Article body

Introduction

During the latest decades, in quite a few detailed explorations of naturally occurring interpreter-mediated interaction or dialogue interpreting (e.g., Apfelbaum 2004, Berk-Seligson 1990, Bjelic 1999, Bolden 2000, Davidson 2000, 2002, Keith 1984, Knapp-Potthoff and Knapp 1986, Mason et al., 1999, Metzger 1999, Roy 2000, Straniero Sergio, 1998, 1999, Wadensjö 1992, 1998, 2000, 2004), the everyday image of an interpreter as someone involved only in translating messages from one language to another, has been dismantled. It has been demonstrated that interpreters’ work is primarily structured by their understanding of the situation, the ongoing activity and its logic, and secondly by the task of translation. Nevertheless, the everyday image of the interpreter as “just translating” indeed appears to influence people’s understanding of interpreter-mediated interaction. It is also part of a professional ideal, and functions as a shorthand explanation of how interpreters perform in face-to-face interaction. This article is about how this ideal image of the interpreter can be confirmed and reinforced in practice, how individual interpreters can, in a certain sense, act and not be seen. It also suggests detailed analysis of naturally occurring interpreter-mediated interaction as a research method, by which it is possible to trace and explain how interpreters “do invisibility,” and how the image of an interpreter as “invisible” is interactively accomplished.

Arguably, how primary participants perceive of interpreters’ involvement in interaction has something to do with how they perceive of one another as partakers in a shared event. For instance, the Swedish radio reporter Lars Borgnäs described an interpreter-assisted interview he once did with Boris Yeltsin. Borgnäs stated:

Yeltsin – just parenthetically, a rather funny thing – Yeltsin had an enormous working capacity and he had this technique to, after delivering an answer, while this was going further via the interpreter and there was a long pause, this technique to lean his head backwards and close his eyes […] That is, it didn’t seem inconsiderate at all, just very professional.”[2]

Ordinarily, if an interviewee repeatedly would disengage in this way from a political news interview it would be considered quite provoking, but apparently, in this situation, it did not. The presence of an interpreter transformed, as it were, the reporter’s understanding, not only of how to put questions, but also of the significance of mutual interest and attention between himself and interviewee.

To my mind, the journalist’s assessment reflects a view of Yeltsin’s and the interpreter’s talk as “source-texts” and “target-texts” respectively and a view of the interpreter as a “translation machine.” Also, it mirrors an image of the interpreter as secondary, someone who, in a certain sense, is not sharing the event with what then stands out as the primary parties – the interviewer and the interviewee. Teasing apart what seems to be a broadly shared image of interpreters, one must note its two, distinctively different theoretical foundations, relating to both language science and sociology. First, when the interpreter is imagined as a simple instrument of language delivery, as a “translation machine,” this idea rests upon a monological model of language and mind. Simply put, it is founded on thoughts of human communication as sets of unidirectional transfer from speakers to listeners and lexical items as carriers of meanings, which speakers “send out” and listeners “receive.” (For further discussion on dialogism and monologism in the context of dialogue interpreting, see e.g., Linell 1998, 2005, Wadensjö, 1998, 2004).

The general approach in this article is to regard interpreting as dialogically organised, social interaction. Nevertheless, the interpreter’s everyday practice is conceived of as a monologising one, in that it implies a continuous de-contextualisation of utterances. The interpreter treats sequences of ongoing talk as “originals,” the corresponding second versions of which they feed into the conversation in progress. Interpreter-mediated interaction implies a specific turn-taking system, different from the one found in ordinary conversation. People communicating through interpreters are to make sense of the content of talk, somewhat disregarding the progression of it (cf. Wadensjö 1998, 2004).

Secondly, the interpreter’s expected behaviour in interaction reminds one of the non-standard behaviour of a non-person, as defined by Goffman (1990 [1959]). The image of a non-person interpreter relates to a model of discrepant social rolesand normal social behaviour. In the following paragraph, I will devote some space to relate the concept of non-person to other Goffmanian concepts designed for explorations of participation in social encounters, before applying them to a close analysis of sequences of talk drawn from the ALL TALK interview. I will examine how, more precisely, the interpreter communicates non-involvement in a more or less marked way, and how he, unnoticeably to the primary participants, promotes the progression of an ongoing exchange. I will also look at how the primary parties’ orientate towards the interpreter, while fitting their own performance with the communicative genre at play.

“Invisibility” and the Role of Non-Person

Goffman (1990 [59]), in his early writings, identifies as a discrepant social role the role of non-person. In his definition, individuals who play this role are people who are “present during the interaction but in some respects do not take the role either of performer or of audience” (Goffman 1990: 150). As a classic type of non-person he mentions the servant. Servants are expected to be present at particular occasions at the same time as they are defined, in certain ways, as not being there. A servant at, for instance, a reception or a dinner party is neither performing as a host nor attending as a guest. Other examples would be photographers and broadcast technicians. In these occupational roles people typically act as non-persons. For instance, during public events like concerts and conferences they normally take liberties to move about while the audience is supposed to take their seats. They can act and talk, carry on with their business, without causing a redefinition of the ongoing event. In some respects, their activities are not expected to contribute to, or be part of the event as such. Those performing on stage will carry on with whatever they are doing, ignoring, in some respects, those who broadcast or film them. Also, the very young, the very old, the sick and, sometimes, the foreigner can be given and can take on the role of non-person. A non-person is someone who is believed not to understand, at some level, talk among people who share the same presuppositions, insights or means of communication.

