Modeling the Shifting Face of the Discourse Mediator

Abstract

This paper proposes a discourse model of translation activity with the mediator (translator/interpreter) as its agent. It captures both oral and written translation, as well as professional and lay translation, and provides a universal explanatory mechanism for shifts in the theories and benchmarks of translation practices over time. The model rests upon the basic claim of the underlying cognitive unity of oral and written discourse, whether monolingual or multilingual and whether mediated or unmediated. The differences between these discourse types, as well as differences in the product of translation activities, follow from differences in the discourse situations, which in turn include the cognitive features (including linguistic and cultural competence, norms, intentions, and identities) of given discourse participants.

Introduction

The impetus for this study comes from considering the implications of the move of translation theory away from primarily linguistically oriented issues of ‘equivalence’ to primarily literature-oriented issues of post-structuralist intellectual discourse. Between the 1940s, when the first theoretical works on translation began to appear (Nida 1947; Brower 1959) and the 1980s, when this move took place, the dominant concern of translation scholars was ‘equivalence.’ Their focus was on the text—specifically, on the linguistic and philosophical possibilities and limitations of achieving ‘interlingual equivalence.’ During this period, some scholars enthusiastically embraced new developments in theoretical linguistics, as they sought ways of providing a solid basis for encoding meaning in languages with differing grammatical structures (Nida 1964; Catford 1965), while others focused on empirical questions of structural differences affecting translation (Vinay and Darbelnet 1977).

When the post-structuralist explosion of theory occurred in literary studies, translation theory in the West followed the lead of literary criticism and turned away from linguistics. The focus shifted from the text to its context (a “cultural turn”) and then to the target text in the receiving society.¹ This was accompanied by a rapid expansion into a number of novel topics, from the exploration of past and present translation practices, to the examination of the political, economic, and cultural contexts of the source and receiving languages involved in a given translation. In a striking departure from the previous approaches that set ‘equivalence’ as their benchmark, feminist translation theorists not only advocated that feminist translators should make their presence in the translated text “visible” by flaunting their presence and agency in it (Godard 1990, 89–91), but also challenged the value of fidelity in translation (Simon 1996, 12–14) and called for retranslation of all works of literature from a feminist perspective.² In a parallel move, queer translation theorists argued that “faithfulness can no longer be regarded as an absolute concept” (Mira 1999, 109) and that homosexuality and gay identity in translations must be brought out of the closet, making explicit any allusions found in the original texts (Keenanghan 1998).

This move away from ‘equivalence’ did not result in the elimination of linguistically oriented studies of translation, which continued to occupy a significant number of researchers. Among more recent developments, the growing accessibility of electronic corpora has notably provided new tools for textual analysis and a renewed impetus for examining the linguistic differences between the originals and translations (Anderman and Rogers 2008). The resulting state of translation studies today is multifarious and eclectic, and this eclecticism itself, to my mind, is a sign of a field in search of an identity. This paper is thus a contribution towards a metatheory of translation that proposes a discourse model of all translation activity, written and oral, as well as a way of modeling various shifts in translation theories.

The Relationship Between Interpreting and Translation

Before we examine the phenomenon of translation, the relationship between interpreting and translation must be addressed. English uses two distinct terms for these activities, but this usage is far from universal: cf. perevodit’, the Russian verb for to translate/interpret, or tafsiri, the Swahili noun for translation/interpretation, or Japanese yaku, a noun that also stands for either concept.³ My first claim is that translation and interpreting are essentially a single communicative phenomenon. Interpreting is to translation as speech is to writing. The difference is in the medium: one is spoken, the other is written; but both are manifestations of linguistic communication based on a single linguistic system of a given language.

That said, the distinction between the two is not insignificant, and most of the obvious features that distinguish oral speech from writing are identical to those that distinguish interpreting from translation. The following familiar contrasts hold, prototypically, between the spoken and written modes; and being prototypical, they in practice allow, of course, some deviations ⁴:

Most of the characteristics in this table are self-explanatory. The only category requiring a special comment is the last one, “receiver”: in OD, the addressee (hearer) As, posited by the speaker in face-to-face communication, and A, the one who actually receives the message, are identical; in contrast, the addressee (reader) As, presupposed by the writer in WD, and the actual reader A are assumed here not to be identical. This difference is a direct corollary of the fact that in face-to-face communication the participants by definition share the time and space of the transaction, whereas in written communication, when the time and space are not identical, the writer (for instance, Leo Tolstoy) has no knowledge of who the reader may be (for instance, I).

