Article body

1. Introduction

The year 2003 saw the publication of Hillary Rodham Clinton’s autobiography Living History,[1] which was followed by the publication of a Taiwan Chinese translation entitled Huo Chu Li Shi[2] (Live out/through history, published by China Times), and then a mainland Chinese version Qin Li Li Shi[3] (First-hand experience of history, published by Yilin Press), all in the same year. Soon after the publication of its translation, the publisher Yilin was accused by the English publisher, Simon & Schuster, of “a breach of contract.” From Simon & Schuster’s news release (2003), [4] the “breach” can be seen as having two-fold implications: one is technical and professional, for “numerous [unauthorized] changes and deletions” were found in the translation, and the other is political or ideological, as “Senator Clinton’s views about China” have been in a way altered. As a result, both the author and the publisher demanded “a faithful and accurate translation.”

Judicial implications apart, it is of interest to see how the production of the two Chinese versions can present itself as a case example for the study of the relationship between a translation and the perceived social context in which it is intended to function, an approach informed, among other things, by the so-called functionalist model under the name of skopos theory developed during the 1980s and 1990s. According to this model, the efficacy of a translation should be viewed in the light of its skopos, i.e., its intended social functions, in the target culture rather than its faithfulness to the source text. The skopos of the target text, as Schäffner (2001: 236-237) succinctly summarizes, is “determined by the initiator’s, i.e., client’s needs,” and “largely constrained by the target text user (reader/listener) and his/her situation and cultural background”; whilst the source text is “just one constituent of the commission given to the translator.”

Viewed from such a social functional perspective, the situation surrounding the pair of Chinese translations is unique in the following senses:

  1. The Yilin version is professed to be largely a rewrite of the China Times version, a practice agreed upon by both Chinese publishers and following the English source text sentence by sentence;[5]

  2. The rewrite was carried out by the same team of translators, the only difference, it seems, is the order in which their names appear on the title/copyright pages of the two publications;

  3. The socio-cultural contexts, in which the two versions are intended to function respectively, are similar in ethnic, cultural and linguistic terms but opposed to each other in political and ideological terms;

  4. One of the renderings has been accepted, or at least acquiesced to, by the original author and publisher, whilst the other has been called into question.

One issue therefore stands out in this case: How the impact of an initiator[6] (or the publisher in our case), as an institutional factor in social functional terms, has been actualized at the level of discourse in the production of a translation, as seen in the Yilin version as against the China Times version and the English original. The present paper is an attempt to examine the matter through a comparative study of the three texts.

In order to highlight the ideological implications of such an institutional factor, some less consequential factors are disregarded. For example, whilst the target language is virtually the same, there are region-specific expressions (analogous to those one may find among ‘varieties of English’) that may be responsible for dialectal idiosyncrasies between the Chinese versions. This allows us to identify and concentrate on any discrepancies among the three texts that may indicate an ideological inclination.

Also, given the size of the parallel corpora and the scale of the current investigation, we have to identify a representative element of the texts for closer examination. In this connection, Simon & Schuster’s press release proves to be a very helpful lead, for it pinpoints two particular chapters where changes and deletions that affect the presentation of the author’s views about China have been found. One of the chapters is Chapter 22, Women’s Rights are Human Rights (p. 298-310). A preliminary comparative reading of the texts shows that deviations in this chapter are related to the following topics: Harry Wu, the Chinese government / officials, and China in general. To keep the study within a manageable yet meaningful scope, we have chosen the topic of Harry Wu as our focus in this project, for, compared with others, it provides a consistent as well as politically sensitive scenario for concentration.

2. The presentation of the Harry Wu topic in the three texts

2.1. The English source text

The Harry Wu topic is included as a prelude to the main event of the chapter: the US participation in the UN women’s conference held in Beijing in 1995. It is presented against the backdrop of the diplomacy between the US and the Chinese government over the former’s participation in this conference, with a detailed account of Wu’s political background, i.e., his first incarceration in Chinese labor camps as a political prisoner (p. 298) and in particular his second imprisonment in the mainland as a spy. Such a presentation is obviously indicative of the author’s concern over the case whilst providing readers with necessary information for relating an individual’s fate to a nation’s diplomacy in a way intended by the author.

