Titela Vîlceanu 8 Evaluating Online Resources for Terminology Management in Legal Translation

The Romanian legal translation market has evolved steadily, expanding after Romania’s accession to the EU in 2007. We acknowledge the quantitative development as demand-driven (translation means service provision, it is commissioned, it does not take place in a social vacuum) as well as quality increase, although the two components show a significantly different degree. Adopting an in-depth approach, we may identify several factors that have contributed heavily to the shaping of the Romanian legal translation market, such as follows: – the need for legal translation – we use the term in its broadest sense, encompassing the translation of all kinds of legal documents, but laying special emphasis on the EU law – the time pressure (enforced deadlines) – the insufficient number of legal translation training programmes: at the university level, Bachelor’s programmes of the major higher education institutions across Romania include a relatively small number of specialised translation and terminology courses – business translation, medical translation, technical translation, audio-visual translation, general terminology course, etc. The situation is replicated at the level of the Master’s programmes, where a module in legal translation comes under the umbrella of specialised translation, and where legal terminology is not in focus – there is one notable exception, i.e., Master’s programme in English and French Languages. European Legal Translation and Terminology, University of Craiova, where translator training underpins development of legal translation skills, legal terminology management skills, on a par with domain specialisation (one-third of the courses is dedicated to familiarisation with different branches of European law and with European bodies) (source: http://cis01.central.ucv.ro/litere/oferta-educationala/ GhidTTJE2013.pdf).

The Romanian legal translation market has evolved steadily, expanding after Romania's accession to the EU in 2007. We acknowledge the quantitative development as demand-driven (translation means service provision, it is commissioned, it does not take place in a social vacuum) as well as quality increase, although the two components show a significantly different degree. Adopting an in-depth approach, we may identify several factors that have contributed heavily to the shaping of the Romanian legal translation market, such as follows: -the need for legal translation -we use the term in its broadest sense, encompassing the translation of all kinds of legal documents, but laying special emphasis on the EU law the time pressure (enforced deadlines) -the insufficient number of legal translation training programmes: at the university level, Bachelor's programmes of the major higher education institutions across Romania include a relatively small number of specialised translation and terminology courses -business translation, medical translation, technical translation, audio-visual translation, general terminology course, etc. The situation is replicated at the level of the Master's programmes, where a module in legal translation comes under the umbrella of specialised translation, and where legal terminology is not in focus -there is one notable exception, i.e., Master's programme in English and French Languages. European Legal Translation and Terminology, University of Craiova, where translator training underpins development of legal translation skills, legal terminology management skills, on a par with domain specialisation (one-third of the courses is dedicated to familiarisation with different branches of European law and with European bodies) (source: http://cis01.central.ucv.ro/litere/oferta-educationala/ GhidTTJE2013.pdf).
In spite of the reform of the higher education system (through the implementation of the Bologna process), of the change to competence-based curricula and the constant endeavour of universities as providers of translator training programmes to map the labour market and achieve harmonised translator's multilayered competence at the European level, having the EMT flagship model at the core (please visit http:// ec.europa.eu/dgs/translation/programmes/emt/key_documents/emt_competences_ translators_en.pdf ) -for instance, 5 Master's programmes in translation delivered by 5 major Romanian universities (more precisely, University of Bucharest - Admittedly, legal translators are authorised following no open competition, and the ever increasing market and societal demand and offer of highly qualified translators seems to be reduced to the natural selection or 'survival of the fittest' principle. At what costs? To our best knowledge, low professional relevance or visibility, a large number of poor quality translations, diminished translator's fees and turnover (also due to the large number of authorised legal translators acting as freelance or in-house ones or cumulatively), and reduced motivation for pursuing professional development come top-most and should spread wider and deeper concerns among policy makers and management authorities.
Needless to say those current and emergent competence requirements at the European level -which, once again, universities seek to meet via competence-based curricula, work placements and employers' consultation -are disregarded in the legal provisions in force. We reinforce the idea by mentioning that within the OPTIMALE framework, the online survey jointly conducted with the EUATC (European Union of Associations of Translation Companies), involving 680 employers from across Europe, revealed that 'A university degree in translation or related areas (and not simply a modern language degree)' is a prerequisite in the translation industry (source: http://www.ressources.univ-rennes2.fr/service-relations-internationales/optimale/ attachments/article/40/Public%20part_report_2010_OPTIMALE%204018-001-001. pdf).

