iiinova — where your digital self keeps evolving

Thierry Poibeau @tpoibeau.bsky.social · 11/09/2026

Reposted by Jean Barré

"Les Mots sans les choses. Philosophie, éthique et impact des grands modèles de langue" vient de paraître aux Editions Rue d'Ulm. Que changent ces modèles à nos façons de penser et d’écrire, et quelles conséquences pour la société et notre autonomie ? presses.ens.psl.eu/les-mots-san...

presses.ens.psl.eu

Les Mots sans les choses

<p style="font-weight: 400;">Philosophie, éthique et impact des grands modèles de langue</p>

133

Reposted by Jean Barré

Merten Kröncke @mertenkroencke.bsky.social · 10/09/2026

Canonicity is central to literary studies, and most researchers have some intuition about which authors or texts are canonical. But how can we measure canonicity in a truly systematic manner? A new article by @agneshilger.bsky.social, @janaeckardt.bsky.social and me tries to theorize this question.

doi.org

Taking a Step Back: Theoretical Fundamentals of Canonicity and Its Empirical Reconstruction

Abstract. Canonicity is central to literary studies. Even if scholars are not explicitly concerned with it, canonicity shapes their selection of texts and authors—and that, in turn, may reinforce the ...

02810

Reposted by Jean Barré

Alexey Koshevoy @alexeykoshevoy.bsky.social · 07/09/2026

Are you interested in lexical competition, agent-based modelling, strong inference and Ukrainian? Then this paper, just published in @pnasnexus.org, is for you. With @oliviermorin.bsky.social and @sobchuk.bsky.social, we asked a simple question: why do some words become more common than others?

24317

Reposted by Jean Barré

Dan Sinykin @dan-sinnamon.bsky.social · 24/08/2026

Fiction has undergone a huge demographic transformation over the last 50 years. Men lost ground, which women gained. But the way some have written about it, you’d think men are, as a NYT headline suggested, disappearing. Hardly! Women have just caught up. sinykin.substack.com/p/where-have...

sinykin.substack.com

Where Have All the Literary Men Gone?

Not too far....

313541

Jean Barré @jbarre.bsky.social · 24/08/2026

3,632 authors, most absent from literary histories. Novel production peaks in the 1880s, golden age of the roman-feuilleton 📈 Metadata errors remain at this scale: I see the Zenodo repo as WIP, versioned as it's corrected/reused.

Histogram of novel counts by year, 1800–1950, showing a peak in the 1880s and a decline after 1900 partly due to Gallica's digitization coverage

020

Jean Barré @jbarre.bsky.social · 24/08/2026

Big thanks to @dorialexander.bsky.social for the upstream corpus. I added OCR filtering, metadata enrichment (Wikidata IDs, author dates, genre tags) and heavy deduplication: 19K → 11k novels.

130

Jean Barré @jbarre.bsky.social · 24/08/2026

Interested by a corpus of 19th/early 20th c. French for your research? Here's the French corpus of 11K! novels I tried to curate during my PhD 😅, now live at JOHD, and downloadable on Zenodo! 📄 doi.org/10.5334/johd.597 📦 zenodo.org/records/20233178

doi.org

Hors d’usage: Curated French Fiction from Gallica, 1800–1950 | Journal of Open Humanities Data

34421

Reposted by Jean Barré

Jean-Baptiste Camps @jbcamps.bsky.social · 07/07/2026

Freshly out in @pnasnexus.org: "On the transmission of texts: Written cultures as complex systems" The first major publication of project @lostma.bsky.social , where we lay down our methodology for studying the transmission, evolution or disparition of texts. /1 #science 🧪

academic.oup.com

On the transmission of texts: Written cultures as complex systems

Abstract. Knowledge of past cultures relies heavily on surviving written material. Over the centuries, texts were copied, altered, and often lost, leaving

24420

The final corpus: 1,391 adventure novels, peaking 1870–1930, matching independent periodization 🎯 A key genre notion, recovered at archive scale. Verne headlines (60 titles), but 68.5% of the 569 authors appear with a single title: a mass production lost to memory 📚

Horizontal bar chart of titles per author for all 569 authors in the final corpus. A handful of headliners top the list — Jules Verne with 60 titles, Paul Féval with 48, Ponson du Terrail with 41 — followed by an extremely long tail: 68.5% of authors appear with a single title, illustrating the mass, largely forgotten production of the genre.

