基于语料库之20世纪20年代英译汉文学文本中的怪异搭配探讨,英语文学论文

Chapter One Introduction

This study looks into odd collocation which is an important long-existed topicbut not well-solved. Since”collocation”has come into being for more than half acentury, researchers have made great achievements in this field, yet few studies havefocused on odd collocations, especially the studies in translated texts. In my research,I would like to make a corpus-based diachronic study to find out the odd collocationalfeatures and interaction between translation and the development of Chinese.The conduction of this study is based on practical and theoretical foundations.Differs from those traditional studies, the application of corpus promotes thetranslational studies from prescriptive one to descriptive one, and enables us toexplore phenomena that previously obscure and to see what distinctive featurestranslated texts have and how they differ from the original and non-translated texts.Because of its advantages, corpus-based translation study (CTS) has become a hotstudy field and my research is such kind of study.However, previous corpus-base translation studies mainly focus on the universalfeatures of translation, namely simplication, explicitation, implicitation, normalizationand sanitization, little attention has been given to the features of translationese whichcan be studied from the lexical, syntactical, collocational and prosodic perspective( Hu & Mao 2012). In other words, if we observe translations from themacro-perspective, some linguistic phenomena may be overlooked, so this studymakes a corpus-based translation study from a micro-perspective, i.e. to study thecollocation.Theoretically speaking, the exploration of collocations is of great importance, forstudying the collocations in translations enables us to (a) better understand how thesyntactical and lexical meaning of a specific language change in the process oftranslation and (b) enrich the study of collocations in translated texts. Meanwhile, thestudy of the relationship between translation and target language can develop thelinguistic evolution theory (Wang & Qin 2012). Though it has been acknowledgedthat translation plays an important role in the development of modern Chinese, thelexical level was not fully described and few studies referred to the interrelationshipbetween translationese and modern Chinese yet. Odd collocation, as the phenomenonof translation, can tell us the diachronic development of translationese as well as therole translation plays in the development of modern Chinese.In general, odd collocation is a phenomenon that can be frequently found intranslations in 1920s, since no studies on odd collocations in 1920s have beenconducted, the corpus-based study of it will be helpful to have a diachronic study ofcollocation in the May 4th period. Besides, as a translational product, the study of oddcollocation shows the interaction between translation and the modern Chinese. Withthe help of corpus, it’s possible to explore the odd collocational features and observethe impact that those collocations have had on the development of modern Chinese. Ihope this study can contribute to a better understanding of odd collocational featuresand the role translation plays in the evolution of Chinese.
.............

Chapter Two Literature Review

2.1 Comparable Corpus-based Studies
Scholars like Mauranen, Olohan and Dayrell make comparable corpus-basedtranslational studies from different aspects, such as the Finnish corpus-based unusualcollocational study, English corpus-based collocational range study and Portuguesecorpus-based systematic study. Based on the Finnish comparable corpus, Mauranen (2000) found two kinds ofcollocations that differ from the normal collocations in Finnish translations. One is theunusual conjunctional collocation and the other is the unusual verbal collocation.From the conjunctional perspective, Mauranen first picked “toisaalta” (means“on the one hand” in English) to see its collocate. Interestingly, she found that themost frequently used collocate with “toisaalta” in translated Finnish was “toisaaltataas” which differs from “mutta toisaalta” that often used in non-translated Finnish.Though they share the same meaning in English, they have different collocates. Thisphenomenon tells us that translation shows different combinational preferences.While from the verbal aspect, contrasting with translated Finnish academic texts, themost often used collocate with “HALUTA” in non-translated Finnish is “HALUTAkorostaa”, with a proportion of nearly 40% which is much higher than 8% intranslated texts. Moreover, collocation like “HALUTA osoittaa” has no appearance innon-translated Finnish but can be found in translated texts.Because of the small size of corpus, Mauranen herself wonders how wider thescope of the untypical word combination really is in translated texts. Besides, the term“untypical” is an obscure concept which needs to be further defined. Dayrell (2004, 2007, 2008) made systematic study of collocational patterns intranslated texts. Two sub-corpora in BPCC (The Brazilian Portuguese ComparableCorpus) are made for analysis, one is fiction corpus and the other is self-made corpusboth with translated and non-translated texts.It is found that collocates of about 80% nodes in translated fiction and that of50% in self-made corpus are less than those collocates in non-translated texts. About ahalf collocational pattern of the translated fiction and 40% collocational pattern of theself-made corpus tend to use the variant of a given collocational pattern. About 36%collocational pattern of the self-made translated texts and 61% that of the translatedfiction texts prefer to use repeated lexical patterns.Dayrell did a good job in the research method, research objectives andconclusions; especially the selection of the node and the extraction of collocate. Asthe co-occurrence of collocates and nodes, Dayrell thought they should co-occur atleast four times. But as to the small or mid-sized corpus, the co-occurrence frequencyless than four times is very common. Therefore, such way of extraction may not suitall the effective collocates.

