Social and cultural anthropology at UPV/EHU

Some universities are remembered for their buildings, their rankings or their most visible figures. Others leave a less immediate, but perhaps more decisive, mark: they make it possible for certain questions to find time, networks, libraries, teaching, fieldwork and continuity.

The social and cultural anthropology associated with the University of the Basque Country / Euskal Herriko Unibertsitatea is one such constellation. For decades, it has brought gender, bodies, cities, language, memory, art, health, heritage and everyday life into conversation. It is neither a homogeneous school nor a list of names. It is a situated intellectual infrastructure.

That is why it offers tools for a question that also matters to AIthropology Lab: what does it mean to digitise a minoritised language? A tradition that has understood language as body, ritual and community — rather than as a mere repository of texts — makes it possible to look differently at the arrival of Basque in language models.

Teresa del Valle and an infrastructure for asking questions

Teresa del Valle Murga (Donostia, 1937–2025) was Emerita Professor of Social Anthropology at UPV/EHU and one of the leading figures behind feminist anthropology in Spain. She carried out more than four decades of teaching and research at UPV/EHU and other universities, belonged to the founding group of the European Association of Social Anthropologists, and received, among other distinctions, the Emakunde Equality Award in 2010 and the Eusko Ikaskuntza Prize in 2018.1

EHU identifies feminist critique in social anthropology, urban and political anthropology, the ethnography of memory and qualitative methodologies among her main lines of work.1 This breadth is visible in an oeuvre that turned gender into a tool for observing how power emerges in urban routes, rites of passage, embodied memory, language, labour and collective participation.

In 1985, she edited the collective work Mujer vasca. Imagen y realidad, a landmark in this tradition. The research did not simply make women’s experiences visible. It challenged narratives that turned tradition into harmony and made it possible to study inequality as something reproduced through institutions, narratives, spaces and concrete relationships.2

Its institutional continuity mattered just as much. The UPV/EHU Women’s Studies Seminar, directed by Del Valle between 1981 and 1994, became a seedbed for feminist research and teaching in the Basque context.2 This matters because intellectual traditions are not sustained by books alone. They also depend on teams, supervision, library collections, courses and the material possibility that new generations may encounter a conversation already under way.

The collective volume Saberes y haceres antropológicos. El legado de Teresa del Valle, published in 2025, reads this inheritance as a feminist genealogy. It considers both Del Valle’s theoretical and methodological contributions and her collaborative practice and work as a teacher and mentor.3 This double dimension — work and transmission — helps explain why her influence remains visible in research on health, bodies, language, cities, festivals, care, memory and heritage.

Andamios para una nueva ciudad (1997) developed this approach to urban space. The city stops being a neutral setting and becomes an organisation of mobility, care, presence and power.4 Her work also made language an embodied and collective question. In Korrika. Rituales de una lengua en el espacio (1988), Basque appears as social practice, shared territory, movement and ritual. The 2025 volume on her legacy returns to this inheritance in a chapter entitled “Korrika, el aliento del cuerpo hablante”.3

Lourdes Méndez and feminist critique as theoretical architecture

Lourdes Méndez was Professor of Social Anthropology at UPV/EHU, associated with the Department of Philosophy of Values and Social Anthropology and the Faculty of Fine Arts in Bizkaia.5 Her trajectory brings together feminist anthropology, fieldwork, visual culture and the analysis of institutions that consecrate art.

In Antropología feminista (2008), Méndez reconstructs the theoretical and methodological foundations of this perspective through a critique of androcentrism. The issue is not to add women to an otherwise intact discipline, but to examine the categories through which sex, gender, kinship, reproduction, labour and power are defined.6

Her later work on the artistic field brought the same critique to systems of cultural recognition. Asking who is recognised as an artist, how value is assigned to a work, or why certain bodies appear as universal allows us to see art as a terrain where prestige, authority and the right to produce meaning are distributed.7

These questions are directly relevant to artificial intelligence. Systems for classification, recommendation and image generation do not merely sort content: they also intervene in the distribution of visibility, cultural value and recognition.

