Skip to main content
Home

Multilinguality from Research Data Management to Publishing

Introduction

How would you ensure output dissemination & data sharing in multiple languages in multiple countries? How would you ensure scientific content is described with appropriate descriptive multilingual metadata to ensure good ‘findability’ and visibility? This resource equips researchers with strategies to manage, publish, and disseminate multilingual research data. It covers data management plans, Linked Data identifiers, FAIR (Findable, Accessible, Interoperable, and Reusable) and CARE (Collective Benefit; Authority to Control; Responsibility; Ethics) principles, and language-aware publishing. Through practical examples and reflective activities, it promotes inclusivity, local community engagement, and reduced English dominance in scholarly communication. It is aimed at Digital Humanists, and researchers in Linguistics or Humanities.

Learning Outcomes

After completing this training resource, learners should be able to:

  • Know the strategies, tools, and standards to manage multilingual research data
  • Plan multilingual research data in order to improve findability and visibility of language diversity
  • Publish and disseminate multilingual research

This module is accompanied by examples, reflexive activities and concrete use cases related to multilingual projects.

1. Strategies, Tools and Standards in Managing Multilingual Research Data

A research dataset typically contains data about entities and/or relations between entities. If your goal is to make your research dataset readable in several languages, you do not need to translate the names of the entities you describe to other languages. You would rather choose to describe these entities using a Uniform Resource Identifier (URI), which enables you to retrieve multilingual labels (and more), and use these as entity descriptor, instead of literal values (names). If you start from an existing dataset containing literal values instead of identifiers, you can use tools like OpenRefine for finding the URI you need.

In the following, we give main recommendations for which sources for web identifiers to choose, together with examples. Clicking on the URI, you get to lists of multilingual entity labels (and more information). If you use a Linked Data database software, you can import all these labels (together with any other information found in these sources) into your own database environment.

1.1 Open Source Wikimedia Tools and Platforms for Multilingual Purposes 

Wikimedia tools are interoperable and multilingual from the start, enabling the integration of researchers’ content and data in the semantic web to expand research impact beyond Academia.

Wikimedia logo

Fig 1. Wikimedia logo

MediaWiki is a multilingual documentation platform with advanced collaborative features that is used by Wikipedia and the other Wikimedia tools.

Wikibase is a set of MediaWiki extensions that enables the use of semantically organized data. This suite supports the Wikidata platform.

It can also be used by researchers with no maintenance costs in a cloud service. Researchers may create Wikibase.cloud service instances to host their own documentation and data and connect it to Wikidata, functioning as an indexer hub, enabling federated queries with organizations like GLAM (Galleries, Libraries, Archives and Museums) institutions (GLAM cultural heritage).

Interior view of the Musée d'Orsay in Paris

Fig 2. GLAM concept in Wikipedia. Interior view of the Musée d’Orsay in Paris

Wikisource is the digital library for hosting texts in public domain and creative commons licences. It integrates powerful OCR (Optical Character Recognition) and HTR (Handwritten Text Recognition)  tools (free credits within Wikisource) to enable digital publishing and content linking.

Wikipedia uses an encyclopedic, simpler language and may be used to reference researchers publications. An interesting academic extension activity is to promote editathons in the language of target communities and reach wider publics.

Set of Recommendations and Examples

  1. Use a database solution that provides a multilingual user interface

Wikibase is a set of extensions to MediaWiki, which is available in 215 languages. Translations are crowdsourced, you can contribute. To learn more about Wikibase, visit DHwiki Wikibase. Wikibase supports URI natively, in a way that users see entity labels in their preferred language, while, under the hood, the entity is represented by a web identifier, which is the same for all its labels, in all available languages.

  1. Do not use literal values (names) for representing entities, and make use of established Linked Data identifiers instead.

Choose identifiers that lead to human readable labels in multiple languages. CESSDA (Consortium of European Social Science Data Archives) lists vocabularies that contain such multilingual identifiers. A general source for multilingually labelled identifiers for entities of all domains is Wikidata. 

  1. Using URI (identifiers on the web) instead of literal values will, by the way, make your dataset ready for Linked Data.

For sharing your research data as Five-Star Open Data, you should use W3C-recommended ontologies for modeling your data. For example, for lexical data, you would use Ontolex-Lemon.

  1. If a Wikidata entity misses a label in one of the languages you want to cover, just add it to Wikidata, so that it becomes available for everyone.
  2. If you have a dataset containing literal values, you can use OpenRefine for finding the corresponding identifiers, e.g. in Wikidata.

1.2 CLARIN Open Source Tools for Multilingual Text Analysis

The Common Language Resources and Technology Infrastructure (CLARIN) platform is a digital research infrastructure that provides language data, tools, and services for researchers in the humanities and social sciences following FAIR principles. This platform is multilingual and supports low-resource and underrepresented languages. It maps Natural Language Processing tools to support text and analysis (e.g., Part of Speech, Named Entity Recognition, Open Large Language Models, dictionaries, parallel corpora). It actively supports projects like LLMs4EU to build open multilingual Artificial Intelligence models for underrepresented European languages.

The Language Resource Switchboard is the CLARIN tool that enables the language detection of uploaded resources and lists all compatible NLP tools available in the CLARIN repository.

The Virtual Language Observatory (VLO) enables researchers to explore linguistic resources, services and tools available within CLARIN and related communities. Its aim is to provide an easy-to-use interface, allowing for a uniform search and discovery process for a large number of resources from a wide variety of domains and providers. 

2. Plan Multilingual Research Data

Making research data multilingual contributes to the implementation of the FAIR principles and supports more inclusive, accessible and equitable research practices. 

Research data are not produced in a linguistically neutral environment. Throughout the research lifecycle, researchers make decisions about which languages to use when formulating research questions, engaging with participants, collecting and interpreting data, collaborating with others, documenting research processes and sharing results. These decisions can influence not only how research data are communicated, but also how knowledge is produced, understood, recognised and reused.

Language is therefore not simply a means of accessing research data; it can also shape what those data mean, how they are interpreted and whose knowledge is represented.

For example, when data are collected in one language but analysed or documented in another, certain concepts, categories or culturally specific meanings may be transformed, simplified or lost through translation. Similarly, the decision to describe a dataset exclusively in English may increase its visibility within international academic infrastructures while limiting its discoverability and usability among the communities from which the data originated.

This highlights the importance of considering the linguistic dimensions of research data throughout their lifecycle, rather than treating multilingualism solely as a matter of dissemination.

Planning multilingual research data also means recognising that language choices are shaped by institutional expectations, disciplinary conventions, available resources and unequal relations of linguistic power. Not all languages enjoy the same levels of visibility, technological support or recognition within research environments. Consequently, decisions about which languages to include, prioritise or exclude may affect who can participate in research and whose knowledge becomes accessible or reusable.

Researchers should consider which communities may benefit from their data, which linguistic resources those communities use, and which elements of the research data require multilingual representation. These may include metadata, documentation, research instruments, terminology, variable descriptions, codebooks and other materials necessary for understanding and reusing the data.

Ultimately, planning multilingual research data is both a practical and a reflective process. It involves anticipating linguistic needs, identifying potential barriers and considering how language choices may enable or constrain participation in the production and reuse of knowledge.

2.1 Include Language Diversity in Your Data Management Plan (DMP) and Use Descriptive Metadata

In your DMP, include your outputs and data in the language of the country of study to make sure it’s accessible and visible to the local communities, and ensure inclusivity (avoid research extractivism). You have to make sure your research process is aligned with CARE principles.

C (Collective Benefit): Data ecosystems shall be designed and function in ways that enable Indigenous Peoples to derive benefit from the data.

A (Authority to Control): Indigenous Peoples’ rights and interests in Indigenous data must be recognised and their authority to control such data be empowered.

R (Responsibility): Those working with Indigenous data have a responsibility to share how those data are used to support Indigenous Peoples’ self- determination and collective benefit.

E (Ethics): Indigenous Peoples’ rights and wellbeing should be the primary concern at all stages of the data life cycle and across the data ecosystem.

In your DMP, translate your researcher profile into the language of the community you will be working with.

Using descriptive metadata is key to ensure findability of research data. For content indexation, libraries use controlled vocabularies instead of freely chosen keywords. The Library of Congress Subject Headings (LCSH) are a widely used controlled vocabulary for that purpose. The use of Linked Data enhances multilinguality. 

2.2 Use a Workflow to Manage a Corpora in Under-Resourced Languages

Four workflows were designed during the workshop “Creating Managing and Archiving Textual Corpora in Under-resourced Languages” co-organised by DARIAH Working Groups Research Data Management and Multilingual DH in August 2024. It brought together experts from different fields of studies working on corpora to discuss how to address pressing issues in the management of under-resourced languages.

