SUPPORTING LEGAL SCHOLARS’ EFFORTS TO COMMUNICATE IN ENGLISH AS A LINGUA FRANCA IN A GLOBALISED WORLD: REVISION OF LEGAL ACADEMIC WRITING

This article examines problems encountered in revising legal academic texts translated into English or written in English as a Lingua Franca (ELF) by legal scholars themselves. It starts by discussing globalisation and the rise of ELF in general and specifically in law and academia, considering the importance of publishing in English for individuals and academic journals. This is followed by a presentation of the Polish perspective on globalisation, as well as the challenges of revising academic legal writing. Specific examples are provided from the author’s professional experience in revising such texts. Two error-prone areas were selected: fixed phrases (including Latin) and legal terminology. Results from several corpora of English are used to justify corrections. Conclusions concern the considerable power and responsibility of revisers, as well as the need for greater awareness of the pitfalls of legal translation and better MT literacy among legal scholars.


Introduction
Globalisation can be narrowly defined as companies' efforts aimed at entering international markets, or more broadly defined as "a social trend that intensifies relations between societies and nations, a process by which decisions, events and activities from one part of the world have strong influence on other distant parts of the world" (Sandrini 2006: 110-111).In the broader sense it is linked with the "dismantling of cultural, disciplinary and national barriers, especially in the context of […] international trade", which also affects language, especially the domain-specific languages used by particular professional communities (Gotti 2007: 21).
Due to globalisation, the growing mobility of people and trade, and the IT revolution, the last seventy years saw "an exponential rise in the demand for linguistic mediation", with an insufficient supply of such services, resulting in non-professionals becoming involved in most kinds of language mediation usually carried out by professionals (Antonini 2021: 171).For translators and interpreters, globalisation meant a period of steady increase in demand for services, which ended with the global crisis of 2008-2009, when the market became increasingly price-driven.Growing volumes of texts for translation with tight deadlines fuelled the development of large language service providers and crowdsourcing platforms, causing the progressive technologisation of translation, wherein humans, mainly freelancers, are often reduced to 'filling the gaps' (Fry 2009).Such an industrial approach is reflected both in the terms used to refer to translation and translators that suggest low status: for example, market or industry rather than profession for the occupation, and vendors, suppliers or resources rather than professionals or practitioners for individuals (Scott 2017: 61, 66), and business practices, termed 'digital Taylorism' (Moorkens 2020) or 'Uberization' (Firat 2021).Translators began to be perceived as being easily replaceable, by both non-professionals and technology (Fry 2009: 16).
The following section discusses the status of English as the language of globalisation, including in academia.Section 3 presents the Polish experiences of globalisation, with references to the author's professional connections as a translator/reviser of academic texts.Section 4 expounds the challenges of revising legal scholarship, claiming that due to the inherent complexity of legal translation it is hardly possible to separate revision ("reading a translation in order to spot problematic passages, and then making or recommending any corrections or improvements" [Mossop 2020: xii]) from content or stylistic editing.To back this claim, practical examples of problems in revising English texts written by Polish legal scholars are provided in Section 5, where translations of problematic phrases and terms are suggested with references to corpora or dictionaries, illustrating the complex decision-making process required in legal translation (Way 2016(Way : 1022(Way -1023)).The conclusion is that revising legal texts entails considerable responsibility-and power-which non-translators may not realise.

