Burkart et al. provide a comprehensive and erudite review offering new perspectives on the burgeoning developments in human and nonhuman animal research on intelligence. I particularly welcome the eventual focus on the relatively recently formulated cultural intelligence hypotheses, which I have collaborated in developing (Whiten & van Schaik Reference Whiten and van Schaik2007; and see Whiten [Reference Whiten, Elison and Serain press] for relationships with the broader, earlier research on social intelligence). However, I have a few comments and puzzlements to share.

The first concerns the authors' conclusion in section 1.2.2 that “natural selection for social learning seems to automatically trigger selection on individual learning and general cognitive ability, suggesting that ontogenetic canalization through social learning may have contributed to enabling the evolution of domain-general cognition” (para. 8). This principle is elaborated further in section 3.3, hypothesising that selection for cultural intelligence offers an explanation for the evolution of greater general intelligence in some species, largely because the rewards consequent on the efficiency of learning from experienced others minimise energetic constraints on encephalization required for greater general intelligence. However, this is an “enabling” explanation rather than one positing positive selection on general intelligence through an emphasis on cultural transmission, and I suggest there is something of a paradox here, or at least a conundrum.

The conundrum is that Burkart et al. propose that cultural learning encourages general intelligence, whereas it is common in the social learning literature to assert, to the contrary, that a core adaptive advantage of this form of learning is that it reduces the costly needs of individual learning. Thus, for example, it appears from the restriction of chimpanzees' nut-cracking to only far West Africa that most chimpanzees have insufficient general intelligence to invent the practice, despite availability of the requisite raw materials (excepting at least one rare innovator, at some stage); however, a suite of experiments has shown that naïve chimpanzees (some from East Africa) can learn the skill following observation of a nut-cracker (Whiten Reference Whiten2015). This suggests that most wild chimpanzees in the West achieve the skill via observational learning, removing selection pressure on the general intelligence necessary to invent the skill.

If this is the case, it suggests that Burkart et al. have more work to do to specify just exactly what aspects of general intelligence they propose may be selected for in such scenarios. They mention practice in this context, which is certainly protracted in the example of nut-cracking (Whiten Reference Whiten2015). But the practice involved in perfecting nut-cracking learned from others seems rather far from the definition that “general intelligence, as defined in either humans or nonhuman animals, stresses reasoning ability and behavioural flexibility” (sect. 1.1, para. 1). In the human case, the phenomenon of “over-imitation,” in which children (apparently unlike other apes) copy others' visibly causally irrelevant actions suggests a marked relinquishing of reasoning and flexibility, commonly interpreted as a correlate of our species' extreme reliance on cultural transmission (Whiten et al. Reference Whiten, McGuigan, Marshall-Pescini and Hopper2009).

Does the authors' emphasis on the potential knock-on effects of cultural intelligence on general intelligence perhaps neglect the direct effects of selection for cultural transmission encouraging other, socio-cognitive enhancements with implications for encephalization? The cultural intelligence hypothesis was originally developed to explain the encephalization and intelligence of the great apes (Whiten & van Schaik Reference Whiten and van Schaik2007), which was not accounted for by broader social intelligence theories that work well for primates in general (Dunbar & Shultz Reference Dunbar and Shultz2007a). Consistent with this, a recent study reported multiple-tradition cultures for gorillas (Robbins et al. Reference Robbins, Ando, Fawcett, Gruiter, Hedwig, Iwata, Lodwick, Masi, Salmi, Stoinski, Todd, Vercellio and Yamagiwa2016) that are consistent with those earlier described for chimpanzees (Whiten et al. Reference Whiten, Goodall, McGrew, Nishida, Reynolds, Sugiyama, Tutin, Wrangham and Boesch1999) and orangutans (van Schaik et al. Reference van Schaik, Ancrenaz, Borgen, Galdikas, Knott, Singleton, Suzuki, Suci and Merrill2003) and appear rich compared to the putative cultures of other animals, although a parallel analysis for spider monkeys, in some ways a New World chimpanzee-like niche, reported a quite similar complexity (Santorelli et al. Reference Santorelli, Schaffner, Campbell, Notman, Pavelka, Weghorst and Aureli2011), and studies of capuchin monkeys suggest something similar may await systematic assessment (e.g., Coehlo et al. Reference Coehlo, Falotico, Ozar, Mannu, Resende, Siqueira and Otonni2015) . Both the latter species are relatively encephalized, and of course the same is true for cetaceans for which a strong evidential case has been made for multiple-tradition cultures including foraging techniques, migration routes, and song (Whitehead & Rendell Reference Whitehead and Rendell2015). The social learning capacities of such animals with heavy dependence on extended cultural repertoires may themselves need to be cognitively sophisticated, including imitative and emulative processes, with neural demands (Whiten Reference Whiten2017; Reference Whiten, Elison and Serain press). In addition, encephalization may be extended simply to facilitate the storage of a greater cultural repertoire. In the human case, this may be very significant when one contemplates the vast scope of the cultural information we assimilate, from language to all aspects of social and material culture.

The latter leads to a related but different comment. The authors tend to run together two threads in the literature when referring to “the cultural intelligence hypothesis” (sect. 3.3, para. 3), and I think it would reduce potential confusion to separate these. The first thread is exemplified by the writings of Tomasello et al. that are cited, such as Moll and Tomasello (Reference Moll and Tomasello2007). The second thread is exemplified by the writings of van Schaik et al. (e.g., Whiten & van Schaik Reference Whiten and van Schaik2007; van Schaik & Burkart Reference van Schaik and Burkart2011). It is this second thread that sets out a cultural intelligence hypothesis addressed originally to the problem of great ape intelligence and encephalization, but in principle relevant to any relevant animal species. By contrast, the first thread was specifically concerned with what makes humans different from all other animals, and was originally and appropriately dubbed “the Vygotskian intelligence hypothesis” (Moll & Tomasello Reference Moll and Tomasello2007, p. 639). This made sense to me, until Herrmann et al. (Reference Herrmann, Call, Hernández-Lloreda, Hare and Tomasello2007) then referred to these ideas as “the cultural intelligence hypothesis” (p. 1360). This was potentially quite confusing insofar as the argument was that it did not apply to nonhuman species. I feel it is important to recognise these differences, whether that is achieved by reverting to the “Vygotskian” tag to distinguish the “human” focused version. This is not, of course, to deny that there is a potentially important linkage between the sets of ideas embedded in these two theories.