Re: Fun with Experimental Design [WAS: Re: The Conjunction Fallacy Fallacy]

From: Michael Vassar (
Date: Mon May 08 2006 - 21:29:07 MDT

My mistake. I should have said "two people who believe one-another to be
rational cannot disagree".
This is old news.
At any rate, it's also something I want to discuss in more detail at the
upcoming AGI conference.
Beliefs about biases in information shouldn't matter in principle, and
better heuristics can make such beliefs irrelevant in practice as well.

For the information of Charles and others, with respect to the updating of
beliefs, though not actions, it seems to me that I have met a grand total of
four people rational enough that they don't have a strong expectation
regarding the direction of disagreement after communicating that
disagreement. It's not much, but it's not zero. Obviously I haven't met
anyone totally debugged of species-typical heuristical weaknesses, but many
people seem to have built module-level systems for pattern matching on
symbolic strings and these seem to have a qualitative impact on general

>From: "Ben Goertzel" <>
>Subject: Re: Fun with Experimental Design [WAS: Re: The Conjunction Fallacy
>Date: Sat, 6 May 2006 13:29:47 -0400
>Michael Vassar wrote:
>>You *can't* rationally disagree with someone who you believe to be
>This is clearly untrue, for a variety of reasons.
>For instance, you might be talking to someone who is rational but
>happens to have formed opinions based on a set of evidence that,
>unknown to him, was highly biased.
>And, when you tell him the evidence his judgments are based on was
>highly biased, he may not believe you ---- because he may believe your
>judgment is flawed because you're not smart enough to assess such
>things, etc...
>In this case, among others, you might disagree with someone who you
>believe to be rational...
>This example is not an exceptional case by any means, since each of us
>has formed our opinions based on different sets of evidence, and none
>of us has time to gather all available evidence...
>-- Ben

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT