LEGAL TRANSLATION – A MULTIDIMENSIONAL ENDEAVOUR

Legal translation is a highly skilled task. It has even been described as the “ultimate linguistic challenge” (Harvey 2002: 177). However, law firms or corporations that procure translations from selfemployed translation practitioners often find the intricacies of the task difficult to perceive. Following extensive fieldwork examining how legal translation is commissioned and performed in ‘outstitutional’ contexts, I have developed a multidimensional model which illustrates the legal translator’s textual agency, aimed at conveying the complexities of translation performance to clients and other stakeholders. It may also serve to train fledgling legal translators, and to heighten practising translators’ awareness of their overall task. The impetus for the model sprang primarily from findings of serious information asymmetry and goal divergence in the market, and evidence that actors involved do not grasp (a) the need for legal translators to be fully briefed, or (b) the layers of skills involved. Juliette SCOTT: Legal Translation – A Multidimensional... 38


Introduction
During a recent global survey of the commissioning and performance of legal translation in "outstitutional" contexts (Scott 2016), it became apparent that the market is severely impaired. Translator briefing is on many occasions negligible and insufficient. In the procurement and 'production' process, legal translation practitioners are frequently relegated to an outlying position. Goal divergences and information asymmetries are rife. Moreover, there is minimal awareness of the extent of competencies required of the professional. In such a setting, it appeared worthwhile to seek out ways of better communicating with market stakeholdersto benefit not only translators, but also their clients via ensuing improvements in quality.
In this paper, I focus on the communication of competencies. I explore the legal translator's textual agencytheir intervention in performance of translation (sub-)tasksseen as the deft handling of four aspects: language, legal system(s), textual genre, and text purpose. Each of these aspects will be reviewed and, as a result of the inherent interdisciplinarity of the legal translation endeavour, wherever opportune mention will be made of related fields of scholarship. Owing the large amount of ground to be covered in the present enterprise, each aspect will be reviewed very succinctly.
I must stress that although these four aspects are discussed separately and in a certain order owing to the standard academic paper's format restrictions, this does not mean that I consider a sequential ordering of the translator's intervention to be appropriate. On the contrary, the cognitive processes involved are of necessity handled quasi-concurrently and contingently in actual performance of the task. For this reason, having appraised a number of portrayals and representations in the course of my research, I propose a multidimensional model that aims to encapsulate the complexities of legal translation performance in practice and communicate them to clients, trainee and practising translators, and other market and educational actors.

Conveying the complexities of legal translation performance
The legal translator's agency as it relates to the text is highly complex. At a first level, that of source and target language, the legal translator's task is analogous to that of translators in other domains, although situated within a specific segment of general language: legal discourse. Such discourse presents a host of linguistic particularities to be mediated.
At the second level, the legal translator's work becomes even more complex, as they 'negotiate' solutions between source and target legal system(s) and their respective concepts, which may be widely divergent.
Third, the legal translator must ensure that the given source genre or subgenre is appropriately transposed into the target genre, e.g., taking into account the relevant sublanguage, macrostructure and/or layout.
The fourth and final layer of difficulty in the legal translator's work is to ensure that the purpose of the source text is correctly reflected in the target text, if such a reflection has been requested by the client. Alternatively, the target text may serve a different purpose.
For example, if the source text is binding legislation, its translation may be requested for a "public website"and the commissioner will then need to specify whether they require the translator to produce a text that is accessible to the general public, and to an extent 'redraft' the text, or whether the translation should be a literal reflection of the source, such as might be produced for study by scholars of comparative law.
Let us consider this highly complex negotiation or 'juggling' of what is effectively eight multifaceted aspects by the legal translator four on the source side and four on the target sideas they progress through their work. The legal translator does not negotiate these multifaceted sub-taskslet us represent them as cubesin a linear manner. As mentioned in my introduction, they are handled quasisimultaneously: time is therefore a non-excludable factor, and we may further specify and refer to "relative simultaneity"in simpler terms n operations happening at the same time but in different frames of reference.
