Reading time: 3”
Every lawyer has at least once dreamed of an ideal translation tool: accurate, efficient and instantaneous. But this “ideal” remains elusive, even in the age of artificial intelligence. So why is translating legal content so difficult?
The difficulties of legal translation
To fully understand the difficulties of legal translation, we must first consider the distinction between general and specialized languages. The former is concerned with everyday, vernacular language, while the latter focuses on “specialty languages”, in other words, technical jargon specific to a given discipline.
This boundary is fundamental to our approach to the difficulties of translation. In the general language, the translator’s challenge is to take into account the dynamism of the “living” language and find the right equivalent to convey the same idea. The case of idiomatic expressions is a good illustration of the “cultural” analogies that need to be made when moving from one language to another. « Il pleut des cordes » becomes « it’s raining cats and dogs » : an identical meaning for a completely different syntax.
In specialized languages, the problem is the opposite: the translator doesn’t have to grope for an equivalent, but has to find the exact and only equivalent that exists. As a general rule, technical terms have a universal semantics, which enables this perfect equivalence between languages. The other side of the coin is that “specialized” translation allows no approximation: you have to find the exact term, otherwise the technical information loses its meaning.
So what about legal language? Let’s just say that this type of language straddles the border between the two. As with translations of everyday language, legal language requires a certain flexibility in translation, due to the distinct “cultural” lexicons of different legal systems. Nevertheless, the translator is obliged to be precise in the terms translated, in order to maintain the thread of legal reasoning, just as in technical translation.
Is machine translation up to the task ?
Like a student, the AI used for machine translation relies on learning. Based on a deep learning system, this type of AI is particularly useful for taking cultural discrepancies into account. By dint of examples, trials, errors and corrections, the system refines its translation capabilities and possible analogies in the event of ambiguous wording. Like human translators, these systems are first trained on the general language before considering specializations.
Let’s take the translation of legal documents as an example : « Only by training AI on data covering different legal fields can we acquire the precision needed in this highly demanding field » points out Jean Senellart, CEO of SYSTRAN, a French publisher of machine translation solutions. « The important thing is to provide professionals with the most accurate linguistic analogy for the legal reasoning being translated. But most of the time, companies don’t have the necessary data to “feed” the machine: this is where we need to provide a turnkey solution that offers customized AI models for each business domain ».
The real challenges of legal translation are also linked to liability and confidentiality. Legal language isn’t just used to convey a message or express an idea: in reality, it enshrines obligations or responsibilities, and each term used can have direct repercussions for the company, the individual, etc.
As with scientific languages, the margin for error is very small, since the right or wrong choice of terms can engage the responsibility of the legal actor. A badly translated contract, an ambiguity in an authenticated deed, and the liability (and reputation) of the lawyer, notary or legal director is engaged. Machine translation, which is a tool under the responsibility of the end user, must above all avoid approximations or even misunderstandings.
Another problem, and not the least important, is that of confidentiality. Using an automatic translator often means sharing data online. When it comes to translating your American cousin’s e-mail, no questions are asked. On the other hand, type in your contract with a listed multinational on a “generalist/general public” translator (i.e. one that will analyze your data because that’s its business model) and you’ll think twice before pressing enter.
What are customers waiting for ? For Tehani Somoguy, of Alerion Avocats, « the solution that’s right for me is a tool that quickly translates all types of documents, even PDFs, into all the customer’s languages, and gives me the most accurate working basis possible. And above all, one that doesn’t exploit the data I’ve translated in any way ».
A performance factor today, a must tomorrow
To sum up, translating legal language means being flexible in its understanding, precise in its translation, safe in its use and secure in its input. Quite a program.
And yet, being able to translate legal documents quickly and efficiently is becoming a real asset in an economy where foreignness is becoming a daily part of doing business. Proof of this need is provided by the fact that the professional translation market (including the legal sector) is now estimated to be worth almost 45 billion euros (2020).
This dynamism has enabled translation professionals to develop new translation tools, integrating the aforementioned constraints.
Returning to the legal market, having the right translation tool has several benefits. The first is to optimize internal processes, thanks to native tools integrated into the work environment of the legal profession. At a time when success depends on responsiveness, integrating these new solutions gives you a head start in terms of time optimization, particularly in the face of competitors who are less well-equipped and, inevitably, less responsive.
The second advantage is cost reduction. By generating a first version, imperfect though it may be, which is already coherent, the right machine translation tool enables you to limit the intervention of an in-house or, what’s more, external translator, who will concentrate on a rapid and effective revision at lower cost.
But the adoption of machine translation tools has an ultimate purpose: the future. As with other digital transformation solutions, the use of translation tools can only increase in the months and years to come. What is a performance factor today will be fundamental tomorrow for any company wishing to keep up with the pace of business.
To gain a better understanding of the translation needs of the legal profession, we’d like to invite you to take part in our (which we promise you will take less than 10 minutes to complete).
Responses are, of course, anonymous, and you will have access to the results of this survey.
