A Corpus-Aided Study of Shifts in English-to-Chinese Translation of Prepositions

This research investigates three types of shifts in English-to-Chinese translation of prepositions, including PP (parallel shift), PP (omissive shift), and PX (transformative shift), across literary, expressive, operative text types using a corpus-analysis approach. The finding shows that there is an overall higher frequency of PP than PP and PX. The translation phenomenon is partially attributed to the differences in the number and function of prepositions between English and Chinese. In addition, different textual functions govern the translation shifts, and variation among four translated novels is diagnosed as relevant to the translator’s stylistic preference. In sum, this research supports the dynamic nature of the translation of prepositions under the impacts of contrastive linguistic differences between source and target languages, textual functions and the translator’s style.


Introduction
Translation shift inevitably occurs in the translation process and has been studied in a number of researches (Catford, 1980;Sager & Hamel, 1995;Al-Zoubi & Al-Hassnawi, 2001).In his book A Linguistic Theory of Translation, Catford (1980: 73) proposes the term shifts and defines it as "departure from formal correspondence in the process of going from SL to TL".Traditional studies of translational shifts, notably Catford (Note 1), are made within the framework of formal and contrastive linguistics, therefore being accused of a "static comparative linguistic approach" without considering pragmatic, extra-textual elements (qtd. in Munday, 2001: 6).Other studies relate translation shifts to extra-linguistic factors, such as text type in Levý (Note 2) (1969), translation style in Popovič (Note 3) (1976), and textual, pragmatic and stylistic elements in Al- Zoubi and Al-Hassnawi (2001).These researches, considering either static comparative linguistic or pragmatic, extra-textual elements, share one feature: a study of universal features and global models of translation shifts with a theory-oriented, prescriptive approach.However, the researches above lack statistical results to support the variation in translation shifts, so there is a need to conduct an empirical, descriptive study to provide statistical evidence.In addition, in discussing the relevance of translation shifts to linguistic factors, textual functions and relevant others, the researches above suggest a comprehensive translation phenomenon, not concentrating on the translation of one specific linguistic feature, such as preposition.Thus, there is also a need to limit a study on the shifts of the translation of a word class.
In light of inadequate researches on translation shifts above, this research investigates preposition translational shifts using a corpus-aided, descriptive approach.The corpus-based analysis moves from samples to theoretical generalization, and the statistical result can provide more objective, systematic and reliable finding than what it is in the qualitative, theoretical studies that moves from theoretical principles or/and norms to supportive examples.Since the main issue is preposition translation, the purpose of this research is twofold: 1) to justify the dynamic nature of preposition translation by comparing three preposition translational shifts across text types and among different translators of four novels through the corpus-based analysis, and 2) to infer the relevance of dynamic preposition translational shifts to both linguistic and extra-linguistic constraints based on some statistical results.Here, the linguistic factor refers to contrastive linguistic discrepancies between English and Chinese and extra-linguistic factors include textual functions, genre conventions and the translator's style.These factors ranging from the micro/textual perspective to the macro/extra-textual perspective, to varying degrees, manipulate the translator's decision-making in the choice of appropriate or/and preferable strategies for translating prepositions from English into Chinese.

Preposition Translation Researches
Translation of prepositions, particularly from English to Chinese, is worth a study because preposition is a distinctive linguistic feature in English but not in Mandarin Chinese.A preposition links nouns, pronouns and phrases to other words in a sentence (MacFadyen, 2007).Its essential function is to indicate the temporal, spatial or logical relationship of its objects to other components of a sentence.In other words, prepositions locate nouns or pronouns or phrases that come after them immediately in certain space and time, or depict how a thing is done or take place and how a relationship is developed (Simmons, 1997(Simmons, -2007)).
The number of English prepositions overwhelmingly exceeds that of Chinese prepositions.In his book entitled 英 漢翻譯比較 [Comparison between English and Chinese and their Translation], Cheng (2004: 99) maintains that "the English language has 286 types of prepositions, including various prepositional phrases, but the Chinese language has only more than ten types based on the statistics made by Curme (1931) in his book Syntax" (Translation mine).Lian (1993) claims that prepositions in Mandarin Chinese are basically loaned from verbs, and cannot be regarded as genuine prepositions.She argues that the Chinese language per se has no prepositions, and there is a huge gap in the number of prepositions between English and Chinese.For this reason, it is hard and challenging to find the same amount of equivalents in Chinese to replace English prepositions in the translation process.

