Bilingual formal meeting as a context of translatoriality

Drawing on the concept of translatorial action by Justa Holz-Mänttäri, this article sets out to analyse the role of translation in a bilingual formal meeting without any professional translation or interpreting. The analysis reveals the central role of translatorial activities: 60 % of the turns include some kind of translatoriality. The chair and expert speakers stand out as producers for most of the translations. Self-translation is the most prominent form of translation, but otherwise the translator role tends to vary dynamically with the role of the source text producer. Three types of translatorial action with varying degrees of replication of content were found: duplicating, summarizing, and expanding. In the meeting context, translatorial action is the primary means of enabling participation for all, regardless of language skills or language background, and this action was used by the participants in flexible and dynamic ways.


Introduction
Over the past two decades, institutional translation and interpretation have attracted increasing attention in translation studies.As a result, our understanding of professional translators' and interpreters' roles and tasks in institutional contexts is gradually increasing (see e.g.Koskinen 2011;Kang 2014).However, not all translation in institutional contexts is carried out by language professionals.On the contrary, there are numerous settings where people with different linguistic backgrounds work together and different languages coexist without the presence of professional translators or interpreters institutionalised to act in mediating roles.
While pragmatic solutions such as working languages are known to be used to reduce linguistic complexity, it is also reasonable to assume that various kinds of ad hoc translating and interpreting performed by other actors may take place in these contexts (see Pilke, Kolehmainen and Penttilä 2015).
So far there has been little research of less formalized translatoriality in institutionally set multilingual situations, and it has often focused on the use of non-translatorial communication is that there are two message carriers present and they have a relationship of relevant similarity with each other.Readers may recognize that we borrow our terminology from the classic treatise of translation theory by Justa Holz-Mänttäri (1984).Her theory of translatorial action is generally known for its defence of the professionalization of the field of translation.However, this article builds upon an alternative approach as it focuses on translatoriality in a context where the role of the translator is taken up by meeting participants who are not trained or paid for their translation or interpreting services but who take them on as a part of or alongside their other professional duties. 1 Although not designed for such cases by Holz-Mänttäri herself, as a versatile and productive approach the theory of translatorial action offers a coherent framework for this purpose as well (see Snell-Hornby 2006, 163).

Aim, data and method
The purpose of this article is to analyse the language practices in the meetings from the point of view of translatoriality.The starting point of the analysis is the assumption that since participants have different language resources and since a formal meeting requires communication to fulfil its goals, some sort of translatorial action is needed in order for the meeting to be successfully carried out.
The study is a subproject of a more extensive research project carried out by the research team BiLingCo (Bilingualism and Communication in Organizations) at the University of Vaasa. 2 Within the context of the Regional Co-operation Group of Ostrobothnia the research team has collected meeting-related materials: audio-and 1 Terminology in this area of research is still unstable.We call this type of activity paraprofessional translation to highlight that it takes place in the context of professional activities and is dependent on the institutionalized professional framework of the participants.By opting to differentiate translation by professionals other than translators in professional contexts from volunteer or non-professional translation entirely outside paid work we postulate, until we have uncovered more details of the nature of both kinds of translation, that there are likely to be differences between them in the translation processes, in the functions and roles of the participants as well as in the constraints posed by the context.This tripartite division follows the one proposed by Tuylenev (2014), but different from him we do not indicate a quality judgement: paraprofessional, for us, only refers to the fact that these professionals translate or interpret alongside their main job or use translating or interpreting to "fulfil their main professional duties" (76). 2 http://www.uva.fi/en/research/groups/management/bilingo/programme/video recordings, meeting documents and a questionnaire.In this article we will focus on the transcripts of the recordings.The transcripts cover 6 hours and 51 minutes of recordings from four meetings by the Regional Co-operation Group held between April 2010 and February 2011.The results of the questionnaire (N=21) will be used for complementing the analysis of the transcripts with the participants' perspective.
As already stated, the meetings we have studied are bilingual: Finnish and Swedish are used in both written and oral modes, case documents are provided in two languages, and the participants may use their own mother tongue during the meeting.As representatives of their background organizations, the participants stand for different language practices, either monolingual (Finnish or Swedish) or bilingual.The meeting thus becomes a site where people from various language backgrounds and with different language skills meet.Although the meeting operates in more than one language, no professional interpretation is provided.In addition, there is little negotiation of language use, and translation is seldom topicalized.One might thus conclude that translating and interpreting are absent from the meeting context.However, this is not the case.Instead, translatoriality is a built-in and taken-for-granted characteristic of the language practices in these meetings.