As Aronsson (1991) demonstrates – drawing upon Goffman’s socio-psychology and with the help of a paediatric case study – what a child says in a paediatric interview will be taken as a contribution to it by the adults, depending on whether the child is treated as fully present to the conversation or not. She draws the conclusion that negotiations of meanings in this kind of multi-party talk – the doctor/parent/child-patient talk – concern not only the meanings of utterances but, directly or indirectly, also personhood or non-personhood (Aronsson 1991: 73).

It would seem, as Goffman (1990[59]) points out, that the role of non-person usually carries with it some subordination (Goffman 1990: 151). However, the person who is given or who takes such a role can also use it to avoid subordination, simply by ignoring others’ definition of the current event, and thus ignore what would be conforming to ordinarily expected behaviour. An individual acting in the role of non-person enjoys the privileges of being able to address anyone or to ignore being addressed.

Interpreters’ Personhood and the Machinery of Translation

Non-personhood, I would argue, is inherent in the social role of interpreter (i.e., irrespective of how individual interpreters behave in particular situations). Those acting as – and those acting through – interpreters in social interaction, are more or less oriented to this specific, culturally established character of the role of interpreter. After all, an interpreter is by definition a kind of servant rather than a main figure. Dialogue interpreters’ physical presence and the image of them as “non-present” creates what we could call a communicative wiggle room, following Erickson’s (2001) metaphor for the variability of interlocutors’ local “readings” of one another in social interaction. Such a wiggle room, I would argue, is characteristic of dialogue-interpreted encounters, and available as a communicative resource to those interacting in them.

A person who understands interpreters to be performing a non-person’s role does not necessarily see them as translation-machines, providing “verbatim translations.” Verbatim translation is in itself an inherently contradictory notion, since translation by definition implies application of another language, which normally involves the use of “new” words. Also, those attending to interpreters’ performance are normally unable to assess it from the point of view of correspondence between participants’ utterances. Still, there is an interesting connection. Interpreters who anticipate criticism in their address, or anything else that could challenge current interaction and their position in between the parties they assist, can utilise the everyday notion of interpreters as “translation machines” to re-establish the interaction order and themselves as someone not fully present to (and hence not responsible for) the ongoing exchange (cf. Wadensjö 1998: 66-67, 118).

Interpreters may also decide to assert their personhood if they feel their personal integrity is being violated while on duty. For instance, Frishberg (1990) quotes an interpreter who had been working in a classroom with a deaf student and a lecturing professor. At one point, the instructor physically touched her in an unexpected and undesirable way. During class, the interpreter suppressed any personal reaction, but later, in an after class situation, when the teacher touched her once more in the same way while she was interpreting, she reacted as a moral agent, turned to him and said: “Please don’t touch me” (Frishberg 1990: 29-30). It seems likely that the interpreter was attentive not only to the need for clarifying her social role[3] but also to matters of timing, communicative space and the shared definition of the current encounter (as lecture and after class situation respectively).

Participation – Three Analytical Levels

People tend to typify social encounters depending on the participants’ social activity roles – such as instructor, student, doctor, patient; show host, interviewee and interpreter. The occupational role of interpreter does not define, say, a medical encounter as medical or a talk show interview as talk show interview. This is one reason why interpreters can be understood as filling the social role of non-person.

Looking at the social identity of those taking part in social encounters is one level of inquiry about people’s participation. The following concern participation in two other senses.

Goffman (1981) in his book Forms of Talk outlines a model of footing or participation framework to explore social encounters with regard to who takes part in them and who contributes to their substance. For instance, in a televised interview, the interviewer and the interviewee(s) normally are the (only) ratified participants. The interviewer, the person in charge, could, for example, occasionally engage someone from the audience, in Goffman’s terms the bystanders, to allow them to ask questions. That exchange would be seen as crossplay, i.e., “communication between ratified participants and bystanders across the boundaries of the dominant encounter” (Goffman 1981: 134), in this case, the interview. If more than two persons are ratified, as participants in the dominant encounter, there are also, in principle, possibilities for occasional byplay, i.e., “subordinated communication of a subset of ratified participants” (Goffman 1981: 134). Exchanges between people off-stage, finally, could take the form of sideplay, i.e. “respectfully hushed words exchanged entirely among bystanders” (Goffman 1981: 134).

The status in interaction of non-persons is by nature flexible. Being physically present, they can initiate and be engaged in occasional byplay, crossplay and sideplay, with or without people taking their talk as contributions to the ongoing dominant encounter.