These characteristics of the receiver in Table 1 are only working approximations, because the actual situation is more complex. In real discourse situations, before undertaking an utterance, the S (speaker) posits the addressee’s state of mind. This cognitive act is called assessment (Yokoyama 1986, 44–52). Since the true state of the addressee’s mind is not accessible to the speaker, S’s assessment of A is always the best possible approximation of reality. Thus A, the real addressee, and As, the addressee posited by S, are strictly speaking, not identical. In face-to-face communication, thanks to multiple visual, contextual, and cultural clues available to S while assessing A’s cognitive state, as well as because of the possibility that S’s erroneous assessments may be corrected on the spot by A, A and As are as close as they can be. This maximal proximity of A and As in face-to-face oral discourse is treated in Table 1 as the ‘equivalence’ A = As, with the understanding that it is only a ‘near-equivalence.’ In written communication, due to the separation in time and space, the gap between the reader as posited by the writer and the real reader is substantially greater than in face-to-face communication; hence in the written mode I have posited A ≠ As. This is also, however, an oversimplification. The gap between A and As in written discourse actually varies: it is minimal, for example, in secure intimate personal correspondence, and maximal when the reading is done by unspecified multitudes at an unspecified place and time (as in political manifestos, fiction, or translated literature).⁵

Table 1 Characteristics of oral versus written modes of communication

While speech and interpreting share the prototypical features of the oral mode, ordinary oral discourse does differ from discourse that involves interpreting. I will call the former monolingual⁶ oral discourse (MOD) and the latter bilingual⁷ oral discourse (BOD)⁸; the features of each are proposed in Table 2 (T stands for the translator):

Table 2 Comparison of ordinary oral discourse and discourse involving oral translation

A few comments are in order about the categories in Table 2. The code is in principle not limited to the linguistic one but includes paralinguistic and cultural codes as well. In MOD all three codes are largely shared, but in BOD linguistic codes are not shared by S and A even though the other two codes may be at least partially shared (since S and A see each other’s body language and hear each other’s intonation, and since they may belong to the same larger cultural realm). The “sharing” in Table 2 refers to the linguistic code. The message being communicated belongs to S in both MOD and BOD. This means that in translation, as in ordinary dialog, it is S whose communicative intention determines the content of his/her utterance, regardless of what A’s expectations or preferences may be and regardless of whether S’s message needs to be translated into another language in order to reach the intended A. The agency of the actual production of the message intended by S, on the other hand, differs sharply between MOD and BOD; in the former, it is S who produces AND delivers the intended utterance, but in BOD, S performs only the first half of the job, as it is T on whom the task of delivering the message to A falls. Consequently, in MOD, there is only one receiver, A, for whom the message is intended and by whom it is received. In BOD, on the other hand, the reception of S’s intended message by A is delayed by T’s intercession, who first receives it as A’s proxy and then conveys it to A in an appropriate code. The first actual receiver T is then only partially presupposed by S (to the extent that S at a minimum posits T’s knowledge of the codes involved) but is not the primary receiver intended by S. Strictly speaking, just as in the case of S’s receiver, T, too, posits an addressee At, who may not have the relevant features A has. Given the position adopted here that in oral discourse the distinction between any posited A and the real A is minimal, the discrepancy between At and A will also be treated as virtually zero in BOD.

The crucial differences between MOD and BOD are thus, not surprisingly, (1) in the presence/absence of a shared code between S and A, and (2) in the presence/absence of the mediator T. Because of these two differences, in MOD the discourse between the interlocutors S and A is direct and a cognitive transaction is concluded without mediation by a third party (S←→A); on the other hand, in BOD the discourse between S and A is mediated, and the cognitive transaction is by definition concluded with the mediation of the third party T (S←→T←→A).