Bodies, care and feminist continuities

Mari Luz Esteban Galarza, a medical graduate and Doctor of Social Anthropology, has taught at UPV/EHU since 1998 and has led the AFIT — Feminist Anthropology — research group since 2013. Her work addresses health and gender, body and image, romantic love and gender relations; among her books are Antropología del cuerpo and Crítica del pensamiento amoroso.8

Her work helps to undo a separation that remains common in technological narratives: the idea that data do not come from bodies, affects, pain, care time and social relations. Thinking about technologies from this perspective means asking which experiences become visible, which forms of distress are minimised and who bears the costs of innovation.

Carmen Díez Mintegui extended this line towards sport, socialisation, labour, festivals and equality. Her work shows that feminist anthropology is not limited to abstract critique: it can analyse the rules, timetables, spaces and expectations through which participation and prestige are distributed.9

Miren Elixabete Imaz represents another necessary continuity. Her doctoral thesis, defended at UPV/EHU in 2008, addressed representations, models and experiences of pregnancy and motherhood among contemporary Basque women.10 In the present of digital health, assisted reproduction and biomedical data, this line reminds us that motherhood cannot be reduced to a clinical category or a predictive variable.

Languages, memory and borders

Jone Miren Hernández García is an Associate Professor of Social Anthropology at UPV/EHU and a member of AFIT. Her 2005 doctoral thesis, supervised by Teresa del Valle, was entitled Euskara, comunidad e identidad. Elementos de transmisión, elementos de transgresión.11 Her work makes it possible to understand language as practice, transmission, conflict and memory.

Together with Jaime Altuna, she has developed the image of the “speaking body” to read Korrika as a situated practice of language, affects, movement and community.12 Her recent publications show that this line remains active. In 2025, together with Joan Pujolar and Noelia Hernández, she published “Agency as doing-together: learning from fieldwork experiences” in an issue devoted to sociolinguistic justice.12

Aitzpea Leizaola adds a decisive vector to this constellation. Her research brings together social and cultural anthropology, border studies, memory, identity, food and linguistic landscape. Her work on the memory of the Spanish Civil War, the Basque cross-border space and signage between Donostia and Baiona opens a crucial question for the digital present: how the borders of a language are constructed, administered and contested.13

Other lines of continuity

Margaret “Maggie” Bullen has researched migration, identities, language and change, as well as gender, festivals and rituals. Her work on conflicts surrounding women’s participation in the Alardes of Hondarribia and Irun shows how traditions are reformulated and politicised through social change.14

Olatz González Abrisketa brings together anthropology of sport, feminism, masculinities, ethnographic film and human–animal relations. Her research on Basque pelota and professional women pelota players expands this perspective towards bodies, ritual and agencies beyond the human. In recent years, she has also contributed to work in more-than-human anthropology.15

Iñaki Arrieta Urtizberea works on cultural policy, heritage and museums. His research reminds us that museums are not neutral repositories of the past, but devices that select what deserves to be preserved, exhibited, funded and remembered.16

The boundaries of this tradition extend beyond this selection. The work of José Ignacio Homobono, from sociology and in sustained dialogue with anthropology, has investigated the city, festivals, sociability, heritage and Basque identities for decades.17 The point is not to turn this radar into a census, but to recognise that UPV/EHU has sustained a plural conversation between feminism, urban life, language, ritual, memory and material culture.

An inheritance for reading artificial intelligence

Let us return to the opening question. The contemporary issue is not only whether Basque has a presence in a digital ecosystem where English remains dominant. It also concerns the conditions under which its data are collected, who decides their uses and which forms of community remain outside computational representation.

Latxa, an open family of language models for Basque, shows that it is possible to broaden the presence of a minoritised language in language technologies. Its models range from 7 to 70 billion parameters and are based on Llama 2, continued on a corpus of 4.3 million documents and 4.2 billion Basque tokens. The work received the ACL 2024 Best Resource Paper Award. In the published evaluations, Latxa substantially outperforms the earlier open models included in the comparison and is competitive with GPT-4 Turbo in linguistic competence and understanding, while remaining behind in reading comprehension and knowledge-intensive tasks.18

The developments of 2025 shift the discussion from the textual base model towards multimodal and educational uses. A November paper used an adapted Latxa variant to train and evaluate image-and-text models in Basque; its authors conclude that a reduced proportion — around 20% — of multimodal Basque data can be enough to achieve solid results in their evaluations.19 In December 2025, a paper accepted for LREC 2026 introduced a public set of 3,200 C1-level essays from HABE and adapted Latxa 8B and 70B for scoring and feedback. These results should be read as work in progress, but they point to a meaningful shift towards transparent and evaluable educational tools.20

