The rules of the game: on the interplay between normative ideas and technology in an online amateur translation community

ABSTRACT

In recent years, translation researchers have paid increasing attention both to the social dynamics within translation communities as well as to how members of these communities view their own translation practices and the normative ideas they formulate for their activities. In this article, I adopt a diachronic perspective by focussing on the processes of construction, negotiation and consolidation of normative ideas on translation and community in an online amateur translation community. Based on a virtual ethnography in a self-organised community of amateur translators, I explore how its members navigate often conflicting ideas of what it means to translate and of what kind of community they want to be. I try to retrace how some ideas meet with resistance while others are agreed upon and eventually established as formalised ‘rules’. Drawing on ‘technography’, a framework from science and technology studies, particular focus is laid on the role of the technology behind the community. In this context, the article shows how the technological design of a community can play a vital part in shaping, consolidating and enforcing the normative views of translation held by amateur translators.

KEYWORDS:

• Translation community
• normative ideas
• virtual ethnography
• technology
• amateur translation

1. Introduction

With the growing interest in translation phenomena in the digital sphere, online translation communities have gained increasing attention in translation research in recent years. Scholarly attempts to categorise these phenomena show the diversity of existing organisational models for online translation initiatives (e.g. ‘top-down’ or ‘bottom-up’; Jiménez-Crespo 2017, 28) and the variety of agendas they pursue (e.g. activist, fan-based groups, open source communities, etc.; e.g. O’Hagan 2011). Such translation communities have not only been studied to explore who their members are, what they do and how they organise their work processes (e.g. Jiménez-Crespo 2017; Risku and Dickinson 2009; Salzberg 2008) but also to examine the social dynamics and views of translation that emerge in them (e.g. Jones 2018; Li 2015; Yang 2020; Yu 2017).

In this context, I explore how the members of an amateur translation community negotiate their own, often conflicting, ideas of what translation means for them and what kind of community they want to be. Such normative views have been the subject of several recent studies. Drugan (2011), for example, compares the community guidelines of different amateurs’ translation initiatives with translators’ professional codes of ethics and finds that many online translation communities put the needs of the community over those of the individual member when establishing their own sets of shared rules (Drugan 2011, 117f.).

However, research on normative conceptions in translation communities has not only focussed on what translators perceive as the ideal community and which rules they establish for interaction and collaboration. Normative constructions can also extend to the translation practices themselves. They can include ideas about how translation is supposed to be done in a specific context, how it should be organised in larger groups, whose decisions take priority in which phases of a collaborative translation process, and which translation strategies are considered ‘appropriate’ in a specific context (e.g. Góngora-Goloubintseff 2021; Yu 2017). In a praxeological study of translation and text editing practices on Wikipedia, Góngora-Goloubintseff (2021) explores how translation standards are negotiated among Wikipedia contributors and whether or how users apply these standards in their own activities. Unlike Drugan (2011), Góngora-Goloubintseff (2021) does not restrict his interest to already established translation and editing standards. Instead, he investigates how such guidelines are adapted over time and incorporated into the users’ translation practices, thus demonstrating the value of adopting a diachronic perspective on community translators’ views. As Jones (2018) shows, the negotiation and regulation processes surrounding such community standards are usually not without conflict, and disagreements are frequently fuelled by the specific communicative conditions in online environments.

In a study of the social processes that contribute to the formation of a certain community identity within a Chinese fansubbing group, Li (2015, 232f.) points towards the particular relevance of technical design choices in promoting a specific vision of the community. A micro-level analysis of the particular effects of technology is, however, outside of the scope of her research. In this context, Góngora-Goloubintseff (2021) provides valuable insights into the role of automation processes in shaping translation practices on Wikipedia. Yet while he focusses on the effects of automated devices and the conflicting views on them held by community members, he does not explicitly draw a connection between the more general notions of translation held by the community and how the technical design of the platform (e.g. the user interfaces) might help stabilise such views. I therefore pursue this question in more detail in this article.

