The present study adopts the approach of social constructionism, which maintains that knowledge is something “historically and culturally relative,” and is constructed through “the goings-on between people in the course of their everyday lives” (Burr 2003: 4). In the context of translation, the principal stance of the present study is the belief that “[t]heory is understanding and explanation, and not only ‘something there and established’” (Susam-Sarajeva 2002: 204). In other words, translation theory is not conceived as owned by academics and students, but as something shared by wider participants in translation processes, including translators, who “are theorizing all the time” even though, in most cases, internally and privately (Pym 2010: 1). Theories express the ways of thinking about translation and its practices as told by anybody who is involved in translation. Therefore, interviews were chosen as the method of the study.

The main interview question was “Have your clients ever had disagreements/issues with the quality of your translation?” This question was chosen because disruption in everyday life, such as having a disagreement with a client, encourages us to try to restore order, and narrative is “a primary means of restoring this sense of order” (Murray 2003: 114). In other words, the topic of disruption was used as a tool to elicit the interviewees’ explanation of their sense of order about translation and I understand these explanations as theorizing.

My role as an interviewer was also important. As the interviewees knew my position as a former professional translator who was conducting academic research on translators, I assumed that they would construct their stories in such a way that the stories were understandable and convincing for me. This is what Gubrium and Holstein (2009: 41) call “activation” of a narrative, where “the facts of experience are locally configured as storytellers and listeners actively take part in discursive exchanges.” I assumed that this activation would work favourably in this study because in this interview setting the interviewees would give their accounts in the most logical and convincing way. In addition, I possessed what Collins (2007: 30-33) calls “interactional expertise” in translation, which is a kind of expertise that encompasses both the language used in the domain and expertise gained through practical experiences. According to Collins this form of interviews in research elicits rich data.

Seventeen professional translators took part in the study. Fifteen were freelance translators and two were in-house translators. All of them translate between English and Japanese, into either or both directions. This language combination was chosen because it is my own language combination, enabling me to undertake a detailed investigation of the participants’ accounts of linguistic matters. The main criterion for selection of interviewees was whether the person was much in demand in the current translation market (i.e., whether they were successful translators in a commercial sense). This was measured in the case of the freelance translators by the average number of words they translate for remuneration every week (on average 9,920 English source text words or the equivalent number of Japanese words, equating one English word to 1.5 Japanese characters, ranging from 5,000 to 16,700 words) and how many enquiries about availability they receive from existing or prospective clients (on average seven a week, ranging from 1 to 20). Both in-house translators were full-time employees who translated for more than five hours a day.

Fourteen interviews were conducted face-to-face and three via Skype. Each interview lasted between 10 minutes and 2 hours, and the total recording time was 957 minutes.

As mentioned above, the main interview question was “Have your clients ever had disagreements/issues with the quality of your translation?” Once the participants started telling their stories, they were encouraged to give details through questions such as “What did your clients say to you?” and “What did you say to them?” These follow-up questions did not aim to elicit true verbal recall, as verbatim memory is notoriously unreliable (Hjelmquist 1984). Instead the aims of the questions were to encourage the interviewees to elaborate on their stories so that more data would be generated. The participants were allowed to present as many stories as they wished. As a result 93 stories were collected from the 17 participants. The story unit used for analysis was one complete episode of an incident in which the translator had a conflict with his/her client, discussion and negotiation took place, and the dispute came to a closure. Each story was labelled with the participant number and story number i.e., Story 01-01 is the first story produced by Interviewee No. 1. The interviews were recorded, transcribed and analysed in the language used (English or Japanese). I translated the Japanese examples for a reporting purpose and this is indicated by [J].

The concept of the Role of Participants is by far the most notable concept used by the participants (198 passages in 51 stories). The participants who were discussed in the interviewees’ accounts include: the translator; the translation agency (and its project manager); the client (this sometimes corresponds to a translation agency and other times to an end-client who commissions a translation agency to translate a document); the direct client (so called when the commissioner of a translation does not use a translation agency); the user of translation; the proof reader; the editor and the source text author.

