By integrating social neuroscience with a second-person approach, Schilbach et al. make a significant contribution, at least to neuroscience methodology. A further contribution, we suggest, would be to more fully spell out the implications of a second-person approach for neuroscience. We take it that the second-person approach is a form of a relational, action-based, or developmental systems approach (e.g., Carpendale & Lewis Reference Carpendale, Lewis and Overton2010; Overton Reference Overton and Lerner2006) because Schilbach et al. state that their second-person approach to knowing minds is based on an “interactive account of social cognition that emphasizes the constitutive role of participating in the social world” (sect. 6, para. 3). From this sort of a perspective, knowledge is interactive rather than representational, and this should lead to a different way of thinking about what the brain does. Neuroscience can only explain cognition as well as the psychological theory it is assimilated to, so there is a second and deeper level at which a relational metatheoretical approach to the nature and origin of thinking could transform neuroscience. Schilbach et al., however, are somewhat silent on these implications. We encourage them to be more explicit in order to avoid the problematic assumptions presupposed in the computational view of the mind.
A relational, action-based, approach would re-conceptualize cognitive and social cognitive development in an embodied and situated context of mutual engagement. According to such an account, children construct knowledge by learning the interactive potential of aspects of the world as experienced, which enables them to anticipate the results of their actions and thereby engage in viable interactions with the world. Cognition, then, does not consist of computations performed on internal symbols, but is a process of anticipatory and adaptive action, and develops through the mastery of physical and social interactions. Importantly, if an agent's knowledge depends on its history of situated actions and interactions in its environment, it simply cannot be characterized in terms of internal representations of an outside world (Marshall Reference Marshall2009).
If meaning, communication, language, and thinking cannot emerge independently of social-emotional interaction, then the brain cannot be viewed in the classical computationalist sense as a detached information-processor, which, in any event, is a vacuous claim because it assumes a homunculus (e.g., Heil Reference Heil1981; Tallis Reference Tallis2004). Rather, neural activation and interconnectivity formed through social experience must enable persons to anticipate increasingly complex interactions within a given environment based on their interaction histories, and to elicit more complex social, emotional engagement. Furthermore, the brain must do this without being described as a thinking or social entity in and of itself, or else a second-person neuroscience would ascribe to the parts of an organism cognitive functions that can only properly be ascribed to the whole, thus succumbing to the mereological fallacy that plagues contemporary neuroscience (Bennett & Hacker Reference Bennett and Hacker2003).
Despite the fact that in the target article Schilbach et al. have pioneered new experimental paradigms, admirably critiqued “spectatorial” theories of social cognition (Hutto Reference Hutto2004), and promoted the neuroscientific study of ongoing social interaction dynamics and social synchronies, the danger remains that even in trying to situate brain activity in a sociocultural context, social engagements will once again be reduced to computational manipulations of purely internal representations in an individual, isolated mind, as has generally been the case in classical cognitivist approaches to social cognition (Frith & Frith Reference Frith and Frith2006; Goldman Reference Goldman2006). Indeed, many of the methods advocated by the authors may be applauded by traditional cognitive neuroscientists who have retained the information-processing assumptions of classical cognitive science. For such theorists, the only shift in thinking required by this target article may be that emotional engagement and ongoing interaction must be included in the experimental design in order to isolate the peculiar social cognitive functions that characterize social encounters. Such theorists might even suggest that social cognitive functions could be better localized in the brain by employing two functional magnetic resonance imaging (fMRI) machines to study both individuals in a dyadic interaction, rather than just one. We suggest that Schilbach et al., in crafting experimental methods to be employed in service of a new, second-person framework for neuroscience, have not clarified the framework governing the interpretation of subsequent experimental results, and have inadvertently improved the methods of a traditional neuroscience that continues to operate on a notion of a detached, information-processing individual mind. This leaves open the possibility that these same second-person methods could just as easily be used by traditional neuroscientists to locate specific brain structures in which social cognition supposedly resides, and to interpret those brain structures as sites of the internal symbol manipulation that supposedly constitutes social cognitive functioning. The innovative methods of second-person neuroscience, therefore, do not protect theorists from the mereological fallacy (Bennett & Hacker Reference Bennett and Hacker2003).
A challenge for a second-person neuroscience is getting from the subpersonal level of neural activity to the personal level of meaning, rationality, and normativity. This distinction is generally neglected in neuroscience, perhaps because it seems difficult to bridge this gap from the perspective of a computational approach. From an action-based approach, however, neural activity makes action possible, and meaning emerges as infants learn the interactive potential of their world. Within infants' interpersonal social and emotional engagement human forms of communication emerge, making self-awareness and thinking possible. Neural pathways are structured through experience in social-emotional interaction and play a causal role in interactivity. Although neurons are necessary, we cannot locate rationality and normativity at that level, nor indeed the interpersonal experience of engagement with others that is so important in the second-person approach. Rather, psychological phenomena at the personal level are emergent, and, as Schilbach et al. state, are constituted through participation in the social world. This needs to be made explicit in a second-person neuroscience.
We encourage Schilbach et al. to further integrate the second-person approach with neuroscience. If the traditional, computational view of the individual mind cannot account for the emergence of meaningful communication, and ignores important aspects of embodied and situated social cognition, then it is necessary to adopt a second-person approach informed by an embodied relational metatheory in interpreting the neuroscientific data on social cognition.