Just as MOD departs from BOD, ordinary monolingual written communication (MWD) also differs from written translation (BWD) in well-defined ways that can be extrapolated from Table 2. The main difference between prototypical oral and written discourse (OD and WD) consists in the removal of the addressee from the time and space of the discourse event. All of the other differences in Table 1 follow from this. The removal in time allows for planning and editing, delays the feedback, and makes the message accessible to an unintended A. Physical distances necessitate the written medium, the production of which calls for literacy, and messages intended for removed addressees favor preservation, which leads to formality. The main difference between prototypical monolingual and bilingual discourse (in Table 2), on the other hand, consists in the crucial involvement of T caused by the lack of a shared linguistic code between the producer of the intended message S and its ultimate receiver A. This is true for both oral and written modes of MD and BD. Thus BWD and MWD depart from each other exactly in the same way that BOD and MOD do: in the absence of a shared code and in the intercession by T. Conversely, the difference between BOD and BWD parallels that between MOD and MWD. This is summarized as follows in Table 3:

Table 3 Discourse modes differentiated by the factors of time, space, and code

The logical continuum instantiated by these features confirms the shared nature of translation and interpreting.

Mediated Discourse

Intuition tells us that a translator or interpreter is positioned in the middle between two parties⁹ who cannot otherwise communicate successfully because of the absence of a common linguistic code. T is thus the third person (i.e. cognitive entity) involved in knowledge transfer between S and A, the two primary interlocutors. Mediators are, of course, not limited to translators or interpreters. Commercial transactions routinely involve middlemen and adversarial situations call for arbitrators, just as many other social and political situations are facilitated by various kinds of intermediaries. Gatekeepers, messengers, spokespersons, and other people¹⁰ playing an intermediary role all deliver another party’s message to the addressee, who is ostensibly communicating with the messenger while accepting into his/her mind the information that originated in that other party to whom the communicative intention belongs. These are all relatively easily discernible “macro” mediator roles, for which we often have distinct lexical labels of the sort just mentioned.

There also exists, however, a much more fleeting and less obvious “micro” mediation, the kind that occurs when a mediator’s intervention is more limited in terms of the role of the mediator and the bulk of the information transmitted by him/her. One of the most common instances of such “micro” mediation occurs in ordinary monolingual discourse, and this kind of discourse mediation, to the best of my knowledge, has thus far received little attention. In the case where knowledge might not be shared between the speaker and all his/her interlocutors in a party of three or more people, the one participant that does share this knowledge with the speaker and is aware that another participant does not share it, may jump in and supply that knowledge to that participant in order to help render the speaker’s transaction successful. For example, in a monolingual discourse situation, when S mentions Stalin and one of the participants (A1) gives S a blank look, another participant (A2) may provide a helpful “footnote” along the lines of “Stalin was a Russian dictator in the last century.” Technically, this “metinformational utterance”¹¹ supplies the referential knowledge of Stalin that S presupposes and A1 lacks.¹² A2 has then mediated the knowledge transaction between S and A1. Similarly, if S, a Japanese woman, disapprovingly says in perfect English that ‘so-and-so served sliced fresh peaches without peeling them,’ and A1, an American woman, looks puzzled; in that case, A2 may supply the associated propositional knowledge¹³ that A1 evidently lacks, i.e., “Japanese prefer eating peaches without their skins on.” In all such cases A2 functions as a “micro” mediator between S and A1.¹⁴

The absence of a shared linguistic code between the interlocutors S and A1 is only one kind of missing knowledge that the mediator A2 may supply. Again, this absence may be total, or it may be partial; it may even be limited to a single word. As for instance in the case of a discourse involving native and near-native speakers, when the near-native speaker A1 might have missed just one word in S’s utterance and A2 jumped in to provide a translation. This, in fact, happens even in monolingual discourse involving only native speakers. When the absence of a shared linguistic code between the interlocutors S and A1 is extensive (and therefore severely debilitating for their communication), the mediation addressing this absence is usually singled out as a special concept: that of translation/interpreting. The mediator who takes on this activity as his/her role is then called a translator or an interpreter, and such a mediator, his/her role being clearly discernible, can now be considered to be a “macro” mediator. Interestingly, in some languages the agents of all discourse mediation, both “macro” and “micro,” are referred to by one and the same word; cf., for example, the Turkish word dilmaç, which means both ‘interpreter’ and ‘clarifier, explainer, commentator’—that is, the mediator who supplies the missing knowledge of the code as well as any other knowledge needed for understanding the message.¹⁵

The third cognitive entity in mediated discourse thus facilitates the transfer of a message. Significantly, the initiative of this message belongs to the speaker and the message is usually intended for the addressee and not for the mediator.

