Carruthers' “mindreading is prior” model holds that we gain knowledge of our own propositional attitudes through applying our mentalizing capacities to our own behavior and inner experience. In evaluating this claim, Carruthers considers the question of developmental asymmetry between self- and other-knowledge. Both the “two independent mechanisms” and the “metacognition is prior” views would predict that metacognition should appear before mindreading. These two models do not, however, exhaust the possibilities for non-interpretative accounts of metacognition (specifically, they leave “one mechanism, two modes of access” as a remaining competitor account). This fact, together with recent evidence for very early mentalizing competence which, for methodological reasons, cannot be matched by data on early metacognition, means that such evidence cannot discriminate between Carruthers' interpretative account and a non-interpretative alternative.
There is one respect, however, in which a developmental asymmetry between mindreading and metacognition could have a bearing on Carruthers' model. Carruthers proposes that self-interpretation can proceed on the basis of information about overt behavior and physical circumstances, along with elements of inner experience such as inner speech, visual imagery, and feelings. He notes that there will be some instances where, behavioral data being lacking (as in the example of someone sitting quietly in their living room), self-interpretation will be based exclusively on information about inner experience.
The question, then, is what sort of information that could support self-interpretation is available to young children. Presumably, young children have the same possibilities for interpreting external information as adults do. But is their inner experience comparable? At least three sources of evidence lead us to scepticism on this point.
First, there are the findings of experimental research on children's understanding of inner experience. For example, consider the findings of Flavell and colleagues (e.g., Flavell et al. Reference Flavell, Green and Flavell1993; Reference Flavell, Green and Flavell2000) that preschool children frequently deny the presence of inner experience in individuals (including themselves) when it would be appropriate to attribute such experience. These results are usually interpreted as evidence that young children have only weak powers of introspection. But an alternative interpretation is that young children do not experience a stream of consciousness in the way that older children and adults do, and that this accounts for their weak understanding of the inner experience of others (Hurlburt & Schwitzgebel Reference Hurlburt and Schwitzgebel2007, sect. 11.1.7.8; Meins et al. Reference Meins, Fernyhough, Wainwright, Clark-Carter, Das Gupta, Fradley and Tuckey2003).
A second line of evidence comes from the limited data from experience sampling in children. Descriptive Experience Sampling (DES; Hurlburt & Heavey Reference Hurlburt and Heavey2006) involves careful interviewing around records made of inner experience shortly preceding a random electronic beep. Hurlburt describes an episode of DES with a nine-year-old boy who reported an image of a hole in his backyard containing some toys (Hurlburt & Schwitzgebel Reference Hurlburt and Schwitzgebel2007, Box 5.8). When asked whether this image was an accurate description of his backyard, the participant replied that he had not yet had time to put all of the toys into the image. Hurlburt's conclusion from this and other instances of childhood experience sampling is that constructing a visual image is a skill that takes time to develop. Although much remains to be done in adapting experience sampling techniques for use with young children, the evidence currently available invites caution in making assumptions about young children's inner experience.
Third, there are theoretical reasons for not assuming that certain aspects of children's inner experience, particularly inner speech, are comparable to those of adults. The most fully developed theory of the development of inner speech is that of Vygotsky (Reference Vygotsky1934/1987). In his theory, inner speech is the developmental outcome of the internalization of social speech via the transitional stage of private speech. Findings from the study of private speech suggest that its transformation into inner speech is unlikely to be complete until middle childhood (Winsler & Naglieri Reference Winsler and Naglieri2003). The view that there is a general shift towards verbal mediation of cognition in the early school-age years is supported by findings that phonological recoding of visually presented material in short-term memory tasks is linked to private speech use at this age (Al-Namlah et al. Reference Al-Namlah, Fernyhough and Meins2006). Speech that is still in the process of being internalized is likely to appear to the child's consciousness as something other than adult-like inner speech (Fernyhough et al. Reference Fernyhough, Bland, Meins and Coltheart2007). Further experimental research on the transition to verbal mediation, complemented by more developmentally sensitive experience sampling studies, should provide a clearer indication of when inner speech can be assumed to be present in young children.
There are reasons, then, for doubting that young children have access to the full range of inner experiences proposed by Carruthers to form the basis of self-interpretation. Because inner speech is one of the main sources of evidence supposed to feed into individuals' interpretations of their own propositional attitudes, the emergence of metacognition should be developmentally constrained by the emergence of inner speech. Other aspects of inner experience, such as visual imagery, are also likely to take time to develop. Given what we know about the timetable for the emergence of mindreading capacities (particularly the evidence for some mentalizing competence in the second year of life), Carruthers' theory should predict a developmental lag between mindreading and metacognition. An alternative for Carruthers would be to argue that behavioral and contextual evidence was sufficient for self-interpretation in young children, but then his account would be indistinguishable from that of Gopnik (Reference Gopnik1993).
