Give orange me give eat orange me eat orange give me

This is very likely an overgeneralization from my personal experience, but I suspect that "belief fixation" is a signal that the speaker/writer has recently read Fodor and has adopted his terminology

I agree that if that happened (and represented real belief change, not expressive responding or even covert reasoning), it would call for real explanation

But you can test this on yourself right now and I expect saying something counterattitudinal won't have much effect

Fair, I'm inclined to be bearish on rationalizations across the board, but they probably do happen sometimes, including via this mechanism

Potentially, at least. It seems plausible to me that writing a counterattitudinal essay makes you consider new arguments + that this could lead to attitude change.

Post-hoc unconscious attitude modulation solely bc you wrote (or *chose* to write) the essay seems less plausible. Compare Vaidis et al

Imo, a lot of this debate turns on whether dissonance is an unconscious, irrational thing that just happens to people (maybe coupled with factors like self-esteem or free choice) or a (quasi)rational effect of looking at one-sided evidence or using auxiliary Hs to protect a view you're confident of

2. Symbolic cognitive processes; some/a lot of access, the mind involves structured computations, not just associations; LLMs don't work like people because they lack compositionality/structure; behavior is *a means of inferring theoretical constructs*

Np! If you're interested, here are a few other recent papers I had in mind:
www.annualreviews.org/content/jour...

www.nature.com/articles/s44...

osf.io/preprints/ps...

FWIW, pretty strong (imo) critiques of the N+W analysis have been around since their original work (e.g., psycnet.apa.org/record/1980-...) and there appears (to my eyes, at least) to be a big trend toward finding more introspective access than we've thought (one ex: mitpress.mit.edu/978026254619...)

• Moè: intentions vs. reality
• Osman: mis/disinfo impacts @osman
• Pierre: consistency puzzles
• Sommer & Oktar: expecting low links @keremoktar.bsky.social @bayesandbounds.bsky.social
• Westaby et al.: behavioral reasoning

Nice video - something to consider:

"what people believed was based not on evidence but on how they felt about the advocates"

I think it looks like this from the outside, but trusting (authoritative) in-group members *is* (defeasible!) evidence. Compare trusting vaccine researchers' testimony

Last new year's, I hung out with my friend Joseph @bayesandbounds.bsky.social to talk about the 100 most influential works in cognitive science. All the big names are here: Turing, Fodor, Marr, Miller, (Joseph's favorite) Herb Simon, and of course - Noam Chomsky!

youtu.be/Zu6ZoZsRGG0