This commentary is an attempt to give a Vygotskian perspective on Carpendale's & Lewis's (C&L's) target article. The article uses ideas that are well familiar to Vygotsky's scholars. However, it develops these ideas further and raises important empirical questions about the role of social interaction in the development of social cognition. The article provides a fresh view on the old problems and frames themes traditional for the English-speaking developmental psychology into a broader international perspective.

Although we agree with the authors' criticism of the reigning approach to children's sociocognitive development, we raise three further issues. First, “mind talk” is not, in fact, any different from the other aspects of the social world about which children learn. Second, there is no choice between either the “single mind” or the “social context.” Finally, there is a spurious separation between organism and environment.

Theories of children's understanding of mind benefit from rigorous interpretations of demonstrations of similar understandings in closely related species. This commentary describes how Carpendale & Lewis's (C&L's) argument could be made more persuasive with a more rigorous interpretation of the studies of chimpanzees' understanding of mind.

Theory-of-mind (TOM) competence is under-specified in Carpendale & Lewis's (C&L's) account. In the neo-Vygotskian alternative outlined below, TOM development is driven by the internalisation of dialogic exchanges which preserve the triadic intentional relations of interactions within the epistemic triangle. On this view, TOM competence stems from children's ability to operate flexibly with the multiple perspectives manifested in internalised dialogue.

It is argued that constructionist theory provides only a partial account of how secure attachment leads to better social understanding. In addition to cooperative parent-child relations, the more efficient arousal and affect regulation system of secure infants, and developmental moderators of the processes of imitation, may play a part in explaining the association and offer clues as to how effective social understanding is generally acquired.

Theories that propose a modular basis for developing a “theory of mind” have no problem accommodating social interaction or social environment factors into either the learning process, or into the genotypes underlying the growth of the neurocognitive modules. Instead, they can offer models which constrain and hence explain the mechanisms through which variations in social interaction affect development. Cognitive models of both competence and performance are critical to evaluating the basis of correlations between variations in social interaction and performance on mental state reasoning tasks.

Individualism is not inconsistent with social interaction; it is required to explain it. Social exchanges, evidenced in gaze monitoring, social referencing, emotional responses, protodeclarative and imperative pointing, pretence, play, and conversation all play a role in development, but the nature of that role is opaque without an understanding of the cognitive mechanisms on which they depend.

Adopting a constructivist stance is not irreconcilable with executive function accounts of emerging social understanding. The executive function view allows for a gradual transition in theory of mind, while specifying the underlying cognitive processes that push that development forward. Executive function abilities can be seen as an important interactional component in the epistemic triangle.

Interpersonal understanding is rooted in social engagement. The question is: How? What features of intersubjective coordination are essential for the growth of concepts about the mind, and how does development proceed on this basis? Carpendale & Lewis (C&L) offer many telling insights, but their account begs questions about the earliest forms of self-other linkage and differentiation, especially as mediated by processes of identification.

Carpendale & Lewis (C&L) provide a convincing argument for how children construct social understanding through social interaction. Certainly mothers are important in family interaction; however, sibling interaction may also be key in the process of developing social understanding. In particular, the highly affective and reciprocal dynamics of the sibling relationship in both positive and conflictual interaction may be critical.

Carpendale & Lewis's (C&L's) interesting and insightful article did not integrate several potentially useful notions from emotion theory and research into their explanatory framework. I propose that emotions are indigenous elements of mind and that children's understanding of them is fundamental to their understanding of the mental life of self and others, understandings critical to the development of social and emotional competence.

We argue that theory-of-mind understanding has developed to facilitate joint thinking and planning, defined as the creation of new mental objects that could not have been created by one mind. Three components of this ability are proposed: the mental architecture indexed by false belief understanding, domain-specific knowledge, and the prioritization of the joint mind over the individual mind.

Carpendale & Lewis's (C&L's) theoretical reconstruction of the “theory of mind” problem offers new hope but still has far to go. The study of atypical development may provide some useful insights for dealing with the work ahead. In particular I discuss three issues – the boundary problem, the question of end states, and the issue of the centrality of triadic interaction.

Carpendale & Lewis's (C&L's) critique of traditional accounts of “theory of mind” is well taken, but the alternative theory they propose is premature at its best, unconvincing at its worst. The proposed theory is ad hoc and confirmatory in its findings; vague and generic in its claims; and unjustified and unnecessary in its (putative) novelty.

Although I am broadly in sympathy with Carpendale & Lewis's (C&L's) version of social constructivism, I raise two issues they might address. One bears on the question of how social understanding develops: Is their resistance to individualism inappropriately combined with a resistance to internalism? A second question concerns a more radical implication of their view for what social understanding is.

Carpendale & Lewis's (C&L's) constructivist account needs greater emphasis on how individual differences in caregivers' impact on the efficacy of epistemic triangle interaction in fostering children's understanding of mind. Caregivers' attunement to their infants' mental states and their willingness to enable infants to participate in exchanges about the mind are posited as important determinants of effective epistemic triangle interaction.

We agree that social interaction is crucial for understanding the development of theory of mind, but suggest that further elaboration of certain issues is needed. Detailed description of the knowledge structure of a developing theory of mind is necessary, and the notion of criteria for the use of mental state terms requires consideration of the sentence structures in which such terms appear.

If anything, Carpendale & Lewis's (C&L's) target article could have gone even further in challenging the view that a radical conceptual shift equips children with a theory of mind. Also, the authors should have elaborated on why their social constructivist account is more plausible than nativism. Their argument against simulation theory is perhaps the least-developed part of their thesis, and does little service to their cause.

Carpendale & Lewis (C&L) contend that correlations between sociolinguistic factors and theory-of-mind performance indicate that social knowledge develops from social interactive processes. However, theory-theory proponents also regard these correlations as compatible with their view of how mental concepts develop. A more fruitful distinction lies in the differences of both accounts in explaining how mental concepts acquire meaning.

Three points extend the authors' comprehensive and provocative argument: (1) The idea of “entering a community of minds” is suggested to replace theory of mind or social understanding; (2) learning words and concepts through a Wittgensteinian process often involves a period of “use without meaning”; (3) concepts based in social interaction are achieved through collaborative – neither individual nor social alone – construction.