131

And the graph corrections are readable: 113 outliers removed (guidebooks, parodies -- Le Captain Cap!), 129 infiltrators added (missed Verne novels, Mayne Reid's frontier, Marryat's maritime). The model captures the genre's material, not its modalization (irony, parody).

Case study: the adventure novel, massively read 1870–1930, largely forgotten today. Stage 1: a classifier trained on 101 novels cited by Letourneux (2010). Stage 2: a k-NN graph refinement over the ~11,000 novels of Gallica.

Flowchart of the two-stage pipeline. Stage 1: a seed list of 101 adventure novels cited by Letourneux (2010) plus negatives sampled from Gallica train a supervised classifier, which produces predictions on the full archive. Stage 2: a k-NN graph refinement (k=10, threshold 80%) removes outliers and adds infiltrators, yielding a final corpus of 1,391 adventure novels.

Sil Hamilton @srhm.ca · 03/07/2026

Small new paper ! 🎉 Genre is a problematic notion in CLS: bibliographically unstable, heterogeneously annotated, yet central to many experiments Our new #NLP4DH paper (ACL 2026) proposes a method to retrieve genre tags automatically in a huge novel archive, from a small seed list only.

Network visualization of a k-nearest-neighbors graph over 10,998 French novels from Gallica. Each node is a novel, connected to its most similar neighbors by cosine similarity of text embeddings. Orange nodes are predicted adventure novels, green nodes are randomly sampled non-adventure novels.

1194

Reposted by Jean Barré

Attending #ACL? Come join us at NLP4DH 2026 on Saturday from 10am to 5pm! Great selection of papers from across the DH & NLP spectrum this year. Schedule: www.nlp4dh.com/nlp4dh-2026/...

Reposted by Jean Barré

Oleg Sobchuk 🇺🇦 @sobchuk.bsky.social · 29/06/2026

📣 I'm hiring! Two positions (PhD student & Postdoc) in my ERC project Macroevolution of European Literature. Let's study the cultural evolution of literature using massive data 📚📈 📍Frankfurt | Apply by August 15 Details: www.aesthetics.mpg.de/en/career/jo... Please repost to help spread the word!

Job announcement of two positions at the Max Planck Institute for Empirical Aesthetics in Frankfurt: postdoc (Macroevolution of European Literature) and PhD student (Computational Literary Studies). Successful candidates will be affiliated with the research group Cultural Evolution of the Arts.

08082

Reposted by Jean Barré

Julia Neugarten @julianeugarten.bsky.social · 22/06/2026

Happy to announce that our article 'The style of scorn: computationally operationalizing the relationship between anger and complexity in 'Ik Ga Leven' by Lale Gül' (together with @jorisvanzundert.bsky.social) is out now with Digital Scholarship in the Humanities! OA link: doi.org/10.1093/llc/...

084

Reposted by Jean Barré

Judith Brottrager @jbrottrager.bsky.social · 19/06/2026

Kicking off the second day of #MappingtheCanon, @jbarre.bsky.social tells us more about the link between genre traditions and the canon.

181

Reposted by Jean Barré

Judith Brottrager @jbrottrager.bsky.social · 18/06/2026

Let the mapping begin! #MappingtheCanon

1134

Jean Barré @jbarre.bsky.social · 11/06/2026

PhDone ! 🎓 Immense gratitude to my jury (including @evelyngius.bsky.social @matthieuletourneux.bsky.social @tedunderwood.com ), my co-directors @tpoibeau.bsky.social & Thomas Conrad, and everyone who supported me along the way ❤️

Six people (Thzierry Poibeau, Thomas Conrad, Marie-Ève Thérenty, Me, Claire Barel-Moisan &Matthieu Letourneux) stand in a row smiling in front of a projection screen for a group photo after a PhD defense. In the centre, a young man with curly dark hair in a dark blazer holds a large bouquet wrapped in kraft paper. The screen behind them reads "Soutenance de thèse" with ENS-PSL and Lattice logos, and the thesis title about computational methods for modelling genre fiction in French-language literature.