The reason why the concept of the Role of Participants was by far the most popular is that the concept was regularly used in conjunction with other concepts. For example, in Story 01-01, the interviewee said that the translation agency gave him a Source Text which was not typed clearly and, as a result, he mistranslated.

[…] the original was not very legible^[1]. There was one place […] where it said this thing needs to be […] tightened but I misread the character as it needs to be loosened. Shimeru (tighten) and yurumeru (loosen), if the character is blurred, it’s easy to be mistaken. And so I translated it in the wrong way. Their client spotted the error […] and they complained to me about this. And I said, “Well fine, give me a better copy next time^[2].”

From Story 01-01

In the early part of the passage (the underlined segment [1]), Interviewee 01 attributed his mistake to the poor quality of the source text. Thus, this segment was coded at the justification method ‘Factors which affect features of translation’ and the concept Source Text. But in underlined segment [2], he developed this concept into an aspect of an action (or here, rather, the lack of it) by the translation agency, saying that they did not supply a clear copy of the Source Text to him. Interviewee 01 said he had said to the agency, “Well fine, give me a better copy next time.” One would not assume that these are the exact words he used, but this utterance can be interpreted as showing that the interviewee attributed his mistranslation to the fact that he did not receive a better copy of the Source Text, i.e., the agency did not fulfil their role.

In Story 02-01, told by a technical translator, the end client did not supply any terminology list for the translation of a text about a clinical data input system so the translator used the terminology of her choice in her translation. After delivering the translation, the translation company contacted her.

They told me how the end client amended the translation and asked me “what do you think about this?” So I thought, “Why are you asking ME?” […] Whether to use “screening failure” or “screening slip” is what the end client needs to decide^[1] . They did not supply me with a terminology list at the beginning^[2] so how can I comment on their amendment?

From Story 02-01 [J] (The capital letters indicate an emphasis in the speech.)

Here, Interviewee 02 used the concept of Authority and Terminology (the client should decide which term is to be used) to justify her translation while explaining the factors which affected the features of translation (underlined segment [1]). Then she developed the concept further into the concept of the Role of Participants by saying that the end client did not do what she thinks they should have done (underlined segment [2]).

The interviewees also talked about their own roles, mostly by claiming that other participants had wrong ideas about them.

I had to go through these the next MORNING and I had no idea of the context until I saw these^[1] and I thought this is SO MUCH OF THIS. I can’t do this in four hours^[2]. But I went anyway and I said, “Well look, I’ll do what I can for you.” And then I was there for two hours looking at these documents and, after two hours the client, the solicitor, said, “I’m sorry. This is too slow.^[3] You are not making any progress.”

From Story 04-01

In this account, Interviewee 04 talked about his experience of being called into a solicitor’s office to translate Japanese documents into English. Here, again, Interviewee 04 justified his translation by explaining what factors affected the features of translation (insufficient time for the job) and the translator’s background knowledge about the particular topic of the translation (legal issues) was limited due to insufficient time for preparation. Here he used the concepts of Time and Effort and Context (the underlined sections [1] and [2]) and then he said the client had a wrong perception of what the translator is capable of ([3]), which was coded as Role of Participants.

Probably because of the interview question (asking about conflict situations), the concept of Role of Participants was most often used in a negative sense, i.e., for explaining that other participants did not fulfil their roles properly. However, the interviewees sometimes used the concept in a positive sense, commending how other participants fulfilled their roles and explaining how that facilitated the translators’ producing good quality translation. In the examples below, the translators appreciate the way the project managers maintained their relationship with their clients, which they assume is the role project managers ought to play. Therefore, those examples were coded under both the concept of Relationship and the concept Role of Participants.

The project manager was very articulate. When I pointed out some problems, she asked me to give her some specific pieces of information like this and this and this, and once the information is supplied, she sent me an email asking me, “I’m going to send this email to the client. Are you happy with this?”

