On categorising online collaborative translation and the consequences for the field of research

ABSTRACT

Translation Studies possesses a plethora of meta-concepts such as ‘online collaborative translation’, ‘community translation’, ‘volunteer translation’, etc. to refer to phenomena like translation crowdsourcing, fansubbing, fandubbing, etc. The existence of these various meta-concepts reflects an obvious need to categorise and subsume such phenomena under a meta-category. The paper will attempt to explain why this is the case, why categorising in and of itself is vital and how it relates to boundaries and undertaking boundary-work, both crucial processes in academia. The paper will advocate, based on boundary-work, for using ‘online collaborative translation’ as a meta-category and conclude by presenting a conceptual map of this meta-category and its various sub-categorisations.

KEYWORDS:

• Online collaborative translation
• categorisation
• boundary-work
• meta-concepts
• conceptual map

1. Introduction

There is a pressing need to categorise phenomena like translation crowdsourcing (e.g. by Facebook¹ or Twitter² or in the form of commercial translation platforms), self-managed and unsolicited online collaborative translations (e.g. Wikipedia-translation), and the various types of online fan translations. This need is expressed in the profusion of meta-concepts in Translation Studies literature for these phenomena and the associated explanations and demarcations. Apart from ‘online collaborative translation’, other meta-concepts have emerged such as ‘user-generated translation (UGT)’, ‘voluntary translation’ and ‘(online) social translation’, to name but a few (e.g. Jiménez-Crespo 2017; McDonough Dolmaya and Sánchez Ramos 2019; Zwischenberger 2022).

This need to categorise stems partly from the topological view of the world we are socialised into and partly from our deep yearning to impose order onto otherwise boundless and undifferentiated environments. Categorising is, thus, a deeply human affair. Only by categorising and thereby carving objects out of their environment can we render them visible. This in turn allows us to more easily pick out single objects or groups thereof and really focus our attention on them. Otherwise, they would run the risk of being overlooked in a boundless sea of free-floating single objects and/or phenomena (Zerubavel 1991).

Categorising is not possible without the setting of boundaries and/or ‘boundary-work’ (Gieryn 1983). The latter consists in demarcating groups of subjects, objects and/or phenomena by attributing certain characteristics that differentiate them from each other. Gieryn (1983) originally employed the metaphor of boundaries to show how science demarcated itself from non-science and attained its position of cognitive authority.³ The same approach can certainly also be applied to analyse how academic disciplines or fields of research within them emerge and demarcate themselves from one another and to identify the rhetorical strategies that make such a differentiation possible.

The paper pursues two interrelated objectives. The first is to explain the seemingly pressing need to categorise the various types of web-based translation mentioned above, referred to in this paper as ‘online collaborative translation’. It will show why categorising and convincing boundary-work are essential for academic disciplines and their internal structuring. This first research objective is formulated against the backdrop of voices within the field of ‘online collaborative translation’ who oppose or at least express reservations about categorising it and/or using a meta-concept. Dwyer (2017) advocates for not applying any categorisations, citing the example of Viki, a commercial fansubbing community that she argues cannot easily be placed into any one box but instead blurs the boundaries between translation crowdsourcing and fansubbing (Dwyer 2017: 173ff.). Desjardins (2021) speaks against categorising the translation phenomena taking place in online and digital spaces on the grounds that they are evolving rapidly, so the goal should be describing them, rather than defining, delineating and applying fixed terminology to them.

The second objective is to advocate for ‘online collaborative translation’ as the meta-concept, taking Zwischenberger (2022) as its point of departure and considering the matter from the perspective of a particular ‘thought community’ (Zerubavel 1996). Certain communities within the field will be shown to align themselves behind certain meta-concepts and to advocate for them. These advocacies will be analysed and made transparent via ‘boundary-work’ (Gieryn 1983), which also underpins the advocacy undertaken in this paper. To conclude, it will demonstrate how the various concepts proposed in the literature are related to ‘online collaborative translation’ and for which forms of online collaborative translation they may actually be suitable. A conceptual cartographical map for online collaborative translation will be introduced to reflect all of this.

2. Online collaborative translation as a field of research and concept

Online collaborative translation is still a very young phenomenon, both as an academic concept and as a field of research. It emerged just over 15 years ago in Translation Studies with research into fansubbing (Díaz Cintas and Muñoz Sánchez 2006) and scanlation, and fansubbing and translation hacking (O’Hagan 2007), and it can be associated with the advent of Web 2.0.

