factual

Priors Are Starting Points, Not Derivable

You cannot replace choosing a prior with another Bayesian sub-problem pushed up a stack to a foundational edge, because priors are starting points reflecting your current credences, not finishing points; you can argue for them only by relating one prior to another and checking internal coherence, but ultimately humans are embedded in the world with pictures of what makes sense, and grounding priors in some absolute foundation is the wrong goal—it is a feature, not a bug.

factualpending

Speaker

Sean Carroll

Evidence Quote

“We human beings are not disembodied, logical reasoners. We're part of the world. We have pictures of what makes sense to us and what doesn't make sense to us, and that's where we get our priors from, and that's okay. That's not a mistake. That's not a bug. It's a feature of how the system works.”

Source

AMA | April 2022— Sean Carroll's Mindscape
Created: 6/13/2026, 12:09:36 AM

Your notes

Loading notes...