You are here

Home

Can we make English-language publishing more accessible? COPE editorial

In this editorial we focus on some of the potential barriers faced by people communicating in a second or additional language, and collate a range of positive actions for the scholarly publishing industry. We have produced this editorial in conjunction with some of our global members and contacts, and some of their own insights accompany this article. We will follow up with more responses and we welcome further insights from readers in the comments section below, or by email. 

Page contents

- Introduction

- Linguistic diversity

- A cycle of discrimination?

- Economic resources matter

- Unspoken language codes

- Solutions

- Guest responses to this editorial

Introduction

The vast majority of modern scholarly literature is written in English: 98% of peer-reviewed science according to one recent estimate. This is changing slowly, but is accompanied by a growing trend for journals in other languages to switch to English because it raises their profile and their utility to authors seeking promotion and tenure. While it’s convenient to have a common language of scholarly communication the fact that this is English is largely a legacy of colonialism and global wealth disparities. Now, growing numbers of authors are pointing out that this is not equitable: it gives some researchers innate advantages over others; it creates an impression that scholarship in English is intrinsically of greater merit; and it brings increased mental and financial toll for authors with English as an additional language. 

The result is a series of structural biases which can accumulate to restrict equal access to scholarly publishing. A letter published in a conservation journal in 2021 found that authors from lower-income countries as well as Russia, Japan, and South Korea are ‘strikingly underrepresented’ in leading journals in ecology, evolution, and conservation. The top three countries represented are all English-speaking, high-income, and related through a history of colonialism (the United Kingdom, the United States, and Australia). This is hardly surprising given that these journals and their peer reviewers are also disproportionately located in North America and Europe. Another recent study of over 700 ecology journals concluded that we are in ‘a grim landscape where most journals [are] making minimal efforts to overcome language barriers’. In short: there is a growing opinion that scholarly publishing needs to become more inclusive when it comes to linguistic and cultural backgrounds. 

The benefits of linguistic diversity are both numerous and obvious, and many were stated clearly in UNESCO’s 2021 recommendations on Open Science. They include equity as a core human goal; a richer and more diverse scholarly literature which benefits human understanding and policy; and equal access to opportunities for personal career development and professional status. Linguistic diversity promotes broader engagement with research and vastly increases the chances that it will reach local and national policy makers, subject populations, and stakeholders. There is even evidence that diversity improves healthcare and health outcomes, increases workplace performance, and encourages innovation. In the words of one of our contributors, it enables scholars to understand the world better and makes our lives more colourful.

Linguistic diversity

Scholars experience linguistic differences in different ways. Some learn additional languages from childhood; others more formally as adults. For some, the mental shift of writing in another language brings benefits: a different logical approach, for instance, or access to a different form of terminology or linguistic structure. For others, fluent expression is difficult to achieve, perhaps because their own language is very different from English, and that shift in communication is jarring. Further, while no other language has the scholarly publishing networks that English does, some (Spanish, for instance) have more outlets than, for example, Mandarin or Portuguese. Although we are focusing here on linguistic familiarity it is also important to note that some scholars find expressing themselves in a certain register a challenge because they are neuro-divergent, or simply because writing is not a part of their work that they enjoy, regardless of the language in which they are working.

A cycle of discrimination?

We should not assume that writing in a secondary or tertiary language is challenging for all scholars. However, a number of studies have pointed out a range of negative impacts at an aggregate level, especially for early career researchers who may lack status, contacts, and experience in academic writing in English. Amano et al found that researchers whose first language was not English spent up to 51% more time writing a paper in English than those who did, for example (study participants came from Bangladesh, Bolivia, Britain, Japan, Nepal, Nigeria, Spain and Ukraine). One of our contributors noted the effort required to identify errors in English which a fluent English speaker will spot and correct easily. Presenting research in English may also be difficult: 33% of participants in Ramírez-Castañeda’s study of Colombian researchers in biological sciences chose not to attend international conferences where they would have to participate in English. Similar outcomes have been found among Taiwanese and Mexican scholars. In addition, it can be expensive to access editorial and translation support, all meaning that researchers less confident in English face more hurdles to publishing in the highest ranked or most widely-read journals.

Opposing views exist too, as is highlighted in some of our contributor responses. Hyland, for example, suggests that all scholars - in English-speaking countries or not - face language barriers to some extent, and that economic, physical and geographical characteristics are more important markers of inequity. It might be wise to consider the intersectional effects of economic, physical, geographic, and even gender characteristics on access to scholarly publishing, and bear in mind that these can be experienced by researchers in all countries.