In Goffman’s (1981) analytical cluster of participation, a non-person (X in figure 1) would appear on the borderline between the dominant encounter and the off-stage region(s), with a potential of being included among the ratified participants (R-s in fig. 1) and the bystanders (B-s in fig 1.) as well as being excluded from both groups, to end up among the overhearers (O-s in fig. 1).

Figure 1

The non-person’s flexible status of participation

The non-person’s flexible status of participation

-> See the list of figures

Goffman’s (1981) model of footing also includes analytical concepts to describe speaker-hearer roles or aspects of self. Several authors have made these aspects relevant in the exploration of the distribution of responsibility for talk and interaction in interpreter-mediated encounters (Wadensjö 1992, 1998, Straniero Sergio 1998, 1999, Mason 1999, Metzger 1999, Roy 2000, Davidson 2002). It may be worth repeating, however, what level of participation these concepts are drawing attention to.

Hence, thirdly, in social interaction, people project various aspects of self. The way in which speakers produce utterances reveals something about in what sense they are prepared to answer for the content and form of it. Goffman identifies three “production formats,” namely animator, author and principal (Goffman 1981: 226). Interpreters normally speak on someone else’s behalf, thus projecting their selves as the others’ animator, as a “sounding box from which utterances come” (Goffman 1981: 226). Moreover, in animating the others’ talk in another language, interpreters evidently also take responsibility in the role of author, “the agent who puts together, composes, or scripts the lines that are uttered” (Goffman 1981: 226). Normally, however, they do not position themselves in the role of principal, i.e., as “the party to whose position, stand and belief the words attest” (Goffman 1981: 226). Interpreters by definition give voice to words that are to demonstrate another person’s outlook and thinking, without necessarily subscribing to this thinking themselves, but appointing the prior speaker as ultimately responsible for the “belief the words attest.” Of course, as studies of naturally occurring dialogue interpreting, such as those mentioned in the introduction to this article have made clear, interpreters do talk a lot from their own point of view and out of their own interest. For instance, they request and provide clarifications. They can then project themselves as principals of the words they utter, and, as a rule, their ownership of these words is perfectly distinguishable to the participant(s) speaking the language then used. Evidently, participants may also understand interpreters to express their own opinions by the way in which they talk on others’ account, irrespective of the interpreters’ efforts to avoid taking stances, let alone showing them. However, the Goffmanian model does not adequately capture this latter aspect.

Explorations of Alignment Work in Talk Show Data

Straniero Sergio (1998, 1999) draws partly on Goffman’s model of footing in his studies of Italian televised talk shows. He states, among other things, that interpreters’ conversational behaviour in the context of talk shows is determined by the alignment work accomplished by the show host (Straniero Sergio 1999: 323). The show hosts in his talk-show data typically, every now and then, addressed the interpreters directly. For instance, they became the target of the host’s teasing, a generic feature in this type of TV event. Straniero Sergio suggests that the host in this way made the interpreters feel obliged to speak for themselves, and shows that they also did so. He explains that, instead of following the normally taken-for-granted principles of acting “invisibly,” from behind the scene, these interpreters here applied what he characterises as “appropriate conversational behaviour” (Straniero Sergio 1999: 323) and exemplifies how they occasionally positioned themselves as principal in Goffman’s sense.

In three of Straniero Sergio’s six examples, the host was addressing the interpreter, calling her or him either informally, by first name, or using the member category, “the interpreter” (l’interpret). In other words, these addressing utterances were “recipient designed” (Sacks et al., 1974), explicitly oriented to the interpreter, who in this way became “personified,” as it were. In two of the excerpts, interpreters were displaying their linguistic authority and expertise without being explicitly invited to do so. When speaking on their own behalf, the interpreters in all of Straniero Sergio’s examples were talking about current talk and talking. One of his examples involves another activity and deserves some additional attention.

Playing With the Interpreter’s Presence and Agency

In a long final example, Straniero Sergio (1999: 320-321) shows how the interpreter took the liberty to speak, when the host was in the middle of a syntactically incomplete unit. The show host acknowledged the interpreter’s utterance. It is phrased as an initiative from the guest and also recognized by the host as talk corresponding to the guest’s prior “original” talk. Then the interpreter adds a statement about the guest in a previous situation, answering directly to the host’s turn. The show host then also included this statement into common discourse, as something he – the host – is actually saying himself.

Excerpt 1 below is drawn from this talk show example. The show host Maurizio Costanzo (MC) and the managing director of a mineral water company (MW) are on stage together (both speaking Italian). A Bulgarian woman (B) and the interpreter (I) are together on an adjacent stage, connected with the main stage through a video link. Current talk concerns brands of mineral water that the woman had been drinking while being on a hunger strike with her family, an event that had led to her inclusion in the present show. The interpreter, responding to an occasional gaze in his direction from the show host, says: “she wanted to say something about mineral water.” The host asks him whether the Bulgarian woman had something to complain about. Being reassured that she had not, he lets the guest speak for herself. Allowed to speak, she cites a doctor who had talked about one particular brand of mineral water – the one produced by MW’s company – as being good for children. Excerpt 1 shows just the end of her contribution, and only in English translation. In reality people spoke Italian and Russian. In the excerpt, italics indicate talk in Russian.

