[FoRK] A Bayesian framework for faith vs. non-faith

Stephen D. Williams < sdw at lig.net > on > Sun Nov 19 17:47:09 PST 2006

Jeff Bone wrote:
> On Nov 18, 2006, at 11:05 AM, Stephen D. Williams wrote:
>> Science != faith.
> +1. But isn't the driver for this discussion the need to identify how 
> they differ?
> So --- I don't have a fully fleshed-out framework for that, at least 
> one that I'm entirely comfortable with. ALL modalities of thought 
> proceed from some axiomatic assumptions. However, intuitively it seems 
> obvious to me how faith and science differ. Faith involves axioms or 
> other propositions for which otherwise reasonable people could 
> entirely legitimately assign different prior probabilities. It is 
> "tainted" (ala Perl variables, if you will ;-) by the existence of 
> such arbitrary probabilities. Thus conclusions from Bayesian or other 
> probabilistic or quantitative reasoning over such values, no matter 
> how rigorous, are not reliable. That is not to say that such 
> conclusions may not be *true* --- merely that they are not as reliable 
> as conclusions from "untainted" or minimally tainted priors.
> Science, on the other hand, rejects (to the greatest extent possible) 
> such arbitrary, subjective priors, and thus yields conclusions with 
> greater reliability. It further admits a willingness to discard any 
> axioms or priors if they are revealed to be of this arbitrary nature 
> --- i.e., it is a meta-method of reasoning, rather than a body of 
> beliefs or a method of reasoning per se --- a flexibility NOT shared 
> by faith-based modalities.
+1 right back atcha.
That is a nice, clean description. I'm sure that I will quote you some day.

> Just a thought...
> jb
> PS - it's entirely possible to take faith-based thinking and 
> deceitfully (or delusionally) wrap it in the cloth of quantitative, 
> even Bayesian reasoning. One good example of this is Stephen D. 
> Unwin's book The Probability of God. Check it out to see what I mean...
> _______________________________________________
> FoRK mailing list
> http://xent.com/mailman/listinfo/fork

More information about the FoRK mailing list