Teaching your kid how to ride a bike is tricky. Take the skill of leaning into a turn. How much should you lean? Well, that depends: Are you riding on pavement, gravel or grass? Is it wet or dry? Do you have bald racing tyres or off-road tyres? How heavy is your bike? Etc., etc. And the results matter: lean too far and you go down, don’t lean far enough and you won’t make the corner.

For some cultural evolutionists, there is a puzzle here. On one hand, if we are to explain how the practice of bike riding stabilises in a population, we need to point to core features of the skill that are reliably transmitted over generations. On the other hand, as we have just seen, the skill must be flexible enough to accommodate variation in the environment and in the equipment used. So, it seems theories which emphasise the rigidity of the transmission process fail to explain the flexibility required for skills to be useful, whereas theories which emphasise flexibility fail to explain the stabilisation of traditions. As Charbonneau and Sperber put it in their introduction to the volume, while “… a tradition reinvented at every generation is no tradition at all”, if techniques are to be adaptive “… they must allow for some flexibility” (p. xii–xiii).

Understanding the trade-off between flexibility and rigidity is the goal of the volume. According to Charbonneau and Sperber, this issue has been neglected by cultural evolutionists. They diagnose the problem as a “morphocentric” bias (p. xii), in that the focus is on the form and function of material culture rather than the techniques required to produce and use it. These models might explain the evolution of bicycles (from penny-farthing to modern downhill mountain bike), but are ill-suited to explain the evolution of the skills required build and operate them. It’s worth noting, however, that Charbonneau and Sperber perhaps overstate the morphocentric bias. Many textbook examples of cultural evolution, such as food processing (e.g. Boyd, Reference Boyd2017, pp. 9–19), are not subject to this bias.

The volume brings together anthropologists, psychologists, primatologists, archaeologists, historians and philosophers to examine the dynamic between flexibility and rigidity. This nicely illustrates the range of methodologies—from experimental work and fieldwork, to phylogenetic modelling, to theory building—required to tackle such issues. The result is a disciplinarily diverse yet topically focussed volume, which makes an important contribution to foundational questions in our understanding of cultural evolution. Most of the material is intended for a specialist audience, though Charbonneau and Sperber’s introduction and the closing discussion by Kim Sterelny will be more accessible to a general readership. The chapters are usefully organized into three thematic parts: (I) Timescales of Technical Rigidity and Flexibility; (II) From Rigid Copying to Flexible Reconstruction; and (III) Exogenous Factors of Technical Rigidity and Flexibility. Part I acknowledges that understanding the trade-off between rigidity and flexibility will require analysis at multiple timescales, and contains chapters assessing cultural change over a single generation, over centuries and over millennia. Part II examines the effects of rigidity and flexibility on cultural transmission, and critically assesses the notion that high-fidelity copying is the central component in social learning. Part III looks at factors exogenous to standard models of cultural transmission, such as ethnic identity, social influence and environmental change, and how these bear on the operation of rigidity and flexibility.

There is not the scope here to provide a comprehensive overview of the various sub-topics covered by the book. I will instead focus on one, which nicely illustrates the three themes above, as well as the need to address questions of rigidity and flexibility from a broad range of disciplinary perspectives.

In the field of cultural evolution there is something of a consensus view that high-fidelity transmission of cultural capital across generations is made possible via imitation. Relatedly, innovation is understood in terms of copying errors between model and learner, with more adaptive traits being conserved over time (e.g. Boyd & Richerson, Reference Boyd and Richerson2005). In the context of the flexibility/rigidity trade-off, this suggests the ability to innovate (via flexible skills) trades-off against the ability to reliably transmit techniques (via rigid copying). Sadie Tenpas, Manon Schweinfurth and Joseph Call put the point starkly: “…an overly flexible group can lead to cultural breakdown[…]” while “…an overly rigid group can lead to cultural stagnancy” (p. 238). The capacity for innovation is thus in competition with the capacity for transmission.

Chapters 1, 4, and the closing discussion respond to this issue. The underlying thought uniting these responses is as follows: while high-fidelity transmission is undoubtedly important in the evolution of techniques, imitation is not the only way to get there. In many cases, techniques can be reliably transmitted via supervised trial and error learning in socially constructed environments. Think again of teaching your kid how to ride a bike. It’s very hard to instruct a student, either verbally or via demonstration, how to lean into a corner. Instead, the best thing to do is to set them up with a bike, a track, and let them work it out for themselves. And if they end up being a better rider than you, it will not be because they copied your technique in an incorrect but more adaptive way; it will be because they did a better job of performing the skill over iterations of trial-and-error learning. It seems there are many domains where imitation is not the only, nor even the best, way of transmitting expertise.