Contrastive Linguistic Analysis of Prepositions
From the functional viewpoint, English prepositions are mainly used to make sentences grammatically correct.Some of them simply occupy the positions but do not mean anything.These prepositions are identified as grammar words, not content words.Deleting some prepositions in the Chinese translation would not distort or change the original message.Chinese sentences without these prepositions remain readable and comprehensible to native Chinese speakers.For example, the preposition of in the expressions be fond of, be tolerant of, be suspicious of and be afraid of can be deleted and these expressions are simply translated into Chinese as 喜歡 [lit: like], 忍受 [lit: endure], 懷疑 [lit: doubt], and 害怕 [lit: fear].Some other English prepositions can only be replaced with Chinese verbs or other forms.For example, the preposition without in the sentence He could not make a success without her help carries the meaning of not having, so that it is translated as a verb 沒有 [lit: do/does not have] in Chinese.Furthermore, the English preposition by in the sentence He traveled by train suggests by means of and taking, but the translation of the first meaning with a Chinese preposition 藉由 [lit: by means of] renders the Chinese translation very awkward and unnatural.Thus, the second meaning is adopted and is translated as 乘坐 [lit: take/ride].More often than not, some English prepositions can be directly replaced by Chinese prepositions.For example, with in the sentence I ask you to join with me and do this can be translated as 與 [lit: with], and the entire sentence translation is 我請您們與我一起行動 [lit: I ask you with me together act].

Three Shifts in the Translation of Prepositions
The above observations lead us to a tentative conclusion that the translation of English prepositions into Chinese involves parallel, omissive and transformative shifts.The parallel shift means that English prepositions are replaced with Chinese prepositions; the omissive shift suggests that English prepositions are omitted in Chinese.The transformative shift refers to the transformation of an English preposition into other forms, such as verbs and conjunctions.These three shifts in preposition translation can be signaled as PP (parallel shift), PP - (omissive shift), and PX (transformative shift).This preliminary generalization can find support from other researches.In the book entitled 科技英語與翻譯 [English in Science and Technology, and Translation], Cheng (1999) declares that translators frequently use three strategies to translate prepositions from English into Chinese: Zhao-I [照譯], Sheng-I [省譯] and Zhuan-Huan [轉換] (Note 4).Yu (2007) proposes a series of strategies of translating English prepositions into Chinese, including omission and modification of prepositions into verbs and other forms.Xia (2004) discusses the similar principles and strategies for translating prepositions and prepositional phrases in English business contracts.Hu and Zhang (2005) address the issue of translating orientation prepositions from English into Chinese from the perspective of a transformation from static to dynamic constructions.Fan (2007) raises some problems of translating the English preposition for in science writings, and proposes omission and transformation strategies to reduce translation errors.In addition, Song (2005) discusses what strategy is most frequently and least frequently used when translating prepositional phrases from English into Chinese.These researches of English-to-Chinese preposition translation concur with my observation that the translation of English prepositions into Chinese may be generalized into three types of shift-replication, omission and transformation.

Research Questions
With the hypothesis above in mind, I investigate whether or not and to what extent the three shifts in preposition translations vary across text types and among different translators within the same genre by analyzing three parallel corpora of 303,998 words.In addition, I will probe how variation in preposition translation shifts is related to some linguistic and extra-linguistic factors, such as English-Chinese contrastive linguistic differences, textual functions, genre conventions and the translator's style.Thus, some research questions for investigation are raised as follows: 1) Which type of shift holds an overall highest or lowest frequency in the translations of three text types together?
2) How do the three shifts vary in relation to different text types and different textual functions?
3) How is the variation in the three shifts among the four translated novels related to the translator's style?
In short, this study attempts to justify the dynamic nature of the translation of prepositions that results from the constraints of contrastive linguistic differences, distinctive textual functions, genre conventions and the translator's style.Although other influential factors, such as the editor's style or the publisher's policy and the content of the source texts are worth notice, they are not discussed in this study.They can be reserved for future studies.

