Bulletin n° 10 - mai 2013

 


Translation Manuals and Drafting Style Guides at the European Commission

Tomáš Svoboda


ABSTRACT

This article is concerned with institutional translation practice in the European Union (EU) Institutions, especially the Directorate General for Translation of the European Commission. The main objective of this study is to examine the role that style guides and translation manuals play in this context. A section is devoted to literature dealing with the topic of such manuals. As a first empirical area of interest, an analysis of the EU Inter-Institutional Style Guide is pursued. Secondly, work-flow at the DGT is studied with special focus on the usage of style guides and translation guidelines, followed by an analysis of guidelines for the translation contractors. The latter shows that the amount of information offered to the external language services providers is extensive and, in some instances, neither harmonised, nor structured. This results in contradicting rules and even meta-guidelines. Thus, a suggestion is made to unify or harmonise those resources to achieve more clarity, better accessibility of the data, and thus to make them more effective. As a fourth field of interest, the future of style guides is sketched with respect to workflow automation and the proliferation of machine translation systems. The concluding remarks discuss some of the areas for future research.

KEYWORDS

institutional translation, drafting style guides, translation guidelines, work-flow, machine translation customisation



Introduction

It is assumed in Translation History research that translation schools (in the sense of groups of translators working in a certain period of time on common translation endeavours) elaborate their own translation rules and guidelines. In recent times, this is true for the Translation Departments of EU Institutions, located primarily in Brussels and Luxembourg, which undoubtedly constitute the largest group of translators under one institution or body ever. The language departments have in-house style guides in place to harmonise language usage. In addition, there is the Inter-Institutional Style Guide (IISG) that covers translation, editing, and drafting. The departments then issue language specific guidelines for their contractors. The guidelines do not follow a common pattern and differ considerably in thematic focus and volume.

In this article, the results of two analyses are presented and discussed, one in respect of on-line resources made available by individual translation departments in the Directorate General for Translation of the European Commission (DGT EC), the other on the language versions of the IISG. Further, the work-flow applied at the DGT is examined. It is broken down into segments, and at every stage, I investigate who the parties involved are and if and what guidelines govern the individual activities.

Before the actual analyses, I summarise the topic of style guides and to what extent it has been studied within the field of Translation Studies so far. To wrap the article up, future prospects of the implementation of style guides are examined with special focus on the ever increasing use of machine translation (MT), the topic of customisation, and workflow automation. Prospects for future research are added.

Research into style guides

Institutional Translation has been studied within the discipline of Translation Studies for more than two decades. However, as regards the specific situation of EU Institutions, K. Koskinen argues that “little research on EU translation has been carried out in translation studies” (Koskinen 2008: 27). In 2009, some were disgruntled with the fact that there is a lack of attention devoted to studies in institutional translation and that little use is made of the material provided by the translation services of the EU institutions (Felipe Boto et al. 2009; the study focuses on the DGT). Since rules are constitutive to the notion of institutionalised translation (cf. e.g. Koskinen 2008: 18, Halverson 2008: 343, and in a less straight-forward way Becker-Mrotzek 1990: 159) as well as translation quality (Sosoni 2011), over time they have been touched upon in several publications (see below).

Historical principles seem to play a distinct role in the research. This is only natural, since it is only when comparing today’s situation with the findings from the past that we can solicit general principles that will help to pinpoint, substantiate, and explain individual phenomena. As Ji-Hae Kang explains, “[s]ome generalizations may be adduced from such historical accounts: institutional translation is carried out by teams of individuals with complementary knowledge and skills, working under established procedures and translating on the basis of explicit principles and language guidelines.” (2009: 142; italics added). By way of example, the translation practices under the King James Bible project are examined with the observation that “[...] scholars [were ...] working with specific guidelines provided by King James I [...]” (ibid.).

Links between older translation school practices and today’s developments in the EU have emerged in Anthony Pym’s papers, although they have not been studied in detail. Pym argues that the cost of multilingualism was high in medieval Spain and was high in the 1990s when EU enlargement was expected. As Pym explains, “[w]e are paying a great deal for linguistic nationalism. The coming enlargement [1] of the European Union could make the price too high for the politics to remain workable.” (Pym 1996: 14).

Various types of “rules” have been described or mentioned: established procedures, explicit principles, glossaries, guidance, (written and unwritten) guidelines, guides, guiding principles, institutional ‘group mind’, institutional doctrines, instructions, manuals, norms, official guidance on (translation) policy, organized procedures, style guides, terminology requirements, translator’s handbooks, algorithms (e.g. automatic TM analysis, pre-translation), codes of practice, (EU) culture, customs, etc.

Ian Mason’s constitutive text (Mason: 2004) mentions several types of guidelines that govern the translational practice, either because they have been issued to the translators by the institution (e.g. glossaries, style guides, codes of practice) or as a result of a “development which grows over a period of years out of shared experience, the need to find common approaches to recurring problems or through advice and training offered to new employees.” (p. 470). He then also uses terms such as “institutional doctrines”, “guidance offered to translators”, “set of guidelines” and, before elaborating on transitivity in institutional settings, he makes an initial summary of the relatively few publications that presents “evidence of official guidance on translation policy” [2] (p. 473).

In Emma Wagner, Svend Bech and Jesús M. Martínez, Translating for the European Union Institutions (2002), although the topic of style guides has not been included as an index entry, there are instances that deal with drafting standards and clear-writing campaigns (pp. 72-75). The work is still a major contribution with respect to its overall relevance. However, considering some of its descriptions and statements pertinent to the present topic, evolving practice has brought about new developments. One example can be the then absence of demand management (p. 72).

Kaisa Koskinen puts forward a rather surprising argument:

There are, for Finnish translators, [...] no translator’s style guides or other documents stating explicitly how the translators are expected to proceed. The more fundamental strategic choices are left to the individual translator to divine from the general ‘climate’ of the institution and previous translations. The collective and intertextual nature of EU translations then ensures that no translator will radically deviate from the general trend. Even if there are no clear strategic guidelines, the translators are not free to use just any strategy they happen to prefer. Instead of planned and carefully considered strategic decisions, there exists a rather haphazard code of practice that most translators would probably not have actively chosen but that now weighs heavily upon them. (Koskinen 2000: 58; italics added.)

Paradoxically, whereas the Finnish presented a pioneering clear-writing manual in 1999 that served as the basis for translating and adapting the Swedish guide in 2001 as well as the Czech translation guidelines Pravidla pro překlad právních předpisů ES probably in 1999 [3], there seems to have been no Finnish translation manual around the year 2000 in place. K. Koskinen then specifies that around 2000, there were explicit “translator’s handbooks for at least Danish, Swedish and English.” (2000: 63).

In 2008 Kaisa Koskinen presented the results of her research within the Finnish DGT Department and again, she tends to mention only implicit “in-house norms” by reproducing a short passage of a dialogue that she had recorded: “C: [...] I again went to ask A that, hrrmmm, ‘olisi’, ‘tulisi’, ‘pitäisi’ [...], one of these three is forbidden here but now again I cannot remember [...]” (p. 143; italics added). The references to the physical movement (“went to ask”) and recalling, instead of consulting a written record (“I cannot remember”) imply that at that time, no written manual was available, where those fundamental issues of terminology and style could have been consulted.

