Language Culture Matters Seminar Series

Convenors: Holly Warner, Aishee Bhattacharya, Dr Clay Becker

When: Every Wednesday, 16:00-17:00 BST/GMT, Weeks 2-9 of Term 1 (15th October 2025 - 3rd December 2025)

Where: A0.23 (Social Sciences Building, University of Warwick) and Microsoft Teams (linked below)

Please note: Seminar on 3rd December 2025 is held in room A1.11 (Social Sciences Building, University of Warwick)

Wednesday 15/10/2025 - 16:00-17:00 BST

Contextualizing High-Dimensional Communication: The Relevance of Linguistic Anthropology for Theorizing Large Language Models

Dr Michael Castelle, University of Warwick

Abstract

The last decade's unfolding scientific and intellectual "revolution" wrought by the use of multilayer artificial neural-network architectures in the field of Natural Language Processing — up to and including large language models (LLMs) – is a result of, in part, their implicit rejection of influential theories in mainstream linguistics and cognitive science. Indeed, many computer scientists in contemporary NLP do not find it necessary to concern themselves with the wide variety of past or present theories of language and learning. I will argue that a better familiarity with the field known as linguistic anthropology can aid scholars from various fields in understanding the current and future successes and/or failure modes of (increasingly dialogical and multimodal) LLMs, as well as helping social scientists and humanists to avoid some common, but misguided, avenues of critique for LLMs. From century-old works of American anthropological linguistics and semiotics; to Michael Silverstein's theories of pragmatics and metapragmatics; to Bambi Schieffelin and Kathryn Woolard's concept of linguistic ideologies, this resolutely empirical and ethnographically well-grounded school of thought provides surprising insights into both the intriguing strengths and fundamental limits of these computational artifacts.

Bio

Michael Castelle is Associate Professor at the Centre of Interdisciplinary Methodologies at the University of Warwick. His research is at the intersection of science and technology studies, history of computing, and linguistic anthropology. He is interested in the use of sociological, anthropological, historical, and semiotic perspectives in recontextualizing and understanding contemporary technological practices, from databases and distributed systems to machine learning and artificial intelligence.

X - https://x.com/mccastelle

Bluesky - https://bsky.app/profile/mcastelle.bsky.social

LinkedIn - https://www.linkedin.com/in/mcastelle/

Wednesday 22/10/2025 - 16:00-17:00 BST

More Digital Technologies, More Mobilization? Mapping 30 Years of Trajectories via the “Technology in Movement” (TiM) Dataset

Dr Jun Liu, University of Copenhagen

Abstract

Does the rise of digital technologies lead to more political mobilization? Are protests today simply more connected, more visible, and more digital? Are protests confined to familiar platforms such as Twitter/X and Facebook—or do technologies extend far beyond these platforms? Although many studies examine individual movements and their use of digital tools, we lack a comprehensive understanding of how digital technologies have transformed political mobilization over time. This project (https://tech-in-movement.ku.dk/) introduces the Technology in Movement (TiM) Dataset, the first dataset to map the intersection of political mobilization and digital technologies over three decades (1990–2020). Built on the Mass Mobilization Data Project, TiM goes beyond case-based studies of particular movements to trace how protest repertoires have shifted with the spread of digital tools.

To answer these questions, we advance a self-training model that leverages Large Language Models (LLMs) to identify, assess, and extract information on the adoption of digital technologies as protest repertoires in news coverage. By positioning digital technology adoption as a key subdimension within Protest Event Analysis (PEA), we develop a semi-automated model capable of extracting contextual semantic elements to systematically capture technology use in protest events. To improve performance and adaptability, we integrate LLMs with syntactic Graph Convolutional Networks (GCNs).

Empirically, the TiM Dataset provides the first large-scale benchmark of technology-mediated protests, where technology functions as a medium or tool of mobilization. By mapping these trajectories, the project sheds light on how digital technologies have reshaped the forms, scale, and dynamics of political mobilization worldwide over the past 30 years. Methodologically, we advance a self-training model that combines LLMs with syntactic GCNs. This semi-automated framework identifies and classifies the adoption of digital technologies as protest repertoires in news coverage, reducing reliance on labor-intensive coding while capturing contextual depth. Together, these innovations allow us to explore how digital technology has shaped, mediated, and expanded political mobilization worldwide over the past 30 years.

Bio

Wednesday 05/11/2025 - 16:00-17:00 GMT

Murmurations of conspiracy: Identifying dynamic belief flows in social systems

Prof Thomas Hills, University of Warwick

Abstract

Beliefs in social communities may follow complex flows. Our beliefs influence our social communities and these in turn influence what we will believe next. Are these flows predictable and can we identify them? In this talk I describe a network and complex systems approach to identifying gateway beliefs among conspiracy theorists. This work is based on a data set from 1000 people in the US self-reporting on 39 conspiracies. The results are suggestive of complex flows through time and multiple belief attractors, with conspiracies promoting further conspiracy belief and other actually preventing further belief. This method is generalizable and could be applied to any system where conceptual information can change over time and is influenced by the social environment.

