From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Mon Jan 08 2001 - 10:51:23 MST
"Michael M. Butler" wrote:
>
> Well, genetically speaking, the ant colony is pretty much one organism.
> The Queen, and to a lesser extent the eggs, are the most treasured
> "organs". The organism as a whole is selfish as much as nature red in
> tooth and claw can be.
Yes, I know. Again, I'm asking whether anyone knows of a case where it's
adaptive for an ant colony, the reproductive unit, to take advantage of
globally emergent qualities created by the operation of pseudo-capitalist
algorithms deriving from the local pseudo-selfishness of the ant units.
Even demonstrating that worker ants who bring home more Oreo crumbs
receive more nutrients in future scouting expeditions would be a case in
point.
> Leafcutter ants and those aphid milking ones are
> two examples of "capitalism" if you squint just right.
How? (Competition between subunits of a reproductive entity only,
please. Competition between reproductive entities doesn't count.)
Damien Broderick wrote:
>
> But this is surely the canonical or Ur-case of sociobiological explanation.
> The key is degrees of relatedness. Sister ants share 0.75 of their genes,
> not 0.5. This is the basis of Hamilton's kin selection theory and altruism.
> Is this not enough for your purposes?
If worker ants share 0.75 but only the queen reproduces, then, from the
standpoint of kin selection, surely the exact proportion shared is
entirely irrelevant? The worker ants are nullentities, selection
nonfocuses, except insofar as they contribute to the reproductive success
of the queen (or the reproductive success of nearby related colonies, for
kin selection).
Anders Sandberg wrote:
>
> However, ant colonies are definitely not socialist in the usual sense
> since there is no central command system deciding what each ant should
> do. Instead they are based very much on a 'spontaneous' order
> (evolutionarily designed).
Yes. This is exactly why I'm interested; because ant colonies are known
to use emergent properties in their design.
-- -- -- -- --
Eliezer S. Yudkowsky http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:04:39 MST