2.2 Parallel Corpus-based Studies
According to Kenny’s study, translated texts are featured as co-existence ofnormalization and foreignization. Based on the GEPCOLT (The German-EnglishParallel Corpus of Literary Texts), Kenny (2001) found that there were unusualcollocations in non-translated German and tried to find out the way English dealt withthose unusual collocations. On the one hand, Kenny thought they were normalized inthe translated texts. “SCHRILLEN TELEFON” which means “shrill telephone” wastaken as an example. Actually, the German word “SCHEILLEN”, meaning theunexpected ring bell, usually occurred with “Alarmglocken” (alarm bells), so thecombination of “SCHRILLEN TELEFON” is a kind of unusual collocation. Though itis somewhat unusual, it is translated into “the phone rang” in the English translation,this phenomenon reflects the normalization of unusual collocations.Contrasting to normalization, Kenny (2001) found there existed foreignization,which means the unusual collocations in the source texts are retained in thetranslations. For example, “mit zwei linen Augen” was literally translated into “withtwo left eyes”. Obviously, there is no “two left eyes” in English. However, only 22%unusual collocations in source texts are normalized in translations, so the assumptionof normalization is partly affirmed.For Kenny’s study, only 22% abnormal German collocations are formalized inEnglish translations, so this partly-affirmed conclusion is not fully convincible. Themost obvious problem is the corpus she conducted, because the study is based on theparallel corpus which is used to compare the source language with the target language,while the comparable corpus is helpful in finding the features of translated texts(Baker, 1995: 235). That is to say, Kenny’s study is not proper to deal with thecollocational study of translations. Nilsson (2004) did a research on the collocational pattern of “av” by applyingESPC (English-Swedish Parallel Corpus). Nilsson chose this word was out of thefrequency in translated texts which was much higher than that in non-translated texts.Then he explored a colligation of “av”, such as “ ndern av”, and found that thefrequency of the non of place in translated texts was three times of that innon-translated texts. The same result can be found in the collocational pattern of“prep+...+av”. It is believed that this unusual collocational pattern is influenced by thesource language, i.e. by English.Similar to Jantunen’s study, Nilsson’s study scope is also very limited. Theconclusion drawn from studying one word, a colligation and a collocational pattern isof less validity.

Chapter Three Methodology ....................... 10

3.1 Data Source ............................ 10

3.2 Instruments ....................... 14

3.3 Research Procedure ............................ 15

Chapter Four Odd Collocations in E-C Literary Translations ....................... 16

4.1 Odd Collocation .............................. 16C

hapter Five Conclusion ............................ 35

Chapter Four Odd Collocations in E-C Literary Translations

4.1 Odd Collocation
First of all, it should be clear what the odd collocation is. Then when analyzingthe odd collocations, it’s necessary to retrieve the odd collocation from thetranslations, so the KWIC (key words in context) is employed to see what context theodd collocation is in. By retrieving the odd collocation, it’s possible to see its usageand the development of the collocations. Scholars in and abroad have defined collocation in different ways. Firth (1957),the father of collocation, defines collocation as the company of words, a mode ofmeaning and an order of mutual expectancy. The Neo-Firthians, Halliday and Hasan(1976:317) argue that collocation is “the co-occurrence of lexical items that are insome way or other typically associated with one another”. Robins regards collocationas the habitual association of words. Nattinger and De Carrico (1992:36) thought it ina more comprehensive way that they defined it as words that habitually co-occur witha mutual expectancy greater than chance. However, these definitions seem to rely onthe linguists’ sense of language rather than the real data analysis. What researchersconsidered only seems to be the co-occurrence of words, Chinese scholar Wei Naixing(2002) gives a more comprehensive definition, that is, collocation is a sequence ofwords which are combined by some restrictions grammatically for a certain semanticmeaning. Each word in this sequence is mutual-expected and co-occurred. In short, acollocation should be characterized by co-occurrence and of both semantic andgrammatical meaning. Consequently, the odd collocation is the deviation of a normalcollocation. It should be pointed out that odd collocation is different from abnormalcollocation which has been well studied by Feng Guangyi (1993), for abnormalcollocation, such as “呐喊的荒野”, “金色的幸福”, is to deliberately created forpragmatic rhetorical effect, it is a deviation from the normal rhetoric, and should beunderstood in context. That is to say, in this study, collocations with figures of speechcould not be considered. Only those violate the semantic restriction, grammaticalrestriction and convention are odd collocations (Wang Dechun, 1987:21). In this section, considering the current obscure meaning of odd collocation, theworking definition of odd collocation is clarified at the very beginning. Thisdefinition distinguishes the odd collocation from abnormal collocation. After definingthe odd collocation, preparation for analysis is made, that is the retrieval of the oddcollocations in the E-C translations in 1920s. In view of the practical context, thecollocational span and the frequency are fixed. All of these make a preparation forfurther analysis in the following section.