The rise of English as a lingua franca
The 'all hands on deck' approach to language mediation did not mean equal growth in demand for all language pairs.Globalisation favoured English as the language of international communication in all fields, from pop music to academic conferences (Gotti 2007: 21).However, when English is used in the intercultural scenario, "different communication styles, different rhetorical patterns" and the fact that "rules and expectations may differ according to [speakers'] cultural conditioning" create the risk of misinterpretation or failed communication (Goddard 2009: 172-173).The picture is even less rosy if we consider that, according to Eurobarometer, 62% of Europeans cannot hold a conversation in English, while 59% only use their native language when communicating on the Internet (Kauliņš 2022).Rayar (1999: 157) warns that when communicating in a lingua franca, "we can make ourselves understood up to a point, but difficulties arise where there are conceptual disparities in the cultures and intellectual worlds of the encoder and the recipient".This is the case with different legal systems, where such disparities need bridging when communicating with people from other cultures.Goddard (2009: 171) points out that globalisation was accompanied by the growth of supranational bodies (UN, NATO, EU and WTO) which produced legal documentation in English, making it the "language of international law".English became the language of international commerce, takeovers, contracts, arbitration, and other cross-border transactions.Despite the European Union's policy of multilingualism, English became a lingua franca in communication between EU officials (Gotti 2007: 21) and the main drafting language, leading to "de jure multilingualism and de facto monolingualism in English" (Leal 2022: 203, 205).
Legal languages are unlikely candidates for globalisation, as they are strictly linked to national legal systems and to the corresponding legal cultures, but they are influenced by intercultural factors, too.Legal texts used locally increasingly implement translated international documents (Gotti 2007: 22).This is connected to the diminishing role of national legal systems and the growing role of transnational ones.Counterintuitively, this does not make legal translation easier as the complexity of the scheme of reference (local, regional, global) also increases (Sandrini 2006: 117, 118).Importantly, "there is no such thing as an international legal language" and legal terminology is multinational rather than multilingual, because the same language can be used in multiple legal systems, e.g.English in the UK, the US, Australia, etc., or German in Austria, Germany and Switzerland (Sandrini 2018: 513, 518).In legal translation, "non-equivalence is the rule rather than the exception" (Kjaer 2007) and there is a particular risk of misunderstanding when lawyers from civil law countries communicate in English, since common law terms do not match the terms from any civil law system (Bogdan 1994: 40, in Goddard 2009: 177).
It should be remembered that national laws emerged relatively recently: the German, French and Austrian civil codes were all adopted in the early 19th century.This clear focus on national law, expressed in the national (or dominant) language, and the development of modern states based on territoriality and sovereignty, happened after a long period of jus commune, based on Roman law and expressed in Latin (Sandrini 2006: 114-115).Seen from this perspective, globalisation is the reverse process: a trend towards a world where capital, labour, goods and ideas move across borders and where global institutions, international treaties or European integration put pressure on national legal systems (Sandrini 2006: 116).The influence of globalisation is particularly noticeable in contract drafting, heavily influenced by the common law style and common law concepts (Vettese 2011), though with different legal force of standard common law contractual provisions in civil law jurisdictions (Cordero-Moss 2011: 353-370); international arbitration proceedings, once United Nations Commission on International Trade Law Model Law on International Commercial Arbitration has been adopted in dozens of countries (Cordero-Moss 2013: 7), as well as competition law, which deals with international competition and cross-border threats to competition (Biel & Sosoni 2019: 212).
Every discipline has "a particular way of communicating and generating specific text types", so translators who work in specific disciplines need passive and active knowledge of how experts communicate.The cultures of many disciplines have become global rather than national: research methods and ethics are generally accepted and research articles are written in a similar way worldwide (Sandrini 2006: 112).Nowadays, when academic writing also occurs in a global market, scholars must decide whether to publish research in their native language or in English, often seen as a synonym for international.After WWII, English became the global language of academic publications, due to the leading position of the US in research investment (Lillis & Curry 2010: 1, 9, 156).It is the main language of articles published in peer-reviewed journals, in both "hard" and "soft" sciences, the most frequently studied foreign language, and the language of instruction at universities (Laurén & Takala 2018: 260, 262).The success of international communication in English as a Lingua Franca (ELF) is stressed (House 2013;Kecskes & Kirner-Ludwig 2019), and there is growing acceptance for non-native writing (Laurén & Takala 2018: 256) and, even more so, for non-native oral communication (Hülmbauer 2009).Yet conference interpreters struggle with non-native English and capture their frustration calling it "bad simple English", "globish" or "desesperanto" (Albl-Mikasa 2017).
This suggests that there is a darker side to the use of ELF.Especially in humanities and social sciences, where argumentation and language matter greatly, it has been observed that familiarity with Anglophone conventions is a benefit and authors who do not conform to Anglophone rhetorical patterns are less likely to succeed (Albl-Mikasa & Ehrensberger-Dow 2019: 48;Horner 2017: 414).Non-native English speakers are said to be discriminated against in international journals, which can potentially "impoverish the creation of knowledge", or are forced to use the costly and time-consuming services of academic 'literacy brokers' (O' Brien et al. 2018: 239).The influence of literacy brokers often goes beyond correcting language errors: they suggest reformulating the text, changing subheadings, removing whole sections, and discussing research results in a certain way (Lillis & Curry 2010: 100-105).Even in a translation studies journal issue devoted to ELF, editors experienced tensions between efforts to achieve "the standard required by a top-ranking international journal" and "individual authors' cultural identity or language ideology", sometimes manifested as refusal to implement stylistic changes (Bennett & Queiroz de Barros 2017: 366).In recent decades, academic English has become more colloquial, perhaps under the influence of plain English initiatives (58)(59)(60)(61), while in other languages long sentences and an indirect way of presenting information are often still considered the norm (O 'Brien et al. 2018: 255).
When chances of career advancement or grant awards depend on publishing in English (Lillis & Curry 2010: 48), scholars no longer think about whether to do it, but rather how to do it in the most economical way.It has been shown that post-editing machine translated texts can lead to the creation of acceptable academic writing in English for dissemination by researchers who are not native speakers of English (O' Brien et al. 2018).Researchers are also known to use MT for research dissemination purposes, therefore calls for developing their MT literacy have emerged.

