Most legal translators doing research are relying on unreliable previously translated texts, often those that were machine translated. The effect on their work and long-term career prospects is destructive because there is no principle for the learning translator to verify the adequacy of those texts and identify a small number of helpful examples. Second, by learning from machine translation tools, translators are focusing on skills that have already been automated.
A broader implication is that most governmental and legal texts ultimately derive from a “Globish” language strategy that is really “Patriotic English”—a teaching strategy adopted in many countries that emphasizes using ESL learning as a vehicle for national pride. Patriotic English is generally only understandable to the domestic audience, not to native English speakers, and is easily automated by machines due to the algorithmic rules defining Patriotic Englishes. In this article, I’ll first introduce the problem posed by patriotic English and then describe how to use a science-based translation approach to ensure high quality for clients.
Patriotic English Is Not for International Use
As has long been widely reported in the media, many countries with universal English education make patriotism in the motherland a core focus of their curriculum. As a result, the English curriculum sacrifices a core quality essential to its success—the authenticity or genuineness of the language being studied.
In China, this manifests in English students devoting a great deal of time to studying the China Daily, a Patriotic English publication where all content is rigorously controlled by Chinese native speakers. The publication frequently uses many China-only English expressions and ordinary everyday things are expressed with vocabulary and style totally alien to authentic English texts.
This situation is not unique to China; it applies to almost every country but is particularly significant in places with strong traditions, like Pakistan, France, China, Korea, and Japan. These countries understandably want to ensure the next generation remains connected with their ancestral traditions, despite forcing the entire country to learn American language and culture. I think Spain’s highly anti-nationalist culture can explain much of the superior quality of Spanish<>English translation compared to the opposite end of the spectrum, such as Russian or Urdu.
Patriotic English Has Anti-colonial Value
My own opinion on the subject of Patriotic English is that it’s a form of resistance to the colonialism or cultural incursion of dominant foreign powers, made necessary to some extent by globalization. From the 1980s to the 2020s, the world economy was very much driven by developing economies adopting low-wage manufacturing, which necessitated entire populations learn English for fairly basic jobs.
However, you will notice that more sophisticated work, such as technological innovation or legal services, wasn’t being done at all in these economies. If you turn the page back on history, you will see that intellectual property infringement was absolutely rampant and bribery—not the legal process—was needed to get anything done. In other words, only very crude English was needed to perform crude tasks: assemble plastic toys, haul them to a ship, and drop cash bribes off at government offices.
In this context, developing a highly patriotic but otherwise very limited English was perfectly feasible, and countries like China and Pakistan were enthusiastically on board. In the case of China, which I am most familiar with, Patriotic English emerged with the help of experts and translation software, scrambling and reassigning words to create what was essentially an entirely new language comparable to Esperanto.
If you’re not familiar with Esperanto, the idea was to combine features of several languages to create a low-difficulty language for international communication. Therefore, China Patriotic English would enable so-called China Experts to engage with the country without ever learning Chinese and enable Chinese speakers to engage with those foreign China Experts without needing to learn English. Instead, speakers of the language could memorize a list of roughly 4,000 words and use them to label each Chinese word to corresponding terms in their native languages. This was very useful in China until the advent of high technology in 2020.
New Technologies Make Patriotic English Obsolete
Electric cars, solar panels, TikTok, and ChatGPT are the four horsemen of the apocalypse for China Patriotic English. Now, ChatGPT can totally automate very routine machine translations, like writing a letter asking to produce 3,000 plastic toys and ship them within two weeks. Additionally, the Chinese economy wishes to focus on high-tech exports under a system that emphasizes the rule of law and respect for procedure, which requires very advanced language skills.
I recall translating a privacy policy for a Chinese electric car maker intended for use in international markets. While reviewing the policy, I noticed language that could cause scandal. The internal team, however, was very insistent on retaining this scandal-prone language, which had apparently come from a machine translation tool. They had English skills, but not ILR5 in legal English.
I predicted that a CNBC or Fox News report could cause substantial brand damage as a result, and the language was ultimately fixed without the semantic changes affecting business operations. Later on, CNBC did air a scandalous report questioning whether Chinese electric vehicles were spying on you and listed several privacy policies that were reviewed—but ours wasn’t on the list.