Competence-related Considerations
The legal translator's multilayered competence underpins both a process-and product-oriented approach, although it seems that the product dimension has taken precedence, being the most visible part of translation as design-oriented, precise and measurable (complying with the specifications). A professional translator should master the methodological toolkit, conceptual frame and related terminology. The legal translator's accountability is expressed in terms of cost-effectiveness (efficiency) and effectiveness. Moreover, it should be understood that effectiveness and efficiency derive not only from an empirical approach, but they are also based on proactive behaviour, i.e., the translator's seeking to anticipate problems and identify re-usable or generalisable solutions. Therefore, the inward-looking perspective of the translator should be complemented by the outward-looking one (against a set of objective criteria).
One important component of the legal translational know-how in relation to the selection and use of available resources is information mining competence (in accordance with the EMT framework), broadly defined as the ability to retrieve and evaluate information in a multiplicity of formats and for a multiplicity of purposes. Legal translators should be able to decide what kind of information they need to gather in order to create a context for the source language text and to fully integrate the target language text into its situational context (ultimately reduced to skopos as translators should master top-down and bottom-up information processing strategies, adopting an action-oriented approach). They should be able to locate primary and secondary information sources by tracing available resources and eliminating unnecessary or outdated sources. There is a wide range of resource types that translators should be aware of (dictionaries, glossaries, catalogues, indexes, workbenches, parallel texts, etc.) in order to extract and manage relevant information.
Effective and efficient legal translators will become translation memory managers in the sense that they will be able to use translation memory software, storing translated texts as reference materials, creating databases with different task-related information sources, thus accelerating the translation process and fostering professional development (long-term orientation).
Another component to be activated in the evaluation of (online) resources for legal translation is represented by thematic area competence, i.e., sufficient field knowledge for functional adequacy. Accordingly, legal translators should develop discovery skills and search for information related not only to the topic area of the document in question (using keywords in the text), but also expand the search to detect the hierarchical structure of themes (in the form of thematic maps). The next step is to acquire knowledge of the general terminology in the field and general stylistic features of the document type, appply such knowledge deductively and/or by association, and further detect specificities.
Thematic analysis underlies an exploratory spirit, mental alertness to the myriad of relations between different themes that make a corpus consistent and intelligible.
Cumulatively, the two components, which we see to be interrelated, lend priority to observation over intuition, to process orientation over product orientation, integrating quantitative analysis to qualitative statements about the usability of resources, language use, terminology management and legal cultures.
Nevertheless, we should not become too optimistic or idealistic. Terminology mining is time-consuming (Picht in Austermühl, 2001: 102, endorses that it takes up 75% of the translation time) and in practice, more often than not, it happens that translators do not have the time to evaluate all the available resources, thus, the selection of reliable tools may be done based on prior successful experience, immediate availability, peer advice, etc. Besides, the dynamics of the European law making is another barrier in the translator's coping with the huge bulk of texts, evolution of legal concepts, detection of the national or supranational character of the legal term in question, etc.

Objectives
Our main aim is to raise critical awareness of the diversity, reliability, and usefulness of the information provided by the different types of online resources from a translationoriented perspective: bilingual legal dictionaries, multilingual legal glossaries (term bases), parallel corpora and interactive tools (discussion forums).
Secondly, we intend to raise awareness of the translators' research potential with a view to informed decision-making and closer mapping of theory to practice.
Thirdly, we aim to provide a toolkit to increase the legal translators' efficiency and effectiveness in the management of terminology.