Under the meta-category or concept of ‘online collaborative translation’, I also subsume all solicited web-based translation such as the various types of translation crowdsourcing for both profit-oriented and non-profit entities, whether voluntary or paid. Translation crowdsourcing is a form of translation solicited by a call from an organisation, institution or company. The process is, thus, a top-down one in which the locus of control and management of the entire translation process remains entirely with the initiating organisation, institution or company. With translation crowdsourcing, the source text is split into chunks, often highly fragmented in nature. In the case of translation crowdsourcing for social media like Facebook or Twitter, for example, the translation unit consisted of a ‘string’ – a word or two, a phrase, or at maximum a sentence or two.

Typical examples of unpaid translation crowdsourcing for the profit-oriented sector include Facebook, Twitter or Skype. Unpaid translation crowdsourcing is also employed by non-profit organisations such as TED (subtitling its TED talks), KIVA (which lends microcredits to students and small entrepreneurs in developing countries) or Translators without Borders (TwB) (which provides translations to the NGO sector).

Paid translation crowdsourcing employed by the translation industry, which emerged around 2008 (Garcia 2015), is a far less (well) researched phenomenon. We may assume that these translation crowdsourcing platforms have been mushrooming ever since, but there is no comprehensive overview available. One of the earliest translation industry platforms offering translation crowdsourcing was Gengo, which was founded in 2008. It focuses on various tiers of professional human translation, charged accordingly (Gengo 2023). Unbabel, founded in 2013, is an AI-platform which delivers exclusively machine translation with human post-editing (Unbabel 2023). Other translation platforms offering paid translation crowdsourcing like Smartling, founded in 2009, offer both human translation and machine translation along with machine translation plus post-editing. Smartling’s translation unit is similarly a string consisting of a word, phrase, sentence or paragraph at most (Smartling 2023).

With unsolicited forms of online collaborative translation, there is no call by an organisation or company. The entire translation process is self-managed by a community or group of translators, who also decide themselves which texts to translate. The process, thus, is bottom-up. One example of this is Wikipedia-translation, which may not be considered ‘conventional’ translating from a source text but rather a mix of translating, collating, summarising and synthesising, drawing from a wide range of sources. Translation on Wikipedia is also an unfinished process, as anyone can edit a text at any time and act as a contributor or collaborator (Jones 2019). Another example would be Translaville (Rogl 2022), a community for exchanging translations among its members. Only members who translate for other members can also request translations themselves. Larger projects are split and divided among translators. The text of the platform’s interface was also heavily fragmented and translated collaboratively by members into 34 languages (Rogl 2022).

All the various types of online fan translations such as fansubbing, fandubbing, scanlation and the translation hacking of video games may also be subsumed under this sub-category. Usually the source texts are split and divided among fan translators, with the various contributions being ultimately recombined into one file, guaranteeing fast turnaround times. Furthermore, these communities are highly hierarchical, with positions like junior translators, translators, revisers and senior translators all contributing to the final text (e.g. Massidda 2015).

These various sub-categories of the main category of ‘online collaborative translation’ may be considered core concepts or categories (Zwischenberger 2022). However, the concept of ‘online collaborative translation’ may be extended, as previously pointed out by Désilets and van der Meer (2011, 28), to also include collaborative terminology resources, translation memory sharing, etc. I propose to extend the category to also include human-machine collaborations like the post-editing of machine translation (MT) or the use of translation memory software (TMS), both of which can be combined in computer-assisted translation (CAT) tools.

Particularly in its cloud-based and free iterations such as MateCat, Wordfast Anywhere or MemSource, TMS may furthermore be regarded as direct instances of online collaborative translation in that many translators constantly feed into their databases.

Many CAT tools additionally offer a built-in function for collaborative translation:

MateCat provides an easy-to-use split functionality to divide large jobs into smaller parts and assign them to multiple translators. It also allows more translators to work simultaneously using the same translation memories, glossaries and even to discuss any issues related to the translation right within the tool itself, thanks to the comments feature. This is useful for large collaborative projects. (MateCat 2023)

2.1 Online collaborative translation as prototypical transcultural communication

While all translation may be regarded as a subtype of transcultural communication, this categorisation holds particularly true for online collaborative translation. Here, all the texts to be translated are fragmented to a greater or lesser extent and sometimes split among dozens, hundreds or, in the case of social-media translation, thousands of translators, who all make their contributions to a final translation product. Furthermore, some of the translators are also involved in voting on, reviewing and revising translations. This means that any translation generated via online collaborative translation is characterised by the confluence and entanglement of various perspectives, skills and resources. The translations generated this way thus feature a huge degree of hybridity, diversity and heterogeneity. This is a perfect manifestation of transculturality and consequently also transcultural communication (Welsch 1999). Welsch (1999) devised his concept of ‘transculturality’ to describe the constitution of entire cultures and also their individual members, both of which owe their formation to various entanglements and blendings. Welsch (2009), however, also includes the products and production processes of cultures and its individual actors in his notion of transculturality:

It would have to be the case that the actors performing a piece not only come from different cultures but that the piece created or performed by them unites various and different cultural patterns, meaning the resulting product is not mono- but rather trans-cultural.