As Straniero Sergio (1999: 322) points out, the host in this exchange expresses his authority, using explicit performatives like “let me make it clear,” “I know” and “I don’t want” (1: 23-24) as a means to regain control of the exchange. Ratifying the interpreter as participant in the ongoing event, he strengthens his own position as currently in charge of the interview. In the above excerpt it can be seen how the host (MC) repeats (1: 29) the interpreter’s (I’s) exact wording (1: 28), adds a “but” and elaborates a narrative about the Bulgarian guest. As Straniero Sergio argues, referring to Orletti (1994), “repetition followed by negation tends to signal conflict or disagreement” (Straniero Sergio 1999: 322). Moreover, as Bockgård (2004) explains, when a speaker (here: MC) repeats in the third turn, a second speaker’s (here: I’s) extension of this speaker’s (MC’s) syntactically incomplete unit, this is a way to claim the ownership of the second turn. Repeating the interpreter’s words (1: 28), the host indicates that what has been said is acceptable as something said in his (MC’s) turn (1: 27), even though it differs from what he had actually intended to say (1: 29). Claiming the ownership of the previous turn, the host makes a statement about the Bulgarian woman’s “new ‘happier’ television experience” (Straniero Sergio 1999: 322), instead of permitting her (and the interpreter) to foreground the hunger strike experience to remain in the talk on mineral water.

For the sake of variety, or for other current practical purposes, the host in Straniero Sergio’s examples was playing with the everyday image of interpreters, using the communicative wiggle room it offers as a resource. His examples showed that the interpreters spoke for themselves and took noticeable initiatives of their own. These mainly demonstrated the interpreter’s view as language expert on something just said. In the last example, the interpreter talked about something other than current talk and the host took over the ownership of his contribution.

Straniero Sergio’s examples illustrate how the interpreter’s talk is available for validation as contributions to an ongoing exchange and for the show host’s appropriation. The following sequences will give examples of yet other ways in which a primary participant can attend to and treat an immediately present interpreter’s talk.

In the ALL TALK interview, there were no apparent interpreter initiatives such as those observed in the Italian data. Nevertheless, this interpreter was also acting as a participating, moral agent, but less apparently so. More obviously, he communicated non-involvement in the encounter, notably, with the help of the interviewer and the interviewee, who were treating him variably as sharing and as not fully sharing the event. In the following, I will demonstrate this, scrutinising two sequences drawn from the talk show interview.

Alignment Work in the ALL TALK Show

In the ALL TALK show, the interpreter was a rare element. The particular broadcast in question here was an exception. It took place in 1996, five years after the Soviet Union was dissolved. The BBC journalist and show host Clive Anderson had invited Michail Gorbachev, at the time a (non-successful) candidate in the recent Russian presidential elections and a popular political lecturer in the West. He was accompanied by Pavel Palazchenko, whose top-level proficiency as a Russian–English interpreter is undisputed and widely acknowledged.

When foreign high-level officials are interviewed for the media, interpreting is sometimes performed into only one of the languages. Interviewees may understand interviewers’ languages and vice versa, but prefer to speak their own. This has important implications for the organisational structuring of the event. In the present case, the interpreter worked into both languages. According to him, Gorbachev did not understand, let alone speak any English, apart from a few words, and Anderson appears to have been equally unqualified in Russian.[4]

The ALL TALK show was modelled on, and was a parody of, an ordinary news broadcast. It normally started with the show host reading fake telegrams during short video clips, with funny commentaries alluding to current news. There would then follow a celebrity interview, in this example, with Gorbachev.

Basically, the host of a broadcast show needs to manage quite a complex interactional activity. Apart from the people interacting on stage and being filmed, there is also a studio staff and occasionally, as in this case, a studio audience. Also, quite obviously, a talk show is ultimately organised as talk for an overhearing audience. The show host is in charge of introducing the viewers, keeping track of the cameras, talking to the audience, asking questions, and so forth.

In contrast to the interpreter in the Italian talk shows mentioned above, this interpreter did not markedly align with the host at any occasion, despite Anderson’s explicit invitations. In Goffman’s terms of “production format” he spoke only from the position of animator and author, but never projected himself as principal. In Wadensjö’s terms of reception format or modes of listening (Wadensjö, 1998), he positioned himself as reporter and recapitulator of the others’ talk. At rare occasions he also acted as a responder (Wadensjö 1998: 91-92) in a way that would be expected of an ordinary conversational partner, however, he did so without coming forth as someone stating his own point of view, or talking out of his own interest. This will be discussed more in relation to Excerpt 3. In Excerpt 2 he is acknowledged as an active participant in the ongoing encounter. Nevertheless, he keeps his professional distance.