This article focuses on following research questions:
1. To what extent does the meeting-talk exhibit translatoriality? 2. Who takes on translatorial roles in the absence of professional translators and interpreters?
3. What types of translatorial action are there?
In order to be able to quantify translatoriality, we have divided the data into countable sequences, here called turns.For the meeting-context, we delineate a turn roughly so that it starts when a speaker takes or is given the opportunity to speak and ends when the speaker stops or is interrupted (see Koskela and Pilke 2016).All turns within one selected meeting will be analysed to see to what extent bilingual and monolingual turns exhibit translatoriality.This will allow us to assess the prevalence of translatorial action in contrast to non-translatorial action in either of the two languages.
In the second phase, we will turn our focus on the actors behind translatorial turns.This will reveal which of the participants take active translatorial roles.Third, types of action are identified within the meeting context.In the absence of explicit norms of completeness, exactness or fidelity, translatorial actors have access to a variety of language practices in completing their task, and these will be discussed.
Through these three phases we aim to arrive at a holistic understanding of the translatoriality of bilingual meetings.Finally, the participants' own perceptions on the functioning of the bilingual meeting will be briefly illustrated.
Since our analysis is transcript-based, we can only observe translatoriality as it takes place in face-to-face verbal interaction, but the framework of translatoriality extends beyond the immediate encounter, to encompass roles played by actors outside the group and translatorial activities conducted before or after the actual meeting.We are well aware of the fact that translatoriality is a multidimensional phenomenon, and only one layer of it is accessible with the transcript data.In order to cover the whole range of translatoriality, a combination of micro and macro level studies would be required including e.g.different modalities as well as pre-and post-meeting activities.
Translatoriality always takes place in a particular cultural and political context, with more or less explicit translatorial elements (Holz-Mänttäri 1984, 26).Briefly, language use in these meetings reflects the language situation, policy and legislation of Finland and the specific characteristics of the region in question.Of the whole population of Finland (5.4 million), 5.4 % are registered as Swedish speakers and 90 % as Finnish speakers.Regarding Swedish in Finland, the majority-minority relationship varies regionally.In the region of Ostrobothnia in the West Coast, 50 % of the inhabitants are Swedish-speaking, 45 % Finnish-speaking, and 5 % speak other languages (Ostrobothnia in Numbers 2014).The region thus has a linguistic profile in which the two national languages stand in a symmetrical rather than asymmetrical relation.Nevertheless, although Swedish is the majority language in the region, many administrative and workplace practices are influenced by the national minority position of Swedish.For example, government documents used in formal meetings are translations from Finnish (Pilke and Salminen 2013, 75).