Globalisation from the Polish perspective
Once part of the European jus commune family of legal systems, Poland lost its statehood in the late 18th century and its different parts were governed by Austrian, Prussian and Russian laws.Until independence was regained in 1918, the Polish legal language was not developing as fast as other national legal languages and the partitioning powers' legal languages left a legacy of numerous "lexical and syntactic calques" (Biel 2014: 333).After WWII, Poland ended up behind the Iron Curtain and it was not until the democratic transformation of 1989 that it could join the mainstream of globalisation.
I personally benefitted from this surge in demand for translation: before graduation I was already involved in translating Polish legislation into English.As a fresh graduate, I was flattered to receive requests from constitutional law scholars to have their conference papers and research articles about the new Polish Constitution, adopted in 1997, translated into English.Soon, discussions about a potential constitution for Europe followed and translation was needed in both directions, as contributions of foreign scholars were also published in Poland.This was my first experience of translating from ELF and of how this potential tool of mutual understanding could "also act as a medium and subject of global misunderstanding" (Spichtinger 2020: 16, emphasis in original).I did indeed find the phenomena described in research on translating and interpreting ELF, particularly the additional cognitive effort required when translating input that deviates from standard English (Albl-Mikasa & Ehrensberger-Dow 2019).
Another important event on the timeline of Poland's globalisation was the efforts aimed at joining the EU, a prerequisite for which was the translation of an impressive body of EU legislation (acquis communautaire) into Polish and the harmonisation of Polish law with this legislation.Before accession, these translations were often criticised for "low quality, translation errors and obscure terminology" (Biel 2014: 333).An example may be the differences in the use of modal verbs and other constructions expressing deontic modality, a fundamental aspect of legislation, which is concerned with regulating people's behaviour by introducing obligations, prohibitions or permissions (Engberg & Rasmussen 2003;Kielar & Miler 1993;Matulewska 2009).Compared to non-translated Polish law, the translations of EU directives and regulations usually contained high frequencies of modal verbs, even though in Polish statutes obligation is often expressed by verbs in the present indicative or by the expression jest obowiązany [is obliged].Normalised frequencies of individual modal verbs also differed between translations of acquis communautaire and non-translated Polish legislation, as evidenced by the verb musi [must], which was 14 times as popular in translated regulations and 41 times as popular in translated directives compared to non-translated instruments.Apart from the potentially foreignising effect, such departure from conventions could also have made these translations more difficult to process, as readers' attention could have shifted from the content to the unusual linguistic patterns (Biel 2014: 342-346, 349-350).
In the early 2000s, I started translating tables of contents into English for one Polish legal journal.Since then, the number of journals that publish titles and abstracts, or even whole articles, in English has been growing steadily.This results from the growing focus on publications in academia, but also from the mounting pressure to publish in English (Kulczycki 2023: 2-6), even though social sciences and humanities-unlike sciences-are considered more "national" in nature because they respond to local claims and need legitimisation from society.Polish journals sought internationalisation by publishing papers in English or by foreign authors (output), or by having foreign board members and referees (management).These actions represented attempts at meeting the requirements of science policy rather than actual successes in internationalisation (Kulczycki et al. 2019: 10, 15-17, 29).I was told unofficially that articles in English attracted fewer clicks from readers than articles in Polish.Yet officially, researchers believe that publishing in English will make their research "more readily indexed, accessed, read, used, and cited", contributing to the creation of a "publish in English or perish" culture (Bowker & Buitrago Ciro 2019: 80-81).
Initially, legal academics only submitted their texts in Polish and the publishing house requested translations.Over time, more and more authors prepared abstracts or whole articles in English for revision and since around 2012 there has been a steady stream of machine-translated texts, sometimes even without the Polish version.The use of machine translation (MT) is never disclosed and authors are unapologetic if confronted about it, perhaps considering the need to prepare an English version of an abstract or article yet another hurdle to publication.Moreover, research grants cannot be expended on translation, but only on revision.Lawyers may not appreciate the complexity of legal translation.This may be confirmed by the fact that commercial law practitioners are the least likely to provide poorly translated texts or use MT, perhaps because they provide services to foreign clients and realise the risk of miscommunication.Elsewhere, people other than language professionals, including lawyers, admit using MT at police stations, during lawsuits and in other legal settings (Vieira et al. 2021(Vieira et al. , 2022)).In a survey about MT use by translators, a number of respondents suspected 'stealth post-editing', i.e., assignments to revise human translations that were in fact MT post-editing (Farrell 2023).
One common feature of the pre-accession translations of EU law and legal academics communicating internationally in English is the mass production of such texts, whether by humans or algorithms, which may lead to low quality.Importantly, MT is unable to transform texts to meet the target language genre requirements or rhetorical expectations (O' Brien et al. 2018).Just as the acquis communautaire was translated hastily to meet tight deadlines, academics often leave the production of the English abstract as the last task before publication.It is not unusual for me to receive a batch of several articles to revise with unrealistic expectations about turnaround time.That time largely depends on source text quality, which often causes problems in translation (especially MT) or, indeed, revision, therefore it merits a closer look, both in terms of legal academic writing conventions and reader orientation.The difficulties result from the fact that authors write in ELF, from differences between legal systems, and from the writing conventions in the legal cultures involved.