(Welsch 2009, 55; my translation)

Furthermore, the outcome of all translation is a communication that goes ‘beyond’ or ‘across’ cultures, as signalled by the ‘trans-’ prefix.

‘Collaboration’ is used here with the deeper definition provided by Organisation Studies rather than the everyday meaning of simply ‘working together’. Viewed through this lens, it becomes clear that ‘online collaborative translation’ is an entirely fitting concept, for the phenomena discussed since transculturality is inherent to it:

Collaboration is a process through which parties who see different aspects of a problem can constructively explore their differences and search for solutions that go beyond their own limited vision of what is possible. […] If collaboration is successful, new solutions emerge that no single party could have envisioned or enacted.

(Gray 1989, 5, 16)

Collaborating means creating something whose strength lies in its hybrid nature and which is produced by compiling individual perspectives, resources and skills. Thus, no single contributor could generate a collaborative result single-handedly. Collaboration constitutes transculturality and is itself constituted by transculturality.

3. The need to categorise

As already pointed out, the concept or category of ‘online collaborative translation’ is far from the only one in use in Translation Studies.

Categorising is something that comes naturally to us. Most of the time we do it spontaneously and without much thought – sometimes even automatically and unconsciously. Categorising, thus, makes us human. To describe the translation of Facebook’s or Twitter’s interfaces, for example, as ‘translation crowdsourcing’ is in fact to categorise it. Even the subordinate act of naming the translation activity as ‘Facebook translation’ is an act of categorisation. Every time we see and name something as a kind of a thing, we are categorising (Lakoff 1987, 5–6).

We need to categorise in order to make sense of and bring order to our world. Otherwise there would be an endless and boundless sea of possibilities that we could not grasp, let alone process. Categorising, thus, is meaning-making: ‘If each of the many things in the world were taken as distinct, unique, a thing in itself unrelated to any other thing, perception of the world would disintegrate into complete meaninglessness’ (Simpson 1961, 2, quoted in Zerubavel 1991, 5). According to Lakoff (1987, 6), without the ability to categorise, we could not navigate or orient ourselves at all, in neither the physical world nor in our social or intellectual ones. The world needs to be divided into digestible bits and pieces, otherwise we would be in a constant state of cognitive and emotional overwhelmedness.

Our need to categorise is so profound that we are prepared to bend phenomena that defy our mental frames and resist classification in order to force them into a category. When such adjustments do not work, we create special categories such as ‘others’ or ‘miscellaneous’ (Zerubavel 1991, 5–6).

Categorising, thus, is an expression of our deep yearning for order, meaning, structure and therefore orientation and guidance in our lives, both of which provide us with security.

3.1 The act of categorising

By categorising, and particularly by creating a higher-order category like ‘online collaborative translation’, we carve out an entity from a continuum and separate it from its surroundings. We create order in an otherwise boundless sea of individual objects and thereby endow this carved-out entity with meaning (Zerubavel 1991).

Carving out these entities or ‘islands of meaning’ (Zerubavel 1991, 5) presupposes two diametrically opposite and yet complementary acts: lumping and splitting (Zerubavel 1996, 421). An island of meaning like ‘online collaborative translation’ is a cluster of phenomena that are regarded as more similar to one another than to anything outside the cluster. As we lump or group those things together into clusters in our minds, we allow their perceived or alleged similarity to outweigh any differences among them. We are therefore focused on finding similarities, which we may even exaggerate in our minds and rhetoric, whilst disregarding or at least downplaying any intracluster differences. The uniqueness of things needs to be ignored through the act of lumping so that we can regard them as ‘typical’ members of clusters. Without this ability, it would be impossible to envision any mental cluster at all. Lumping presupposes our ability to envision our reality topologically (Zerubavel 1996, 421ff.).