The use of the conjunction although and the modifier valid to describe Wu’s arrest in the following line gives an example of the author’s position on the matter: “Although he had a valid visa to visit China, he was charged with espionage and thrown in jail to await trial” (p. 298). The presentation about Wu’s conviction and eventual release after a sham trial (p. 300) indicate a more explicit ideological stance when the author stated that the Chinese government could not bear more bad publicity (p. 300) over Wu’s case.

On the other hand, a close connection is established between Wu and the US government to illustrate the US concern over China’s human rights situation. For example, in the following lines, the use of thrown into limbo and a diplomatic bind stresses the US commitment by linking Wu’s case to the US participation in the conference; and that of privately and publicly stresses the extent to which the US has made its commitment known.

  1. Months of meetings and strategy sessions with representatives from the United Nations and other countries were thrown into limbo after Wu’s imprisonment.

  2. We were in a diplomatic bind in which Harry Wu’s imprisonment and my attendance at the conference were linked.

  3. Our government continued to state privately and publicly that I would not attend if Mr. Wu remained under arrest.

Clinton 2003a: 300

In the end, with the settlement of Wu’s case (i.e., Wu’s release from the Chinese prison), the US government, including the White House, State Department, and the President, finally decided that the author should attend the conference held in Beijing. The main event of the chapter is then brought in. The final mention of Wu’s case associates the topic once again with the women’s conference via a positive evaluation of the consequences.

2.2. The China Times version

Similar to the English text, the story develops from the protagonist’s political background into a diplomatic struggle between the US and the Chinese government over the topical event and the women’s conference held in Beijing. The struggle ends in Harry Wu’s release and the US final decision to attend the conference. In this process, apart from the listed expressions in the above section, renderings deviating from their English counterparts, as exemplified in italicized expressions in Table 1, are especially notable in revealing the author’s ideological stance over the topical event and China’s human rights status.

Table 1

Deviations in the China Times Version

Deviations in the China Times Version

-> See the list of tables

2.3. The Yilin version

At the very beginning of the chapter, the tone for the story is set via the author’s evaluation on Wu’s imprisonment, to the effect that under normal conditions, it would have been overlooked by the American media (possibly because it was assumed to be an internal affair of China). What follows is a brief introduction about the topical event. A link between the topical event and the US preparation for the women’s conference is then created when the former acts as the temporal foil for the latter. Finally, Wu’s case was settled (i.e., Wu was expelled from China) prior to the conference, and the US decided that the author should make the trip, despite Wu’s objection.

Noteworthy is, among other things, the deletion of all such mentions in the English version as listed in Table 1, which is indicative of a distinct ideological position over the topical event and China’s human rights status.

3. Mediation, selection, and discursive presentation

According to the functionalist conception, translation is a purposeful social functional activity which involves other parties besides the translator proper. Translation conceived as such extends from textual composition as the translator’s personal concern to text publication as a public enterprise in which such other parties are more actively involved. As Fawcett (1995: 189) observes, “the text of a translation in the case of published books is rarely all the translator’s own work” as it has to be processed by “a copy editor or other translation reviser” before the final publication. And more often than not, editors and revisers represent the institutional publisher (publication by definition is an institutional act) or the initiator, rather than being disinterested individuals. In the case of Living History, apart from being the initiator, the Taiwan publisher also plays the role of the translator (all translators of the China Times version are editors of China Times),[7] whilst the mainland publisher takes the sole responsibility for editing and revising. In this connection, we may assume that the deviations displayed respectively in the two Chinese versions from the English original, as illustrated above, can be more meaningfully studied as a result of the publisher’s positive discursive mediation than as the consequence of passive negligence of nuances in linguistic form and subject matter.

Mediation as a discursive practice has been a long-standing topic in Critical Discourse Analysis (CDA) and Text Linguistics. At the very beginning of his seminal book, Language in the News, Fowler (1991: 1-4) acknowledges language as “a highly constructive mediator” rather than an innocent verbal instrument, since language intervenes in the construction of social realities, rather than registering them neutrally, by “imposing a structure of values, social and economic in origin on whatever is represented.” Taking this as its basic argument, CDA, a “textually oriented discourse analysis” (Fairclough 2003: 2), aims to “investigate critically social inequality as it is expressed, signaled, constituted, legitimized, and so on by language use” (Wodak and Meyer 2001: 10). Such a critical view of language, deeply rooted in Foucault, also underlies the postcolonialist approaches to discourse. In postcolonialist terms, a text as discourse in communication “maps a territory and determines the features of what it charts, thus projecting a domain [or for that matter, a ‘world’] to be lived in” or “guides the way we live within this world” (Iser 2006: 172, 175). This projected domain or world is closely linked with “[one’s] model of the current communicative situation” as an element of textuality, by which mediation is to be determined as “the extent to which one feeds one’s own beliefs and goals into” such a “situation model” (Beaugrande and Dressler 1981: 163). In this light, both text production (i.e., writing) and text processing (i.e., reading), two fundamental components of translating, can be seen as operations of mediation between the existing real world one is living in and the projected discursive world to be lived in. To the existing world, mediation can in fact be a discursive practice of violence, as in Michel Foucault’s words, “We must conceive discourse as a violence that we do to things, or, at all events, as a practice we impose on them; and it is in this practice that the events of discourse find the principle of their regularity” (Foucault 1984: 127); according to Iser (2006: 173), “this regularity [of discourse] organizes what ‘things’ are meant to be.”

The presentation of the Harry Wu topic in Living History can thus be seen as creating a situation model into which the author feeds her own beliefs and goals, and projecting onto the reader’s mind “what ‘things’ are meant to be” (Iser 2006: 173). Mediation in the presentation of a situation model might be further observed at the level of text composition in terms of the interplay between two operations which are termed as situation monitoring and situation management in Text Linguistics. A text monitors when its “dominant function […] is to provide a reasonably unmediated account of the situation model,” and manages when its “dominant function is to guide the situation in a manner favourable to the text producer’s goals,” as defined by Beaugrande and Dressler (1981: 163). But the authors go on to point out that:

The borderline between monitoring and managing is extremely fuzzy and can vary according to the views of the individual participants. Indeed, people seem to prefer disguising their managings as monitorings, creating the impression that things are going the desired way in the normal course of events.

Beaugrande and Dressler 1981: 163

This line between the two is fuzzy because, whilst managing is apparently a textual operation of overt imposition, or violation, monitoring is by no means as innocuous as it seems to be. Monitoring operates by the selection, or inclusion in Fairclough’s (2003: 135-137) terms, of entities and events for attention in “a reasonably unmediated account” (Beaugrande and Dressler 1981: 163) of the real world we are supposed to be living in. But selection, or inclusion, entails exclusion: whilst some elements are being selected and included in the configuration of a situation before it is projected as a discursive world meant to be lived in, other elements are excluded from such a world. In line with Fowler (1991: 19), selection of events is “an ideological act of interpretation” by the selector, as the events selected “can be seen in a certain light of representation.” When it manifests in a discursive form of unmediated monitoring, it becomes a covert, thus more powerful, mode of managing. The selection or inclusion of the Wu event, of all events in the real world that could be of relevance, in Living History, and the configuration of the event in association with America’s attendance at the UN women’s conference held in Beijing (another instance of selection), are thus ideologically meaningful. Once selected, any overt signs of mediation, namely, those discursive deviations in the account of the event as we observed in Section 2.1 are but further indications functioning “to guide the situation in a manner favourable to the text producer’s goals” (Beaugrande and Dressler 1981: 163).

Mediation, being fundamental to discursive operations, underlies both writing and translating. But translating, as the inter-lingual re-writing of an existing text in another language for another culture, presents a more complicated picture; for in translation, with two languages coming into contact, two systems of values are brought into contest, or even conflict.

In theory, an author can presumably have a free hand in selecting any entity for presentation, in deciding upon its occurrence frequency and its discursive configuration, and in opting for a mode of textualization that either monitors or manages the situation in bringing its constituent entities or events into a coherent interplay. But such freedom is not absolute; instead, it is governed, among other factors, by the social purpose or function, or the skopos (which, incidentally, should not be confined to the study of translation), of the text. For a translation producer, the situation model has been given in the source text, and has been, in its own socio-cultural milieu, deemed appropriate, desirable or effective, as “an ideological act of interpretation” (Fowler 1991: 19) by all parties involved in its production. It is “an offer of information,” according to Nord (2001: 31-32), in the translation process, based on which the translation producer selects events/entities, configures their relations, and chooses a monitoring or managing mode to present a specific element textually. As such, any discursive (as opposed to grammatical) deviation from the source-text situation model in a translation can be construed as a sign of mediation that indicates a certain “ideological act of interpretation” (Fowler 1991: 19) similarly deemed by all parties involved in its production as appropriate, desirable or effective in the socio-cultural context perceived by these parties.