There are some enlightening contributions in Terminologie et Traduction, e.g. Obrová, Pelka (2001), on the practices that had been applied when the Czech version of the Acquis Communautaire was translated prior to the accession of the Czech Republic to the EU. Other articles in the same volume cover the situation in some of the other candidate countries. As part of a clear-writing campaign, Wagner and Martin jointly published Fighting the Fog at the European Commission (1998).

Equally informative are the texts in Perspectives – Studies in Translatology of 2001, studies published on-line at the DGT portal [4] or Huiping Wu: Das Sprachenregime der Institutionen der EU zwischen Grundsatz und Effizienz (diss., 2004).

The obvious aim of external guidelines is to improve the quality of the delivered texts and ultimately to save time and resources on the part of the DGT as well as minimise the likelihood of extensive revision work. For example, the Guide for external translators uses the following words to express this:

Its main aim is to provide the contractors with practical information to help them with the translation work assigned by the DGT and to facilitate the communication between the contractors and the Commission (DGT’s Language departments and External Translation Unit), by laying down certain rules for standardisation (word-processing software, layout) and for the use of information technology. (2008: 4)

Having summarised some of the relevant literature on style guides, the focus will now turn to the drafting standard for all the EU Institutions.

IISG – A Manual for Drafting and Translation Processes

The Inter-institutional Style guide (IISG) – “a reference tool for written works for all European Union institutions, bodies and organisations” [5] – can be accessed via a link on the homepage of the Publications Office of the European Union (Publications Office). Its aim is the standardisation of linguistic practices [6] and it “contains uniform stylistic rules and conventions which must be used by all the institutions, bodies and agencies of the European Union.” (Guide for external translators 2008: 12; italics added.) The Publications Office attempted to narrow the gap between standardisation/harmonisation on the one hand, and specificities of individual languages on the other [7]. Printed copy versions of the individual language versions of the IISG are in the process of being finalised to mark a completed stage, however, the reader is reminded to follow the electronic version closely, since “the [...] work [...] undergoes a process of continual update” [8].

The IISG consists of four main parts and other sections (including, among others, the News section and Annexes). Our brief analysis of the structure and content of the IISG concentrated on Part One (the rules for editing the acts published in the Official Journal) and Part Four (language specific rules) and is based on 2010 data. The results are shown in Charts 1, 2, and 4.

Analysing Part I

In the printed version, Part I of the IISG typically has around 50 pages. Despite the fact that it is structured according to a common pattern, differences can be found in all the language versions at various levels. Five language versions of the guide were compared: English (EN), French (FR), German (DE), Czech (CS), and Slovak (SK). In comparing the five versions, four principles were examined: Additions, Specifications, Omissions, and Alterations.

CHART1

IISG, Part I

Language version

Additions

Specifications

Omissions

Alterations

Total Differences

EN

4

1

5

FR

4

2

2

4

12

DE

6

6

1

1

14

CS

7

3

2

12

SK

11

2

2

1

16


Under Additions, those phenomena are summarised that are unique to one particular version. As Chart 1 shows, the highest number of additions can be found in the Slovak version, whereas the English and French versions contained the same amount of added information. Applying the procedures of comparatistics, this category shows that the older the resource was the fewer additions there were (the Czech and Slovak versions were produced in the early 2000s, whereas the other three had been created some ten years earlier). [9]

With Specifications, features such as explicitations, elaborations, and explanations were solicited, showing that the highest number by far is found in the German version. This could be explained by the level of detail that is inherent in the tradition of producing similar text types, which is culture-bound.

The category of Omissions could be expected to show complementarity towards the one of Additions. Nevertheless, this circumstance was not confirmed, and it was found that there are categories that occur in all guides, with the exception of one or two. Since those shared categories were never missing in the English version, the assumption was made that the English served as the source for creating the other language-specific versions. There are not many variations to the Omissions category, and thus no significant differences were found between the versions.

In examining Alterations (e.g. different references used as compared to other versions), the English version, with the least number of alterations, seems to support our assumption that it is the source text. The other style guides show a very low number of alterations, with the exception of the French style guide.

Putting the four categories together, from the total number of differences between the style guides, the English version appears to be the source document with only five varying features. The scores for the French and Czech versions show an equally modest stance for producing a unique, non-derived model (a total of 12 instances each), whereas German with 14, and specifically Slovak with 16 differing features, show the most apparent signs of localisation/adaptation.

Analysis of Part I – Discussing the results

The French version contains the highest number of alterations, indicating that opposition or distance was sought with respect to the other language versions (at the time of its creation). On the other hand, an overall tendency towards independence and/or adaptation was not clear-cut when taking all categories together.

The analysis showed further that the German text draws on both the English and French versions, most likely to take advantage of the available information and, at the same time, to contribute with their own specific know-how (with the highest number of Specifications). The German style guide is very explicit and user-friendly.

The Czech version was the only one to incorporate explicit multilingual notes/remarks. It, therefore, had the second highest number of additions and specifications (e.g. punctuation, grammar). The Slovak style guide contains the highest number of additions due to references to court proceedings (especially the Court of Justice). It draws on the English, French, and German versions, whereas the Czech style guide rarely draws on any other version except the English one.

The language versions of the IISG are a clear example of a localisation undertaking, in that they avoid what is specific for other linguistic situations and/or add what is specific for the language in question. The assumption that the structure of the style guides in Part I is the same for all the versions has been verified with the five-language versions sample. At the same time, no dichotomy between the so-called “old” (EU-15) and “new” (EU-10 at that time) languages was found when comparing the sample of the five language versions.

Analysing Part IV

In Part IV (language specific rules), the structure is obviously much less regimented. The liberty taken here in creating language-specific manuals is apparent already in the titles the authors used for the sections:

– CS: Publikace v českém jazyce (Publications in Czech);

– DE: Richtlinien für den Schriftsatz (Guidelines Governing Printed Material);

– EN: House rules for the preparation of the text;

– FR: Présentation formelle du texte (Formal Representation of Texts);

– SK: Slovenské publikácie (Slovak Publications).

Major differences occur in the volume of the text as well. Whereas the German text is strikingly short (14 pages of the webpage printout), English and Czech ranked second with a comparable length (18 and 20, respectively). The French and Slovak versions are the longest with 24 and 43 pages, respectively. Comparing the DE and SK versions, it can be observed that the Slovak version is more than three times longer than the German version.

CHART 2

IISG, Part IV

Language Version

Number of pages

EN

18

FR

24

DE

14

CS

20

SK

43


Further analysis focused on the linguistic features that are under scrutiny in the manuals. Chart 4 of the Annex lists the categories as they appeared in Section IV of the IISG.

All in all, the five language-specific style guides were concerned with 32 categories of phenomena. When looking at the individual groups (Punctuation – Grammar – Style – Conventions/typography – Miscellaneous), the most represented group is Punctuation (14 categories, three of which are of interest to all five of the language versions). The areas of common interest, which are the same in all five style guides, include the following:

Nine of the categories (including Quotation marks and Ellipsis) are repeated in any four of the five selected style-guides. Interestingly, with the exception of one category (Hyphen), the other eight categories concern phenomena listed in all style guides except for the German one. Elsewhere, many other criteria are part of the style guides that are represented in one, two, or three of the style guides only. Singular occurrences include Question mark, Exclamation mark, Apostrophe (EN), Prefixes (FR), Style (CS), Capitalisation (SK), among others.

