Social cognitive abilities in infancy: Is mindreading the best explanation?

Abstract

I discuss three arguments that have been advanced in support of the epistemic mentalist view, i.e., the view that infants' social cognitive abilities (SCAs) manifest a capacity to attribute beliefs. The argument from implicitness holds that SCAs already reflect the possession of an “implicit” and “rudimentary” capacity to attribute representational states. Against it, I note that SCAs are significantly limited, and have likely evolved to respond to contextual information in situated interaction with others. I challenge the argument from parsimony by claiming that parsimony per se favors neither a mentalist nor a behavior-reading account. Finally, I argue that early SCAs do not develop continuously into four-year-olds' belief attribution abilities. Accordingly, the argument from developmental continuity is empirically inadequate. Careful analysis of both the empirical data and the theoretical assumptions leading to the epistemic mentalist view is needed in order to improve our understanding of SCAs in early infancy.

Keywords::

• Belief
• Developmental Continuity
• False Belief Test
• Mindreading
• Social Cognition
• Tacit Knowledge
• Theory of Mind

Acknowledgements

I wish to thank Ian Apperly, Cameron Buckner, Susan Carey, Jay Garfield, Mitchell Herschbach, Albert Newen, Josef Perner, Silvano Zipoli Caiani, and three anonymous reviewers for useful discussion and comments on previous versions of this article. I would like also to thank audiences in Bochum, Cambridge, and Glasgow. Preparation of this article was supported by a Short Research Grant for Doctoral Candidates and Young Academics and Scientists from the German Academic Exchange Service (DAAD) and by a visiting fellowship from the Center of Mind, Brain, and Cognitive Evolution at the Department of Philosophy, Ruhr-University Bochum.

Notes

^[1] The central tenet of the epistemic mentalist view is that infants specifically can attribute beliefs. This is why it is labelled “epistemic.” Mentalist interpretations concerning infants' capacity to attribute other kinds of mental states (e.g., desires and intentions) are not the matter of discussion in the current debate.

^[2] Thus, in what follows, I will use ‘non-epistemic interpretations’ to refer to belief-like mindreading, perceptual mindreading, and non-mindreading accounts. In doing this, I will deliberatively gloss over the fact that belief-like and perceptual mindreading accounts are still mentalist—although in a weaker sense than the epistemic mentalist view—while non-mindreading accounts oppose both epistemic and non-epistemic mentalist interpretations (where by ‘non-epistemic mentalist interpretations’, I mean belief-like and perceptual mindreading accounts).

^[3] “False belief understanding should be taken as explanandum instead of explanans. That is, we should try to explain different forms of false belief understanding by appealing to an explanatory terminology that does not include ‘false belief’ as an explanatory concept” (de Bruin & Newen, 2012, p. 254).

^[4] “If the principles stated by the axioms of a formal theory of a task are embodied as tacit knowledge in a processing system that performs the task, then the causal structure of the processing mirrors the derivational structure of the (canonical) proofs in the formal theory” (Davies & Stone, 2001, p. 153).

^[5] Significantly, no experiment has ever assessed infants' sensitivity to others' beliefs originating from auditory information (but see Fenici, 2011 for a suggestion about how to design such a study).

^[6] I am in debt to Corrado Sinigaglia for this observation.

^[7] See also Mandler (2012, p. 438): “[15-month-olds] merely think that the other person has seen something and will make use of that to locate it, just as they do themselves. Infants of this age also use what other people see to predict their actions even if it differs from what the infants themselves see (Luo & Baillargeon, 2007; Sodian, Thoermer, & Metz, 2007). All of this is seeing as knowing, and it does not imply theory of mind.”

^[8] See Low and Watts (2013) for evidence of signature limits in infants' alleged capacity to attribute representational states.

^[9] Marr (1982) distinguished three levels of description of a (cognitive) function. At the more abstract computational level, a function is extensionally defined by an association between a set of activation patterns that can be provided as input to a cognitive mechanism and the set of produced responses. The algorithmic level specifies the computational one by adding a formal description of the procedure by which the function at the computational level is computed. Finally, the implementation level also describes how the algorithmic level is instanced in a physical device or a biological organism.

^[10] Though see Luo (2011) and Luo and Johnson (2009) for evidence on younger infants, and Fenici (unpublished manuscript) for a critical discussion.

^[11] See Dennett (1983) for a similar point about the cognitive capacities of vervet monkeys.

^[12] Baillargeon, Scott, and He (2010, p. 115) argue that SR-FBTs “involve only the false-belief-representation process” but do not impose significant inhibitory demands. They do not explain however why this should be the case, nor what counts as “significant” inhibitory demands.

^[13] This presupposes that these populations have the same SCAs in infancy of typically developing subjects, an assumption that seems certainly false in the case of autism. We do not have reason to doubt this in the case of other populations, however. For instance, because social cognitive development is considered invariant across different cultures, there is no reason why the development of SCAs' of infants in Asian countries should be different from that of infants in Western countries. Moreover, studies on deaf children demonstrate that only kids raised by hearing parents, who are not fluent in sign languages, are delayed in the capacity to pass ER-FBTs; in contrast, their peers raised in non-hearing families present typical development of this ability (Schick, de Villiers, de Villiers, & Hoffmeister, 2007). Considered that both groups present normal inhibitory capacities at age four, this result is at odds with the hypothesis that SCAs develop atypically in deaf infants. If that were the case (and the maturation of inhibitory skills was the essential factor affecting four-year-olds' performance on ER-FBTs), one would expect all deaf four-year-olds either passing or missing ER-FBT depending on the relevance of early SCAs to it.

^[14] And indeed, Senju et al. (2009), who also endorse a compensatory analysis of the capacity of adults with autism to pass ER-FBTs, also conclude against the argument from developmental continuity that their results “raise the surprising possibility that an early developing form of the cognitive ability to mentalize, evident in spontaneous looking behavior, is not a necessary precursor of the later developing form of mental-state attribution, which supports explicit reasoning” (Senju et al., 2009, p. 885).

^[15] In fact, they write that “the evidence for task specificity emerging from the present study appears to support lean, rather than rich accounts of ToM in infancy, since a rich account of infant ToM would predict generalized developmental relations with conceptually equivalent tasks” (Thoermer, Sodian, Vuori, Perst, & Kristen, 2012, p. 183).

Additional information

Notes on contributors

Marco Fenici

Marco Fenici is a Postdoctoral Research Fellow at University of Florence.