From Story 02-0 [J]

The agency totally accepted what I said. They said, “We understand now. We’ll send your comments back to the client and see how they react.” So I think the agency forwarded my report directly back to the client.

From Story 10-1 [J]

Judging from its quantitative prominence, the concept of Role of Participants situates itself at a higher level than other concepts. In other words, in translators’ discourse, Role of Participants is a higher-order concept to which other narrower concepts (or sub-concepts) relate. The link between Role of Participants and other concepts was measured quantitatively by counting the number of overlaps of coding (how many segments of texts were coded by the concept of Role of Participants along with other sub-concepts). The top ten concepts are shown in Table 2.

What follows is a description of how these sub-concepts relate to the concept of the Role of Participants.

Section 2 described prominent classroom discourses (theories) and Section 3 explained how commercially successful translators theorise their translation. When these two discourses are compared, three points come to our attention: two concerning similarity between them, and one concerning dissimilarity.

First, the translators’ discourse presents a clear similarity with the principles of the theory of translatorial action (henceforth TTA) developed by Holz-Mänttäri (1984) in that they both focus on the actions and responsibilities of participants in translation, rather than on the products of translation, claiming that, in order for a successful production of translation, participants involved in the translation need to “agree (explicitly or implicitly) on who is responsible for which tasks” (Schäffner 2011: 159). Although TTA has been criticised for presenting an ideal translational situation, not a description of a real situation, and hence as lacking in empirical evidence (Schäffner 2011: 160), the translators’ discourse identified in the present study can contribute such empirical evidence. In the educational setting TTA is being taught by 27.8% of translation teachers, being the 19^th (out of 39) most popular theory in the classrooms (Sakamoto 2013: 54). Although TTA is not one of the most prominent theories according to my survey, it can be assumed that, because of its nature, the framework of TTA is actually being used more often in teaching, not in the form of an explicit theory, but as a tacit framework for more practical teaching of translation such as role-playing exercises, where students play different roles in the translation process such as translator, proofreader or project manager, to simulate a professional scenario. Role-play exercises are now one of the “standard methods or techniques used with small groups, … being the most frequent or most useful in translator training” (Kelly 2005: 101). Therefore, it would be reasonable to deduce that TTA serves as a conceptual contact point between the subcultures of professional translators and scholars and teachers in current translation education in the UK.

Second, the fact that the concept of Natural versus Literal Translation was the second most prominent concept in the translators’ discourse echoes the survey result that a majority of translation teachers teach this dichotomy, though under different labels (87.5% as “domesticating vs. foreignising translation,” 83.3% as “dynamic vs. formal equivalence,” 58.3% as “semantic and communicative translation,” 41.7% as “overt vs. covert translation” and 37.5% as “documentary vs. instrumental translation”) (Sakamoto 2013: 54). This suggests that this dichotomy of Natural versus Literal Translation is shared by both translation teachers and practicing translators, occupying an important place in their respective discourses. However, my analysis also suggests that the way the translators explained their practice is rather complex. For example, a factor such as Authority (how much authority the translators feel they have in the translation project in relation to other translation participants) seems to affect their translation decisions about the degree of naturalness/literalness of the textual features of translation. If such external factors are affecting the textual features of translation in the real world, it would be important to recognise the relationships between the factors and textual features and reflect them in classroom teaching. The concepts identified in the present study will offer a list of such factors.

