Re: Memes and Things

John Wilkins (
Fri, 29 Jan 1999 10:09:11 +1000

Date: Fri, 29 Jan 1999 10:09:11 +1000
From: John Wilkins <>
Subject: Re: Memes and Things
In-Reply-To: <>

"Hans-Cees Speel" <> wrote
>> Some ideas on the perennial question:
>> 1. The meme is not a thing, it is a relationship of similarity between a
>> number of things. It is, in other words a statistical concept which
>> classifies behaviour, artefacts or events in terms of their recurrence
>> (their familiarity). This is the only way we can identify it or even speak
>> of it.
>In my book a meme is not just similar, but also a unit of heridity. If
>this is not so, I do not call it a meme. Just my thoughts, but
>important I think, if you want to stay in the evolutionary paradigm

This goes to the heart of a major problem in modern evolutionary theory: the
debate between morphological and genetic conceptions of evolution (plus units
of selection, reductionism versus holism, yadda, yadda). In the 1960s, Sokal
and Sneath introduced a cluster analysis of phenotypic traits method of
classifying species (what they called Operational taxonomic Units), later
called "phenetics". The problem with morphological similarities is that the
resulting groups are highly sensitive to the traits selected. To use a crude
example, if you measure femur length, you may get one group, but measure hair
distribution and you may get another. Phenetics as a movement was eventually
supplanted by a different technique, known as "cladistics", which treats the
morphological data as markers of Mendelian genes, and so selects data that is
less sensitive and depends more on the systematist's knowledge of the taxon.
In fact, cladistic analyses are now common on molecular data. But I digress.

The issue here is whether we must treat memes as high fidelity units of
heredity, because Darwinian evolution requires these "replicators" in order to
proceed (otherwise, it's not Darwinian evolution). In fact, I would now argue
that the notion of a replicator is a *neo-*Darwinian concept (due to the work
initially of Weismann and Fisher), and while sufficient to cause Darwinian
evolution, is not necessary. One case of ordinary Darwinian evolution -
selection without replicators but with loose reproducers - would be the
initial selection on autocatalytic hypercycles at the beginnings of life. In
other words, the first "living" systems did not have a gene-phenotype split.
It had to evolve, quite Darwinianly, in the absence of high fidelity

Now, I know that Dawkins and Hull think that memetic evolution is replicator
based. Certainly Hull's book focuses on easy to replicate and distinguish
units of transmission (scientific theories and parts of theories, citations).
But a lot of culture has always struck me as hard to characterise in this way.
The epidemiological notion of transmission is often based on the lack of hard
replicators (I'm thinking here of the views of Dan Sperber [Sperber, 1995,
1996], who argues that culture is transformed when transmitted; ie, is low
fidelity). I cannot think of an easy and nonsubjective way to characterise the
way kids teach each other how to do yoyo moves, for example. What's the meme
there? They copy each other, and the copy is low fidelity because the
standards for assessing a "good" move are varibale from schoolyard to
schoolyard. Yet, yoyoing evolves (I couldn't do what today's kids do - it's
way past what yoyos were supposed to do when I was a lad).

So, I think that there is a sense in which a type is low fidelity but
transmitted without a replicator, and that this is still within the
evolutionary paradigm, so construed. I also think that there is an operational
issue: we can best investigate the high fidelity replicators, and that is
where we should start, as I argue in my forthcoming JoM-EMIT reply to

Sperber, Dan. 1996. Explaining culture: a naturalistic approach. Oxford, UK;
Cambridge, Mass.: Blackwell. Chapter 5.

Sperber, Dan, David Premack, and Ann James Premack, eds. 1995. Causal
cognition: a multidisciplinary debate. Oxford, UK: New York: Clarendon Press;
Oxford University Press. Sperber's response to Atran.


John Wilkins, Head, Graphic Production The Walter and Eliza Hall Institute of Medical Research Melbourne, Australia <mailto:wilkins@WEHI.EDU.AU><> Homo homini aut deus aut lupus - Erasmus of Rotterdam

=============================================================== 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: