What is the nature of infants’ earliest abilities to track briefly unperceived physical objects? And how, over the course of development, do infants get from these early forms of cognition to knowledge of simple facts about particular physical objects? The leading, best supported conjecture is that infants’ core knowledge of objects consists in a system of object indexes (Leslie et al. 1998; Scholl and Leslie 1999; Carey and Xu 2001; Scholl 2007). Accepting this conjecture confronts us with two challenges. The first is to explain a pattern of discrepancies involving multiple measures (violation-of-expectations vs manual search, for instance) and multiple types of disappearance (occlusion vs endarkening). The second challenge concerns the interface between core knowledge and thought. Object indexes have only limited, indirect influences on thought and action. How then could their operations explain the findings of violation-of-expectation experiments? And how could they play a role in explaining the emergence, in development, of knowledge of physical objects? This talk introduces the challenges and proposes a novel, two-part response. We should recognise that infants, like adults, sometimes represent objects motorically. Core knowledge is not one thing but a composite of two things that are, to an interesting degree, independent of each other: object indexes and motor representations. Further, object indexes can give rise to metacognitive feelings. The developmental transition from core knowledge of objects to knowledge proper may have such a protracted developmental course because only metacognitive feelings connect core knowledge to knowledge proper.