Third, I would like to go beyond the comparison of discourse between professional translators and academics and teachers for a moment and consider the translators’ discourse in the context of recent industry trends, which seems to be particularly revealing, and also alarming. The translators’ discourse presents a set of norms which they believe should be in operation in professional environments in order for them to do a good job. The norms are particularly related to the role of the project manager of a translation company or their client, which can be summarised to be: 1) proper provision of work material, 2) provision of clear instruction; 3) appropriate personnel and time management; 4) implementation of adequate quality control (QC) systems; 5) maintenance of necessary communication between all parties (including translators and end-clients). With regard to their own roles, the interviewees often talked about how their roles were misunderstood by other participants in translation. The misunderstandings include: i) A translator can translate very quickly; ii) A translator is willing to work for small remuneration; iii) A translator understands what kind of translation the client wants without proper instructions; iv) A translator can produce a translation even if the ST is not in order. Removing these misunderstandings would produce their ideal working conditions, e.g., translators should be given sufficient time to produce good quality translation, etc. When this set of ‘ideal’ conditions presented by the translators is compared with what seems to be the recent trend in the industry, some discrepancies are notable. For instance, my interviewees stressed the importance of good relationships between the participants in translation for the production of successful translations. Translation (at least non-literary translation) is now increasingly undertaken by a virtual team (Rodríguez-Castro 2013; Stoeller 2011), where the translator is located remotely from other participants and only connected virtually, for instance by email. In these conditions, shaping good personal relationships with other participants may be an unrealistic wish on the part of translators. The advent of the paid crowdsourcing business model has also had significant impacts on the translation industry (Garcia 2015), including on communication between participants in the translation process. Where this new model is used, translator bids for translation assignments via the translation company’s Intranet interface. If the bid is successful, they are commissioned to undertake the assignment, again via the Intranet interface. In this digital environment, even an exchange of email is unnecessary, and it is difficult for translators to establish good personal relationships with the project manager or the client in this environment. The issue of speed is another area to note. My interviewees often used the concept of Time and Effort as a factor which influences the textual features of translation. However, the demand for quick turnaround of translation is extremely high in the translation industry currently, particularly given the increasing use of machine translation. How can translators secure the time they think is necessary to do a good job in this increasingly speed-conscious working environment?

There is an emerging body of literature which highlights this gap between translators’ ideal working conditions and actual professional environments. This work covers, for example, Danish translators’ frustration about the lack of recognition of their expertise by the general public (Dam and Zethsen 2008), the importance of the physical location of translators and their contacts with other participants (e.g., readers) for their self-esteem and professional pride in the context of Finnish EU translators (Koskinen 2008), and Finnish freelance audiovisual translators’ lack of agency when they are forced to play roles assigned to them by their employers (Abdallah 2011). The research foci of these studies overlap with some of the concepts identified in the present article, such as Knowledge of Language (like the Danish translators, my translators feel that other participants sometimes do not acknowledge that translation requires more than just good knowledge of language), Relationship (like Koskinen’s EU translators, the remote location of the translators from other participants and the lack of close and cooperative Relationships between participants adversely affect the self-esteem of translators) and Time and Effort and Authority (like Abdallah’s audiovisual translators in Finland, my translators showed a passive attitude about having Authority in translation projects while claiming Authority where they think natural-sounding translation is suitable for the target text. They also feel that they are sometimes under pressure to work in an unfairly short timeframe).

These studies situate themselves in sociological translation studies, dealing with such things as agency of actors in translation or the network of the actors. In my survey of translation teachers, only 10% said they explicitly teach sociological translation theories (eight out of 93 teachers said so with seven of them mentioning Bourdieu and two mentioning Latour) (Sakamoto 2013: 54-55). Of course sociological notions have been discussed in more traditional translation discourse such as in norm theory of translation (e.g., Chesterman 2000; Toury 1995), but the more recent sociological turn in translation studies has shed light on translators and, importantly, other actors in the translation process more explicitly. The translation industry is going through rapid changes, and in this climate, I agree, like Jääskeläinen, Kujamäki, et al. (2011: 152), with Bernardini’s opinion that the core aim of translation education is to “favour the growth of the individual, developing her cognitive capacities, and those attitudes and predispositions that will put her in a position to cope with the most varying (professional) situations” (Bernardini 2004: 20; my emphasis). Considering how my interviewees, who are commercially successful translators, theorised their translation, more explicitly sociological-oriented theories may deserve more attention in classroom discourse so that the teaching can not only prepare students for the recent trends of the industry, but also to nurture their capacities and attitudes to cope with any future changes in their professional life. Sociological theories of translation can play an important role as a meeting point for the professional world and academia.