An important paper from Nicolas Claidière et al., “How Darwinian is cultural evolution?” appeared recently in Philosophical Transactions of Royal Society B.
Claidière and coauthors adopt a useful schema from the philosopher Peter Godfrey-Smith to distinguish “how darwinian” an explanatory framework can be considered. At the more general level, an explanation can be considered “populational” if considers “a system (such as culture) as a population of relatively autonomous items of different types with the frequencies of types changing over time” (Claidière et al. 2014). A second step is needed to qualify the explanation as “evolutionary”, that is, the frequencies of those types at time t are a function of their frequencies at time t-1. All cultural evolution explanatory frameworks are both populational and evolutionary (I do not consider here vast amounts of socio-cultural anthropology).
Things get more interesting at the next two levels, i.e. the “selectional” and the “replicative”. The former implies that the above types should exhibit variation, heritability, and fitness differences, while the latter adds a further constraint, that is that they should also replicate themselves. While only strict memeticists claim that cultural evolution is darwinian in all four senses, Claidière et al. argue that the correct explanatory framework for cultural evolution is not even the selectional one.
The dispute is about the fidelity of cultural transmission. For memeticists, the problem does not exist, as fidelity is sufficient to consider cultural transmission as a proper process of replication (as it is considered in biological evolution). For “standard” cultural evolutionists, rates of mutations of cultural items are higher than in genetic transmission, but it is still useful to consider in general the cultural success of an item as a result of selection among competing variants. According to Claidière et al. however this is, in the majority of cases, incorrect, as cultural traits do not propagate trough a process of copying (with more or less fidelity) but they are reconstructed each time. For example, I am trying here to report with high fidelity the cultural trait “How Darwinian is cultural evolution?”, but I am certainly modifying it, making it shorter in respect to the original version, focusing on parts that I am particularly interested, misinterpreting and misunderstanding (my fault) what authors wanted really to transmit. Even worst, what you, reader, will recall of this post, will be even different (if anything). All this, for a transmission chain of two or three passages.
Claidière et al. reformulate and extend a concept that the anthropologist Dan Sperber proposes from a long time: cultural evolution forces are better thought in terms of attraction then in terms of selection. The various forces that transform cultural items during the reconstructive processes are not random, but they tend to act in a consistent, even when weak, direction, making cultural outputs converge toward particular “attractors”. Sperber and colleagues generally focuses on cognitive forces. For example, some features of a story are easier to remember than others, and they will serve as attractors when the story is re-constructed. Eric Havelock (about which I read in Rubin, 1995) suggests that this is indeed a function of heroes and gods in epics and ballads that were orally transmitted. Heroes and gods are “bags of attractors” (that’s neither Havelock nor Rubin expression), as they have fixed and concrete features that are easy to image and to remember, which help narrators both to describe abstract concepts and to organise the narration in familiar sequences (Levi-Strauss often cited aphorism on animals “good to think with” should go in a similar direction).
While I am generally sympathetic with this proposal (as well as a Dan Sperber’s fan), I think there are few points worth to discuss. The first concerns the notion of “attraction”, and the fact that is often understood in different ways (and the effort of Claidière et al. to present a clear definition is certainly welcome), being sometimes equated to Boyd and Richerson’s content bias (by me, for example). The second, more interesting, concerns the issue of reconstruction versus
replication preservation. The question whether cultural propagation is better described as one or the other is an empirical one, and the answer varies for different cultural domains. Claidière et al. agree with that, but claims that “a large number” of cultural propagations are results of reconstructive processes. Perhaps. Without entering in technical details on how exactly copying fidelity should be considered (which are anyway useful), there should be a continuum between ideally pure individual innovation and virtually error-free copying (this continuum being wider than in the biological case), and, depending on the domain we consider, we may want to “move” in the Godfrey-Smith’s schema. I’ll try to write more on that soon.
[Much more interesting stuff below in the comments. Thanks in particular to two of the authors of the paper discussed, Thom Scott-Phillips and Dan Sperber, and to Alex Mesoudi and Tim Tyler. Also, the discussion continues here.]