Whilst deviation at the selection stage can be evidenced in the inclusion or exclusion of events/entities in the target text, at the presentation stage the best linguistic tool to account for it, as Fowler (1991: 68-70) suggests, is Halliday’s functional model,[8] in which ideational and interpersonal metafunctions that language performs are “especially valuable” for analysis of “mediation of social relationships and values.” Among all linguistic representations of these two meta-functions, transitivity, lexis and modality are most relevant to our current study, especially to our observation of the overt managing in the following section. Transitivity includes the process described by the verb, participants in the process and circumstances associated with the process; lexis represents patterns of extralinguistic experience; whilst modality, realized by finite modal operators (e.g. must be, will be and should have) or modal adjuncts (e.g. usually, probably and always), is “an expression of the speaker’s opinion” (Halliday 1994: 89). Sifting through these linguistic features in translation selection and presentation will enable us to identify where and how discursive deviations occur.

In the case of the Harry Wu topic in Living History, the two Chinese versions presumably share the situation model in the source text as their common discursive provenance. But as we have seen in the preliminary observation in Section 2, each of them has demonstrated various deviations, which can be explained as ideology-driven mediation, the interplay between managing and monitoring, at stages of selection and presentation. This will be further expounded in the following section.

4. Operations and effects of mediation in presenting the Harry Wu topic in the Chinese versions

This section attempts to bring our theoretical contemplation of mediation onto an observation of the actual realization of mediation in the translated texts. Our data consist of all mentions relating the topical event in the three texts. For simplicity and clarity, they are put in parallel in the appendix, along with brief explanations of their contribution to the development of the topical event placed in brackets in the column of the source text. Expressions that will be discussed in the ensuing part are underlined, and each mention of the protagonist is highlighted in italics.

Note that our analysis of the Yilin version is carried out on the basis of its comparison with the English source text. This is because, as noted in Section 1, during the editing process of the Yilin version, the China Times version was checked against the English text sentence by sentence and revised accordingly. In this connection, the China Times version was taken as a point of reference, rather than a prototype for revision.

4.1. Mediation via the selection of the topical event and the configuration of the situation model at the stage of selection

Both versions maintain the primary components presented in the source text situation model, i.e., the protagonist, the US, the Chinese government and the women’s conference. But the occurrence number of the protagonist is reduced to 9 in the Yilin version, as against 21 in the English text and 20 in the China Times version.

In addition, in the China Times version, all mentions concerning the topical event are selected. Apart from this, the inclusion of “踐踏”(tramping on) in Mention 13 is striking, given its absence in the source text. The added word highlights the negative judgment on the status of the mainland human rights. The deletion of His in Mention 14 conceals the fact that the Californian Congress woman was on Wu’s side. In consequence, attacks on the US decision to participate in the conference may appear to be not only from Harry Wu but also from other parties.

The Yilin version, on the other hand, chooses to exclude six mentions (Mentions 2, 7, 8, 9, 14 and 16) completely and another seven (Mentions 1, 3, 4, 10, 12, 13 and 15) partially, which are primarily about the protagonist’s political background and the interrelation between the US and the topical event. What are included are those that serve to suggest the lawfulness of the action taken by the Chinese government as in parts of Mentions 3 and 10. Also included are those that alone, or together with the ensuing passages, function to mitigate the significance of the topical event as in Mention 11 and part of Mention 1. Additionally, Mentions 5 and 6 that help create a temporal link between the topical event and the US preparation for the UN women’s conference are fully selected. Mention 6 in the source text, along with Mentions 7, 8, and 9, plays a role in accentuating the significance of the topical event. Such a textual function is, however, less prominent in the Yilin version, since the zero translation of its co-texts has reduced its textual connection with the rest of the text.

