Thomas proposes a daring account of early relationships representations, which will lead to exciting future research in developmental science and beyond. However, Thomas leaves vague the definition of relationships. Indeed, Thomas includes discussions of transient interactions among (groups of) people that fall outside widely agreed upon defining features of relationships as dyadic interchanges that occur over time (Hinde, Reference Hinde1979), raising questions about the specificity of Thomas’ theory to relationships. Moreover, Thomas draws exclusively on Fiske’s (Reference Fiske1992) Relational Models Theory. Although influential, this theoretical focus fails to incorporate rich theory and evidence from the field of relationship science (Berscheid, Reference Berscheid1999; Finkel & Simpson, Reference Finkel and Simpson2015; Fletcher et al., Reference Fletcher, Simpson, Campbell and Overall2019; Simpson & Campbell, Reference Simpson and Campbell2013), particularly as informed by social developmental research (Laursen & Bukowski, Reference Laursen and Bukowski1997; Sroufe, Reference Sroufe2000; Underwood & Rosen, Reference Underwood and Rosen2013). Here, we identify three critical issues with Thomas’ account, highlighting the importance of considering infants’ social world in formulating an intuitive theory of relationships.

First, Thomas argues that a core characteristic of relationships is strength. However, from a relationship science perspective, relationships are defined in terms of their quality, with close relationships (e.g., parent-child attachments, friendships) distinguished from other types of relationships (e.g., peers, work relationships). One widely accepted definition of close relationships is that they are characterized by a high degree of interdependence that endures over an extended period (Kelley et al., Reference Kelley, Berscheid, Christensen, Harvey and Huston1983). Thus, close relationships are distinguished from other relationships by their nature or quality, which strength fails to capture.

Take, for example, attachment relationships. Attachments are typically the first relationships infants establish and thus, relevant to Thomas’ proposal. Characterizing such relationships in terms of strength results in a fundamental misunderstanding of attachment (Sroufe & Waters, Reference Sroufe and Waters1977). Attachments are defined as an enduring psychological bond between the infant and caregiver that serves protective functions and evolved to promote survival (Bowlby, Reference Bowlby1969), highlighting defining characteristics that separate attachments from other relationships that cannot be reduced to strength. For example, having “strong” relationships with kin (e.g., parent, grandparent) does not indicate that such relationships are attachments. Indeed, an attachment relationship is evidenced by preferential direction of attachment signals toward a relationship partner in times of uncertainty (Bowlby, Reference Bowlby1969).

Second, Thomas argues for a core knowledge system of relationships, analogous to core knowledge systems of objects and agents (Spelke, Reference Spelke2024). However, further information is needed to support claims of relationship representations as core knowledge that is “…innate, skeletal, largely automatic, and evolutionarily old” (p. 7). Although comparative studies relevant to specific relationship cues are reviewed, this falls short of providing an evolutionary account for the premise of core knowledge of relationships. Why would humans evolve to have innate knowledge of relationships and to organize it around the proposed core components? This is addressed in supporting knowledge in other core domains (Cosmides & Tooby, Reference Cosmides, Tooby, Hirschfeld and Gelman1994). For example, innate knowledge of the physical and social worlds might have evolved to allow humans evolutionary advantages of locating objects, navigating, foraging, hunting, as well as facilitating living with others and in social groups (e.g., Hamlin, Reference Hamlin2013; Völter et al., Reference Völter, Tomašić, Nipperdey and Huber2023; Wynn & Bloom, Reference Wynn, Bloom, Killen and Smetana2014). A core knowledge system of relationships might afford comparable adaptive benefits. However, situating the presence and organization of intuitive relationships knowledge in evolutionary theory is necessary, highlighting another way in which Thomas’ theory would benefit from greater attention to social developmental and relationship science (e.g., Simpson & Belsky, Reference Simpson, Belsky, Cassidy and Shaver2008).

Moreover, the extant infant cognitive evidence cited by Thomas fails to support core knowledge of relationships. Instead, the results can be interpreted as infant sensitivity to relationship cues. For example, infants are shown one interaction between two individuals. The presence of certain cues (e.g., imitation, saliva-sharing) evidently leads to infants making predictions about the individuals’ (inter)actions in a new context. Being able to interpret dyadic interactions based on certain cues, however, is not the same as having innate representations of relationships. Whereas it is possible that infants are born with a sensitivity to cues of relationships, it is premature to suggest that infants also have an innate concept of dyadic, interpersonal relationships guiding their reasoning about the social world. Rather, as infants are born into relationships with their parents, they may learn from early experiences (e.g., interpersonal synchrony and saliva-sharing between parents and infants) the importance of certain cues and use them to make predictions about social interactions.

Third, Thomas implies that relationship representations can supersede or precede sociomoral knowledge (Baillargeon et al., Reference Baillargeon, Scott, He, Sloane, Setoh, Jin, Wu, Bian, Mikulincer, Shaver, Borgida and Bargh2015; Hamlin, Reference Hamlin2013). As noted above, the extant evidence for infant relationship representations remains tenuous, while there has been a wealth of evidence for infant moral knowledge (Ting et al., Reference Ting, Dawkins, Stavans, Baillargeon and Decety2020; Woo, Tan, & Hamlin, Reference Woo, Tan and Hamlin2022). We therefore suggest that innate relationship representations, should they be present, complement moral understandings in the core social domain. For example, infant sensitivity to relationship cues might help them reason about dyadic interactions, as shown by Thomas and colleagues (Reference Thomas, Saxe and Spelke2022a; Reference Thomas, Woo, Nettle, Spelke and Saxe2022b). Group membership markers, even minimal cues (Jin & Baillargeon, Reference Jin and Baillargeon2017), can help infants reason about situations involving more than two individuals and evaluate interactions among them, based on morality (Ting et al., Reference Ting, Dawkins, Stavans, Baillargeon and Decety2020). The interplay of group representations, relationship representations, and morality thus underpins social cognition.

In summary, we believe Thomas’ theory will inspire more work that aims to examine multiple facets of infants’ social cognition. However, just as we suggest that Thomas’ theory would benefit from closer attention to relationship science, work on infants’ social cognition would also benefit from closer attention to the social world (Tasimi, Reference Tasimi2020). Socioemotional developmentalists have extensively studied how experiences and individual characteristics shape development (e.g., Borowski et al., Reference Borowski, Groh, Bakermans-Kranenburg, Fearon, Roisman, van IJzendoorn and Vaughn2021; Groh et al., Reference Groh, Fearon, Bakermans-Kranenburg, Van IJzendoorn, Steele and Roisman2014). These factors must be included when examining how infants construct representations of the social world, including their representations of relationships. As an example, we have started collaborative research on infant attachment representations, drawing on our shared expertise (Luo, vanMarle, & Groh, Reference Luo, vanMarle and Groh2024). Collaborations such as this are necessary to chart the developmental course of infant representations as they pertain to the social world and critically evaluate theories regarding the role of experience and innate knowledge.