This study ties in with previous research on the views of translation and the role of technology in online communities on various counts. It is based on my own PhD thesis (Rogl 2022) on the interplay between technical structures and social practices in an amateur translation community (which is referred to by the pseudonym ‘Translaville’). In this article, I focus on the notions of translation and community that emerge from Translaville over time. While I assume, based on Drugan’s (2011) work, that the members may act with a view to their shared values and give precedence to the needs of the community, I do not emphasise established, formalised sets of norms but rather the processes of negotiation and consolidation of normative ideas over time. Unlike Wikipedia or the fansubbing groups explored in the aforementioned studies, Translaville is not driven by the requirements of a specific field – it is a community that places the very activity of translating centre stage. Accordingly, the first part of this article focusses on how normative ideas of translation and community are negotiated on Translaville and how some of them become more established and formalised. It also takes a look at some of the instances of resistance to such rules. The second part is dedicated to the interplay between normative ideas and the technical structures in which the translation practices on Translaville are embedded. After all, its members are only able to experience and participate in the community’s activities by means of the platform and its design. In this context, I delve deeper into the questions of how normative ideas held by the members of Translaville are inscribed into the technical design and the extent to which technology can contribute to solidifying amateurs’ normative views. Translaville appears particularly promising in this context as its technical design is custom-built by the community itself and technical changes are initiated by members. It can therefore be assumed that there is a clear connection between the members’ ideas and expectations and the technical structures they design for their community activities.

2. The interplay between constructions of meaning and technology

This contribution deals with normative conceptions as part of reflections on meaning-construction and sociotechnical agency. In this sense, I follow recent sociologically-oriented approaches which have been regarded as a response to criticism against notions of norms inspired by descriptive translation studies: the most notable piece of criticism being that they place too much emphasis on rules and restrictions without taking resistance or negotiation processes into account, which in turn entails a blind spot for translators’ agency and for the question of how norms might evolve over time (Yannakopoulou 2008, 3ff.). I focus on normative ideas as articulated by the members of Translaville themselves, either in interviews, when expressing and negotiating their views in the community forum or when establishing shared guidelines. Following Chesterman (2017, 189), ideas are considered normative if they exhibit evidence of normative force, taking the form of (1) belief statements (i.e. members’ justifications for their actions), (2) explicit criticism of unexpected or undesirable behaviour (criticism that is considered justified by the rest of the community), and (3) norm statements (i.e. codified community rules).

The aim of this study is to gain a better understanding of the relationship between translation practices, amateur translators’ views (on translation and their community) and the technology in which their practices are embedded. To do so, I draw on a micro-sociological framework referred to as ‘technography’ that was developed in the field of science and technology studies, most prominently by Rammert (2016; cf. also Rammert and Schubert 2006). The term ‘technography’ might suggest an exclusive focus on a research method. However, while technography does propose a specific methodological approach to ethnographic research into sociotechnical practices, it also provides a theoretical framework on sociotechnical agency that helps to explore how actions are distributed across complex sociotechnical constellations, how cultural or symbolic meanings are inscribed into technologies and what effects are produced when technology acts or repeatedly operates in a specific way.¹

According to Rammert (2016, xii f.), constructions of meaning (e.g. a certain understanding of how to do something) allow actors to carry out practices in the first place. For him, actors have not already fully formed their ideas before they act. Instead, they construct them collectively in their interactions and in dealing with technologies in concrete situations (Rammert 2016, xii f.). Constructing meaning, developing a specific view or observing something is therefore not an abstract state of being but rather a bodily experience reserved to humans because they interact with other people, objects or their natural environment (Rammert 2016, 89f.). In this context, I understand normative ideas as one form of meaning construction. As such, they are part of our wider implicit, practical knowledge which we use to symbolically organise our reality (Reckwitz 2002, 246). When I refer to such ideas as ‘rules’ below, I do so following my ethnographic approach, reconstructing the very notions that are used in the field itself to designate a phenomenon. When members of Translaville speak of ‘translation or community rules’, they usually refer to generally agreed upon, formalised ‘community rules’ (i.e. the above-mentioned ‘norm statements’, Chesterman 2017, 189f.).