The workflow ‘Managing corpora in under-resourced languages’ provides a complete set of steps to manage your corpora while staying compliant with FAIR and Open Science principles. Best practices, dedicated tools and references to other workflows are provided. 

  • Preparing material
  • Documenting your Corpora
  • Creating metadata
  • Using standardised vocabularies
  • Updating your Data Management Plan
  • Assigning (open) license
  • Versioning

2.3 Take the CARE Principles into Account in Your DMP

Collective Benefit and Authority to Control is most of the time enabled by providing key research processes and results documentation in the language of the community that is being studied.

  • For digital content that is culturally sensitive, researchers may use a baseline license from Creative Commons: Creative Commons Attribution-Non Commercial-No Derivatives (CC BY-NC-ND)
  • For signalling culturally sensitive and communicate community-defined protocols for its use, researchers may use Traditional Knowledge Labels - TK Labels
  • To ensure openness does not compromise collective benefit, consider adopting Less Open Equity-based Licenses - NOODLE 

Implementing licenses and practices that align with CARE principles to ensure community-centric control may also imply using the community’s own hardware and human maintenance infrastructure. Linked open data ecosystems like Wikidata, Wikibase.cloud, and Wikimedia Commons and other public infrastructure operate under very strict open licensing frameworks that strictly enforce FAIR principles. The Very Small GLAM project describes the implementation of Wikimedia-based tools on-premise. Even if an on-premise setup may be ideal as a sovereign solution, it may still require too many resources. A sub-optimal alternative may be to use other hosting services.

2.4 Sociolinguistic Approach of Multilingual Research

Planning multilingual research data involves more than translating datasets or metadata after research has been completed. It requires recognising linguistic diversity as an integral dimension of research data management from the outset.

This is particularly relevant when research involves communities whose languages differ from those predominantly used in academic institutions, scientific publications or research infrastructures. If a researcher’s target community communicates in a different language, members of that community should, wherever possible, be able to discover, access, understand and reuse the research data in ways that are linguistically meaningful to them.

Communities that currently lack the institutional, financial or technical resources to conduct similar research independently may nevertheless possess valuable linguistic, cultural and contextual knowledge that enables them to interpret, enrich and reuse existing data. Providing multilingual research resources can help create opportunities for these communities to engage with research findings, develop new applications and contribute to further knowledge production.

3. Publish and Disseminate Multilingual Research

Currently, 95% of all global scientific publications are published in English. The second language used is Spanish which accounts for 1% of the total (González Alcaide et al., 2009; Badillo, 2022; Instituto Cervantes, 2023). This has many implications at many levels. Citizen engagement with scientific discoveries across disciplines is highly dependent on access to the appropriate content in their mother tongue. The progress of new technologies in machine translation of texts opens up a new range of opportunities, and challenges, for the use of languages other than English in research. However, the absence of a significant body of (scientific) publications in a particular language is hindering the research and development of the necessary AI tools. An increasing number of scholars are drawing attention to the challenges posed to researchers, science, and society by the emphasis on using English as the primary language for scholarly communication.

3.1 Activity “Inside a Multilingual Research Journal”

Academic publishing is a key site in which decisions about language, audience, visibility, and scholarly legitimacy are made. Journals working in multilingual fields must negotiate the role of English as a dominant language of international academic communication alongside the languages that constitute their objects of study and scholarly communities.

Recommendations and Good Practices

Use this worksheet to reflect on your own experiences, and type your responses directly into this worksheet. You can then download and save it to your computer for future use.

As you explore a potential journal for publication, do not focus only on the language in which full articles are written. Consider all the different spaces in which languages appear.

Step 1 of 2Use the following questions to guide your investigation
Use the following questions to guide your investigation

3.2 Reflexive Activity

Choose one research project (your own or a hypothetical one) involving more than one language. Imagine that you are preparing an article based on this research.

You should prioritise journals where abstract and keywords in a second language should be mandatory.

Concluding Remarks

The rich environment of Wikimedia and CLARIN tools and standards is central to support researchers in making their research dataset readable in several languages but also to ensure findability, accessibility, and reproducibility of multilingual data. 

Refining multilingual research processes in your DMP or using specific workflows are also very helpful to manage your corpora while staying compliant with FAIR and Open Science principles.

These recommendations and best practices contribute to promoting inclusivity, local community engagement, and reduced English dominance in scholarly communication.