Discourse Mediators and Their Faces

A pivotal question for our analysis of translation concerns the neutrality of discourse mediators. Mediators are often thought of as neutral, faceless, and lacking influence from their own values and interests. Translators in particular have been conceived as transparent colorless glass, on one side of which information enters in one language and emerges on the other side in the other language. Such translators do not refract the incoming ray or color it in any way, but just pass it through without altering its nature. Gulliver (1979, 217) challenged the widespread notion of mediators’ neutrality, calling it a myth born in Western societies. If he is right, no faceless translators should exist. Yet the notion of an impartial faceless discourse mediator persists. This putative impartial discourse mediator is, moreover, expected to command perfect competence in both languages and cultures and holds ‘interlingual equivalence’ as his/her benchmark.

There is indeed some evidence of translators’ facelessness in our cultural history. The very custom of identifying the translator of a written text, for example, is relatively new. Western history has preserved the names of only a few translators, those of culturally important texts, and they all happen to be more famous for other activities—Cicero, St. Jerome, Luther, and Tyndale. Translators of culturally less important texts, such as royal epistles, were non-persons who remained faceless, fading before the monarchic authority of the originals. Equally anonymous are the translators of government decrees, administrative and legal documents, or commercial contracts today. These professionals have effaced themselves as persons while continuing to strive for ‘functional equivalence.’ The tradition of the self-imposed benchmark of “equivalence” has continued as the ideal stance of translators well into the twentieth century. “Faithful” translators have deferred to the text’s or its original creator’s authority, and “faithful” professional interpreters today embrace the professional ethics standards stressed during their training, with their requirements of neutrality and impartiality (Baaring 1992, 60 et seq.).

However, the impartiality of the “faceless” translator is seriously undermined if we examine non-professional oral translation. Echoing Gulliver’s claims about the myth of the impartial mediator, Valero-Garcés (2007) challenges “the myth of the invisible interpreter and translator” providing many examples of partiality. None of the expectations of facelessness, neutrality, or, for that matter, perfect bilingual competence can be maintained with respect to those engaged in language brokering, volunteer community interpreting, and casual personal discourse mediating by family members and friends. Work by “partial” mediators occurs daily in multilingual communities today, most of it in face-to-face discourse. Many bilingual discourse mediators are untrained and only partially competent in at least one of the languages they mediate (Tse 1996); many of them are children, heritage speakers of their family’s languages, mediating for their monolingual family members.¹⁶ Considerable research has appeared in the last two decades on these “community/liaison interpreters,” “language brokers,” and other partial discourse mediators (Eksner and Orellana 2005; Morales and Hanson 2005; García-Sánchez 2010), providing ample evidence of Ts motivated by their own sympathies, views, or agenda.

Let us consider some examples of partiality in discourse mediation done by an interpreter in Sweden taken from Wadensjö (1998):

(1) S: […] zdes’ medicina vse-taki ne tak ¹⁷ razvita, kak u nas v Sovetskom Sojuze.

‘medicine here is not as advanced as in the Soviet Union, after all’

T: […] medicinsk utveckling är inte på samma nivå som i Sovjet

‘medical development isn’t at the same level as in the USSR’ (157–8)

(2) S: jag vet inte om du förstår skillnaden mellan att vara- ha en viss nationalitet och att va medborgare?

‘I don’t know if you understand the difference between being- having a certain nationality and being a citizen?’

T: odno delo sčitat’ … po nacional’nostjam a drugoe delo sčitat’ sebja graždaninom kakogo-libo gosudarstva

‘one thing is to consider in terms of nationalities and another thing is to consider oneself a citizen of some state’ (111)

In (1) and (2), a Russian-speaking immigrant obstetrician applying for a residence permit in Sweden is being interviewed by an immigration officer. In (1), T introduces ambiguity into the immigrant’s comparison between the levels of Soviet and Swedish medicine, changing the original statement asserting the relative excellence of Soviet medicine into one that states only that there is a difference between the levels in the two countries. In doing so, T removes a potentially problematic assertion by the applicant. In (2), T removes the potentially offensive implication in the officer’s wording that suggests ignorance on the part of the applicant. In both cases, T evidently strives for the role of the applicant’s facilitator, if not protector, rather than being a “machine” that merely produces ‘interlingual equivalents.’