At the same time, the Yilin version resorts to two added cohesive expressions to repair, so to speak, the textual coherence that is weakened by the exclusions and to manage the logical connection between clauses to project a situation somewhat different from that portrayed in the source text. With the exclusion of the source text concessive clause in Mention 3, the added subordinator “因此” (so) creates a causal relationship between the two remaining clauses. The protagonist’s arrest thus turns into a lawful action on the Chinese government’s part; whilst with the absence of the source text minor clause, Mention 12 loses its connection with the above sentence. As compensation, the conjunctive adverb, “这时” (at this moment), is included as both a structural and content word. As a structural word, it serves to connect the clause with the above context, maintaining the textual cohesion and making the target text less out of place or strange; as a content word, it carries some sort of implication when denoting the temporal relationship between the US decision to attend the conference and Wu’s release.

4.2. Mediation via managing at the stage of presentation

Once selected, how these events/entities are presented in the target text becomes significant. At this presentation stage, the overt signs of mediation are exemplified in discursive deviations, which guide the development of the situation in a manner different from that in the original. This sub-section is to detail these deviations in terms of transitivity, lexis, and modality. As is discussed above, our focus is on discrepancies that carry ideological implications.

Table 2

Discursive deviations at the presentation stage

Discursive deviations at the presentation stage

-> See the list of tables

As can be seen from Table 2, the distinction between the two versions is obvious. Except for one modality and one lexical deviation that are shared by both versions, most deviations at levels of transitivity and lexis are only found in the China Times version.

4.2.1. Transitivity deviation

According to Halliday (1994: 109), the concepts of process, participants and circumstances in transitivity are semantic categories which explain how phenomena of the real world are represented as linguistic structures. Hence, any deviations in these aspects demonstrate varied interpretations of the real world.

In the China Times version, four mentions are found to deviate in transitivity: two (Mentions 2 and 7) choose to insert a process that is not found in the English text, one (Mention 3) avoids the process in the source text and one (Mention 8) selects a different linguistic representation rather than the original process, thus altering textual nuances. In the source text, Mention 2 relates the identity of the protagonist in a static manner; whilst the China Times version depicts him dynamically as the patient of a material process “被打為” (be hit as, read: be labeled and prosecuted); additionally, the agent of the action is kept unsaid in the process, probably because it is inferable from the context. Such textual management strengthens the author’s (or rather, the target text producer’s) ideological stance over Wu’s case; in Mention 7 in the China Times version, the inserted mental process introduced by “讓⋯懷疑” (give reason … to doubt) specifies the supposed negative effect of the US participation in the women’s conference on the release of the protagonist; in Mention 3 in the source text, the material process thrown denotes a momentary action; however, when the version chooses to avoid the mention of the process, the temporal boundary for the protagonist’s imprisonment becomes blurred; in Mention 8, the relational process in the source text undergoes materialization in its target text counterpart. The target text material process “陷” (sink) indicates a more dynamic and interdependent relationship between the US and Wu’s case, and thus endows the topical event with a more central position.

4.2.2. Lexical deviation

Another notable device of mediation in terms of managing is the replacement of the original words/phrases with expressions carrying different connotations. In Mention 6, the use of the verbal group bethrown into limbo seems to accentuate the effect of Wu’s case in the source text. Its textual function decreases when it is replaced with “顯得浮動不安” (seems to be floating restlessly) in the China Times version; and increases when it is translated into “被抛到了脑后” (be put out of mind) in the Yilin version, although the strengthened textual function is not brought into play due to the reason we have discussed in Section 4.1.

Apart from this, in the Yilin version, the replacement of thrown with “拘押” (detained) in Mention 3, which makes the arrest sound more official and more of a lawful action; in the China Times version, five more cases of lexical deviation (Mentions 4, 7, 8, 14 and 15) are found. In Mention 4, the source text attribute widely known implies a wide publicity of Wu’s case; whilst its corresponding part, “家喻戶曉” (become a household name), plays up the influence of the protagonist’s case, giving the topical event a more prominent place in the US diplomatic policy; in Mention 7, the source text modifier confused sounds neutral, whereas the corresponding expression “錯誤” (wrong) commits the text to an explicit judgment; a similar case can be found in translating the word confront in Mention 15 as “質疑” (to question [the validity/legality, etc.]); in Mention 8, compared with its original part linked, the selection of “扣在一起” (be interlocked) in the translation tightens up the connection. In Mention 14, the source text attribute coup suggests that the US participation was a turnaround for the Chinese publications; whereas its Chinese translation “晴天霹靂” (a bolt from the blue), often collocated with unpleasant events, implies that my presence is a situation of extreme unexpectedness to the Chinese public relations.

