The giant shoulders of English

“EVERYONE who matters speaks English.” So say many in Britain and America. In fact, a lot of people do not. But in some domains, this crude approximation is true: in globalised enterprises the world’s single scholarly language is increasingly indispensable. Among those global enterprises is science, in which more and more work is being done in English. This is not always good.

A scientific lingua franca has advantages. A few moments imagining scientists toiling away in different countries unaware of each other’s successes and failures is enough to show that. For centuries, Latin allowed the Copernicuses, Keplers and Newtons of Europe to stand, in Newton’s words, “on the shoulders of giants” who had preceded them. With the rise of European vernaculars as “serious” languages, an educated person was expected to read several; German was a leading language of science.

Now, non-Anglophone scientists learn English; English-speaking scientists hardly bother with other languages at all. The rise in perceived need for more STEM subjects (science, technology, engineering and maths) has made schools squeeze anything that looks dispensable, and in the English-speaking world that includes foreign languages. Legislators in Florida have even proposed letting schoolchildren learn a computer language to satisfy schools’ foreign-language requirements.

Three scientists have raised an alarm about English-only science in a paper in PLOS Biology, a journal. Tatsuya Amano, Juan González-Varo and William Sutherland looked at fields where local knowledge matters, such as ecology and conservation. They found that 64.4% of papers on Google Scholar mentioning “conservation” or “biodiversity” were in English. The second most common language, Spanish, was far behind, with 12.6%.

Monolingual ghettos are bad for science. In 2004, work on the transfer of H5N1 flu from birds to pigs languished unread in Chinese while critical time was lost. In the study’s sample, only half of Spanish-language papers and a third of those in Japanese even had abstracts in English. Those that did, unsurprisingly, were more likely to be published in prestigious, peer-reviewed journals. But the bird-flu case shows that that hardly includes all the science that matters. Some good scientists still can’t write in English.

The solution is not to replace English, but to encourage multilingualism wherever practical, and require it when needed. This can be an advantage for non-native English-speakers. Studies have shown that writing and thinking in a second language can encourage a deliberate mode of thinking. Working in your native language encourages the fluid kind. A bilingual person can have the best of both.

Multilingualism is needed in other ways. In disciplines including psychology, biology and medicine, university-based researchers will work with subjects (patients, for example) and data-gatherers (say, remote experts in local flora and fauna) in other languages. The bilingual scientist who can later write all this up in English has a competitive advantage.

More and more young scientists will speak English as a matter of course. They should ensure that clear English abstracts and keywords from their papers are available; this may be more important than the original abstract itself. But Anglophone scholars and institutions can also play a role. Where work is of particular importance to a particular country or region, they too should make sure that abstracts and keywords are available in relevant languages. Groups of scholars can share the cost of full high-quality translations.

Changing practices takes time. Until then, some technological tools can help. Machine translation (MT) has improved in recent years. And specialised MT systems—say, those designed specifically to handle texts in a field like ecology—are far more accurate than general systems that are designed for all kinds of text (like those that are free online). Building such systems is getting cheaper and easier. If scientists could support the development of such MT systems for their fields, they could increasingly get usable gists of abstracts instantly, and find out which work might be worth full translation.

The alternative is a future in which all work is done in English. In such a world, other languages would fail to develop the kinds of technical vocabulary and expressions needed for science. They would be used socially and at home, but not for serious work. That would be a shame.