Splitting is the act of viewing different clusters as separate from one another by widening the perceived gaps between them, i.e. exaggerating intercluster differences. If we could not perceive such mental gaps between clusters, we would not be able to envision any islands of meaning at all (Zerubavel 1996, 424).

Clearly, no category can ever be fully insular – indeed, categories may also overlap and certain phenomena may fall into multiple categories simultaneously. Such is the case, for example, for the various types of online fan translations, which certainly may be performed collaboratively but also individually. The same applies to translation requests on Translaville, mentioned above, where individual and short translation requests are usually dealt with by individual translators, whereas larger translation projects or the translation of the platform’s interface are undertaken collaboratively by the community. Furthermore, the voluntary translators are requested to participate in reviewing and evaluating others’ translations via polls (Rogl 2022, 176, 246, 251).

Family resemblance, to follow Wittgenstein, suffices to group things together into the same category. This means that members of a category may be related to one another without all members possessing 100% of the properties that define the category (Lakoff 1987, 12ff.). Fansubbing certainly involves a different kind of collaborative process to, for example, translation crowdsourcing. There is also a gradation in the degree of collaboration, in the social closeness of the individuals and in the hybridity of its resulting product. Whereas the translation crowdsourcing employed by Facebook may have involved thousands of translators contributing concurrently, thus generating a translation of enormous hybridity, the same cannot be said of, for example, paid translation crowdsourcing within the translation industry or fansubbing. With fansubbing, however, people collaborate and have closer social interactions, developing an identity of ‘we as a group’ (e.g. Li 2017), whereas translation crowdsourcing initiatives usually involve more anonymity among members of the crowd. Nevertheless, in both cases the translations created are the hybrid sum of a multitude of contributions. Some examples within the category of ‘online collaborative translation’ may be more central or prototypical than others (Rosch 1983). There is, thus, asymmetry within categories (Lakoff 1987, 12ff.).

This presupposes categorising with a flexible – not to be confused with fuzzy – mind rather than a rigid one (Zerubavel 1991, 1995). A rigid mind would see discrete, insular and mutually exclusive entities of meaning clearly separated from one another by wide gulfs, whereas the opposite, a fuzzy mind, would see no contours and thus establish no order at all. Categorising with a flexible mind, by contrast, means envisioning categories that have contours but no definitive edges. To the flexible mind, categories may certainly overlap and a phenomenon may span more than one category – in other words, categories are anything but intellectual silos (Zerubavel 1995).

4. Setting boundaries

To categorise is also to set boundaries – to specify what lies inside and what lies outside the bounded space. Boundaries help demarcate things and/or groups of things from one another. To categorise and thus to set boundaries or draw lines is also to define. In fact, ‘to define’ derives etymologically from the Latin finis, meaning ‘boundary’. To define is to mark boundaries and erect a mental fence around a certain space (Zerubavel 1991, 2).

The boundary metaphor is also intimately related, indeed essential, to the idea of space. Academia is a bounded space split into disciplines, faculties, departments, fields of research or areas of specialities (Carlson and Lewis 2020, 130). The category of ‘online collaborative translation’ may be assumed to represent an emerging field of research or an area of speciality for scholars. Boundaries bring enclosure.

4.1 Boundary-work

The concept of ‘boundary-work’ may be traced back to Gieryn (1983, 1995, 1999), who demonstrated how science successfully professionalised and demarcated itself from non-science. The concept is not new in Translation Studies, but seldom has it been employed as an analytical instrument.⁴ It was introduced by Grbić (2010, 2023), who used it to study the professionalisation processes of sign language interpreters in Austria.

Boundary-work is defined by Gieryn (1999, 4–5) as: ‘the discursive attribution of selected qualities to scientists, scientific methods, and scientific claims for the purpose of drawing a rhetorical boundary between science and some less authoritative residual non-science’. Boundary-work is a credibility and ideology contest whose goal is to rhetorically demarcate one category from another and identify one as superior (Gieryn 1999). It depends on persuasion and the associated argumentation: ‘The justification of knowledge production is not purely an intrinsic matter but a social one that occurs through the process of boundary work in the form of “credibility contests” among competing parties’ (Carlson and Lewis 2020, 124). The various competing meta-concepts outlined in section 6 may be seen against this backdrop.

There are various ways of doing boundary-work. Gieryn (1999, 15–16) distinguishes between expulsion, expansion and the protection of autonomy. Expulsion is about excluding and putting discrepant claims outside a certain map or authoritative cultural space. This may also include strategies to delegitimise the other. Expansion is the act of integrating a certain epistemic domain as part of a certain bounded space. Finally, the protection of autonomy is about defending the autonomy of a certain bounded space against external forces that would claim it for themselves.