Euskorpus places this discussion on an institutional scale. On 22 July 2025, the Basque Government approved a direct grant of up to five million euros for the Euskorpora association, within an announced total funding of 10.55 million euros for 2025–2027. The plan includes three phases: definition and planning, corpus collection and the development of open models, and transfer to companies, institutions and the public.21

Technological presence, however, does not exhaust the anthropological question. A language is not only the set of texts that can be assembled into a corpus. It is also a situated practice, a shared memory and a repertoire of affects, rhythms, bodies and relationships. The embodied orality of bertsolaritza and the collective ritual of Korrika are dimensions that a textual corpus can scarcely capture in full.

The question, then, is not only whether Basque enters language models. It is who takes part in building them, which practices are represented, which uses are prioritised and what capacity for decision-making its speaking communities retain.

The other works in this constellation sharpen the view in other areas. Esteban’s critique of romantic thought offers a framework for reading affective chatbots without falling into either enthusiasm or panic. Méndez helps us recognise the hierarchies of value that image-generation systems may reproduce. González Abrisketa and Arrieta remind us that every classification of bodies, species or memories is also a cultural politics.

To honour this tradition is not to place it in a display case. It is to use its tools again in order to read a present in which images, data, languages, bodies and memories are being rapidly reorganised, while artificial intelligence redistributes — never neutrally — visibility, authority and the capacity to decide.

Sources and bibliography

Footnotes

  1. EHU, “Fallece la antropóloga Teresa del Valle”, 8 April 2025. ↩ ↩2

  2. Jone M. Hernández, Margaret Bullen, Carmen Díez, Mari Luz Esteban and Miren Urquijo, “In memoriam: Teresa del Valle, antropóloga precursora y maestra”, EHU, 11 April 2025. ↩ ↩2

  3. Jone M. Hernández García et al., Saberes y haceres antropológicos. El legado de Teresa del Valle, Los Libros de la Catarata, 2025. ↩ ↩2

  4. Teresa del Valle, Andamios para una nueva ciudad: lecturas desde la antropología, record in Dialnet, 1997. ↩

  5. Lourdes Méndez Pérez, EHU scientific output profile. ↩

  6. Méndez, Lourdes. 2008. Antropología feminista. Madrid: Síntesis. ↩

  7. Méndez, Lourdes. 2009. Antropología del campo artístico. Del arte primitivo al contemporáneo. Madrid: Síntesis. ↩

  8. Mari Luz Esteban Galarza, EHU profile. ↩

  9. María Carmen Díez Mintegui, Dialnet profile and publications. ↩

  10. Department of Philosophy of Values and Social Anthropology, EHU doctoral theses. ↩

  11. Jone Miren Hernández García, EHU scientific output profile. ↩

  12. Jone Miren Hernández García, Dialnet profile and publications. ↩ ↩2

  13. Aitzpea Leizaola, EHU profile and ResearchGate research profile. ↩

  14. Margaret Louise Bullen, EHU scientific output and Dialnet profile. ↩

  15. Olatz González Abrisketa, EHU profile. ↩

  16. Iñaki Arrieta Urtizberea, EHU profile. ↩

  17. José Ignacio Homobono Martínez, Dialnet profile and publications. ↩

  18. Julen Etxaniz et al., “Latxa: An Open Language Model and Evaluation Suite for Basque”, Proceedings of ACL 2024, pp. 14952–14972. ↩

  19. Lukas Arana et al., “Multimodal Large Language Models for Low-Resource Languages: A Case Study for Basque”, arXiv, 2025. ↩

  20. Ekhi Azurmendi, Xabier Arregi and Oier Lopez de Lacalle, “Automatic Essay Scoring and Feedback Generation in Basque Language Learning”, accepted to LREC 2026; preprint available on arXiv:2512.08713, December 2025. ↩

  21. Basque Government, “El Gobierno Vasco destina 5 millones de euros para impulsar el desarrollo de la inteligencia artificial en euskera mediante el proyecto Euskorpus”, 22 July 2025. ↩