From: John Conover <john@email.johncon.com>
Subject: Re: EMT - Product and mtkg meeting Thurs 1/11
Date: Wed, 10 Jan 1996 13:10:23 -0800
Tim Wood writes:
>
> Real interesting stuff you've passed along the last few days John.
> Would you elaborate on "(in)transitivity of priorities?" Does it
> have to do with dependencies among priorities or with the subjective
> thought processes that determine them?
>
Kind of. What it means is that there can never exist a process (that
is finite, and logically consistent,) that can order priorities,
(specifically, if there are more than two priorities at issue.) See
Kenneth Arrow, and the Impossibility Theorem. For an introductory
approach, see:
@book{Hoffman,
address = "New York, New York",
author = "Paul Hoffman",
publisher = "Fawcett Crest",
title = "Archimedes' Revenge",
year = 1993}
chapters 6 and 7. Arrow was an economist that was working on
optimization of the social welfare function by game-theoretic means at
RAND in the late 50s. He proved that priorities, (when there are 3, or
more,) can not be ordered in a society, (ie., they are intransitive
since we can never determine that priority 1 < priority 2 ...) He got
the Nobel for it in the early 60's. In general, the game-theoretic
outcome of an attempt to order priorities is that everyone will get
their least favorable priority as their first priority.
For a more in depth approach, see:
@book{Luce,
address = "New York, New York",
author = "R. Duncan Luce and Howard Raiffa",
publisher = "John Wiley & Sons",
title = "Games and Decisions",
year = 1957}
There is a lot of empirical evidence taken in the 60's by RAND (from
the Congressional records and personal interviews,) that the theorem
has merit, BTW. Kind of puts the Willy/Newt arguments into
perspective. Usually, when you get into these scenarios of priority
determination, it is that you are not thinking at a high enough
strategic level, BTW.
John
--
John Conover, john@email.johncon.com, http://www.johncon.com/