Revising academic legal writing in ELF
Translating and interpreting ELF input was found to be challenging for language professionals, especially interpreters, due to time constraints, while ELF speeches were often less understandable for the audience than professional interpretation into their native languages.The main difficulties with the comprehension of ELF compared to standard English input include non-standard syntax, lexis and other cross-linguistic transfer from L1 (which may be partly alleviated by the shared language benefit), as well as problems with textual cohesion (Albl-Mikasa & Ehrensberger-Dow 2019: 50, 53).Even for translators, non-standard English source texts were found to require longer processing time and the resulting target text solutions were twice as likely to be inadequate compared to the translations of texts edited to conform to the norms of English, while having the same L1 did not always help with problems such as cognates or unclear lexical expressions (Albl-Mikasa et al. 2017: 380-381).There have been no studies on ELF revision, but the ideas expressed by language mediators in surveys resonate with my experience of revising academic legal writing.
As Rayar (1999: 161) points out, an important aspect of the translation or revision of legal texts is the orientation towards target readers.If a text written for domestic readers is translated for foreign readers, it is important to consider the presupposed knowledge, since "system-specific information relating to the author's domestic legal system may very well be new information to foreign readers".It is necessary to have background knowledge of underlying ideas and intertextual or interdiscursive links in order to understand the message of legal texts.The requirement that addressees have "implicit professional information" exists in any professional sublanguage, including the legal one (Azuelos-Atias 2018: 106).When the author intends the text to be read abroad, but writes in his/her native language, the translator needs to check if enough explanation has been provided.However, Rayar (1999: 162-164) observes that the most challenging scenario is revising a legal text written for foreign readers by a non-native speaker.It may involve fixing more than linguistic problems, though even problems resulting from the author's insufficient command of the foreign language can only be fixed if the intended meaning can be decoded.Comprehensibility may be difficult to achieve due to conceptual differences and system-bound terms, while implied information may need to be made explicit.Finally, translated texts may also seem "foreign due to expressed attitudes: spending 'taxpayers' money'" on helping the less fortunate may need to be advocated for in the UK, while it is mainstream in France or Germany (Goddard 2009: 184).
As for terminology, Rayar uses revision to mean both improving the quality of translations and of texts written in ELF.However, Mossop (2020: 115) defines revision as a "function of professional translators in which they find features of a draft translation that fall short of what is acceptable", which includes fixing "omissions, major mistranslations, gross translationese, nonsensical passages, terminology errors, and departures from the rules of the standard language", sometimes also improving the text by stylistic editing and "minor adjustments in meaning to better reflect the [source text]".Another activity he mentions is editing non-native English, which seems close to Rayar's third revision scenario.Mossop (2020: 23) observes that even good speakers of English may have poor writing skills, which results in the need to correct micro-errors, such as wrong lexical choices, the use of false friends, mixing formal and informal language or using excessively informal style, or "failures in English composition", that is, poor text organisation.In the remaining part, like Rayar, I use revision as a broad, umbrella term, but try to show that it comprises editing as understood by Mossop, though the boundaries between the two are blurred.
As law is a language-based practice, the language used by legal scholars merits attention.The obvious problem in translation is terminology, particularly terms without equivalents (system-bound) or partially equivalent ones (Šarčević 1997).Even though there are bilingual legal dictionaries, their utility is limited (Sandrini 2018;Šarčević 1990;Szemińska 2014).Unlike terminologists, who look for solutions applicable in all cases (one concept-one term principle), legal translators need pragmatic equivalents for a given translation context, so comparative law often provides more help than terminology (Engberg 2013: 14-18;Sandrini 2018: 516).In addition, there are different linguistic conventions in civil law and common law countries, neatly summarised by Cao (2007: 28-29): the former use 'concise' language, with principles formulated in general terms, while the latter use 'precise' language, listing all applicable cases and exceptions.When legislation is co-drafted in Canada, the length of equivalent provisions in French and English can vary considerably (Gémar 2013: 163).
Some problems may result from the different degrees of development of plain language, with plain Polish still being a relatively new style of writing compared to plain English (Setkowicz-Ryszka 2022a).Yet, against the rules of plain language, English legal writing is still often syntactically complex, with "intricate patterns of coordination and subordination" and "syntactic discontinuities", a strong tendency to use deverbal nouns and passive constructions or verbs in third person to make the text impersonal (Williams 2004: 113-115).Importantly, academic legal texts are intended to be read by other legal academics or practitioners, therefore they need not be written in plain language unless the author wishes to make them more accessible.Frequent problems that make Polish legal texts, including academic ones, difficult to understand concern: (1) complex syntax, especially syntactic homonymy or secondary syntactic connections, both of which lead to ambiguity, (2) poor punctuation, and (3) excessive use of words of foreign origin (Kurek 2015: 305-309).

Revising academic legal texts of Polish scholars in English: examples of problems
This section examines some problems encountered while revising academic texts written in ELF.The Polish term used by the legal publishing house is korekta, i.e., revision, which suggests that these texts are seen as translations.It should be stressed that it is often impossible to tell whether a text was translated by a human using a dictionary or machine-translated and subsequently post-edited.The worst texts I have revised in recent years must have been human translations 1 .There have also been texts whose original Polish versions I could recover from DeepL translation engine 2 .In 1.After the initial frustration with the poor quality of MT output in the early 2010s, I came to appreciate the predictable errors in neural MT, especially compared to poor human translations.2. Meaning that authors used free DeepL engine and the Polish texts were retained by the provider.
what follows, I will not distinguish between ELF texts, non-professional translations, and machine-translated texts, assuming that regardless of how their authors produced the English versions, they were satisfied with their quality and took full responsibility for their contents.The areas discussed in the following part include fixed phrases of academic legal language and terminology, both of which have received considerable interest in legal translation studies.The examples should demonstrate that during linguistic revision one also deals with cases where revision involves some form of editing.