4.2.3. Modality deviation

As a cover term, modality expresses probability and usuality, indicating the speaker’s opinion. In this sense, when the source text modal verb is replaced by a different one in the target text, the degree of the probability or usuality will be changed, which will, in turn, reflect different ideological orientation.

In Mention 1, the use of the finite modal operator might have and the modal adjunct scant in the source text, along with modal adjuncts not unusual, illustrates that the possibility for the topical event to receive attention in the American media is slight, since the arrest of a dissident is not uncommon in mainland China. Yet, their corresponding parts in the China Times version, “可能沒有” (might not have), reduce such a possibility to almost nil. Negativity as such, together with “並不稀奇” (not rare and strange), the counterpart of not unusual, accentuates the prevalence of the arrest of a dissident, and thus suggests a less favourable situation of human rights in the Chinese mainland; on the contrary, their translations in the Yilin version, “原本不会…太多” (would not have…too much) indicate that it is normal for Wu’s arrest to be left unnoticed in the American media. The importance of Wu’s case is, therefore, downplayed.

4.3. The mediated discursive worlds projected in the two translations

In the China Times version, the effects of the target text producer’s mediation in translation are three-fold:

  1. giving more prominence to the topical event and its connection with the US diplomacy (Mentions 4, 7, and 8);

  2. foregrounding the status of human rights in China by adding a derogative expression (Mention 13), changing modal operators (Mention 1) and inserting a material process that foregrounds the topic as a political prisoner (Mention 2);

  3. reinforcing the author’s negative attitude towards the Chinese government by making lexical changes (Mentions 7 and 14).

As a result, the translation presents a discursive world with a more conspicuous ideological commitment. By comparison, the publisher’s mediation in the Yilin version is mainly seen in

  1. deemphasizing the significance of the topical event by decreasing its occurrence frequency, modifying the author’s attitude over Wu’s case (Mention 1) and excluding the relevant knowledge of Wu’s political background and his connection with the US diplomacy;

  2. emphasizing the authority of the Chinese government over Wu’s case by deselecting derogative expressions about the Chinese government and making a lexical change (Mentions 3 and 10);

  3. creating a seemingly unmediated account of the topical event by adding two cohesive expressions (Mentions 3 and 12).

The mainland readers are thus given a discursive world where Wu’s imprisonment in the Chinese mainland is legitimate, and where the topical event does not weigh much in US-China diplomacy.

5. Impact factors of mediation in the translation process

From the social functional point of view, the translation process is largely constrained by the target text recipients and their situation and cultural background. In other words, the recipients’ needs as perceived by translation producers play a crucial role in deciding how and to what extent the translation should be mediated. From this postulate, a further refinement to our understanding of the position of the target text receiver is needed for a more comprehensive account of the mediation action.

According to Müller (2004: 22), there are two groups of recipients that a writer (also a translator as rewriter) in a sociopolitical context should have in mind: the‘professional’ readers involved in the censoring process and the intended reading public. The professional readers refer to censors with “formal qualifications,” “work experience” or expertise in their field (Müller 2004: 22-23). They are stereotyped by the government/government agencies and are endowed with privileges in accessing and inspecting the translation before/after publication and in imposing and enforcing constraints on the reading public’s access to the book by facilitating or hindering its publication. Their censorship[9] is practised in accordance with the ideology and the power order that the authority wants to maintain, or in other words, the “wide national patterns of taste” (Billiani 2007: 15) embodied in various laws and regularities. Any contents contravening these laws should be “explicitly prohibited” (Bourdieu 1991: 138). Censorship, in this sense, becomes “an authoritarian control over what reaches the public sphere” (Müller 2004: 12) and a type of power for “establishing a given cultural authority […] by exercising the power of punishment and the right of surveillance” (Billiani 2007: 15). Compared with the professional readers, the reading public are relatively passive in the sense that they can only read those censored translations and have, by and large, no access to the censored message. As the dominant market/cultural force, they are more concerned with the readability of the translation.