Analysis of Part IV – Discussing the results

In all five cases, liberty was taken in coining the individual titles for Part IV. “New” languages seem to be less concerned with explicitly stating national attributes in a European Style guide, indicating the language name in the title.

As regards volume in this part of the IISG, it can be concluded that the countries having a language policy in place exert considerable effort to protect their linguistic domain. This is true for both one of the “old” languages (FR) and one of the “new” languages (SK). However, a significant as well as difficult-to-explain exception is the shortest German version.

In terms of linguistic categories that are of interest to the creators of the style guides, approximately one sixth is shared across the board and roughly another quarter is mostly of importance to four of the style guides, with the exception of German.

Overall conclusions for the IISG analysis

Unlike in Part I where an identical structure was applied (with the most distinct alterations observed in the Slovak style guide), Part IV showed many imbalances due to the language-specific nature of the texts. For example, the French and Slovak versions made extensive use of “codifying” their linguistic preferences or introduced specific categories, as the length of the style guides shows. As was the case in Part I, the German version is distinct in Part IV as well. In the former case, the German style guide is specific and, at the same time, draws on the two other major language versions, whereas in the latter, it is rather restricted in volume and omits a high number of linguistic categories that the other guides in the sample apply.

The analysis shows a static picture of the situation in 2010. The IISG is, however, a dynamic medium, and the on-line version is updated almost every month. In addition, changes stemming from the Treaty of Lisbon have recently been accounted for, resulting in the creation of a printed version – the 2011 edition – which has become a new “base version”. Thus, as soon as more up-to-date research is conducted on the individual differences in the style guides, it will be possible to identify trends and tendencies in this respect.

Style guides and Translation Guidelines in the DGT workflow

Whereas the IISG aims primarily at the creation/drafting of texts and is binding for all Directorates General (DGs) of the Commission, the DGT has introduced manuals to govern the translation process. In the following, the workflow involving the translation process in its broader sense is presented and the applicable guidelines are incorporated in it.

The rules that determine both the drafting and the translation process in an institutional setting may be explicit and/or implicit. I have made an attempt to break down the process of both drafting and translating documents at the DGT EC and the requesting DGs. At every stage, I have sought to determine the agents involved and the explicit rules that apply. To this end, I used my own experience from a three-year in-house employment with the Czech Language Department at the DGT EC in Luxembourg. To obtain more recent data (my knowledge of internal DGT procedures dates back to 2004–2007), I have contacted the Czech Language Department and the S 3 division of DGT (Multilinguisme et Etudes de traduction). In a process of personal consultations and exchange of e-mails, I have arrived at the scheme that is presented in Chart 5 of the Annex. In broad terms, the scheme can be regarded as representing general practice in the DGT.

Translation Workflow

Initially, in my endeavour to map the translation process at the DGT [10], it was necessary to define its scope. In the narrower sense, the process of translating per se commences after a translator has opened an original text and either starts typing the target-text rendering or opens the first segment with his/her CAT tool. However, the translation process, when regarded as translation workflow, begins with the choice of texts for translation. [11]

DGT-External Procedures

Demand management helps to determine what texts will be translated and what texts will be communicated in one language version only [12], thus being subject to so-called “non-translation”.

To some extent, due to the complexity of the communication situation with EU texts, the very fact that a text is intended for translation means that the potential translation becomes implicitly part of the drafting process as well. For example, restrictions apply as regards the size of certain text parts as well as the text as a whole, and terminology is managed and harmonised in the drafting stage to facilitate future translation. With texts intended for on-line publication, translators in the Web Translation Unit [13] provide linguistic advice even before the start of the drafting procedure as well as editing and linguistic revision at all stages prior to finalisation of the original text. Although beyond the DGT’s practices, the CDT manual “Writing for Translation” is important for this topic:

For a text to reach its readers, it is essential for writers in a multilingual environment to keep the translator in mind. [...] Translation is not done in a vacuum: you [the writer of EU texts] and the translator can help each other. (CDT 2010: 1, italics added)

The key person in the drafting process is obviously the author. Very often, collective authorship is a wide-spread practice with regard to the importance of the texts produced (e.g. legislative documents). Drafting an original text is a dynamic exercise, since it involves consultations, recursive processes, amendments, feed-back, changes, comments, and editing/review at the originating department level. Under certain conditions, the product – the original-for-translation – tends to be far from edified/stable as well, and quite often, requests for additions, deletions, explanatory remarks, reference materials, and parallel texts are sent to the translation service during the course of translating.

Guidelines that apply to this stage: the Clear Writing strategy (drafting principles); The Joint Practical Guide for the Drafting of Community Legislation; IISG; Keep it Short and Simple (KISS); Fight-the-Fog; How to Write Clearly; Writing for the Web; automated or written tools/material; LegisWrite as well as any standard forms.

In larger projects, the team of authors can be joined by the Early Lead Translator, who is responsible for liaising with the client and acting as a contact person for translators in the future. A Terminologist may be invited to participate as well, whose role is to contribute to demand management and harmonisation of terminology in the original. Whereas the Early Lead Translator most probably does not have any guidelines at his/her disposal, the Terminologist does.

The inter-section of DGT-Internal and External Procedures

A draft is submitted to the Lawyer Linguists, who revise legal texts, provide advice on English legal terminology, and conduct legal analysis. One example of guidelines here would be the IISG. The draft original is also sent to the Editing Unit of one of the language departments responsible for the procedural languages (i.e. English, French, and German) according to the language of the original. The largest Editing Unit is operated by the English Language Department [14] and a number of style guides apply: For example, Clear Writing. In addition, a draft is received by Planning, an organisational structure that solicits requests for linguistic processing, e.g. translation, proof-reading, editing, gisting, summarising, etc., with guidelines in place: e.g. Demand Management. In these multiple stages, the draft becomes an original.

At the next stage, it is decided if the original will be treated by the DGT in-house translators or if it will be sub-contracted. Procedures external to the DGT and the Commission are filed in Tréfle, a facility maintaining contacts with free-lancers. Cf. below and Annex 5.

Process management systems are used to handle the large quantity [15] of texts for translation at the DGT: Poetry and Suivi. Only the Commission’s staff has access to the systems. Whereas the latter is a DGT-internal workflow system (to be discussed below), Poetry is “the software used for the electronic transmission of translation requests from clients to the Translation DG”. (DGT 2002: 4.) By entering the request into the system, the on-line interface generates a project (electronic folder) containing the translation request, the original document, and any reference documents deemed necessary.

The concept of originals as understood by clients tends to be fuzzy or dynamic. Many times, parts and/or versions of it are sent to the DGT after the submission of a core original, sometimes even very close to the deadline for the translation. Another instance of misusing the time allotted for translation is when an original is sent to both the translation departments and the Editing Unit at the same time. In the workflow, this will save the extra time that is necessary for the Editing Unit to polish the text, however, translators must add corrections to the text segments that they have translated meanwhile. This, obviously, is not an example of good practice, since it results in resources being wasted: one deleted word will multiply to 22 amendments in translations. Every change has repercussions for the context, which produce additional time expenditures, and increase the chances of generating errors. If a deadline remains unchanged, any added text segment would put more pressure on the translators; deleted text segments cause frustration, giving the translators the feeling that time and effort have been wasted.

