Re: reductionism in memetics

Rob Clewley (
Thu, 21 Jan 1999 16:10:16 +0000 (GMT)

Date: Thu, 21 Jan 1999 16:10:16 +0000 (GMT)
From: Rob Clewley <>
Subject: Re: reductionism in memetics
In-Reply-To: <000201be44d1$8a29c740$ea293fce@rbrodie>


Paul Marsden just posted some good stuff about reductionism in memetics,
but I thought I'd like to add a couple of specific points (especially
since he mentioned my recent work) that Richard Brodie and If Price

On Wed, 20 Jan 1999, Richard Brodie wrote:

> I am excited by the interaction of the "software" of the mind --
> memes -- with the "hardware" layer that is being called evolutionary
> psychology. I see that all kinds of really interesting predictions can be
> made from just that model, with no understanding whatsoever of how ideas are
> stored neurologically.

Of course interesting predictions can be made from such "single-level"
models, but they do not necessarily have an ounce of accuracy or use
unless they are very carefully tied in to the understanding of the
"lower levels"...

Richard Brodie also wrote:

> Now that in itself is a fascinating field, but not
> necessary for working in memetics any more than an understanding of chip
> electronics is necessary for designing a web page.

..... even in your computing example, I think my point still holds true,
because your analogy does not do the subtly of the problem justice.
There are many "levels" of analysis at which we may attempt to
understand human behaviour. Each are appropriate on different occasions,
and even though they can never quite be arranged in a perfect hierarchy
(there is overlap), there is at least some sense in which some levels
are "closer" to others. Closeness means something like "degree of
abstractness of representation". Now it is indeed absurd that
understanding quantum physics will help us understand human behaviour,
certainly not in the near future. This I think corresponds to the story
for microelectronics vs. web-page design, as you suggest.

But understanding the genetic constraints on our perceptual and
cognitive neural systems (e.g. by evolutionary psychology) *will* help
us understand the nature of social interactions. This is taking two
levels of representation much "closer" together, and much better suited
for comparison. In your computer analogy, this would be more like
knowing the operating system so that we might understand the
system-specific constraints or extra features of high-level programming
language packages.

You say you are excited by the interactions between the social level and
the evolutionary psychology level, so surely you acknowledge that evol.
psych. constrains the form of social interactions (loosely speaking). So
why do you then claim that we can proceed in memetics without any
understanding of lower levels?

And If Price wrote:

> I am starting to wonder (work submitted to JoM-EMIT) whether we
> should see 'choosing' - making different responses to environmental
> stimuli - and 'learning' as basically part of the hardware - a good
> trick in genetic possibility space, and 'stuckness' (all the stuff of
> group think, social contagion, mental models etc) as memetic.

I think any model that bases its concepts in the software-hardware
analogy to the mind-brain will soon find itself unable to fit into the
"graded" picture that is developing of higher brain function from the
neurosciences and cognitive psychology (etc.). The (dis)analogy has
caused artificial intelligence crippling problems in its attempts to
understand human minds (as part of cognitive science) and it will do the
same for memetics unless we avoid the laziness of a dichotomic
differentiation between "high-level" and "lower-level" brain function.

The problems with the s/w-h/w analogy have been written about before,
but the falacy persists and so needs re-addressing regularly! I discuss
the problems with particular reference to memetics (along with some
related points) in my forthcoming paper. A version awaiting a few
stylistic changes is on my web page, before submission to JoM. However,
the section relevant to my points here appears in its final form.


Rob Clewley.

Applied Nonlinear Mathematics, and Neural Computing Groups,
Dept. of Engineering Maths, University of Bristol,
Queen's Building, University Walk, Bristol
BS8 1TR, United Kingdom.

(+44) (0)117 928 9798 / 9743
( Fax ) (+44) (0)117 925 1154

"Most general statements are false"

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)