From: Daniel Radetsky (daniel@radray.us)
Date: Sat May 28 2005 - 01:28:03 MDT
On Sat, 28 May 2005 16:46:37 +1200 (NZST)
Marc Geddes <marc_geddes@yahoo.co.nz> wrote:
> This looks good on paper. It does superficially
> appear to be the case that deduction is just a special
> case of Bayesian reasoning.
Not 'superficially.' It *is* the case.
> Some problems though. For one thing, there's more to
> intelligence than just calculation of probabilities -
> there is also the goal system. The thing about
> deduction is that there is more going on than just
> calculation of probabilities. There is a switch to
> descriptions of reality at higher levels of
> organization where classifications are being made that
> go beyond mere associations. This implicitly
> introduces an 'extra' ingredient over and above mere
> calculation of probabilities.
The higher level descriptions are a problem for any system, including
"Geddesian Epistemology," although you have failed to substantiate what that
would be.
> I wouldn't be so quick to embrace the religion of
> Bayes if I were you. Reading more widely in
> philosophy reveals that there are dissenting opinions
> as to whether the Bayesian framework really is the
> ultimate epistemology that it is touted to be. For
> instance I recommend the excellent book 'Theory and
> Reality' by leading philosopher of science
> Godfrey-Smith. He devotes a whole chapter to the
> philosophy of Bayesianism, beginning with this:
>
> 'Although Bayesian is most popular approach to solving
> these problems today, I am not in the Bayesian camp.
> Some parts of Bayesianism are undeniably powerful, but
> I would cautiously put my money on some different
> ideas. These will be introduced at the end of the
> chapter...'
>
> Go read the book if you want to see what ideas he's
> referring to:
Perhaps you should summarize a few of these ideas, so this won't be just a
naked appeal to authority.
Daniel
This archive was generated by hypermail 2.1.5 : Tue Feb 21 2006 - 04:22:56 MST