Supporting the central claim of Phillips et al. that knowledge rather than belief representation constitutes the more basic cognitive capacity, we highlight emerging evidence from developmental research on infants' communicative use of such knowledge representations. This commentary suggests that infants actively and selectively seek epistemic input from more knowledgeable others and, in turn, transmit such information to less knowledgeable others. As active participants in the bidirectional exchange of knowledge, infants take an interrogative stance and also themselves act as informants, using developmentally available tools from their preverbal communicative repertoire.

To actively solicit information, infants have been shown to socially reference adults who were more likely to help them resolve an epistemically uncertain situation (Bazhydai, Westermann, & Parise, Reference Bazhydai, Westermann and Parise2020c; Goupil, Romand-Monnier, & Kouider, Reference Goupil, Romand-Monnier and Kouider2016; Harris, Bartz, & Rowe, Reference Harris, Bartz and Rowe2017; Stenberg, Reference Stenberg2013; Vaish, Demir, & Baldwin, Reference Vaish, Demir and Baldwin2011) and point to objects they want to learn about in the presence of a knowledgeable rather than an uninformed person (Begus & Southgate, Reference Begus and Southgate2012; Kovács, Tauzin, Téglás, Gergely, & Csibra, Reference Kovács, Tauzin, Téglás, Gergely and Csibra2014; Lucca & Wilbourn, Reference Lucca and Wilbourn2018). To actively transmit information in situations where infants themselves were more knowledgeable than their social partners, they have been shown to use informative pointing (Liszkowski, Carpenter, Striano, & Tomasello, Reference Liszkowski, Carpenter, Striano and Tomasello2006; Liszkowski, Carpenter, & Tomasello, Reference Liszkowski, Carpenter and Tomasello2008; Meng & Hashiya, Reference Meng and Hashiya2014; O'Neill, Reference O'Neill1996) and deliberate action demonstration (Bazhydai, Silverstein, Parise, & Westermann, Reference Bazhydai, Silverstein, Parise and Westermann2020a; Flynn, Reference Flynn2008; Vredenburgh, Kushnir, & Casasola, Reference Vredenburgh, Kushnir and Casasola2015) as communicative tools indicative of early emerging, proto-teaching strategies (Strauss & Ziv, Reference Strauss and Ziv2012).

Not only do infants represent others' states of knowledge versus ignorance, but they also form epistemic expectations and actively seek explanations or clarifications when adults do not act in accordance with those prior expectations (Harris, Koenig, Corriveau, & Jaswal, Reference Harris, Koenig, Corriveau and Jaswal2018). For instance, infants expect to learn from previously knowledgeable informants (Begus, Gliga, & Southgate, Reference Begus, Gliga and Southgate2016) and look longer toward adults who provide inaccurate labels for familiar objects (Koenig & Echols, Reference Koenig and Echols2003) or object location information incongruent with their true knowledge (Galazka, Gredebäck, & Ganea, Reference Galazka, Gredebäck and Ganea2016). Such enhanced attention to a speaker can be plausibly interpreted as indexing a violation of the expectation that social partners are, by default, reliable rather than misleading in their information provision (Sperber et al., Reference Sperber, Clément, Heintz, Mascaro, Mercier, Origgi and Wilson2010). Furthermore, infants are less likely to subsequently learn from previously untrustworthy informants (Brooker & Poulin-Dubois, Reference Brooker and Poulin-Dubois2013; Koenig & Woodward, Reference Koenig and Woodward2010) or from those who provide information incongruent with what was asked of them (Begus, Gliga, & Southgate, Reference Begus, Gliga and Southgate2014).

These early behaviors indicate that in social situations of epistemic uncertainty, infants act to close both intra- and inter-individual knowledge gaps, ultimately achieving an equal distribution of knowledge upon its social transfer (Harris, Reference Harris2017; Strauss & Ziv, Reference Strauss and Ziv2012). Prominent theories of epistemic curiosity (information gap and learning progress; for a review, see Bazhydai, Twomey, & Westermann, Reference Bazhydai, Twomey, Westermann and Benson2020b) conclude that the information being sought is inherently factual, as it is in the case of curiosity in social learning (Begus & Southgate, Reference Begus, Southgate, Saylor and Ganea2018; Harris, Reference Harris, Butler, Ronfard and Corriveau2020). When providing information, evidence to date shows that infants transmit factual information (e.g., where a hidden object is located or how to make a new toy play music).

Notably, and in accord with the proposals by Phillips et al., the ability to actively exchange knowledge does not presuppose a full-blown mentalizing ability: Although infants exert control over their information seeking and information provision, their behaviors are likely proto-metacognitive (but see e.g., Goupil & Kouider, Reference Goupil and Kouider2016, for evidence of metacognitive sensitivity in infancy) (Harris, Reference Harris, Butler, Ronfard and Corriveau2020; Heyes, Reference Heyes2016; Strauss & Ziv, Reference Strauss and Ziv2012). For example, when infants indicate what they know to ignorant others, there is no evidence so far to suggest they realize that they possess unique transferable knowledge, or that they deliberately reason about the nature of their communicative behaviors, which nevertheless perform an informative function, as a result of which others can also know. Thus, the intra- and inter-individual epistemic gaps do not have to be realized as such for active social learning to occur.

Crucially, and dissenting from the picture painted by Phillips et al., we argue that this deliberate process of “asking for” knowledge and spontaneously taking steps to pass it on is a distinctively human ability. Although nonhuman animals represent others' knowledge and act in accordance with those representations, in contrast to infants (Harris & Lane, Reference Harris and Lane2014; Ronfard & Harris, Reference Ronfard and Harris2015), we see little evidence of active information solicitation in them. Similarly, evidence of information transmission remains limited and less diverse and flexible than that of humans (Burdett, Dean, & Ronfard, Reference Burdett, Dean and Ronfard2017; Strauss & Ziv, Reference Strauss and Ziv2012; but see Musgrave et al., Reference Musgrave, Lonsdorf, Morgan, Prestipino, Bernstein-Kurtycz, Mundry and Sanz2020, for new evidence of teaching-like behaviors in wild chimpanzees). For example, infants have the capacity to exchange cultural information (e.g., the label or function of an artifact) as opposed to exclusively functional information (e.g., the location of a food source), thereby distinguishing human infants' information exchange from that of any other nonhuman primate. Thus, among the various social-learning strategies that involve the transmission of knowledge from one social partner to another, and which we share to a large extent with nonhuman animals (imitation, emulation, and observation), the active and selective seeking and provision of information among conspecifics appears to be unique to humans.

We are excited about the directions outlined in the paper's call to arms and emphasize the need to investigate the developmental foundations of information seeking and transmission. In light of the likely connections between curiosity and teaching in cultural evolution (van Schaik, Pradhan, & Tennie, Reference van Schaik, Pradhan and Tennie2019), future studies should investigate how the active solicitation of information impacts its subsequent transmission, examining whether the motivation to seek knowledge rather than belief makes information sharing more likely. If knowledge representations are primary, knowledge- rather than belief-based information would be more likely to be both sought and further propagated.

In summary, we support the knowledge-as-more-basic view and propose to strengthen the account by adding to the list of signature properties of knowledge representation, the ability of infants to engage in active social learning as manifested in information seeking and information transmission. These emerging findings support the proposal that knowledge representation as a basic capacity may be shared with other evolutionarily close species, whereas the active communication of knowledge evolved in humans to optimize learning from others and informing others.