Technography focusses on the questions of how technical objects are made sense of in the course of practices (e.g. when creating a new technical design, using a certain tool) or, conversely, how meanings are inscribed into ‘physical points of reference’ (i.e. a specific technical object) (Rammert 2016, 131). More specifically, it explores how and why certain individual meanings gain solidity over time or when spread to other contexts, thus becoming collective patterns of meaning. It assumes that the process of constructing meaning around sociotechnical practices is always partially disrupted by what is referred to as the ‘resistance of things’ (Rammert 2016, 131, my translation). Similar to Pickering’s (1995) concept of the ‘mangle of practice’,² a constant mutual process of adjustment between human planning, meaning making and expectations on the one hand, and the materiality of the physical world on the other, is assumed (Rammert 2016, 131). Technographic studies reject an artificial separation between ‘a cultural world of meaning-making and a technical world of blind rule-following’ (Rammert 2016, 62, my translation): Collective expectations and knowledge are granted the same role in the creation of social reality as technical objects, sign systems, nature and the human body (Rammert 2016, 47f.).

Since (normative) constructions of meaning are usually considered to be only one aspect that informs sociotechnical practices, they are often explored as part of studies on agency or human-machine interaction. In this article, I place particular emphasis on the views and expectations of the platform’s users and subsequently delve into the interplay between normative ideas and technology.

3. Case study and data

Translaville, the amateur translation community selected for this case study, was founded in 2005 and was highly active for more than a decade until it gradually registered less and less activity in recent years. At the time of writing (2022/23), it has around 260,0000 registered members.³ The data analysed for this study (see below) indicate that most members do not have professional translation training or experience. The founder of the group and designer of the platform is a mobile games developer who started working on Translaville as a hobby. The community is a special case in that its translation activities do not revolve around an external agenda or interest in a particular type of source texts, as is the case, for example, in fan-based or activist translation initiatives. Instead, it builds on a principle of mutual exchange of translations: every member who provides translations receives points in return and can redeem these points to ask for translations themselves. There are no restrictions on the content of source texts, which are generally quite short and can range from poems written by the users themselves and everyday texts to love letters or mobile phone app documentation.

To capture the individual and collective experiences of the community members and at the same time allow for a micro-analysis of the technical environment of the platform, I conducted a virtual ethnography (Hine 2015) of Translaville. In this approach, the methodological principles of classical ethnography are applied to the study of digital practices. Hine’s (2015) approach to virtual ethnography complements the theoretical framework used in this research particularly well, as it also places a particular focus on the material and medial aspects of digital practices. Online spaces are conceptualised both as cultures as well as cultural artefacts, which are shaped by people’s ideas and expectations of virtual worlds and interactions (Hine 2015, 30).

Participant observation was crucial in this research design and was not only used to gain an emic perspective of the community practices: For many subsections of the platform, participating as a translator myself was the only way to obtain access to the web interface for all steps of the translation and evaluation process. The observations were conducted over a period of 18 months in 2018 and 2019. My double role as a translator and researcher was made transparent prior to all observations and interactions with other community members. The vast majority of the data stem from the community’s online forum, and for the analysis, I selected the five thematic sections most relevant to my research question. To complement and corroborate the insights from the forum data, I conducted four expert interviews with experienced members of the community who assume different roles on the platform (i.e. regular translator, translation expert or evaluator, administrator, super user). The interview guide (constructed according to Helfferich 2009) included 20 main questions serving as narrative prompts, with a series of optional detail questions. Table 1 shows the data types and amount of data analysed.

Table 1. Data analysed for this study.

All data were analysed using qualitative content analysis according to Schreier (2012). The category system for the coding process was initially based on predefined, deductive categories, which were then continuously expanded and refined in the course of the coding process, i.e. supplemented by inductive categories. The deductive first and second-level categories covered the three main conceptual focal points of the overall research project, namely (1) practices/acting/agency, (2) ideas/expectations/constructions of meaning and (3) technological structures. For this paper, I revisited my analysis of the categories (2) and (3). In order to better manage the large amount of data and the different data sources, data processing and analysis were carried out using the MAXQDA software.

Implementing ethical standards for this research project was challenging for several reasons: High volatility in membership made it impossible to ask all members for consent to use this material. I was in contact with several of the community’s gatekeepers and introduced both myself and my research via the community forum prior to any field observations and interview requests. Although all forum data used in the study are publicly accessible, I took into account that some users might feel exposed by seeing what they considered to be private conversations made available to a wider public. I also considered the members’ online identities to be as worthy of protection as their offline identities. Therefore, I anonymised my data set, removed all personal information, and use the pseudonym Translaville instead of the community’s real name. Direct quotations were slightly altered (in word order, one or two words changed to synonyms) to avoid them being traced to the community by backsearching the quotations in a search engine (cf. Markham 2012).⁴

4. The construction and consolidation of normative ideas on Translaville

In this section, I take a look at the different ideas regarding translation and the community which circulate on Translaville. Before entering into a discussion of what they encompass, I first examine how these ideas are actually constructed and negotiated. In doing so, I try to retrace how some ideas meet with resistance while others are agreed upon and ultimately established as formalised ‘rules’. I also endeavour to identify what can cause an attempt to change a rule at a later date.

