Fausto Carcassi

This is *very cool*

Can LLMs use ToM to genuinely persuade you, or do they just use good rhetoric? In our new preprint, we use the MINDGAMES framework to test this. Surprisingly, LLMs like o3 can be incredibly effective persuaders *without* actually understanding your mental states. 🧵👇

Come to my and Tom Stephen's course at ESSLLI 2026! 2026.esslli.eu/courses-work... We'll consider typological data to formulate tightly fitting empirical constraints on the operation of semantic composition in natural language, in the tradition of Generalized Quantification Theory

my course notes on a bayesian workflow for (single agent) cognitive modeling are now fully revised and online: fusaroli.github.io/AdvancedCogn...

Predictive checks, updating checks, sensitivity analyses and simulation based calibration in @mc-stan.org

Feedback is very welcome!

A new preprint, co-authored with @johnwkrakauer.bsky.social:

The Deliberation Taboo

Cognitive science is, nominally, the science of thinking. We argue that the field has no theory of what thinking is and, even worse, that the topic has largely dropped out of focus. 1/

osf.io/preprints/ps...

**Postdoc position in human category learning**

@thecharleywu.bsky.social, Frank Jäkel and I are seeking a postdoctoral fellow to lead a joint project on human category learning at the Centre for Cognitive Science @tuda.bsky.social.

www.career.tu-darmstadt.de/tu-darmstadt...

New preprint

"Human-Like Coarse Object Representations in Vision Models"

arxiv.org/pdf/2602.12486

Drop an album that was important to you when you were nineteen.

Come work with us!!

Two full substitute professorships for Computational Linguistics (1 year) and General Linguistics (1.5 years) at the University of Tübingen. @unituebingen.bsky.social

uni-tuebingen.de/universitaet...

uni-tuebingen.de/universitaet...

I binged @tonytula.com's two books and want more so now I'm forced to read "Remote Research: Real Users, Real Time, Real Research" as a piece of narrative in the Tulathimutte-verse

you know there is one thing everyone could have that billionaires now have: "a meaningful say over their work, their lives, and the places they live". don't think Altman will like the story of how we get that for everyone tho

TIL Fats Domino's eight children were called: Antoine III, Anatole, Andre, Antonio, Antoinette, Andrea, Anola, and Adonica.

I agree! I am more worried for students. I guess I should have said: the fear is AI will absorb anyone who *could learn* what a function application is

I think the fear might be that formal semantics will disappear because AI will absorb everyone who knows what a function application is

Sadly, as a field it's just slightly too close to the star of AI research. Whether it will orbit or fall and crash is unclear (a lot of pessimism around though afaict).

Personally in teaching it I think we should emphasise more: (1) How strangely language behaves even in apparently simple domains like Boolean connectives (2) events events events!

So I think that, much like logic, TCS lives a split but stable existence between a topic students learn to make their thinking about language a bit more precise, and a research field that connects more and more with other fields (typology, cognitive science)

The field has matured so that the intro course (usually H&K, using S at the top) is quite far removed from the standard picture in the research lit (some kind of neo-Davidsonian event semantics w/ a rich verbal spine), though there's some attempts to realign (Coppock & Champollion textbook)

(1) a tool to state typological generalisations or describe underdocumented langs, (2) a framework to formulate precise empirical predictions to test (Jacopo Romoli has great stuff here), (3) a way to study how certain kind of meanings are realized and dealt with (e.g. degrees), ...

In my experience most people aren't as committed to (/interested in?) the big foundations that motivated it in the old days (I think Davidson (?) somewhere sums it up as: a systematic account of compositionality+entailment patterns+logical form), and instead use TCS in more applied ways e.g., as:

Our large-scale interdisciplinary research initiative @unituebingen.bsky.social on "Common Ground" now has a website: uni-tuebingen.de/en/research/...

We are also hiring! Check out recent postings here and the continually added job ads here:
uni-tuebingen.de/en/universit...

#linguistics + friends

Excited to share a new CogSci paper co-led with @benpry.bsky.social!

Once a cornerstone for studying human reasoning, the think-aloud method declined in popularity as manual coding limited its scale. We introduce a method to automate analysis of verbal reports and scale think-aloud studies. (1/8)🧵

First day notes now posted: juliangrove.github.io/nasslli-2025...

I'm packing for NASSLLI 2025. The lineup is terrific! I'm looking forward to enjoying Seattle for the first time and climbing again with @shanest.bsky.social, who organized the school (kudos!). nasslli25.shane.st/schedule/

Yeah, I am in the process of metabolizing that particular shock...

If I think of the time I squandered, Battiato comes to mind:

"Se penso a come ho speso male il mio tempo
Che non tornerà, non ritornerà più"

Is it blasphemy to say that pre-WWI biology (filtered through the humanistic eyes of Thomas Mann) has a FEP vibe to it?

which could in all purity make beauty perceptible to the senses. Rather was it conveyed and shaped by the somehow awakened voluptuousness of matter, of the organic, dying-living substance itself, the reeking flesh.

and which became form, beauty, a lofty image, and yet all the time the essence of sensuality and desire. For this form and beauty were not spiritborne; nor, like the form and beauty of sculpture, conveyed by a neutral and spiritconsumed substance,