By integrating multilingual considerations into research data management from the beginning, researchers can contribute to FAIRer research practices while supporting greater linguistic diversity, more equitable access to knowledge and broader opportunities for meaningful data reuse.

Organisations Involved in this Resource 

  • DARIAH-EU 
  • The University of Edinburgh
  • Spanish Researchers in the United Kingdom (SRUK/CERU)
  • Fundación Española para la Ciencia y la Tecnología (FECYT)
  • The Etxepare Basque Institute
  • King’s College London
Bibliography

Colin-Arce, A.,  Amell, B.,  (2025), “Multilingualism in Open Access Scholarly Publishing”, HSSCommons, https://doi.org/10.25547/3GS3-6R44

Curry MJ, Lillis T. Multilingualism in academic writing for publication: Putting English in its place. Language Teaching. (2024); 57(1): 87-100. https://doi.org/10.1017/S0261444822000040 

Erzsébet Tóth-Czifra, Marta Błaszczyńska, Francesco Gelati, Femmy Admiraal, Mirjam Blümm, Erik Buelinckx, Vera Chiquet, Rita Gautschy, Peter Gietz, Péter Király, Maria Vivas-Romero, Walter Scholger, Bartłomiej Szleszyński, & Ulrike Wuttke. (2023). Research Data Management for Arts and Humanities: Integrating Voices of the Community. Zenodo. https://doi.org/10.5281/zenodo.8059626 

Grallert, T., Horvath, A., Spence, P., Battaner Moro, E., Boot, P., Candela, G., del Rio Riande, G., Khemka, A., le Rell, A., McGinn, E., Mende, J.-K., Miyagawa, S., Müller-Laackman, J., Nagasaki, K., Sanz, A., Sigurðarson, E. S., Tekgürler, M., Tharsen, J., Viglianti, R., … Wrisley, D. J. (2026). Manifesto for Multilingual Digital Humanities. Zenodo. https://doi.org/10.5281/zenodo.21431692 

Kátia Monteiro, Eliana Hirano. (2026). Balancing the North and the South: A Brazilian perspective on multilingual publishing, Journal of English for Academic Purposes ; Volume 81, https://doi.org/10.1016/j.jeap.2026.101678

Kulczycki E, Guns R, Pölönen J, et al. (2020). Multilingual publishing in the social sciences and humanities: A seven-country European study. J. Assoc. Inf. Sci. Technol. ; Volume 71: 1371–1385,  https://doi.org/10.1002/asi.24336  

Lindemann, D., Candela, G., Marchetti, A., Pellizzari di San Girolamo, C. C., Olea, I., Varvantakis, C., Assis, T., Moitinho de Almeida, V., Obregón Sierra, Á., Salgado, A., Santiago Faria, A., Schöch, C., & Trigo, L. (2026). The Wikibase Ecosystem in DH and GLAM. Transformations, Volume 2 “The Past” (Research paper), https://doi.org/10.46298/transformations.17200

Lorella Viola, Paul Spence (Eds). (2024). Multilingual Digital Humanities, Routledge & CRC Press,  246 p. ISBN 9781032494173

Olea, I. (2026, March 24). Very Small GLAM. Wikimedia Meta-Wiki. https://meta.wikimedia.org/wiki/Very_Small_GLAM 

Pölönen, J. (Ed.), Garnier-Rizet, M., Greco, G. M., Kulczycki, E., Brasil Varandas Pinto, A. L., Conceição, M. C., Cots, J.-M., Dantas, L., Dendrinos, B., García-Sánchez, S., Laihonen, P., Soler Montes, C., Ochsner, M., Peruginelli, G., Rafols, I., & Urbanc, M. (2025). Implementation Proposal for Language-Aware Assessments. CoARA WG on Multilingualism and Language Biases in Research Assessment. https://doi.org/10.5281/zenodo.17881907 

Domain
Social Sciences and Humanities
Language
English
Published to DARIAH-Campus
19/11/2026
License
CC BY 4.0
Sources
DARIAH

Cite as

Françoise Gouzi, David Lindemann, Carlos Soler Montes, Luis Trigo and Sara Franco Ortega (2026). Multilinguality from Research Data Management to Publishing. Version 1.0.0. DARIAH Campus [Training module]. https://hdl.handle.net/21.11159/01a1f375-a9f2-7568-ad9a-54bbfbf9d140

Reuse conditions

Resources hosted on DARIAH-Campus are subjects to the DARIAH-Campus Training Materials Reuse Charter.