Textual Function and Genre Convention
Since textual function and genre convention are used as extra-linguistic factors to explain the variation in preposition translation shifts across text types, the basic concepts of textual function and genre convention must be illustrated at some length.Textual function refers to the way in which a certain type of text is presented or used for the fulfillment of a certain purpose.Textual function determines text types.Bühler's (1990) theory of functional typology proposes a three-way distinction, labeling three text types as expressive, informative and vocative, depending on the foci of "the producer (emotive), the subject-mater (referential) or the receiver (connotative)" (qtd. in Bell, 1991: 204).Similar to this functional typology is Reiss's (1977Reiss's ( /1989Reiss's ( , 2000) ) three-way division of texts into informative texts, expressive texts and operative texts based on their functions (qtd. in Hatim, 2001:77).From Reiss's (1977Reiss's ( /1989Reiss's ( , 2000) ) viewpoint, informative texts are used to convey information; expressive texts communicate thoughts in a creative way, and operative texts attempt to persuade the audience.To be more specific, informative texts fulfill an informative purpose while expressive texts have a literary purpose and operative texts have an emotive, persuasive purpose.Thus, the successful translations of informative, expressive and operative texts respectively fulfill the purposes of informational transmission, aesthetic appreciation, and emotional appeal.
Different textual functions also impose some constraints on the linguistic presentations of text types.We may define this constraint as genre conventions that bring about conformity and expectation (qtd. in Kussmaul, 1997: 68;Lewis, 1969: 78).In Kussmaul's (1997: 68) words, "when people produce specific text types, such as business letters, they can be expected to conform to specific regularities and rules".People use the conventional words in a specific genre because "they are normally expected to use them in the same way as other people do" (Kussmaul, 1997: 68).If writing or speaking does not conform to the generally-agreed regularities in certain genres or/and text types, "communication may turn out to be difficult or even break down" (Kussmaul, 1997: 68).Thus, when the translation of technical manuals, for example, are targeted at the non-technical audience, the prepositions that present locations, purposes, means, directions, and movement need to be explicitly presented.In a different manner, Chinese speeches emphasize the persuasive effect, so that verbs rather than prepositions are paid attention to, and the Chinese translator is likely to highlight verbs or nouns that present actions and events at the cost of prepositions.

The Translator's Style
Since the translator's style has been regarded as one potential factor to govern preposition translational shifts, there is a need to briefly introduce its main concept.The translator's style generally refers to the translator's personal preferences and idiosyncrasies.House (1997) interprets the notion of style as "variation in the level of formality" and as "patterned choices across all linguistic levels" (qtd. in Baker, 2000: 242).This remark stresses that the style of a translator stems from specific choices or translation strategies made by the translator.In other words, a translator's style can be his/her consistent use of specific strategies and shows his/her particular linguistic habit in translation.In Baker's (2000: 245) view, a translator's style is "a matter of patterning" and "preferred or recurring patterns of linguistic behavior" of a translator.For example, translator A is identified as presenting a more concise, terse style than translator B because the average sentence length of translator A is shorter than that of translator B. In other cases, translator C is regarded as having the flowery style and translator D shows a plain style because translator C uses larger varied types of nouns and adjectives than translator D.
There are numerous methods of distinguishing a translator's style.In this research, the translator's style is defined as the translator's characteristic use of prepositions in the Chinese translation.This suggests that one translator preferably uses the strategy of PP, but others use the strategies of PP -or/and PX when they translate the texts of the same type.The use of different preferable strategies suggests the effect of the translator's personal style although the contents of the translated texts are more or less related.Noticeably, the translator's style will not influence the good or bad quality of a translation.

Methodology
This research adopts a corpus-aided approach (Note 5) to the investigation of preposition translational shifts.The use of a huge volume of authentic data is expected to obtain the more objective and more reliable result.Furthermore, this study efficiently fulfills a statistical survey with the aid of some corpus-processing tools (Note 6).Nonetheless, appropriate corpus design and specific search criteria are required if we want to obtain a complete and relevant result.These two components will be introduced as follows.