Having searched exhaustively for an adequate model, and bearing in mind the points raised by Chesterman (2013) in his chapter on modelling translation processes, I should like to refer to a concept from mathematicsthat of the tesseract, also known as hypercube. The tesseract is the four-dimensional analogue of the cubein other words the tesseract is to the cube as the cube is to the square (Darling 2004: 316), where time is the fourth dimension. Given that time is one of the dimensions, a static geometrical representation (even if it were three-dimensional) is totally inadequate to model the figure. I therefore need to include here an animated video model (https://commons.wikimedia.org/wiki/File:8-cell.gif#/media/File:8cell-orig.gif). The rotating perspective projections being wrapped and unwrappedconstantly interacting with and mapping on to one another -can be a convenient way to convey the legal translator's mental processes and the indissociable and synchronous nature of the fields of their textual agency. The shifting net of the polytope's vortices, mapping onto each other in an ever-evolving way accentuates the dynamic transformation of the text from source to target. I must add, however, that in this case the tesseract takes no account of the mathematical arguments involved; it is used purely to model complex cognitive activity. 1 Figure 2 below is a simplified representation of that polytope.
1 If this seems incongruous to the reader, it is worth noting that neural network modelling is at the heart of research leading technological advances in machine translation and speech recognition (e.g., Sutskever, Vinyals and Le 2014). For a visualisation of how billions of neurons interact in neural networks: https://www.youtube.com/watch?v=vyNkAuX29OU. In the above figurea venture to set down the multidimensional model on paper for the purposes of discussionthe flows represented by the arrowheads further extrapolate the legal translator's evolving quest to negotiate and transfer the linguistic/legal systemic-conceptual/generic/purposive aspects of the source and target texts in the most appropriate way. The myriad directionality of the flows seeks to underline the fact that translators' agency may itself have an effect on the fields involved. For example, "contact-induced issues" arise in Eurolects -EU legal language variants generated through translation and/or transposition (Strandvik 2015;Biel 2014). Similarly, textual agency through translations may affect legal systems. A good example of this is the case law of the Court of Justice of the European Union (McAuliffe 2013). Although the practice of "genre bending" (Bhatia 2014) has been evoked mainly in the context of changing societal values and practices (Garzone and Ilie 2014), the bending of legal genres may also occur through the 'weight' of or availability/access through translation: e.g., arrest warrants (Garrido Rodríguez 2012); Pakistani birth certificates in Spanish translation (Mayoral Asensio 2003)or crime fiction in the case of Nordic noir. Lastly, whilst the translator cannot retroactively affect the purpose of the source text, their very enterprise will greatly impact the extent to which the target text purpose is fulfilled.
I have already determined that time constitutes the fourth dimension in this model composed of three-dimensional 'cubes' representing the multi-faceted textual and contextual aspects negotiated during legal translation performance. In order to embrace fully the 'negotiating flows' in the previous paragraph we may take the model a step further and introduce a fifth dimension. I suggested earlier in this section that the translators' sub-tasks are handled in a relatively simultaneous mannerbut they are also handled contingently. Moving away from the tesseract for a moment, let us consider the gravitational pull exerted by planets in a solar system, each with its own specific path. In the tesseract model, we can embrace such pull or influence by introducing the idea of "mapping tension". Such tension allows the model to take into account the fact that the language, legal system, genre, or purpose may have a greater or lesser influence in different translation situations.
Readers will note that in this paper I do not enter into discussions of translation or linguistic equivalence, and limit myself to a brief review of scholarship concerning legal equivalence. In view of my focus here on cognitive "juggling", I have adopted less controversial terms such as "flows", "convey", and "negotiate". As grounds for this approach I refer readers to Šarčević 1997;Snell-Hornby 1988Gentzler 1993;and Baker 2001: 5-6. I take the view that the term 'equivalence' itself sits uneasily within a legal translation context: its inherent duality is (a) at odds with the multidimensional nature of the task, and (b) needs to embrace the basic tenets of comparative law (e.g., De Groot 2006: 423-433). Moreover, it is likely to embroil communication with non-specialist market players owing to its absolutist connotations. Indeed, Cao claims that "no exact equivalence or complete identity of understanding can be expected or is really necessary" (2007: 35), while Kjaer observes that "[e]stablishing equivalence between legal texts across languages is as impossible as squaring a circle" (2008: 67). As Gémar, whose work on the theory and practice of legal translation extends over several decades, recounts: "So, is any translation method to be recommended that guarantees full equivalence? Clearly not, and in legal translation even less so." (2012: 71, my translation, maintaining original emphasis). 2

Hacking through the thicket of legal language(s)
The translation of a legal text transfers the natural language in which it is written into the natural language required. This aspect of the legal translator's work is similar to the work performed by translators in all fields. As in many other translation specialisms, the source and target language variantsfor example, Belgian French, Swiss French or Canadian Frenchshould be specified. This is especially important in legal texts, because of the related problems engendered by differences in source and target legal systems, styles, concepts, and terminology.