Translatorial action and translatoriality
Justa Holz-Mänttäri's (1984) theory of translatorial action concentrates on professional conduct.However, it also allows for analysing translatoriality in contexts such as in our present data where the role of the translator is taken up by paraprofessional translators.Holz-Mänttäri defines translation as an 'action' (translatorisches Handeln).Actions are, by definition, goal-oriented and purposeful.
When communication needs to cross cultural and linguistic boundaries, translatorial action is needed (82).The function of translatorial action is to produce a messagecarrier designed to enable a transcultural transfer of messages needed for steering purposeful actual and communicative cooperation in a particular context of action (162).
Translatorial action is thus always context-dependent, and its functions, as well as the final format and mode that the carrier takes, are subordinate to the higherlevel functions of the formative actions of that context.In other words, there is little point in expecting one-to-one correspondence between the source text content and form and the target message that has been strategically drafted to fit the particular purpose(s).
In the case we are reporting here, the framework of translatorial action is that of a formal meeting, and it thus follows that translatorial action is expected to support the overarching goals of the meeting, both in the pragmatic terms of enabling the meeting to be opened and closed, discussions to be had and decisions to be made, and in the more abstract terms of forwarding the agenda set for the meeting and the general aims of why it has been reconvened.
In Holz-Mänttäri's model (e.g. 1984, 84) crossing a cultural boundary is a defining feature of translatorial action, and linguistic barriers are seen as secondary although relevant.However, each situation needs to be viewed individually.The bilingual nature of the meeting in our case implies that some crossing of cultural boundaries may take place, and that the action of translation will therefore be needed.Still, the participants are not so much divided by the two linguacultures, Finnish and Swedish, both national languages in Finland and locally used in the region, but by their differing acculturation to either bilingual or monolingual action in general and bilingual or monolingual meeting practices in particular.All participants are, as officials from different municipalities and as representatives of various organisations, well aware of the general framework of a formal meeting, but depending on the linguistic background of their organisation and the organisational culture they have been socialised into, they have different attitudes as well as varied repertoires of activities concerning bilingualism.Their behaviour in terms of roletaking and translatorial action needs to be interpreted against this cultural background (see Holz-Mänttäri 1984, 33).
The concept of translatorial "action" foregrounds the "actors".Holz-Mänttäri (1984, 109) identifies six key roles as follows (see also Schäffner 2011, 157-158 These translatorial roles will be analysed below with respect to the immediate interaction of the participants.In particular, we will focus on observing who takes the role of the translator and how they then act.Each actor when acting in the role of translator engages in translatorial assessment and translatorial decision-making.
This assessment leads to a strategic choice of language practices that carry the meaning from one language to the other so that the goals of the communicative situation are best served.Because the role of translator is not bound to a designated person in the situation, there is no predefined division of roles.Thus in our data, the roles of the source text provider and translator often blend and overlap creating a dynamic interplay of different language practices.
For the purposes of our analysis, translatorial action is defined as all types of action by anyone who takes the role of the translator in a bi-or multilingual situation with the purpose of delivering spoken or written content in one language in somehow related content expressed in another language to receivers with potentially different needs and uses of translation.The bilingual nature of the group's meetings results in a variety of language practices that are employed to enable participation in both languages.We will describe these by modifying the terms used by Reh (2004) in her study of the use of written language in a linguistic landscape.
First, what we call 'a duplicating language practice' involves cases in which content or meanings that have previously been expressed in one language are reproduced in the other in full.Second, 'a summarizing language practice' includes language use in which at least some of the content is summarized in the other language.Third, an 'expanding language practice' describes cases in which some of the content is repeated in the other language but also augmented with additional content.Fourth, 'a complementary language practice' means that there are no overlaps between the contents produced in the two languages, i.e., no translatoriality.In Reh's (2004) original categorization there are two categories, fragmentary multilingualism and overlapping multilingualism, for the language practice where only a part of the information is repeated in the other language(s).
However, in our spoken data such a differentiation between a summary and partial exact translation cannot be made.The classification leads to a discussion of the limits of translatoriality because the degree of fidelity with the 'original' varies, and sometimes it can be difficult to decide whether "relevant similarity" can or cannot be considered to exist between message-carriers.
In the next section, we will first analyse how much translatoriality there is in the meetings (4.1), who takes on translatorial roles (4.2), as well as describe the types of translatorial action in the bilingual meeting context (4.3).These steps of analysis will provide a deeper understanding of translatoriality as a characteristic feature of bilingual meetings.Finally, this understanding is reflected against the participants' perceptions (4.4).

Language practices in bilingual formal meetings
The language profile in the four meetings in our data is relatively balanced between the use of Finnish and Swedish.In the four meetings, 53 % of the time is allotted to Finnish and 47 % to Swedish.In the four meetings, there are 207 monolingual and 188 bilingual turns, in total 395 turns.By bilingual turn we understand a turn including language alternation, i.e. alternate use of Finnish and Swedish.The minimum requirement for a bilingual turn is one word in the other language.The majority of the monolingual turns is in Swedish (69 %), whereas Finnish (54 %) gets more time in bilingual turns.
In order to understand how translatorial action fits into this overall picture, we studied one meeting (Meeting 1) in more detail.The meeting was held in 16 April 2010, it lasted 1 hour 39 minutes and encompasses in total 103 turns of which 41 (ca 40 %) were monolingual and 62 (ca 60 %) bilingual.This meeting was chosen for a closer analysis because it was of average length, it included both presentations and discussion, and it was carried out according to normal routines; i.e. there were no changes to the agenda or other exceptions.

