Humans are often observed using multipurpose smartphones to listen to podcasts, surf the Web, and plan international travel. By contrast, even the most sophisticated non-human primates only use single-purpose tools for situation-specific purposes. Whereas wild chimpanzees, for example, use reeds to fish for termites, they never build tools with multiple components and they never use tools in ways that diverge from the situation-specific purpose for which they were created. Put simply, there are undeniable, significant, and manifest differences in the tool-using behavior of human and non-human primates. Vaesen maintains that such differences are best explained by reference to evolutionarily discontinuous cognitive mechanisms. He argues that our comparative advantage in eight cognitive capacities suffices to establish “a major cognitive discontinuity between us and our closest relatives” (sect. 1). We disagree.

The term capacity has a variety of distinct meanings in the cognitive and biological sciences: It can denote a trait, ability, or mechanism; and although evolutionary pressures sometimes call for the evolution of novel mechanisms, it is generally less expensive to integrate, redeploy, or recalibrate existing mechanisms than it is to build new ones from scratch (Gould & Vrba Reference Gould and Vrba1982; Shubin & Marshall Reference Shubin and Marshall2000; Simon Reference Simon1996). On the related assumption, that permutations “of the old within complex systems can do wonders” (Gould Reference Gould1977), even an evolutionary gradualist can acknowledge unique skills while rejecting appeals to new mechanisms. To establish more than the banality that there are uniquely human traits and abilities, Vaesen must demonstrate that these traits and abilities depend on phylogenetically novel wetware. But we hold that his argument is inconclusive, because it ignores a salient explanatory alternative: namely, the hypothesis that cognitively sophisticated tool use depends not on phylogenetically novel wetware, but on the appropriation of social and environmental scaffolding. We invite Vaesen to consider this explanation, for it is simpler than the appeal to unique mechanisms, and therefore preferable even by his own standards.

To make the case for this, we must note that non-human primates use tools in ways suggestive of several (at least) proto-human cognitive capacities. Wild chimpanzees and capuchins use tools to obtain food that is out of reach, crack nuts with “hammers,” and sponge liquid with leaves (Boesch & Boesch-Achermann Reference Boesch and Boesch-Achermann2000; Fragaszy et al. Reference Fragaszy, Izar, Visalberghi, Ottoni, Gomes and Oliveira2004; Whiten et al. Reference Whiten, Goodall, McGrew, Nishida, Reynolds, Sugiyama, Tutin, Wrangham and Boesch1999); and although neither vervets nor cotton-top tamarins use tools in the wild, both can be trained to do so in the laboratory (Santos et al. Reference Santos, Miller and Hauser2003). Of course, many non-human primates fail to represent the functional properties of their tools (cf. Povinelli Reference Povinelli, Reaux, Theall and Giambrone2000). But wild chimpanzees use different tools at different kinds of termite nests, show selective preferences for different materials, and repeatedly visit nests with reusable tools (Sanz & Morgan Reference Sanz, Morgan, Lonsdorf, Ross and Matsuzawa2010) and recent data suggest that they use multi-functional tools (Boesch et al. Reference Boesch, Head and Robbins2009). Furthermore, captive capuchins can discriminate between functionally appropriate and inappropriate throwing tools (Evans & Westergaard Reference Evans and Westergaard2006); and looking-time methods reveal that cotton-top tamarins and rhesus macaques perceive changes in functional properties as relevant to tool use, but color change as irrelevant (Santos et al. Reference Santos, Miller and Hauser2003; for vervets and lemurs, see Hauser & Santos Reference Hauser, Santos, Laurence and Margolis2007). Finally, repeated experience with tools appears to lead to a more sophisticated understanding of their functionally relevant features (Santos et al. Reference Santos, Miller and Hauser2003, p. 280).

Next, we contend that an ecologically valid approach to cognition requires attending to both the environment in which traits are expressed and the complex relationships between organisms and their embedding environment. Although it is sometimes legitimate and productive to focus on internal mechanisms, cognitive processes (including categorization, inference, and reasoning) are often better understood by reference to coupled organism-environment systems (Hutchins Reference Hutchins2008). Consider two uncontroversial examples: When chimpanzees are trained to exploit abstract, symbolic resources, they show a pronounced increase in executive control and inhibition (Boysen & Berntson Reference Boysen and Berntson1995). Similarly, when human beings supplement their internal capacities for working memory and mathematics with external resources such as pens and paper, we are capable of executing a significantly wider range of computations than we otherwise could (Carruthers Reference Carruthers2002; Rumelhart et al.1986). As an organism's capacities are delineated by the tasks it is able to perform, we contend that many capacities are likely to depend on environmental scaffolding (Barrett Reference Barrett2011; Clark Reference Clark2008).

We suggest that Vaesen should consider the merits of a more ecological approach to uniquely human traits. Relatively minor modifications of primate neuroanatomy (underwritten by the increase in volume of the prefrontal cortex and intimately coupled changes in evolutionary niches) might have enabled humans to cash in on the informational potential of their environments. Both humans and non-human primates engage in acts of epistemic engineering, construct cognitive niches, and exploit the cognitive potential of their environments (Clark Reference Clark2006; Sterelny 2003. This dynamic relationship with the environment can increase the sophistication and power of existing cognitive capacities (cf. Beer Reference Beer2000; Kelso Reference Kelso1995). In many cases, unique capacities are parsimoniously explained by reference to coupled organism-environment systems; and where the environment plays the role of cognitive catalyst, we have little reason to appeal to exotic evolutionary mechanisms (Pinker Reference Pinker2010). Finally, such ecological explanations make sense in light of the highly competitive context in which human cognition evolved. Just as corporations often increase their productive capacities by outsourcing production, human brains increase their computational capacity by outsourcing cognitive operations onto the environment. In both cases, we find a marked increase in capacity without expensive changes in internal structure.

We assume that none of these claims will surprise Vaesen, and moreover, that they are not particularly controversial. So perhaps our disagreements are merely terminological. Not only does capacity have numerous meanings; so, too, does unique. Every species is genetically, behaviorally, and developmentally unique in some respect or other, but Vaesen presumably regards these facts as beside the point. At any rate, we hold that terminological clarity and consistency are crucial for discussions of evolutionary uniqueness. Because we are unsure that Vaesen's treatment lives up to this standard, we would like to see a clearer statement of his key terms. For only in light of such a statement could we accurately evaluate the ultimate force of his arguments.