These strategies have been adopted by the advocates of the various meta-concepts for defining the various translation phenomena outlined in section 2.

5. Consequences of categorising or creating bounded spaces

Categorising single phenomena like Wikipedia-translation, translation crowdsourcing, etc. under a meta-category like ‘online collaborative translation’, and thereby carving them out as a wider entity from reality, renders them visible. This is undoubtedly an act of social construction (Zerubavel 1991). When grouped together, these various translation phenomena are seen more readily as potential objects of research and may consequently attract more attention than if they are left outside and isolated. Together they have more clout. An emerging field of research can thereby gain real momentum and become institutionalised within an academic discipline. This has consequences in terms of academic recognition, which manifests itself in conferences and publications, study programmes, single courses, academic positions, etc. being dedicated to a particular category and attracting third-funding possibilities. This again helps to foster an academic discipline, which is important for any discipline, and particularly for younger ones like Translation Studies.

Attributing certain characteristics to the translation phenomena mentioned above, such as their collaborative and consequently also transcultural nature, unites them into the category of ‘online collaborative translation’, which has a distinct identity: ‘Social identity is always exclusionary, since any inclusion necessarily entails some element of exclusion as well’ (Zerubavel 1991, 41).

Categorising and setting boundaries is certainly an arbitrary activity. Lines can always be drawn differently, but they need to be drawn convincingly, making analytical thinking and convincing argumentation imperative (Gieryn 1999; Zerubavel 1991, 116). This also entails a well thought-through and reasoned structuring of categories into sub-categories, etc. Otherwise a newly carved-out entity will stand no chance of persisting.

Categorising also always shines a light on a particular group of objects or phenomena. Categories work like social lenses that place the focus on something while disregarding its surroundings.

However, we do not look at the world through social lenses individually but rather as collectives. As pointed out by Zerubavel (1991, 76–77), ‘when we cut up the world, we usually don’t do it as […] individuals, but rather as members of society’. Being socialised or acculturated entails knowing not only how to behave, but also how to perceive reality in a certain way. Being a member of a society, i.e. a certain ‘thought community’ (Zerubavel 1996, 427–428), entails ‘seeing’ the world through special mental lenses or from a certain ‘horizon’ (Zerubavel 1993). I and other Translation Studies’ scholars like, for example, Jiménez-Crespo (2017) see collaboration as the great uniting force, and have argued our case for seeing ‘online collaborative translation’ as the most suitable meta-category for these phenomena. At first glance, we may seem to come from the same thought community, and yet our conceptions of collaboration are seemingly informed by different backgrounds. To me as a scholar in Translation Studies and transcultural communication, collaboration – especially when undertaken by such a multitude of actors, yielding translations characterised by hybridity and entanglements – is a perfect example of transculturality and/or transcultural communication. As mentioned, I also draw from Organisation Studies’ definition of collaboration, whereas for Jiménez-Crespo (2017) and other Translation Studies scholars, collaboration seems to be associated more with the technological possibilities granted by Web 2.0 and the possibilities it provides for the collaborative and interactive use of platforms.

6. Categories and their (de-)construction through boundary work

Zwischenberger (2022, 4ff.) presented a rather comprehensive overview of alternative meta-concepts but no discussion of how they relate to ‘online collaborative translation’ based on boundary-work, nor of how these alternative concepts were negotiated as top-level concepts by various authors and their demarcation strategies. There was also no map illustrating these structures, now remedied in Figure 1. In it, when a concept fully applies to a sub- or sub-sub-category, it is given in normal font weight, whereas a lighter font indicates only partial applicability.

Figure 1. Conceptual map of online collaborative translation.

The meta-category of ‘community translation’ (e.g. O’Hagan 2011) is inevitably associated with ‘community interpreting’. O’Hagan (2011) shows herself to be aware of this association but insists that community translation has taken on a specific meaning in the context of Web 2.0. O’Hagan (2011) dismisses further concepts like ‘volunteer translation’, as favoured by Pym (2011), due to the fact that not all types of translation made possible by Web 2.0 are voluntary in the sense of unpaid. The remuneration paid via translation industry platforms makes ‘volunteer translation’ an unfitting meta-concept, as also pointed out by Zwischenberger (2022, 5). For the rest of the sub-categories, and thus also the various types of unsolicited translation but also for unpaid translation crowdsourcing ‘volunteer translation’ fully applies (Figure 1).