Corpus Design
Three parallel corpora of translations are used in this research for analysis.One is composed of informative texts with their Chinese translations, and contains a total of 101,693 words (hereinafter PCIT).The other is composed of the first chapter of the four English novels with their Chinese translations, and contains a total of 100,089 words (hereinafter PCN).Another is composed of English speeches and advertisements with their Chinese translations, and contains a total of 102,216 words (hereinafter PCSA).A combination of these three corpora has 303,998 words.
The authors and translators of informative texts are anonymous.In addition, the publication time of these texts is not known, so only the time for collecting these materials is provided.Appendix 1 shows the internal structure of PCIT.The translators of the four novels are Liu Zhu-hai [劉珠還], who is an English Professor at the College of Foreign Languages of Anhwei Normal University, Zhuo Miao-rong [卓妙容], who is Accounting and Business Administration majors, Shi Qing-zhen[施清真], who is Journalism and Mass Media majors; and Ho Wen-jing [何 文敬], who is a British and American Literature major.The four novels are randomly selected and their translations are easily understood.Most importantly, the publication time for the four translated novels is similar, ranging from 2002 to 2006.Appendix 2 shows the internal structure of PCN.
It is not easy to collect the bilingual (English-to-Chinese) translations of speeches and advertisements that are published or released during the similar period of time, so the materials in PCSA show a larger gap in the translation time (from 1997 to 2006) than other two corpora do.Appendix 3 shows the internal structure of PCSA.

Search Criteria
This research explores preposition translational shifts, so six categories of prepositions are selected as key words for search.These prepositions carry out diverse functions in English.Table 1 shows the six categories of prepositions for a corpus-aided investigation.The six categories of English prepositions play a very important role in English writing and speaking.Omitting a preposition where necessary makes an English sentence grammatically incorrect.However, native Chinese use more verbs than prepositions in daily communication and thereby the translation of English prepositions into Chinese must comply with Chinese linguistic conventions and undergo some grammatical shifts.

Findings
The result of the investigation showed that PP -held the highest frequency (46.35%) in the three types of texts together.Nevertheless, in PCIT (parallel corpus of informative texts), PP took up a higher frequency (44.63%) than PP -(38.96%) and PX (16.41%).In PCN (parallel corpus of novels), PP also took up a higher frequency (47.34%) than PP -(45.9%) and PX (6.76%).Unlike them, PP -had the highest frequency (54.2%) in PCSA (parallel corpus of speech and ads.), and PP took up 36.6%;PX, 9.2%.Figure 1 shows the statistical result of the three shifts in the translations of prepositions across text types.APC means all parallel corpuses, a combination of PCIT, PCN and PCSA.It is noted that omissive translation of prepositions is the most preferable, while the transformative translation of preposition is the least preferable in overall English-to-Chinese translations.To further support above statistical results, "Mann-Whitney U Test" was conducted by selecting "type" (PCIT, PCN, PCSA and APC) as independent variables, and the three shifts as dependent variables.The result showed that the p value reached the significant level (p < 0.05).Thus, the result rejected the null hypothesis that there was no significant difference among PP, PP -and P→X among PCIT, PCN, PCSA and APC.It is confirmed that there is a significant difference among the three shifts in the translations of prepositions across three text types and in all text types together.
The proportions of the three shifts also showed variation among the four translated novels.The Human Stain (THS) and The Beloved (TB) had a higher frequency of PP than PP -.Oppositely, The Book of Salt (TBS) and The Bonesetter's Daughter (TBD) had a higher frequency of PP -than PP.In THS, PP held 50%; PP -, 43.47% and PX,6.53%.In TB, P P held 50.14%;PP -, 42.1% and PX,7.76%.In TBS, PP held 44%; PP -, 47% and PX, 9%.In TBD, PP held 37.42%; PP -, 58.49% and PX, 4%. Figure 2 shows the different proportions of the three shifts among the four translated novels.To support the above statistical result, the "repeated-measure one-way ANOVA test" was used to confirm that the p value was less than 0.05, so that the null hypothesis was rejected that there was no significant difference in PP, PP -and P→X among four translated novels.Actually, the implications of the findings are more important than the statistical figures, and the following section discusses the relevance of preposition translational shifts to the impacts of languages, text types and the translator's stylistic preference.

Discussions and Implications
Identifying varied proportions of the three shifts across text types and among four translated novels is not an end in itself.The statistical results will be used to infer and discuss the relevance of parallel, omissive and transformative translations of prepositions to some linguistic and extra-linguistic constraints.