A man with curly dark hair, wearing a dark blazer over a white shirt and beige trousers, sits at a desk working on a laptop during his PhD defense. Behind him, a large projection screen shows a video-call interface with two remote participants (Evelyn Gius & Ted Underwood) in separate windows. A small conference-room camera device stands on a table to the right.

5382

Reposted by Jean Barré

Judith Brottrager @jbrottrager.bsky.social · 03/06/2026

Very excited about a very special event: in two weeks, some of the most interesting quantitative work on the literary canon and canon formation comes together in Darmstadt 🧵

26025

Reposted by Jean Barré

Alexey Koshevoy @alexeykoshevoy.bsky.social · 19/05/2026

🚨New paper alert!🚨 In our new paper with @oliviermorin.bsky.social, Rowan Hall, and Marie Halo published in the Journal of Language Evolution, we explored the relationship between word length and ambiguity on a large-cross linguistic dataset. 1/6

2246

Jean Barré @jbarre.bsky.social · 12/05/2026

Defending my PhD on June 5 at ENS Paris ! "Modeling Genre Literature: Computational Methods and Genre Evolution in French-Language Fiction"—detective + adventure novels, 1866–2020. Jury includes @evelyngius.bsky.social, @tedunderwood.com & @matthieuletourneux.bsky.social. Open to all !

0194

Jean Barré @jbarre.bsky.social · 12/05/2026

Je soutiens mon Doctorat le 5 juin à l'ENS à 14h, salle des Résistants (45 rue d'Ulm) : « Modéliser la littérature de genre. Méthodes computationnelles et évolution des genres dans la fiction de langue française » une aventure qui se termine ! 🎉

Page de couverture de la thèse de doctorat de Jean Barré (PSL / ENS, laboratoire Lattice), intitulée « Modéliser la littérature de genre : Méthodes computationnelles et évolution des genres dans la fiction de langue française ». Soutenance prévue le 5 juin 2026. Jury : Marie-Ève Thérenty (Présidente, Montpellier III), Evelyn Gius (Rapporteure, TU Darmstadt), Matthieu Letourneux (Rapporteur, Paris-Nanterre), Claire Barel-Moisan (Examinatrice, ENS Lyon / CNRS), Ted Underwood (Examinateur, Illinois Urbana-Champaign), Thomas Conrad (Co-encadrant, ENS-PSL), Thierry Poibeau (Directeur, ENS-PSL / CNRS).

151

Reposted by Jean Barré

Computational Humanities Research @comphumresearch.bsky.social · 22/04/2026

📢 CfP for #CHR2027 is out! Submissions on all aspects of computational humanities research welcome; all details available here: 2027.computational-humanities-research.org/cfp/ And yes, CHR2027 will take place a little later than usual. So you *could* start later… but we suggest starting now!

03629

Jean Barré @jbarre.bsky.social · 07/04/2026

yippeeee ! bravo ! Avoues ils t'ont filé le diplôme quand ils t'ont vu dans cet accoutrement

Nathan Lambert @natolambert.bsky.social · 02/04/2026

Reposted by Jean Barré

Google dropped 4 different Gemma open-weight models! I'm most excited that they're finally adopting a standard Apache 2.0 open source license. huggingface.co/collections/...

211021

Reposted by Jean Barré

Katarzyna Anna Kapitan @kakapitan.bsky.social · 25/03/2026

#jobklaxon PhD Fellowships available at @psl-univ.bsky.social! Come & work w/us in CultureLab (w/ @jbcamps.bsky.social @oliviermorin.bsky.social et al not bsky). Deadline 31/05 Starting 01/10 Duration 3 yrs (fixed-term employment with benefits) Salary: sad. www.culturelab.psl.eu/en/news-cult...