In his reference work on legal linguistics Mattila points out that, whilst legal language is based on "ordinary" or "natural" language, it exhibits: "linguistic norms (phraseology, vocabulary, hierarchy of terms and meanings)" and specific "morphosyntactic, semantic and pragmatic" features (2006: 3). Owing to limitations of space, I will discuss only a small selection of these features with particularly strong relevance for legal translation. Whilst the examples of general linguistic features of legal discourse in this section are mainly taken from English, there are, of course, parallels in other languages (e.g., Mattila 2006;Galdia 2009).
Collocations play a large part in the acceptance or refusal of translated legal texts by those receiving them (as noted, e.g., by Biel 2010b). To stress their significance in legal discourse: "collocations with a specialised legal sense are the types of word combinations that are most frequently found in legal texts of all genres" (Yunus andAwab 2011: 159, citing Kjaer 2007: 509). As well as arising in running legal text, collocations may occur in terms referring to legal concepts or in legal maxims. I use the term "collocations" to refer to "recurrent word combinations" (Goźdź-Roszkowski 2006: 139), also called clusters, lexical bundles, word partners, compound terms, ngrams or colligations. 3 Collocations are not only recurrent in legal texts, they are quite inflexible. There is a specific type of 'inseparable' collocation in the legal domain: doublets/triplets or binomials/multinomials, e.g., from Brazilian Portuguese: "perdas e danos". Unlike many terms that can collocate with others or be used alone, in certain contexts these cannot be used without their 'partner'. The historical reasons for and value of such apparent redundancy will not be discussed here (see Tiersma 2000). A practical example raised by a translator participating in a 2011 study (Scott 2016: 64) shows why collocations are important in legal translation: a translator may know part of a target-language term, but not its collocate -such as whether to use "hold harmless from" or "hold harmless against" (or indeed "hold harmless from and against"the term often contains the doublet). Metaphorical expressions may be appended to this group as they too are composed of several lexical units and occur within a close span of a word. They will be discussed later in this section.
In a number of cultures, perhaps one of the most obvious features of legal discourse to neophytes is the use of archaic language: compound adverbs such as "heretofore" or "therein", and prepositional phrases such as "notwithstanding" and "pursuant to" (Alcaraz and Hughes 2002: 7-9). Tiersma refers to "antiquated morphology" (e.g., "witnesseth") and "formulaic subjunctives" (i.e., "be it known") (2000: 87-95). A related feature, owing to the origins of many legal systems and some 'cross-fertilisation', is the presence of Latinisms, although the frequency of their use depends on the natural language or legal culture in question.
Translation issues may also arise with terms that are monosemic, e.g., "estoppel", "tort", or "usufruct"; and those that are polysemic, with separate everyday meanings such as "consideration" meaning payment, "construction" meaning interpretation, or "issue" meaning heirs. When translated, as Alcaraz and Hughes point out, terms may move from being monosemic to polysemic and vice versa (2002: 17).
While the terms above cause confusion because different concepts can be expressed with the same word, confusion in legal discourse can also arise, particularly when translated, as a result of the use of synonyms as 'redundant' words, occurring in multinomial expressions (often strictly collocated), or as a result of drafting stylethe practice of "elegant variation". This may cause issues in translationfor instance where a synonym does not exist in the target language, or problems of understanding where textual cohesion and/or the drafter's intention is not clear, or where a translator elects to 'clarify' by replacing occurrences of synonyms with the same word, thus unilaterally eliminating ambiguity in the source text. Partial synonymity further complicates matters.
Despite some evidence of a shortening in recent years (Barnes 2016), long and complex sentences remain a prominent feature of some legal genres, as do unusual word order, the omission of articles, nominalisation, passivisation, use of "shall", and layers of embedded clauses. Such sentences may cause translators difficulty in deciphering anaphora, and in maintaining cohesion and coherence. Depending on the language pair in question, some measure of 'localisation' may be needed, according to the intended purpose of the translated text.
