name_and_address_supplied{at}hotmail.com (Name And Address Supplied) wrote
in news:bnve1j$2smj$1{at}darwin.ediacara.org: 

> William Morse  wrote in message
> news:... 
 
>> I also remember a very interesting 
>> article in Scientific American last year discussing research on
>> policing, in which one person could determine the split of a pot of
>> money between the two subjects and the second person could accept the
>> split (with both receiving the split share) or reject the split (with
>> both receiving nothing). It was made clear that this was a one time
>> interaction, so there was no incentive for the second subject to
>> reject based on a future negotiation with the first subject. Under
>> game theory the first subject should propose a lopsided split in
>> their favor, and the second subject should accept such a split (also
>> note that the amount of money was enough to represent a significant
>> windfall for many of the participants). In fact the first subjects
>> typically proposed splits close to 50:50 and the second subjects
>> rejected splits that were significantly lower than that. The only
>> reasonable explanation for this behavior has to be based on group
>> selection. 
 
> But where are the groups?  As far as I can see, we have a group of two
> individuals, behaving in a manner which reduces the group's success. 
> If group selection is at work, the second player should always accept,
> as their dyad generates a payoff rather than not generating a payoff.

> I assume that you refer to group selection in our past, having shaped
> the human psyche in such a spiteful manner. So really what you are
> suggesting is that the subjects are behaving irrationally, and haven't
> really divorced this artificial situation from the context in which
> they have evolved.  You seem to suggest that at some level they
> perceive the existence of a group which will ultimately benefit from
> this decision.  Why should this perceived beneficiary be a group?  Why
> rule out the possibility that player two, at some irrational level,
> perceives that (s)he will benefit from this course of action due to a
> possible run in with player 1 in the future?  Because player two is
> told by the experimenters that the game is a one-off?  I'm concerned
> that in order to accept your conclusion I have to simultaneously
> believe that humans are irrational products of their evolutionary past
> (i.e. they believe that being spiteful in a two player interaction in
> a research institution can somehow enhance the fitness of their social
> group), yet their rationality reappears so that they able to fully
> comprehend the meaning of a one-off game, even though these games must
> have occurred very infrequently during the social evolution of early
> humans.
> 
> Also, can we really rule out fear of retaliation?  I assume that the
> studies were carried out with complete anonymity, but can the subjects
> really be 100% sure that what their behaviour in that room, and their
> identity, will not be released in such a way as to impair their social
> standing at some point in the future?  Even if there is absolute
> assurance, should the subjects trust the experimenters?  And again,
> this is such an artificial set up, bearing no relation to any kind of
> social encounter in our evolutionary past.  Even if the subject is
> really convinced that there will be no consequences of their actions
> other than the monetary payoff, they may be religious and believe that
> some deity is looking on, ready to pass judgement.

Let me note that Tim Tyler raised some similar objections in a subsequent 
follow, and this reply attempts to address his concerns as well. Also let 
me note that the research discussed in Scientific American was with 
humans. In the lead-in to my discussion of the Scientific American 
article I mentioned research on "fairness" with capuchin monkeys, which 
may have led to some confusion.I don't have any details on the research 
on capuchin monkeys, other than that the monkeys rejected rewards that 
were obviously unfair compared to rewards received for similar work by 
other monkeys. The following comments are limited to the research on 
_humans_ reported on by Scientific American:



You raise two main issues, (apart from that of design of the experiment, 
which I haven't commented on because the article was just a summary of 
the results without details on the protocols).

The first issue (which Tim also raised) was that the behavior was best 
considered as being explained by reciprocal altruism theory, with no need 
to invoke a group and with the explanation that even though the 
experiment was set up as a one-time anonymous interaction, the players 
assumed (or were genetically programmed to assume) that anonymity was not 
guaranteed and that future interactions were likely. 

I don't want to  dismiss the argument that most interactions between 
animals that live in groups will not be one-time - I think this is an 
important influence on behavior.  On the other hand, I think we all 
recognize that we behave very differently in situations where we expect 
either anonymity or limited future interactions. But even assuming no 
anonymity, trying to explain the behavior as enforcement of reciprocity 
between two individuals is unconvincing to me, because the situation did 
not involve failure to pay back a favor. Any split higher than a minimal 
percentage to the second person should satisfy the requirements of 
reciprocal altruism - 20% should be considered generous - but in fact the 
second person often rejected splits less than about 60-40. 

(Actually one of the really fascinating results was that among members of 
one tribe - I don't recall which one - the typical split proposed was 
greater than 50:50 in favor of the second person, who sometimes rejected 
it if the split was too much in their favor!)

  
The second issue you raise is that my explanation for the behavior 
implies "irrationality" (by which you seem to mean social vs. individual 
interest) on one hand combined with "rationality" (purely individual 
interest) on the other hand. But by the logic of group selection,the 
behavior you call "spiteful" (I would call it policing) is not irrational 
- it will help produce fairness for the group as a whole. And another 
result of the research I discussed indicates that people do in fact treat 
such situations rationally (to the extent that we ever do): when the 
second subjects were told that the split was determined randomly or was 
based on relative performance on a test they were much more likely to 
accept highly skewed percentages.
  

Yours,

Bill Morse
---
þ RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info{at}bbsworld.com

---
 * RIMEGate(tm)V10.2áÿ* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 11/3/03 8:20:36 PM