The degree of translatoriality
As the first step of analysing to what extent language practices in the meetings are translatorial, we categorised turns in Meeting 1 into translatorial and nontranslatorial ones.Non-translatorial turns are such that they do not include any obvious signs of repeating something in one language that had already been said in the other.This minimalist definition of translatoriality allows us to focus on translating actors, but it obviously overlooks other forms of action and other kinds of actors involved, for example, source text producers (one could well argue that also those monolingual turns which function as source text to translators are in fact translatorial by nature; this is an aspect we chose to overlook for the sake of simplicity).The analysis was carried out so that each of the three authors analysed the data independently identifying occurrences of translatoriality and the results were then compared.The results were mainly unanimous, but some cases of relevant similarity were negotiated until a consensus was reached.
Concentrating on entire turns does not differentiate what happens within each turn, but nevertheless it gives an overview of how common the practice is.The results of the analysis show that 60 % of the turns were translatorial whereas 40 % of the turns were not.As Table 1 shows, translatorial action in this meeting concentrates on the 62 bilingual turns.Only 3 monolingual turns were categorized as translatorial and only 3 bilingual turns as non-translatorial.In the cases where a monolingual turn is translatorial, it forms a pair with a previous turn that is located either immediately before the translatorial turn or close-by.
The translatorial action within bilingual turns was divided into turn-internal translatoriality, i.e. turns including self-translation of some part of the speaker's own speech, and turn-external translatoriality, standing for situations where something said in previous turns (by someone else, or by the speaker) was reproduced in the other language.There were also turns (marked [3] in Table 1) that included both types of translatoriality.In Table 1, these were counted only as turn-internal (56), but have been visualized in the turn-external category under the subcategory "own" [3].
The results of the analysis showed that turn-internal translatoriality is the most common type in the material as it occurred in 56 out of 103 turns (see Table 1).The high number of turn-internal translatorial actions reflects the strict structuring of the formal meeting.An important part of the language practices in the meeting consists of the chairperson's and experts' turns that deliver basically the same content in two languages.Next, we will take a closer look at who takes on translatorial roles in the absence of professional translators and interpreters.

Actors taking up translatorial roles
Since we are dealing with paraprofessional translation, it is not automatically obvious who, if anyone, will play the role of the translator.In previous research, translators' agency has been defined as their "willingness and ability to act" (Koskinen and Kinnunen 2010, 6).The definition seems apt here: willingness becomes highly relevant as translation is no-one's explicit duty, and although we are not dealing with professional translation practice, successful carrying out of the role requires many of the same abilities.Holz-Mänttäri (1982, 30) gives a list of necessary qualities such as team player skills and flexibility; creative curiosity; wide interest and willingness to learn; analytical textual skills; good language skills in the working languages, particularly in terms of cultural and situational knowledge, and readiness to accept criticism.In the context of bilingual meetings, the translator role requires language resources in both languages as well as an ability to quickly grasp the essential content and to rephrase it in a form suitable for the purposes of the meeting.It also presupposes an active attitude and curiosity towards the issues at hand as well as flexibility since there is often no possibility for advance preparation.
The meeting framework endows the chairperson with some directive power over when, how, and by whom translation takes place.As Table 2 shows, in most cases, the translation is carried out either by the chair or by the experts.
Table 2. Actors taking up the translatorial role.Mostly the other speakers only engage in self-translation, while the chair also translates the other participants' talk.The participants in their turn mainly exercise their right to use their own language only.
In Holz-Mänttäri's (1984, 66) model, translatorial action is not seen as communicative in itself in the sense that the translator is relaying someone else's message without their own need to communicate anything.This is true of professional translation, but does not cover our data.This is illustrated clearly by the predominance of self-translation.
In some cases the participants also negotiate over the practicalities of language use.In the local context of the meeting, any participant can take the role of the "client", for example by signalling the need for a translation or reminding the speakers of the bilingual nature of the meeting as in example 13 .
[Literally: Expert: Does somebody want this in Finnish, yes?
Participant: You could.
Expert: Yes, I thought so.] While the initiator, i.e. the formal bilingual meeting context, remains the same, the source-text producer, the translator and the one presenting the text orally might or might not be the same person (see Holz-Mänttäri 1984, 109).From the point of view of the recipients this can make a difference: reading aloud might be more difficult to follow than spontaneous speech.Also the source-text producer role can and does fall on any of the participants, either in advance or in the discussion.
Similarly, all translations are received by all and used for whatever purpose each participant wishes to use them.

