The increasing dominance of English as the language of research often leads to the suggestion that scholars from emerging countries need to publish in the language to increase recognition.

But from a survey of researchers for whom English is not their mother tongue indicates that a large body of journal submissions are still made in other languages, even in the natural sciences.

For the study, researchers from the University of Ottawa surveyed almost 900 non-native English-speaking researchers publishing in Springer journals across a range of disciplines.

Although answers from the respondents, mostly based across Europe, Latin America and Asia, confirmed that English was the most used language for journal submissions, ¡°its dominance¡± was ¡°far from complete¡±, the paper, published in Scientometrics, says.

É«ºÐÖ±²¥

On average, those surveyed said?that about 40 per cent of their recent article submissions were in a language other than English, and only 16 per cent?reported that they had sought to publish everything in English.

The perception that publishing in English might improve the reputation of an individual scholar appeared to play a greater role than any pressure from institutions or funders, the study also suggests.

É«ºÐÖ±²¥

On average, those agreeing that publishing in English was good for reputation submitted 63 per cent of their articles in the language compared with 52 per cent for those disagreeing, but the gap was?only 5 percentage points for those who reported institutional pressure to publish in English and those who did not.

A researcher¡¯s likelihood of seeking publication in English also seemed to increase along with other factors including their age, whether they were working for a university (rather than another organisation like a thinktank) and their location. For instance, on average European researchers submitted 61 per cent of their articles in English, compared with 54 per cent in Latin America and 57 per cent in Asia.

However, perhaps the most striking result came when looking at the difference between disciplines.

Although it is often assumed that publication in English is much more likely in the natural sciences, on average researchers in these fields still submitted about?35 per cent of their articles in another language, a smaller percentage than the social sciences (44 per cent) but not a huge difference.

É«ºÐÖ±²¥

Michael Wigginton, a PhD student in the School of Political Studies at Ottawa, who co-authored the study with Daniel Stockemer, associate professor in the school, said that he was ¡°fairly surprised¡± by the overall average for research submissions in English because he ¡°had expected the number to be quite a bit higher¡±.

¡°I was particularly surprised to see the relatively small difference between the social and natural sciences, as my impression based [on] previous research had been that English was far more dominant in fields such as biology and chemistry, compared with sociology or political science,¡± he said.

He added that the research also highlighted the potential problems with trying to compare the citation impact of research in different languages, which risked being ¡°an apples-to-oranges comparison¡± given the relative size of the audiences.

¡°Knowing that researchers publish so much in other languages, I think it¡¯s worthwhile to consider how relying too much on these metrics risks undervaluing the importance and impact of non-English scholarship,¡± Mr Wigginton said.

É«ºÐÖ±²¥

simon.baker@timeshighereducation.com