relevant work from @andreaeyleen.bsky.social and myself:
How Computational Modeling Can Force Theory Building in Psychological Science doi.org/10.1177/1745...
On Logical Inference over Brains, Behaviour, and Artificial Neural Networks doi.org/10.1007/s421...
...
What Makes a Good Theory, and How Do We Make a Theory Good? doi.org/10.1007/s421...
A Metatheory of Classical and Modern Connectionism doi.org/10.1037/rev0...
Thanks! Full table here (below in post is published version, but my preprint version is more readable to me) comparing types of connectionist modelling in cognitive science
bsky.app/profile/oliv...
For the linking idea, known as mediation, also see the great work of Morgan and Morrison (1999) who basically created the modern pragmatic view on modelling
www.cambridge.org/core/books/m...
seop.illc.uva.nl/entries/stru...
Really great work!
CFP, AI Resistance, Refusal, Reclamation & Reimagining: Ethical Imperatives and Emerging Practices! Note "This collection is focused ...on the strategies & actions of individuals, communities, organisations & collectives to actively resist, refuse, reimagine & reclaim 'artificial intelligence."
Absolutely too busy to lead something for this, but this call for papers looks great...
> CfP: Topical Collection on AI Resistance, Refusal, Reclamation and Reimagining: Ethical Imperatives and Emerging Practices
Oh, and a related talk too from last week, kindly recorded by @bcbl.bsky.social, in which I analyze ANNs when used as scientific instruments of study and when functioning as emergent arbiters of the zeitgeist in the cognitive, computational, and neural sciences.
www.youtube.com/watch?v=4DHg...
🎬 🍿
Everyone, go watch this amazing talk!
🧠 🖥️
You can thank me later 😌
👇
Are you following @olivia.science? You should be.
that quote reminds me of the opening quote we used @andreaeyleen.bsky.social here:
[A]ll science would be superfluous if the outward appearance and the essence of things directly coincided. (Marx, 1894, p. 592) doi.org/10.1007/s421...
so even a baby knows "correlation does not imply cognition" 🙃
some resources for cognitive scientists — especially for junior scholars who ask me wonderful questions and want to learn more — on theorising and metatheorising olivia.science/theory/ (not 100% finished & more to come, but all my work is freely available here: olivia.science#publications as usual)
Love to show these to students. I used to teach them about the elephant clock by al-Jazari in class in AI as a Science
doi.org/10.5281/zeno...
broke: citing work from 2027
woke: citing work from 1987
BESPOKE: citing work from 1887
Louisa S. Cook's Geometrical psychology, or, The Science of Representation: An Abstract of the Theories and Diagrams of B. W. Betts @publicdomainrev.bsky.social
publicdomainreview.org/collection/g...
If I had lived in 1887 I would have definitely been studying mathematical psychology as well.
some reflections here if useful bsky.app/profile/oliv...
By the same token, our path model allows us to delineate and discuss where computational modeling itself has been compromised by QRPs occurring at the specification and implementation levels. A typical case of this is when authors report only partial results of implementing a specification of their theory; for example, only some implementations show the required or predicted patterns of behavior (Guest et al., 2020). As mentioned, the solution is to cycle within the red area of Figure 2 to ensure theory-, specification-, and implementation-level details are indeed assigned to and described at the correct level. Failing to do that, we propose, is a type of QRP.
Example of a case in which "authors report only partial results of implementing a specification of their theory; for example, only some implementations show the required or predicted patterns of behavior."
"The solution is to cycle within the red area of Figure 2."
