Doesn't see, doesn't know: is anticipatory looking really related to understanding of belief?

Clements and Perner (Cognitive Development, 9 (1994), 377-397) reported that children show, understanding of a story characters belief in their anticipatoi-y looking responses before thev show this in their answers to test questions. According to Clements and Perner the anticipatory looking responses provide evidence of implicit understanding of belief. This paper examines the possibility that the anticipatory looking measure is indicative of (a) children using a seeing = knowing rule, i.e. children linking not seeing with ignorance rather than a sensitivity to belief, or (b) a tendency to associate the protagonist with the left-hand container. Thirty-two children aged between 2 years 11 months and 4 years were told a false belief story similar to that used in Clements and Perner (1994) except that three containers were used instead of two. The protagonist first looks inside the middle box but then puts the object in the A,,ft-hand box. In his absence, a second character moves the object unexpectedly to the right-hand box. If children's anticipatory looking was based on sensitivity to belief then they, should have looked clearly to the left-hand box. If it was based on an association bias or sensitivity to the character not knowing then they should have looked equally to the left-hand and middle boxes. The results were consistent with the former prediction suggesting that children's anticipatory looking responses may indeed be governed by an implicit sensitivity to belief.