Contrastive Linguistic Impacts
Analysis of the three types of shifts in the translations of prepositions in the three text types together shows a notable tendency of P= P -than other two.From the viewpoint of English-Chinese contrastive linguistics, this phenomenon occurs because of the distinct differences in the number, functions and meanings of preposition between English and Chinese.It is often found that when some English prepositions are literally translated into Chinese, they are functionally redundant and thereby can be deleted without ruining or distorting the original message in the Chinese translation.This is viewed as optional shifts or individual shifts.According to Bakker, Koster and Leuven-Zwart (1998: 230), this kind of shift is usually "prompted by the stylistic propensities and the subjective idiolect of the individual translator".In contrast, some English prepositions cannot find their Chinese equivalents and can only be transformed into other forms.This phenomenon is identified as obligatory shifts.
Popovič (1976: 16) defines obligatory shifts as an "inevitable shift that takes place in the translation as a consequence of differences between the two languages, the two poetics and the two styles of original and translation".In this research, the obligatory shifts mainly result from the lack of a Chinese equivalent for the English preposition, whereas optional shifts are attributed to an attempt to achieve some textual functions or the translator's idiosyncratic preference.The following examples, respectively extracted from the three corpora, support optional shifts (Examples 1-2) and obligatory shifts (Example 3-5).However, the translator omitted the English preposition without ruining the original meaning.This is regarded as the optional shift, and suggests that English sentences rely on prepositions for grammatical accuracy or/and semantic clarity, but Chinese sentences can make use of correct word order to clearly present the information.
Ex.2 SL: How strong it was he had never entirely realized until he counted all the people, department by department.Still difficultly, the preposition at in this example cannot be literally translated as 在 [lit: at]; otherwise its translation would be awkward to the native Chinese audience like 如同面對在一起嚴重的高速公路事故、一場 大火、一場爆炸….[lit: as if confronting at a severe highway accident, a big fire, an explosion….].Thus, the preposition at must be omitted when the English segment being present is translated as a verb confronting.This translation is identified as the obligatory shift, and suggests that English uses much more prepositions than Chinese.In brief, whatever the omission of English prepositions in Chinese translation is the obligatory or optional shift, it suggests that some contrastive linguistic differences between English and Chinese have governed preposition translations.

Textual Function Impacts
This section discusses how textual functions affect the variation in preposition translational shifts across three text types.The finding shows that despite an overall higher frequency of PP -, PP presents a higher frequency than PP -in both informative and expressive texts.In addition, PP -holds a higher frequency than PP and PX in operative texts.This variation can be inferred as derived from different textual functions and specific genre conventions.
Since the translation of an informative text aims to clearly transmit information, it has to explicate the time, space and logical relationship of a message.English prepositions are intended to perform these functions, so the translator literally renders them into Chinese to achieve the original textual functions.This way of translating English prepositions with the PP shift helps the audience to easily grasp the time, location, purpose and other relationships.The preposition translations of following examples support this point.
Ex.6 SL: Select Switch calls to switch between the two calls.

TL: 選擇切換通話，在兩通電話間切換。
[Select Switch calls, at two calls between switch.] In this example, the literal translation of the English preposition between enables the target audience to clearly understand spatial information.The parallel translation aims to achieve the function of informative texts-to clearly inform how to operate or use a device.
Ex.7 SL: You can save a maximum of ten codes in the code memo.

TL: 您可以在密碼備忘錄內儲存最多十個密碼。
[You may at code memo inside save maximum ten codes.] The above example shows that the literal translation of the prepositions in helps the audience to immediately know where to switch calls and where to save codes.The higher frequency of PP in technical translation is conducive to the explicit transmission of information, meeting Hatim and Munday's (2005) argument that the successful translation of an informative text must clearly convey information.
The expressive text similarly shows a higher frequency of PP than other two shifts, but it has a different reason for the phenomenon.The prepositions in the novels are intended to create aesthetic effect other than transmitting information.In order to reproduce the original artistic form or to preserve the author's creative style, the translator tends to literally translate English prepositions into Chinese.Examples 8 and 9 support the parallel translation.
Ex.8 SL: Past the storeroom, past the cold house, the privy, the shed, on around to the porch.
TL: 經過儲藏室、經過冷藏室、廁所、柴門，一直繞到門廊。 [Past storeroom, past cold house, privy, shed, all the way detouring to porch.] In this example, the second preposition past could have been omitted in the Chinese translation for conciseness, but the translator preserves it to replicate the original literary from.
Ex.9 SL: The water running over their heads made it worse.

