Message-Id: <3.0.1.32.19990228111223.00b05214@popmail.mcs.net>
Date: Sun, 28 Feb 1999 11:12:23 -0600
To: memetics@mmu.ac.uk
From: Aaron Lynch <aaron@mcs.net>
Subject: Re: Generalizing symbolic memetics
I meant to give the reference to the paper in which I introduce the
symbolic representations of memes and their events. It is Lynch, A. (1998)
Units, events and dynamics in memetic evolution. Journal of Memetics -
Evolutionary Models of
Information Transmission 2
http://www.cpm.mmu.ac.uk/jom-emit/1998/vol2/lynch_a.html.
I also wish to point out to anyone trying to give just a cursory read to
the paper that the event diagrams are not representations of propositional
operations going on inside the brain, but transitions going on within the
population. The focus is on developing a population theory, not a theory of
mind.
Prior post:
Derek and others interested in studying behavior or artifacts have raised
objections to the symbolic representations of memory items ("mnemons"), or
any attempt to model a memetics of beliefs, ideas, memory content, or other
internal information. Nevertheless, they have engaged in discussions
labeling individuals as "Hutterite," "atheist," etc., and even discussed
questions of how groups grow or die out. An important part of cultural
selection processes consists of increases or decreases in numbers of people
having (labeled as having) particular cultural attributes. (As a philosophy
of science point, it involves abstractions either way, in that we are still
grouping people together by labeling them as having "the same" attribute.)
So for behaviorists, I wish to point out that symbolic representations are
merely a shorthand for labeling that social scientists already do, and that
such labeling does not inherently have to refer to internal information. It
may thus be generalized to include behaviors with little or no modification
except in the plain English paraphrasing. For example, one can represent
the internal knowledge of how to play baseball with the symbol B and the
internal knowledge of how to play tennis with the symbol T. Someone with
both skills is then represented with the symbol B*T. Yet we could also
label people by patterns of behaviors, and use B as a shorthand for
"baseball player" and T as shorthand for "tennis player."
Conversion events can also be represented:
B*~T + ~B*T --> B*~T + B*T (baseball-playing non-tennis-player plus
non-baseball-playing tennis-player yields baseball-playing
non-tennis-player plus baseball-playing tennis-player.)
Counting such events is a method of tracking the increments and decrements
in the population of those to whom we apply various labels, whether based
on internal information indirectly observed or directly observed behaviors.
The above event increments B*T by 1 while decrementing ~B*T by 1. The rate
of occurrence of the event thus gives information on the rates of change of
populations of labeled individuals--a starting point in population
memetics. The forms of such events and occurrence rates are useful in
constructing computer simulations of movements or developing mathematical
models thereof, to give the actual quantitative population memetics.
In order to move to a formal *general theory* of population memetics, one
needs to identify isomorphisms across different applications. Substitute
basketball for baseball, and ping pong for tennis, and one can write an
isomorphic event diagram. That in turn, implies that a simulation or
equation developed for the baseball/tennis case can be used again for the
basketball/ping pong case--only with different parameter values. So while
the choices of symbols are arbitrary, the purpose in theory development is
not.
Note that one should, expect to re-use population genetic equations for the
simple reason that genetic replication events are not *generally*
isomorphic to memetic replication events. Hence the rationale for devising
new diagrams, new equations, and new simulations.
It is not necessary to become fixated on whether or not we are labeling
people for their internal information or overt behavior in order to proceed
with population memetic computations. We might say that it is
"self-evident" that groups having high rates of certain events will grow,
but that does not substitute for computing how much they will grow as a
function of time and making specific quantitative predictions. Even the
Hutterites, with their huge family sizes, admit of specific calculations
based on timings of births, loss rates, etc.
We might take the term "Hutterite" as a label designating a behavior
pattern rather than a belief, but this does not stop on from using
shorthand such as 2H --> 2H + ~H --> 3H to represent the multi-stage
process of "two people labeled as Hutterite yield two people labeled as
Hutterite plus a baby labeled as non-Hutterite yields three people labeled
as Hutterite."
One can likewise use labels and associated shorthand based, for instance,
on what artifacts one owns. It is, however, only designed for use in
modeling the processes of change in numbers of labeled people, not numbers
of labeled artifacts or labeled behavior instances. So it is not intended,
for instance, to be used in enumerating entities (whether they are called
"memes" or not) that are considered to be instantiated sometimes by
artifacts and sometimes by brains. (People interested in modeling
information items instantiated sometimes or always by artifacts are, of
course, free to develop their own formalisms, propagation equations, and
computer simulations.)
--Aaron Lynch
http://www.mcs.net/~aaron/thoughtcontagion.html
===============================================================
This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)
see: http://www.cpm.mmu.ac.uk/jom-emit