Schilbach et al. argue that their “second-person” approach to social cognition – which emphasizes the centrality of interaction and emotional engagement in facilitating social understanding – differs fundamentally from what they term “spectator theories.” Under the latter heading they include not just “Theory Theory” (TT) and “Simulation Theory” (ST), but also what they aptly term the “social perception” approach. Although we suspect similar questions could be raised with respect to TT and ST, we focus on the question of whether it is true that the social perception approach is “committed to an observer epistemology” (sect 1.2, para. 2) and thus is a “spectator theory.”

The social perception approach comes in different forms (see, e.g., Duddington Reference Duddington1918; Gallagher Reference Gallagher2008; Hobson Reference Hobson2008; Scheler Reference Scheler and Heath1954; Stout Reference Stout2012; Wittgenstein Reference Wittgenstein, Luckhardt and Aue1980). Common to all is the idea that we can sometimes perceive (and thus need not infer) the mental states of others (Krueger Reference Krueger2012; Krueger & Overgaard Reference Krueger, Overgaard, Miguens and Preyer2012; Overgaard Reference Overgaard and Zahavi2012). But contra Schilbach et al., it is not clear that this commits the social perception approach to a “spectator theory of how we relate to other people” (sect 1.2, para. 2). For, first of all, there is no reason to construe the social perception theory as a general theory of “how we relate to other people.” Not only is the theory fully consistent with the idea that social life is more a matter of interacting with, than merely observing others. Additionally, it can also maintain that when we perceive others' emotions and goals, say, we typically do so in order to facilitate or enable interacting with them. For example, I may perceive anger in another person's facial expressions and gestures; the former expresses specific features of the anger whereas the latter articulates its intensity and level of arousal (Ekman Reference Ekman1965). But this need not involve a static observational stance. Rather, I see this anger as affording (or perhaps constraining) further interactive possibilities. My social perception is thus typically part of an ongoing process in which I use what I see in others to further negotiate my engagement with them.

More importantly, the social perception account can maintain that the ability to perceive others' mental states may be a capacity that is developed and/or refined within interaction, for example, the face-to-face interaction characterizing early infant-caregiver exchanges (see, e.g., Hobson Reference Hobson2002; Legerstee Reference Legerstee2005). Moreover, even in adult perceivers, our exercise of the ability may be greatly facilitated by current interaction with others – and may in some cases be hindered in the absence of interaction (see, e.g., Gallagher Reference Gallagher2008). For example, certain socially-salient details – gestures, intonation, posture, shifts of spatial proximity, and so forth – may only become available for perception as an interaction unfolds over time. And when we perceive facial gestures (e.g., a smile or frown), we automatically attune to these gestures by involuntarily mimicking them (Dimberg & Thunberg Reference Dimberg and Thunberg1998). As one of the target article authors notes in another paper, this mimetic response seems to indicate that “the process of perceiving faces always includes an ‘enactive’ element through which we engage with and respond to stimuli instead of a mere ‘passive’ perception of face-based cues” (Schilbach et al. Reference Schilbach, Eickhoff, Mojzisch and Vogeley2008a, p. 45). The social perception approach is fully compatible with this observation.

But Schilbach et al. are right that at least standard versions of the social perception approach suggest that social cognition is “something that could in principle occur in the presence of a one-way mirror, where a detached observer reads out the mental states of another person” (sect. 2.1, para. 2). However, it seems highly plausible that “detached” social perception of the “one-way mirror” sort could in principle occur; indeed, it might be something that happens fairly frequently – as, for example, when we perceive the emotions expressed by people on TV, while watching one's fellow diners in a restaurant, or when viewing others from a distance.

So Schilbach et al. are faced with the following dilemma. Either they maintain that social perception is in principle impossible in the absence of current social interaction, as, for example, in the case of the one-way mirror. This is a very strong claim: A single instance of social perception without interaction (e.g., successfully perceiving a diner's expression of happiness as such across a crowded restaurant) would suffice to falsify it. Or else they acknowledge that the latter is a possibility. But then they are wrong to suggest that their “second-person” approach is distinguished from the social perception approach by the “spectatorial” nature of the latter. For, the sort of purely “spectatorial” social cognition that the social perception approach as such makes room for (e.g., recognizing emotional expressions at a distance or in the absence of interaction) is one that Schilbach et al. themselves want to accommodate.

In our view, it is crucial to recognize the broad scope of our social competence, encompassing situations in which we are active participants, as well as situations in which we are passive bystanders. Despite their justified emphasis on the primacy of interaction, Schilbach et al. recognize this complexity and attempt to develop their view accordingly. However, in doing so, they fail to show that the social perception approach to social cognition is genuinely committed to the sort of “observer epistemology” they see their own view as opposing.

Although we have here focused on the social perception approach, our commentary also raises the more general question of whether the theoretical commitments of existing approaches really are so opposed to the “second-person” approach championed by Schilbach et al. Defenders of TT, for example, have long emphasized the importance of social interaction for the development of desire and belief understanding (e.g., Repacholi & Gopnik Reference Repacholi and Gopnik1997; Ruffman et al. Reference Ruffman, Perner, Naito, Parkin and Clements1998). “Development in social cognition depends on two-way traffic between self and other,” as Meltzoff et al. (Reference Meltzoff, Gopnik, Repacholi, Zelazo, Astington and Olson1999, p. 19) put it more than a decade ago. So perhaps it is not just the social perception approach that in reality is less “spectatorial” than Schilbach et al. make it appear to be.