Jean Barré
@jbarre.bsky.social
PhD student @ École Normale Supérieure in Paris. Working in the Computational Literary Studies field on literary evolution of novel subgenres & canonization process + fr-BookNLP implementation w/ @labolattice.bsky.social
https://crazyjeannot.github.io/
https://crazyjeannot.github.io/
This leads to our key question: despite high accuracy & strong unity, how does the archetype evolve across genres and time?
So we model its Semantic Trajectory : from a cold, rational reasoner to an intuitive, emotionally engaged figure.
💻 Interactive viz → crazyjeannot.github.io/pdfs/detecti...
So we model its Semantic Trajectory : from a cold, rational reasoner to an intuitive, emotionally engaged figure.
💻 Interactive viz → crazyjeannot.github.io/pdfs/detecti...
November 4, 2025 at 5:36 PM
This leads to our key question: despite high accuracy & strong unity, how does the archetype evolve across genres and time?
So we model its Semantic Trajectory : from a cold, rational reasoner to an intuitive, emotionally engaged figure.
💻 Interactive viz → crazyjeannot.github.io/pdfs/detecti...
So we model its Semantic Trajectory : from a cold, rational reasoner to an intuitive, emotionally engaged figure.
💻 Interactive viz → crazyjeannot.github.io/pdfs/detecti...
Over the 19th and 20th centuries, the detective becomes both more frequent and more central in French fiction.
Our model predicts not only a growing ratio of detective characters but also a rise in the proportion of text devoted to them.
Our model predicts not only a growing ratio of detective characters but also a rise in the proportion of text devoted to them.
November 4, 2025 at 5:36 PM
Over the 19th and 20th centuries, the detective becomes both more frequent and more central in French fiction.
Our model predicts not only a growing ratio of detective characters but also a rise in the proportion of text devoted to them.
Our model predicts not only a growing ratio of detective characters but also a rise in the proportion of text devoted to them.
What happens when we model the detective archetype at scale? 🕵️♂️📚
Our new paper, accepted for #CHR2025 combines literary history and computational modeling to trace how the figure of the detective evolves across 150 years of French fiction.
arxiv.org/pdf/2511.00627
Our new paper, accepted for #CHR2025 combines literary history and computational modeling to trace how the figure of the detective evolves across 150 years of French fiction.
arxiv.org/pdf/2511.00627
November 4, 2025 at 5:36 PM
What happens when we model the detective archetype at scale? 🕵️♂️📚
Our new paper, accepted for #CHR2025 combines literary history and computational modeling to trace how the figure of the detective evolves across 150 years of French fiction.
arxiv.org/pdf/2511.00627
Our new paper, accepted for #CHR2025 combines literary history and computational modeling to trace how the figure of the detective evolves across 150 years of French fiction.
arxiv.org/pdf/2511.00627
I had fun presenting some of my PhD obsessions about the french detective novel in Würzburg.
Thank you @fotisjannidis.bsky.social for the invitation ! The whole team is impressive, brand new building and talented people, the future of DH is actually here 🤩
Thank you @fotisjannidis.bsky.social for the invitation ! The whole team is impressive, brand new building and talented people, the future of DH is actually here 🤩
June 16, 2025 at 9:39 AM
I had fun presenting some of my PhD obsessions about the french detective novel in Würzburg.
Thank you @fotisjannidis.bsky.social for the invitation ! The whole team is impressive, brand new building and talented people, the future of DH is actually here 🤩
Thank you @fotisjannidis.bsky.social for the invitation ! The whole team is impressive, brand new building and talented people, the future of DH is actually here 🤩
What drives this tense-revolution? Our OLS models reveal it’s genre above all, with time period and canonicity as secondary factors. 🔍
May 6, 2025 at 5:30 PM
What drives this tense-revolution? Our OLS models reveal it’s genre above all, with time period and canonicity as secondary factors. 🔍
This article evaluates different drivers of change: genre conventions, canonicity, or simple temporal drift?🤔
First, we mapped the passé simple’s co-movement over 200 years 🔄, revealing a clear split between past tenses (imparfait, plus-que-parfait) and present tenses (présent, futur, passé composé)
First, we mapped the passé simple’s co-movement over 200 years 🔄, revealing a clear split between past tenses (imparfait, plus-que-parfait) and present tenses (présent, futur, passé composé)
May 6, 2025 at 5:28 PM
This article evaluates different drivers of change: genre conventions, canonicity, or simple temporal drift?🤔
First, we mapped the passé simple’s co-movement over 200 years 🔄, revealing a clear split between past tenses (imparfait, plus-que-parfait) and present tenses (présent, futur, passé composé)
First, we mapped the passé simple’s co-movement over 200 years 🔄, revealing a clear split between past tenses (imparfait, plus-que-parfait) and present tenses (présent, futur, passé composé)
Little baseline for french - 14 years of mean absolute error
Good old sklearn linear regression, 3000 novels