Bio

Thomas Hills is a Professor of Psychology at the University of Warwick. Trained as a biologist, his research focuses on applying principles from behavioural ecology to the human behavioural sciences. His work involves methods such as network science, natural language processing, and computational modeling, with a strong emphasis on cognitive theory. He has worked at the University of Basel, Indiana University, and University of Texas at Austin, following his PhD from the University of Utah.

Wednesday 12/11/2025 - 16:00-17:00 GMT

Do you sound softer than you think? The social life of voice quality

Xinran Gao, University of Warwick

Abstract

This talk explores the intricate relationship between voice quality and social meaning. In a narrow sense, voice quality encompasses features such as pitch and phonation and plays a crucial role in how individuals are perceived in social contexts. Drawing on recent research in phonetics and sociolinguistics, I will discuss how variations in voice quality can influence judgments about a speaker's personal traits and emotional state. The talk will highlight the impact of breathy and creaky voice quality on social interactions in different languages, including varieties of English and Chinese languages, leading to linguistically and culturally specific interpretations of different voice qualities. By understanding the social life of voice quality, we can gain deeper insights into human communication and the subtle ways in which our voices shape our social experiences and reveal about who we are.

Bio

Xinran Gao is a PhD candidate in Applied Linguistics at the University of Warwick. She received her MA from University College London. Her research focuses on phonetics and sociolinguistics, particularly how voice quality influences social interactions and reflects identities. In addition, she also collaborates with colleagues to explore lifespan sound change of individuals and the voice and life experiences of female academics. Aside from her research, she has extensive interdisciplinary GTA teaching experiences across three departments/centres at the University of Warwick, demonstrating her versatility and commitment to higher education.

LinkedIn - https://www.linkedin.com/in/xinrangaoxg/

Link opens in a new win

Wednesday 19/11/2025 - 16:00-17:00 GMT

‘I always feel guilty when I’m not prioritising one thing’: Co-constructed narratives of student-parenting and the wellbeing of academic mothers

FECARE, University of Warwick

Abstract

This talk draws on our forthcoming book chapter examining how PhD student-mothers narrate and negotiate the intersecting demands of caregiving and academic work. Adopting De Fina’s (2021) narratives-as-practices framework, we analyse focus-group storytelling as a form of collective meaning-making, where emotions such as guilt and inadequacy are not merely expressed but interactively constructed, shared, and sometimes resisted. Our analysis shows how participants mobilise and reframe dominant discourses of the ‘good mother’ and the ‘successful academic’ to make sense of conflicting temporalities, care responsibilities, and institutional expectations. We argue that these narratives illuminate the moral and affective labour underpinning academic life and reveal how peer interaction can transform individual struggles into collective reflection. Ultimately, the study reimagines wellbeing as relational and discursive, calling for research cultures that legitimise diverse academic trajectories and integrate care into the fabric of higher education.

Bio

Cagla Karatepe, Xinran Gao, Dr Mingzhi Li and Dr Yanyan Li are researchers in Applied Linguistics at the University of Warwick. Together, they lead the FECARE project (Towards More Inclusive Practices: Female Postgraduate Researchers and Early Career Researchers with Caring Responsibilities), which explores how academic mothers navigate care, identity, and wellbeing within higher education. Cagla’s research focuses on impoliteness, humour, and socioprosodics; Mingzhi’s on gender, language, and social inequality; Xinran’s on phonetics, sociolinguistics, and language variation; and Yanyan’s on multimodal communication and conversation analysis. Their collaborative work is united by a shared commitment to understanding how language constructs social identities and to promoting more inclusive and compassionate research cultures across academia.

Website - https://warwick.ac.uk/research/research-culture-at-warwick/fecare/

Wednesday 26/11/2025 - 16:00-17:00 GMT

Celebrating National Linguistics Day – Three Cheers for Multilingualism: Language learning, public health and a decolonial critique (with a focus on African languages)

Prof Lutz Marten, SOAS, University of London

Abstract

To celebrate National Linguistics DayLink opens in a new window this year, in this talk I will offer three case studies – or vignettes – of recent research revolving around multilingual practices from different domains, all of which involving African languages. The first looks at learners of the South African language Zulu in the UK, and in particular at their linguistic repertoires – which is very rich, and students typically speak two or three other languages. This high level of multilingualism has positive effects in students’ motivation and learning progress (Marten and Mostert 2012, Marten 2021, Marten and Zhang forthcoming). The second case study looks at how London’s multilingual communities used their access to different discourses for interpretating and understanding the Covid-19 pandemic, showing that multilingual practices are key to understanding (and future responses to) public health crises (Marten and Sato-Rossberg). Finally, I will adopt a more theoretical perspective and show how the model of Ubuntu Translanguaging (Makelela 2016, Marten forthcoming) can be harnessed to study multilingual realities and dynamics.

