Status and technology in the professionalisation of translators. Market disorder and the return of hierarchy
Anthony Pym, David Orrego-Carmona and Esther Torres-Simón, Universitat Rovira i Virgili
The professional status of a translator is traditionally indicated by a set of social signals including previous experience, academic qualifications, professional accreditation and membership of associations. When those signals shifted from print and word-of-mouth to electronic media, some degree of market disorder resulted with respect to globalisation of translator-client contacts, the growth of volunteer translation, access to free online machine translation, and the corresponding motivation to steal the identities of professional translators. Three case studies of websites and forums that have been associated with market disorder (ProZ.com, a comparison of aRGENTeaM and GrupoTS, and the Translator Scammers Directory) indicate that the initial disorder has been challenged and in some cases significantly corrected, with new forms of signalling appearing within the electronic environments. ProZ.com has instigated its own accreditation system, the volunteer subtitling communities have developed elaborate internal hierarchies of control, and the stealing of translators’ identities has been challenged through more sophisticated use of the same electronic media that allowed the thefts. In the new configuration of signals, however, it would seem that academic qualifications have less market value than does verifiable professional experience, while professional accreditation still has value but can be forged. For many segments of the translation market, the return to market equilibrium will require greater attention to new signalling mechanisms, with more sophisticated uses of electronic communication.
Translator status, signalling mechanisms, translator professionalisation, electronic media
We understand “translator status” to be the set of values produced by social signals that indicate the relative trustworthiness of a translator. When those signals are controlled by groups of translators themselves, indicating who is in the group and who is not, we can talk about a degree of professionalism. This approach is particularly important in the case of translators because, more than most other services, the value of a translation is commonly not easily attested by the person paying for the service: in many situations, when you pay someone for a translation, it is because you cannot do the translation yourself. Clients are thus largely dependent on the external signals of a translator’s status.
The signals of status are limited in range and type. They can be classified as attested experience, academic qualifications, professional qualifications, membership of associations, and evidence of reliability (often in the form of word-of-mouth recommendations). Different social situations will have different configurations of those signals.
Here we delve into instances where traditional signals of professional status, based on print and word-of-mouth, have been challenged by electronic media (websites, online forums, email, electronic payment systems). More specifically, we will be looking at cases of relative de-professionalisation. Just as Foucault (1975) analysed the history of the French prison system in order to understand the nature of internalised surveillance throughout the whole of the society, and Toury (1995) scrutinised pseudotranslations in order to grasp the norms of all translations, so we look at relative market disorder in order to trace the signals of translator status. We will be considering three broad cases, all marked by the possibilities opened up by electronic communication: online translator-client marketplaces (ProZ.com), online communities of volunteer translators (comparing aRGENTeaM and GrupoTS), and the online challenging of translator scammers (the Translator Scammers Directory).
2. Market disorder and the study of signals
One way to study professionalisation is to select a group of people who seem to be professionals, then you ask them questions, and you might also ask other groups questions about the target group. This basic approach can effectively test initial hypotheses about professionalisation. It nevertheless remains methodologically problematic in that it presupposes the pertinence and veracity of certain signals of status, first in order to select the professional group (e.g. translators with a postgraduate degree in translation plus three years’ experience) and second in the assumption that everyone is telling the truth. This becomes awkward when, for example, you are an academic who trains translators and you ask people about the relative value of formal training for translators — they will want to tell you it is very important. Or again, if you ask a group of translators who all have formal training, they will also tell you that training is of high value, precisely because they have made personal investments in that kind of signal. And then, in a slightly different way, you might ask translation companies whether they prefer speed or accuracy when hiring translators, and they will all tell you that accuracy is far more important, since to say otherwise would mean signalling that the company produces translations with mistakes. The relative trustworthiness of signals (especially when the signals are produced precisely to indicate trustworthiness) is problematic even within the basic mechanics of research.
One way to mitigate this problem is to assume, as does information economics, that everyone is always lying to everyone, more or less. More exactly, everyone has an interest in overstating the value of their skills, services or products. When a translator negotiates with a client, the translator will overstate the value of their skills, just as the client has every interest in understating how much the translation will be worth in its actual use. There are then complicated ways of measuring just how much the reported value can deviate from actual exchange value, and how certain lee-ways can still allow for cooperation (mutual benefits) and market equilibrium to be achieved (see Pym et al. 2012/2013: 150ff.).
What most concerns us here, however, is what is called “asymmetric signalling”, which is where one party is assumed to have significantly more or better information than the other (after Spence 1973). Translators know more about their own skills than do their clients, just as clients usually know more about what will be done with the translations than do the translators. This asymmetry is particularly pertinent to the translation industry because in many cases (and in the general case we shall be assuming here), the client does not know the languages that the translator is working with. The product (the translation) is thus relatively opaque to the client, who can mostly only assess its value on the basis of external signals of the translator’s trustworthiness. In this, the buying of a translation could resemble the buying of a used car in the days when the market was unregulated. According to Akerlof’s classical model (1970), if there are no reliable signals of the car’s quality, then the risk of buying a bad car (a ‘lemon’) is greater. The overall price of used cars will thus be low, and good used cars will not be put on the market. This is ‘adverse selection’, since bad products drive good products out of the market. It is also a prime instance of ‘market disorder’, which is the general term for situations where signalling mechanisms are weak, contradictory or otherwise unable to promote relations based on trust. Once a reliable signalling mechanism is in place, however, the buyer’s risk decreases, the price of the cars increases, and good used cars enter the market and might be valued at something near what they are worth. A degree of market equilibrium is thus restored.