A rushy but successful experiment in front of students 🪄
Good old sklearn linear regression, 3000 novels
A rushy but successful experiment in front of students 🪄
December 23, 2024 at 7:23 PM
Little baseline for french - 14 years of mean absolute error
Good old sklearn linear regression, 3000 novels
A rushy but successful experiment in front of students 🪄
Good old sklearn linear regression, 3000 novels
A rushy but successful experiment in front of students 🪄
🖋️ Featuring:
- Literary NER
- Coreference resolution
Results may not match standard NLP benchmarks for NER, but hey, it's literary NER—so it's tough! 😢
+ Coreference resolution at the novel scale! (No 512-token limitation here, 😏😊)
- Literary NER
- Coreference resolution
Results may not match standard NLP benchmarks for NER, but hey, it's literary NER—so it's tough! 😢
+ Coreference resolution at the novel scale! (No 512-token limitation here, 😏😊)
November 21, 2024 at 4:54 PM
🖋️ Featuring:
- Literary NER
- Coreference resolution
Results may not match standard NLP benchmarks for NER, but hey, it's literary NER—so it's tough! 😢
+ Coreference resolution at the novel scale! (No 512-token limitation here, 😏😊)
- Literary NER
- Coreference resolution
Results may not match standard NLP benchmarks for NER, but hey, it's literary NER—so it's tough! 😢
+ Coreference resolution at the novel scale! (No 512-token limitation here, 😏😊)
For genres -we took the example of synthetic "Adventure novels" labels- the pattern is less obvious, but we partly have what we expected i.e. higher similarity around the publication date
October 25, 2024 at 4:20 PM
For genres -we took the example of synthetic "Adventure novels" labels- the pattern is less obvious, but we partly have what we expected i.e. higher similarity around the publication date
It is when we added Canonical tags that things went wild : Look at this crazy trend where canonical works (blue line) distinctly separate from the archive around the date of publication, showing a higher cosine similarity with what comes after them
October 25, 2024 at 4:19 PM
It is when we added Canonical tags that things went wild : Look at this crazy trend where canonical works (blue line) distinctly separate from the archive around the date of publication, showing a higher cosine similarity with what comes after them
This gives us individual similarity trends - Here is Victor Hugo's "Les Misérables" pattern over time
And collective ones (Here is the mean similarity with years before and after)- This graph confirms earlier research on the "style of a time" trend.
And collective ones (Here is the mean similarity with years before and after)- This graph confirms earlier research on the "style of a time" trend.
October 25, 2024 at 4:18 PM
This gives us individual similarity trends - Here is Victor Hugo's "Les Misérables" pattern over time
And collective ones (Here is the mean similarity with years before and after)- This graph confirms earlier research on the "style of a time" trend.
And collective ones (Here is the mean similarity with years before and after)- This graph confirms earlier research on the "style of a time" trend.
better late than never 😋
September 17, 2024 at 9:47 AM
better late than never 😋
- Corpus scale: proportion decreases over time, from 30% to 15% in 2 c. Correlation with the adventure subgenre distribution ? or is it non fiction raising ? or even the change in medium for the daily dose of adventure ? No idea tbh, but further work is exciting !
March 13, 2024 at 4:41 PM
- Corpus scale: proportion decreases over time, from 30% to 15% in 2 c. Correlation with the adventure subgenre distribution ? or is it non fiction raising ? or even the change in medium for the daily dose of adventure ? No idea tbh, but further work is exciting !
- Subgenre scale: Stereotypical adventures scenes are most common in.. travelogue genre ! Here we clearly see the limitation of gpt annotations, whith the "change of scenery" having too much impact in the definition of the adventure scenes.
March 13, 2024 at 4:37 PM
- Subgenre scale: Stereotypical adventures scenes are most common in.. travelogue genre ! Here we clearly see the limitation of gpt annotations, whith the "change of scenery" having too much impact in the definition of the adventure scenes.
Then we ran inferences on 3 scales:
- Novel scale w/ a stereotypical adventure novel : Vingt Milles Lieues sous les mers and an adventures novel very far from being adventurous... L'Éducation sentimentale
- Novel scale w/ a stereotypical adventure novel : Vingt Milles Lieues sous les mers and an adventures novel very far from being adventurous... L'Éducation sentimentale
March 13, 2024 at 4:36 PM
Then we ran inferences on 3 scales:
- Novel scale w/ a stereotypical adventure novel : Vingt Milles Lieues sous les mers and an adventures novel very far from being adventurous... L'Éducation sentimentale
- Novel scale w/ a stereotypical adventure novel : Vingt Milles Lieues sous les mers and an adventures novel very far from being adventurous... L'Éducation sentimentale
New blog post ! "LLMs in Quest for Adventure"
👉 huggingface.co/blog/crazyje... 👈
First experiments with LLMs, trying to evaluate how they can be useful for computational literary studies. Here we focus on the detection of stereotypical scenes in adventure novel.
