,1999.
Biber, D., S. Conrad&V. Cortes. If you look at...: Lexical bundles in university teaching andtextbooks [J]. Applied Linguistics,25(3),2004.
Bolinger, D., Meaning and Memory[J]. Forum Linguisticum1(1),1976.
Buck, Pearl S. The Good Earth [M]. New York: John Day Company,1931.
Buck, Pearl S. Sons[M]. New York: John Day Company,1932.
Buck, Pearl S. A House Divided[M]. Michigan: Reynal&Hitchcock,1935.
Buck, Pearl S. All Men are Brothers[M].New York: The George Macy Companies, Inc.,1948.
Dent-Young, John and Alex. The Marshes of Mount Liang[M]. Hongkong: The ChineseUniversity of Hongkong,1994-2002.
Firth, J. Papers in linguistics[M]. London: Oxford University Press,1957.
Frawley, W. Prolegomenon to a theory of translation.[A]. In W. Frawley(ed.). Translation:Literary, Linguistic and Philosophical Perspectives[C]. London: Associated University Press,1984.
Gellerstam, M. Translations as a source fro cross-linguistic studies[A]. In K. Aijmer, B. Altenberg&M. Johansson (eds.) Language in contrast: papers from a symposium on text-basedcross-linguistic studies, Lund, March1994[C]. Lund: Lund University Press,1996.
Halliday, M.A.K. An Introduction to Functional Grammar[M]. Beijing: Foreign LanguageTeaching and Research Press,2000.
Halliday, M. A. K. Corpus studies and probabilistic grammar.[A]. In Aijmer&Altenberg (eds.),1991.
Halliday, M. A. K. Language as Social Semiotic: The Social Interpretation of Language andMeaning [M]. London:Arnold,1978/Beijing: Foreign Language Teaching and ResearchPress,2001.
Hartmann, R. Contrastive textology [J]. Language and Communication,1995(5).
Huston, S. Corpora in Applied Linguistics[M]. Cambridge: Cambridge University Press,2002.
Jackson, J. H. Water Margin[M]. Shanghai: The Commercial Press, Ltd.,1937.
Jackson. J. H. Water Margin [M]. Hong Kong: The Commercial Press, LTD.,1963, volume one,volume two.
Jackson. J.H. The Water Margin: Outlaws of the Marsh [M]. Tokyo: Tuttle Publishing,2010.
Laviosa, S. Core patterns of lexical use in a comparable corpus of English narrative prose [J].Meta,1998(43).
Laviosa, S. How comparable can 'comparable corpora' be?[J]. Target,1997(9).
Leech. Introducing corpus annotation[A] in R. Garside, Leech and A. McEnery (eds) CorpusAnnotation.[C] London: Longman,1997.
Louw, B. Irony in the text or insincerity in the writer? The diagnostic potential of semanticprosodies [A]. M. Baker，G. Francis&E. Toni-Bonelli (eds.). Text and Technology: InHonour of John Sinclair[C]. Amsterdam: John Benjamins,1993.
Luz, S.&M. Baker. TEC: A toolkit and app lication program interface for distributed corpusprocessing[EB/OL]. Paper presented at the workshop on Web-based LanguageDocumentation and Descrip tion12-15December2000, Philadelphia, USA. http://www. ldc.upenn. edu/exp loration/exp l2000/papers/luz/,2000.2004-12-04.
Martin, J. R. English Text System and Structure [M]. Beijing: Peking University Press,2004.
McEnery, T., Hardie, A. Corpus Linguistics: Method, Theory and Practice[M]. Cambridge:Cambridge University Press,2012.
McEnery, A.; Wilson, A. Corpus linguistics (2nd edition)[M]. Edinburgh: Edinburgh UniversityPress,2001.
McEnery, T., Xiao, R. and Tono, Y. Corpus-based Language Studies: An advanced ResourceBook [M]. London/New York: Routledge,2006.
McEnery, A.; Xiao, Z. Domains, text types, aspect marking and English-Chinese translation[J].Journal of Languages in Contrast,2002,2(2).
Olohan, M. Introducing Corpora in Translation Studies[M]. London:Routledge,2004.
Partington, A. Patterns and Meanings: Using Corpora for English Language Research andTeaching[M]. Amsterdam; Philadelphia: J. Benjamins Publishing,1998.