Economic resources matter

Nonetheless, Hyland’s emphasis on economic disparities bears further examination. For example, a 2004 study quantified a 2.17 unit increase in publication rates with each increase in national funding for research in medicine in ‘developed nations’. This was higher than the association with language proficiency (an increase of 1.56 units), though together these two measures explained 71.5% of observed variation in publication rates. One of the reasons is that lower national income tends to go with lower research funding. Less money for research frequently translates into lower university rankings which in turn, means fewer opportunities for career advancement and collaboration on an international stage. This can hinder access to resources and publication in high ranked journals. The same 2004 study, for example, found that Asian countries have generally the lowest publication rates in medical studies involving human participants (an expensive field of study). 

One high-impact solution to language disparities is professional editing services. However, these are costly, and as one of our contributors points out (below) demand has soared with increased article submissions. 59.2% of researchers in Ramírez-Castañeda’s study reported having paid for editing services and 28.6% had paid for a translation. Critics would argue that no one should have to pay for support services which many others do not need, and one contributor told us that they have sometimes provided editing and translation services for free when needed. However, this is not an affordable business practice, and it deflects attention from more sustainable responses from across the industry. Furthermore, it is scholars from lower income countries who are least able to access paid-for services - as well as potentially having less well-resourced research. This translates to the ironic position that publishing is more expensive for the least well-funded researchers. Add in sometimes prohibitive article processing charges (up to half of a doctoral student’s monthly salary in Colombia, for example) and often restricted waiver schemes, and this can become a substantial barrier to publishing. The effect is to marginalise work produced in lower and middle income countries - as the associate editors of one Wiley journal pointed out when announcing a work stoppage over the use of high APCs for open access journals.

Unspoken language codes

Fluency in academic writing requires more than formal language acquisition. Also important is ‘linguistic capital’, which broadly means knowledge of unspoken rules, tone, and idioms. Researchers who lack this capital (which is generally gained via experience) will likely find it harder to achieve ‘publishable’ levels of English. The study by Amano et al, cited above, shows that researchers from countries with low-to-moderate English proficiency experience higher rates of rejection overall because of language issues. Authors lacking linguistic capital are also at greater risk of being deceived by emails from predatory publishers because many warning signs are based on subtle linguistic clues. 

The situation is not helped by the fact that so-called publishable academic writing is very ill-defined and varies from one English-speaking country to another. Few instructions for authors define what is meant by academic English, and interviews with peer reviewers have suggested that it is based on internalised knowledge - or, simply reviewer preference. Authors with good support and information networks (‘social capital’) are more likely to be able to bypass these ambiguities because their connections act as a guarantee for their credentials. This makes anonymous review policies even more important, alongside diversity in reviewer pools.

Solutions

True action requires large-scale cultural change; however, there are smaller steps that will make a difference, and everyone involved can think about what actions they can take. One of the biggest opportunities on offer lies with the advent of artificial intelligence (AI). AI-driven tools like automated translation and editing services can significantly reduce the time and effort required for manuscript preparation and ensure that the focus remains on the content rather than the language proficiency. As machine learning algorithms improve these tools are becoming more accurate and specialised, and hence can play a valuable role in democratising language support and bridging socioeconomic gaps. The editors of one medical journal based in Chile have reported their successful move in 2021 to bilingualism (Spanish and English) and near-simultaneous publication in both languages, utilising AI tools to aid translation and publication processes. Although this required a substantial financial investment which was made possible by income from journal fees and matched grant funding, the editors are keen to stress the feasibility of such initiatives. They also point out the benefits of partner collaborations between lower-income countries (their technology partners are based in India and Chile). 

However, like all areas of AI, translation and editing tools should be handled carefully and with consideration of unintended consequences. These could include the standardisation of academic expression, potentially stifling individual writing styles and cultural nuances; concerns about data privacy; and the possible impact of uploading content to AI platforms which could be repurposed without sufficient originality or critical insight. Education of all stakeholders is needed on the current (and often quickly evolving) strengths and limitations of AI tools in different languages and on the need for careful checking and refinement of the output. One of our contributors reports a growth in highly technical manuscripts that were produced using AI-powered translation or editing tools which contain numerous scientific errors, usually because the authors have not noticed (or do not have sufficient proficiency in another language to recognise) linguistic errors or ambiguities. However, as part of a suite of initiatives and with careful use, they can have huge benefits in moving towards an equitable environment for scholarly publishing - as long as they are accessible to everyone. The editors of the Chilean journal cited above use AI alongside English-speaking and subject-specialist translators who cross check the final version in both languages, and report a huge rise in consistency and quality over translations done by authors. However, it is only the better-resourced publishers and journals who can access custom-built AI tools to bypass concerns about confidentiality, or the paid-for services increasingly built in to services like ChatGPT or Perplexity. Again, concerted action from across the industry will be required to ensure that AI tools remain accessible and don’t become an additional source of inequity

- Publishers

In terms of actions for stakeholder groups, publishers have a particularly big opportunity for impact. Publishers set the tone, the business model, and often the author policies of their journals. If they would publish clear accessibility statements about how language will be assessed this would be an important, but easy-to-achieve statement to authors of all backgrounds that intellectual content is valued above expression. 