👉 huggingface.co/blog/crazyje... 👈
First experiments with LLMs, trying to evaluate how they can be useful for computational literary studies. Here we focus on the detection of stereotypical scenes in adventure novel.
March 13, 2024 at 4:32 PM
New blog post ! "LLMs in Quest for Adventure"
👉 huggingface.co/blog/crazyje... 👈
First experiments with LLMs, trying to evaluate how they can be useful for computational literary studies. Here we focus on the detection of stereotypical scenes in adventure novel.
👉 huggingface.co/blog/crazyje... 👈
First experiments with LLMs, trying to evaluate how they can be useful for computational literary studies. Here we focus on the detection of stereotypical scenes in adventure novel.
The findings reveal that canonical authors tend to contribute slightly more to literary change than those from the archive. Tentative explanations in the paper! Spoiler: really need more work to have the final word on this question.
November 27, 2023 at 1:16 PM
The findings reveal that canonical authors tend to contribute slightly more to literary change than those from the archive. Tentative explanations in the paper! Spoiler: really need more work to have the final word on this question.
While this image may appear to be random noise, each peak represents a significant level of innovation within a specific cohort when compared to the preceding one, with birth dates plotted on the x-axis.
November 27, 2023 at 1:14 PM
While this image may appear to be random noise, each peak represents a significant level of innovation within a specific cohort when compared to the preceding one, with birth dates plotted on the x-axis.
Is it possible to identify key authors and novels that have had an impact on literary change? Are those names present in our contemporary literary canon? New article out w @tpoibeau.bsky.social: "Beyond Canonicity. Modeling Canon/Archive Literary Change in French Fiction"
ceur-ws.org/Vol-3558/pap...
ceur-ws.org/Vol-3558/pap...
November 27, 2023 at 1:11 PM
Is it possible to identify key authors and novels that have had an impact on literary change? Are those names present in our contemporary literary canon? New article out w @tpoibeau.bsky.social: "Beyond Canonicity. Modeling Canon/Archive Literary Change in French Fiction"
ceur-ws.org/Vol-3558/pap...
ceur-ws.org/Vol-3558/pap...
Take a look at the fantastic workshop scheduled for December 5th in Paris! It's being held in coordination with the CHR conference! 👉https://workshop-llms4cls.github.io/👈
October 29, 2023 at 2:04 PM
Take a look at the fantastic workshop scheduled for December 5th in Paris! It's being held in coordination with the CHR conference! 👉https://workshop-llms4cls.github.io/👈
or Balzac/La Comédie humaine 😆
October 25, 2023 at 9:34 AM
or Balzac/La Comédie humaine 😆
I am struggling with top2vec to get novel representation w/ topic modelling, half of the resulted topics are clean but half are clustered character names... from the same author/narrative cycle. That is not what I want but it's really interesting. For example Zola & les Rougon-Macquart:
October 25, 2023 at 9:28 AM
I am struggling with top2vec to get novel representation w/ topic modelling, half of the resulted topics are clean but half are clustered character names... from the same author/narrative cycle. That is not what I want but it's really interesting. For example Zola & les Rougon-Macquart:
Working at the novel scale, the same pattern emerged : certain works within an author’s oeuvre might align more closely with the established norms and criteria of the literary canon, while others might deviate or be less congruent with those norms. For example, Colette's works :
October 19, 2023 at 4:18 PM
Working at the novel scale, the same pattern emerged : certain works within an author’s oeuvre might align more closely with the established norms and criteria of the literary canon, while others might deviate or be less congruent with those norms. For example, Colette's works :
We then analyzed features importance from our SVM inferences. Spoiler : hard work, but raw insights are: dialogs, facts vs stylistic overdose :p 5/
October 19, 2023 at 4:18 PM
We then analyzed features importance from our SVM inferences. Spoiler : hard work, but raw insights are: dialogs, facts vs stylistic overdose :p 5/
It's out ! Very excited to present "Operationalizing Canonicity: A Quantitative Study of French 19th and 20th Century Literature." by me, @jbcamps.bsky.social and Thierry Poibeau. This pic shows the main results, but let me tell you the full story : 1/N
t.co/y3FtrrEqQn
t.co/y3FtrrEqQn
October 19, 2023 at 4:16 PM
It's out ! Very excited to present "Operationalizing Canonicity: A Quantitative Study of French 19th and 20th Century Literature." by me, @jbcamps.bsky.social and Thierry Poibeau. This pic shows the main results, but let me tell you the full story : 1/N
t.co/y3FtrrEqQn
t.co/y3FtrrEqQn