Piao, Scott Songlin. Word alignment in English-Chinese parallel corpora[A]. Literary andLinguistic Computing,17(2), Oxford, UK.
Rayson, P. and Garside, R.(2000). Comparing corpora using frequency profiling[A]. InProceedings of the workshop on Comparing Corpora, held in conjunction with the38thannual meeting of the Association for Computational Linguistics (ACL2000)[C].1-8October2000, Hong Kong, pp.1-6.
Rayson, P., Archer, D., Piao, S.&McEnery, A. M.. The UCREL semantic analysis system[A].Proceedings of the beyond named entity recognition semantic labelling for NLP tasksworkshop[C]. Lisbon, Portugal,2004. Lisbon p.7-12..
Rayson P. From key-words to key semantic domains[J]. International Journal of CorpusLinguistics13(4):519–549,2008.
Scott, M. Clearing the undergrowth and marking out trails… challenges in investigatingkeyness[A]. Keyness in Text Conference. Certosa di Pontignano, Siena,2007.
Scott, M. PC analysis of key words–and key key words[J]. System25(1):1–13,1997.
Scott, M. Problems in investigating keyness, or clearing the undergrowth and marking outtrails…[A]. In Bondi, M.&Scott, M.(eds.) Keyness in Texts[C]. Amsterdam/Philadelphia:John Benjamins Publishing Company,2010.
Scott, M. What can corpus software do?[A] in O’Keefe and McCarthy (ed.) The ReoutledgeHandbook of Corpus Linguistics[C]. London: Routledge,2010.
Shapiro, Sidney. Outlaws of the Marsh [M]. Beijing: Foreign Languages Press,2003.
Sinclair, J. Looking up: An account of the COBUILD project in lexical computing[M]. London:Collins,1987.
Sinclair, J. Corpus Concordance Collocation[M]. Oxford: Oxford University Press,1991.
Sinclair, J. The Search for Units of Meaning[M]. Textus9(1),1996.
Sinclair, J. Sinclair, J. Reading Concordance: A introduction[M]. London: Pearson EducationLimited,2003.
Teubert, W. Comparable or parallel corpora?[J]. International Journal of Lexicography,1996,9(3).
Tognini Bonelli, E. Corpus Linguistics at Work[M]. Amsterdam: John Benjamins PublishingCompany,2001.
Tymoczko, M. Computerized corpora and the future of translation studies[J]. Meta,1998,43(2).
Xiao. R. How different is translated Chinese from native Chinese? A corpus-based study oftranslation universals [J]. International Journal of Corpus Linguistic,2010,(15).
包惠南.文化语境与语言翻译[M].北京：中国对外翻译出版公司，2001.
陈伟.翻译英语语料库——新型翻译研究的利器[J].华中农业大学学报(社会科学版)，2009,(1).
陈智淦.从《水浒传》的“吃”字翻译谈起[J].龙岩师专学报,2004,(2).
崔素花.论沙博里《水浒传》英译本中言内意义的翻译[J].太原理工大学学报（社会科学版）,2009,(2).
董琇.译者风格形成的立体多元辩证观——赛珍珠翻译风格探源[D].上海：上海外国语大学，2009.
冯光武.语用标记语和语义/语用界面[J].外语学刊，2005,(3).
冯庆华．母语文化下的译者风格[M]．上海：上海外语教育出版社，2008．
桂诗春.基于语料库的英语语言学语体分析[M].北京：外语教学与研究出版社，2009.
郭梅,亦歌.《水浒》英译三种比较研究[J].长江学术，2006,(4).
胡天赋.从人物的再现看赛译《水浒传》的后殖民主义色彩[J].河南大学报，2006,(5).
胡壮麟等.系统功能语法概论[M].长沙:湖南教育出版社，1989.
黄昌宁、李涓子.语料库语言学[M].北京：商务印书馆，2002.
黄国文、葛达西.功能语篇分析[M].上海:上海外语教育出版社，2006.
李晶.翻译与意识形态——《水浒传》英译本不同书名成因探析[J].外语与外语教学，2006,(1).
李赛红.解构英国国家语料库[J].外语教学与研究，2002,(4).
李小龙.中国古典小说回目的叙事功能[J].文艺理论研究，2011,(3).
廖七一.语料库与翻译研究[J].外语教学与研究，2000,(5).
刘丹青.汉语是一种动词型语言——试说动词型语言和名词型语言的类型差异[J].世界汉语教学，2010,(1).