By integrating social neuroscience with a second-person approach, Schilbach et al. make a significant contribution, at least to neuroscience methodology. A further contribution, we suggest, would be to more fully spell out the implications of a second-person approach for neuroscience. We take it that the second-person approach is a form of a relational, action-based, or developmental systems approach (e.g., Carpendale & Lewis Reference Carpendale, Lewis and Overton2010; Overton Reference Overton and Lerner2006) because Schilbach et al. state that their second-person approach to knowing minds is based on an “interactive account of social cognition that emphasizes the constitutive role of participating in the social world” (sect. 6, para. 3). From this sort of a perspective, knowledge is interactive rather than representational, and this should lead to a different way of thinking about what the brain does. Neuroscience can only explain cognition as well as the psychological theory it is assimilated to, so there is a second and deeper level at which a relational metatheoretical approach to the nature and origin of thinking could transform neuroscience. Schilbach et al., however, are somewhat silent on these implications. We encourage them to be more explicit in order to avoid the problematic assumptions presupposed in the computational view of the mind.
A relational, action-based, approach would re-conceptualize cognitive and social cognitive development in an embodied and situated context of mutual engagement. According to such an account, children construct knowledge by learning the interactive potential of aspects of the world as experienced, which enables them to anticipate the results of their actions and thereby engage in viable interactions with the world. Cognition, then, does not consist of computations performed on internal symbols, but is a process of anticipatory and adaptive action, and develops through the mastery of physical and social interactions. Importantly, if an agent's knowledge depends on its history of situated actions and interactions in its environment, it simply cannot be characterized in terms of internal representations of an outside world (Marshall Reference Marshall2009).
If meaning, communication, language, and thinking cannot emerge independently of social-emotional interaction, then the brain cannot be viewed in the classical computationalist sense as a detached information-processor, which, in any event, is a vacuous claim because it assumes a homunculus (e.g., Heil Reference Heil1981; Tallis Reference Tallis2004). Rather, neural activation and interconnectivity formed through social experience must enable persons to anticipate increasingly complex interactions within a given environment based on their interaction histories, and to elicit more complex social, emotional engagement. Furthermore, the brain must do this without being described as a thinking or social entity in and of itself, or else a second-person neuroscience would ascribe to the parts of an organism cognitive functions that can only properly be ascribed to the whole, thus succumbing to the mereological fallacy that plagues contemporary neuroscience (Bennett & Hacker Reference Bennett and Hacker2003).
Despite the fact that in the target article Schilbach et al. have pioneered new experimental paradigms, admirably critiqued “spectatorial” theories of social cognition (Hutto Reference Hutto2004), and promoted the neuroscientific study of ongoing social interaction dynamics and social synchronies, the danger remains that even in trying to situate brain activity in a sociocultural context, social engagements will once again be reduced to computational manipulations of purely internal representations in an individual, isolated mind, as has generally been the case in classical cognitivist approaches to social cognition (Frith & Frith Reference Frith and Frith2006; Goldman Reference Goldman2006). Indeed, many of the methods advocated by the authors may be applauded by traditional cognitive neuroscientists who have retained the information-processing assumptions of classical cognitive science. For such theorists, the only shift in thinking required by this target article may be that emotional engagement and ongoing interaction must be included in the experimental design in order to isolate the peculiar social cognitive functions that characterize social encounters. Such theorists might even suggest that social cognitive functions could be better localized in the brain by employing two functional magnetic resonance imaging (fMRI) machines to study both individuals in a dyadic interaction, rather than just one. We suggest that Schilbach et al., in crafting experimental methods to be employed in service of a new, second-person framework for neuroscience, have not clarified the framework governing the interpretation of subsequent experimental results, and have inadvertently improved the methods of a traditional neuroscience that continues to operate on a notion of a detached, information-processing individual mind. This leaves open the possibility that these same second-person methods could just as easily be used by traditional neuroscientists to locate specific brain structures in which social cognition supposedly resides, and to interpret those brain structures as sites of the internal symbol manipulation that supposedly constitutes social cognitive functioning. The innovative methods of second-person neuroscience, therefore, do not protect theorists from the mereological fallacy (Bennett & Hacker Reference Bennett and Hacker2003).
A challenge for a second-person neuroscience is getting from the subpersonal level of neural activity to the personal level of meaning, rationality, and normativity. This distinction is generally neglected in neuroscience, perhaps because it seems difficult to bridge this gap from the perspective of a computational approach. From an action-based approach, however, neural activity makes action possible, and meaning emerges as infants learn the interactive potential of their world. Within infants' interpersonal social and emotional engagement human forms of communication emerge, making self-awareness and thinking possible. Neural pathways are structured through experience in social-emotional interaction and play a causal role in interactivity. Although neurons are necessary, we cannot locate rationality and normativity at that level, nor indeed the interpersonal experience of engagement with others that is so important in the second-person approach. Rather, psychological phenomena at the personal level are emergent, and, as Schilbach et al. state, are constituted through participation in the social world. This needs to be made explicit in a second-person neuroscience.
We encourage Schilbach et al. to further integrate the second-person approach with neuroscience. If the traditional, computational view of the individual mind cannot account for the emergence of meaningful communication, and ignores important aspects of embodied and situated social cognition, then it is necessary to adopt a second-person approach informed by an embodied relational metatheory in interpreting the neuroscientific data on social cognition.