DGT-Internal Procedures

The editable original then enters the stage of automatic pre-treatment. One or more translation memories are generated based on the textual characteristics. This is carried out automatically or semi-automatically by the Euramis Service (EURopean Advanced Multilingual Information System). Apart from its capacity as a central translation memory where all the translation segments of all language pairs are stored, this linguistic resource database makes accessible a number of other very valuable functions (e.g. extraction of references from a document, extraction of titles or text of corresponding legal acts from EUR-Lex, automatic alignment of downloaded acts, the Concordance function, and translation memories generated according to filter settings).

Once the request is filed in the system and automatic pre-treatment has finished, the language departments can access the relevant files and perform pre-treatment tasks at their own level (e.g. alignment of reference documents).

At the Department level, according to the topic and requesting DG, coordinators assign the task to a specific division (a Unit), and there, in a further step, it is assigned to a translator. The assignment is then entered into the DGT internal workflow system referred to as Suivi, which is “the software [named “Management of linguistic processing requests”] used for the electronic management of translation requests within the DGT. The program monitors the progress of a document and sends translations back to requesters.” (DGT 2002: 4.)

The translation process is substantially influenced by the Translation Brief. At DGT it bears the French title Fiche de Travail and in line with the relativistic definition of quality (depending on client’s expectations), a number of categories are specified for every assignment.

CHART 3

The Translation Brief (Fiche de Travail)

General categories

Process-specific categories

Processing Language Department and Unit

Source language

Requesting Directorate General

Target language(s)

Requesting Service

Original file format

Contact Person

Desired file format

Authorised Person

Type of service (Translation, Revision, Proof-reading, etc.)

Descriptors (Year, Number, Version, Part)

Quantity (Number of pages)

Original Document Title

Mode of communication (e.g. Publication in the Official Journal)

Date received with Planning

Document Type (e.g. Directive)

Date due

Type of Translation (e.g. Legislative)

Inter-service Consultation (Yes/No)

Quality Classification

Inter-institutional Procedure (Yes/No)

Note

 

Track Changes applied (Yes/No)

 

Translation aids available (e.g. alignment, pre-translation, MT, terminology)

Note: Own translation. The categories have been grouped for the sake of clarity; a standardised form is used with the description of categories in French; the table does not reflect the actual structure of the form. The Translation Brief accompanies every service request.

The Translation Process in the narrower sense

At this stage, the actual translation (eventually referred to as “interactive” when CAT tools are used) is carried out. Here, all the editing, search, CAT, and quality assurance tools come into play. An important part of the workflow is the Translator’s Desktop,anon-line interface and file management system where all the files related to the assignment are accessible for download: the original, either as one file or in parts, versions, reference documents, document comparisons, comments files, all the other language versions (translations), a note facility, and, of course, ongoing translations as well as documents already released and archived. The Translator’s Desktop is an interface that is closely associated with the translation process in the narrower sense. Since the translation file is opened from within this interface, the translator gets back to this interface every time they need to consult other language versions in process.

Translation Memory (TM) import is automated, and a pre-generated TM (retrieval) and other TMs (e.g. reference TM) get downloaded with the click of your mouse. Relevant software is used, i.e. a text editor, a TM system, and eventually other programs. A search is performed by using an Internet browser, including a meta-search engine (Quest). Some dictionaries are available off-line (pre-installed on the hard drive of the machine/work station).

Translators can order printed material they deem appropriate to have available in their office, including reference books and dictionaries; other such resources are available in the DGT library.

Telework is enabled at certain translation departments of EU Institutions (e.g. Parliament or the Translation Centre) where translators have secured Internet access to the restricted intranet and can use all the on-line tools as in the office. They are equipped with hardware (typically a notebook) with all the pre-installed software and authorisation facilities to be able to connect to the intra-Community network.

In-house translation problems are discussed with peer translators, and meetings are held on Unit/Department levels. Training is available at the EC/DGT/Department level, and specialised training may be obtained (e.g. sometimes national expert lecturers are invited). Networks with national experts in specific areas are in place to deal with terminology challenges.

As input devices, a keyboard and a mouse are used and dictating (voice recognition) HW and SW can be obtained on request. Scanning facilities and OCR software are available as well.

In terms of internal translation guidelines in this phase, the situation differs from one Language Department to the other. In the Czech language department, the choice is relatively wide: DGT Czech Style Guide, HR matters (output volumes), the Guide of the Coordination and Revision Centre (KRC) in Prague, a brief guide for new-comers, and others. Nevertheless, comparing this choice to the guides available on-line for external contractors, the latter are much more numerous. Thus, the following concept is highly probable: In-house translators are required to know all the relevant information that is pertinent to them and that is part of the style guides and translation guidelines available to external contractors. Apart from that, they are expected to be familiar with the in-house guidelines.

As a current matter of interest, a machine translation system [16] is being piloted. As opposed to the previously used engine, the current system enables processing of a much wider array of languages, including EU-10 (EU-12) languages.

Spell-check and QA software is available, either integrated into the text editor or the CAT tool. A specific format checking tool is the LegisWrite macro template.

Leaving the translation-proper phase

Before a translation is finalised and uploaded into the Translator’s Desktop file-management system, it is, as a rule, checked, i.e. revised or proofread. Under specific circumstances, it can be checked by an internal or external expert. In-house checks are performed by peer translators. With regard to their experience or specific expertise, certain colleagues tend to be preferred for revision skills as opposed to translation.

The translators receive their revised material back and add the proposed changes to the text according to their discretion or upon further consultations with the revisers. Either using an integrated function in the text editor or by manually sending an e-mail, the translators create a translation memory (typically “cleaning” the document), and in the Translator’s Desktop, they sign the assignment off and change the status of the file. They then inform the assistant who performs a final check on the text (completeness check, final verification). For evaluation purposes, the revision on paper may need to be archived.

Some Language departments have conducted a meeting on best practice in revisions, and the materials of this workshop form a part of the relevant guide list. For the final check, X-Bench helps considerably with the follow-up of translations with respect to their originals.

When the specific need arises, a Horizontal Review may be performed to achieve “multilingual concordance”. This involves an inspection of all language versions to harmonise their structure and to determine if the numbers and proper names match across the board. This can take place either at the level of the Planning Unit (TraDesk), the Lawyer Linguists, or the Secretariat General.

The TM produced is stored in the central TM system (Euramis) and is thus made available for future use. Eventual further uses of a finalised text include terminology extraction by the department’s terminologist. This person may also be asked to introduce some new terminology into the IATE term bank or to perform what is called “sentence management” in Euramis, which includes changing or harmonising the TM segments in the central database. For these purposes, a manual exists.

Using E-Greffe, another file/workflow-management tool, the text is then disseminated to its requesting DG. This stage marks the down-stream interface to the DGT-external workflow. If destined for publication (either in the Official Journal or otherwise), the text can be forwarded to the Publication Office. Some restrictions – as stipulated in guidelines – can apply here, for example, a communication embargo governing the exact mode of communicating a document or a group of documents.