Pym (2011, 97) demarcates ‘volunteer translation’ from other possible meta-concepts such as ‘community translation’ and ‘collaborative translation’ because to him the lack of financial remuneration is the defining characteristic. He explicitly includes translation crowdsourcing under ‘volunteer translation’. Similarly, O’Hagan (2011, 13) explicitly expands her concept to include translation crowdsourcing. Thus, she uses both expulsion and expansion as rhetorical strategies.

‘Community translation’ is not considered a fitting candidate for the various types of online translation phenomena presented above, as it presupposes a group that is bound together by a feeling of ‘we as a group’. This certainly does not hold true for translation crowdsourcing where the group is usually composed of a large anonymous crowd (Zwischenberger 2022, 4). However, ‘community translation’ applies as a sub-category to contain the various types of unsolicited online collaborative translation that actually involve a community and self-manage the translation process, relying on social bonding and mutual engagement (see Figure 1).

‘Non-professional’ or ‘amateur translation’ (Pérez González and Susam-Saraeva 2012) also do not apply as meta-categories in Zwischenberger’s (2022) view, since the various types of translation subsumed under ‘online collaborative translation’ are not performed by non-professional translators or amateurs exclusively, as proven by various empirical studies in which the translators were asked to self-identify (Dombek 2014; McDonough Dolmaya 2012). Categorising with a flexible mind, the two concepts are not completely expulsed from the conceptual map, but they are stated in lighter font, as they do not fully fit into any of the categories as a whole but rather traverse and intersect with them (see Figure 1). Pérez González and Susam-Saraeva (2012), however, go far beyond the translation types subsumed under online collaborative translation by them. They offer an account of how Translation Studies in general has expanded to embrace the various types of non-professional or amateur translation in its research agenda.

These two concepts are closely linked to ‘user-generated translation (UGT)’ (e.g. O’Hagan 2009; Perrino 2009) as a possible top-level concept. O’Hagan (2009) does not demarcate the concept from any other concepts but instead seems to place UGT and community translation on the same categorical plane: ‘The article describes the evolution from unsolicited fan translation to solicited community translation now called crowdsourcing and considers them in the framework of user-generated translation (UGT)’ (O’Hagan 2009, 94). She defines UGT as ‘[…] a wide range of Translation, carried out based on free user participation in digital media spaces where translation is undertaken by unspecified self-selected individuals’ (O’Hagan 2009, 97). Perrino (2009) defines UGT in the same vein as ‘the harnessing of Web 2.0 services and tools to make online content […] accessible in a variety of languages. It […] implies the collaboration between users – be they amateurs or experts’ (Perrino 2009, 62). Perrino (2009) does not explicitly demarcate ‘UGT’ from any other competitor top-level concepts but instead expands it to include the concept of ‘collaboration’. Like O’Hagan (2009), he attributes the characteristics of democracy and empowerment to UGT.

However, as pointed out by Zwischenberger (2022, 4), basing her argument on media theorists Kaplan and Haenlein (2010), content – and thus also translation – must be created outside of professional activities and without a commercial intent in order to be truly ‘user-generated’. Neither of these can be assumed for translation crowdsourcing for the profit-oriented sector. Furthermore, all online collaborative translation practices may also involve professional translators or those trained in the field of translation or interpreting. Similarly, the hierarchical structure and guidelines within fansubbing groups imitate the workflows of commercial practices. Consequently, ‘UGT’ is included as a sub-category in a lighter font for all categories in Figure 1, except for ‘translation crowdsourcing for the profit-sector’, from which it is entirely expulsed, as it is does not apply at all in this sub-category.

One other meta-concept that centres on democracy and empowerment is ‘participatory translation’ (Gambier and Kaspere 2021; Jones 2021), clearly derived from ‘participatory culture’ (Jenkins 1992). Jones (2021) explicitly proposes the concept as an alternative to ‘online collaborative translation’, thereby completely expulsing the concept on the grounds that he sees ‘collaboration’ as incompatible with the often heated conflicts observed in Wikipedia-translation (Jones 2019). However, he does not underpin his conception of ‘collaboration’ with any definitions. In fact, collaborations are usually described in Organization Studies as conflict-induced and not necessarily culminating in a harmonious result (Gray 1989).

