Using Theory of Mind

— The ability to flexibly predict others’ behaviorshas been ascribed to a theory of mind (ToM) system. Mostresearch has focused on formal conceptual definitions ofsuch a system, and the question of whom to credit with aToM. In this article, I suggest shifting perspective fromformal definitions to a usage-based approach. Thisapproach views action within human interaction as cen-tral to the emergence and continuous development of theability to flexibly predict others’ behaviors. Addressing thecurrent debate about whether infants have a ToM, I illus-trate how infants use flexible action expectations tointeract with others appropriately. I also discuss the con-tinuous development of ToM and its natural structure froma usage-based perspective. KEYWORDS— action prediction; usage-based theory; falsebelief; prelinguistic communication THEORY OF MIND: WHAT FOR? Theory of mind (ToM) is a cognitive system that “can be used tomake predictions about the behavior of others” (Premack WRakoczy, 2012). Other mental-state concepts, like intentions,desire, and knowledge, are considered precursors to a full, rep-resentational ToM (e.g., Wellman & Liu, 2004), but these stillfollow the logic of formal conceptual definitions of ToM.One problem with formal conceptual approaches is that thepostulated definitions of ToM and their empirical operationaliza-tions influence each other. The validity of the widely used stan-dard false belief tests has been questioned because additionaltask demands like language and executive skills conceal con-ceptual understanding (Bloom & German, 2000). When remov-ing these demands, apparently even 1-year-olds understandfalse belief (Baillargeon, Scott, & He, 2010). For other research-ers, however, the very conceptual definitions require some of theremoved demands, such as language to entail hypothetical —notonly practical—belief scenarios (Perner, 2010), executive skillsto enable flexibility in perspective shifting (e.g., Moses, Carlson,& Sabbagh, 2005), or other general reasoning skills to get frompremises to conclusions. This results in the ambiguous use ofthe term ToM (Rakoczy, 2012) and leads to auxiliary termslike infant ToM and the postulation of separate ToM systems(Apperly & Butterfill, 2009).A second problem is that formal approaches bewitch us to con-ceptualize ToM as a stage-like all-or-nothing affair, obscuring