TL: 從頭頂上流下來的水讓情況更糟。
[From head top running down de water makes situation worse.] In Example 9, the preposition over could have been deleted and the Chinese translation could be 頭頂上流下來 的水 [lit: head top running down de water] without ruining the original message, but the preposition is translated as 從[lit: from].The English preposition over is replaced with another preposition from.The adoption of parallel translation helps retain the original preposition, concurring with Reiss's (1977Reiss's ( /1989Reiss's ( , 2000) ) view that the translation of the expressive text aims to seek the identical effect and to achieve the goal of aesthetic appreciation.Retaining as many English prepositions as possible in the Chinese translation can keep intact the original literary form.
In contrast to the above two, PP -shows a higher frequency in the translation of an operative text.This translation phenomenon partially results from the translator's attempt to create the powerful, persuasive effect.As mentioned earlier, operative texts are used to persuade the audience or to inspire the audience, and thereby the translator tends to omit prepositions to shift the audience's attention from the surrounding conditions to some events and core messages in speeches.The following two examples demonstrate that the translator may find Chinese equivalents for English prepositions, but omit them to change the audience's attention.
Ex.10 SL: Last year every American was horrified and heartbroken by the tragic killings in Jonesboro, Paducah, Pearl, Edinboro, Springfield.

TL: 去年，每個美國人都為薑斯堡等地的悲劇性槍殺案感到很震驚和傷心。
[Last year, every American all for Jonesboro etc places de tragic killing case felt very horrified and sad].
This translation could have called the reader's attention to the place when the preposition of location (where the shooting incident occurred) is directly translated as 在薑斯堡等地的悲劇性槍殺案 [lit: at Jonesboro etc places de tragic killing case].Notwithstanding, to highlight the severe event, tragic killing, the translator omits the preposition in and tries to move the audience's attention from the location to the event.In fact, the purpose of the speech is to call people's attention to the terrible killing and to refrain from it in the future.The information of the location for the killing is not important.Let us look at another example.
Ex.11 SL: With tax credits and more affordable student loans, with more work-study grants and more Pell grants, with education IRAs and the new HOPE scholarship tax cut that more than 5 million Americans will receive this year, we have opened the doors of college to all Americans.
TL:今年，將有 500 萬人受益於稅收削減、助學貸款及各種獎學金的增加。受高等教育的大門終於向全體 美國人敞開了。 [This year, there will 500 million people benefit from tax cut, educational loan and various scholarship de increase.Receiving high education de gate finally to all Americans open le.].
This translation highlights the good news of higher education offered to an increasing number of Americans, and the preposition with [lit: due to] is a grammar word.To shift the focus from the grammar word to the number of people who can benefit a new policy, the translator uses a different word order to convey the information and omits the preposition with.This omissive translation suggests that the translation of a political speech aims to draw the audience's attention to some policies, news and events, and other secondary information can be modified.The omissive translational shift meets Hatim and Munday's (2005) argument that the successful translation of operative texts is to create some psychological appeal to the audience.
In short, distinct textual functions, though not absolutely but subconsciously, have governed preposition translational shifts.The higher frequency of PP in the translation of informative texts aims to clearly present spatial, temporal or conditional information whereas the translation of expressive texts attempts to reproduce literary forms with the similar higher frequency of PP.Differently, the translation of operative texts shows a higher frequency of PP -because omission of prepositions could shift the audience's attention from spatial and temporal conditions to specific events and core messages in an operative text.