Methodology
We shall use a mixed research methodology, combining both quantitative methodscollection of empirical data, statistics, and qualitative ones -review of mainstream literature, analysis of the reliability and usefulness of online resources, identification of patterns and relationships between online resources.
Our approach is mainly descriptive, yet, we would like the final recommendations to be considered from a norm-oriented perspective as guiding legal translators' behaviour due to their evidence-based nature. In this respect, the overview of online resources shall focus on the general features understood as strategic costs and benefits.
In the applied part, for the purpose of the current paper and due to space constraints, we shall restrict our investigation to two legal terms, i.e., court and tribunal, which may be considered partial synonyms, but which, nevertheless, are not freely interchangeable and describe different legal systems.

The burden of proof Bilingual legal dictionaries: English-Romanian, Romanian-English
Generally speaking, bilingual legal dictionaries provide decontextualised correspondents or standard equivalents, with little detailed information on typical collocations, legal areas, contexts of use, examples and differences between the source language concept and the target language counterpart. Even if equivalents provide a viable solution, their cope of application is not the same. At this point, we make the terminological distinction between correspondent and equivalent, the former being decontextualised and predefined, whereas the latter involves a context-embedded decision-making process and a gradient -there are no absolute or perfect equivalents, they may range from non-equivalents, partial equivalents (minimum degree) to near or optimal equivalents (maximum degree) (see Catford, 1965;Newmark, 1988;Sandrini, 1999;Šarčević, 2000;Groot & Laer, 2006, etc.).
Another shortcoming is related to polysemous items that may become a stumbling block for translators in the attempt to disambiguate. As noted by different scholars, bilingual dictionaries of law are even 'less informative than monolingual dictionaries' (Šarčević, 1989: 277). 'The dubious quality of legal dictionaries' (Groot & Laer, 2006) is assessed based on the system specificity of legal terms and on the need to frequently revise and update them.
A search of English-Romanian/Romanian-English legal dictionary detected a general dictionary incorporating legal entries (http://dictionare.4rev. net/?Get=tra&Dictionary_base=0&Dictionary_list) and EUdict (http://www.eudict. com/?lang=engrom&word=court). Entry: court In the first dictionary, section Standard dictionary, English-Romanian, we see that equivalents are provided at the level of the phrase. Comment: as far as the polysemy of court is concerned, its different meanings are listed alphabetically (with no specification of domain). The legal meaning is exemplified as follows: admiralty court -tribunal maritim court of arbitration -arbitraj (tribunal, curte) court martial -curte marțială court of justice -tribunal law-courts -tribunal main court -curte de onoare maritime court -tribunal maritim police-court -tribunal corecțional Comment: in subsidiary, an item belonging to the lexical field of tribunal is listed: ușier (la tribunal) -bailiff, usher of the court. Furthermore, we shall see, after consultation of glossaries (IATE), parallel corpora (EurLex) and interactive tools (Proz.com), that the equivalence of bailiff to ușier is referentially inaccurate.

EUdict -English-Romanian section:
(the) Law Courts; <~de ocol/pace aprox.>country court -judecătorie (în Londra) 1. jur. to rule smb. out of court. 2. fig. to exonerate smb. (from blame) -a scoate pe cineva din cauză country court -judecătorie (mai mică) court-martialled -tradus în fața curții marțiale in court -la judecată in the first court/instance -în prima instanță jur. to discharge an order of the court -a revoca o hotărâre a curții judecătorești to lay a matter before the court -a ridica o chestiune în fața tribunalului law court -judecătorie (sediu) court martial -curte marțială EUdict English-Romanian Dictionary: curte -tribunal tribunal a fi citat la tribunal -to come up before the bench ușier (la tribunal) -bailiff ușier (la tribunal) -usher to the court Comment: both noun phrases and verb phrases are taken into consideration. Note: In the case of EUdict, there is a disclaimer: 'EUdict (European dictionary) is a collection of online dictionaries for the languages spoken mostly in Europe (414 language pairs). These dictionaries are the result of the work of many authors who worked very hard and finally offered their product free of charge on the internet thus making it easier to all of us to communicate with each other. Some of the dictionaries have only a few thousand words, others have more than 250,000. Some of the words may be incorrectly translated or mistyped'. However, it also aims, rhetorically, to create a feeling of togetherness, of a community of practice, and it counts as a covert invitation addressed to those valuing the 'hard work' to contribute to this tool development.