While Jones (2021) expands his notion of ‘participatory translation’ to all the web-based translation phenomena also subsumed under ‘online collaborative translation’ in this paper, Gambier and Kaspere (2021) conceptualise it as a synonym for ‘translation crowdsourcing’. Gambier and Kaspere (2021) consider ‘collaborative translation’ a category of its own characterised by the sharing of resources, working on the same document from diverse locations, revision, proofreading, etc. However, the exact same happens in translation crowdsourcing as well, which Gambier and Kaspere (2021) consider a sub-category of ‘participatory translation’.

As pointed out by Zwischenberger (2023) drawing on Fuchs (2014), ‘participatory culture’ and consequently ‘participatory translation’ as concepts are directly linked to participatory democracy theory. ‘Participatory translation’ would entail a democratic participation in the power and decision-making processes surrounding organisation and structure and also economic considerations (Carpentier, Duarte Melo, and Ribeiro 2019; Fuchs 2014). This is definitely not the case with the top-down processes inherent in translation crowdsourcing, where translators have absolutely no influence over organisational or economic decisions. Similarly, the pronounced hierarchies within the various types of online fan translations resist such a characterisation. Nor does it fully apply in Wikipedia-translation, where certain groups often seem to have the upper hand in text production, as shown by Jones (2019). ‘Participatory translation’ therefore fundamentally cannot apply to translation crowdsourcing, which always involves a top-down process. However, it is not completely expulsed for the various types of unsolicited online collaborative translation. For these, ‘participatory translation’ is thus given as a sub-category in a lighter font.

A further top-level concept used in the Translation Studies literature is ‘social translation’ (Jiménez-Crespo 2017, 28). McDonough Dolmaya and Sánchez Ramos (2019) add the qualifier ‘online’ to make ‘online social translation’. In Jiménez-Crespo’s (2017, 28) conceptual map, ‘social translation’ is the top-level concept, subsuming all other concepts, including ‘collaborative translation’, as sub-categories. This is interesting given the prominence of ‘online collaborative translation’ in the title of and throughout his monograph. Zwischenberger (2022) points out that the top-level concept of ‘social translation’ is too nebulous and therefore unconvincing, since all translation done by humans for humans is ‘social’ by definition. Furthermore, ‘social translation’ may also evoke the connotations of the translation concept as used by the Social Sciences. In general, the term ‘social’ is viewed with great scepticism by social media theorists. It is considered somewhat misleading, since the goal of social media is not human connectedness but rather the connectivity that social media generate so they can sell the ensuing data to corporations for targeted advertising (Van Dijck 2013). We must also view ‘online social translation’, which derives its name from ‘social media studies’ (McDonough Dolmaya and Sánchez Ramos 2019), in this light. McDonough Dolmaya and Sánchez Ramos (2019) also argue that it is built on ‘community translation’ as introduced by O’Hagan (2011). However, ‘community translation’ as a meta-concept is expulsed by McDonough Dolmaya and Sánchez Ramos (2019) due to its associations with the field of ‘community interpreting’. Collaboration is also a factor here: ‘[…] collaboration is a visible and inherent feature of online social translation’ (McDonough Dolmaya and Sánchez Ramos 2019, 131). Given all the above, ‘(online) social translation’ is expulsed from the map (Figure 1).

A very similar line of argumentation is advocated by Hebenstreit (2019) who, however, does not just refer to social media but explicitly proposes ‘social-media driven translation’ as a meta-concept. His search for a suitable meta-concept starts with Nuopponen (2007) and her model of various criteria for analysing dynamic concepts. He concludes that there are three basic criteria that play a dominant role in the Translation Studies literature, namely agent (the user), way of doing (collaborativeness) and instrument (social media technology). Despite his conclusion that ‘collaboration’ is the most frequent attribution in descriptions, he nevertheless chooses ‘instrument’ and thus ‘social media-driven translation’ as a meta-concept. He does so on the assumption that social media technology as an instrument is independent of its users and their collaborations: ‘[…] “users” and “collaborativeness” are not independent characteristics, they depend on social media technology’ (Hebenstreit 2019, 149). Zwischenberger (2022, 6) argues that social media are equally dependent on their users and their collaborations in order to exist and have any purpose. Furthermore, most online fan translation groups operate via fora originating in Web 1.0 and even earlier, rather than within the social media technology that emerged only with Web 2.0. Similarly, wiki technology was introduced in the mid-1990s and thus in the era of Web 1.0. ‘Social-media driven translation’ is therefore suitable as a sub-category but not as the main one (see Figure 1).