4.1 The discursive construction and evolution of normative ideas

Because Translaville is a self-organised group, all members are encouraged to participate in discussions on the design and organisation of the community’s activities irrespective of their experience and role. These discussions take place primarily in the member forum, to which all users have access. In the early years, especially, Translaville members often held lengthy discussions about their ideas of what constitutes a good translation (e.g. criteria such as completeness and correctness), what they want to stand for as a community (e.g. a space for free exchange of translations) and what they expect of each other (e.g. reliability and respectful behaviour). As a result, many of the data examples provided below stem from the period between 2006 and 2008, when many of the practices that members might now take for granted were just forming.

While members were expected to base their translations on the principles they had more or less agreed upon in previous public discussions in this initial phase (as evidenced by discussions posted under individual translations), there was soon a call for greater harmonisation:

As becomes clear from this extract, these shared ‘rules’ represented more than just an attempt to harmonise the evaluators’ approach to providing feedback and offer translators an idea of what is expected of them. They also served as a way to save the community the trouble of having to renegotiate basic ideas of how to translate for every target text.

In the following example, which dates from a little more than a year after the previous example, the community members appear to have started collecting and wording these ‘rules’. As can be seen in the post, the community made an effort to negotiate these rules publicly and word them with a view to accommodating different expectations.

In this way, a formalised set of shared norms was established with the help of Translaville’s forum and eventually published in a Wiki. Interestingly, the ‘rules’ on this list were by no means the only significant ones that translators and evaluators seemed to apply to their practices and decisions. As the data analysis revealed, informal but still widely accepted norms also continued to exist, even after multiple revisions of the ‘translation rules’. Apparently only those normative ideas that offered the least margin for interpretation – and were thus considered easier to apply in practice – were ultimately put into written form. Conversely, more complex ideas scarcely found their way into the formalised set of ‘rules’ (e.g. a harmonised approach to applying the rating scale for translation evaluation, which source texts to reject for ethical reasons; see discussion below).

Table 2 provides an overview of the main normative views that were collectively expressed in the data. It includes ideas that most members seemed to agree upon but which were not formalised (light grey shading) as well as those which at some point became part of the written guidelines (dark grey shading).

Table 2. Formal and informal norms on Translaville.

While most of the formalised ‘rules’ for submitting source texts and producing translations concern product-related criteria, process- and community-related conceptions are also prominent in the data set, even if they were less likely to be included in the official ‘community rules’. As can be seen in Table 2, most normative views on translation products concerned constructions of translation quality, in particular textual criteria (e.g. the correct use of diacritics, grammar, completeness). Interestingly, not using diacritics or accents seemed to be regarded as far more serious than orthographic errors because omitting diacritics was attributed to a lack of care, while orthographic problems were interpreted as a lack of linguistic competence, which was more readily excused. The members’ views on what they consider a good translation also contained ideas on the aim of a translation (i.e. to convey the meaning, to create a target text with the same effect as the original, a certain idea of target reader orientation and an expectation to meet the author’s intentions for the translation). There was one additional product-related idea, namely that there should be a relatively stable correspondence between the lengths of the source and target texts in every language combination (irrespective of the text type or length and aim of the translation). I take this idea up again in section 4.2 to discuss how certain ideas are inscribed in technical structures and can thus help restructure practices.

Other than product-related expectations, the data indicate a variety of ideas that relate to the translation processes on Translaville. These concerned, among other things, the tools that community members were encouraged to use (e.g. online dictionaries), those they were prohibited from using (e.g. MT) as well as the main goal of the evaluation process (i.e. a focus on constructive feedback to support language learners). Particularly interesting in this context is the fact that, in discussing these views of translation, the members of Translaville seemed to construct a certain idea of commitment or duty to the community, which proved to be a recurring theme both in individual user accounts, where they stated what they expected of themselves, and in collective constructions, e.g. the community’s FAQs. There appeared to be a palpable sense of obligation to finish the translations in a ‘reasonable time’, to be present regularly so as not to ‘leave others hanging’ (with some members even appointing people to deputise for them while on holiday) and to participate actively in order to contribute to a vibrant community culture.