culturelab.psl.eu

2026 CultureLab PhD Grant Call - Grands Programmes de Recherche-Culture Lab

068

Reposted by Jean Barré

Katarzyna Anna Kapitan @kakapitan.bsky.social · 20/03/2026

#jobklaxon 10 postdocs available at @psl-univ.bsky.social in #AI. Working language: English. Btw. #digitalhumanities also happen under the AI hype these days, so get in touch and walk by this very building on your way to work. Join me in being #touristatwork #juniorProfInParis

12020

Reposted by Jean Barré

Folgert Karsdorp @folgertk.bsky.social · 28/02/2026

Out in Evolutionary Human Sciences! With @mikekestemont.bsky.social, @jbcamps.bsky.social, @remcosleiderink.bsky.social & Anne Chao New work on unseen species models for cult heritage to the question: how many stories were _shared_ between medieval French and Dutch literature? lnkd.in/exyAWtir

lnkd.in

This link will take you to a page that’s not on LinkedIn

0158

Reposted by Jean Barré

Ryan Heuser @ryanheuser.com · 26/02/2026

I'm on a 38(!)-author paper just published in Frontiers in Artificial Intelligence, "Computational hermeneutics: evaluating generative AI as a cultural technology". We splice Schleiermacher and hermeneutic theory into AI debates, arguing AI are "context machines". www.frontiersin.org/journals/art...

frontiersin.org

Frontiers | Computational hermeneutics: evaluating generative AI as a cultural technology

Generative AI (GenAI) systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be...

55420

Reposted by Jean Barré

Andrew Lampinen @lampinen.bsky.social · 18/02/2026

What is the relationship between memorization and generalization in AI? Is there a fundamental tradeoff? In infinitefaculty.substack.com/p/memorizati... I’ve reviewed some of the evolving perspectives on memorization & generalization in machine learning, from classic perspectives through LLMs.

infinitefaculty.substack.com

Memorization vs. generalization in deep learning: implicit biases, benign overfitting, and more

Or: how I learned to stop worrying and love the memorization

415629

GitHub - lattice-8094/fictional-character-ontology: Repository for the data & code of the paper "Toward an Ontological Representation of Fictional Characters"

A bit ironically, we end up learning as much about the researchers (and their implicit representations) as about the characters themselves. Kudos to @antoine-bourgois.bsky.social , first author & 1st-year PhD student 👏 📂 Data + code → github.com/lattice-8094/fictional-character-ontology

github.com

Repository for the data & code of the paper "Toward an Ontological Representation of Fictional Characters" - lattice-8094/fictional-character-ontology

050

This flips the similarity question on its head. It's no longer: "Are these two characters similar?" But rather: "Along which dimensions did the scholars who designed this benchmark implicitly define similarity between these two characters?"

We then built on @dbamman.bsky.social et al.'s (2014) triplets protocol and introduce CharaSim-fr, a brand new benchmark for characters similarity Key method: we exhaustively compute all possible mixes of concatenated ontological classes (131,054 combinations!) to find the optimal similarity signal

Box plot showing accuracy scores for all combinations of 1 to 17 ontological character attribute classes. Three trend lines track maximum (green), mean (red), and minimum (orange) accuracy. The key finding: maximum accuracy peaks at ~0.96 with just 4–5 carefully selected classes, while using all 17 classes simultaneously yields no improvement over the mean (~0.64). A small, well-chosen subset of ontological dimensions outperforms the full representation — more is not better.

So we built an ontology of characterization w/ 17 classes grounded in narratological theory: actions, emotions, personality traits, relations, cognition, objects, body parts, ++ The goal: stop treating characters as bags of features and start asking along which dimensions they resemble each other.

≠ Rosanette: less refined, more eroticized, defined largely through dialogue ≠ Mme Arnoux: the central figure — the other two are satellites who exist mainly by contrast with her ≠ Mme Dambreuse: far more effaced, valued by Frédéric mainly for the social access she provides — little interiority

The story starts with an argument among co-authors about three women in Flaubert's L'Éducation sentimentale: Mme Arnoux, Rosanette, Mme Dambreuse. The question: who is the most dissimilar in this triplet? Three defensible answers That is exactly where the tears began

Sweat, because we tried to grasp what it actually means for two characters to be "similar" measured across 100,000+ characters. The standard approach? Character embeddings, cosine similarity, done. We were so naive. So wrong.