Reference corpora
The selection of lexical items discussed in Section 5 are items that I frequently encounter in practice.To confirm their prevalence, I will use the frequencies of these items from the National Corpus of Polish (NKJP), a balanced corpus of general Polish language, queried using PELCRA search engine (Pęzik 2012).Subcorpora (a) and (b) were analysed using corpus tools available on the English Corpora website.Since they are both corpora of general language, a decision was made to use the same corpus tools for the relevant COCA and BNC subcorpora for greater comparability, even though BNC is also available on Sketch Engine.The two specialist (sub)corpora (c) and (d) were chosen due to their focus on legal and academic language, respectively, especially the separate law and law and politics sections.Finally, a corpus containing what is presumably translated Englishthe Directory of Open Academic Journals (DOAJ), also available on Sketch Engine-is used for comparison with data for Polish and for native English obtained using the remaining corpora.The corpus consists of 3.35 bn tokens in total, while the law and politics subcorpus (DOAJ Law&Pol) comprises 16,134,544 tokens.(Sub)corpora (c), (d) and DOAJ were analysed using Sketch Engine tools.
It should be stressed that any data concerning the phrases and terms of interest, especially their relative frequencies, should be taken with a pinch of salt.Firstly, it is impossible to ensure full comparability of the corpora due to the differences in their composition.Secondly, it is impossible to tell whether my intuitions about corpora containing "native English" or "translated English" are accurate, since the DOAJ corpus may contain articles written by native speakers of English, while OCAE may include articles which have been translated or written in ELF.What is presented below is more akin to the findings of extensive research conducted for a revision assignment than typical academic research.Results from Google searches are not included, regardless of their practical utility (Biel 2008;Giampieri 2018).Nevertheless, it is hoped that even such rough comparisons can shed some light on the complex decision-making process in legal translation (Way 2016) and the need to 'cognitively juggle' legal languages from various cultures (Scott 2017a).

Phraseology of legal scholarship
Legal language abounds in fixed phrases.The presence of formulaic expressions and standard phrases, such as binominals, in legal language may be rooted in the oral tradition of law, but also in the dialogue between legal experts and judges that takes place in academic legal writing of civil law countries and influences legal phraseology (Kjaer 2007: 510).As for English, certain expressions tend to be perpetuated in legal language because their meaning is well-established and change might result in ambiguity or even litigation (Williams 2004: 118).In terms of degrees of stability, Kjaer (2007: 512, 513) distinguishes (1) word combinations directly or (2) indirectly prescribed by law, failure to use which may invalidate a text or affect its legal force, (3) word combinations based on implicit quotations from other texts, and (4) other routine, habitually used phrases.The last two types of word combinations are important for concept formation, for ensuring continuity in law, and for speeding up text production.They are important in academic writing as their use may signal membership of a community of lawyers, as opposed to non-lawyers (Goldstein & Lieberman 2016: 20).

Latin maxims
Fixed phrases include Latin maxims used in only one legal system or in multiple ones.Some of them survived the fall of the Roman empire, others were coined locally and more recently.Lawyers from different countries have different repertoires of Latin maxims.Popular ones used in Polish, such as lex retro non agit [law is not retroactive] and in dubio pro reo/tributario [(when) in doubt, it should be resolved in favour of the defendant/ taxpayer], are not even included in Black's Law Dictionary (BLD).Another maxim, nullum crimen (nulla poena) sine lege (poenali) [no punishment without a law authorising it], appears only in the form nulla poena sine lege in BLD.The dictionary also explains that ne bis in idem [not twice for the same thing] is the civil-law equivalent of the rule against double jeopardy (Gałuskina & Sycz 2013: 15-17).
The absolute and-where applicable-relative frequencies per million tokens (RF) of the above maxims across (sub)corpora are presented in In NKJP, maxims 1, 2b, and 3ab are also found in subcorpora of dailies, weeklies and monthlies, though less frequently than in Press_other.In COCA, maxims 1 and 3ab do not occur at all, maxim 2b has 6 occurrences, yet just one in an academic text, while maxim 4 appears only once and not in an academic text.None of these maxims are found in BNC or BLaRC.
In OCAE, there are no instances of maxim 1, single instances of in dubio pro reo and in dubio pro fisco (curiously, both in the sociology section), two instances of maxim 3-even when clipped to just nulla poena, as in English Latin maxims are often clipped in English (Setkowicz-Ryszka 2022b: 223, 226)-and one instance of maxim 4 (see Figure 1).A look into the DOAJ Law&Pol subcorpus reveals that although Englishmedium texts published in German, Polish, Romanian, etc., journals contain no trace of maxims 1 and 2a, there is a single instance of 2b (although there are 21 instances of this maxim in the full corpus), 5 instances of maxim 3a, including 3 instances of the full version, 3b (although there are 17 and 13, respectively, in the whole corpus, see Figure 2).Meanwhile, maxim 4 appears 16 times in legal texts, with 13 instances from a journal published in Poland (see Figure 3), but 206 times in the entire DOAJ (RF = 0.06).This begs the question about the intended readers: the above frequencies suggest that other civil-law lawyers might understand Latin maxims, but few native speakers of English will.Yet would they read an English abstract (or full article) about Polish law in the first place?So, when one "translates" Latin maxims to match the conventions of legal English, is it not a disservice to the actual readers?