“Well done Pavel”

Excerpt 2 demonstrates how Clive Anderson at one point, in a side comment, addresses the interpreter directly, bringing attention to and recognising his presence at the current moment, as well as in a past, talked about event. Palazchenko (PP in the transcripts) subsequently reports to Gorbachev what the host just said, including this comment in his own address. The sequence starts eight minutes into the interview. Clive Anderson (CA in the transcripts) puts a question referring to something Gorbachev (MG in the transcripts) wrote in his autobiography. (AUD in the transcripts stands for the audience). Excerpt 2 shows talk in the original English and Russian. Under each numbered Russian line is my English “close” translation (in italics).[5]

Excerpt 2 shows Anderson prefacing a new topic, addressing his interviewee, and adding a comment to it: “wha- what struck me is something else you mentioned in your book was the- the fact tha- I think it was at Reykjavik, when you were having disarmament talks, man to man, just you and president Reagan. I dare say you were there as well, mister translator” (lines 2: 322, 323, 325, 327, 329, 331). The interpreter’s rendition of this in Russian largely comes as parallel discourse (lines 2: 324, 326, 328, 330 and 333).

Pointing out the expression “man to man” in the context of an interpreted (triadic) exchange goes in line with the host’s normal way of playing with the ambiguity of words and expressions. The audience reacts as expected – they burst out in laughter (2: 332). The interpreter slightly changes his position, leaning a bit back and forth. There may also have been a glimpse of a smile on his face, however, his attention to a point in front of him and the pitch of his voice stays the same. He latches on immediately, speaking in Russian, with his usual “poker face” (2: 333). The audience’s laughter (2: 332) makes it impossible to hear exactly what he says, but clearly Gorbachev can hear him, and undoubtedly the interviewee gets to know what the audience was laughing about – Anderson’s linking with “mister translator.” Catching on directly to what Palazchenko says on the host’s behalf, Gorbachev looks at Anderson and points at the interpreter.[6]

Notably, Gorbachev does not speak in the slot potentially designed for the interpreter. Commenting without words, he avoids taking the turn when the interviewer speaks and he avoids answering to something not recognisable as a question. He doesn’t really break, although he doesn’t altogether conform to the ordinary convention of an interview, presupposing the interviewer to allocate turns, and select the next speaker (Heritage & Greatbatch 1991). Gorbachev does not self-select, but the audience’s reaction (2: 338) to his non-verbal pointing in effect converts it into a substantial turn.

Figure 2

Anderson: I dare say you were there as well, mister translator.

-> See the list of figures

Figure 3

Gorbachev points at Palazchenko, the interpreter.

-> See the list of figures

Alignment with the Audience

If we look at this in terms of the model just presented, Anderson would seem to have ratified the interpreter as participant of the current encounter when addressing him as “Mister translator” (2: 331). The term of address is the occupational one, relevant in both the Reykjavik event referred to and in the current interview. By pointing at the interpreter, Gorbachev puts focus on him as well, confirming his physical presence in the current and in the talked about situation. At the same time, he casts him as a bystander to the ongoing exchange by looking away from him while catching the gaze of Anderson. The initiative subsequently transforms to become part of a side commentary sequence, playing – as part of the show – with the personhood of the interpreter.

When the host addresses the interpreter again with “well done Pavel” (2: 337) Anderson’s gaze and body movements indicate that he is not expecting the interpreter’s alignment. Looking straight into the camera, the host addresses his faithful allies – the overhearing audience(s) – and gets the expected, normalising reaction – we laugh. The audience is tuned in on Anderson’s turns, listening for the exact spots where to laugh, here, as well as during the rest of the interview. And while the laughter continues, Anderson holds on to his turn, addressing Gorbachev “and the: : the: : (.) and you” (2: 337). Arguably, the show host’s prosody, gaze and body orientation, plus his choice of the informal address – Pavel – is used to bracket the current dominant activity (the interview) and thus accomplish what at this point clearly stands out as a side commentary sequence. (Cf. David Letterman’s short exchanges with one of his broadcast technicians, coming as time-out activities in the Letterman show.)

The show host did not receive alignment from Palazchenko and failed to make him act as co-interviewee. Gorbachev, by retaining the attention to the interpreter’s physical presence, demonstrates that he – the non-English speaker – is with Anderson. The guest acknowledges and joins the host’s playing with the non-person aspect in the interpreter’s occupational role. By pointing at Palazchenko and at the same time looking away from him, Gorbachev demonstrates his understanding of the interpreter as not fully present to the ongoing exchange. His pointing gesture also evidenced, as it were, that Palazchenko had “just translated” and translated “everything” Anderson had said, also the part that wasn’t produced for Gorbachev’s ears.