The Translator's Stylistic Impacts
After the impacts of contrastive linguistics and textual function are discussed, this section emphasizes how the translator's style could be a factor to cause the variation in translational shifts within the texts of the same genre.
A comparison of the four translated novels shows that The Human Stain and The Beloved have a higher frequency of PP than PP -, but The Book of Salt and The Bonesetter's Daughter have a higher frequency of PP -than PP.This variation can be inferred from the perspective of four translators' stylistic preferences.The two translators who are language or literature majors are inclined to follow the principle of faithfulness in literary translation, and they opt for P P. The following two examples illustrate that some prepositions could have been deleted in the translation without ruining the original message but are literally translated by the translator.
Ex.12 SL: Yet Coleman, by abandoning a draft of a book as bad as the draft he'd finished, had somehow managed to swim free not only from the wreck of the book but from the wreck of his life.
TL: 然而，柯爾曼僅以甩掉一部像他剛完成的那樣拙劣的書稿，就不僅成功地從自己書的殘骸中，而且 還從自己生活的殘骸中，遊出水面，重獲自由。 [Nevertheless, Coleman, simply by getting rid of a like he just finished de that poor book draft, and not only successfully from own book de wreck inside, swam out water surface, regain freedom.] In Example 12, if we delete the preposition by and the Chinese translation becomes 柯爾曼甩掉一部….[lit: Coleman got rid of a….], it does not change the original message and remains understandable.However, the translator uses the parallel translation by retaining the preposition.This can be attributed to the translator's personal choice.In Example 13, the preposition from can be deleted and the translation can be …親愛的， 這是 『甜蜜之家』 的保羅四 [lit: …Honey, this is "Sweet Home" de Paul D.] without changing the original meaning.However, the translator adopts the parallel translation by retaining the original preposition.This explains the translator's stylistic preference.
In contrast, the other two translators who are business-specific majors opt for free translation and henceforth use the omissive strategy more often than other strategies.The following two examples support their preference for omissive translation.
Ex. 14 SL: Four, pick up the girls from the rink and drop them off at their friend's house on Jackson Street.

TL: 四，到溜冰場接小孩，然後帶她們到傑克森街的朋友家。
[Four, to rink pick up kids, then take them to Jackson Street de Friend house.] In Example 14, the preposition on could have been literally translated and the translation could be …，然後帶她 們到在傑克森街的朋友家。[lit: …and take them to Friend's house on Jackson Street].However, the translator adopts omissive translation, not parallel translation.Perhaps the translator views the parallel translation of a preposition on as redundant and thereby omits the preposition.This can be attributed to the translator's stylistic preference.In Example 15, the preposition from could have been literally translated, and the translation would be 從過去的 經驗她知道，每次….[lit:From the past experience she knew, each time….].However, the preposition from is deleted in Chinese translation.Obviously, the translator opts for the omissive translation of the preposition.
Without changing the original meaning, either omission or preservation of prepositions reflects the translator's stylistic preference in addition to the governance by the textual function or contrastive linguistic differences between source and target languages.Some translators choose to retain original prepositions while others focus instead on the actions and specific issues by omitting or/and transforming prepositions into other word classes.
The personal choice concurs with Popovič's (1976) concept of individual shifts that are "related to the subjective idiolect of the individual translator" (qtd. in Bakker, et al., 1998: 230).In this respect, the translation practice has been justified as partially constrained by the translator's subjectivity or his/her own idiosyncratic decision.

