precis on bayesian rationality
The February 2009 issue of the journal Behavioral and Brain Sciences has a great paper and exchange on bayesian reasoning (pdf, not sure if its gated): “Precis on Bayesian Rationality: The Probabilistic Approach to Human Reasoning,” by Mike Oaksford and Nick Chater. (The paper was published along with 22 short responses — again, I love that format: what a great way to have a meaningful scholarly conversation!)
orgtheory.net 
thanks!
LikeLike
Michael Bishop
May 15, 2009 at 5:46 pm
I’ve a question:
if human reasoning works as Bayesian rationality suggests (i.e. from probabilities, rather than logical deduction), how does “bounded” or ecological rationality come into the picture? Are the two (Bayesian and Bounded rationality) compatible at all?
I’ve read literature that suggest the two are incompatible and mutually exclusive. But it seems to me that they can work together to explain how people reason in conditions of uncertainty. Bayesian consideration of prior information doesn’t have to include perfect (i.e. all) prior information, right? Can greater weightage be allocated to most recent prior information?
LikeLike
Alex
May 16, 2009 at 4:08 am
A nice point. A quick stab — in my mind Bayesian intuition offers one way (among many others) to think about bounded rationality (priors essentially presume BR), and a very useful way to also operationalize it in terms of modeling.
LikeLike
tf
May 16, 2009 at 4:27 pm
Ariel Rubenstein’s _Modeling Bounded Rationality_ treats Bayesian updating briefly. He also notes a lecture by Reinhold Selten in 1989 — “Evolution, Learning, and Economics Behaviour” which I have not seen.
LikeLike
REW
May 16, 2009 at 6:44 pm
REW: Thanks!
LikeLike
tf
May 16, 2009 at 8:38 pm