The argument here is insufficiently clear on the mechanism by which priors get trapped. I'm thinking the argument goes something like this: a Bayesian updater

The argument here is insufficiently clear on the mechanism by which priors get trapped. I'm thinking the argument goes something like this: a Bayesian updater takes in new evidence, and the adjusts their priors based on that evidence. Crucially, the information that is fed back into the Bayesian machine should be independent of the prior. People do not have direct access to evidence. They only have their perception of that evidence, and the content of those perceptions depend in various complicated ways on people's priors. Therefore, evidence that should move people's priors may move them less than they should. Sounds plausible. I think reinforcement learning algorithms like temporal difference learning are probably a more accurate description of how people learn tha Bayesian update, but the issue described would apply regardless.