The ALL TALK show differs on an important point from those Italian talk shows investigated by Straniero Serigo (1998, 1999) and others (e.g., Chiaro 2002, Mack 2002). In these, a side-stage sofa was connected with the main stage through a video link. Clive Anderson could not use different constellations of guests to initiate quick frame shifts. Instead, for this purpose, he could utilise the studio audience.

Charles Goodwin (1986), in an issue of Text, wrote about an audience’s option to creatively assign meanings to what is being said on a stage. He observed that the audience is both shaped by the talk it is attending and is helping to shape what will be made of that talk (Goodwin 1986: 311). Indeed, the studio audience’s reaction to Anderson’s and Gorbachev’s toying with the interpreting business, illustrated in Excerpt 2, helped shape it as something as equally “non-serious” as the rest of the talk show. The ALL TALK show interview is a hybrid communicative event because it appeals to the way “real” political interviews are accomplished, at the same time as it is displayed as non-serious, as entertainment. Excerpt 2 illustrated a moment in which the interpreting matter was also “hybridised,” adapted to the entertainment mode of the interview. This was accomplished both by the primary parties, overtly violating the normative expectation that interpreters in interaction are to be treated as people present but not there, and by Palazchenko, by his non-compliance with “ordinary” social behaviour, while meeting the terms of normal interpreter behaviour.

In the following example, the primary participants do not point to the (supposed) “invisibility” of the interpreter, however the resulting communicative wiggle room is no less decisive for what is accomplished in interaction. To understand how, we need to look at the interpreter’s way of feeding renditions into common discourse.

Alignment Work and Interpreting Techniques

Palazchenko was switching between consecutive and simultaneous interpreting when working into Russian as well as into English. Irrespective of language, the technique applied seems to have depended on his anticipation of the length of the participants’ ongoing utterances. However, there was an important difference between the languages. When interpreting into English he spoke loudly. When interpreting into Russian he adapted his voice to the ears of Gorbachev, sitting right next to him. The interpreter’s presence was thus much more noticeable when he spoke English (on Gorbachev’s behalf) for the host and the audience, compared to when he spoke Russian (on Anderson’s account) for Gorbachev. This meant that the Russian interpreting to a large extent could be performed as byplay, as a subordinate activity going on in parallel with the commonly focused interaction. The audience could hear it only occasionally during the show.[7] Palazchenko’s interpreting into English rather took the form of crossplay. The interpreter, evidently, did not count as a contributor to the interview but as a mediator of communication. Hence, when he talked in English this was conceived of as talk originating from Gorbachev, a state of affairs that was indeed supported by Gorbachev’s listening behaviour. The interviewee occasionally could “freeze” in a speaker’s position, keeping eye contact with the host, attending to Anderson’s reaction to his words, during the interpreter’s delivery of them in English.

In the following, I will analyse an instance where Anderson and Gorbachev compete for the floor and Gorbachev, who ultimately gets it, follows through with his communicative project – the telling of a joke – with the unspoken support of the interpreter.

“I hope this joke is very good, I must say”

Where Excerpt 3 starts, we are well beyond the middle of the interview. Anderson prefaces a new question, again referring to Gorbachev’s newly published memoirs.

Before the show host has formulated a proper question (3: 495-496), Gorbachev takes the mentioning of his book (3: 497) as an opportunity to suggest a new topic – his alcohol and drinking policy in the USSR (3: 498). Notably, Palazchenko starts rendering the question in overlap with Anderson (in byplay with Gorbachev). The show host’s marking his ownership of the current turn “but you- you” (3: 496) never occurs in the rendition (3: 497), before Gorbachev comes in with his initiative (3: 498), which the interpreter subsequently renders into English (3: 499-500). Evidently, in interpreter-mediated interaction, the technique of interpreting has an impact on the shaping of transition-relevance places (TRPs) (Sacks et al 1974).

TRPs in interpreter-mediated talk

According to research on interpreter-mediated interaction (e.g., Englund-Dimitrova 1997, Wadensjö 1998, Davidson 2002) interpreters seldom, if ever, provide second versions of what they understand as back-channelling behaviour. The reason is fairly obvious. A single feedback token, such as “yes,” “mhm” or “okay,” when repeated, stands out as a marked contribution to the current exchange, rather than as unmarked conversational acceptance or encouragement (cf. Wadensjö 1998: 273).

Hearing the interpretation of Gorbachev’s initiative (3: 500), Anderson immediately reacts with what sounds like and comes to function as a minimal responsive act ”yes” (3: 501). Palazchenko doesn’t translate it, what thus should be in line with expectations, according to the above-mentioned research. But he does not leave the floor open. Instead, compensating, as it were, for the inbuilt delayed progression in interpreter-mediated interaction, he elaborates on his latest rendition (3: 499-500) by specifying something understood in Gorbachev’s preceding original (3: 498) (irrespective of the fact that the rendition (3: 499-500) would seem complete). Going on talking (3: 502) he responds in the way conversational partners ordinarily do to feedback of this kind. However, what he says stands out as the English version of Gorbachev’s prior talk and not as something originating from himself.