So how do we test this idea? One important thing to note is that the trade-off between rigidity and flexibility operates at multiple spatiotemporal scales. To paraphrase Charbonneau and Sperber (p. xvii), on one hand rigidly preserved traditions operating over broad timescales may be instantiated by flexibility in use and learning at the scale of an individual lifetime; for instance, the flexible trial and error learning that kids go through in learning how to ride a bike sustains the tradition of bike riding in our population. On the other hand, flexible use and learning at an individual level can be maintained by rigidly preserved traditions operating at broad timescales; for instance, teaching our kids how to ride a bike can be maintained by traditions such as building BMX tracks or training wheels, or indeed by non-material culture, such as the enduring popularity of the Tour de France. This means investigating the processes involved will require employing evolutionary methods at both micro- and macro-timescales.

In their contribution, Valentine Roux, Blandine Bril, Anne-Lise Goujon and Catherine Lara design an innovative field experiment (combining experimental and ethnographic techniques) to investigate how the flexibility/rigidity trade-off operates within a single generation. They looked at the capacity for potters—from India, Ecuador, Kenya, Ethiopia, and Cameroon—to adapt their technique to create novel pot-forms. Valentine and colleagues’ results show that those able to produce the most accurate culturally standardised pot-forms (i.e. experts) also create more accurate reproductions of novel designs. Less capable (but still skilful) practitioners, however, were less able to adapt to new designs. As such, both innovation and high-fidelity transmission look to reside in expertise. But this raises an important question: every society has experts, and hence innovators—why then do we see certain techniques stabilise across generations? Roux and colleagues argue stabilisation occurs because there is social demand for particular artefacts, which leads to experts biasing the techniques that best produce those artefacts. This evidence counts against the commonly held view that there is an inverse relationship between the capacity for a population to innovate and the capacity for a population to produce stable traditions.

Evolutionary archaeologists use phylogenetic methods to analyse long-term trends in cultural evolution (e.g. O’Brien & Lyman, Reference O’Brien and Lyman2000). Typically, this involves assessing similarities and differences in the “traits” of artefacts and then identifying the most parsimonious tree that fits this data. As “traits” are understood as sub-parts of an artefact, evolutionary archaeology is typically “morphocentric”. In an innovative contribution, Sébastien Manem adapts phylogenetic methods to the analysis of techniques; namely the production sequences involved in pottery-making in the Duffaits Bronze Age culture over several millennia. This involves taxonomizing the various stages comprising the production process, which become the “traits” we can then use to produce a phylogeny. In particular, he distinguishes technique—the physical modalities used in the production process—from method—the actual set of operations used in production sequence. His results suggest that techniques and methods contribute to branching in lineages, and that innovation in method is much more dynamic than in technique. Manem analyses this in terms of a “slippery stability”, whereby a traditional technique is conserved, but “slides” to different parts of the pot in response to demand for different forms. This elegantly illustrates, at a macro-evolutionary level, the way “…potters juggled tradition and novelty…” (p. 66).

An essay by Kim Sterelny closes the volume. In his view, over the deep time of hominin cultural evolution, there is a change from the transmission of specific procedures to the transmission of expertise. Consequently, once this transition has taken place, “…there is no trade-off between the reliable transfer of the community’s informational capital and its improvement”, because “[t]he social and cognitive mechanisms that support expertise support both” (p. 253). Sterelny distinguishes techniques applied to the physical and biological environment from those applied in the context of social interaction. The former, he argues, are particularly ill-suited to transmission via imitation. Rather, such skills are better transmitted via socially structured trial and error learning. Ethnographic work on forager education systems lends weight to this claim. Experimentation through hands-on, self-guided learning is prioritised, while the “copy-template” model of skill transmission via imitation looks comparatively rare. Sterelny thinks the problem here is that cultural evolution has traditionally been modelled too closely on genetic evolution. Genetic transmission is largely a template-copying process, as local environmental effects do not modify the genes inherited across a single generation. But not so in the case of skill transmission. Of course, there are good motivations for cultural evolutionists to build models that resemble genetic evolution as closely as possible. The latter is tried and tested, and the less analogous cultural change looks to biological change the more we start to worry if the models will work. However, meeting this heuristic should not come at the expense of denying the reality of cultural learning.

Charbonneau’s volume addresses a key issue in cultural evolution, and nicely illustrates how innovative interdisciplinary methodologies can shed light on the dynamic between flexibility and rigidity. It is highly recommended reading.