Texts produced externally

The sub-contracting workflow must also be considered. The reasons for this are manifold. First of all, the DGT outsources a large proportion of its work (28%, i.e. approximately 590,000 pages in 2011; cf. DGT’s presentations on-line). The DGT is responsible for the final outcome, regardless of whether or not part of the procedure had been outsourced, and the majority of the DGT’s guidelines are made available on-line and are addressed to its subcontractors. Some of the guides include an explicit statement relating to observance of the guidelines contained in these guides as well as other guidelines. [17]

If a text or a part of an assigned project is subcontracted according to demand management or any procedures established based on good practice, it is processed by the DGT’s External Translation Unit. This alternative sub-procedure is summarised in Chart 6 and it replaces the “Translation” segment under the role of the Translator in Chart 5 (as marked by a frame). The description presented in this sub-chapter follows a tentative pattern. It does not represent the workflow applied at any specific Language Services Provider (LSP), nor can it function as an exhaustive account of agency-internal processes. Its aim is rather to compare/contrast, in general terms, the agency/LSP setting to that of the DGT. The model implies agency mediation, although individual translators can become free-lance contractors for the DGT as well.

The interface between the DGT and its external contractors is the eXtra Portal. Contractors can follow the whole life cycle of assignments, including new job announcement, allocation of translation order, delivery of the translation, quality assessment, and payment. On eXtra, it is also possible to check one’s own profile data and current ranking, as well as access instructions and other useful material on the DGT’s procedures.[18]

Upon receipt by the LSP/agency, a text is pre-treated and assigned a translator. Using an array of tools and resources, they then produce a target text rendering and revise it. For this, some will use their own manual (e.g. a new user of CAT software with notes from relevant training).

After post-treatment, the translation is sent back to the LSP, where revision and a new post-treatment take place. The agency’s Project Manager or Assistant perform the final check and upload the text to the eXtra portal, thus closing the sub-process. However, recurrences in what might appear to be a linear model can still readily occur, since consultations, quality feedback, and requests for improvement or updating are rare in the case.

For the above processes, an agency certified under the European Standard for the provision of language services (EN 15038) is obliged to have a Quality Manual. Even an agency or LSP not certified under the standard will have produced/formulated their basic requirements. The file, upon completion of the assignment and returning it back to the DGT, is expected to be in line with the Guide for External Contractors.

Contrasting the translation sub-processes carried out within and outside the DGT, one observes the reduplication of some stages, e.g. pre-treatment, post-treatment, final check. All, however, serve a relevant purpose (e.g. keeping TM within the agency, final check to fulfil one’s own responsibility).

After examining the DGT workflow, taking into account the impact that translation exerts on the drafting process, as well as depicting the translation process for both in-house and external translations, the study will now focus on guidelines aimed specifically at external contractors.

DGT Guidelines for Translators – Contractors

The Case of Czech

External and internal translators and revisers working for the DGT are expected to adhere to a vast number of guidelines. Taking the example of Translation Resources for contractors translating into Czech, currently there are 32 resources listed on the contractors’ webpage. [19] The quantity of documents/links suggests that the client strongly believes that the translation process needs to be managed in order to achieve the desired result (“high-quality translations”). While some of the materials showing good practice and using accumulated knowledge are clearly descriptive in nature (e.g. some of the material under the Contractors’ seminar presentations), the vast majority of them are clearly marked by normative features.

Despite the fact that the webpage is well structured (there are three main sections, each of which is divided into two or three subsections), difficulties could arise when trying to distinguish between the titles of the individual materials (cf. Handbook for translating into Czech vs. Brief guidelines for translating into Czech) or when trying to understand what their main contribution is (cf. Amendments – standard phrases). Some may point to identical targets, yet either linked directly, or indirectly, e.g. a link to the List of directorates-general and departments of the European Commission, which is part of the IISG with its own entry in the list of resource materials. Another example is the Translation checklist, which is presented both as a separate guideline and as part of the Handbook for translating into Czech.

Material in Czech is posted on the page back to back with resource materials in English. Surprisingly, there are instances of texts in English dealing with Czech language phenomena geared to translators [20], who are translating into Czech, cf. Translating online content.

Sometimes the links have the form of hyperlinks leading directly to the addressed page or file; sometimes there are files available for download, either in zipped packages or as standard pdf or Word files. The page is subject to review/updating, however, there is a tendency to increase the amount of data rather than reduce it.

Other web pages available for translators

Taking a more general look, language-specific web pages for translators/drafters using individual languages differ considerably in terms of the amount of data/references available as well as in the structure applied.

The large amount of rules/guidelines is made obvious by the fact that the German webpage features a document called Using style guides — in what order?[21]. It postulates the hierarchy of guidelines translators are expected to refer to and admits that “so far, [some regulations ...] are contradictory to one another”. (Own translation.) A hierarchical order of relevant guidelines is presented, the authority of which, however, depends on what institution produced the legal act in question. The need to unify terminology and procedures as perceived within the Language Department is demonstrated by the Übersetzungshandbuch (DG Translation German in-house style guide), which is 1,009 pages [22].

Other findings from examining the web pages pertaining to individual (target) languages include interesting files that are accessible to the public, e.g. Common translation problems – overview (in Polish) or the fact that there are language departments that publish bulletins/newsletters: the Polish (Co brzmi w trzcinie) is published on an annual basis, the Italian (Inter@lia>) approximately two to three times a year, and the Spanish (Puntoycoma) more frequently – every other month. Other bulletins include one in Lithuanian (Forum LTF), Maltese (l-aċċent), Portuguese (a folha), and Slovak (Network STS).

Reference sites covering languages that consist of language variants (such as French and German) typically devote extra resources to them. Furthermore, many of the published guideline texts do not indicate the version, date of production/publication, and/or author. This can easily result in losing track of the current and binding versions.

In this respect, it is rather a challenge for an external translator (freelancer) or agency/LSP to find their way through the material, and it may be suggested to unify/sort the valuable information that the DGT offers on-line, to allow for more intuitive and effective search procedures, resulting in a user-friendly interface and more accurate search results.

Pedagogical concerns come into play as well. When guiding students through the manuals and guidelines, as soon as they are confronted with the sheer quantity and size of the publications, they easily become discouraged when it comes to actually making use of the valuable resources.

Having analysed the IISG, the work-flow at the DGT, and guidelines for the translation contractors, the following chapter takes a step ahead and attempts to sketch possible developments in the field.

The Future for Style guides

First of all, development in the DGT will focus primarily on further workflow automation, the integration of language interfaces and services, closer cooperation at the inter-institutional level, as well as further improvement of the Translator’s Desktop to enhance the use of translation technology.

As the drive for Machine Translation continues to escalate, style guides will play an increasingly important role in the domain in the future. Chapters on controlled language will supply style guides with all the necessary rules governing the creation of MT-friendly originals. Although, today, translation style guides are communicated separately from style guides on the proper creation of original texts, this will no longer be sustainable in the future, especially in an institutional setting, as the creation of original texts is increasingly becoming a part of the translation process in its wider sense.[23]

Decisions will have to be made as to the quality vs. unacceptability of machine-translated segments of texts. Some MT systems ceased to be offered for free when used for commercial purposes (e.g. Google Translate API), and MT customisation [24] will most likely become a payable service as well.

