[extropy-chat] Re: Overconfidence and meta-rationality

Robin Hanson rhanson at gmu.edu
Sat Mar 12 16:18:36 UTC 2005


At 07:53 PM 3/10/2005, Hal Finney wrote:
>Aumann basically says (neglecting the part about priors) that it is
>impossible for a rational person to believe that he has a persistent
>disagreement with another person whom he believes to be rational, where
>the other person also believes the first person is rational.  Note that
>this says nothing about what the second person actually believes.  It all
>has to do with what the first person believes about the world.  It says
>that a certain combination of beliefs is impossible for a rational person
>to hold.

This is true, given that "rational" means "perfectly rational".  And of
course this is why Aumann's initial result seemed so easy to dismiss --
of course we are not sure that they are sure that we are sure that ...
we are both perfectly rational.  But more recent results are harder to
dismiss this way.  For example, this paper of mine,

<http://hanson.gmu.edu/disagree.pdf>For Bayesian Wannabes, Are 
Disagreements Not About Information?
<http://www.kluweronline.com/issn/0040-5833/>Theory and Decision 
54(2):105-123, March 2003,

allows for arbitrary deviations from perfect rationality.  It says that
if you both are reasonably confident that you both satisfy a few easy to
compute belief constraints, then you have to each believe that the other
person is biased in a certain direction *on average*, while each person
believes themselves to be unbiased.

>This perspective frees us from the competitive aspect of meta-rationality,
>the "mine is bigger than yours" dynamic that sometimes arises in
>discussions of this issue.  It's not a matter of one person being wrong
>in a disagreement, or one person being more meta-rational than another.

Well, but it is in part.  Most dimensions of "bigger" are not directly
relevant. It is not directly about who knows more or thinks faster or
makes fewer cognitive errors.  It is only directly about calibration and
arrogance -  who on average better adjusts their estimates to account for
the fact that other people might be more right than they.  How various
mental "bigger" dimensions correlate with calibration remains an open
question to me.



Robin Hanson  rhanson at gmu.edu  http://hanson.gmu.edu
Assistant Professor of Economics, George Mason University
MSN 1D3, Carow Hall, Fairfax VA 22030-4444
703-993-2326  FAX: 703-993-2323  





More information about the extropy-chat mailing list