The same also applies to the concept of ‘concurrent translation’, which was recently introduced by Gough et al. (2023) but has never claimed to be a main category. Gough et al. (2023, 3) see ‘concurrent translation’ as a sub-category of ‘online collaborative translation’. ‘Concurrent translation’ occurs in cloud-based environments:

[…] only a scenario where one text is simultaneously translated by a number of translators – whether by splitting a text and assigning segments to individual translators or by allowing translators to select segments on a ‘first come first served’ basis – is taken into account in this study. The key notion is that all participating translators have access to the same document and work synchronously (concurrently).

(Gough et al. 2023, 2)

Gough et al. (2023) rightly subsume all sorts of translation crowdsourcing platforms under this sub-category, along with CAT tools that have a specific built-in function for this. However, wiki-technology and thus also Wikipedia-translation show that cloud-based environments are not essential to ‘concurrent translation’ (see section 2 and Figure 1).

The above can be summarised in the following map.

Greater contextualisation is needed in order to place the extended notion of ‘online collaborative translation’ involving human-machine collaborations (e.g. post-editing of MT, the use of TMS or the combination of MT and TMS in CAT tools) within the subcategories. The post-editing of machine translation certainly can be voluntary translation, community translation, amateur translation, etc. These extended forms of ‘online collaborative translation’ may also represent an integral part of the core forms of ‘online collaborative translation’ such as translation crowdsourcing, which may involve collaboration not only between humans but also between machines and humans (see section 2).

7. Conclusions

The conceptual map presented here is the result of categorisation and boundary-work, presenting ‘online collaborative translation’ as the most suitable meta-category candidate. This was achieved through a specific argumentation process that fully engaged with the various alternative candidates proposed in the literature, which were all expulsed as suitable meta-concepts. Most of the alternative meta-concepts, however, are included as sub-categories of ‘online collaborative translation’, though some were expulsed from the conceptual map altogether (e.g. ‘social translation’). ‘Online collaborative translation’ was also presented as the most expansive concept since it can be expanded beyond the various types of solicited and unsolicited online translation types to also cover 1) human-machine interactions (MT and CAT-tools) and post-editing, and 2) cloud-based TMS and CAT tools that allow projects to be fragmented and translated collaboratively, yielding translation results that are highly hybrid in nature. The category of ‘online collaborative translation’ in this paper was also linked to transculturality and/or transcultural communication. It therefore derives from a certain horizon or thought community that presents the highly hybrid and heterogeneous nature of these translations – generated through the confluence of various contributions into one text through collaboration – as the great uniting characteristic of these various types of online translation.

Categorising is an important act for rendering carved-out entities visible and imbuing them with meaning so that they can be grasped, and therefore attract attention, investigation and discussion. These entities underpin the emergence of institutionalised fields of research and therefore help advance academic disciplines. Categories play an important role in providing a joint framework for discussion. Categorisations also generate debate and discussion themselves, and they emerge as a result of debate and discussions. Categories and their boundaries are never fixed. The aim of the categorisation and boundary-work and the ensuing conceptual map presented in this paper is to present a possible common framework and take this as a point of departure for inviting further discussions and debate in order to advance research into the phenomena subsumed here under ‘online collaborative translation’.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Cornelia Zwischenberger

Cornelia Zwischenberger is Professor in Transcultural Communication at the Centre for Translation Studies at the University of Vienna. Prior to her appointment as professor at the University of Vienna in March 2020, Cornelia Zwischenberger held a professorship in Translation Studies at the University of Graz in Austria. Professor Zwischenberger has published numerous contributions on both Translation and Interpreting Studies. Cornelia Zwischenberger’s current research focuses on the use of the translation concept beyond Translation Studies from a transdisciplinary/transcultural perspective and on online collaborative translation as a prototypical form of transcultural communication. Together with Alexa Alfer she has been working on the blended concept of translaboration for several years now. She is the leader of the research group Transcult.com. Furthermore, her research also revolves around scientific theoretical questions such as the use of the appropriate concepts to narrate the evolution of the Translation Studies discipline.

Notes

1. Facebook stopped its translation crowdsourcing in mid-April 2022 and has not revealed any further details regarding this move (Translate Facebook Team 2022).

2. Twitter stopped its translation crowdsourcing in November 2017 and has since switched entirely to neural machine translation (Translate Twitter 2022).

3. The contesting of empirically proven scientific facts during the Covid pandemic shows that boundaries, and thus any clear authority over scientific truth, are not permanent and unassailable, but rather that boundary-work is and must be continuous.

4. For an overview and critique of the use of ‘boundary-work’ in Translation Studies, see Grbić (2023, 115–116).