Spelke (Reference Spelke2022) presents an intriguing account of the ontogenetic origins of human knowledge by reviewing research on human infants using diverse methods. Although the evidence put forward to support core knowledge systems in the physical domain, including places, objects, number, and geometry, is compelling, research examining the neuroscience of social cognition and especially the neural bases of joint attention in infancy has been overlooked. Specifically, there now exists a mounting body of evidence, showing that young infants, like adults, recruit medial prefrontal cortical regions supporting sophisticated social-cognitive functions, including joint (triadic) engagement (see Grossmann, Reference Grossmann2013, Reference Grossmann2015, for reviews). In adults, medial prefrontal cortex (mPFC) has been shown to play a key role in attributing and interpreting mental states (Amodio & Frith, Reference Amodio and Frith2006). There also is behavioral evidence suggesting that already newborns show a basic sensitivity to eye gaze cueing of object locations foundational to joint attention (Farroni, Pividori, Simion, Massaccesi, & Johnson, Reference Farroni, Pividori, Simion, Massaccesi and Johnson2004) and that, at least by 3 months of age, infants discriminate between dyadic and triadic social interactions (Striano & Stahl, Reference Striano and Stahl2005). First insights into the neural underpinnings of joint attention came from a series of event-related brain potential (ERP) studies. Specifically, Striano, Reid, and Hoehl (Reference Striano, Reid and Hoehl2006) examined the ERP correlates of joint engagement in 9-month-old infants in a paradigm in which an adult interacted live with the infant in two contexts. In the joint-attention context, the adult looked at the infant and then at the computer screen displaying a novel object. In the non-joint-attention context, the adult only looked at the chest of the infant and then at the novel object presented on the screen. Objects presented in the joint-attention context, compared to objects in the non-joint-attention context, were found to elicit a greater negative component (Nc) over frontal and central electrodes, an ERP component known to be generated within the mPFC (Reynolds & Richards, Reference Reynolds and Richards2005). Based on these neural-level findings it was concluded that infants are sensitive to joint-attention interactions. Critically, this ERP paradigm has also been used to examine joint attention in younger infants (Parise, Reid, Stets, & Striano, Reference Parise, Reid, Stets and Striano2008). This study reported that, already by the age of 5 months, infants show a selectively enhanced Nc during the joint-attention condition.

Relatedly, Grossmann and Johnson (Reference Grossmann and Johnson2010) examined localized brain responses in 5-month-old infants during triadic social interactions using functional near-infrared spectroscopy (fNIRS). In this study, infants were presented with interactive scenarios in which a social partner (virtual agent presented on a screen): (a) Engaged in joint attention by gaze cueing the infant's attention to an object after establishing eye contact [joint-attention condition], (b) gaze cued the infant's attention to an empty location [no referent condition], or (c) looked at an object without prior eye contact with the infant [no eye contact condition]. Only in the joint-attention condition, infants selectively recruited a brain region within the mPFC, demonstrating that 5-month-old infants are sensitive to triadic interactions. Moreover, this study showed that 5-month-old infants employed a similar region of mPFC as seen during joint attention in human adults (Schilbach et al., Reference Schilbach, Wilms, Eickhoff, Romanzetti, Tepest, Bente and Vogeley2010). Furthermore, there is neuroscience evidence to show that young infants are sensitive to when a social partner follows their gaze. Grossmann, Lloyd-Fox, and Johnson (Reference Grossmann, Lloyd-Fox and Johnson2013) examined 5-month-olds' sensitivity to when a social partner follows their gaze by measuring infant brain responses using fNIRS during scenarios in which a social partner either followed the infants' gaze to an object that they had previously looked at (congruent condition) or a social partner shifted attention to look at a different object (incongruent condition). The fNIRS results of this study demonstrated that a selective region in the mPFC displayed an enhanced response to the congruent condition, suggesting that infants are sensitive to when someone follows their gaze. This finding provides early developmental evidence for theories, positing that brain processes are flexibly engaged by self- and other-initiated social interactions, including during joint-attentional engagement (Schilbach et al., Reference Schilbach, Timmermans, Reddy, Costall, Bente, Schlicht and Vogeley2013).

Taken together, the findings summarized above suggest the early developmental emergence of the brain system involved in joint attention by at least 5 months of age (Grossmann et al., Reference Grossmann, Lloyd-Fox and Johnson2013; Grossmann & Johnson, Reference Grossmann and Johnson2010; Parise et al., Reference Parise, Reid, Stets and Striano2008; Striano et al., Reference Striano, Reid and Hoehl2006). This is also in agreement with behavioral evidence showing that young infants, by around 5 months of age, reliably follow pointing gestures, suggesting that the sensitivity to joint (triadic) engagement extends beyond eye cues to human-unique, gestural means of triadic communication (Bertenthal, Boyer, & Harding, Reference Bertenthal, Boyer and Harding2014; Rohlfing, Longo, & Bertenthal, Reference Rohlfing, Longo and Bertenthal2012). These findings, demonstrating the early emergence of the brain system involved in joint attention as a sophisticated form of human social cognition, challenge Spelke's account arguing for the relatively late emergence of such social-cognitive skills around 10 or 12 months of age. These neural-level findings are difficult to reconcile with, or integrate into, Spelke's current account, unless a new core system is added, or the core systems (agents and social partners) implicated in social cognition are revised or extended to adequately include existing evidence from the overlooked line of neuroimaging research on joint attention with infants. Moreover, this line of evidence from neuroimaging research with young infants also challenges Spelke's claims that the purported social-cognitive changes at the end of the first postnatal year of life are underpinned by “the mastery of language.” Indeed, the neuroscience work overlooked by Spelke is much better aligned with alternative theoretical accounts, assigning developmental primacy to nonverbal, social-cognitive capacities, namely joint or shared attention/intentionality, being of foundational importance for the acquisition of spoken language and other communicative, cooperative, and cultural feats (Grossmann, Reference Grossmann2017; Kuhl, Reference Kuhl2007; Tomasello, Reference Tomasello2019). In sum, this commentary challenges Spelke's view on the early development of social cognition from a neuroscience perspective, presenting an alternative of what may be considered uniquely human forms of social cognition in the service of cooperation, communication, and culture.