This model of asymmetric signaling, adverse selection and market disorder has been applied to the professionalisation of translators by Chan (2008, 2009) and has been developed in Pym et al. (2012/2013). The basic concept of market disorder is also used by Tseng (1992) and Ju (2009) with respect to the professionalisation of interpreters, in a model where professionalisation moves from ‘market disorder’ to ‘consensus and commitment’, ‘formal networks’ and finally ‘professional autonomy’. In this model, the nirvana of professionalisation is the state where formal ethical standards are established, there is control over who is admitted to the profession, and professional organisations work with the various stakeholders to achieve market control and influence legislation and certification – broadly speaking the professionals control the signals of their status.
The heuristic advantage of this general approach is that one can ask not just how far towards professionalisation a particular occupational group may be (on any of the parameters just mentioned), but also what evidence of market disorder there is with respect to the general occupational activity. Further, instead of looking at numerous sociological or political features of the occupational group, as does sociological trait theory (as in Witter-Merithew and Johnson 2004) one can ascertain degrees of market disorder quite elegantly by looking at the signals of professional trustworthiness, at how they are produced, by whom, and what relative value they are accorded on the market.
Does this approach mean that translators and their clients will somehow reveal their true values to us? Not at all. What it does assume is that, when people’s lies move too far out of kilter, this will be picked up in degrees of market disorder. And market disorder is the thing that professionalism is supposed to overcome.
We focus on market disorder here because we suspect that, contrary to Tseng’s model of linear progress from disorder to professional autonomy, there have been significant instances where the translation market has relapsed into degrees of disorder. That is, the supposed historical progress towards ever-greater professionalisation has been seriously challenged. And a major challenge, we propose, has been the technological shift in the nature of the signals themselves, from stamped documents and word-of-mouth to electronic communication. This concerns developments on many fronts. In the first place, electronic communication has allowed the translation market to become truly international, rather than local or national, which complicates the checking of signals. Second, it has allowed segments of the market (notably for audiovisual products) to be seriously challenged by volunteer translators (which might be seen as an instance of relative de-professionalisation). Third, electronic communication has presented us with free online machine translation, thanks to which anyone can produce something that looks like a translation (allowing for significant moral hazard). And finally, electronic communication allows the more traditional signals (CVs and certificates) to be copied easily and used by non-translators, who sell unedited machine translations in a globalising market where the lack of geographical proximity, as we have said, makes checking processes more difficult.
Our aim here is to explore the kinds of market disorder that have resulted from these effects of electronic communication. Our methodology is based on three case studies, each presenting evidence of at least one of the instances of disorder just mentioned:
1) The online translator marketplace ProZ.com, based on a model where clients propose jobs and translators bid for those jobs, initially drove down the average price of translations and thus forced trustworthy translators to abandon the sites. This would appear to be a case of classic adverse selection.
2) The online communities of volunteer subtitlers GrupoTS/TusSeries and aRGENTeaM, egalitarian and anti-consumerist in ethos, would appear to have de-professionalised the subtitling of many mainstream entertainment products.
3) The Translator Scammers Directory provides information on instances where translators’ signals of status have been stolen by false translators, who then send clients raw machine-translation output. This gives us further indications on how the signals of status are operating, if only because it suggests which signals are considered valuable enough to steal.
Our data in all these cases is drawn principally from the recent history of websites, accessed with Wayback Machine (which caches images of websites at various points in their history). This does not include information from members-only areas of the websites or password protected sections. For each case study, our main research questions have been as follows:
- What are the main signals of translator status?
- What communication media are used?
- What is the importance given to educational qualifications?
- Do previous clients/employers play a role in the formation of status?
- Do professional exams play a role?
- Does membership of professional associations play a role?
- Does citizenship of any country play a role?
- Does the group have its own internal signals of trustworthiness?
- Does it have an internal hierarchy? How is it signalled?
- What do members give to the community?
- Has the configuration of signals has changed during the life of the group?
- Are there other signs of de-professionalisation or re‑professionalisation in the historical evolution of the group?
In all these questions, we are looking at signals of status. And we are aware that all the signals could be lies.
3. Case study 1: ProZ.com
ProZ.com is a profit-making company that was launched by Henry Dotterer in 1999. In 2014 it has offices in Syracuse (United States), Buenos Aires (Argentina) and Kharkiv (Ukraine) and it has some 18 employees. Three months after its launch in 1999 it had 6,000 registered translators; in March 2014 it claimed to have 687,530 registered “users” (apparently including translators, translation agencies and translation companies) (see Figure 1). This is in a world where we estimate there is only a need for 333,000 full-time translators and interpreters who are professional enough to declare themselves as such when paying taxes (Pym et al. 2012: 132). Not surprisingly, ProZ.com can be considered the world’s largest translator community.