Non-customised MT systems are main-stream nowadays. However, the future is bright for customisable systems, with customisation patterns as specific as those governed by individual users. Terminology databanks will be fed into the rules settings along with style guides [25]. Alternatively, the MT user will be able to set (e.g. using bars) mixed levels of components: adding word-level accuracy against idiomatic fluency (setting the paradigm to prefer narrower collocations/contexts/structures), focusing more on terminology consistency vs. stylistic variations, and showing multiple choices instead of opting for speed.

Customisation will gradually become more of a real-time experience. Should the translator decide that the MT output for a segment they are translating is not well placed, he/she can change the settings and re-enter the request. Users will be able to trial-error the settings as long as they arrive at the most beneficial renderings.

Admiringly, with style guide import, the real-time feature will lag, as those rules will have to be converted from an in-house-memory type of brochure into a standardised file (e.g. the “*.ini” type). This, most likely, will be carried out by specialised companies or experts. Once converted, however, legacy style guides will become obsolete and customisation (in a company/agency setting) will be a shared commitment for both linguists/Translation Studies experts and IT staff or trained specialists.

As human (bilingual) revision and (target-text) proofreading are differentiated, different rules will also have to be created for post-editing MT output (PEMT) as post-MT revision or post-MT proofreading.

Although the Czech DGT Department has recently banned the use of freely accessible MT engines [26], other EU sources speak quite a different language, though without reference to public MT:

As “[m]achine translations are a key feature of European Union policy”, and as “[g]iven the technological development, the maximum period for the development of high quality machine translations cannot be considered to exceed 12 years” [27], the EU drafting and/or translation manuals/style guides will sooner or later incorporate rules in this domain.

Outlook for future research

To extend the analysis of the IISG further, it is appropriate to seek consultations with the Publications Office to arrive at plausible and accurate explanations of phenomena that are difficult to derive from the linguistic, textual, and meta-textual spheres only. Research in the future might eventually focus on the following topics: What is the situation in the European Union Institutions when it comes to style guides and translation guidelines? Is it still true that there are language departments that have no explicit internal translation manuals? Where there are internal guidelines, what is the proportion between the information contained therein and that contained in manuals intended for contractors (free-lance translators)?

On a more general note, what are the challenges that translation practitioners face, the solution of which typically finds its way into style guides? In what way are revisions represented (bearing in mind their future, more prominent usage due to emerging MT)?

As there is an IISG binding for the drafting process in all languages (with language-specific chapters) as opposed to a high number of language-specific guidelines on translation (apart from the Guide for Contractors, which covers contractual circumstances in particular), will there be a tendency to standardise practices and come up with a uniform (albeit localisable) manual for translators as well? Will it be possible to determine the overall translation policy of the Institutions/DGT from such a guide? Are there national policies in translation across the EU? What will the relationship between national and European policies be?

To what extent is the policy of non-translation (as related to Demand Management) applied in individual institutions?

Finally, if in-house practices are codified, one of the most crucial questions is, do translators actually observe those rules? [28]

Moreover, since manuals cover areas such as Translational Quality, work-flow, linguistic challenges, reference materials, form, etc., inquiry into these distinct topics will need to be carried out to arrive at tangible results that can serve as a basis for further investigation.

Summary

Research on the EU’s institutional translation practice has been primarily carried out on EU-15 countries, i.e. before the enlargements of 2004 and 2007. Therefore, my attempt has been to make an inquiry into the question of whether the 2004 enlargement, which introduced massive changes in the DGT’s translation policy (cf. Mackiewicz 2008), has left the service with two rather distinct translation cultures, in other words, if there are differences between the “old” and “new” languages in terms of their institutional practice or codification of drafting/translation rules.

The analysis of the drafting style guides (outside the DGT) for the three procedural languages (DE, EN, and FR) on the one hand and Czech and Slovak on the other, showed no significant differences between those hypothetical groups. Rather the opposite was true: FR grouped with SK in certain aspects and CS with EN in other.

However, based on my personal experience (in-house translator in the CZ Department between 2004 and 2007) and a reference in the relevant literature (cf. Koskinen 2000), a different picture emerges. Whereas a relatively “new” department (CZ) started to create its in-house guideline from its inception, an experienced team (FN) seemed to have shared all the required knowledge implicitly. Should this be the case even now after more than a decade, and in other departments as well, difficulties could arise when employing new staff. There could be no explicit rules to adhere to, and the record of best practice could get lost with experienced colleagues leaving, taking with them the wealth of “house memory”.

There is no lack of guidelines addressed to external staff and contractors. On the contrary, with the majority of departments, relevant web pages show an abundant, though perhaps too complex choice of guides, manuals, style guides, resources, and instructions. This situation should be improved so that users can find the information they require quickly and in an unambiguous manner.

The prospects for the usage of style guides seem to indicate a paradigm shift. With the spread of freely accessible MT services and with their integration into CAT tools, the actual work pattern of a translator will gradually shift from “interactive” translating to PEMT. This change will have to be reflected in the style guides and guidelines and their mode of existence may potentially change from a brochure-like or an on-line resource to a set of strings and settings ready for import under customisable MT services.



Words of Thanks

I would like to thank Ms Catherine Vieilledent-Monfort of DGT S 3 (Multilinguisme et Etudes de traduction) and Mr Jiří Basch, Translator, Reviser, and Terminologist of the Czech Language Department of the DGT EC, for their valuable comments and for updating me on the current practice at the DGT.



Bibliographical references

Becker-Mrotzek, Michael (1990): Kommunikation und Sprache in Institutionen. Ein Forschungsbericht zur Analyse institutioneller Kommunikation. Teil I: Sammelbände mit Arbeiten zur Kommunikation in Instituionen und Monographien zu Beratungen in Institutionen. Deutsche Sprache. 2/1990: 158–190, 3/1990: 241–259.

CDT (2010): Writing for Translation. Luxembourg: Centre de Traduction, CdT [Translation Centre for the Bodies of the European Union].

DGT (2009a): Translating for a multilingual community. European Commission, Directorate-General for Translation. Available from <http://ec.europa.eu/dgs/translation/publications/brochures/index_en.htm>. (Retrieved on 04-06-2012).

DGT (2009b): Translation tools and workflow. European Commission, Directorate-General for Translation, Communication and Information Unit. Luxembourg: Office for Official Publications of the European Communities. Available from <http://ec.europa.eu/dgs/translation/publications/brochures/index_en.htm>. (Retrieved on 04-06-2012).

Drugan, Joanna (2004): Multilingual document management and workflow in the European institutions, University of Leeds 2004 <http://www.mt-archive.info/Aslib-2004-Drugan.pdf>. (Retrieved on 29-05-2012).

Felipe Boto, Maria del Rosario, Antolín, Martín José Fernández, et al. (2009): La traducción de textos normativos en el ámbito institucional de la Unión Europea [The translation of normative texts in the institutional field of the European Union]. Actas del Congreso Mundial de Traducción Especializada. Unión Latina y Red Mundial para la Diversidad Lingüística Maaya. Paris, 56–64. <http://dtil.unilat.org/cmte2008/actas/Actas%20CMTE.pdf>. (Retrieved on 29-05-2012).

Guide for external translators (2008): European Commission, Directorate-General for Translation. <http://ec.europa.eu/translation/documents/guide_contractors_en.pdf>. (Retrieved on 28-05-2012).

