{"id":92,"date":"2022-11-16T10:08:10","date_gmt":"2022-11-16T10:08:10","guid":{"rendered":"https:\/\/sites.edgehill.ac.uk\/corpusresearchgroup\/?page_id=92"},"modified":"2026-03-06T14:51:58","modified_gmt":"2026-03-06T14:51:58","slug":"past","status":"publish","type":"page","link":"https:\/\/sites.edgehill.ac.uk\/crg\/past\/","title":{"rendered":"PAST EVENTS"},"content":{"rendered":"\n
\n

<\/p>\n\n\n\n

\n\n\n\n

2026<\/strong><\/h2>\n<\/div>\n\n\n\n
\n\n\n\n
\n
MEETING #20: Friday 6 March 2026, 10:00-11:30 am<\/span><\/h6>\n\n\n\n
Topic: LLMs, Corpus Linguistics, and Language Learning<\/em><\/h6>\n\n\n\n
Peter Crosthwaite<\/strong><\/a> (University of Queensland, Australia)<\/h5>\n\n\n\n
Corpora, Prompts, and Pedagogy: Human-AI Text Comparison in Applied Linguistics<\/strong><\/em><\/h5>\n\n\n\n

SLIDES<\/a><\/strong><\/p>\n\n\n\n

Generative AI has fundamentally disrupted corpus linguistics by making it possible to create large, readily generable corpora of machine-produced text, challenging long-standing assumptions about what constitutes a \u201cnatural\u201d language dataset. In response, a growing body of corpus-based research has begun to examine the linguistic characteristics of AI-generated texts, often comparing them with human writing to identify similarities, differences, and potential risks for language education and assessment.<\/p>\n\n\n\n

Drawing on my own corpus-based analyses, this talk first revisits findings showing that AI-generated academic texts do not reliably approximate human-like stance, particularly in how evaluation, commitment, and authorial presence are expressed. These results complicate claims that AI writing is simply more \u201cexpert-like\u201d or rhetorically mature than student writing. However, more recent work extends this line of inquiry by demonstrating that many of the features attributed to AI texts are strongly conditioned by the prompts used to generate them, including task descriptions, genre cues, and implicit assessment criteria.<\/p>\n\n\n\n

Taken together, these studies suggest that differences between human and AI writing are neither fixed nor intrinsic, but emerge from the interaction between prompts, tasks, and evaluative expectations. The talk concludes by considering whether this insight opens productive possibilities for language teaching, asking whether corpus-driven data-driven learning approaches can be meaningfully combined with AI-mediated tools to support learner noticing, agency, and writing development, rather than positioning corpora and AI as competing paradigms.<\/p>\n\n\n\n

\n\n\n\n
MEETING #19: Friday 6 February 2026, 2:00-3:30 pm<\/h6>\n\n\n\n
Topic: Discourse Oriented Corpus Studies<\/em><\/h6>\n\n\n\n
Dan Malone<\/strong><\/a> (Edge Hill University, UK)<\/h5>\n\n\n\n
From Global Uncertainty to Domestic Danger: The lone wolf terrorist as a topos of threat in (poly)crisis discourses<\/strong><\/em><\/h5>\n\n\n\n

Crises arise amidst uncertainty and are characterised, alongside urgency, by a sense of threat (Lipscy 2020). Few figures embody uncertainty more vividly than the lone-wolf terrorist. Acting in isolation and without formal ties to organised groups, the lone-wolf terrorist has become increasingly prominent in public discourse over the past 15 years (Malone, 2025).<\/p>\n\n\n\n

In this talk, I explore how the lone-wolf terrorist emerges in the construction of polycrises in the UK press. Polycrises are understood here, following Krzy\u017canowski et al. (2023: 423), as the \u201ccombination of many, more or less simultaneous and overlapping, crises whose repercussions unfold in a cumulative manner\u201d. Their discursive construction thus relies on how crises are presented as interacting, exacerbating, and reshaping one another (cf. Janzwood & Homer-Dixon, 2022: 4). Specifically, I focus on how the lone-wolf terrorist is rhetorically employed as a topos<\/em> of threat, operating under the premise that if there is a risk or danger, action must be taken to prevent it (Wodak 2001; Boukala, 2016).<\/p>\n\n\n\n