Figure 1. ProZ.com registered users (1999-2014)
Originally the company provided no more than a networking platform where translators could meet new clients and share experiences with fellow translators. It was organised by language pair and offered spaces to post and bid for translation jobs, share and resolve terminological doubts, and buy and sell books. The platform had an early growth of about 1,000 registered translators per month (Risku and Dickinson 2009: 58) and a later boom with the inclusion of agencies and companies. Currently, the platform also has a comprehensive set of tools for translators and agencies: glossaries, conferences in various countries, software discounts, feedback on clients, on agencies and on fellow translators, personal domains and e-mails, and, most recently, accreditation.
Although ProZ.com offers several services, not all of them are for the same type of user. The terminology and translation discussion forums are open to both registered and non-registered users. Job offers, bidding and access to virtual conferences is possible for registered users only. Registered users also have a different homepage, which they can customise.
Registration is free, but paying registered members obtain better positioning when bidding, receive exclusive job offers, and enjoy extra services like access to the BlueBoard (feedback on translation agencies and other possible clients), discounts on translation software, website hosting and e-mail, access to training courses, and online invoicing. It is difficult to estimate how many paying members there actually are, or how many participate in the various community activities. McDonough (2007: 805) studied TranslatorsCafé, which has a similar configuration, and found that “fewer than a quarter of the members actually visited the site in a 30-day period” and only seven per cent of registered members “had ever posted a question in the discussion forum.”
Translation jobs are either offered directly to specific translators that clients locate through the translator directory and whose status can be checked in the user profile pages, or posted publicly asking for quotes (bids). The KudoZ network is an open forum for users to help each other with terminology or other translation-related doubts. Registered users receive points for providing popular answers (accepted as correct and voted as the best). The visibility of the exchange works as an effective mode of peer recommendation. The ProZ.com forums are the place for translators to compare techniques, obtain and provide technical support, and pose general questions unrelated to terminology. ProZ.com encourages participation in the community by awarding BrowniZ to users who help translate the site, introduce new members, or organise or attend powwows (face-to-face meetings).
In ProZ.com, translator status is directly connected with the possibility of obtaining a translation job. The higher the translator’s status, the higher their profile appears in the directory and the better chances they have of winning bids.
Membership level marks the first difference. Currently there are five levels of membership: registered users (free membership), student membership, partial membership (6-month registration), full membership (12-month registration) and PRO certified member (full member with accreditation). Full members obtain better positioning in the directory.
A second signal of status is provided internally on the basis of work invested in the community. This is measured in terms of BrowniZ and KudoZ points. KudoZ points signal expertise and provide status with respect to clients, while BrowniZ represent contributions to the community and can be traded for community rewards (for example, 4,000 BrowniZ points can be exchanged for a discount in membership fees).
Feedback from previous clients and translators is one of the most important signals in the selection of translators, although it does not provide better positioning in the directory. While translators can choose whether clients’ comments are added to their profiles, showing feedback is highly recommended by the company, and daily reminders are sent to translators who have not yet provided any.
Membership of a professional association (like the American Translators Association) has been accepted as another signal of trust since 2005. Within the general configuration, however, most signalling is mediated by ProZ.com itself. The company carries out regular checks on a new member’s identity, credentials and membership, providing prospective clients with a minimum set of signals for building a relationship. Moreover, offering prospective translators in a list form assigns a certain status to the translators at the top of the list. Clients presumably use the list as a ranked indication of trustworthiness, although they might not always be aware that the top positions were obtained by full-paying members (that is, to a certain extent “bought”) and then classified by the number of KudoZ points (that is, expertise shown in the forum) and BrowniZ points (investment in the community). In effect, paying members are always ranked above non-paying members, regardless of their expertise.
Despite this distinction — understandable enough in a commercial operation — the combination of these three signals (payment, expertise and community involvement) seems to work fairly well as a filter of trustworthiness. Less qualified translators might buy membership, and even obtain BrowniZ by translating parts of the sites or attending powwows, but KudoZ points are difficult to accumulate without active participation in the community and advanced knowledge of the working languages and specialised terminology. Still, when surveyed, users stated that the presence of non-professional members (non-qualified translators or translators working outside their language pairs) was a drawback that bothered them (Risku and Dickinson 2009: 65).
There are some signs that point at changes in the configuration in order to restate imbalances provoked by the transition to electronic communication and rapid growth of the site.
Initially checks of identity, mother tongue and qualifications where services provided for paying members or accessed by registered users for a fee. Currently, identity check and mother-tongue accreditation is compulsory. Identity accreditation was one of the issues that originally remained linked to face-to-face interaction — it could only be done by attending a powwow — but it is now checked by paying a fee by credit card. The identities of paying members are automatically checked when they register.