Halverson, Sandra L. Translations as institutional facts: An ontology for “assumed translation”. In Anthony Pym, Miriam Shlesinger and Daniel Simeoni (eds.), Beyond Descriptive Translation Studies. In Homage to Gideon Toury. John Benjamins.

Kang, Ji-Hae (2009): “Institutional Translation”. In Mona Baker and Gabriela Saldanha (eds.), Routledge Encyclopaedia of Translation Studies. London and New York: Routledge, 141–144.

KOM (2011): Proposal for a Council Regulation implementing enhanced cooperation in the area of the creation of unitary patent protection with regard to the applicable translation arrangements. KOM 2011 (216) final. accessible via Eur-Lex. <http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2011:0216:FIN:EN:HTML>. (Retrieved on 04-06-2012).

Koskinen, Kaisa (2000): Institutional Illusions. Translating in the EU Commission. The Translator. 6: 1, 49–65.

Koskinen, Kaisa (2008): Translating Institutions. An Ethnographic Study of EU Translation. Manchester: St Jerome.

Léger, Sylvie, ed. (1996): Vers un agenda linguistique: regard futuriste sur les Nations Unies / Towards a Language Agenda: Futurist Outlook on the United Nations. Ottawa: Canadian Centre for Linguistic Rights, University of Ottawa.

Mackiewicz, Wolfgang (2008): Translation as a strategy for multilingualism. 2nd DGT-Universities Conference.

Mason, Ian (2004): Text Parameters in Translation: Transitivity and Institutional Cultures. In: Lawrence Venuti (ed.) The Translation Studies Reader. Second Edition, London & New York: Routledge, 470–81.

Obrová, Pavlína, Pelka, Jiří (2001): Překládání práva ES do češtiny [Translation of EC Law into Czech] Terminologie et Traduction. 2.2001. Luxembourg: EC. (Bilingual parallel text CS–EN.)

Pravidla pro jednotnou úpravu dokumentů (2011): Brussels, Luxemburg: Publications Office. (Printed edition.)

Pym, Anthony (1996): The Price of Alfonso’s Wisdom. Nationalist Translation Policy in Thirteenth-Century Castile. <http://usuaris.tinet.cat/apym/on-line/intercultures/1996_alfonso.pdf>. (Retrieved on 30-05-2012).

Sosoni, Vilelmini (2011): Training translators to work for the EU institutions: luxury or necessity? The Journal of Specialised Translation. 16, 77–108.

Svoboda, Tomáš (2008): Ubi sunt homines? Poznámky k řízení kvality překladů u Generálního ředitelství pro překlad Evropské komise [Ubi sunt homines? On Translation Quality Management at European Commission’s Directorate General for Translation]. In: A. Ďuricová (ed.). Od textu k prekladu II. Praha: JTP, 143–151. (ISBN 978-80-7374-047-4).

Svoboda, Tomáš (2004): Barokní normy překladu se zřetelem k česko-německým literárním kontaktům [Baroque Translation Norms with Regard to Czech-German Literary Contacts]. Prague: Charles University, Institute of Translation Studies. (Diss.).

Tonkin, Humphrey (1996): Language hierarchy at the United Nations. In: Léger 1996: 3–28.

Wagner, Emma, Svend Bech and Jesús Martinez (2002): Translating for the European Union institutions. Manchester: St Jerome.

Wagner
, Emma, Tim Martin (1998): Fighting the Fog at the European Commission. Terminologie et Traduction 2.1998, Luxembourg: EC, 131–139.



Annexes

CHART 4

Those categories that were represented in all five language versions are reproduced with white font against a dark background. Categories that where shared by four language versions are highlighted with a darker background colour, whereas those represented in two to three versions are not highlighted. Unique categories are highlighted in a lighter background colour.

 Svoboda

CHART 5

Texts produced internally

Stage

Explicit rules?

By whom?

Draft (Translation in mind)

Clear Writing; Fight-the-Fog; (KISS); The Joint Practical Guide for the Drafting of Community Legislation; IISG; How to Write Clearly; Writing for the Web; (Automated/Written); LegisWrite + standard forms

Author(s)

None (Co-operation with client, future contact person for translators)

Early
Lead Translator

Demand Management

Terminologist

Original (Translation in mind)

IISG; data not available (NA)

Lawyer Linguists

IISG; Clear Writing

Editing Unit

Demand Management

Planning

Pre-Treatment / TM

Guidelines for the integration of Euramis/TWB in the workflow of translation units; Automatic (TM analysis, generation: algorithms, Pre-Translation); Semi-Automatic (OCR)

Euramis Service (“Pre-Treatment Team”)

Pre-Treatment / Department

e.g. on Alignment

Assistant

Translation

HR – Production guidelines, such as output volume per day/year, etc.

Department-specific.

Czech Department: IISG, DGT Internal Style Guide, “KRC” Style guide (by Czech Government Translation Service), LegisWrite, Standard forms, Terminology requirements/resources

Translator

Revision

General evaluation guidelines (working group ‘Optimising the evaluation of freelance translation’) and a computer-assisted marking aid

Horizontal Review – TraDesk

None

Final Verification, Completeness Check

e.g. X-Bench, LegisWrite

Assistant /
Translator

Horizontal Review (“Multiling. Concordance”)

NA

Lawyer Linguists

[Horizontal Review (“Multiling. Concordance”)]

NA (this stage is rare)

[Secretariat General]

Post-Treatment: TM

Automatic (algorithms)

Euramis Service

Post-Treatment: TM/Terminology

e.g. on Updating Euramis – “Sentence Management”

Terminologist

Post-Treatment: TM

e.g. on Updating Euramis – “Sentence Management”

Assistant

Distribution

e.g. communication embargo

Secretariat General; Respective DG /
E-Greffe

(Pre-Publication)

(IISG, other)

(Publication Office)


CHART 6

Texts produced externally: LSP (e.g. agency)

Stage

Explicit rules?

By whom?

The disseminating unit

Guidelines for external contractors

DGT:
External
Translation Unit

Pre-Treatment

Automatic: e.g. TM analysis, LegisWrite;
Manual: OCR, Pre-Translation, Terminology;
Written: DGT Guide for External Contractors, agency’s Quality Manual

Project
Manager (Assistant)

Pre-Treatment

e.g. on OCR, Pre-Translation, Terminology, LegisWrite

Translator

Translation

DGT Guidelines, agency’s Quality Manual

Revision

DGT Guidelines, agency’s Quality Manual

Post-Treatment

None

Revision

Agency’s Quality Manual

Reviser

Post-Treatment

None (Post-Alignment; de-briefing, Terminology

 

Project
Manager (Assistant)

Final Check

Agency’s Quality Manual, LegisWrite

Sending out

Guide for External Contractors (e.g. file formats)

The receiving end

 

DGT:
External Translation
Unit

 


1 - The so-called “Big-Bang” enlargement of the 10 new member-states in 2004, the biggest in the EU’s history.  [retour]

2 - He names four publications: Koskinen (2000), Wagner et al. 2002, Munday 2001, and Kidd 1997.  [retour]

3 - This seems to contradict the statement in Obrová, Pelka: “Pravidla
[] byla vypracována na základě švédské a finské verze teprve v r. 1998 [...]” (“Rules for Translation of EC Law were elaborated based upon the Swedish and Finnish versions.” (2001: 98; own translation.).  [retour]