Types of translatorial action
In this section we will discuss the three types of translatorial action that can be identified within the meeting context.The following sections discuss duplicating, summarizing, and expanding described above.Each type is illustrated with examples from our material.

Duplicating
Duplicating, i.e. producing full content in two message-carriers, is regularly used in two types of turns: in those where the chair steers the meeting forward according to the agenda, and in presentations given by experts.
Example 2 includes several excerpts from the course of Meeting 1 that illustrate how the chair takes care of the formalities as he routinely repeats the key phrases in both languages.
(2) a) Chair: Avataan kokous sitten tuota, toivotan tervetulleeksi.Mötet öppnat.Önskar er alla välkomna.[Literally: The meeting is closed.The meeting is closed.] This type of translatorial action brings into forefront the bilingual character of the meeting by means of a balanced use of the two languages, producing a sense of belonging to both language groups and also giving a specific character to a bilingual meeting.However in some phrases it can be asked if the chair really engages in cognitive translatorial action or if these phrases are grammaticalized to the degree that they can best be seen as frozen translations.This type of translatorial action is mostly turn-internal as the chair repeats himself.However, it can also be turnexternal and dialogic.In example 3, the chair and an expert discuss the time for the next meeting.
(3) Chair: Här och klockan?[Chair: At thirteen.Thirteen ou ou] The chair's second turn in example 3 represents turn-external translatoriality because the time was originally mentioned by the expert in Swedish, then translated to Finnish by the chair.In this case, the repetition of the time in Swedish can be categorized as either self-translation or repetition of the previous turn.If translatorial action is defined as reproducing something in the other language that has been said previously in the same context in the other language, producing these phrases will count as translatorial action.But if we assume that translatorial action requires not only willingness and ability but also a specific type of cognitive effort, these are border-line cases.
Duplicating as used by experts takes a different shape as they have chosen to [Literally: Expert: […] At the point when we continue further in this process, then we will take up both key projects and packages of measures.So when we continue forward again to more concrete issues, so in this Regional Programme key projects and packages of measures will be entered in order to reach these goals.And according to the prioritized topics, we have written down focus areas.And the focus areas in the first (strategic) priority are as follows, securing the functional conditions for the strong international industries in Ostrobothnia, a developed innovation system, and strengthened entrepreneurship as well as developing the support for new, innovative and actively expansive companies and so on.So in these prioritized topics we have again taken these focus areas, in this first one the focus areas concern, as you can see, this entrepreneurship of ours, innovation system and these key clusters of ours.In this way we get, we go towards more concrete issues.And finally when we draft this implementation plan, we come to more concrete measures for projects.] In example 4, the stretches in each language are rather short, and the speaker's strategy is to be as faithful to the original as possible.This can be achieved by relying on prepared documents that have been translated by a professional translator and by using the support of PowerPoint.In our material, the extreme case for following this strategy is an expert presentation in which the speaker changes language 44 times within one turn (compared with the average of 4 times per turn in our data).However, the goal of faithfulness can also be reached by using longer stretches before producing the message-carrier in the other language.
Duplicating translatoriality complies with folk notions of good translation, as "everything" is conveyed in the other language.In examples 2 and 3 it functions as an efficient method of inclusion that is dynamically laced into the linguistic fabric of the meeting.However, as example 4 shows, when employed rigidly and in a prepared manner, duplicating may become tedious.This is especially true of long turns, and is aggravated in contexts such as ours where all participants have some ability to follow both languages.As Holz-Mänttäri (e.g.1984, 84) repeatedly emphasizes, successful translatorial action is not a question of repetition but reshaping new message-carriers that best fulfil their expected functions, and professional translators need to be trained to think beyond duplicating.It is thus interesting to see how the paraprofessional translators in our data deal with nonduplicating translatoriality, i.e., summarizing and expanding.

