From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Fri Jul 11 2003 - 05:57:05 MDT
Wei Dai wrote:
>
> This all sounds very reasonable, but there is still a big piece of the
> puzzle missing: why does believing that one is acting for the good of
> the tribe win public support? Why doesn't evolution make the public
> realize that the person is self-deceived (or "evolutionarily deceived")
> when he says that the public good is served by him taking absolute
> power?
>
> We seem to still lack a theory of self-deception that explains why it is
> an evolutionarily stable strategy. Without this I'm starting to have
> doubts on the whole concept. Perhaps the dictator erroneously believing
> that taking absolute power is good for the public is simply a case of
> maladaptation, not self-deception, and in ancestral environments it was
> actually good for the tribe?
Who says it's an evolutionarily stable strategy? Linguistic argument
about politics is pretty recent. We may be looking at a simple attack
that requires a complex defense, one that would require another coupla
hundred thousand years to evolve. Evolution is slow; it plays long, slow
games; and we may just be at a midpoint of it.
Or another possibility: If you've got a mutation that leads you to be
suspicious, and you're the *only* one with that mutation, it ain't
necessarily a reproductive advantage. If anything, people seem to have
actively strong faith in tribal leaders, and given what happens to people
who openly express doubts, I'm guessing that the evolutionary force behind
that emotional faith has nothing to do with the tribal leaders actually
being good guys - more like a heritable tendency that was a reproductive
advantage because its bearers ended up joining the winning side. Think
Nash equilibria.
The thing to remember is not just that some emotions are amazingly enough
adaptive, but that *every* emotion is there because it was adaptive, and
moreover, we have that set of emotions which was given us by a force which
was thinking *only* about the number of surviving grandchildren, and no
other criterion. That is, it's not that human emotions have to be post
facto rationalized as adaptive - it's that we have those emotions, and
only those emotions, that you would expect to find in a design built
purely around adaptiveness. It takes a long while to realize that this is
what you are looking at when you see a human... but the amazing thing is
that it really is true. There is nowhere else that emotions could
possibly come from. The design forces involved are just very... indirect.
-- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Fri Jul 11 2003 - 06:10:20 MDT