Metaphor also figures widely in many genres of legal discourse. I tend to disagree with Mattila when he holds that "in modern legal language, metaphors in particular are rare"he himself asserts that "[a]dvocates in the Romance countries use these images fairly often" (2006: 75-76). Vespaziani goes as far as to claim that "there is no such a thing as a non-metaphorical legal language" (2009: 1). Similarly, M. R. Smith found that metaphor was so widely used and studied in legal discourse that he set out to identify different "levels of metaphor" (2007: 921). Unfortunately for legal translators, metaphorical language is also highly challenging to translate (e.g., Schäffner 2004).
My final point in this short selection of legal linguistic features that are particularly sensitive in translation concerns rhetoric and rhetorical devicesessential linguistic tools for lawyers in many legal cultures. To cite a few examples: "paradoxes", "deliberate stylistic faults, plays on words", "surprise arguments", and "deliberate howlers" (Mattila 2006: 38-39). Additionally, multiple negatives, layered embedding of clauses, and ambiguous anaphora can all be called upon to muddy the waters. Such confusing mechanisms are in danger of multiplying their effects when this kind of discourse needs to be translated, particularly if the translator is not informed whether the intended end-user is 'friend or foe'for example whether the target text is for the adverse party in litigation or for a colleague.
Equivocality and language risk "Law is language and language is imprecise" (S. A. Smith 1995, citing Wesel 1992. The latter phrase was not, as one might imagine, uttered by a translator lamenting their lot, but by a law professor. There is a huge body of literature and constant debate on the interpretation or construction of the law. Apart from the possibility of ambiguous meanings being conferred to terms by those enunciating them, we must also add the complexity of what those receiving them understand. This is clearly explained by Engberg in the context of statutory interpretation and in a critical analysis of strong language theory: "The problem is that the meaning of texts can only exist as a construction in the minds of individuals, built on the basis of perceived underspecified textual signs and existing mental models"; hence "only if sender and receiver have near identical systems are they able to understand words in the same way" (2004: 1142).
Linguistic indeterminacy is broken down by Endicott (2000) into the following terms, inter alia: imprecision; incompleteness; incommensurability; immensurability; contestability; family resemblances (relating to common sets of features); and dummy standards (provisions presupposing a standard but not laying down a standard), while linguistic indeterminacy in American statutes has been categorised by Solan, using examples from case law, into areas such as "syntactic ambiguity, semantic ambiguity, ambiguity of reference and vagueness" (2011: 2). These analyses, however, draw examples mainly from the interpretation of laws by the courts rather than from texts drafted by lawyers such as contracts or pleadings where indeterminacy and equivocality may even be strategic.
Indeterminacy may arise not only out of wording, but also out of choice of language or language variant. There is a growing area of legal research devoted to language risk examining, for example, which party carries such risks of interpretation in contract litigation, or what happens when terms cross borders. This is referred to by French comparatists as "risque linguistique" (e.g., Mauro 1998), and in German as "Sprachrisiko". A well-known example is the "Socks Case" in German case law. 4 Whilst there are legislative protections in place regarding language and the right to a defence, under private law where parties are free to choose their language regime, either the language of one party or the other, or a third-party neutral language, may be selected and various issues of conflicting interpretation can arise. The various types of ambivalence described above may be exacerbated when translated, as exemplified by Salmi-Tolonen when discussing the language of contracts: "Words are not containers whose contents are transferred from one interlocutor to another […] unchanged" (2006: 86).
The potential risks of plain language For some years now, plain language movements, 5 the impetus for which came, in English-speaking jurisdictions, to a large extent from Mellinkoff (1963), have been campaigning to 'simplify' legal languageoften referred to in that context as legaleseand, inter alia, render it more accessible to the general public. However, some legal scholars such as Phillips hold that "on the contrary, the development and maintenance of the law's special language can be justified" (2003, preface). Offering a wide range of suggestions for the improvement of legal drafting, Pollman also stands in favour of an informed approach to legal jargon (2002), as does Crump (2002). A recent and ongoing study by Barnes (2016) provides a thorough inventory of scholarly and institutional research on plain language in legislative contexts.
A recurrent concern by those hostile to plain language approaches is the "under-specification of legal scope"i.e., the risk that plain language may in fact infringe rights, especially in common law jurisdictions where laws and legal documents are subject to interpretation by the courts, by handing power over to individual judges (Bhatia 2010: 9;Solan 2011). To address calls for simplification without compromising legal robustness, Bhatia has made an alternative proposal: the "easification" of legal language (e.g., 2010). This aims to make legal texts more accessible to their "intended readership", without compromising the "depth of specification" (Bhatia 2010: 10-13) that might occur if the text were simplified.