The names used for the various community participants constitute another mode of signalling. Originally the site was for translators only, so in 2000 ProZ claimed it had “6,000 translators”. One year later, the community had accepted companies and thus presented itself as “14,287 freelancers and agencies.” Then, in an attempt to combine all translation structures under one umbrella, in 2002 they talked about “30,000 language professionals,” but only two years after that we find them as “63,391 professionals and agencies”. In 2009, the size of the site — which in 2008 had already claimed to be “the world’s largest community of translators” — is measured in terms of “translation jobs posted”. Since 2010 there has been a distinction between “registered freelancers” and “registered users,” and since 2012 the site claims to have “over 300,000 professional translators and translation companies”.
The need to accurately denominate the participants is certainly related to the public being addressed, but it also shows the importance of differentiating between different kinds of status.
A recent modification to the homepage introduces a basic bifurcation between registered and non-registered users. The homepage met by non-registered users provides few details on specific translators or job offers and instead links to a section for prospective clients or future members. Registered users, on the other hand, have a homepage that links to job offers, KudoZ questions, featured translators, etc.
In 2009 ProZ.com launched a new level of accreditation that aims to signal competent translators: the “Certified PRO network”. Members who obtain this certification are awarded a special badge and can access a private forum where they can find fellow “screened professionals” to work with. To attain this status, freelancers have to show basic translation ability, business reliability and “good citizenship”. The first item is defined by the EN 15038 standard for quality in the translation industry and requires competence in the source and target languages, research competence and cultural competence. These skills are signalled by industry credentials (like membership of ATA or the Chartered Institute of Linguists) and peer revision, with other possibilities for less common language pairs. In accordance with EN 15038, these competences may also be acquired and tested through formal higher education in translation, equivalent qualification in any other subject plus a minimum of two years of documented experience in translating or at least five years of documented professional experience in translating. Business reliability is signalled by client and peer reviews. Good citizenship is signalled by paid-up membership, showing a complete profile and adhering to site rules. These changes have been openly applauded by members, which implicitly suggests that the previous filters were not efficient.
ProZ.com’s trial-and-error evolution clearly indicates that a plurality of signals is desirable when signalling a translator’s professional status: willingness to invest money, proven language ability, feedback from previous clients, and investment in the community, with academic qualifications playing a minor role. Since most of these signals are publicly available, they function relative trustworthiness both internally (among peers) and externally (for prospective clients). The signals are thus helping to restore some degree of order to the online marketplace.
4. Case study 2. Online collaborative subtitling
The emergence of collaborative subtitling is strongly associated with the democratisation of technology. The first subtitling communities appeared in the 1980s as a part of a transgressive movement that was trying to import more and uncensored Japanese anime into the United States (cf. O’Hagan 2009). Despite the technological challenges of the time, volunteer communities managed to produce and distribute fansubbed versions of their favourite anime. It was not until the late 1990s, however, that interactive technologies started to fuel the production of amateur subtitles on a large scale, allowing the activity to include underground films, films produced in countries with small uncommercialised cinematographic industries, and TV series and films that, although produced by the major companies, were subject to extended delays for international distribution.
Here we are not concerned with the conflicting interests of the professional and volunteer subtitling communities (it is possible that they have quite complementary social roles), and we have no reason to believe that there is any significant difference in the quality of the subtitles they produce (Orrego-Carmona 2011). Instead, we are interested in way the online volunteer communities signal the trustworthiness of their subtitles, mainly for popular mainstream TV series and films. Communities need to establish a good public reputation in order to keep users returning to their websites and using their subtitles, and this basically requires that the users trust the particular subtitling community. At the same time, relations of trust have to be established within the online subtitling community, so that members have confidence in the quality of each other’s work. On both these levels, the signals of status are almost purely electronic.
The Internet has been essential for the existence of these groups. On the one hand, it helps bring together geographically isolated members of the community, enabling like-minded people from all around the world to interact with each other. On the other, the significant increase in bandwidth makes it possible to distribute large video files with relative ease. These two conditions allowed the development of international audiences for audiovisual products. Fans all over the world now access the content as soon as it is released.
Volunteer subtitling groups usually have two main components: translation and social interaction. The translating is generally performed off-line, although some communities do have online subtitling tools. The social interaction is online and it is organised around forums on topics ranging from the translation of specific series, to genre preferences, film critics and audiovisual habits. The forums are thus an important part of the community (Orrego-Carmona 2011). They allow for the exchange of long messages, with each member having a profile with the information they want to provide. All messages and exchanges can be systematically arranged, stored and re-organised and information is indexed and becomes searchable. Although the results of the translation activity, the subtitles, are publicly available to anyone on the Internet, the records of the social interaction are normally only available to those who are registered.
The forums are also where the online communities, initially predicated on opposition to official commercial culture, begin to form internal hierarchies. In principle, each forum has at least three types of members: administrators, who manage the technical details of the site and the rest of the members; moderators, who mediate the discussions, arrange and organise the threads and keep the forum clean of spam; and users or members, who can post and reply in the forum. Each level has different “permissions” to act in the site.