In another example, a different interpreter omits ‘BC’ when translating an Armenian refugee’s statement about Armenia’s Christianization:

(3) S: [v Armeniju] vveli xristianstvo uže v 301 g. do našej èry.

‘(to Armenia) Christianity was brought already in 301 BC.’¹⁸ (203)

In this case, the omission was motivated by T’s realization of the ridiculousness of this statement and was aimed at preventing a comic effect that would have been embarrassing to the speaker and distracting for all those involved. This example, too, shows that this T did not strive to merely produce ‘interlingual equivalents.’ The subtle and, at first glance, inconsequential departures from the originals in (1–3) affect the discourse in ways that significantly alter A’s perceptions of S, potentially producing significant differences for the parties involved.

The “partial” discourse mediator has always existed and will always remain with us. The first such discourse mediator was born when the first two human dialects diverged to become sufficiently mutually unintelligible to benefit from a bi-dialectal mediator. Such a T predates professional interpreting and literary translation, and in fact s/he predates literature itself. I suggest that to be “with a face” and “partial” is the quintessential condition of all Ts.

Translators “with a Face” in BWD

I now return to the question posed at the beginning of this paper: How can we account for the post-structuralist shift in translation theory? I will argue in this section that the shift is only apparent, and that translators of the post-structuralist era do not constitute a new or extraordinary phenomenon. Taking a bird’s-eye view on the cultural progression in Europe over the past millennium, I will claim that these “opinionated” translators have predecessors in European history. I will examine the motivations of these translators (to the extent that this is possible) in two areas: religious or ideological convictions (today we might call them “identities”) and social or cultural standards.

Translators with Strong Ideologies

Well-documented challenges to the authorities’ attempts, to regulate translation, began in Europe in the Late Middle Ages. Along with the rise of vernacular literatures throughout Europe, the movement to translate the Scriptures into local vernaculars began, in which the translators followed their convictions, ignoring Rome’s proscription of unauthorized Bible translations. John Wycliffe of England translated the New Testament from Latin into English vernacular in the late fourteenth century, influencing Jan Hus, a Bohemian humanist, to translate sections of the scriptures and liturgical texts into Czech. While Wycliffe died a natural death in 1384, Hus was less fortunate and met his end at the stake in 1415.¹⁹

As the Reformation spread through Europe, it provided a continuing impetus for translating the scriptures into local vernaculars. Martin Luther’s translation of the Bible, liturgical texts, and hymns into German by 1534 was made possible by the intervention of Frederick the Wise, who protected Luther from persecution, sheltering him in his castle and providing him with the time and security to engage in his translation work.²⁰ Luther used his freedom from Rome’s control of his translation activity, to promote his theological position, adding the word “alone” into the phrase faith justifies (Romans 3:28). This suited his theology of salvation, an important point on which he diverged from Rome (Mullet 2004, 148–150). Around the same time, the less fortunate but no less “opinionated” William Tyndale was burned as an unauthorized heretical translator of the Bible into English.

Translating the classics into European vernaculars could be just as risky, provided these translations had the potential of introducing objectionable ideas. Thus, Etienne Dolet was burned at the stake around the same time for the heresy of adding, in a translation of Plato into French, the adverbial intensifier rien du tout ‘not at all’ to an already negative sentence about there being no life after death (Christie 1899).

These are some well-known examples of “opinionated” translators who followed their convictions in the face of life-threatening control by the political (or ecclesiastic) powers in Europe in the late Middle Ages and early modern period. A different sort of ideology motivated Alexander Ross to translate the Qur’an into English (from French, since he did not know Arabic) in 1649. As he stated in the foreword to the translation, Ross took the initiative to translate the Qur’an, which he disdained, because it could underscore “the health of Christianity.”²¹ Although far from risking his life for this translation (and evidently undaunted by the fact that he did not know Arabic and that his knowledge of French was limited), as one of the “opinionated” translators of his time Ross responded to the cultural and political curiosity (if not hostility) of the young British Empire towards Islam; especially, at a time when the Ottomans were at the height of their power and influence on the European continent and in the Mediterranean basin.