In any event, whatever their own view on plain language, the legal translator must not unilaterally render legal discourse plainer and thereby potentially distort the legal meaning and legal scope of the text: it is essential that they adhere to their brief. Hence a legal translator cannot permit themselves to be a gatekeeper of public access to justicein other words if the source text consists of dense, unreadable legalese then, subject to a sensitive negotiation of what is appropriate in that target language, it should stay that way in the target text, unless there has been a specific instruction to the contrary in the brief. This is an important matter to be addressed within legal translator training programs, and can often be a problem with inexperienced translators or those that accept legal work without being specialised in the field.

Contending with cross-jurisdictional asymmetries
Unfortunately for legal translators, the world does not have a unique legal system. Although similarities exist enabling groups to be composed, experts cannot agree on a single classification of the world's various systems into "families" (e.g., Hertel 2009; Monjean-Decaudin 2010b; Samuel 2014). To further complicate matters, a legal system "may, for example, be allocated to a different legal family as regards civil law than as regards administrative law. Even the law of companies may be characterised differently from the general civil law" (Hertel 2009: 2). Thus in addition to asymmetries between legal systems themselves, there may be inconsistencies between "different branches and fields of law" (Pommer 2008: 18). Comparative legal scholars may adopt a "macro" approachcomparing a whole system with another -but they may also adopt a "micro" or "meso" approach drilling down into individual legal concepts within each system and examining their similarities and differences. For translators, the differences between superficially and seemingly similar legal concepts add more layers of complexity to their task, and the latter conceptual comparisons may be of greater service to them when translating a specific termalbeit always seen in the context of the wider system in question.
Concerning the ways in which boundaries between legal systems are crossed, a distinction can be drawn between "vertical" and "horizontal" legal translation. Monjean-Decaudin (2010b), building upon Folena's work (1991), determines "vertical legal translation" as translation of legal texts in a language seen to be of higher status into a language deemed of lower status, such as European legislation to be transposed into the law of the Member States, while she describes "horizontal legal translation" as a communication channel opened between two legal systems and two languages of ostensibly equal status. Monjean-Decaudin gives three examples of contexts in which horizontal translation might occur: (a) where individuals or firms assert and wish to have recognized rights or legal status by a second State's authorities; (b) to acquire knowledge of or to disseminate a given country's laws; (c) legal translation carried out on behalf of the judiciary -either to facilitate dialogue among different States' authorities, or between judicial authorities and a citizen who does not understand the language of proceedings (2010b: 702-703). Lamalle, on the other hand, views horizontal translation as the transfer of legal concepts into another legal language or system (2014: 299), and, following Flusser (2002: 194), sees vertical translation as the transfer of a concept from one field of knowledge to another: e.g., "from theology to law" (Lamalle 2014: 299).
It is important to note that "legal terminology [and legal discourse] is system-bound, tied to the legal system rather than to language" and hence "multiple legal languages can exist within the boundaries of a natural language, depending on how many legal orders make use of that same language" (Pommer 2008: 18). For example one natural language, French, is used to vehicle different many legal systemsinter alia, those of Belgium, France, Luxembourg, Monaco, some African countries and Switzerland, not to mention its use as the working language of the case law of the Court of Justice of the European Communities. Equally, there are "plurilingual states and regions with different legal systems or a mixed legal system" "such as Canada, India, Sri Lanka, Israel, South Africa, and, more recently, China (Hong Kong and Macau)" (Šarčević 2012: 193). On the other hand, translations may in some cases, as noted by Galdia, also be carried out between closely related languages such as Danish and Norwegian, where the countries also have similar legal systems, and issues related to legal systems/concepts may even disappear completely in multilingual States with a single legal system such as Finland (2003: 2). Glanert and Legrand note that "the extent of the challenge is much wider than might readily be expected" and point out that one may even need to "translate Englishthus, "privacy" in the UK is not "privacy" in the US" (2013: 516).
