3 ms·
Suppose I am a perfectly-rational Bayesian. Over the course of my long life I have acquired a false belief with a high degree of confidence. Through sheer stati
by jerf 16y ago
Suppose I am a perfectly-rational Bayesian. Over the course of my long life I have acquired a false belief with a high degree of confidence. Through sheer statistical flukery and the vicissitudes of living in the real world, a perfectly-rational Bayesian should actually expect that in their enormous stock of beliefs lies at least one of these. We all have access to only a tiny shred of the full totality of reality, and it isn't even a uniformly-randomly-chosen shred.
So you come up to me and present to me a fact that demonstrates I am objectively wrong in my belief. This being the real world, let us suppose that it isn't entirely unexplainable in my belief system. I take your fact, discount it with my strength of belief in your reliability, and update my Bayesian belief. It can stay quite strong, even if I behave with perfect rationality. With the way the Bayesian rule works, you can only utterly destroy my belief with something that is utterly impossible under my belief structure, but your ability to provide that evidence is capped by my inability to believe you that strongly in the first place. I must consider that you might be trying to fool me, or that you yourself might be wrong and bringing your own fallacious interpretation into what you may very well believe are objective facts.
If even a perfectly rational person may not be convinced by you simply telling them a few facts, how do you expect mere humans to behave?
I'm not necessarily defending people. If you think this country is going to hell in a handbasket because we're funding too many Bridges to Nowhere, but I provide evidence that Bridges to Nowhere (for whatever your definition is) account for only .04% of the Federal budget, you've still got at least a little cover left in wondering whether or not I'm using a different definition than you or just lying. Of course, this sort of belief should be susceptible to change over time as you are presented with numerous falsifications of your belief, but my point is not that people are secretly acting rational, but rather that this isn't really that mystifying, nor can one imagine it changing.