Translators with Convictions About Social Norms

In the case of Dolet, one might argue that the fatal decision he made was more of a linguistic than theological nature. A translator is faced with linguistic decisions in every sentence or utterance s/he translates and is naturally aware of this. Luther was not only an “opinionated” theologian but also the author of Sendbrief von Dolmetschen, a treatise on translation.²² With his interest in theoretical questions regarding translation and given his respect towards the vernacular, Luther continued the line of prominent translators of antiquity like the much earlier Bible translator St. Jerome who, a millennium earlier, made the difficult decision (for which he was criticized) to perform the Bible-translation task entrusted to him by the pope by translating it into Latin vernacular; this he did despite the fact that he had a full command of Ciceronian Classical Latin. Linguistic decisions of this sort were based on the translators’ views of language—not only on its semantics and structure, but also its registers and its diachronic development. Moreover, translators obviously differed among themselves in these views. Let us consider here two such opposing views leading to different decisions, and due, I suggest, to shifting social norms over the period of the last century.

Ackroyd, who in 2009 “retold” the Canterbury Tales in an intralingual translation, followed his views on translation and modernization when he took an unapologetic step towards an explicitly lower register than those into which his predecessors had translated the classic. He sacrificed fidelity in favor of modernization, rendering Chaucer’s […] for what profit (purpose) was a wight (body) y-wroght (made)?[…] they wer (were) nat (not) maad (made) for noght (nothing) (Hopper 1970, 391) as Cunts are not made for nothing, are they? (Ackroyd 2009, 149), and Love me at-ones (instantly), or I wol (will) dyen (die) (Hopper 1970, 209) as Fuck me or I’m finished (Ackroyd 2009, 84). The social norms of our times allow, or perhaps even encourage, such decisions in order for the book to appeal (and be marketed) to the target readership.

Under pressure from the power of taste and propriety, decisions in the opposite direction have been made as well. Here are a few examples of such decisions made in the Victorian age by well-known Russian translators of Shakespeare and published in 1902 by the exclusive and authoritative Brokgaus and Efron publishing house under the general editorship of S.A. Vengerov:

(4) Original::

a pissing while

Translation::

skol’ko nužno, čtoby vysmorkat’sja

‘time needed to blow nose’

(Two gentlemen of Verona, Act IV, sc. 4; Vs. Miller)

(5) Original::

When I have fought with the men, I will be cruel with the maids – I will

cut off their heads. […] Ay, the heads of the maids, or their maidenheads.

Translation::

Lupi mužčin, da pronimaj i bab! Im ne ujti ot menja celymi!

‘Smack the men and get the women, too! They won’t get away from me

unharmed!’

(Romeo and Juliet, Act I, sc. 1; Ap. Grigor’ev)

(6) Original::

Thou Protector of this damned Strumpet

Translation::

Ty pokrovitel’ merzkoj ved’my

‘you, protector of the disgusting witch’

(Richard III, Act III, sc. 4; A.V. Družinin)

The endnote for (4) provides the original “crude” English idiomatic expression a pissing while (‘a short time’), which was replaced by the translator as ‘time needed to blow one’s nose’; the note suggests that the motivation for this replacement was the crudity of the original (549). The series of puns built around the theme of rape in (5) is obscured and considerably shortened in the translation. The endnote explains this, and supplies a full literal translation of the omitted “crude” lines (554). Regarding (6), the translator writes that the original had not witch, “but a different, salacious word.” For those who wonder, he adds: “Lovers of precision may want to substitute it, because the verse will remain as felicitous” (561).²³ In some cases, improper material was simply left untranslated, as is explicitly noted at the end of the translation of Love’s Labor’s Lost: “Some of the too crude and indecent witticisms have been omitted” (553).

The pressure on the translators to conform to the culturally acceptable norms of the polite society these gorgeous Russian editions of Shakespeare targeted was particularly strong, evidently resulting in self-censorship of the kind we see in these examples. Although they are ostensibly opposed to one another, what the decisions by these Russian translators and those by Ackroyd have in common is the conviction underlying them regarding the need to adhere to the social norms of their assumed readership.

