Monday 10 February 2014

Reply to Red Dog on memes

RDFRS commenter Red Dog claimed recently (here and here) that he hadn't encountered replies by meme advocates to a particular criticism made by Scott Attran in THE TROUBLE WITH MEMES : INFERENCE VERSUS IMITATION IN CULTURAL CREATION.

In that article, Atran Writes:

A general theory of the evolution of replicators under natural selection requires: fecundity and variation, heredity and high-fidelity, longevity and fitness, competition for survival-enhancing resources.
This seems like the same misunderstanding that Boyd and Richerson made of memetics in the article: Replicators are not necessary for cumulative adaptive cultural evolution".

I have already replied to this criticism in: The claim that evolution doesn't require replicators.

"Fidelity", "Fecundity" and "Longevity" were originally presented by Dawkins as attributes of successful replicators, not as prerequisites for evolution.

Dawkins explicitly repudiated them being prerequisites in 1982 - writing:

But a candidate should be regarded as an actual replicator only if it possesses some minimum degree of longevity/fecundity/fidelity (there may be trade-offs among the three).

Atran writes a lot of other things in the 41 pages of the article, prominently including the claim the memetics is "mindblind". I don't have time to go into it all here.

To look more specifically at what Red Dog says:

It's not just that natural selection will favor genes that replicate with high fidelity it's that we wouldn't even get natural selection if the genes didn't replicate with a high degree of fidelity.
Yes you would. Natural selection is ubiquitous.

If memes don't replicate with high fidelity then the concept of "meme as an idea replicator as gene is a replicator for traits" is dead on arrival.
This seems like a confusion. Some genes are copied with high fidelity. Others are not - leading to an error catastrophe / mutational meltdown. Similarly some memes are copied with high fidelity and others are not. Evolutionary theory is flexible enough to work in both cases - it has models in which the mutation rate is a parameter which can take a wide range of values.

When I've heard Dawkins defend the meme to this kind of attack, which has been very little, he never says "memes don't need to replicate with high fidelity"
If it helps at all, I will say that: "Memes don't need to replicate with high fidelity".

I am sure Dawkins would agree that genes and memes need to be high fidelity for the concepts to be viable.

I wrote a 2011 book on memetics. IMHO, neither genes and memes need to be copied with high fidelity for them to count as such. It's just a fact that heritable information in DNA and heritable information in culture are sometimes not copied with high fidelity. You could stop calling them "genes" and "memes" at that point - but then you would lose these terms as a general-purpose basis for evolutionary theory - and you would need new, similar terms to replace them that are more general and can deal with low-fidelity copying. I think that it is simpler, easier and better not to bother with this fiasco - and just use the terms "gene" and "meme" to refer to the more general concept in the first place.

This is pretty-much what most scientists studying cultural evolution have done - although they often use some meme synonym instead.

No comments:

Post a Comment