The broad range of commentaries addressed central debates in the origins, nature, and development of relationship knowledge. These debates mirror those that have played out in other domains: do infants possess innate, abstract primitives for representing relationships? Or is their understanding primarily the result of learning? Should we consider infants’ behavior in experiments a reflection of abstract cognitive concepts or as reflexes or embodied knowledge? The commentaries also raise questions about other possible cognitive primitives as well as the evolutionary roots of such knowledge. What is the result of statistical learning, and what are inductive biases? Ultimately, these commentaries examine how innate representations of relationships may facilitate the resolution of other learning challenges, including the acquisition of cultural norms, institutional roles, group dynamics, and moral principles. Ultimately, I argue that advancing our understanding will require integrative approaches drawing on developmental, comparative, computational, and cultural research. This exchange clarifies many pressing empirical puzzles, setting the stage for a research program that I hope others across disciplines and theoretical standpoints will join in.

Thomas suggests that evidence suggestive of an innate moral core can instead be explained by relationship inferences and evaluations. We argue that the analogy between other demonstrations of prosociality (e.g., helping) and imitation is insufficient, and additionally highlight key findings that appear inconsistent with a relationship-centric view of infants’ sociomoral capacities.

Cognitive scientists have long investigated infants’ abilities to reason about others’ minds. Here, we identify a limitation to research on early mental state reasoning: that this research has mostly challenged infants to reason about the minds of strangers. In this commentary, we call on cognitive scientists to consider the role of social relationships when studying mental state reasoning and cognition more broadly.

Thomas proposes that infants classify social relationships using the categories of Relational Models Theory. Attachment theory and research describe an innate, early-emerging suite of psychological processes by which infants identify and bond selectively to caregivers. Classification of others’ relationships might build upon these same mechanisms, with implications for future research on social cognition across development.

Understanding innate core cognition for social relations as a deep, generative social grammar that helps solve the learnability problem of discovering who relates to whom in what ways “amounts to asking how human social life is understandable, and hence possible” (Thomsen & Carey, 2013, p.17). Expanding the focus beyond enduring interpersonal relationships allows us to understand how social and societal meaning (e.g., in ideology) reflect fundamental forms of relational coordination ultimately grounded in cross-species affordances.

Thomas proposes that two basic computations determine how infants represent social relationships: categorizing the connection by model and computing its strength. I argue that the concept of relational strength is ill-defined, its empirical evidence in early development unconvincing, and its putative operations better explained as categorical inferences about narrow relational schemas (e.g., caregiving).

Navigating the social world requires representing how people relate to one another and how people are connected in a network. Though both are forms of relational knowledge, it is difficult to unify them in a single theoretical framework. I hypothesize that the statistics of observation may be sufficient to learn both network structure and an intuitive theory of relationships.

In this commentary, we express enthusiasm for Thomas’s article and research efforts, but question her proposed division between social groups and social relationships. Joining other social scientists, we view relationships as fundamental to – and sometimes synonymous with – social groups. Further, empirical studies focused on infants and young children highlight the tight link between social groups and relationships.

Equality matching describes a mode in which reciprocal relationships are maintained. While Thomas argues for core representations that can support equality matching, we argue that two forms of reciprocity must be considered and may require distinct representations: positive and negative direct reciprocity. We review both evolutionary and developmental evidence in support of this proposal.

Core social knowledge is not limited to humans; it has evolved in other species. A comparative, evolutionary perspective prompts one to ask: what, exactly, has been favored by natural selection? A rigid, intuitive theory that matches each species’ society, or a more general tendency to search for social patterns, leaving the details and theory to emerge during development?

Integrating Relational Models Theory with Carey’s core cognition theorizing is possible, promises to be productive, but requires clarifications. We sketch how such an integration can strengthen social psychological theorizing in three areas.

In addition to core knowledge, the prenatal environment, parental care, and early life experience play substantial roles in the establishment of social cognition. Given the evolutionary importance of early life care in humans and mammals in general, there are innate reward and stress based neuroendocrine mechanisms that are critical to maternal infant bonding as well as later social behavior.

Thomas proposes that infants use relational models to reason about interactions between relationship partners. However, many basic social interactions do not contain enough information to differentiate between models. We propose that both basic interaction reasoning and the formation of diverse, culturally relevant relationship categories are better accounted for by positing relationship features that can be flexibly combined.

Thomas’ paper provides a compelling account of the nature and development of mental representations of social relationships. Understandably, it focuses on dyadic and triadic relationships. We emphasize the need to elucidate the development of another key component of humans’ understanding of social relationships: “stitching together” knowledge of social ties into representations of social networks that shape one’s thoughts and actions.

Thomas proposes the existence of core knowledge that supports an innate, or early developing, intuitive theory, where social relationships are understood in reference to three models: communal sharing, authority ranking, and equality matching. We consider possible precursor-representations that could support infants’ discovery of relationships in the social world and how these representations relate to the basic models discussed by Thomas.

The target article notes ways infants can use direct interactions between individuals to recognize relationships. We ask whether infants might also recognize relationships between individuals indirectly, by observing how they interact with other people. We suggest that indirect information might be used to distinguish between communal sharing and equality matching relationships and discuss how it might support inferences of relationship strength.

This commentary proposes a dynamic approach to the construction of relationship representations and mentalization. We argue that these processes result from early caregiver–infant interactions, particularly through mechanisms of synchrony, parental sensitivity, and attachment. This extends the author’s model beyond cognitive and innate perspectives by emphasizing the role of early interactions in constructing representations that are embodied, affective, and dynamic.

Infants and young children appear to have core knowledge structures related to social dynamics. However, I think intuitions about group categories are too easily dismissed or ignored in the target article. More substantively, I argue for a theoretical distinction between intuitive sociology and intuitive social psychology.

Ashley Thomas’s account does not detail mechanisms for processing structural regularities, organizing sets of social relationships. I argue that combinatoriality and relational reasoning are essential for processing social structures. These abilities, foundational to processing social relationships, enable early, abstract, and flexible understanding of social hierarchies and networks, supporting adaptive social inference across species and development.

Thomas compellingly argues that infants are especially attuned to relationships. We agree but suggest her claims can reach further in her discussion of infants’ reactions to distributional events. Thomas overlooks an interpretation of this literature in which unequal allocations might reflect relational inferences, challenging some of her claims regarding equality matching relationships and whether infants possess an “innate moral core.”