Apparently, while native English-speaking lawyers drafted the policies with an eye to the court of law, they overlooked the broader implications in the court of public opinion. In this case, the use of China Patriotic English could have caused millions of dollars in brand equity damage to the company—not just by framing them as another “evil tech company,” but by fueling suspicions of nefarious foreign spying.
Therefore, Patriotic English no longer serves the needs of a modern economy—only highly authentic English, or Chinese, or French will suffice. As Anthony Pym (2023) notes in his chapter “Automation,” advances in translation technology may eventually eliminate the need for universal English education. This is good news as it ensures the unique cultures of China, Pakistan, and France can be preserved and will not be whitewashed away!
How to Base Translational Language Use in Authentic Documents
When modeling language, a translator needs to use authentic and genuine texts and be capable of distinguishing them from artificial ones. Usually, drawing a line between natural and artificial things is easy, such as recognizing a tree as natural and a telephone pole as artificial. In this case, since all written texts are human-made, making this distinction is harder. However, identifying what’s clearly artificial, and thus inauthentic, is relatively easy.
Texts produced through translation are basically always artificial and not reliable for any kind of linguistic emulation. That should also be extended to texts produced that incorporate translated material. For example, if a journalist in New York reports on China using documents translated in Beijing, then the resulting English news report is going to include a lot of translated material and, thus, be unreliable due to a lack of authenticity.
Similarly, artificial intelligence tends to produce very noticeable “AI-ese,” which readers can often recognize quite easily. Nevertheless, despite its unnatural phrasing, AI-ese is still generally of much better quality than ESL Englishes. That can tell us something about the kinds of factors that make for authentic English.
A simple four-factor test of the source text can be applied to ensure that a reference text is suitably authentic to be used in translation:
- The text should be written by native speakers from the target language region. For example, when looking at a contract, check the name of the law firm and its contact address in the notices section to find out who wrote it. Often, you’ll find that a foreign company doing business in the United States imposes translationese on local business documents, as seen in the JAC Motors lawsuit case.
- The text should be about things occurring in the language’s native area. For instance, a big law firm might file leases for both New York and Paris office buildings. However, there is a huge chance that a Paris lease will contain translated text or concepts woven into it.
- The text should be talking about the specific subject matter within its professional field. For example, if working on a commercial lease, you should not be referencing contracts for buying corporate stocks and bonds—an office is not a stock.
- The text should have no foreign elements in it. You will need to check names of companies, products, and places that don’t appear in the target language region. When doing research with a search engine, be mindful that there is a high chance the search engine will recommend translated texts for someone who is researching how to do a translation. This last point is confusing so requires some additional explanation.
It may seem like most English is going to be authentic, so avoiding inauthentic English should be easy. However, modern search engines have made that an unrealistic expectation. The first factor is that keyword searches will often be based on translations, thus increasing the likelihood of seeing more translations rather than native texts. Second, to add insult to injury, search engines use geolocation, meaning that you need to have specific configurations for the target region to get good search results. Otherwise, search engines will send you geographically local materials and, if you’re not in the United States or United Kingdom, you’ll be referred to a lot of documents with less than useful English.
The Science-Based Reasons for Using Authenticity for Translations
There are numerous science-based reasons for using authenticity for translations. A good starting point is in the second language acquisition field, where research consistently shows that learners who engage with authentic texts when learning their second language perform better at communicative tasks. A translator using reference documents is, to a large extent, using them to learn the expert-level language relevant to their target language. Learning how to “talk like a lawyer” or “talk like a CEO” is not easy and requires substantial exposure to the authentic language used within those professional fields.
A second science-based reason for using authentic texts comes from recent advances in sociolinguistics, particularly around the discourse community concept. The idea of a discourse community is now so powerful that entire scientific fields, such as discourse analysis, and journals are based around inquiries into the discourse used in communities. By following the approach I outlined above, you can tap into the knowledge and experience of the relevant discourse community and deploy that knowledge to your translation.
For example, if your client belongs to the discourse community of New York securities lawyers, then following these steps ensures that the translated document she’s reading will have all the conceptual and communicative characteristics of that legal community. A foreign lawyer in Mexico City or Shanghai will thus be able to interact and collaborate with the target reader seamlessly, free from linguistic or conceptual barriers.
Learn from People, Not Translations
To sum up what we’ve learned here, translators—like any other kind of professional—need to learn continuously. The challenge for translators is finding the right sources for learning, and the way to do that is to select authentic, genuine, and relevant resources that can provide a strong foundation for your translation project.
