>I'm arguing that mere truth is unknowable without faith.
Could you please define "mere" truth?
>From a bayesian perspective, one of our priors must be that there is some probability that the very mind that is doing the thinking may be of a nature that is unable to effectively reason.
Nonetheless, that probability can be very low, putting us firmly in normal epistemological territory and not requiring any "faith". In fact, millions of people regularly make the rational, but clouded, assessment that they are too drunk to drive, and take a taxi home.
(Also, from a Bayesian perspective, the form of the generative model and priors is up to the experimenter. From a formal epistemology perspective, you can either try to designate some optimal universal prior, or you can take a variational approach and simply say you want to minimize KL divergence of your predictive posterior from your available data.)
Could you please define "mere" truth?
>From a bayesian perspective, one of our priors must be that there is some probability that the very mind that is doing the thinking may be of a nature that is unable to effectively reason.
Nonetheless, that probability can be very low, putting us firmly in normal epistemological territory and not requiring any "faith". In fact, millions of people regularly make the rational, but clouded, assessment that they are too drunk to drive, and take a taxi home.
(Also, from a Bayesian perspective, the form of the generative model and priors is up to the experimenter. From a formal epistemology perspective, you can either try to designate some optimal universal prior, or you can take a variational approach and simply say you want to minimize KL divergence of your predictive posterior from your available data.)