Group members are appointed to the posts of the hierarchical structure on the basis of experience and performance within the group. Since no previous qualification is required and the groups follow a learning-by-doing approach, it is assumed that the required skills are developed by carrying out activities for the group. In most cases, it is possible to rise in the hierarchy thanks to the amount of work performed and the time invested in the activities. In general, the more a member works with the community, the higher they rise in the hierarchy.
The recognition of volunteer communities is essentially linked to the amount of content they translate and the impact of this content. Groups depend on what they produce in order to attract new users and to become part of the non-professional subtitling landscape. Taking into account that one of the main reasons for the existence of non-professional subtitling is to overcome the lengthy delays in international distribution, groups should also produce the subtitles as fast as possible. There are clear signs of competition and tension between the groups as they strive to produce subtitles that are the fastest and/or of the highest-quality, and they thus aim to win the highest numbers of downloads and registered users. In these aspects, the communities are working rather like localisation companies (where speed is also a key factor) and television channels (where the user’s general preference is what counts).
Drawing on a qualitative analysis of two collaborative subtitling communities that have been operative for more than ten years, our study suggests that the growth of the community entails the unfolding of a more complex hierarchical structure. The communities analysed are aRGENTeaM and GroupTS/Tus Series. The two communities were formed in 2003, coinciding with the popularity of TV series produced in the United States that heavily influenced the creation of international audiences, such as Friends, Six Feet Under and The Sopranos and the release of others that rapidly gained global renown, such as Two and a Half Men, Lost, House and Desperate Housewives.
The aRGENTeaM forum (argenteam.net) was launched in 2003 and has been in continuous operation since then. It is based in Argentina and initially started as an Argentinean forum. The group produces subtitles for popular TV series and films, mainly from the United States and Europe. Most of the subtitles are translated from English into Spanish. The subtitles are publically available for free and users do not need to be registered in the forum in order to download them. The membership has grown steadily, with a significant increase 2007–2009 (see Figure 2). Currently the group has 534,348 registered members, which includes the people involved in the production of the translation and the users of the translation.
The forum was initially in the main page of the website. From 2006, however, the main page has offered the possibility to go either to the forum or to a search engine to look for the subtitles. The group thus clearly differentiates between the internal forum, for registered users only, and the external site for the wider audience.
Participation in the group is open to everyone who is interested. All registered users can access the forums, read the previous exchanges between members and post messages. Members who are actively engaged in the forum may then become part of the organisational staff, with a vote in the decision-making process of the group (Orrego-Carmona 2011). Staff members are also engaged in training new members and ensuring the quality of the subtitles. This is where the hierarchical organisation becomes very clear.
The group currently has six different types of roles in the production staff: administrators, moderators, revisers, translators, pre-revisers and junior translators. As the group grew, it was necessary to include levels that would reflect the progress people made within the group. In February 2009 the role of junior translator was created for active participants who had translated some segments regularly and could now participate in team projects. Junior translators can then be promoted to the position of translator. The training process also extends to the reviser role: the pre-revisers are translators who want, or are expected, to become revisers. These translators are assigned the tasks of a reviser but are supported and supervised by an official reviser. The reviser acts as a guide to make sure the subtitles comply with the group requirements and processes.
The status of members is also made very clear to other members. Each post includes an avatar, a nickname, the number of posts the member has made and the date when they joined the forum. When they change their type of membership (when a regular user becomes a translator, for instance), a label next to their avatar indicates this and the colour of their nickname also changes. In all, this information constitutes a fairly subtle type of signalling, enabling relations of relative trust to be established within the group.
GrupoTS (grupots.net) started under the name of TusSeries (“your series”) in 2003. Unlike aRGENTeaM, it was initially set up as a discussion forum only to share opinions about TV series mainly distributed in Spain. The subtitle section was a forum that ran parallel to the discussion threads but was less active. Later, the discussions about English-language TV series were made part of the main forum, as there was increasing activity by members doing subtitles or ripping them from DVD versions of the TV series. The group currently has 763,065 registered members (see Figure 2).
Figure 2. Users registered in aRGENTeaM and GrupoTS/TusSeries, 2003–2013
The site has always been a forum. Members have to navigate through the discussion threads to find the subtitles they want. One thread lists all the TV series that the group has subtitled, while information about the series is provided in the TV series thread. A team of translators is defined for each season of the TV series to be translated. Each team includes a reviser and at least six translators, and they may adopt group nickname for themselves, suggesting their wish to be identified as members of a collectivity rather than as individuals. The subtitles always include the name of the forum and, in some cases, also the nickname of the group that produced the subtitles.
Participation in the group is also open to everyone willing to participate. Collaborators in the translation section should be willing to commit to the tasks they are assigned and learn how to perform them, devote enough time to perform them and have the necessary linguistic knowledge to do the translation.