Conclusion
Contrasting with the preposition researches that discuss translational shifts from the theory-oriented perspective using a prescriptive approach, this research has identified the variation in translational shifts across text types and among different translators using a descriptive, corpus-aided approach.The findings justify the dynamic nature of preposition translational shifts.The implication is that shifts in the translation of prepositions could not be fixed without distinction.Instead, they vary as the result of linguistic, functional and stylistic impacts, including English-Chinese contrastive linguistic differences, textual functions of different text types and the translator's personal styles.The overall higher frequency of PP -than PP and PX relates to different ways of using prepositions between English and Chinese.In addition, the variation in the three types of translational shifts across text types is inferred as being governed by different textual functions and different genre conventions.Set within the same text type, the variation among four translated novels is closely relevant to the translator's style.
Although the result of this research, compared to other studies to date, is less subjective by using a corpus-aided approach, I believe that a more genuine result can be found if there is a larger size of corpora.However, this research has shed light on some significance that preposition translational shifts are closely related to a host of factors such as contrastive linguistic differences between source and target languages, textual functions and the translator's style.As final suggestions, this research can be replicated for other linguistic features (e.g., adverb or pronoun) and other language pairs (e.g., English to Japanese or Chinese to Spanish and others) to gain more significant insights in future translation researches.Catford (1980) proposed level shift and category shift as the universal modes.Level shift is "the replacement of a SL term with a TL equivalent" at grammatical, lexical, graphological and phonological levels (qtd. in Munday, 2001: 60-61).Category shifts are subdivided into class shifts, structure shifts, intra-system shifts and unit shifts.Structural shifts denote the replacement of a grammatical structure in the SL with a different one in the TL; intra-system shifts suggest the selection of a non-corresponding term in the TL although the SL and the TL belong to the same language system.In addition, unit shifts indicate the substitution of a unit in the SL for a unit of the different size in the TL, whereas class shifts refer to the transformation of the class of one lexical item into another in the translation process.
Note 2. Levý (1969) notices that translation shifts depend on the text type that has its characteristic linguistic features, including "denotative meaning, connotation, stylistic arrangement, syntax, sound repetition (rhythm, etc.), vowel length and articulation" (qtd. in Munday, 2001: 62 and 19).In Levý's (1969) view, denotative meaning is most important in the translation of a technical text and must be reserved in the TL text, but poetry translation emphasizes rhythm, connotation and form.
Note 3. In Dictionary for the Analysis of Literary Translation, Popovič (1976) considers translation style an influential factor in determining translation shifts, and then distinguishes between constitutive shifts and individual shifts.This division concurs with a distinction between "obligatory shifts" and "optional shifts" (qtd. in Bakker, et al., 1998: 228).
Note 4. The first strategy preserves the original form of an English preposition in Chinese; the second, transposing the English preposition into another word class in Chinese, and the third, omitting the English preposition in Chinese.
Note 5.The corpus-aided, descriptive approach, initiated by Mona Baker, is used in modern translation studies to investigate the universal features of a translational language, such as explicitation, simplification and conventionality or normalization.
Note 6.The author used the ParaConc tool to align SL and TL texts and obtain the concordance output through a query for some key words.ParaConc, a commercially released software, provides the functions of basic concordance search, bilingual search, bilingual KWIC, collocations, and others.The concordance output contains the queried preposition in the English sentence and its corresponding Chinese translation.

Figure 1 .
Figure 1.Frequencies of three shifts in the translations of prepositions across text types

Figure 2 .
Figure 2. Frequencies of three shifts in the translations of prepositions among four translated novels standby, press and hold a volume key for one second.TL: 在待機狀態中，按住一個音量鍵一秒。 [In standby situation, press a volume key one second].The preposition for in Example 1 could have been translated as 長達 [lit: last], rendering the Chinese translation as…按住一個音量鍵長達一秒 [lit:…press a volume key long to one second].
she is my Denver.This is Paul D, honey, from Sweet Home."TL: 「瞧，她就是我的丹佛。親愛的，這是從『甜蜜之家』來的保羅四。 」 [Look, she is my Denver.Honey, this is from『Sweet Home』come de Paul D.] knew from experience what happened whenever she asked her mother to render Chinese characters into English.TL:過去的經驗告訴她，每次她請媽媽把中文翻成英文，結果準沒好事。 [Past experience told her, each time she asked Mother get Chinese translate into English, result exactly nothing good.]
[This power exactly has how powerful; he always did not understand, until he one department one department calculated exactly has how many people].The preposition by in Example 2 could have been translated as a verb in Chinese 接著 [lit: follow/link] and the translation would be …直到他一個系接著一個系地計算出究竟有多少人[lit:…until he one department following/linking one department calculated exactly there is how many people].But the translator directly omits by and the Chinese translation is still clearly understood by native Chinese.This optional shift suggests that adverbial expressions in Chinese have flexible forms and English has to use prepositions to connect preceding and proceeding linguistic components for adverbial expressions.Select different concerning picture effects].We can only but omit for after we find it infeasible to translate this preposition as the Chinese verb 給 [lit: give] or 提供 [lit: provide].The omission of the preposition is identified as the obligatory shift, suggesting a contrastive linguistic difference that English uses a preposition for adverbial expressions (noun--prep.--location),butChinese presents it using a kind of possessor (location--de/of--noun).The preposition at in the phrase at a time cannot be literally translated as 在一次 [lit: at one time] because it does not conform to Chinese linguistic convention.Thus, it is translated into another adverbial expression 逐一 逐次地 [lit: successively].Transformative translation of the English prepositional phrase is arbitrary and thereby the translation belongs to the obligatory shift.This shift suggests that English has many prep-led phrases for adverbial expressions, but Chinese preferably uses varied forms of adverbial expressions without a preposition to connect words.

Table 2 .
The internal structure of PCIT

Table 3 .
The internal structure of PCN