Polish-Latin fixed phrases
In legal academic texts in Polish, there are also mixed Polish-Latin phrases, such as: uwagi/wnioski de lege lata [remarks/conclusions about the law as it stands], postulaty/wnioski/uwagi de lege ferenda [suggestions/conclusions/ remarks as to what the law should be]-both of which appear even in the titles of articles-or wyrazić expressis verbis [to state expressly], and warunek sine qua non [a sine qua non].
As for the first two phrases, in the entire NKJP one finds: -one occurrence of de lege lata with 'remarks' and three with 'conclusions' (77 occurrences of the Latin element alone, RF=0.051); -28 occurrences of de lege ferenda with 'suggestions', 58 with 'conclusions', and three with 'remarks' (212 occurrences of just the Latin element, RF=0.139).
Neither of the above Latin expressions is explained in BLD or found in BNC, while their nominative forms can be found in COCA Law&Pol and BLaRC, but extremely rarely.There are also some instances in OCAE Law&Pol (see Figure 4).There are many more instances of these four expressions in the DOAJ Law&Pol subcorpus and in the full DOAJ, although due to the sizes of these corpora, all relative frequencies are low.The frequencies in Englishlanguage (sub)corpora are presented in Table 2.These Latin expressions seem far more popular in non-English-speaking countries; thus they find their way into English-medium journal articles written by non-native speakers.Although my usual solution is to translate the above Latin expressions into English, again this might be unnecessary if most readers come from civil-law countries.
In the NKJP, the phrase wyrazić expressis verbis appears thrice.However, expressis verbis alone is not found in any corpus of native English, compared to a total of 904 instances in NKJP Full (RF=0.593).Most of them -536 (RF=4.962)-appear in the Press_other subcorpus, but there are also instances in daily, weekly and monthly press.In DOAJ, there are 39 occurrences (RF=0.01),which might suggest a transfer from the authors' native languages (see Figure 5).Consequently, I always replace it with 'expressly' when revising English texts.In the case of sine qua non, as shown in Table 3, in native English it is rarely accompanied by condition or conditio, but in DOAJ this happens slightly more often.Revisers of English-medium texts are advised to reduce this expression to just the Latin part.In Polish, the expression is popular in non-academic texts, therefore for this comparison full corpora were used.Finally, when the Latin expressions are used in both Polish and English, they are usually inflected to match the Polish grammar (de lege lata/ferenda) but in the nominative in English (lex lata/ferenda), a non-inflected language (Gałuskina & Sycz 2013: 20).Therefore, the Latin part of the Polish title: Jurysdykcja krajowa w sprawach wynikających z recepti arbitrii na gruncie rozporządzenia brukselskiego I, translated as: Jurisdiction in disputes arising out of receptum arbitrii under the Brussels I bis Regulation (the basic form used in Polish), had to be corrected to receptum arbitrum (the basic form used in English).Interestingly, no native English corpus features this expression in either form, while DOAJ has two occurrences of receptum arbitrii (both from a South African journal).

Polish fixed phrases
Finally, there is a group of purely Polish fixed phrases, among which w doktrynie i orzecznictwie [in legal scholarship and case law] and zgodnie z powszechnie/ogólnie obowiązującymi przepisami [pursuant to generally applicable laws and regulations, as opposed to local ones] stand out as particularly frequent.NKJP Full contains 30 instances (RF=0.02) of the first phrase and 7 instances (RF=0.005) of either variant of the second one.
Legal scholars often use the former expression when they refer to views of other lawyers, in the output of legal scholars and commentators or judicial decisions.In their English versions, they often render it as in doctrine and jurispridence.Doktryna corresponds to an already obsolete meaning of doctrine as a "body of principles or tenets, a system of beliefs; a theory; a branch of knowledge" (Brown 1993: 719), while orzecznictwo can be translated as case law or jurisprudence in the sense of "the decision of a court" (Brown 1993(Brown : 1465)).Therefore, in this expression, it is mainly the first part that needs changing from doctrine to legal scholarship or similar expressions.However, considering that jurisprudence can also mean "science which treats of human laws" or "a body of law" (Brown 1993(Brown : 1465)), I tend to also change the second part to case law for clarity.Meanwhile, legal scholars seem to prefer English words of Latin origin.Of all the Englishlanguage corpora used here, only DOAJ has multiple instances of doctrine and jurisprudence (152 in the full corpus, 2 in the Law&Pol subcorpus) or doctrine and case law (12 and 1, respectively).COCA has one instance of either expression, while BNC, BLaRC, and OCAE have none.Thus, it might be used as a consequence of transfer from the native language, rather than a phrase known in legal English.
The main difficulty with the second expression is that it is sometimes rendered as pursuant to common law, since common is a possible equivalent of powszechny.This rendition is misleading because it refers to common law, which is an alternative to civil law (a legal system) rather than local laws and regulations.As for the expression generally applicable law(s), without pursuant or according to, it appears in COCA (23 times) and DOAJ (3 times), but not in the BNC, BLaRC or OCAE corpora.