In parallel to Drugan’s (2011) insights, community-related ideas seemed to be accorded particularly high priority. These included expectations towards fellow community members’ personal qualities or views of how to communicate and interact with each other. Lack of language skills was often readily excused if the translators in question showed a certain willingness to accept feedback, learn from their mistakes, participate actively and generally be helpful and support others. Community-related ideas also concerned members’ perceptions on where to draw the line on misconduct and abuse. Their understanding of what constitutes an abuse of Translaville related mainly to three areas of contention: (1) a waste of community members’ time (e.g. by requesting ‘pointless translations’ or exhibiting destructive behaviour); (2) a lack of fairness (e.g. no contribution to the community activities or a registration of multiple accounts to be able to request more free translations); and (3) the upload of source texts that are considered inappropriate. The data include a series of intriguing debates on morally ambiguous source texts, with rather heated discussions on what to do with texts that contain misogynist, homophobic, racist, pornographic or vulgar content. These discussions focussed primarily on whether the community should: a) translate such texts, because requesters should have the right to know their content, hateful as it may be, or b) delete them to avoid spreading content considered unacceptable to a broader public. The debate did not reach a more concrete conclusion than to leave the decision on whether to translate such texts to the discretion of the individual members and to ban texts that curtail human dignity.⁵

Translaville actively encourages what Drugan (2011, 118) refers to as ‘community policing’. As the group grew in size, issues relating to what is considered abuse in the community also seemed to be on the rise (cf. Rogl 2016 for a more in-depth discussion). On Translaville, nobody is able to keep the whole translation process in view for any given translation. There is often little context or information on the source texts, and the different members involved in the translation process (e.g. the submitter of a source text, the translator, the evaluator) do not necessarily have to interact with one another. The community’s response to the sometimes opaque structures, interactions and requests that will likely be encountered in any virtual space of this size thus seemed to be to assign responsibility to assist administrators with sanctioning violations of the shared rules to each individual member.

What will not become visible when an analysis is restricted to established sets of community rules is the fact that the normative views discussed above – even those published in a Wiki and thus formalised – are in no way stable. The forum discussions reveal that many of the shared ‘rules’ were subject to changes or adaptations at one point or another, especially within the first 10 years after Translaville was first launched. A recurring reason for such adaptations was the realisation that a rule was not as self-explanatory as expected. One such clarification concerned the terms used to designate the language into and from which users can translate (used both in the registration form and in passages on expectations of language proficiency in the ‘translation rules’). After a forum debate in which several members complained about low quality translations, the term for the language into which users are allowed to translate was changed from ‘main language’ to ‘native language’ to stress that this ‘is supposed to be [their] strongest language’. Another change was made for working languages other than the user’s first language. While the registration form initially asked users to specify the ‘list of languages [they] can read and understand’, the wording of this passage was later changed to ‘languages [users] can read and understand at least 75% at first reading it’. It is, however, debatable whether this solution really managed to clarify the community’s expectations of users’ language proficiency.

In other cases, already formalised rules were challenged because users found them restrictive or inappropriate. Figure 1 is a screenshot of ‘Rule No. 4’, which concerns the rejection of source texts that only contain isolated words or incomplete sentences.

Figure 1. Translation rule no. 4.

In the forum post below, user B draws attention to the fact that a text can indeed be comprehensible without a verb. He tries to demonstrate this by also phrasing his post in a way that does not contain a conjugated verb. He argues for the fostering of an approach that allows translators to exercise their own judgement of ‘the translatability’ of texts rather than forcing them to blindly follow rules that may not be appropriate in all contexts.

The next example (Figure 2) concerns the prohibition of machine translation, which is phrased in absolute terms in the community guidelines.

Figure 2. Translation rule prohibiting the use of machine translation.