Makalela, Leketi. 2016. Ubuntu translanguaging: an alternative framework for complex multilingual encounters. Southern African Linguistics and Applied Language Studies 3(3). 187-196.

Marten, Lutz. 2021. From the Margins: Appreciating the African language renaissance as learners of African languages. African Language and Culture Studies 1: 41-62.

Marten, Lutz. Forthcoming. Historical Linguistics and Ubuntu Translanguaging: Towards a model of multilingualism, language change and linguistic convergence in the Bantu linguistic area. In Jadranka Gvozdanović (ed.) Historical Linguistics 2023. Selected papers from the 26th ICHL, Heidelberg, 4-8 September 2023. Amsterdam: Benjamins.

Marten, Lutz and Carola Mostert. 2012. Background languages, learner motivation and self-assessed progress in learning Zulu as an additional language in the UK. International Journal of Multilingualism 9: 101-128.

Marten, Lutz and Nana Sato-Rossberg. 2026. Languages, Cultures, and Health in a Global City: Translating and communicating Covid-19 among London’s multilingual communities. London/New York: Routledge.

Marten, Lutz and Chungai Zhang. Forthcoming. African languages and area studies in higher education in China and the UK: Contexts, pedagogies and perspectives. Journal of the UK-China Humanities Alliance 3

Bio

Lutz Marten is Professor of General and African Linguistics at the School of Oriental and African Studies (SOAS), University of London. He is interested in how language is structured and used, how languages differ and change over time, and how language is linked to culture, society, history, nature and other domains of human life. Most of his work is related to African languages and he has worked with researchers and communities across the continent for many years. His recent and current research projects include Morphosyntactic Variation in Bantu, Cultural Translation of Covid-19 in London’s Community Languages, the Description and Promotion of the Kenyan Bantu Language Kitaveta, and Variation in Swahili.

LinkedIn - https://www.linkedin.com/in/lutz-marten/

Wednesday 03/12/2025 - 16:00-17:00 GMT

Global Standards, Local Realities: Aligning the CEFR with Arabic through Hamza Test

Dr Ebtesam Abdulhaleem, King Salman Global Academy for Arabic Language

Abstract

The Common European Framework of Reference for Languages (CEFR) has become the global reference point for language assessment, yet aligning it with Arabic reveals both challenges and opportunities. This presentation focuses on Hamza Test, an Arabic proficiency assessment for non-native speakers, designed to capture the language’s distinct linguistic and sociolinguistic features. The alignment process raises important issues: adapting CEFR descriptors to Arabic’s diglossic ecology, rich morphology, and script-specific demands; and ensuring accurate translation of the CEFR into Arabic, where semantic equivalence, terminological consistency, and cultural nuance remain pressing challenges. Beyond technical adaptation, these efforts also required training Arabic experts to apply the framework meaningfully. At the same time, the CEFR opens pathways for transparency, international recognition, and comparability of language proficiency worldwide. By reflecting on these dynamics, Hamza Test demonstrates how global frameworks can be localized without compromising validity, offering a model for context-sensitive, reliable, and internationally trusted Arabic assessment.

Bio

Dr. Ebtesam Abdulhaleem is an applied linguist specializing in language testing and assessment. She holds a PhD from the University of Warwick and a postgraduate degree in Language Testing from Lancaster University. Her research spans language testing and assessment, corpus linguistics, academic writing, and teacher professional development.

Dr. Abdulhaleem has served in several academic leadership roles, including Head of the Scientific Research Unit and Vice Chair of the English Department at King Saud University. She led the Teachers as Researchers initiative, which fostered reflective practice and evidence-based inquiry among educators, leading to sustainable professional growth.

Currently, she is the Head of the Test Development Unit at King Salman Global Academy for Arabic Language, where she directs the design, development and validation of Arabic language proficiency assessments and other Arabic tests. Her work contributes to advancing fair and innovative assessment practices and strengthening global dialogue on language education and testing.

X - @ebtisamhaleem

LinkedIn - https://www.linkedin.com/in/ebtesam-abdulhaleem-phd-78877a35/

See the recordings of our previous seminars here (please note that not all seminars are recorded):

Language Culture Matters Seminar RecordingsLink opens in a new window

See the list of our previous seminars here:

Language Culture Matters Seminar ArchiveLink opens in a new window