Terminology
When revising Polish legal scholars' writing, much effort is geared towards making the texts understandable for foreign recipients.As mentioned in Section 2, an obvious problem is concepts without (full) equivalents in other languages.Considering the number of countries where English is the (an) official language and its status as the language of international law instruments, the number of possible equivalents of a single Polish term can be considerable (Matulewska 2017: 129-133), making polysemy another source of problems.When one is asked for a linguistic revision, few clients realise how often it involves terminological changes (micro-level content editing) and how much information may need to be elicited to resolve linguistic issues.

Terminological problems
Problems with legal terminology start with general clauses.Common law uses the standard of reasonable man, whereas civil law makes reference to dobre obyczaje [good customs].These are the accepted and laudable standards of behaviour in various social groups (Kopaczyńska-Pieczniak 2016;Biskup 2007).But this all-encompassing character is precisely the problem.One scholar translated it as good morals and good manners in a single text.However, it is much more than just manners, especially as it applies to business entities, too.In this context, business integrity might be appropriate, though it would exclude individuals.Good mores, good customs, good practices, and honesty are all only partial equivalents.The descriptive term recognised ethical standards might be more appropriate, although it cannot be guaranteed to fit all contexts, especially as the meaning of this term is contextualised by each scholar.
Sometimes the whole argument hinges on the choice of words.In a recent judgment, the Polish Constitutional Tribunal (the official name of the constitutional court) pressed the view that the Polish Constitutionwhich distinguishes between courts and tribunals-precludes the European Court of Human Rights from considering it as a court determining human rights or a court at all.The Tribunal claimed it was outside the scope of application of the European Convention on Human Rights, referring to the Polish version of Article 6(1), which mentions niezawisły i bezstronny sąd ustanowiony ustawą.The word sąd literally means court while the Polish name of this judicial body translates literally as tribunal.However, the same provision in English reads: an independent and impartial tribunal established by law, so the argument was lost in English.It was necessary to explain the confusion between the two language versions.Insufficient command of English may have been the reason why one scholar wrote about penalties for untimely delivery of goods.The Polish word nieterminowy is rather formal and means late, often in the context of discharging obligations.Meanwhile terminowy can be translated as timely.A query about the collocates of untimely in BNC confirmed that it usually co-occurred with death, end, and demise (top three collocates, see Figure 6) and a COCA query added passing to the list (see Figure 7).The final version was penalty for late delivery.In my collaboration with a commercial law journal, it has become standard practice that whenever the text mentions spółka, the Polish word covering both company and partnership, I ask for more specific information.In Polish one can distinguish between the two types of entities by adding adjectives: spółka kapitałowa means company, while spółka osobowa corresponds to partnership.The statute regulating both kinds of entities is called Kodeks spółek handlowych in Polish, which, considering its contents, requires translation as the Code of Commercial Partnerships and Companies rather than the frequently used Commercial Companies Code.Therefore, when texts mention spółki without an adjective, it is necessary to establish if they refer to both companies and partnerships or just one type of entity.But when the Polish version mentions spółki posiadające osobowość prawną [companies/partnerships which are legal persons], the whole expression can be reduced to companies, since partnerships are not legal persons.In texts written by Poles, it is fairly common to find the incorrect rendition *capital companies instead of simply companies.
The sentence: The principle of freedom of choice and the principle of fair elections should be then clearly established was offered as a translation of the Polish sentence: Należy również jasno określić zasadę wolności wyborów i zasadę uczciwych wyborów.Here, freedom of choice is a grammatically correct expression, so on a strict understanding of language revision and considering that the author is a subject-matter expert it may have been left unchanged.Yet, the proximity of elections suggested a problem with a polysemous word: the Polish word wybory can mean choices, but also an election.Therefore, the revised version mentioned the principle of free and fair elections.
More humorously, someone wrote about mystery of proceedings (literal translation of tajemnica postępowania) instead of legal privilege or about public comrades of land registers, where the surprising mention of comrades must have resulted from consulting a dictionary entry for wiara [faith, trust], which mentions an outdated meaning of wiara as company, friends, comrades.However, sometimes texts created with the help of a dictionary are submitted without the Polish version and despite the shared language benefit it is hard to decide what the author intended to say.This forces the reviser to guess, or to suggest possible intelligible versions, making them a ghost co-author.Unfortunately, in these cases, as Mossop (2020: 26) warns, accidental changes of meaning are likely to happen when the reviser is focused on fixing language issues.

