Language diversity in research and in research evaluation is a key issue for the social sciences and the humanities (SSH). Over the past years, a lively and interesting professional discussion has been developing regarding the specificities and the relative merits of multilingualism and of plurilingualism. However, such specialised debates in the SSH world should not lose sight that one common principle —More than one language— is a key starting point in addressing language diversity issues and that this principle needs to be actively promoted as a core value of scientific knowledge advancement in the SSH disciplines. If this principle gets a stronger recognition as a feature of scientific excellence, it will be considered in the current process of “reevaluating evaluation” of the SSH contribution and impact, to use Prof. Milena Zic-Fuchs’ terminology in her opening remarks.

There is a marked differentiated situation in the SSH compared to exact or experimental sciences as far as language as a scientific tool is concerned, adding to other significant epistemic and methodological differences raised by Prof. Andrea Bonnacorsi in his contribution. In SSH, understanding the construction process of situated concepts and categories, using national and local historical sources, and interacting with living people for interviews and surveys require the mastering of national and local languages to deliver high-quality research on a wide range of scientific and societal issues. This is all the more true for “language-embedded” disciplines or for those that are as concerned with questions of the form of an argument as other disciplines are of the function, to rephrase Prof. Jennifer Edmond’s point. The representation of questions by agents is local, plural, impacted by history, and dependent from cultural and linguistic characteristics. As noted by Prof. Riccardo Pozzo, this situation as no equivalent in so-called « hard » sciences.

It may also be noted that the situation is differentiated across the SSH spectrum. Some disciplines, notably those that make an intensive use of models and of data (such as economics), and those scholars who make an intensive use of models and of data in their research practices (whether in the fields of psychology, linguistics, demography, political science, sociology, etc.) tend to privilege English as the main language for scientific publications. If we look at the Nobel Prize Award in Economics over the past 35 years, all winners (whatever their nationality and mother language) published in English and were recognised for their ground- breaking articles in leading journals published in English.

More surprisingly, and this is related to the well-known language biases in expert-assessment, metrics and impact assessment underlined by Prof. Janne Pölönen in his paper, over 80% of the winners of the Holberg Prize, which is often considered as the Nobel Prize for SSH, publish in English. The award winners were from a large set of disciplines ranging from sociology to philosophy, from history to law, from social anthropology to semiotics. This underlines the question of the dominance of English-speaking publishing houses that organise the market for printed monographies as well as the market for academic journals. And, as a consequence, the twin questions of the visibility and the impact of publications in other languages than English, and that of their integration in a more balanced scientific assessment of scholarship achievement. These key issues will be discussed in other sessions of the conference.

One way forward is to support advancement of open digital infrastructure in a way that would ensure — no matter the language— discoverability, accessibility and legibility of relevant data, resources and material published on a specific topic. By doing this, three main issues of the SSH communities would be addressed at the same time: (a) fairness in contents access, (b) dissemination of open scholarly communication practices, (c) enhancement of SSH impact in scientific assessment.