4 - Cf. <http://ec.europa.eu/dgs/translation/publications/studies/index_en.htm>.
(Retrieved on 03-06-2012).  [retour]

5 - <Http://publications.europa.eu/code/en/en-000900.htm>.
(Retrieved on 23-05-2012).  [retour]

6 - Obviously, this was not easy to achieve: “Bringing about a process of linguistic standardisation with such a large number of languages appeared unachievable.” Martine Reicherts; Preface; <http://publications.europa.eu/code/en/en-000700.htm>. (Retrieved on 23-05-2012).  [retour]

7 - “
[A]ll the conventions and common working practices contained in this work have been elaborated by our specialists while according the greatest respect to each language’s particularities” (Pierre de Boissieu; Foreword); <http://publications.europa.eu/code/en/en-000600.htm>. (Retrieved on 23-05-2012).  [retour]

8 - Introduction; cf. <http://publications.europa.eu/code/en/en-000900.htm>. (Retrieved on 23-05-2012).  [retour]

9 - On the creation of a Czech Style guide cf. Obrová, Pelka (2001). As regards the creation of the style guide as a whole, cf. IISG (Pravidla 2011 or on-line). An interesting circumstance – although outside the current scope of DGT procedures – is the role of the Finnish and Swedish style guides: They were used as a starting point for CDT 2010: “The booklet is based on the Swedish document Tala för att tolkas. Skriva för att översättas, published by the Swedish Government Offices, Ministry for Foreign Affairs
[in 2001; 32 pp. in total]. The Swedish document is a translation and adaptation of the Finnish guide, Käännetäänkö tekstisi, tulkataanko puheenvuorosi? [Helsinki - Monila 1999; 36 pp. in total], written to celebrate the Finnish EU presidency during the second half of 1999.” (p. 1; italics added).  [retour]

10 - This endeavour is to a limited extent parallel to the account in Wagner et al. (2002, mainly pp. 83–91) and to that of Joanna Drugan (2004). Its purpose here is to put forward an updated version, because a decade has passed since the release of the former publication and six years since the letter. Obviously, there are new practices in place, both in terms of organisation and technology. Moreover, Wagner et al. do not aim at presenting the workflow of one institution, as their aim is to cover all the EU Institutions (cf. the title of the book and p. 83). Drugan’s portfolio, on the other hand, focuses on tools and technologies present in the workflow. The main motive here has been to use the description of the translation process as grounds for linking it with the discourse on style guides and guidelines. For diagrams of the process, cf. e.g. DGT 2009a or DGT 2009b: 23. An analysis of the translation process with regard to Quality Management cf. in Svoboda (2008).  [retour]

11 - With the usage of IT tools in mind, Joanna Drugan sees the process set in at a quite different stage: “The process
[the typical translation workflow for in-house translations] begins simply enough with a translation request.” (Drugan 2004: 4). It is yet to be noted here that the present study assumes source texts as originating in the requesting DGs, as opposed to texts that were drafted outside the DGs and still are sent for translation at the DGT.  [retour]

12 - Cf. “Selective translation” <http://ec.europa.eu/dgs/translation/translating/index_en.htm>. (Retrieved on 24-05-2012).  [retour]

13 - More information on the Web Translation Unit (Division DGT.D.2) in this context is available via <http://ec.europa.eu/ipg/basics/management/day_to_day/dgt/index_en.htm>. (Retrieved on 24-05-2012). Interestingly, even the Web Translation Unit published two guides on its website: Translation Requestor guide and Translation Dashboard guide[retour]

14 - The reason for this is the fact that a vast majority of originals, 85 per-cent, are produced in English (cf. DGT’s presentations, e.g. <http://ec.europa.eu/translation/czech/guidelines/documents/contractors_seminar_presentations_prague_20110317_cs.zip>. (Retrieved on 04-06-2012).  [retour]

15 - In 2011, the DGT’s output was 2.11 million pages (1,500 characters not including spaces; cf. <http://ec.europa.eu/dgs/translation/faq/index_en.htm#5>. (Retrieved on 24-05-2012).  [retour]

16 - The SYSTRAN system has recently been replaced by ECMT (short for European Commission Machine Translation). For the current developments, cf. DGT 2002b. Other tools, such as voice recognition or smart search tools, are available for the translation stage proper as well.  [retour]

17 - For example, the Guide for external translators: “The contractors are required to consult all background or reference documents or glossaries supplied or recommended by the Commission and to make every effort to consult any other information sources recommended by Commission’s Language departments. <http://ec.europa.eu/translation/documents/guide_contractors_en.pdf>. (Retrieved on 28-05-2012).  [retour]

18 - Cf. <https://webgate.ec.europa.eu/dgt/freelance/index/home.cfm>. (Retrieved on 28-05-2012).  [retour]

19 - Cf. <http://ec.europa.eu/translation/czech/guidelines/cs_guidelines_en.htm>. (Retrieved on 04-06-2012).  [retour]

20 - Whereas, according to the mother-tongue principle that is mostly applied at the DGT, these will typically be native speakers of Czech.  [retour]

21 - Hierarchie der Leitfäden für die Abfassung von Rechtstexten. <http://ec.europa.eu/translation/german/guidelines/documents/using_style_guides_german_de.pdf>. (Retrieved on 03-06-2012).  [retour]

22 - As on 23 May 2012. An explanatory file (Explanation) is attached again to comment on the usage of this both huge and valuable resource. Exceeding 1,000 pages, we believe this to be the most detailed and voluminous translation style guide to date.  [retour]

23 - Many originals are consumed by fewer recipients compared to their translations, and the translation component is much more tedious and costly than the actual production of the original.  [retour]

24 - Customisation to be distinguished from MT training (which, typically, takes place automatically today, monitoring user behaviour – acceptance or dismissal of a proposed solution, etc.).  [retour]

25 - An example of the request for customisation as regards house-style rules can be found in a European Commission’s call for tenders for clear writing tools: “call for tenders for clear writing tools – chiefly stylechecking software to assist individual authors, that is designed to promote readability and can be customised (for example, to incorporate house style rules)”. Cf. <http://ec.europa.eu/dgs/translation/workwithus/calls/planned/index_en.htm>. (Retrieved on 30-05-2012).  [retour]

26 - Cf. <http://ec.europa.eu/translation/czech/guidelines/documents/guidelines_czech_cs.pdf>. (Retrieved on 23-05-2012), point No. 10.  [retour]

27 - Statement made in the context of European patent applications and specifications; KOM 2011; cf. also, on the other hand: “Such machine translations should serve for information purposes only and should not have any legal effect.” (ibid.).  [retour]

28 - Ian Mason elaborated on this question in his article (Mason 2004) for the specific situation of the European Parliament. Obviously, further studies on other institutions must be carried out in this respect. However, from past empirical findings, we already know, for example, that during the Baroque period (cf. Svoboda 2004), the clash between theory and practice (norm and actual practice) was massive.  [retour]

ISIT - CRATIL

39 bis rue d'Assas
75006 Paris
+33 (0)1 42 22 33 16 
Design: Page18 Interactive
Le Bulletin du CRATIL - ISSN 2263-7591 2015-Tous droits réservés ©