A user who claims to have professional knowledge in the development of machine translation systems suggests that this rule be reworded to ask translators to use MT responsibly:

The founder of the group defends the prohibitive nature of the wording by referring to a need for the simplest possible rules for beginners, while also conceding that it is tacitly accepted that experienced members flout some of these rules. The same reasoning may have been behind the decision not to modify the prohibition of incomplete sentences (see above). The freedom to disregard rules seems to become more acceptable with a user’s increasing experience on the site (or probably also higher status due to longer membership in the community). The example also shows that formalised rules ultimately tend to oversimplify the actual issues at stake. The community thus seems to actively avoid establishing formal rules for the more complex issues with which members struggle. Interestingly, technological innovations such as more recent generations of machine translation have not prompted a modification of the ban of MT in more recent years. Instead, easy access to quick translation on the web seems to have led to an overall decrease in translation requests on Translaville, which in turn made many members leave the community. The members who stayed on Translaville, however, strongly insist on their self-perception of providing human, ‘slow translation’.

4.2. Normative ideas and technology

The second strand of inquiry into normative ideas on Translaville focusses on the role that technology can assume in shaping, consolidating and possibly even enforcing the normative views of translation held by amateur translators.

The interplay between community norms and technology becomes particularly noticeable in what happens if a user ignores the community rules. Translaville offers a range of possibilities for sanctioning perceived misbehaviour. Administrators can issue verbal warnings, reject source or target texts, or ban members from the site. One interviewee, an administrator himself, confirmed that he did indeed make use of the whole range of these measures (‘I did the work of a cop’). At the same time, the platform is designed in a way that prevents users from taking certain actions in the first place. It can be set to automatically block someone from working with certain languages and is also able to automatically detect multiple accounts or ban certain IP addresses. The range of possible actions can be placed on a continuum from more communicative to more technologically assisted or even automated ways of sanctioning. As forum discussions about specific incidents (e.g. spammers or users who repeatedly try to obtain translations without contributing themselves) depict, the more serious the community considered a violation of a community norm, the more willing it was, in most cases, to use a so-called ‘technological fix’ (cf. Johnston 2018) to deal with it. From a technographic perspective, this can be described as a redistribution of agency towards technical structures.

However, technography draws our attention not only to the question of how agency is distributed between less and more technologically assisted courses of action, it also delves into the consequences of automating sequences of action. Automating sanctions against users certainly eased the workload of Translaville administrators. But, automated solutions are also harder to reverse (as only the developer of the platform has the authority to implement major design modifications) and make it impossible for administrators to decide on a case-by-case basis how they want to deal with a problem. By creating an interface that privileges preferred actions and by implementing automated sanctions for unwanted behaviour, specific normative views were inscribed into the user interface. Because technical structures tend to be less easily reversible than an agreed-upon rule, these views and agendas also acquired greater solidity in the community than views that were only communicated as guidelines.

In other words, some of these normative views were fixed in technical form. As such, these technical structures (e.g. platform elements, user interfaces) can also help to (re)structure translation practices. This becomes visible in the idea (see section 4.1) that something must be wrong with a translation if it deviates by more than a small margin from the established correspondence in length between source and target texts. Figure 3 shows an extract from Translaville’s list of 74 languages and their automatically calculated ‘equivalence ratios’, which indicate how many characters a translation from a 100-character English source text should have in the respective target language.

Figure 3. Automatically calculated length equivalence ratios.

The community developers made use of this idea to create design features that prevent users from performing a series of unwanted actions. For instance, the system automatically detects if users add too little or too much information to the input field reserved for translations (e.g. if they include more than one translation solution or provide comments). If this is the case, an error message pops up (Figure 4).

Figure 4. Error message for not meeting the calculated length correspondence.

The translators can override this message (which requires several clicks) and upload their text anyway. However, as the forum data show, this can still mean a disruption to their translation process:

Because it requires more time and effort to translate in a way that defies this ‘technically privileged’ form of action, many of the translators try to keep their translations as close to the specified length correspondence as possible. As a consequence, the platform design ends up favouring translation practices that confirm rather than challenge a rather conservative notion of translation.

The users of Translaville do not, however, always accept and accommodate technological constraints without objection. Instead, they sometimes find creative workarounds to such constraints through what Rammert (2016, 43, 60) would refer to as a ‘creative or playful appropriation’ of technology. For example, the data show that some translators found it restrictive to have an automatic time limit for every translation (see Figure 5). When a translator starts to work on a source text, it is reserved for them for a certain period of time in order to ensure there is always only one translation for each source text.

Figure 5. Warning message on the time limit for an ongoing translation.