Since 2009 the configuration of the translation staff has been divided into two groups: the CheckTeam and the SubTeam. The CheckTeam are the revisers, who are in charge of managing the project and correcting the translations. The SubTeam rank includes the translators. Considerable time and effort is needed to receive these ranks. The group has defined the requirements for each rank in a clear-cut way: twenty translated parts from twenty different episodes are required to become part of the translation team, SubTeam. Members who achieve this rank are authorised to use a special member logo within the forum. This change was guided by the need to offer something in exchange for the time invested in the translations and to motivate others to translate more. The translator receives specific feedback from the reviser to adapt to the forum standards. To become a reviser, a translator needs to create a translation team for a TV series and ask for permission to translate the season. Members then need to complete the revision of six full seasons in order to become part of the CheckTeam. It also offers a rank identification badge, along with more permissions to handle threads in the forum.
Active and continuous participation is expected from the members of the SubTeam. If a member is not part of an active translation project at any given moment, they risk losing their position within the group.
These two examples indicate the extent to which subtitling groups are organised like professional subtitling projects. Participants have to comply with tight schedules and perform specific tasks within a clearly defined environment, which resembles the constant time pressure experienced by professional translators. Additionally, they are required to develop the necessary technical skills and are subject to constant assessment from peers. Since the decision-making process is still collective, interaction with peers becomes of great importance for the progress of the group.
Both aRGENTeaM and GrupoTS/TusSeries aim to produce subtitles with a quality similar to that of professional subtitles. They have defined and revised a set of subtitling guidelines that, although not as comprehensive as those of most subtitling agencies, indicate their concern for standardisation and compliance with professionally accepted practices. They have also proved themselves able to coordinate a tremendous amount of work carried out by a large group of people under considerable time pressure. In short, they might be considered professional in all but pay.
The strict control over the production process and the hierarchical system that both communities have developed indicate the way in which they present themselves to users as a structured unit. We should also bear in mind that users are increasingly more active and more demanding with the products and this affects their selection of subtitles (Orrego-Carmona 2014). In short, users’ trust in the product (indicated mainly by a steady number of downloads of the subtitles) reflects the status of the subtitling community, to the extent that the community brand subtitles acts as a quality label for the subtitles.
5. Case study 3: The Translator Scammers Directory
Electronic communication has developed several new ways of signalling and checking a translator’s status, both in the professional and volunteer sectors. The same electronic means have, however, also made it relatively easy to hi-jack the signals of status, enabling non-translators to present themselves as competent professionals. This is of interest to us here because it says something about what kinds of status are now considered to have market value, and are thus worth stealing. At the same time, we find sites that denounce the scams, indicating that those same signals are considered valuable enough to protect.
There have always been people who lie about their qualifications, in translation activities as elsewhere. Identity theft on a large scale nevertheless appears to be a fairly recent phenomenon in the translation industry. It has been spurred on by electronic technologies in three ways: online documents (such as professional certificates are easy to copy and modify; the online provision of translations means that face-to-face communication is no longer necessary (the client does not need to meet any translator); and online statistical machine translation means that raw translations can be produced for free and presented as if they were professionally postedited translations. Put those three together and you have a situation where fake “translation agencies” simply take the curriculum vitae of a legitimate translator, change the name and email on it (sometimes only the email), use it to get translation jobs, then send the client an unedited machine translation. These are the scammers of most interest to us.
Some of these activities can be tracked through the Translator Scammers Directory, which is website (translator-scammers.com) and a Facebook page, the latter being created in August 2013. The website is authored by the Translator Scammers Intelligence Group and lists 25 “listening posts” (presumably contacts who send information), mostly in North America and Europe. This is the part of the world that has an interest in revealing scammers and thus protecting signals. There are no estimates of how many scammers there might be in actual operation, but in April 2014 the website published 5,408 emails used by scammers, and the Facebook site was adding up to seven scammers a day. One company reports that “in one day we received 10 CVs, of which 8 were fake” (Scammers Directory Facebook, 6 April 2014). So this is a sizeable phenomenon. Indeed, the extent of the scamming is such that there is at least one online service (Kenax) that offers to help clients filter out false translator identities, thus adding a signal of legitimacy to the translators it promotes. In reality, each “translator” pays the company $99 to have their CV sent to “more than 15,000 email addresses” (this is called “CV Blast”). There is no sign of any filtering except the $99 paid, which appears to be the sole signal that the translator is not a scammer:
CVs sent through this service, about three a month, use this blue template (although custom is also possible) and whose subject always begins with [TA:], like this one, so that you know the applications are from real translators, who are serious about entering the industry and who have paid $99 for this service. They are provided with a free training package on how to provide a quality and reliable service, although many of them are already established translators simply seeking to expand their client base. (Kenax, at homeworktranslationjobs.com, consulted 30.04.2014).
So how long will it take a scammer to copy the template? In the meantime, like computer viruses, scamming makes money both for the perpetrators and for some of those who would appear to be fighting against them. Indeed, the Scammers Directory itself could make some money, if ever it wanted to accept bribes of the kind offered by this desperate scammer:
Your are [sic] a big problem of me and i wanna you to stop, we can make a deal. Stop of bothering me and let my work go, and we can agree about monthly payment for you!! (Translator Scammers Directory Facebook, 31 March 2014)
The Translator Scammers Directory presents a fascinating anti-world of professionalisation, where we see how the signals accrued by translators are effectively stolen and used in order to sell machine translations. The theft can be as simple as taking a certificate of ATA accreditation and changing the translator’s name (Figure 3). Simple enough, and quite effective: ATA accreditation does have a definite market value, at least in the United States, and this forgery is implicitly affirming that value.
So what other signals are considered worth stealing? Here we seek no more than a cursory overview of a rapidly transforming phenomenon. Of the 13 CVs immediately available to us on the Translator Scammers Directory, almost all make some vague mention of academic training but only four claim academic training in translation (two say they have it but do not mention where from, one has a copied academic degree, and the remaining one has been copied but claims it has been “verified”). In the same sample, only one CV has stolen membership of a professional association, while five give false lists of previous employers. If we were to use this as an indication of what signals the scammers prefer to steal, the order of preference might go as follows: references from previous employers, academic qualifications in non-translation fields, academic qualifications in translation, and membership of professional associations.
Of course, a complicating factor here is that the scammers are not going to give references that are easily checked. The forged ATA certificate is given away by the ATA’s online directory of members, where “Xiao Ruan” does not appear (strangely enough, neither does “Bin Li”, the translator supposedly scammed – don’t believe anything!). Similarly, reference to a specific degree (discipline, institution and year) can usually be checked with the institution. For the same reason, the “previous employers” tend to be multinationals that are so large that checking on one particular translator is virtually impossible. So should we disregard this evidence entirely? But then, the appeal of non-traceable signals applies to all the types, so the scammers’ preferences for certain signals over others should still be telling us something about which signals have market value.
Figure 3. Forged ATA accreditation
(Translator Scammers Directory Facebook, 25 March 2014)
There is also a significant geographical dimension to the areas in which this activity is being detected. Taking the first 100 scammers on the Directory, we find that the ten most scammed countries (the ones the identities have been stolen from) are the United States (11), Italy (10), France (9), Argentina (6), Germany (6), Spain (6), Denmark (5), Russia (5), the United Kingdom (5), and Brazil (4). These are all relatively rich countries with developed translation industries. The top scamming countries in our sample are Palestine (48 per cent of the sample – perhaps a transitory phenomenon), China (7 per cent), Egypt (7 per cent), India (7 per cent), with less than 4 per cent each for Romania, Argentina, Indonesia, Jordan, Korea and Malta. These might be seen as countries that are not quite so rich, with translation industries that are perhaps not as developed as are the computer prowess of the scammers. We note, however, that Argentina appears in both our lists, so we should probably not read too much into the details of this small sample. The general suggestion can only be that the most trusted signals tend to come from the richer parts of the world, and are imitated elsewhere. Poor countries steal identities from rich countries, while European names have high status (they are the pseudonyms that are most popular in the Middle East and Africa) and the false photographs tend to be of attractive young Caucasians. A white face with a European name in a rich country would seem to be worthy of trust, but this is perhaps also because the prospective clients are similarly white, European and rich.
So how are the scammers caught? In most cases they are given away by the IP address of their computers, which are not in the countries they claim to be in. Similarly, the country codes of the phone numbers also tend to be not where the translator is supposed to be located. Another clue is the PayPal address used for payments (“Please pay to my friend in the Middle East”), which is never the same as that of the “translator”. In other cases there are quite elementary mistakes, and not just in the atrocious English: one translator claims to live in “High Street” in Paris; another is a member of the “Chattered Institute of Linguists”; and over here we have a PhD in Nuclear Physics from MIT who will translate for 5 cents a word. And so on.
While official professionalism promotes an image where translators are equally trustworthy all over the world, the scam translators suggest that professional prestige is very asymmetrically distributed, operating in favour of rich economies. Our brief passage through the scam world also shows that, while electronic communication has obviously enabled it to flourish, the same electronic communication enables the scammers to be caught and denounced. The medium has its revenge, and some degree of trustworthiness may yet be restored.
These three instances suggest a common model: a change in communication technology led to initial market disorder, but over time the same technologies have enabled new hierarchies to develop and certain degrees of trustworthiness thereby to be signalled. ProZ.com has been responsive and creative in developing its internal signals, recognising the failings of its early models. The volunteer subtitling communities have been similarly adept at developing their own internal signalling mechanisms, effectively ensuring that non-professionals can produce work that is of “professional” quality. And the scammer community, which has been able to exploit the specificities of electronic communication on numerous levels, is effectively challenged by ethical use of the same electronic means.