I draw on analyses from the Lone Wolf Corpus (Malone, 2026), a topic-specific corpus of UK newspaper articles featuring the lone-wolf terrorist, focusing on the years 2020 to 2024. The analytical approach employs three interrelated stages to identify collocations and semantic preferences of the lemma lone wolf<\/em>, as well as discourse prosodies, that is, implicit and explicit attitudes (Stubbs, 2001) towards the lone-wolf terrorist. In the discussion, I pay particular attention to the role of metaphor clusters, viewing metaphor as a device for expressing evaluation implicitly (Martin & White, 2005).<\/p>\n\n\n\n

Findings show that the lone-wolf terrorist is a recurring evaluative resource through which crisis events are connected and presented as threatening. From Iran\u2019s nuclear ambitions, Hamas\u2019s activities, and Russia\u2019s aggression in Ukraine, to mass migration, the climate crisis, and fears of radicalisation fuelled by AI and intensified by COVID-19 lockdowns, the lone-wolf terrorist emerges to recontextualise international events as matters of domestic security, transforming global uncertainty into a sense of danger for UK audiences. In this way, otherwise discrete events are discursively construed and rendered as potential polycrises through imagined terroristic violence.<\/p>\n\n\n\n

References<\/strong><\/p>\n\n\n\n

Boukala, S. (2016). Rethinking topos in the discourse-historical approach: Endoxon seeking and argumentation in Greek media discourses on \u201cIslamist terrorism\u201d. Discourse Studies, 18<\/em>(3), 249\u2013268. https:\/\/doi.org\/10.1177\/1461445616634550<\/a><\/p>\n\n\n\n

Fairclough, N. (1989). Language and power<\/em>. Longman.<\/p>\n\n\n\n

Hjermann, A. R., & Wilhelmsen, J. (2025). Topos of threat and metapolitics in Russia\u2019s securitisation of NATO post-Crimea. Review of International Studies<\/em>, Advance online publication, 1\u201320. https:\/\/doi.org\/10.1017\/S0260210524000937<\/a><\/p>\n\n\n\n

Hoey, M. (2005). Lexical priming: A new theory of words and language<\/em>. Routledge.<\/p>\n\n\n\n

Krzy\u017canowski, M., Wodak, R., Bradby, H., Gardell, M., Kallis, A., Krzy\u017canowska, N., Mudde, C., & Rydgren, J. (2023). Discourses and practices of the \u201cnew normal\u201d: Towards an interdisciplinary research agenda on crisis and the normalization of anti- and post-democratic action. Journal of Language and Politics, 22<\/em>(4), 415\u2013437. https:\/\/doi.org\/10.1075\/jlp.23024.krz<\/a><\/p>\n\n\n\n

Janzwood, S., & Homer-Dixon, T. (2022). What is a global polycrisis? And how is it different from systemic risk?<\/em> Discussion paper. Cascade Institute.<\/p>\n\n\n\n

Janzwood, Scott, and Thomas Homer-Dixon. 2022. What Is a Global Polycrisis? And How Is It Different From Systemic Risk? Discussion Paper.<\/p>\n\n\n\n

Lipscy, P. Y. (2020). COVID-19 and the politics of crisis. International Organization, 74<\/em>(S1), E98\u2013E127. https:\/\/doi.org\/10.1017\/S0020818320000375<\/a><\/p>\n\n\n\n

Malone, D. (2025). The discourse presentation of the lone-wolf terrorist in the British press, 2000\u20132019: A corpus-based study<\/em> (Doctoral dissertation, Edge Hill University).<\/p>\n\n\n\n

Malone, D. (2026). Topic-specific corpus compilation: A componential approach to query formulation. Applied Corpus Linguistics, 6<\/em>(1), 100180. https:\/\/doi.org\/10.1016\/j.acorp.2025.100180<\/a><\/p>\n\n\n\n

Martin, J. R., & White, P. R. R. (2005). The language of evaluation: Appraisal in English<\/em>. Palgrave Macmillan.<\/p>\n\n\n\n

Stubbs, M. (2001). Words and phrases: Corpus studies of lexical semantics<\/em>. Blackwell.<\/p>\n\n\n\n

Wodak, R. (2001). The discourse-historical approach. In R. Wodak & M. Meyer (Eds.), Methods of critical discourse analysis<\/em> (pp. 63\u201395). Sage.<\/p>\n\n\n\n

