Defining translation and web localization
The development and growth of web localization practices have meant that for more than a decade, industry experts have attempted to define what web localization entails. This was done primarily to justify the added value and added cost of web localization in the language industry. These efforts have often been built on simplistic implicit or explicit models of translation that have been considered as a generic or less complex process. Metaphorical devices are commonly used to define translation in relation to web localization, in which the former is viewed as a process focused on ‘texts’, ‘words’ or ‘language’. The most pervasive industry conceptualization is that there is adaptation added to the translation, in which web localization evidently entails the translation of texts plus an added level of adaptation, often called ‘cultural’. With some exceptions, the technological adaptations to the back end or programming of the website are not mentioned, even when this is the part that few translators can complete themselves. The objectives of these efforts to define translation vis-à-vis web localization is clearly different from those of any consolidated discipline like international marketing or business globalization.
For the purposes of the advancement of research at ABC Translations, we recognize the argument that that industry-based definitions and models are often of little use for our research purposes. The supposedly distinct features of web localization are normally part of other regular translation processes, whether it’s cultural adaptation, multilingual management, the adaptation of icons and graphics, or the use of technology. All these elements are part of a large amount of most translation projects nowadays. The question therefore arises: can the we offer a definitional foundation that can have an impact in the industry? Is a definition necessary at all to advanced ABC research in this area considering existing debates on the necessity of a definition of translation? Web localization is a distinctive translation modality that represents an objective reality. It exists independent of both any potential observer and theorizations or research efforts. Prototype approaches can help, as previously mentioned, in the daunting task of identifying a working operational prototype for research purposes through the identification of more or less central characteristics of web localization, assuming the existence of fuzzy boundaries and membership gradience. Adopting this approach means that in choosing to adopt a prototype category, we do not merely surrender the notion of being able to define our object now (which is what most people assume when they adopt a working definition). On the contrary, we surrender the notion that we will ever be able to provide an absolute definition.
That is, by adopting a prototype approach we consciously surrender the task of identifying a single, working definition which fits all definitions of web localization in the shifting world of online technologies.
A prototypical approach to web localization
The adoption of a prototypical approach means that the central features of the prototype need to be explored. For the purposes of providing this type of approach, I will propose several core characteristics which can be used to identify what web localization represents within the wider network of translation-related phenomena. The main guiding criterion has been precisely finding key features that separate it from other types of general or specialized translational practices.
Web localization operates exclusively on digital web genres, such as a corporate website, a promotional website, a social networking website or a dating website.
- Web localization is a digital WWW mediated activity. However, not necessarily all Internet mediated communications that are the object of translational activities are part of it, such as chat or email translation.
- Web localization normally entails processing interactive hypertexts, and therefore an analysis of new textual models on the web is required.
- Web localization entails a specific set of technological and management processes not shared with other translation practices (Web Content Management Systems and other web-specific technologies).
- Web localization in the industry encompasses both professional and non-professional or crowdsourced models.
Web localization is a challenging new process in which a myriad of translation types and modalities converge. Translation types are defined here as categories related to socio-professional environments (medical, legal, political, scientific, religious, literary, etc.), while modalities depart from the type of transformation and variation between the source and target text, like simultaneous interpreting, sight translation, audiovisual translation, audio description, etc.
Two main features effectively distinguish web localization from any other translational activity. First of all, web localization occurs fully on the World Wide Web, and other processes that are paper-based or Internet related such as chats, interpreted video chat, and email translation cannot be considered to be part of the core features of web localization. The second one, and probably the most important distinguishing feature, is the fact that web localization occurs primarily on a number of digital web genres associated with it, which according to genre theory, are genres that are used on the web. These genres, which include corporate, institutional, travel and even social networking sites, have become mainstream and are now popular and well known in modern societies.
Additional characteristics of this phenomenon are shared with other translational activities, and therefore, we cannot be considered them as a defining feature of web localization per se. Nevertheless, they are of great significance if we combine with any of the two core features above. First, texts are produced, translated, distributed, and consumed in a digital world. Other texts that are produced for paper or audiovisual distribution but merely posted online cannot be considered of significance to define web localization. This distinction tends to be difficult to interpret because hypertexts are open in nature. Consequently, any type of file can be uploaded or linked within a website. The translation of a piece of news that goes both on printed paper and on the website would not be considered a prototypical exemplar of web localization. In these cases, only the superstructure of any website, that is, the main hyperlinked textual structure that is conventionally associated with any exemplar of a web genre is of significance for our purposes. Rather, the focus has to be placed in the overall hyperlinked structure of the digital newspaper, together with its navigation menus and web functionalities. Consequently, the digital nature of any text or its distribution via the WWW cannot be considered per se a prototypical feature of web localization, due to the large number of other types of digital texts in our technological world, such as a GPS, smartphone apps or software.
