Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site grkermit.UUCP
Path: utzoo!linus!decvax!genrad!grkermit!chris
From: chris@grkermit.UUCP (Chris Hibbert)
Newsgroups: net.philosophy
Subject: Re: (ir)rationality and (un)selfishness
Message-ID: <670@grkermit.UUCP>
Date: Sun, 6-Nov-83 13:58:05 EST
Article-I.D.: grkermit.670
Posted: Sun Nov  6 13:58:05 1983
Date-Received: Tue, 8-Nov-83 03:16:27 EST
References: <462@dciem.UUCP>
Organization: GenRad Inc., Concord, MA
Lines: 50

...	"...Craver's recent arguments about the Prisoner's Dilemma
	problem were very neatly countered a few months ago by Hofstadter
	in his Scientific American articles. Extrapolating from those
	papers to the real world may be a bit dangerous, but the results
	clearly suggest that altruism is the most selfish course *until you
	discover that your partner/opponent is not altruistic*. Also,
	rational action in a non-zero sum game may frequently look altruistic
	while being selfish in fact. Cooperation gets everyone further than
	obvious selfishness."

	Martin Taylor


I have to disagree with your use of the world "altruism" here.  It's
quite possible that we agree on what you meant, and just disagree about
whether that's the right word.  However, I want to talk a bit about
some of what Hofstadter said, so I'm using this as a starting point.

As they have been used on the net, "cooperation" and "altruism" are not
the same thing.  The "tit for tat" algorithm described by Hofstadter
(see the "Metamagical Themas" column in the July or August (I think)
issues of *Scientific American* for more explanation of "tit for tat"
and its context.) isn't at all altruistic.  It succeeds for pretty much
the same reasons that people give tips in restaurants that are sized
relative to the quality of the service they got.  The reason that
tipping succeeds as a means of feedback (besides common courtesy and
the crowd mentality) is that it is the superrational thing to do.

Hofstadter defines "Superrationality" in the following way:  When a
rational person is deciding on the proper course of action in a
situation in which there are multiple actors he or she realizes that
all the other rational decision makers must come up with the same
plan.  A superrational person is able to make the additional assumption
that all the other actors are rational (and superrational as well.)
This forces the decision to be something that will work, especially when
everyone does the same thing.

Thus, tipping seems to be at least one case where most people act as if
they were superrational even though a superficial evaluation of the
"rational" action suggests a different course.  If you don't expect to
return to a particular restaurant, then the size of your tip can't
effect any future service you might get.  "Rationally", you can
conclude that you can save yourself the cost of the tip with no loss.
Superrationally, you conclude that cooperating with people you will
never have a chance to meet (and formally agree with) pays off for all.

                    Chris Hibbert

{decvax,allegra!linus}!genrad!grkermit!chris
          harpo!eagle!mit-vax!grkermit!chris