4.2 Odd Collocational Features
Several corpus-based studies on collocations in the translated texts have beenconducted as mentioned before, yet little attention has been paid to the oddcollocations. Researches done by Mauranen, Dorothy Kenny, Maeve Olohan, QinHongwu and Wang Kefei, Wu Guangjun and so on certainly shed light on theexploration of the odd collocational features. Are the odd collocational featuressimilar to those of the normal collocational features, or they have their own features?In this section, in view of the former studies on collocation, odd collocational featureswill be explored from the following aspects. Generally speaking, translation is to some extent a third code, when it comes tothe studies of E-C translational features, two points should be taken into consideration.One is the contrastive studies of the overall features of translated and non-translatedChinese texts, while the other is the comparison between the original English and thetranslated Chinese. Consequently, collocational patterns will be explored with the helpof Corpus A, B, C and D. The observation of Corpus A is to tell the odd collocationalfeatures from the inner perspective. The comparison between Corpus A and Corpus C( in Table 5) aims at telling the odd collocational features in the 1920s’ texts from adiachronic perspective, while studies among Corpus A, B, C and D may tell the oddcollocational differences, especially the usage of these odd collocations in differentages. From the above table, we can easily find that though there are 34 kinds of oddcollocations in Corpus A, they are quite different in frequency. This is in consist withthe findings of Dayrell (2004) who once discussed the collocational patterns in bothtranslated and non-translated texts, and found that the translated texts tend to use acertain collocational patterns, not all the types of collocation. And in this study, onlyfive types are of significance, namely v+n, adj+n, num+quantifier+n, adv+v andv+adj. It is obvious that odd collocations in Corpus B, C, D are very similar in the oddcollocational patterns (see Table 7 and Graph 2), such as “num+quantifier+n”, “v+n”and “n+adj”, only different in frequency (especially the frequency of “v+n” and“n+adj” in Corpus D and “adj+n” in Corpus C) and odd collocational types (Corpus Bhas six odd collocational types while Corpus C and D have five types). But when itcomes to the patterns in Corpus A, there is a sharp difference, as shown in Graph 3.Graph 2 Pattern Distributions in Corpora B, C, D

Chapter Five Conclusion

This paper discusses the odd collocational features of the E-C literarytranslations in 1920s and its contributive factors to the oddness. Since thisis acorpus-based study on Chinese odd collocations, three referential corpora areemployed to find out the features of odd collocations in 1920s and to testify the roletranslation and Chinese development play in the forming of odd collocation. Somefindings are listed as follows:The first one is that in the 129 odd collocations in the 1920s’ E-C literarytranslations, only one odd collocation is caused by the English odd collocation, therest odd collocations have the normal corresponding English collocations. That is tosay, most odd collocations are caused by Chinese translations rather than the Englishsource texts. Comparing to the English texts, Chinese translations have more oddcollocations.Second, with the help of other three referential corpora, it can be found that thereare 34 kinds of odd collocations in the 1920s’ E-C literary translations. These 34 kindsof odd collocations are mainly lexical collocations rather than the grammatical ones.Besides, the most frequently used odd collocations are v+n（34.11%）, adj+n（17.83%）,num+quantifier+n（5.43%）, adv+v（4.56%）, v+adj（3.87%）.Third, these odd collocations are mainly those violate the semantic, grammatical,logic and non-conventional rules. To be detail, in a semantic odd collocation, it isusually the former collocate that breaks the expectation of the latter collocate, while inthe “  的  ”structure, there will be the mutual unexpectation of the collocates.Grammatically speaking, the odd collocations are usually those with “v+…” structure.In logical odd collocations, there will be logic conflict between the former and thelatter collocate. Furthermore, the non-conventional collocations are featured as“num+quantifier+n”structure.The last finding is that some odd collocations appeared in 1920s’ E-C literarytranslations but disappeared in 1920s’ non-translated texts, while some oddcollocations also appeared in 1920s’ non-translated texts.Limited by the Chinese knowledge about the collocation, the illustration of someinstances may not so professional. However, just as other corpus-based studies, thisresearch can only draw a tentative conclusion, but it may be helpful for further study,especially for this long-existed but few studied issue.
..............

参考文献（略）