The show host at this point comes up with a counter question “which did- which didn’t wo: : rk, did it?” (3: 503) The audience’s immediate reaction – the laughter (3: 504) – answers to this and marks it as a playfully offensive act. The question is interpreted into Russian (3: 506). The interpreter feeds his rendition of it quickly into discourse. At the same time, Gorbachev is in the middle of a suggesting a joke (3: 505) following up on the topic he just introduced (3: 498), and he pursues the suggestion (3: 507) after a hardly appreciable interruption during the prior rendition (3: 506).

Interestingly, the interpreter’s rendition of Gorbachev’s suggestion to tell a joke (using a questioning intonation already in the first mentioning of razkazath (“tell”) (in 3: 505), transforms it into an announcement of the telling: “I’ll tell you a joke” (3: 508). Palazchenko is, one could argue, already a step ahead. Anticipating the telling of the joke he reduces the current need for management of interaction. Simultaneously, Anderson’s questioning project is nipped in the bud. Arguably, the interpreter’s action is tuned in with Gorbachev’s casting of Palazchenko’s overlapping speech (3: 506) as interruptive, by postponing the turn completion “tell? you a joke” (3: 507) (cf. Lerner 1989).

Palazchenko is not necessarily and deliberately siding with Gorbachev, I would argue, however, he is evidently alive to what is happening in interaction from moment to moment, in terms of turn allocations and turn shifts, and seems to have a primary orientation towards supporting a shared focus between interviewer and interviewee and, at the same time, limit his own involvement.

Gorbachev, in turn, seems to have been expecting, at this point, some kind of warrant for the telling. In the absence of this, he fills in that the joke relates to his current topic (3: 509).

The pause and the audience’s reaction shape Anderson’s news receipt “oh good” (3: 510) as a marked, ironic sort of approval (cf. Heritage 1984), which is immediately appreciated by the audience (3: 511). When Anderson says: “I’ll buy you a drink” (3: 513), thus making the exchange resemble what Marjorie Goodwin, analysing forms of disputes, has termed a “paired counter” (Goodwin 1990: 241), they laugh again (3: 514).

The show host’s pretended suggestion to buy a drink could also be understood as a pre-sequence to what is about to come – alluding to what would be a standard phrase at a pub – Anderson’s occasional time-out from the interview. Sipping from his glass of water and literally turning his back on the interviewee, Anderson announces the next extended turn to be Gorbachev’s: “I hope this joke is very good, I must sa: : y “ (3: 517).

Meanwhile, the irony of Anderson’s “oh good” (3: 510) has been neutralised in Palazchenko’s confirming choroshó (“good,” “okay”) (3: 512). The paired counter, “I’ll buy you a drink” (3: 513), makes the interpreter hesitate and search for words (3: 516), and the comment about Gorbachev’s storytelling, “I hope this joke is very good, I must sa: : y” (3: 517), passes without being interpreted into Russian. In the slot available for the rendition, Palazchenko slightly coughs (3: 519). Even if Anderson was partly talking in overlap with him, the interpreter must have heard the provocative sanctioning of the storytelling, and subsequently would have been able to render it into Russian, but he doesn’t. The floor is left open. His renditions in this sequence (3: 499-500, 3: 502, 3: 506, 3: 508, 3: 512, 3: 516) are all backing one communicative project – Gorbachev’s telling of a joke (3: 520), and so is arguably also his non-rendering into Russian of Anderson’s “hope” (3: 517). Most likely however, neither Gorbachev nor Anderson, nor the members of the audience(s), could have realised at the time what the above detailed analysis suggests.How, more precisely, he was acting in his own interest, as mediator of the shared event, passed unnoticed to all people present, probably including Palazchenko himself.

Concluding Discussion

I started this paper by referring to a radio interview with Boris Yeltsin, describing how he was able to ignore the ordinary courtesy of an interlocutor and the ordinary attention of an interviewee, by leaning his head back and closing his eyes, while the interpreter was speaking on his behalf, without the interviewer thinking of this as rude or strange, but as “just very professional” (cf. above). In interpreter-mediated, face-to-face interaction, how the primary parties perceive of the interpreter’s agency appears to be directly dependent on how they perceive their own and the other party’s participation in a shared event. People communicating via an interpreter almost by definition do not share the same presuppositions, insights and means of communication. Yet, as the examples explored in this article indicate, in these kinds of conversations, people can also confirm and appreciate one another as “ordinary” conversational partners, e.g., through mutual attention and shared gazes, irrespective of a language in common.

Bruce Anderson (1976), in his early explorations of interpreter-mediated courtroom hearings, clarified the need for including non-verbal behaviour in the analysis of the work of interpreters. A lot of people seem to, in principle, agree, but there are very few systematic, empirical investigations of how non-verbal communication – gaze, body movements, mimicry and so forth – in effect work in actual cases of dialogue interpreting. This issue was only briefly touched upon in the present paper and would deserve more thorough attention in future research.