That said, none of these case studies should provide much comfort to those who seek a status like that of the liberal professions (doctors, lawyers, engineers and the like). The kind of legally protected title that is based on an academic degree and membership of an association seems not likely to happen anytime soon in the case of translation. This is firstly because of the relatively high numbers of freelance translators and part-timers — estimated at 74 per cent and 60 per cent respectively in Europe (Pym et al. 2012) — along with a significant fragmentation of specialisations and the corresponding multiplicity of associations (Pym 2014). There are significant hierarchies and divisions of professionalisation within the sector, and they work against the fictional uniformity of a professional title. Yet it is also because, in the electronic age, the available signals of trustworthiness are not entirely trusted, with academic qualifications being of less market value as signals than is verifiable previous experience (as was also found by Bowker 2005 and Toudic 2012). New forms of signalling are being developed, and some of them are coming precisely from sectors that have been seen as the greatest threats to market equilibrium.
The message to be gleaned from our three case studies is that the professional status of translators, both above and below legitimate trust, will have to adjust to the signalling mechanisms favoured by electronic communication.
- Akerlof, George (1970). “The market for lemons: Quality uncertainty and the market mechanism.” Quarterly Journal of Economics 84(3), 488–500.
- Bowker, Lynne (2005). “Professional Recognition in the Canadian Translation Industry: How Is It Perceived by Translators and Employers?” Translation Watch Quarterly 1, 19–116. http://goo.gl/xIoGkZ (consulted 9.04.2015).
- Chan, Andy Lung Jan (2008). Information economics, the translation profession and translator certification. Ph.D. Thesis. Intercultural Studies Group, Universitat Rovira i Virgili. http://tdx.cat/handle/10803/8772 (consulted 30.04.2014).
- — (2009). “Effectiveness of Translator Certification as a Signaling Device: Views from the Translator Recruiters.” Translation and Interpreting Studies 4(2), 155–171.
- Foucault, Michel (1975). Surveiller et punir. Naissance de la prison. Paris: Gallimard.
- Ju, Elma Mingli (2009). “The professionalization of interpreting in Taiwan: A critical review of Tseng’s model.” Compilation and Translation Review 2(2), 105–125. http://ej.naer.edu.tw/CTR/ (consulted 30.04.2014).
- McDonough, Julie (2007). “How do language professionals organize themselves? An overview of translation networks.” Meta. Translators' Journal 52(4), 793–815.
- O’Hagan, Minako (2009). “Evolution of User-generated Translation: Fansubs, Translation Hacking and Crowdsourcing.” The Journal of Internationalisation and Localisation 1, 94–121.
- Orrego-Carmona, David (2011). The empirical study of non-professional subtitling. Minor dissertation. Universitat Rovira i Virgili. http://goo.gl/qB6v5D (consulted 9.04.2015).
- — (2014). “Subtitling, video consumption and viewers: The impact of the young audience.” To appear in Translation Spaces 3, 51–70.
- Pym, Anthony (2014). “Translator associations – from gatekeepers to communities.” Target 26(3), 466–491.
- Pym, Anthony, François Grin, Claudio Sfreddo, Andy Lung Jan Chan (2012/2013). The status of the translation profession in the European Union. Luxembourg: European Commission. Revised edition: London: Anthem.
- Risku, Hanna and Angela Dickinson (2009). “Translators as Networkers: The Role of Virtual Communities.” Hermes – Journal of Language and Communication Studies 42(2), 49–70.
- Spence, Michael (1973). “Job market signalling.” Quarterly Journal of Economics 87, 355–374.
- Toudic, Daniel (2012). Employer Consultation Synthesis Report. OPTIMALE Academic Network project on translator education and training, Université Rennes 2, Rennes.
- Toury, Gideon (1995). Descriptive Translation Studies and beyond. Amsterdam and Philadelphia: John Benjamins.
- Tseng, Joseph (1992). Interpreting as an emerging profession in Taiwan — A sociological model. MA thesis. Fu Jen Catholic University. http://goo.gl/ivSdg1 (consulted 9.04.2015).
- Witter-Merithew, Anna and Leilani Johnson (2004). “Market disorder within the field of sign language interpreting: Professionalization implications.” RID Journal of Interpretation 2004, 19–55.
Anthony Pym holds a PhD in Sociology from the École des Hautes Études en Sciences Sociales, Paris, and is Professor of Translation and Intercultural Studies at the Universitat Rovira i Virgili. He is also a fellow of the Catalan Institution for Research and Advanced Studies, Visiting Researcher at the Middlebury Institute of International Studies at Monterey, Professor Extraordinary at the University of Stellenbosch, and Walter Benjamin Visiting Professor at the University of Vienna.
David Orrego-Carmona holds a BA in Translation and a PhD in Translation and Intercultural Studies from the Universitat Rovira i Virgili, Tarragona, where he also teaches translation. His current research is on the reception of professionally and non-professionally subtitled audiovisual content, using eye-tracking and exploring the possible convergences between professional and volunteer translation.
Ester Torres-Simón holds a PhD in Translation and Intercultural Studies from the Universitat Rovira i Virgili, Tarragona, where she is currently teaching and organising research activities. She carries out research on the sociology of the translator, translation and reception of cultural items, Korean translation flows, and new technologies in the teaching and communicating of research.