Multilingual legal glossaries
Terms should be envisaged as depositories of structured knowledge in a particular field. With strict reference to legal terms, their meaning may be shaped and stabilised intertextually (by legislation); therefore, when two legal systems are involved, the source language term and its equivalent in the target language will have a different semantic potential (for instance, narrowing or extension of meaning). Moreover, the equivalent should provide information about the legal system as a knowledge base. Equivalents have their own lifecycle: they are created, adopted by a speech community or a community of practice and if used repeatedly, they entrench cognitive routines, becoming established (Molina & Hurtado Albir, 2002: 510). IATEthe EU's multilingual term base, interactive terminology for Europe (source: http://iate.europa.eu/) is a web-based tool developed by language services with the main aim of benefitting institutional translators and language staff in their tasks. IATE is inter-institutional in nature, as indicated by the fact that the entries specify the institution from which an input is retrieved. Besides, the reliability of each term and the date of registration are also made available, on a par with the definition of the original term. On the other hand, definitions of the equivalent term and illustrations of use in the target language are not common practice. Focussing on the analysis of the different renderings of the term court, we identified 72 entries, pertaining to different domains and sub-domains. The following examples are illustrative:

LAW, EU institution [Council]
Rules of procedure of the General Court (rated by 3 -reliable). Note: when the information is extracted from Eurlex, we notice the simultaneous recording of the source language term and target language term.

Parallel corpora
Parallel corpora (also known as translation corpora) are identified as source language texts aligned with the corresponding target language texts or translations. We focus on the EU law as overrepresented in online translation resources in comparison with other legal genres; therefore, in line with Kasirer (2001), we advocate legicentrism.
There are skeptical voices claiming that such corpora are of little assistance to achieve a high degree of 'descriptive adequacy' (Granger, 2003: 19) and that they have a limited applicability to forensic linguists and consequently, to legal translators, by virtue of the fact that legislation 'does not often need large corpora since, because of its conservatism and "formulaic form-function correlations", it may be "equally efficient and reliable" to conduct a manual analysis, for example on a single legislative act' (Bhatia, Langton & Lung, 2004: 207). They go further as to say that 'there is very little need for comprehensive or automatic linguistic frequency measures, as they are easily identifiable manually ' (2004: 212). To our mind, it may be the case with linguists and other language researchers (translation theorists, included), but it is unlikely that the industry will ever embark on such a task. Nevertheless, the authors admit that these corpora are most useful in 'researching intertextuality within and across a particular genre' (ibidem) -we infer that they may be explored and exploited to assimilate the characteristic features of legal sub-genres. What the authors overlook is qualitative analysis, which is able to provide data about institutional, social and cognitive factors, sharpening the legal translator's awareness of similarities and differences.
We acknowledge the existence of several reliable and useful parallel corpora provided by the EU, briefly described below, yet, we believe that EurLex deserves special attention as easily available, expanding in real time with the EU law text production and highly reliable for the validation of online dictionaries and glossaries search results.
The JRC-Acquis (https://ec.europa.eu/jrc/en/language-technologies/jrcacquis), freely available to be used and distributed for research purposes, claims to be the biggest parallel corpus (full-text documents, paragraph aligned) in existence, comprising selected legislative texts from 1950s up to date; as far as the pair English-Romanian is concerned, the revised corpus amounted to 19,211 texts in 2009, and the Romanian texts were processed by the Romanian Academy of Sciences, which makes them an authoritative source.
The DGT-Acquis (https://ec.europa.eu/jrc/en/language-technologies/dgtacquis), the exclusive property of the EU, a multilingual parallel corpora (full-text documents, paragraph-aligned) extracted from the Official Journal of the European Union, consists of documents collected from the mid-2004 to the late 2011.
EUR-Lex (http://eur-lex.europa.eu/), as stated on the home page, contains 'EU law and other public EU documents, authentic electronic Official Journal of the EU in 24 languages'. The full-texts are organised by domain, sub-domain, year of document, type of procedure, author (i.e., European body), and the type of act. There is constant concern for improvement -latest developments are announced with respect to quick search (for instance, by document number), expert search (re-grouping of fields in a more logical manner, addition of alphanumerical sorting and of the zoom functionality -the search terms are highlighted, etc.), higher visibility of national implementing measures.
With respect to our term search, curte is cited 68,782 times and tribunal registers 32,619 occurrences.

Interactive tools -discussion forums
We identified two relevant global translation communities of practice, namely Proz. com and Translatorscafe.com, where, inter alia, specialised translators engage in collaborative and reflective work by asking other peers terminological questions (often concerning context-dependent phraseological units) and getting their reasoned contributions so as to reduce uncertainty. One of the major strategic benefits of such forums is that they allow for an insight into what happens in the industry, while also providing solutions to recurrent terminological problems of professional translators. More experienced translators may also provide links to 'the recruitment of background knowledge' (Evans & Green, 2006: 160) and relevant legal knowledge structures, thus enhancing a more standardised translation.
We shall exemplify the proz.com search. As far as court is concerned, the total number of (KudoZ) results is 74. The information is not so easily retrievable as the domains are not ordered or organised in a coherent way. In the field of Law / Patents, it seems that the majority of queries concern the UK legal system (not our focus with legal translators' training). What is worth noticing is that the suggested equivalents are accompanied, in many cases, by explanations shedding light on the meaning of the culture-bound items. For instance, Court of First Instance is equated to both Judecătorie and Tribunal, and the author explains that both are concerned with 'first instance'. Some other voices seem also hesitant, while others clearly favour Tribunal, but the translator is finally left alone to make his/her own decision. The English term Tribunal (5 results) is also envisaged beyond the word level (in phrases and sentences). For instance: the award of the arbitration tribunal shall be binding upon the parties -decizia instanţei de arbitraj va fi obligatorie pentru părţi -clauze polița de asigurare . The Romanian counterpart is based on neutralisation (translating back as instance), and there is a specification of the sub-domain (insurance policy). More opaque culture-bound items are translated via explicitation, i.e., the abbreviation is conveyed in Romanian by the full name: HM Courts and Tribunal Service -Serviciul Majestății Sale pentru curți și tribunale, and clarification work is done via indication of the term reference: https://e-justice.europa.eu/content_fundamental_rights-176sc-ro.do.
It is obvious that such an empirical approach needs to be refined and become more systematic in order to provide more viable results.

Conclusions and Recommendations
Terminological problems illustrate the need to analyse changing situational factors and macro-textual parameters for decision-making at the micro-textual level. We advocate a more global and efficient intersystemic and interlingual information retrieval and processing in legal terminological work. The acceptability of the functional equivalent term from a reader-oriented perspective becomes the keyword 'to meet formal or conceptual correspondence priorities' (Prieto Ramos, in Cheng et al., 2014: 124). In other words, bilingual and multilingual resources are expected to be descriptive rather than prescriptive in nature, informing the translator's decision.
To wrap up, reliable online resources for terminology management in legal translation should provide macro-level (comparative) information about the legal systems and branches of law (text types and sub-types) in question, meso-level information about the institutionalised practices (roughly equated to standardisation work, external and internal terminological coherence), and micro-level information, such as language affinities (collocations), pragmatic contexts of use, reformulation options, etc.
Undeniably, parallel corpora provide a fresher look at lexicography, terminology and terminography since they reflect recognised translation practices (although the source language text and target language text are equally authentic from a legal point of view), containing translation units and their equivalents in the form of single and multiple units (concern for phraseology), and should be used by legal translators to complement and validate dictionaries and glossaries, especially because dedicated extraction software (CAT tools) accelerate terminological work.
Driving the agenda forward, we hope that in the near future, not only multilingual glossaries, but also online bilingual dictionaries will be based on parallel corpora to reflect the legal language and terminology in use. Last but not least, it is our hope that online terminological resources will boost the legal translators' research-driven performance for practical integration and systematisation.