BBS peer-commentaries often favour criticism, including my own. Here I start instead with a broad note of approval. I judge Osiurak and Reynaud (O&R) are essentially correct in arguing that pervasive claims for special forms of human social learning constituting the primary explanation for our distinctive cumulative technological cultures have neglected an “elephant in the room”: non-social cognition. In an article cited by O&R, Whiten et al. (Reference Whiten, Horner and Marshall-Pescini2003) reviewed accumulating evidence of chimpanzees learning diverse kinds of object manipulation from others, implying similar capacities in our common cognitive ancestry. This led us to suggest that at the time ancestral hominins left evidence of early cumulative culture in the shape of knapped stone artefacts, “the really significant change … was not so much in social learning mechanisms themselves, but rather in the kinds of behaviour patterns that were invented as increasing brain size made more sophisticated sequences of action possible. Existing capacities for imitation would be brought into play to acquire these from their inventors, for as encephalization made possible more complex innovations, so would it allow the imitator to represent more complex actions for copying” (p. 103).

O&R do the field of cultural evolution a real service by elaborating on this last point, that sophisticated technological cognition may facilitate not only the innovations essential to cumulative culture, but also the representation of novel behaviours performed by others, such that they can be copied. However, not all of the authors' specific propositions are compelling. I highlight some limitations further below.

First, I highlight another observation made in Whiten et al. (Reference Whiten, Horner and Marshall-Pescini2003) that supports the target article's core argument. This refers to hominin Acheulian stone tool making, which persisted with little cumulative change for over 1 million years. Experimental work on the acquisition of the skills necessary to knap Acheulian-type bifaces (e.g., Stout et al. Reference Stout, Rogers, Jaeggi and Semaw2019) indicates this requires at least imitative copying, if not some form of teaching. As Whiten et al. (Reference Whiten, Horner and Marshall-Pescini2003) and earlier, Mithen (Reference Mithen1996a) commented, this implies that such high-fidelity copying was not sufficient to create cumulative lithic culture; what was necessary was the invention of more sophisticated tool-making techniques, thence copied by others.

I wonder, however, why the authors of the target article are not more ambitious in addressing the elephant in the room beyond only cumulative technology? Human cumulative culture has spanned diverse domains, from language to rituals, religions, dance and greeting gestures, to name a few; doesn't the elaboration of these also require both faithful cultural transmission and bouts of innovative non-social cognition?

With this in mind I'd question the authors' reference to “reason” as critical to the non-social cognitive component – does one need to “reason” to innovate a new dance step or a novel ritual flourish? Even considering the authors' insistence (sect. 2.1) that “technical reasoning is critical for the learning of technical content from social interactions as well as for improving this content”, numerous empirical findings appear in conflict, including the claimed unique manifestations in humans. For example in the original finding nowadays referred to as over-imitation, Horner and Whiten (Reference Horner and Whiten2005) discovered that young chimpanzees omitted visibly unnecessary elements of a tool use demonstration when their turn came, thus apparently displaying more “technical reasoning” than children, who blanket-copied both necessary and causally unnecessary elements. Such findings seem at odds with the authors' assertion “that technical-reasoning skills allow humans to extract relevant technical information from a ‘social demonstration’ (imitation), [and] to reject irrelevant information” – in our experiment, the latter is what the chimpanzees did, but the children did not.

This is but one instance in which I suggest more work is needed to more accurately delineate the nature of the non-social cognition that the authors urge us to focus on to better understand cumulative culture. Below there is space to list just a few more, with a focus on human/chimpanzee cognitive comparisons.

In Section 2.1, it is said that humans are the only species to display cognitive transfer skills. What then of cases such as that reported by Tennie et al. (Reference Tennie, Call and Tomasello2010) in which chimpanzees were shown by a human that pouring water from a bottle could lift a peanut in a cylinder, to retrieve it? Later, presented with this problem but no bottle, several chimpanzees transferred their knowledge of getting water from their drinker and spitting it, to instead fill the cylinder by this means: a real and imaginative innovation on their part, achieved by this transfer. In reviewing animal tool use, Visalberghi et al. (Reference Visalberghi, Sabbatini, Taylor, Hunt and Call2017) cite over 60 papers in their section on “transfer.”

It is nevertheless true that chimpanzees often fail to achieve cumulative social learning in contexts where humans do. In the first demonstration of this, young chimpanzees learned a simple technique of using a probe to extract honey from an artificial foraging device, but then failed to adopt a more complex and profitable demonstrated technology, which other chimpanzees who had not learned the simple technique did discover (Marshall-Pescini & Whiten Reference Marshall-Pescini and Whiten2008). The first group thus seemed to get “stuck” on the simple technique, whereas children later tested in a similar context did “step up” and adopt the more sophisticated and productive option after they witnessed it (Whiten et al. Reference Whiten, McGuigan, Marshall-Pescini and Hopper2009). More recent experiments have however revealed some instances of cumulative technology through chimpanzees' combination of discoveries made by others (Vale et al. Reference Vale, Davis, Lambeth, Schapiro and Whiten2017), and where cognitive loads involved in stepping up to a more complex technology are reduced (Davis et al. Reference Davis, Vale, Schapiro, Lambeth and Whiten2016; Reference Davis, Schapiro, Lambeth, Wood and Whiten2018).

The authors have made a valuable contribution to the field of cultural evolution, but more refined, and sometimes more nuanced, understanding of the aspects of non-social cognition key to cumulative culture are called for. Comparative studies focused on cumulative culture issues such as those cited above are clearly relevant, but so may be new syntheses and thinking on such core-related topics as problem solving (Seed & Mayer Reference Seed, Mayer and Call2017), innovation (Kuczaj Reference Kuczaj, Call, Burghardt, Pepperberg, Snowdon and Zentall2017) and causal reasoning (Völter & Call Reference Völter, Call and Call2017).