Systemic and/or conceptual asymmetries are thus a major constraint on the legal translation process. In De Groot's view, "the level of difficulty of a legal translation does not primarily depend on linguistically determined differences, but rather on structural differences between legal systems" (translated from the German and cited by Galdia 2003: 2). Several translation scholars have called upon the field of comparative law to advise how legal translators should negotiate differences between systems and concepts (e.g. , Pommer 2008;Šarčević 1997;Monjean-Decaudin 2012, Sandrini 1999. Other scholars have also made efforts to categorise types of terminological equivalence across legal systems. Šarčević suggests the following three categories: "near equivalence", "partial equivalence", and "non-equivalence", and holds that the above enable categories to be changed "depending on the use of the term in context" (1997: 237). Nielsen, referring to a projected bilingual dictionary of contract law, goes into more detail: he lists full equivalence "where an L1 equivalent has exactly the same semantic and pragmatic properties as its L2 lemma", but adds that instances of such equivalence are "few and far between"; partial equivalence which he subdivides into three sets; and zero equivalence where there is no corresponding legal concept in the foreign system in questionhe adds, however, concerning this category that the lack of a corresponding concept does not mean that a "suggested translation equivalent" cannot be offered (1994: 162-169). 6 6 A few examples of highly system-specific concepts posing translation difficulties must suffice: the common law concept of "equity" which is held to have no equivalent in civil law (Department of Justice Canada, bijural terminology records, 2012); the "concept of 'faute', which is well known in French law, has no direct equivalent in other legal systems (in particular, English and German law)", Joint Practical Guide for the drafting of legislation within the Community institutions; usufruit (Kasirer 2001); "viager" (Février, Linnemer and Visser 2004); trusts, particularly with regard to Italian law (Lupoi 2013); or tort/delict (Schroth 1986: 57-58). An even more basic example from company law is the lack of correspondence between types of legal entity under foreign laws (Rogers-Glabush 2009: 514-523).
In establishing a set of postulates on legal translation, De Groot asserts: "In practical translation, an approximate equivalence of concepts is sufficient when deciding whether one concept may be used as translation for another" and: "Whether an approximate equivalence exists or not depends on the context and goal of the translation" ( : 229-230, translated by Engberg 2013. De Groot thus renders both the degree of concepts' legal equivalence and their use (or not) contingent on the intended purpose of the translated text.
A number of projects are currently being developed to create terminological references for the translation of legal concepts across different legal systems, not all of which, unfortunately, are accessible to practitioners who work outside institutions. Examples include the Textual and Terminological Database for the Portuguese Parliament [BDTT-AR]; TermWise for the Belgian Federal Justice Services (Heylen et al. 2014); Italian and German projects at the Institute for Specialised Communication and Multilingualism at the European Academy of Bolzano (Ralli 2009).
In most countries, there are no legislative guidelines providing for differences between legal systems or concepts. Some efforts are being made in this regard, for example as part of the harmonisation and approximation 7 of European legislation, and at global level by UNIDROIT (the International Institute for the Unification of Private Law).

Preserving the integrity of genres and subgenres
Having discussed in the above sections how language and legal systems need to be taken into account in legal translation performance, the third aspect of my model concerns genre. Genre is defined by Swales as a "class of communicative events [sharing a] set of communicative purposes", forming the "rationale for the genre" (1990: 58). That rationale is enforced by the parent discourse community, leading to constraints on "choice of content and style" (Swales 1990: 58).
Legal genres have been classified by scholars in numerous ways. Bhatia (1987: 227) outlines a structure differentiating the main legal genres by their "communicative purposes", which is highly pertinent in its implications for translation and the intended user of a target text. In a more recent work, Bhatia (2006: 6-7) distinguishes "primary" genres -legislation; "secondary" genrese.g., judgments and case reports; "enabling academic genres"such as textbooks, critical essays, etc.; and "target genres"e.g., contracts, affidavits, insurance documents etc. We may align these categories with the different environments in which translation is carried out: "primary genres" within institutions; "enabling academic genres" often by authors themselves or in close collaboration with their translator; and "target" and "secondary" genres frequently outsourced. This recalls Trosborg who divides legal text types according to "external factors pertaining to the situation of use " (1987: 20). A further classification refers to the sublanguages of legal professions, which may vary according to the country involvedexamples given include the "notarial profession" in Europe, "legal authors", "judges", and "counsel" (Mattila 2006: 4-5). This provides a further taxonomical facet in terms of the drafters of the source text and the intended user(s) or receiver(s) of the translated text. As an alternative, Mattila suggests that "legal language can be divided into subgenres on the basis of branches of law" (2006: 5)such as criminal law, property law, and tax law, while Monjean-Decaudin asserts that division of texts in such a way has proved to be "tedious and of little relevance" (2010a: 4, my translation). 8 However, it may well be a more accessible approach for the market.
