Friday, 30 September 2011

Tim Tyler: Can memetics predict meme fitnesses?


Hi! I'm Tim Tyler, and this video addresses a common objection to memetics. The idea is that "survival of the fittest" being a circular tautology is a problem for memetics, because we lack a predictive theory of meme fitnesses.

One fairly common objection to memetics is that it is vulnerable to the idea that "survival of the fittest" is a tautology - since the fittest are defined as being those that survive - and that is a circular argument.

David S. Wilson makes this point in an article titled "Flying Over Uncharted Territory" - which was a review of Blackmore’s book The Meme Machine. He says:
The oft-repeated accusation that natural selection is a tautology fails because fitness is not defined in terms of whatever evolves but in terms of the properties that enable organisms to survive and reproduce in their environments. Moths that are colored to match their background have a high fitness with respect to bird predation, but cryptic coloration may not evolve if the appropriate mutations either do not arise or are lost by genetic drift. The ability to define fitness independently of what evolves saves the concept of natural selection from being a tautology. For the meme concept to escape the same problem, we must define cultural fitness independently of what evolves. If the first four notes of Beethoven's fifth is a powerful meme only because it is common, we have achieved no insight.
Serenely and Griffiths (1999 p.334) also voice this complaint, saying:
With the possible exception of scientific ideas, we have no explanation of the nature of the fitness of ideas, nor do we typically understand why they differ in fitness. We can call a tune "a meme with high replication potential" rather than "catchy" if we like. But without source laws, this adds nothing to our understanding of musical trends.
Maria Kronfeldner also gives this objection in her 300 page critique of memetics. In her summary (p.290) she writes:
Either the analogy is heuristically trivial, because it loses its main claim, namely that memetics presents an alternative to the traditional explanation, which is given in terms of properties and interests of humans, or the explanatory units of selection analogy is trivial in explanatory terms, because it is tautological – it does not explain anything, since it merely states that those memes that have a high actual survival are those memes that have a high propensity for survival, without explaining where this high fitness emerges from.
Massimo Pigliucci gave much the same objection in a recent podcast:

[Massimo Pigliucci speaks]

We do, in fact know some things about which memes are fitter than other ones. This area is part of what is often called "Applied Memetics" (an area which also includes memetic engineering). It seems to be one of the better studied areas of the field to me - simply because social media marketing and advertising departments need to be able to predict what spreads and what doesn't in order to be able to rapidly construct successful viral marketing media without the expense of doing a lot of testing. Popular writers and musicians know what people are likely to share. Scientists often have some idea about what scientific ideas are more likely to persist. Good editors know what will sell and what will not. Military propaganda creators know which rumors are most likely to spread. Often meme fitness variations are obvious - for example, if you put "Lady Gaga" in the title of a video you create, it is likely to be viewed more than if you use "Joe Bloggs" instead. However, the area of what makes ideas spread has also been studied by social scientists.

Francis Heylighen's 1998 paper "What makes a meme successful? Selection criteria for cultural evolution" gives an early introduction to the field - breaking the meme lifecycle down into assimilation, retention, expression and transmission and observing that successful memes need strengths in each area. More recently, there have been many studies of social media on the internet that look into why memes spread. For instance see, Dan Zarella's "The Science of ReTweets" or his "The 8 Elements of Contagious Ideas".

Of course we don't know everything about this field, but it is often a challenge to determine the fitness of genes without testing them in organic ecology too.

Even if it was true that social scientists were totally clueless about what spreads and what does not - which is far from the case - then the correct attitude would not be to declare the whole enterprise of predicting meme fitnesses to be hopeless, but rather to actually do some field work - and figure out the factors that make some ideas spread while other ones do not.

The objection that "survival of the fittest" is a tautology is exactly as much of a fallacy in memetic evolution as it is in organic evolution for exactly the same reasons.


The rest of Massimo Pigliucci's critical comments are here.

For another example of this criticism, see: Jerry Coyne's review of The Meme Machine:

memetics seems completely tautological, unable to explain why a meme spreads except by asserting, post-facto, that it had qualities enabling it to spread. One might as well say that aspirin relieves pain because of its pain-relieving properties. The most interesting question — why some memes spread and not others — is completely neglected. Why did Christianity take hold during the waning days of the Roman Empire? You won’t find the answer,or any way to attain it, in memetics. (This, by the way, makes memetics utterly unlike biological evolution. The spread of genes through natural selection is not tautological because one can predict their fate through their known effects on replication and the reproduction of their carriers.)
Edouard Machery offers a similar complaint:
First, and most important, memetics has usually very little explanatory power. Too often, memetic explanations boil down to the uninformative claim that a given meme has spread in a population because it has reproduced itself successfully. The essential issue, What makes it the case that a cultural variant is preferentially transmitted?, is often dodged, or, when it is not, is addressed with sheer speculation.

There are now some nice books on the topic of why some ideas spread, while others do not:

I've a feeling that this is another criticism of memetics that simple progress will firmly assign to the dustbin of history.

No comments:

Post a Comment