From: John Conover <john@email.johncon.com>
Subject: Re: Re[2]: iefbr14 etc
Date: Thu, 4 Aug 1994 01:48:50 -0700 (PDT)
Raymond Werner writes:
>
> 2. Game theoretic issues: This is the e-mail I like best!
>
The point I was going to make in the "mystery email" was that the game
theorist have proposed some interesting conclusions concerning issues
in social administration. For the most part, game theory has been
developed to address context, or qualitative, descriptions and models
in economics. (Those that believe that mathematics and logic are
quantitative disciplines have some epistemological issues with the
concepts-and there is still some controversy.) After the application
to economic theory, (which it revolutionized) the theory was applied
to the nuclear arms race in an attempt to deduce a "best" strategy, in
how to "play the game" of nuclear armament. (This work was done at
RAND Corporation in the 1950's, and it is not clear how much of the
work was used by policy makers since the documentation to such
decision making is still classified-one thing that is clear is that
J. Von Neumann was at RAND at this time, and he invented game theory,
and was also government advisor.) One of the RAND researchers involved
at the time was a young economist named Kenneth Arrow (now at Sanford
U.,) who was to derive the so-called "Impossibility Theorem," for
which he was to win the Nobel prize. IMHO, it is possibly the crowning
achievement of 20'th century intellectual pursuits (surpassing even
the quantum mechanics, theory of relativity, Godel's Incompleteness
Theorem, etc.).
Here is why I make that statement.
What the theorem states is that the determination of social priorities
by a group of people is intransitive. It means that there exists no
logical process by which a group of people can rank a set of social
priorities in order. Those that have an epistemological commitment to
democracy operating in a logically consistent fashion are kind of
SOL. Bear in mind that this does not mean that democracy won't
work-simply that it can't work in a logical fashion. (If you think
about it, it is astonishing that a logical process could prove this.)
According to Kenneth Arrow, he was not the first to prove this
(although his proof is nice, compact, and general.) It was first
proven by the political philosopher and probability theorist,
Condorcet, in 1785, and was promptly forgotten. (Although, there is
some speculation that those attending the Constitutional Convention
were aware of the issues, since some of the axiomatic methodology used
in the U.S. Constitution addressed the issues of determination of
priorities, eg., balance of power, etc., which were new for the time,
and, obviously clearly understood by the authors of the Constitution.)
It was rediscovered again by Duncan Black in the Journal of Political
Economy, and promptly forgotten. It was then, again, rediscovered by
Abram Bergson in a 1938 paper, where he described the "social welfare
function," and forgotten, until the works by Arrow.
If you consider the implications of the theory, it means that there is
no way that society can operate in a logically consistent fashion. So,
how do you design an administrative system that must operate in a
logically inconsistent environment: the authors at the Constitutional
Convention did an astonishing job. (I think it is important to point
out that they architected not only a democracy, but a liberal
democracy, re: Jefferson's "certain unaliable rights" clause in the
Declaration of Independence, if I remember correctly.)
The implications are significant:
1) We have a great deal of disdain for elected officials, but when
we consider that we send them to D.C. to operate logically (ie.,
represent our priorities, which we consider "logical") in a
non-logical environment ...
2) How can we expect the Congress to operate in a logically
consistent fashion ...
3) Is politics the "engine" by which we determine priorities in a
logically inconsistent environment ... if it is not, how do social
systems determine priorities
4) What about the operators of social administration (ie.,
attorneys,) in these intransitive-they would seem to be caught in
the middle of a bunch of illogical processes
BTW, the game theorists also analyzed dictatorships, etc. The
situation is worse in these cases (you don't have diversity of
priorities in these cases, which is good if the social system is not
complex, a disaster if it is)-just so you won't jump to conclusions.
Note that the "Impossibility Theorem" does not really imply anything
that we do not already know (democracy, congress, social systems,
companies, etc., have never operated in a logical manner.) It just
formalized that they can never do so, so perhaps we should stop
expecting that to be the case.
Just some thoughts,
John
--
John Conover, john@email.johncon.com, http://www.johncon.com/