It is worth pointing out that commissioning clients, whether from law firms or corporate entities, translation companies/agencies, and translation practitioners do not generally adopt the above scholarly classifications (see Biel 2011: 166). The need to embrace other text typologies has been recognised, for example, by Prieto Ramos who notes that "subdivisions are ultimately determined by the lens through which textual realities are observed" (2014: 263).
8 Several other categorisations of legal genres have been proposed by, inter alios, Cornu (1990) and Bocquet (2008), but given space restrictions and the aims of this paper I do not explore this point further here.
Genre provides translators with insights in a number of ways. The contribution to be made by an awareness of and reference to both source and target genres/subgenres when translating includes but is not limited to: appropriate structure and terms; participants and their relationship(s); and the context of a communication act (e.g., Montalt Ressurrecció et al. 2008). The following statement by Gotti on the relevance of genre for discourse analysts applies equally well to legal translators: "not only to get a better understanding of the linguistic characteristics of texts, but also of the macrostructure of these texts, which appears to be organised according to genre expectations and conventions", enabling them moreover to learn "how genres are constructed, interpreted, used and exploited in the achievement of specific goals in highly specialized contexts" (2012: 61). Furthermore, textual standardisation is "stronger in legal genres" (Gotti 2012: 60).
In the light of such standardisation, corpus analysis is a powerful technique to obtain the insights described above. These methods are truly useful in providing translators with multiperspective, fast and reliable access to a given legal genre pool (e.g., García Izquierdo and Borja Albi 2008;Biel 2010a). More and more genre-based studies employ corpus techniques to investigate legal, business and financial discourse, such as Pontrandolfo (2015)

Distinguishing and addressing purpose
The fourth and final aspect of the legal translator's textual agency that I shall review here is the task of taking into account the purpose of the source text and addressing the purpose of the target text. Concerning the suitability of this functionalist approach to legal translation, Garzone concludes that: 'the degree of equivalence to be achieved in the translation of a given text is not absolute, but depends first and foremost on the TT [target text] intended function as well as on the nature of the ST [source text]; the whole process is governed by a principle located at a sufficiently high level of generalisation as to be suitable for virtually all types of legal texts.' (2000: 9, emphasis added).
Despite her use of the word "virtually", Garzone (2000) does not specify which types of legal text might be unsuitable. It is interesting to note in passing that, in addition to a body of translation scholars, functionalist methods have also been supported by a leading member of the judiciary -Justice Pigeon of the Supreme Court of Canadawho has agitated strongly against literal translation with regard to court documents (Pigeon 1982).
As Munday explains: "knowing why a ST [source text] is to be translated and what the function of the TT [target text] will be are crucial for the translator" (2008: 79). Without lessening the importance of the function of the target text, the latter assertion includes both source and target texts, as does Garzone's claim above, and embraces circumstances whereby translators may either have to transpose the source text purpose, or produce a target text fulfilling a different purpose. The target text purpose is also relevant when assessing performance, as Nida claims when expressing his view on adequacy: The relative adequacy of different translations of the same text can only be determined in terms of the extent to which each translation successfully fulfils the purpose for which it was intended. (1976: 64). This is particularly salient in view of the current adoption of fitness-for-purpose as a quality benchmark, e.g., by the European Commission, and the inclusion of the specification by clients of text purpose in translation standards such as the Deutsches Institut für Normung (DIN) 2345 and the ASTM (American Society for Testing and Materials) standard F2575-06 (Scott 2016: 124-126).
The following selection of observations made by Pym reviews how functionalism can help with performance in practice: '[Functionalism] recognizes that the translator works in a professional situation, with complex obligations to people as well as to texts. It liberates the translator from theories that would try to formulate linguistic rules governing every decision. It forces us to see translation as involving many factors, rather than as work on just one text. ' (2010: 56) In sum, when negotiating the various aspects of legal translation performance, which are inextricably linked and dynamically connected, knowledge of the target text purpose is a crucial and indispensable prerequisite for success.