Publishers could also consider:

  • promoting double-anonymised review protocols and diversity in editorial boards and reviewer pools.
  • actively seeking out submissions from diverse geographical and socioeconomic areas.

At a higher level of resource they could also:

  • provide free proof-reading services and offer support for coaching and language support for authors with English as an additional language.
  • provide or back AI tools for translations and editing (alongside clear guidance on its responsible and ethical use, and checks of the resulting translations).
  • consider more expansive open access and fee-waiver schemes for authors from lower-income countries and institutions. 
  • consider expanding open access and copyright-free agreements to encourage translations, again with checks in place to judge the accuracy of the translation.
  • set up mentoring programs to support people writing in additional languages.

- Journal editors

Journal editors also have an important part to play in setting policies for authors and reviewers, monitoring linguistic and other forms of diversity on their editorial boards and reviewer pools, and considering new forms of content which are accessible to other languages. For example:

  • publishing all author and reviewer guidelines, as well as journal policies, in a range of languages.
  • reinforcing any publisher policies, or creating their own policies, on the use of AI for editing and translation, and on assessing submissions solely on intellectual content. If there is an expectation about minimum standards of English then these could be specified clearly in author guidelines.
  • encouraging the submission of abstracts and keywords, if not full articles, in other languages than English.
  • dedicating part of their space to announcements, news items, literature reviews, or letters in languages other than English.
  • summarising important contributions published elsewhere in other languages (with links to the original journals).

There is a role for other stakeholders too, especially if we are to see a bigger change in cultures around publishing in English. 

- Funders and research institutions

Funders and research institutions could:

  • give more weight, credit, and financial support to publications in local and other applicable languages.
  • encourage international collaborations and alternative forms of dissemination and output than conventional research articles.
  • encourage co-authorship with people who have been the subject of the research (a positive step for diversity and agency more generally).
  • support translations, and offer training in academic writing and research, and publication ethics.
  • offer grants to encourage publications from researchers in lower-income countries.
  • build in incentives for researchers to engage with multilingual audiences, for example by supporting conference bursaries in grant bids.
  • support or subsidise open access fees for researchers from lower-income countries.
  • set up mentoring programs as suggested above, alongside workshops and online resources on academic writing and publication processes (in multiple languages).

- Conference organisers 

Conference organisers can:

  • more routinely offer simultaneous translation and publication in other languages.
  • offer fee reductions and waivers to promote accessibility.

The promotion of multilingualism in scholarly publishing requires capacity building and shared communities of practice. There are encouraging moves in this direction, including the Helsinki Initiative on Multilingualism in Scholarly Communication. There are also growing numbers of initiatives by individual journals or organisations; some examples are highlighted in this 2024 study of journals in biology, which also sets out a range of solutions. One IOP journal has introduced Editorial Fellowships for scholars from historically excluded groups; eLife has set up a Global South Committee for Open Science; and some ACS journals are trialling a free AI-based language editing service. We are grateful to members of the C4DISC Community of Practice group for publicising initiatives like these. 

All of these efforts should be applauded, and will hopefully in time contribute to a more multilingual, multicultural and accessible industry. As one of our contributors put it as we prepared this commentary, ‘ultimately, embracing linguistic diversity is not just an ethical imperative but also a catalyst for innovation and a broader dissemination of knowledge’.

We hope that this editorial will encourage debate from across our readership and we thank our contributors for adding their own perspectives to these issues.

Guest responses to this editorial

COPE is working on a new DEIA policy which will set out our commitments to improving diversity in our leadership and membership, and enhancing the accessibility of our resources to audiences from a range of linguistic backgrounds.

Conceptualisation: Alysa Levene, in conjunction with COPE’s Diversity, Equity, Inclusivity, and Accessibility subcommittee.

Writing - original draft preparation: Alysa Levene

Writing - review and editing: Rania Abdelghani, Tiziano Innocenti (GIMBE), Tony Alves, Vesna Skrbinjek, Haseeb Irfanullah, Lisa DeTora, Tash Brown, and Muhammad Sarwar, Maryam Sayab, all of the Asian Council of Science Editors, Caryn Jones (ThinkSCIENCE - Japan), Professor Lvxiang (Roger) Deng, Mabel Chew, Trevor Lane, Itamar Ashkenazi, Anubhav Pradhan

Related resources and further reading