The Ascending Agency of T with a Face

By the second half of the twentieth century, various expressions of identity grew more pronounced, producing new kinds of “opinionated” translators. At the time of the Berkeley Renaissance in the 1950s, when Jack Spicer translated García Lorca’s homosexual images more frankly than they were worded in the originals, it was still an act of uncommon courage comparable to that of the post-Renaissance theologians; but by the time the feminist movement was in full swing and Mary Phil Korsak had begun to retranslate the Bible incorporating the new position of women in modern societies,²⁴ gender and sexuality had become one of the central topics of postmodern humanistic discourse. The drive to convey what these new “opinionated” translators thought was or should be the message of the originals resulted in translation theories that called for conveying these translators’ positions, rather than the ostensible “surface” meaning of the original wording. The translator felt compelled not only to show his or her face but also to assert his/her identity and values through his/her work.

There are two crucial points to note here. The first point is that although the motivations of all the “opinionated” translators considered above may have arisen from convictions of varying nature—from religion to identity, to politics, to linguistics—and although their convictions did not expose them to physical danger to the same extent, the kind of force these motivations exerted on their choice of what and how they translated was in essence the same. The second and even more important point that needs to be made here is that the “opinionated-ness” of these translators, cognitively parallels the “partiality” of the language brokers and community interpreters considered in section “Discourse Mediators and their Faces”. Whether it be clerics convinced of their understanding of the Scripture, or queer activists striving to bring homosexuality out of the closet, or children translating for their immigrant parents in a multicultural community, all of these discourse mediators are crucial agents who not only make communicative transactions possible but who at the same time are persons with their own views, allegiances, and values. Based on these values, they choose their stance vis-à-vis the transaction—that is, they decide what and how to translate, and even whether to assert their “faces” or not. It is this variety of T’s stances that a metatheory of translation must capture, and this is what I will attempt to do in the next section.

Modeling Translation

I now turn to a model that captures, I suggest, the variation in T’s “faced-ness” in different societies at different times in both BOD and BWD. The proposed overarching translation theory captures translation activity using the Transactional Discourse Model of Yokoyama (1986).

The Transactional Discourse Model is a dynamic face-to-face dyadic model that tracks the process of knowledge transfer as it is manifested in the concrete linguistic choices the speaker makes, while assessing the addressee and with the addressee’s active engagement. This model is general, and as such it has been shown to account for various knowledge transactions involving the interlocutors’ identities and opinions.²⁵ In this model, the interlocutors’ sets of current concern (their attention sets) contain the awareness of mutual availability for contact at a given time and place.²⁶ This shared awareness consists of the referential knowledge of I, you, here and now (i.e. {I}, {you}, {here}, and {now}, abbreviated as {DEIXIS}). Each item of referential knowledge is accompanied by a set of associated knowledge items in the form of propositions. Consider some possible propositions associated with {I} and {you} as seen by S in Table 4:

Table 4 Examples of propositional knowledge associated with {I} and {you} in MD

The propositions activated in any given discourse transaction are not a random list but are also related to the knowledge associated with {here} and {now}, such that, for example, [[I am a customer]] but not [[I am a professor]] would normally be activated when {here} happens to be associated with [[this is a grocery store]], while the reverse will be the case when {here} happens to be associated with [[this is a university library]]. Some items of associated knowledge are in an implicational relationship with some other items.²⁷

Possible extensions of this to BD are given in Table 5; the associated propositions represent, of course, only some of the possibilities corresponding to the examples of translators discussed above.

Table 5 Examples of propositional knowledge associated with {I} in BD

The proposition [[I am a faceless translator]] may correlate with [[here is the UN]] and may further imply [[I strive for ‘equivalence’]]. This won’t be the case when [[I am a community interpreter]] and [[I like this immigrant]] are correlated with {I}, {you} is associated with [[you are an immigration officer]], and {here} is correlated with [[this is an immigration office]]. Such a combination would lead the “partial” T to translate not as advanced as not at the same level, the same choice the interpreter tactfully (but inaccurately) made in example (1) above. Similarly, when the propositions [[I am a proper person]] and [[you are a proper person]] correlate with {I} and {you}, and {here} is correlated with [[this is a polite society]], T would translate pissing as nose blowing, the choice Miller made in example (4). It is easy to see that the propositions [[I am a feminist]], [[I am gay]], [[I am a humanist/Renaissance man]], [[I am a Czech/German nationalist]], [[I am anti-Muslim]], when correlated with appropriate propositions associated with {you}, {here}, and {now}, would lead T to the decisions made by Korsak, Spicer, Dolet, Hus, Luther, and Ross, and would result in translating faith justifies as faith alone justifies, or love me as fuck me, and so on.