Differentiation of receivership and differentiation of status
The end-user of a target text is closely related to its intended purpose, and needs to be taken into consideration, as asserted by Reiss & Vermeer: "information about the target-text addressee […] is of crucial importance for the translator" (1984: 101, cited from the German by Nord 1997: 22). Despite her 1997 criticisms of Skopostheorie, and whilst maintaining them, Šarčević entitled her 2000 paper "Legal Translation and Translation Theory, A Receiveroriented Approach", and in that paper unambiguously claims: "[l]ike other areas of translation, the translation of legal texts is (or ought to be) receiver oriented" (2000: 1). She broaches the differentiation of readership and a corresponding variation in translation strategies, and uses the terms "addressees" and "receivers" synonymously. Citing Kelsen (1979, in the German), she distinguishes between direct addressees (specialists) or indirect addressees (including the public). Gémar has also subdivided readers, into four groups: laymen; those who are 'lettered'; practising legal professionals; and legal scholars. He argues that depending on its destination, a translation will be constrained and informed by knowledge of its intended readership (2002: 168). I use the term "end-user" because, given the length of the Chain of Supply in outsourced contexts for example (Scott 2016: 36-43), the immediate recipient of the translation may be an intermediary or a reviser, and not its useri.e., not the reader or receiver as understood by Gémar or Šarčević.
The differentiation of translators' services according to the target text's status has been put forward, e.g., by Chesterman and Wagner (2002), although the latter does not refer specifically to legal texts. Recently, the translation market has begun to see initiatives by translation agencies offering different "service levels" (Scott 2016: 146-148). I have proposed the following tripartite classification as a first step: translations for gist; translations that are not to be legally binding; translations where the target text will be legally binding (2016: 85-86). The triad is resolutely succinct, with the aim of making it accessible to those commissioning translations, and for the same reason worded as simply as possible. The proposed classification was tested in fieldwork relating to the outsourced market, and positively received (Scott 2016).
In order to further elucidate receivers' expectations, it is useful to adopt the distinction between covert and overt translations first highlighted by House, whereby "an overt translation is one which must overtly be a translation" (1977: 106) and "a covert translation […] enjoys or enjoyed the status of an original ST in the target culture" [… -a] ST and its covert TT have equivalent purposes" (1977: 107). I have offered a covert-overt cline for legal translation, and some proposals for its practical implementation (Scott 2016: 86-88;Scott forthcoming). For example covert legal translations may include contracts for signature, banking terms and conditions for publication, or calls for tender being issued. Overt legal translations, where literal approaches are favoured and where the 'transparency' of the target text will allow the source to 'shine' through a translation, pertain rather to comparative law exercises (Baaij 2014), access to foreign legislation, official translations of certificates or diplomas, and certain judicial genres depending on their destination (Monjean-Decaudin 2012).

Conclusion
In closing, I would like to underscore the reasoning behind the multidimensional model offered in this paper. Scholars are increasingly in agreement that legal translation theory ought to be induced from professional practice (e.g., Bocquet as early as 1994;Šarčević 2000;Biel and Engberg 2013;Prieto Ramos 2014). The model aims to make a contribution to describing effective performance in the field of legal translation, while taking into account Toury's recommendation to balance "the mental [and] the environmental, i.e., the situation in and for which the act [of translation] is performed" (2012: 67).
The paper has emphasised the pivotal nature of each of four fundamental aspects of the legal translator's textual agencywhere the word "pivotal" emphasizes their crucial importance in producing high quality legal translation, and highlights the non-static and nonlinear performance of the translator's sub-tasks. To reiterate these building blocks of difficulty:  transferring one legal language to another is a most intricate affair;  in negotiating solutions between different systems and cultures with their specific concepts the legal translator enters the highly sensitive and complex domain of comparative law;  genre and/or sub-genre compliance must also be respected, subject to rejection of the target text;  the purpose of the translated text -encompassing the text's end-user, its status and level of covertness/overtnessis a crucial aspect that must determine the outcome, and one that is often used as a quality benchmark for the work produced.
Most importantly, only when we, as legal translation theorists, trainers, and practitioners, fully appreciate the complexity of the cognitive juggling and negotiating involved, and then communicate it to others, will we be able to convey the importance of the legal translator's task, raise their status, enhance interaction with clients and other market actors, and, who knows, attract even brighter stars to the profession.