It is at the data and hypothesis levels that preregistration and similar methods attempt to constrain science to avoid QRPs (e.g., Flis, 2019; Szollosi et al., 2019). To ensure scientific quality, however, we propose that preregistration is not enough because it serves only to constrain the data and hypothesis spaces. Researchers who wish to develop their formal account of a capacity must ascend the path instead of, or in addition to, for example, preregistering analysis plans. Preregistration cannot on its own evaluate theories. We cannot coherently describe and thus cannot sensibly preregister what we do not yet (formally and computationally) understand. Indeed, theories can and should be computationally embodied and pitted against each other without gathering or analyzing any new data. To develop, evaluate, and stress-test theories, we need theory-level constraints on and discussions about our science. Figure 2 can serve as a first step in the right direction toward such an ideal.
“To develop, evaluate, and stress-test theories, we need theory-level constraints on and discussions about our science.”
By @olivia.science & @andreaeyleen.bsky.social (2021): doi.org/10.1177/1745...
Monday, my AI class will discuss @olivia.science, @irisvanrooij.bsky.social, and Marcela Suarez's Toward Critical AI Literacies.
Cannot begin to guess how the conversation will go but excited to see where students are on this!
“Two figures showing a pencil drawn map with points A and B and a mountain range. On the left, the mountain range is in between points A and B. On the right, the part containing the mountain range has been cut out, showing the author’s desk through the hole cut in the map.” Figure 12 in: van de Braak, L., van Rooij, I., Dingemanse, M., Toni, I., & Blokpoel, M. (2025). Understanding misunderstanding: How quick-fix solutions undermine explanation. Zenodo. https://doi.org/10.5281/zenodo.17152893
New preprint from the lab!
🗣️ 🤔 💻 🗺️
📝 van de Braak, van Rooij, Dingemanse, Toni & Blokpoel (2025). Understanding misunderstanding: How quick-fix solutions undermine explanation. doi.org/10.5281/zeno... 📝
cc @irisvanrooij.bsky.social @dingemansemark.bsky.social @blokpoel.bsky.social
📚 ☀️
If you cannot attend, but am looking for online resources, please check out @olivia.science ’s amazing Critical AI Literacy website: olivia.science/ai/
5/🧵
📚 ☀️
Deadline for early bird fee is March 31. You can apply here: www.ru.nl/en/education... We have limited space, and will select based on motivation. We especially encourage women and/or minoritized people to apply.
4/🧵
📚 ☀️
The Summer School is a 5-day course, designed to foster critical AI literacies and to empower participants to develop ways of resisting or reclaiming AI in their own practices and social context.
1/🧵
Summer School: Critical AI Literacies for Resisting and Reclaiming irisvanrooijcogsci.com/2026/02/18/s... cc @olivia.science @marentierra.bsky.social
i don't know how to put this politely but stating that your work is approved/supported by companies like open ai, google and anthropic doesn't give you the credibility you think it does.
for me, this is a clear sign that i am not interested in engaging or working with you
Yuuuup... I'm so tired of explaining that maybe that's not actually good work if it's minimally giving the impression of a conflict of interest and maximally just an advertisement
bsky.app/profile/oliv...
For those who know what I think there is probably nothing new, but it might be useful to link to in discussions to help others. I just needed to collect these points as for me they keep coming up so much that I think it's easier to have them on my website. For example:
bsky.app/profile/oliv...
2/
I compiled a sort of follow up here olivia.science/before/ on the situation with ppl thinking we're in some special moment:
"We have certainly been here before. Many many times in the past, companies — just like artificial intelligence (AI) companies now — have lied to us to sell us products."
1/
guardrails are a scam...
as I said before: AI & any concept relating to it like so-called guardrails are a scam in the deepest sense like a perpetual motion machine or a ouija board — and not only a scam like a pyramid scheme which is a possible way to make money if you are first in first out
This is an important resource created by my colleague @olivia.science, useful for all who want to understand the origins of the Turing test. The gendered dimensions may surprise you (or not) and are still relevant today, for the ongoing abuses of correlationalism for the dehumanisation of women.
📚 Reading Women in Cognitive Science 📚
“Recommendations for readings are welcome, especially in the history of cognitive science (prior to 1950s, and the older the better).”
irisvanrooijcogsci.com/2026/02/15/%...