In the Chinese mainland, the professional readers are represented by a three-level censorship network including the publisher, provincial Bureau of Press and Publishing Bureau, and the General Administration of Press and Publishing, operating under the Regulations on the Administration of Publishing[10] and Notice Regarding Further Strengthening the Administration of Censoring Books. The network is established to ensure that no publications should contain contents against the basic principles confirmed in the Constitution or harmful to security or interests of the country. An interesting example of the way the censorship policy works in translation publishing is provided in Chang (1998). Chang’s Chinese translation of Yes Prime Minister was rejected by about ten mainland publishers in the year of 1990. This is because his translation, as Chang (1998: 255) acknowledges, functions as “a form of ideological resistance” and a “political satire,” touching upon the “dangerous and often forbidden ground” (Chang 1998: 251), such as sex, anti-communist and anti-Chinese sentiments, in the field of literary translation in the Chinese mainland. In Taiwan, on the other hand, the role of the professional readers is played by Government Information Office, Republic of China as well as the police. Publications are rated in accordance with classification rules[11] and shall not be against laws. In this connection, the published translation can be viewed as, in Bourdieu’s (1991: 137) words, a “compromise” between the author’s “expressive interest” and the “censorship.”

Different needs of the target text readers pose challenges to the translation producer, who must, on the one hand, keep the translation politically and ideologically consistent with the taste of the professional readers, and on the other, make the translation interesting enough to attract the reading public. These challenges become more serious when there is a political, ideological or cultural conflict between the social and cultural societies the source and the target texts respectively belong to, and when the conflict results in a number of textual inconsistencies with conventional linguistic norms of the target society. It is at this juncture that the “internalized […] forms of perception and expression” (Bourdieu 1991: 138) of the translation producer, the publisher in the case of Living History, start to work. To put it simply, in order to comply with the professional readers’ desires, the translation producer practises self-censorship by mediating the translation process, i.e., by including/excluding/revising expressions that touch upon the forbidden ground, prior to the official censorship. It is through such coordination between the professional readers and the translation producer that a broad ideological and cultural consensus can be molded in the target society.

6. Discussion

Before examining mediation in the translation process, let us take another look at the source text, which was once viewed as “heavily politically invested” by more than half of its American readers.[12] The inclusion of Harry Wu as a topical subject itself can be taken as a classic case laden with ideological and political information. The presentation of Wu’s story, in particular, demonstrates the author’s strong concern with the protagonist and her negative stance over China’s human rights situation. Nevertheless, as a bestseller, the English text was predicted to have great potential to satisfy the Chinese reading public’s curiosity about the inside stories of the Clintons and so its translation would have brought about substantial profits.[13] Yet, to get published, the target text should, first of all, satisfy the professional readers’ curiosity, or address their possible concerns.

In general, the ideological and political information carried in the source text conforms to the Taiwan censorial system, and Taiwan society that shares similar political and ideological interests with American society. This can be verified by the fact that most contents about the topic are included in the China Times version. The constant tendency of elevating the significance of the topical event and strengthening the author’s negative attitude toward the Chinese mainland revealed in mediation at both translation stages also evidences ideological opposition between the two Chinese societies. The publication of the China Times version proved that these reinforced ideological implications served the Taiwan professional readers. Also, they did not provoke hostile reactions from the author and the American publisher.

By contrast, in the Chinese mainland, the venue of the topical event depicted in the text, the situation is complicated, and all the more complicated because the Harry Wu topic is politically sensitive and has been banned from the media. In other words, the author’s expressive interest in the protagonist does not fit in with the mainland censorial system. Hence, in theory, the publisher should mediate the editing process in order that the mediated version can avoid the “dangerous and often forbidden ground” (Chang 1998: 251) and maintain an ideological coherence with the mainland professional readers. In practice, instead of wiping out all accounts of the topical event, Yilin chose to include the Harry Wu topic, but “make some revisions and a few technical changes whilst keeping the integrity of the book.”