Other problems with terminological choices
An issue that is on the margins of terminology, but that matters from the point of view of academic language conventions, can be found in the phrase legislative intervention by the legislator offered as the equivalent of interwencji legislacyjnej ustawodawcy.Here, the figure of the legislator (ustawodawca meaning legislature, parliament) merits attention: Polish legal scholars often speak about what the legislator changed rather than what the law now provides (Setkowicz-Ryszka 2023).In native English writing, one can also find references to the whole parliament as the legislator or lawmaker, but these terms are used to refer to individual members of such bodies, too.Meanwhile, Polish legal scholars often speak about the legislator's actions, intentions, successes, or failures.The above phrase could simply be reduced to legislative intervention.
Sometimes grammatically correct sentences can require intervention, even though-considering that the publishing house only requests linguistic corrections-they may be left unchanged.In translation from a language without definite/indefinite articles, one is suspicious of expressions such as: the addressee of law should also be able to understand himself as the author of law.The problem here is whether the scholar means the author of law (the only author) or an author of law (one of the authors, in addition to the legislator).After consultation with the author (of the article), the indefinite article was used.
A seemingly correct expression raw materials (products) of mining origin turned out to have negative prosody.Nothing was found in the corpora, but a Google search: "of mining origin" site:uk yielded such examples as: pollution of mining origin, landscape changes of mining origin, a void of mining origin, debris of waste rock and post-washing material of mining origin, and seismic shock of mining origin.Meanwhile "products of mining origin" had only five hits, all on websites from non-English-speaking country domains.The edited version mentions products obtained through mining.

Conclusions
As shown by the above examples, considerable responsibility for the final wording and meaning of articles in legal journals is placed on language professionals with a low status compared to lawyers (Scott 2017).Considering the complexity of legal translation (Way 2016: 1010-1015), constant vigilance and life-long learning are a must for the task of revision.Apart from the linguistic-plus-content fixes discussed above, it is often necessary to ask authors for disambiguation when guessing might distort the meaning.Moreover, such questions often seem to win the reviser the authors' respect.Due to the intertextual nature of legal writing, titles and quotations need to be checked, and quotations from judgments or legislation may even have to be detected, because they are often unmarked.Once, a historical consultation was needed to determine if a legal scholar who lived more than a century ago published his works in Lemberg, Lwów or Lviv-all of which refer to the same city at different times.Quotations (back)translated by authors need to be detected in original publications or in existing English translations.Sometimes the corrections can save the authors' reputation or at least eliminate unintended humorous effects.Part of the job of the reviser includes some plain language editing, as plain English is more advanced than plain Polish, including in the field of law and academia.There are many instances when revision cannot be separated from editing.
Meanwhile for many legal scholars, especially if they only need an English abstract and title, this last hurdle before publication requires nothing more than a generic MT engine.Interestingly, the more an author is likely to have had contact with foreign clients (commercial law and tax law experts, as opposed to judges), the less they tend to use MT or translate themselves with a dictionary, in a word-for-word manner.Perhaps they are more aware of the risk of miscommunication.I am uniquely positioned to observe the trends as I cooperate with several journals and revise large numbers of abstracts on a regular basis.However, since law tends to be a local discipline, domain experts may be unaware of the pitfalls of legal translation.All this confirms that MT literacy is indeed much needed among lawyers, particularly among those who seem to believe in the overstated claims about MT capabilities (Vieira 2020) or fail to realise the ethical dimension of asking for "revision" of machine-translated texts.Much has been said about the ethics of translators (Chesterman 2001), and there are a growing number of discussions on various aspects of MT ethics (e.g.Moniz & Parra Escartín 2023), meanwhile the ethics of using the services of translators or revisers are yet to be discussed.Moreover, the contribution of language experts to the internationalisation of legal scholarship needs proper recognition.It is my hope that this article helps appreciate this.

Figure 6 .
Figure 6.Noun collocates of the adjective "untimely" in BNC (source: English Corpora, access 12/05/2023) The corpus has 1.5 bn tokens and includes texts collected in 2008-2010.It is not possible to narrow down the search to just academic texts, but data can be broken down by channels.The channel that is of potential interest-Press_other-has 108,024,296 tokens.

Table 1 .
Frequencies of Latin maxims across corpora of Polish and English (source: own compilation)

Table 2 .
Absolute and relative frequencies of Latin expressions in five (sub)corpora of English (source: own compilation)

Table 3 .
Absolute and relative frequencies of variants of "sine qua non" across corpora (source: own compilation) Supporting legal scholars' efforts to communicate in English as a lingua franca… 3795.3.2.PolysemyProblems caused by polysemy start from the very word prawo, meaning both law and right, sometimes distinguished by adding adjectives.For precision, prawo przedmiotowe [law] and prawo podmiotowe[right]are used, but this may lead to translation errors such as *subjective right [instead of legal right].In certain contexts, such as nadużycie prawa [abuse of rights/ law] or prawo autorskie [copyright (law)], the distinction can be difficult even for lawyers.