Summarizing
Summarizing, i.e., condensing the gist of a message-carrier in another language, is in our data mainly used in turn-external translatoriality.It can be linked to either one's own talk, or more often somebody else's talk as in the chair's turn in example 5.
[Literally: Expert: Yes, chair, this is a project that concerns a grocery shop, a super market we will get, and the situation today is that since the money for investments has been cut down to zero, so we cannot invest in projects like this without getting financing somewhere else.Prisma will open a large supermarket that will bring about a lot of traffic which is why we need a roundabout in order to get the traffic going, and these plans are ready and the cost estimate is 168000 euro.The financing is such that the municipality will pay half and ELY money the other half.
Chair: So this concerns securing the traffic arrangements of a commercial centre and development of the centre of Luoto, these traffic arrangements, and the lack of money leads to the need of using regional development money for this type of purposes.] Example 5 illustrates how the expert's presentation of an issue in one language activates the chair to engage in summarizing to make sure the Finnish-speaking participants are up to speed.The case concerns a traffic arrangement in a Swedishspeaking area, which is probably why the presentation was only given in Swedish.
Another reason why a summary can be considered to be sufficient is that the decision has been presented in the agenda in both languages.The summary presented by the chair covers the main features of the presentation.Summaries can sometimes be very general, as example 6 from the discussion part of the meeting illustrates.In example 6, the chair prepares for an answer to a statement made by a participant by asking an expert to comment on the issue.The participant presents the question in Swedish, and the chair basically only gives the topic of the issue in Finnish (this condition of the roads and the lack of money).4However brief the summary, by engaging in translatorial action the chair makes it possible for the Finnish-speakers to follow the discussion that was initiated in Swedish.
Summarizing can be used to support translatoriality, but it is also a regular feature of meetings.The chair often needs to sum up a discussion and formulate a conclusion irrespective of the number of languages used.In example 7, the chair summarizes a message carrier that has been presented in both languages in the previous turn.
[Literally: Expert: Yes, chair.About this still that when this round of comments has been completed and when we have made all the changes based on it, then the Co-operation Group will get this back for discussion.So that before this then finally goes to the regional council we will have the possibility to once again discuss this.
Chair: Ok, so could we now note to N's comment that this is in that way still an open issue that we will have the opportunity to comment on this and it comes here for discussion once more, so that this is not in any way final yet.] As is typical of translatorial action, the chair frames his summary linguistically (okei, elikkä=ok, so…) and makes a conclusion of what has been said before in order to round up the discussion.However, it seems that in this case the chair engages routinely in translatorial action where it actually is not necessary.