\n\n\n\n

2025<\/strong><\/h2>\n\n\n\n
\n\n\n\n
MEETING #18: Friday 19 December 2025, 2:00-4:00 pm<\/h6>\n\n\n\n
Topic: Philosophies of Language and Corpus<\/em> Linguistics<\/em><\/h6>\n\n\n\n
Alan Partington<\/strong> (SiBol Group<\/a>\/CoLiTec<\/a>, Italy)<\/h5>\n\n\n\n
Language Distrusted, Language Ignored, Language Recovered: From Plato to Corpus Linguistics and Beyond<\/em><\/strong><\/h5>\n\n\n\n

SLIDES<\/a><\/strong> || VIDEO<\/a><\/strong><\/p>\n\n\n\n

I wish to discuss the Long Essay, A Short History of the Philosophies Underpinning Corpus Linguistics: From Aristotle to AI<\/em>, which traces the intertwined histories of linguistic and philosophical thought that shaped\u2014and sometimes resisted\u2014the emergence of corpus linguistics. The heavily revised, updated and now illustrated edition is free to download: https:\/\/doi.org\/10.5281\/zenodo.17998000<\/a><\/strong><\/p>\n\n\n\n

New themes include:
\u2022 Metaphorophobia in classical philosophy.
\u2022 How and why CL can fruitfully cohabit and collaborate with AI\/LLMs
\u2022 Ontological and epistemological differences between CL, CaDS and LLMs. LLMs: just artefacts or self-organising organisms (Kant)?
\u2022 How AI learns metaphorical usage and evaluation. Are there patterns of creative language (including humour) that AI can acquire then use?
\u2022 CL as a physical and a human science: causality (Bacon) versus teleology (Aristotle)
\u2022 CL\/CaDS and the revenge of evaluation, from Hunston to evaluative cohesion.<\/p>\n\n\n\n

Alan Partington has taught linguistics at the Universities of Bologna and Camerino, Italy. He is the author of Patterns and Meanings<\/em>, co-author of Patterns and Meanings in Discourse <\/em>(both John Benjamins) and author of The Linguistics of Political Argument<\/em>, The Linguistics of Laughter <\/em>and The Language of Persuasion in Politics and the Media <\/em>(all Routledge). He was the co-founding Editor-in Chief of the Journal of Corpora and Discourse Studies<\/em>. Contact: partington.alan@gmail.com<\/a><\/p>\n\n\n\n

\n\n\n\n
MEETING #17 Thursday 13 November 2025, 3-5 pm (GMT<\/a>)<\/h6>\n\n\n\n
Topic: LLMs and Corpus Tools<\/em><\/h6>\n\n\n\n
Mark Davies<\/strong><\/a> (English-Corpora.org<\/a>, USA)<\/h5>\n\n\n\n
Integrating information from AI \/ LLMs into English-Corpora.org<\/strong><\/em><\/h5>\n\n\n\n

SLIDES<\/a> <\/strong> || VIDEO<\/strong><\/a><\/p>\n\n\n\n

In March 2025 I released seven detailed studies<\/a> that discuss how well the predictions from LLMs (Large Language Models) match the actual data from large, well-known, publicly-accessible corpora (like those from English-Corpora.org). The seven detailed studies dealt with word frequency, phrase frequency, collocates, comparing words (via collocates), genre-based variation, historical variation, and dialectal variation.<\/p>\n\n\n\n

But it\u2019s probably not a question of \u201ceither\/or with corpora and AI; rather it is probably an issue of \u201cand\/with\u201d. Why not take the strengths of AI \/ LLMs and integrate them right into the corpus interface? As the comparisons between corpora and AI\/LLMs indicate, what LLM are really good at is classifying and explaining linguistic data.<\/p>\n\n\n\n

So as of September 2025, English-Corpora.org allows users to combine the depth and reliability of corpus data with the analytic power of LLMs like GPT, Gemini, Claude, Perplexity, Llama, Mistral, and DeepSeek. With just one click, the corpus can send collocates, frequency patterns, phrase lists, or concordance lines to an LLM via an \u201cAPI call\u201d, and then the LLM instantly groups, explains, and interprets the data, and returns that to the corpus. These AI-powered insights appear directly in the interface, alongside the original corpus results (while still keeping it very clear which is the corpus data, and which are the AI categorizations or analyses).<\/p>\n\n\n\n

The following are the types of analyses \/ categorizations that are now available to end users:<\/p>\n\n\n\n