刘宏照.试析《水浒传》部分专名的英译[J].湖北社会科学，2008,(2).
刘泽权、田璐.《红楼梦》叙事标记语及其英译——基于语料库的对比分析[J].外语学刊，2009,(1).
卢红梅.汉语语言文化及其汉英翻译[M].武汉：武汉大学出版社，2011.
卢惠惠.古代白话小说语言形式的程式化特征[J].明清小说研究，2007,(1).
卢艳春.语用学与翻译--《水浒传》中粗俗俚语的翻译之管见[J].内蒙古农业报，2005,(3).
路东平.试谈《水浒传》中称谓的使用及其翻译[J].社科纵横，2003,(1).
马红军.为赛珍珠的“误译”正名[J].四川外语学院学报，2003,(3).
马文熙等.古汉语知识辞典[M].北京:中华书局，2004.
毛翠.试谈沙译《水浒传》中方言词汇的英译[J].科教导刊（中旬刊），2010,(10).
穆雷.翻译研究方法概论[M].北京：外语教学与研究出版社，2011.
钱歌川.翻译的基本知识[M].长沙:湖南科学技术出版社，1981.
冉永平.论语用元语言现象及其语用指向[J].外语学刊，2005,(6).
沙博理.《水浒传》的英译[J].翻译通讯，1984,(2).
申丹.有关小说中人物话语表达形式的几点思考[J].外语与外语教学，1999,(1).
石麟.章回小说回目的来源演变及其文化意蕴[J].明清小说研究，2004,(1).
孙建成、温秀颖.后殖民语境下《水浒传》英译本的文化特征——登特杨新译本批评[J].外语研究，2008(2).
孙建成.水浒传英译的语言与文化[M].上海:复旦大学出版社，2008.
唐艳芳.赛珍珠《水浒传》翻译研究——后殖民理论的视角[M].上海:复旦大学出版社，2010.
田畔.水浒国骂大观[J].社会科学论坛，2009,(5).
王克非.语料库翻译学——新研究范式[J].中国外语，2006,(3).
王克非等.双语平行语料库研制与应用[M].北京：外语教学与研究出版社，2004.
王克友、王东升.叙述方式的转换与小说翻译效果——以《水浒传》第47回三个译文为例[J].外语教学，2005,(4).
王丽虹.中国章回体小说的艺术特色及沙博理对《水浒传》的英译[J].语文学刊，2005,(4).
王守元，张德禄.文体学辞典[M].济南：山东教育出版社，1996.
王佐良、丁往道.英语文体学引论[M].北京：外语教学与研究出版社，1987.
卫乃兴.词语学要义[M].上海:上海外语教育出版社，2011.
魏莉.《水浒传》三个英译本的比较研究——兼谈文学翻译中译者的“创造性叛逆”[J].内蒙古农业大学学报（社会科学版），2009(3).
文军、罗张.国内《水浒传》英译研究三十年[J].民族翻译，2011(1).
肖忠华、戴光荣.翻译教学与研究的新框架:语料库翻译学综述[J].外语教学理论与实践，2011,(1).
肖忠华.平行语料库与可比语料库在语言研究中的应用[J].中国英语教育，2003,(1).
徐学平.试谈沙译《水浒传》中英雄绰号的英译[J].湛江师范学报，2001,(5).
许燕.赛珍珠英译《水浒传》的陌生化取向——试析赛译本成功的原因[J].华东理工大学学报(社会科学版)，2009(2).
许燕.二十年来《水浒传》的英译研究[J].山东外语教学，2008,(2).
姚君伟.我们今天研究赛珍珠什么？[J].江苏大学学报(社科版),2003,(4).
张彩凤.从沙博里《水浒传》英译本看文学风格的可译性（英文）[J].内蒙古师范大学学报（哲学社会科学版），2006,(1).
张德禄、刘汝山.语篇连贯与衔接理论的发展及应用[M].上海:上海外语教育出版社，2003.
张世君.明清小说评点叙事概念研究[M].北京：中国社会科学出版社，2007.
郑公盾.水浒传论文集[C].银川：宁夏人民出版社，1983.
朱永生、严世清.系统功能语言学多维思考[M].上海:上海外语教育出版社，2001.
朱永生等.功能语言学导论[M].上海:上海外语教育出版社，2004.