In the debate about how to interpret Violation-of-Expectation (VoE) based false-belief experiments, it has been suggested that infants are predicting the actions of the agent based on more or less sophisticated cognitive means. We present an alternative, more parsimonious interpretation, exploring the possibility that the infants' reactions are not governed by rational expectation but rather of memory strength due to differences in the allocation of cognitive resources earlier in the experiment. Specifically, it is argued that (1) infants' have a tendency to find more interest in events that observed agents are attending to as opposed to unattended events ("interest contagion"), (2) the object-location configurations that result from such interesting events are remembered more strongly by the infants, and (3) the VoE contrast arises as a consequence of the difference in memory strength between more and less interesting object-location configurations. We discuss two published experiments, one which we argue that our model can explain (Kovács etal., 2010), and one which we argue cannot be readily explained by our model (Onishi and Baillargeon, 2005).
Keywords: attention; development; false belief; interest contagion; memory; theory of mind.