Many community members commented that they often did not have enough time to work on a translation. The forum discussion below shows how one user came up with a way to trick this system:

The existence of such workarounds indicates that translators find creative ways of appropriating the technological environment in which their practices are embedded. This example illustrates what I discussed above as a theoretical assumption, namely that there is a constant interplay between resistance and adaptation (Rammert 2016, 31).

5. Conclusions

In this article, I explore how the members of an amateur translation community construct, negotiate and sanction their own normative ideas of what translation or a translation community ‘ought to be’. For them, ‘good’ translation is, above all, a certain moral obligation towards others – in the sense that they want to provide translations in a reasonable time and not fail their fellow community members. These findings coincide with Drugan’s (2011) conclusions that users place a special focus on respectful behaviour in the community. Success is seen more in the success of the community rather than the fulfilment of individual goals.

My attempt to retrace the trajectory of normative ideas on Translaville from the moment they are first brought to the debate, and later re-negotiated, to the point when some of them are formalised in shared guidelines and, in some cases, ultimately inscribed into technological form, reveals that such ideas are neither stable nor remain unchallenged. At the same time, many widely agreed upon but more complex normative ideas (e.g. a harmonised approach to applying the rating scale for translation evaluation) were deliberately never formalised because the community could not find a way to break them down into simple, easy-to-follow ‘rules’. This means that members are left without much guidance in those areas where they probably need it most, while the ‘rules’ that do exist are seen to be disregarded by more longstanding community members as they see fit.

The findings also show that some of the central constructions of meaning in the community are inscribed into technical form over time in order to favour desirable forms of use and restrict others. This makes them more durable and stable than normative ideas that are merely verbalised or fixed in writing. However, it appears that (albeit well-intentioned) technical solutions can produce side effects that have the potential to considerably restructure or disrupt translation practices.

The translators appear to deal with such restrictions in different ways. In some cases, they adapt their work processes to the technical design, thus renouncing part of their agency. Others find creative, resistant new forms of appropriating technical structures and working around what they perceive as constraints. Some take such overly restrictive technical structures as motivation to ask for design adaptations, which may again produce unexpected effects. Following technography, this process can be described as an ongoing redistribution of agency. However, it not only seems relevant how agency is distributed in socio-technical systems, but the abovementioned findings also raise the question of who can ultimately influence this distribution of agency. On Translaville, only the developer of the platform has the authority to implement design changes (albeit mostly at the suggestion of the community), and only a small circle of tech-savvy members in specific user roles is involved in the technical maintenance of the platform. This creates a certain imbalance of power between the ‘regular’ translators and members with access to technical design tasks. To a certain degree, technology thus serves to lock in social relations and to perpetuate power relations in the community.

Rammert’s (2016) central idea of a constant back and forth between processes of accommodation and adaptation with respect to human sense-making and materiality seems to provide an apt possibility for conceptualising the interplay between normative ideas and technology in digital community translation practices.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Regina Rogl

Regina Rogl holds a PhD in translation studies/transcultural communication and is currently working as a post-doc researcher at the University of Vienna. She is a member of the research group SocoTrans (Socio-Cognitive Translation Studies: Processes and Network) and an associate researcher on the third-party funded research project Rethinking Translation Expertise: A Workplace Study (RETREX), led by Hanna Risku. Her research interests include digital translation practices, socio-technical conceptualisations of translation, non-professional translation/interpreting, and workplace research.

Notes

1. The framework ties in with actor-network theory (ANT) by adopting its central tenet of the methodological symmetry of human and non-human agency. However, it adds to ANT in that it not only distinguishes agency and non-agency, but different degrees of agency. In contrast to the approach of the social construction of technology (SCOT), technography does not only focus on the stage of the development and subsequent adaptation of technologies, but brings the development and use of technologies (and notably also the autonomous agency of technology) together in one analytical framework, trying to retrace socio-technical practices on a micro-level (Rammert 2016, 44ff.).

2. See Olohan (2011) for an application to translation studies.

3. This is a cumulative count. Since inactive profiles are not automatically removed from the platform, it is impossible to determine how many users are still active at any given point in time.

4. The proposal for the PhD thesis this article is based on was approved by an interdisciplinary faculty committee. A dedicated ethical clearance was not required by my university at the time.

5. The latter was decided after an administrator called attention to the French Youth Protection Act, since Translaville is registered in France.