From: janke@unixg.ubc.ca
Date: Fri, 26 Jul 1996 06:46:19 +0800
To: tcmay@got.net (Timothy C. May)
Subject: Re: Schelling Points, Rights, and Game Theory--Part II
In-Reply-To: <ae1bb4100d021004937f@[205.199.118.202]>
Message-ID: <m2g26grv3f.fsf@clouds.heaven.org>
MIME-Version: 1.0
Content-Type: text/plain



Hello, Tim,

I found your essay interesting, but would like to describe a 
hypothetical situation and my ideas of how your notion of Schelling 
points applies to you to see if I am correctly following your ideas:

Suppose that I live in a rural area and 
I know that my neighbour beats his children because I have seen them with 
bruises before and too many times just to be from household accidents. 
Since I am far enough away from him, the beating does not make enough noise
to distrub me from any of my activities. I am also planning on 
moving in three years, so there is little danger that I will be a
victim if the children develop into violent criminals due, in part,
to their abuse. In this case, the "least action" reasoning seems to 
tell me to do nothing. 

On the other hand, the state might do some sort of calculation like
the following: 

(probability the children will become violent criminals) x
(cost of dealing with violent criminals)
-(cost of taking the children from the parent)

to see if intervention is warranted. This is already a simple
application of utilitarianism, however, so that the introduction of the 
notion of Schelling points to explain state intervention seems
unnecessary. 

One the hand, the theory does seem distinguishable from utilitarianism for 
explaining the likely behaviour of the other neighbours of the beater:
Cosinder ones who will be living in the area for quite some time. They
are more likely to later become victims of violent crime (either from the
beater or the children) so would have a greater probability of 
intervening than others. However, it seems that no intervention can be 
justified on notion of Schelling points unless

(probability the children will become a violent criminal) x
(cost of children's crimes to me if they do)
-(cost of person acting now)

is positive. (Of course, people will have diffent estimates of the costs
and probability affecting whether or not they act). A utilitarian, 
however, would have to do the same calculation as the state to determine 
whether or not it was right to act, so would be more likely act,
because the utilitarian needs to consider the cost to everyone, not
just to him or herself. 

Am I following your ideas ok? :) 

-- 
Leonard Janke (janke@unixg.ubc.ca)
NEW pgp key id 0x6BF11645 (0xF4118611 eaten by /dev/fd0 :( )