Expanding
Expanding as the third type of translatorial action challenges the definition of translation.In the context of bilingual formal meetings, the paraprofessional nature of translation does not follow any stated norms of completeness or exactness, and translatorial actors have freedom to choose their preferred type of action.These are often ad-hoc solutions produced spontaneously during the course of the meeting.
The translatorial action of combining translation and comment is inevitably connected with power relationships, especially when the translatorial action concerns the words of others, but sometimes also when it concerns self-translation.
As already stated, it is most often the chair that has the power to add comments in the other language.In example 8, a participant exceptionally chooses to translate his own question to Finnish and expand the message with additional comments.
( As seen in example 8 the participant engages in self-translation with expanding.
The example illustrates how the source-text producer becomes the translator and then returns to the source-text producer role in the other language.In this way, the translatorial roles are interchangeable and translatorial action is dynamically intertwined with the progress of posing the question.
Expanding is also often combined with summarizing.Example 9 shows an illustrative case with respect to the blurred distinction between a complementary language practice and translatorial action as well as of the flexible co-existence of the two languages.
( [Literally: Chair: Ok, other questions?Other questions, comments?If not, we can state that the idea is good, but there is too little money, but we will go on with this, so that the idea is good as such.We do great programs but the resources are a little smaller than last year so it is "as long as it is broad" but.But this is how it is, so the proposal, proposal is that the Co-operation Group states…] Example 9 is taken from the end of a paragraph.First, the chair offers to continue the discussion, and as there are no more comments he then starts summarizing in Finnish the bilingual discussion that has been carried out during the discussion of the paragraph.However, in the middle of sentence he changes to Swedish and repeats his conclusion in the other language (good idea, but little money).After this, the chair continues with his own evaluation in Swedish, which again is continued in Finnish in the middle of a sentence (this is how it is).This is complementary language practice that enables the participants to follow the discussion, but raises the question to which degree language practices need to be translatorial to enable participation for both language groups.It also begs the question of power: professional translation strategies routinely include ideas of explaining and explication, but expanding in ways illustrated by examples 8 and 9 requires a status of authorship rarely accorded to mediating translators and interpreters.

Participants' testimonies
In order to tap meeting participants' own understanding of the language practices we sent out a brief questionnaire to the members of the Regional Co-operation Group.We received 21 answers (43 % Finnish, 57 % Swedish).The results have been reported in full in Pilke and Salminen (2013), and here we will summarize the results concerning translatorial action.
A vast majority of the participants found the linguistic practices of the meetings functional (52 % fully agree, 33 % partly agree).As a negative point, using two languages was felt to prolong the meetings by 57 % of the respondents, who fully or partly agreed with the statement.Interestingly, 42 % of the Swedish-speaking respondents fully disagreed with this while none of the Finnish-speaking respondents did so.The statement that using two languages makes the discussions of issues more versatile again divided the language groups: 92 % of the Swedishspeakers fully or partly agreed, while only 33 % of the Finnish-speakers thought so.
Swedish-speakers thus appear more positively attuned towards bilingual practices than their Finnish-speaking colleagues.
The questionnaire also included an open question on the best and the worst experiences of bilingual meetings.One of the most important positive issues commented on was the possibility of using one's own language.This was felt to increase accuracy and allow everyone to speak irrespective of language skills.From the point of view of translatorial action, this is essential because it increases the demand for translatorial action from the part of other participants.Other positive issues include mentions of practical functionality, of the socially inclusive effect, equality and naturalness of bilingual practices.
Negative issues reported included a repetition and excessive use of time.
Additional problems described were leaving out something relevant in the other language or saying something only in one of the languages.These comments indicate that the participants recognize both the necessity of translatorial practices and the risks involved in each strategy.

Conclusions and suggestions for future work
The purpose of this article has been to analyse the language practices of a formal meeting from the point of view of translatoriality.Even though Holz-Mänttäri's (1984) theory has been designed for describing translation as a professional practice, our analysis illustrates that it has wider applicability and can also help understanding paraprofessional translation as a language practice.The goal of paraprofessional translation in bilingual formal meetings is first and foremost to involve all participants in the discussions and to enable participation.Translatorial action is the primary means of doing this.As our analysis shows, there is variety in the frequency and type of translatorial action, which reflects the institutionalized context of the meeting, the roles of the participants, but also the individual speaker's own skills, preferences and values.
The results of the analysis show that translatoriality is concentrated on bilingual turns.Turn-internal translatoriality is the most common type.The translatorial role is most often taken up by the chair who tends to engage in routine production or self-translation of fixed phrases.Also expert presentations normally include selftranslation that can be realized in various ways.In discussion sections again, there tend to be non-translatorial turns by the participants asking questions and both translatorial and non-translatorial answers to these questions.Only the chair has the power to engage in turn-external translation.In Holz-Mänttäri's (1984, 66) terms, the chair has the power of being communicative even when relaying someone else's message.In the chair's role both ability and willingness to translate are transformed to serve the duty of chairing, and in addition to linguistic skills, the ability to see the relevant in each turn, alertness and language awareness come into play.
In the material, three types of translatorial action were distinguished: duplicating, which is closest to the prototypical idea of translation, summarizing, which is a linguistic practice common in meetings irrespective of language, but is reshaped in the bilingual context so that it becomes a major factor in involving the participants in the meeting, and finally, expanding, which is the practice which allows the translator to add something in the other language to what was said already.Expanding also captures the dynamism of changing of translatorial roles of source-text producer and translator.Of these types, summarizing and expanding illustrate how message-carriers can be dynamically reshaped in order to fulfil their expected functions (see Holz-Mänttäri 1984, 84).
The blurred author-translator roles entailed in self-translation have so far mainly been studied in the context of literary translation (e.g.Hokenson Walsh and Munson 2007), but our data indicates that self-translation may have central functions in many other multilingual contexts as well.As for paraprofessional translation the constant shifts from duplicating to summarizing and expanding, and from one language to another, testify to creative language practices unhindered by preconceived notions of translating.In the case of fansubbing, the research community has already begun to acknowledge the creative and reinvigorating potential of new, unorthodox practices (Pérez-Gonzáles 2012; Secară 2011).This data suggests that, in a similar manner, new viewpoints to institutional translation can be found by observing paraprofessional multilingual practices in institutional contexts, as previously recorded in the case of journalistic translation or "transediting" as exercised by journalists (Stetting 1989).
This study has its limitations.Importantly, we have only categorized types of translatorial action at the level of whole turns.However, as our examples show, there is another level of translatorial action to be discovered within the turns.This forms a challenge for further research as does the use of several semiotic modes for translatorial purposes.Another question that arises from our analysis is how and whether paraprofessional translation of the type studied here differs from both professional translation and entirely non-professional forms of translation.In order to paint a bigger picture of the explanatory power of the concept of translatorial action in multilingual contexts, more in-depth research in all these fields is needed.
In particular, the replication of our analysis with different types of meetings including languages pertaining to different families and representing greater cultural differences would help to shed further light on the dynamic nature of translatoriality.
): a) the initiator, who, for his or her own purposes, needs a translation b) the client, who commissions translatorial action c) the source-text producer d) the translator e) the user of translation f) the recipient, who is the end-user of the translation.
total 50 turns used by the chair (49 % of all turns in the meeting) 20 % (10 turns) are non-translatorial.These are mostly short utterances directly addressing the next speaker when handing over a turn.Most of the chair's turns (34) contain turn-internal translatoriality, and 6 contain turn-external translatoriality.Although the latter number is small, it is worth noting that the chair is the only one engaging in turn-external translatoriality.All in all 32 turns (of 103=31 %) are used by experts presenting issues.The most prominent of the experts is the county chief executive who uses in total 17 turns of which 14 turns are turn-internally translatorial and only three are nontranslatorial.Other experts stand for 15 turns of which 9 are non-translatorial and 6 include turn-internal translation.Of the 21 turns used by other participants only two include turn-internal translation.These numbers reflect the statutory duty of the chair and experts to include all meeting participants and enable their participation.
…] det finns inga pengar och det finns int ens lite pengar heller […] alla anser om det här vägunderhållet och basvägnätet och de här små projekten.[…] Man kommer överens om kostnadsfördelningar för cykelvägar, trafiksäkerhetsarrangemang dom gäller och bas-landsvägar, stamvägar.[…] Att det är ju som man frågar sig om det är värt att räkna upp de här landsvägarna och byvägarna nå mera […] Chair: Jo.Här då kommenterar N.N.Hän kommentoi tätä tuota teiden kuntoa ja vähää rahamäärää.[Literally: Participant: […] there is no money, not even a little money […] everybody has an opinion about keeping up the roads and networks of main roads and these small projects […] We decide on how to divide costs for cycle ways, traffic safety arrangements for highways and main roads […] So it makes one wonder if they are worth listing up these main roads and village roads anymore […] Chair: Yes.Here comments N.N.He comments on this condition of the roads and the lack of money.]

Table 1 .
Number and type of non-translatorial and translatorial turns in one meeting.