At the selection stage, the exclusion of more than half mentions that are primarily about Wu’s political experiences and the significance of his case in US-China diplomacy suggests the impact of the mainland censorial system on the translation producer. A knock-on effect is the inclusion of two cohesive expressions in Mentions 3 and 12, which, on the other hand, bespeaks the translation producers’ efforts to cater to the intended reading public. Another salient example is the modality deviation at the presentation stage, which registers a different tone from the source text in presenting the topical event. After being mediated as such, the ideological stance of the author over the Chinese government and the topical event is somewhat neutralized. The compromise achieved between the author’s expressive interest and the censorship was approved by the mainland professional readers and finally led to the publication of the translation, an operation that was, however, viewed by the author and the American publisher with bitter resentment.

Apart from this, it is worth noting that some deviations may seem to be inconsistent with the overall ideological tendency as seen in the related operations of mediation. Two explanations suggest themselves in this connection: firstly, deviation may not always be prompted by a conscious, ideologically coherent intention of mediation; secondly, deviation may be caused by different interpretations of the source text. The translation of the phrase be thrown into limbo in Mention 6 is a case in point. In the China Times version, it is given as “顯得浮動不安” (seems to be floating restlessly, a state of uncertainty between two courses of action) obviously following the extended usage of limbo in its first sense, i.e., “a region supposed to exist on the border of Hell as the abode of the just who died before Christ’s coming, and of unbaptized infants”;[14] whilst in the Yilin version, contrary to its overall ideological drift, the impact of Harry Wu’s case sounds graver as if it were threatening to push those sessions with representatives from the United Nations and other countries out of the minds of those concerned – “被抛到了脑后” (be put out of mind), a translation more in keeping with the figurative meaning of limbo, that is, “any unfavourable place or condition, likened to Limbo; esp. a condition of neglect or oblivion to which persons or things are consigned when regarded as outworn, useless, or absurd.”

Since mediation is found in both translations and the ideological implications of the source text have been modified accordingly, albeit to differing degrees, the conventionally held concept of faithfulness of the target text to its source text cannot be a real issue in this case. As our investigation suggests, what makes the difference between the two Chinese versions, instead, is in whose interest discursive mediation has been performed. And it is the divided ideological and political “loyalty” (Nord 2001: 123) displayed in the translation producer’s mediation and the discursive worlds the resultant texts project that give rise to the different responses from the American author and publisher to the two Chinese versions. Whilst their acceptance of the China Times version indicates their identification with the ideological and political interests of Taiwan society, their criticism of the Yilin version has the mainland professional readers, the censorial system and the dominant ideology and political power they represent, as its true, or ultimate, target.

7. Concluding remarks

The present study investigates the textual operation of mediation in translation and its relationship with the socio-cultural context where the target text is supposed to be consumed. Inspirations from Text Linguistics and Halliday’s functional model enable us to examine how mediation is realized by means of inclusion and exclusion of information at the selection stage and, at the presentation stage, by means of discursive deviations at levels of transitivity, lexis, and modality. The detailed textual analysis helps delineate a scenario of how the translation is oriented, in a manner favorable to the target text producer’s goals, to project a discursive world, differing from that presented in the source text. Concepts from skopos theory and CDA provide insights in discussing the effect of mediation in relation to the target text recipients, i.e., the professional readers and the intended reading public, and their socio-cultural backgrounds. Our discussion shows that for the target text producers whose translation is under constraints of readers’ needs, mediation is a powerful means to make the translation commensurate with the ideological interests of the professional readers, i.e., the representatives of the dominant social power. Mediation action itself embodies the power control of the professional readers over the reading public in the target society. Our argument in this article is that an investigation of translation from the perspective of mediation helps reveal the power struggle underlying the translation process. The detailed description of the textual realization of mediation, in particular, helps create an awareness of various ways the target text producers mediate the translation process step by step to achieve the compromise between the author’s interests and the professional readers’ political concerns so as to get their translation published.

The present study represents our initial effort to analyze the textual actualization of mediation in translation and to discuss how mediation functions to interact with socio-cultural institutions that may have an impact on the final product of the translation by drawing and integrating concepts from various disciplines. Much more effort is needed to depict the analytical model in greater detail so as to facilitate exploration of translation from such a perspective.