The use of technologies is also important. In the early days of localization, one of the main defining features of localization often was the use of translation technology tools and the technological process. Nevertheless, the widespread use of translation memory and other content management systems has gone mainstream in the professional world of translation. Nowadays, a technological process cannot be considered the main distinguishing feature that separates web localization from other translational phenomena. In fact, now all computer-assisted translation tools can process tagged files of many kinds, allowing professional translators to work on html and other web-based files. Other localization types, such as software localization, do often require the use of localization specific tools like Catalyst or Passolo.
Another feature is the fact that web localization requires the cooperation of several agents to complete any web localization project. In large-scale projects, localization managers, localization engineers, localizers, freelance translators, and QA specialists might cooperate to transformation and variation between the source and target text, like simultaneous interpreting, sight translation, audiovisual translation, audio description, among others.
Web localization occurs fully on the World Wide Web, and other processes that are paper-based or Internet related such as chats, interpreted video chat, email translation, etc. cannot be part of the core features of web localization. Probably the most important distinguishing feature, is the fact that web localization occurs primarily on several digital web genres associated with it. These genres, which include corporate, institutional, travel and even social networking sites, have become mainstream and are now popular and well known in modern societies.
There is also the question of multiplicity of agents. This is due to the number of interrelated processes required to fully localize a large website. The fact that localization represents several interrelated processes with different agents involved is part of the definition of some analysts at ABC Translations. For example, some separate the technological and managerial stages in the process from those related to common translation: the actual translation of textual strings or textual chunks. From the twenty-nine steps or processes identified in web localization processes, ten of them are shared with other translational activities, such as terminology research or the quality analysis. Nevertheless, the multistep process that these scholars highlight should not necessarily be placed at the very core of the prototype, mostly since often one or two agents can localize a small website. Alternatively, in crowdsourced approaches the final resulting product might be the work of a massive number of participants, such as the cases of Facebook or Wikipedia.
Another feature at this level is the fact that web localization requires cultural and other adaptations. Cultural adaptations have consistently been presented as the additional component that separates translation from web localization in industry-based publications, even when cultural adaptation is an integral part of all translational processes. For example, publications by industry experts have claimed that localization entails the linguistic and cultural adaptation of digital content, even when the type of cultural adaptation has always been a normal part of many other translational activities, such as technical translation. Nevertheless, web localization can (but does not always) require a level of cultural adaptation; it can never be assumed to be radically different from other types of cultural mediation processes found in translation. It is also of interest to note that the industry has moved in the opposite direction, “delocalizing”, products from the start and making web texts intended for localization as culturally neutral as possible. This is to guarantee a smooth and efficient web localization process, free of expansive and time-consuming adaptations. Some recent studies have shown that the lack of cultural adaptations tends to be the norm rather than the exception.
The last two features are the fact that web localization is multimodal in nature and performed by professionals. The prototype of web localization, as happens with translation, assumes the professional status of agents. Nevertheless, recent developments in web localization have opened up to what is known as ‘crowdsourcing’, defined as volunteer translation produced in some form of collaboration by a group of Internet users forming an online community, often using specific platforms. Classic examples of websites produced using a crowdsourced approach are Facebook or Twitter. Other cases of non-professional web localization can be found in websites of non-profit and other small organizations that localize their websites for speakers of other languages within multicultural societies. Some efforts have even been made to create web platforms that can facilitate volunteer, non-profit social translation such as the one created by the Rosetta Foundation. Thus, even when the web localization prototype is often associated with a professional process, it cannot be forgotten that the participatory nature of the Web 2.0 allows any individual or collective to produce and distribute localized websites.
A of features inhabit the fuzzy area that separates debatable cases of what can or cannot be web localization. These are issues such as the use of volunteers and crowdsourcing models, or the fact that text is often extracted and sent to translators and then reassembled on the website by localization engineers. In this case, freelance translators often process texts in a linear fashion like other translation processes. Therefore, from a translation perspective, translators often work in a similar fashion to what is done in other processes. Nevertheless, the result is a web localization product resulting from the collaboration of several agents.
Keeping these prototypical features in mind, this approach allows us to categorize real-world exemplars of the category being described as more or less central to the prototype, or even as outliers that do not belong to the category. There are possible categorizations of real-world examples of localization processes following this prototypical approach. At the core of the web localization prototype would be a professional web localization of a consolidated digital genre, such as a corporate website or a search engine. These cases can be considered widespread agreed-upon exemplars located at the core of the prototype. Around this core exemplars might be cases such as the crowdsourced version of a social networking site, the localization by volunteers or bilingual employees of a small website, or even the localization of a small NGO website by one single person using translation technologies. There are cases of digital texts distributed through the WWW that would not be directly part of the prototype, such as the translation and reposting of user generated content such as tweets or subtitled YouTube videos or the translation of e-texts. Finally, it should be acknowledged that in the fuzzy grey area, some practices, such as a web localization using FOMT machine translation might not, as previously mentioned, be considered a case of web localization by scholars. Some industry practitioners or the public might classify it differently.