In this paper, I took as a point of departure that the role of interpreter is culturally established as something of a non-person (Goffman, 1990 [1959]) in that interpreters are supposed to act and not be seen. In analyses of sequences drawn from two talk show interviews, I observed how this non-person aspect helped create a certain communicative wiggle room, and how those interacting with each other utilised it to exercise certain frame control and initiate frame shifts.

Excerpt 1, drawn from Straniero Sergio’s (1999) Italian talk show data, demonstrated how the show host first confirmed an initiative from the interpreter (phrased as originating from an interviewee) as part of common interaction. When the interpreter then said something about the interviewee, the show host appropriated the ownership of his turn, ratifying and dismissing the preceding speaker at the same time. The presence of an interpreter in interaction makes his talk available in the main exchange, as talk of a ratified participant. It can also be cast as non-valid, as talk of a bystander or an overhearer, in Goffman’s (1981) terminology.

In contrast to the interpreters in the Italian data, Pavel Palazchenko, interpreting in the BBC interview, avoided overtly aligning to the show host’s explicit summons in his address. As demonstrated in Excerpt 2, he “just translated” something Clive Anderson said, irrespective of the fact that part of it was addressing himself and wasn’t meant for Gorbachev’s ears. This action, when “reviled” by the interviewee, could be seen as confirming the image of the interpreter both as non-person, enjoying specific rights in social situations, and as translation machine, not siding with anyone, just enabling interaction (by making the reason for the current laughter shared knowledge).

After this instance, during the whole interview, Clive Anderson never again explicitly addressed the interpreter. He treated him as present to the encounter, but only in the capacity of intermediary, rendering others’ talk. Also Gorbachev treated him as a ratified participant only in this sense. At no occasion did he address the interpreter directly. This could be understood as a strategy to defend, in front of the host and the audience, his own position as a full-fledged participant, fully sharing the event with the others, as was shown in Excerpt 2.

In practice, a non-person’s “invisibility” is first and foremost in the eyes of the beholder. As demonstrated in Excerpt 2, the primary participants were playing with the interpreting business, adapting this aspect of the interview to the generally hybrid character of the talk show. In this way they were highlighting the restricted and specific agency principally and formally expected of interpreters.

Interpreters’ “invisibility” is also “real” in the sense that they perform actions that remain unnoticed by the parties they assist. The third excerpt showed an instance where Clive Anderson was in the process of formulating a question, when Gorbachev proposed to tell a joke. In the end, he gets to tell it, not exactly encouraged and not really discouraged by the host, but tacitly supported by Palazchenko’s intermediary actions. These were occasionally performed from the position of responder (Wadensjö 1998: 91-92). The interpreter responded directly, anticipating pragmatic features of turn composition as an “ordinary” participant, for example to some instances of parallel talk, to back-channelling tokens and to turn allocations through (lack of) gaze-directional addressing. These actions helped disambiguate the current turn-taking order in favour of Gorbachev’s projected trajectory of talk. The interpreter’s backing of Gorbachev was also communicated through regular interpreter turns. Palazchenko neutralised, as it were, Anderson’s ironic approval, by timely delivering his renditions of it in an unemotional, unmarked style.

To my mind, it would be a simplification to argue that the interpreter was intentionally siding with Gorbachev, being associated with him as his employee. This is an issue that of course is open for speculation; however, people’s intentions can hardly be proved empirically. What can be observed, however, is that Palazchenko’s technique of interpreting, in combination with the physical set up of the ALL-TALK-show interview, in some sense privileged Gorbachev’s participation. Speaking on his account, for the host and everyone else to hear, the interpreting, displayed as (the second version of) Gorbachev’s talk, became part of the focused and shared communicative event. In contrast, the Russian interpreting (on Clive Anderson’s behalf) could go on with less demand on communicative space and so permitted the listener an exclusive and more immediate access. Sitting close to Gorbachev, Palazchenko could perform his Russian interpreting as byplay (Goffman 1981: 134).

As Roy (2000: 67 ff) has pointed out, in cases of competing communicative projects introduced more or less simultaneously by the primary parties, the interpreter, in order to secure a joint focus and a shared trajectory of talk, needs to foreground one, at the expense of the other. Further detailed analyses of real-life dialogue-interpreted discourse data will reveal more about how interpreters read situations, including hints from the primary participants’ communicative behaviour, when it comes to giving priority to one before the other. It will also disclose how this is achieved to a more or less noticeable extent to those taking part in the event.

To conclude, the image of an interpreter that emerges from an interpreter-mediated event naturally depends on his or her proficiency in the two working languages. Moreover, it depends of how all people present to the encounter engage in the language game that uniquely characterises dialogue interpreting – the intermittent de-contextualising and re-contextualising of utterances and the resulting “artificial” separation between progression and content of interaction. In this paper, I have shown how the primary participants, taking the illusory “invisibility” of the interpreter seriously (e.g., by joking about it), helped him act as someone holding no stake in interaction, and, in line with this, appear as someone “just translating.”