Vaesen presents a compelling and comprehensive overview of the cognitive abilities underpinning human tool use. Across diverse domains, Vaesen argues for important differences between humans and other primates in all but one. Here we focus on this last domain, body schema plasticity, which Vaesen suggests may not differ substantially between humans and apes. Although we agree that the fact of body schema plasticity characterises both human and non-human primate cognition, recent results have revealed a highly complex relation between plasticity of body representations and tool use in humans. We suggest that there are likely to be fundamental differences in such mechanisms between humans and other primates, with important implications for tool use and its relation to other cognitive abilities. In particular, we focus on two main issues: (1) the time course of plasticity in humans occurs flexibly across multiple time scales, and (2) multiple body representations coexist in the human brain, responding with differential plasticity in the context of tool use, and accounting for the different kinds of experience associated with different types of tools.

Apes and monkeys in the wild rarely use tools spontaneously, and they learn to do so only after long and laborious training (Iriki & Sakura Reference Iriki and Sakura2008). In humans, however, tool use induces plasticity at multiple time scales, showing long-term learning in the case of specific expertise, but also flexibly changing over just a few seconds in experimental situations. For example, some studies have varied tool use on a trial-to-trial basis, finding clear modulation of peripersonal space representations depending on whether or not a tool is used (Holmes et al. Reference Holmes, Calvert and Spence2007) or what length tool is used (Longo & Lourenco Reference Longo and Lourenco2006), demonstrating that tool use induces nearly instantaneous plasticity. Other recent studies have demonstrated long-term plastic changes associated with expertise for specific tools. In blind cane users, for example, merely passively holding the cane extended auditory-tactile interactions along the length of the tool; in control participants, by contrast, active training with the cane was required to induce such extension (Serino et al. Reference Serino, Bassolino, Farnè and Làdavas2007). Analogous findings have been reported for everyday use of the computer mouse (Bassolino et al. Reference Bassolino, Serino, Ubaldi and Làdavas2010): Merely holding a mouse in the hand habitually used to control the mouse (the right) extended auditory-interactions to the space near the screen; whereas such effects were found only when the mouse was actively used, and not just passively held, in the hand not habitually used to control the mouse (the left). These results demonstrate that tool-induced plasticity is highly complex, occurring across multiple time scales and levels of abstraction.

Although the human brain certainly treats wielded tools at some level as if they were extensions of the body, distinctions between the body and tools must also be made, and at several levels. For example, Povinelli et al. (Reference Povinelli, Reaux and Frey2010) rightly point out that one important function of tools is to allow actions that would otherwise be prohibitively dangerous, such as reaching into a fire or stirring a pot of boiling soup. In such cases, effective guidance of the tool may require it being treated as part of the body, even as safety considerations may necessitate it being strongly distinguished from the body. Such conflicting requirements highlight the need for multiple body representations, maintaining parallel, and potentially inconsistent, representations of the body with or without the tool.

This flexibility appears much less pronounced in non-humans primates: In monkeys, long-term tool use trainings induce structural changes in neural body representations, which are rigid and persist whether the animal is tested with the tool or without (Quallo et al. Reference Quallo, Price, Ueno, Asamizuya, Cheng, Lemon and Iriki2009). In humans, conversely, long-term tool use expertise develops multiple body representations, which can be selectively activated depending on the presence/absence of the tool. In blind cane users, for example, peripersonal space representations were extended towards the far space, or limited around the hand (as in sighted subjects), depending on whether blind subjects held their cane during testing (Serino et al. Reference Serino, Bassolino, Farnè and Làdavas2007).

It is also interesting to note that in humans, the subjective experience of wielding a tool is strikingly different from that of illusions, such as the rubber hand illusion (Botvinick & Cohen Reference Botvinick and Cohen1998), in which external physical objects are treated as being part of the body. This dissociation suggests that the tool is “embodied” at a lower, more implicit level, what De Preester and Tsakiris (Reference De Preester and Tsakiris2009) refer to as “body-extension,” distinguishing it from the higher-level, more conscious “body-incorporation” seen in the rubber hand and related illusions. An interesting, intermediate case is that of prosthesis implantation: A prosthesis is a tool, extending action potentialities of an accidentally limited body; but prostheses also replace the shape of the missing limb, hence restructuring the physical body. There seems to be wide variability in amputees' experiences of their prostheses, from those who experience the prosthesis as a corporeal structure to those who consider it an artificial device (Murray Reference Murray2004). It is probable that both functional (level of motor control) and cosmetic (level of anthropomorphism) features of the prosthesis underlie such differences. Recent findings suggest that the sense of ownership over a prosthesis can be enhanced by illusory (Ehrsson et al. Reference Ehrsson, Rosén, Stockselius, Ragnö, Köhler and Lundborg2008) or physical (Marasco et al. Reference Marasco, Kim, Colgate, Peshkin and Kuiken2011) sensory feedback to the stump. This level of abstraction in the experience of body incorporation of artificial objects cannot be investigated in non-human primates (Graziano Reference Graziano1999).

We suggest that different levels of body schema plasticity characterize human cognition and might account for the different experiences associated with the multiplicity of complex tools used by humans in everyday life. These and other recent findings have provided fundamental insight into the role of plasticity of body representations in human tool use. Together, they suggest that body schema plasticity is a highly complex, flexible, and task-dependent process, which should not be thought of as simple “present or absent” in an organism or species. Therefore, we believe Vaesen has too quickly excluded an important role for this factor as an important source of differences between human tool use and that of other primates.