Dr Nilo Pedrazzini, BA (Pavia), MSt DPhil (Oxford)

Profile

I am a Postdoctoral Research Associate within the Schmidt Sciences-funded project Text Machine: Computing Literary Innovation.  Before joining QMUL, I was a Turing Research Fellow at The Alan Turing Institute (2023-2026), a Guest Researcher in linguistics at the University of Oslo (2025-2026), and a Research Associate in corpus-based digital humanities (2022-2023) at the Turing Institute within the Living with Machines project. 

I obtained my BA in Russian and English in Pavia (Italy), where I also received my linguistics training, before pursuing an MSt in Slavonic Studies (2018) and a DPhil in Linguistics (2023) at the University of Oxford.  

Research

Research Interests:

My main research areas are computational linguistics, historical linguistics and language typology. I am particularly interested in the study of systematic cross-linguistic variation in the grammatical encoding of concepts and pragmatic functions. Some of the topics I work on are: 

• Computational modelling and statistical analysis of historical and low-resource languages 

• Subordination and clause-linkage 

• Formal-semantic representation of discourse structures 

• Linguistic change at the syntax-pragmatics interface 

Recent and on-going projects:

At QMUL, I adapt language models to different domains and historical periods, using semantic space modelling and mechanistic interpretability tools to detect shifts in contextual representations and study how adapted models encode meaning across time, genre, and text sources. A central aim is to exploit usages that are atypical under one period- or domain-specific model but not another, as a way of identifying items relevant to broader questions of literary or social change. One strand looks at changing representations of suicide motives in Victorian newspapers, building on work begun during my time at the Turing Institute. 

Besides my research at QMUL, I have worked on the study of cross-linguistic variation in the grammatical encoding of concepts and pragmatic functions, with a focus on underrepresented and historical varieties, including early Slavic  and Latin American and Caribbean varieties. As part of an ongoing collaborative project at the University of Oslo (where I am a Guest Researcher at the linguistics department), I have been investigating the extent to which multilingual large language models encode systematic structural variation among languages and have been using massively parallel multilingual corpora as a source for typological analyses. 

Besides my research in computational humanities and linguistics, I have recently spent a year as part of the Turing Institute’s Research Engineering Team, where I worked on the development of pipelines and models in environmental data science (e.g. Quartz Solar AI Nowcasting  and IceNet), and have collaborated on several projects around open humanities data, and digital sustainability and data reproducibility in the humanities, arts, and social sciences. 

Publications

A full list of publications can be found on my webpage:  https://npedrazzini.github.io/ 

Selected articles and book chapters:

Pedrazzini, Nilo. Forthcoming. Early Slavic when-clauses in cross-linguistic perspective. In Łukasz Jędrzejowski (ed.), On the diachrony of adverbial clauses (Brill's Studies in Historical Linguistics). Leiden & Boston: Brill. 

McGillivray, Barbara, Nilo Pedrazzini, Arianna Ciula, Jon Lawrence, Tiffany Ong, Mia Ridge & Miguel Vieira. 2025. Analysing the language of mechanisation in 19th-century British newspapers. In Ruth Ahnert, Emma Griffin, Jon Lawrence & the Living with Machines Team (eds.), Living with machines: Computational histories of the age of industry. London: University of London Press. 

Pedrazzini, Nilo. 2024. Mapping 'when'-clauses in Latin American and Caribbean languages: An experiment in subtoken-based typology. In Manuel Mager, Abteen Ebrahimi, Shruti Rijhwani, Arturo Oncevay, Luis Chiruzzo, Robert Pugh & Katharina von der Wense (eds.), Proceedings of the 4th Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP 2024), 24–33. Mexico City: Association for Computational Linguistics. Available here: https://doi.org/10.18653/v1/2024.americasnlp-1.4  

Stopponi, Silvia, Nilo Pedrazzini, Saskia Peels-Matthey, Barbara McGillivray & Malvina Nissim. 2024. Natural language processing for Ancient Greek: Design, advantages and challenges of language models. Diachronica 41(3). 414–435.

Haug, Dag & Nilo Pedrazzini. 2023. The semantic map of when and its typological parallels. Frontiers in Communication 8. 

Stopponi, Silvia, Nilo Pedrazzini, Saskia Peels, Barbara McGillivray & Malvina Nissim. 2023. Evaluation of distributional semantic models of Ancient Greek: Preliminary results and a road map for future work. Proceedings of the Ancient Language Processing Workshop, associated with RANLP-2023, 49–58. 

McGillivray, Barbara, Paola Marongiu, Nilo Pedrazzini, Mandy Wigdorowitz & Eleonora Zordan. 2022. Deep impact: A study on the impact of data papers and datasets in the humanities and social sciences. Publications 10(4). 39.

Pedrazzini, Nilo. 2022. One question, different annotation depths: A case study in Early Slavic. Journal of Historical Syntax 6(7) (Special issue: Annotating historical corpora). 1–40. Available here: https://doi.org/10.18148/hs/2022.v6i4-11.96 

Pedrazzini, Nilo & Barbara McGillivray. 2022. Machines in the media: Semantic change in the lexical field of mechanization in 19th-century British newspapers. Proceedings of the 2nd International Workshop on Natural Language Processing for Digital Humanities (NLP4DH), 85–95. Association for Computational Linguistics.

Pedrazzini, Nilo & Hanne Martine Eckhoff. 2021. OldSlavNet: A scalable Early Slavic dependency parser trained on modern language data. Software Impacts. 100063. Available here: https://doi.org/10.1016/j.simpa.2021.100063  

Public Engagement

Blog Posts:

Fairytales on trial: The good and the beautiful in early-Soviet children's literature (Mar. 2019). British Library, European Studies Blog. Available here: https://bl.iro.bl.uk/concern/generic_works/f166c876-b4b9-4579-b6c8-14542da981b7 

A reluctantly modern voice from the 17th-century Russian storm: Archpriest Avvakum and the Life written by himself (June 2015). British Library, European Studies Blog. Available here: https://bl.iro.bl.uk/concern/generic_works/c191b34a-db60-4dc7-b943-f227eb3d6418 

Media:

Living with Machines: On Collaboration. Docuseries episode. 2023.

Living with Machines: The Language of Mechanisation. Docuseries episode. 2023.

Crowdsourcing:

Living with Machines crowdsourcing projects on lexical semantic change in 19th-century English. Zooniverse. Data from this project fed into Living with Machines: Human Stories from the Industrial Age (2022), an exhibition at Leeds City Museum (curated by Mia Ridge and John McGoldrick).