The Transactional Discourse Model can thus formalize the discourse situations that give rise to the translation decisions made by various Ts, whether faceless or with a face, in the past and in the present, and in oral as well as in written communication. This model has sufficient explanatory power to account for the existence of various translation theories, the question we set out to answer at the beginning of this paper. The variations in T’s stance are shown in this model to be determined in correlation with T’s perception of him/herself and the addressee, as well as with the time and place of the given discourse. Cultural assumptions are formalized in this model in the form of propositions associated with the {DEIXIS} involved in a given transaction, and as such they constitute a formalization not only of the interlocutors’ psychological stances and their identities but also of their cultural notions.

A quick glance at a few concerns examined in well-known translation theories will suffice to illustrate this point. Toury’s (1978) concept of norm can be accounted for by this formalism, as can his 1995 thesis regarding the translations being embedded within their social and historical context, since these are cultural notions. T’s decision whether to provide cultural “footnoting” and how to accomplish it is thus explained by the knowledge associated with {DEIXIS} activated at the point when T assesses At. This would answer, for example, Hardwick’s (2000) concerns about providing cultural frameworks for the benefit of the reader. The translator’s purpose (and his/her assessment of the clients’ purpose) in Vermeer’s (1983, 1989) skopos theory is also easily formalized in this model as propositions associated with {I} and {you}. In this way, the model in Table 5 captures all of the possible “hats” a translator may wear, depending on the assumed addressee and the context in which the transaction takes place. Translation thus emerges as a discourse activity shaped by its agent T, whether as a “faceless,” “impartial” mediator striving for ‘equivalence,’ or a “partial,” “opinionated” mediator asserting his or her identities, allegiances, and philosophies through the translation choices s/he makes.

Conclusion

I have argued here for unifying all translation theories on the basis of the discourse nature of mediated discourse. I have proposed a formal discourse model that combines the top-down Durkheimian and the bottom-up Malinowskian approaches and captures the shifting visibility of the translator, incorporating his/her context and cultural background as well as the fleeting propositions that determine the linguistic choices T makes at any given moment. The translator’s shifting benchmarks are natural consequences of the factors captured in this model. As a consequence of this analysis, I have been able to confirm the essential similarity of the cognitive mechanism of Bible translators during the Reformation, Russian translators of Shakespeare in the Victorian age, the ideological stances of post-structuralist translation, or community translators and language brokers in oral translation.

The Transactional Discourse Model was used here to reveal the characteristic features of translation discourse in a consistent fashion that allows for the contrastive examination of other discourse types as well. Formalism is a powerful heuristic device that forces a rigorous logic and invites strong falsifiable hypotheses, and this, in turn, leads to discovery. By assuming that BD is a type of discourse, the model cogently demonstrates that the way T translates is never cast in stone but is a function of time, place, and the self, and the addressee as T sees it. ‘Equivalence’ is only one of the options T may pursue when engaging in BD.

The model also allows us to peek into the future. The “democratization” brought about by the Internet goes hand in hand with the vernacularization of S’s voice. It contributes to the growth of S’s power to reach across time and space, a development that began with the invention of writing systems and progressed with book printing, the spread of literacy, and the affordability of printed media. In the current multilingual globalizing world, the discourse mediator, too, has more opportunities than ever before for asserting his or her identity and opinion. Along with S, T can now also be “heard” by the whole world, while eluding the watchful eye of the powers that be, be they political or cultural. The as yet unknown magnitude of the developments that began two decades ago will continue to alter our perception of ourselves and of our interlocutors. They will also continue to make accessible the interlocutors’ feedback, and, as a result, the proposition sets associated with {DEIXIS} will gain in variety. Better conditions have never existed for being “partial” and “opinionated,” regardless of whether T is engaged in MWD or BWD. It is thus to be expected that translation will continue to become more and more variegated and creative.