I believe all they are inferring from this experiment is that infants have some awareness that others (people, caterpillars,etc.) have internal mental states and processes, and previously they thought this didn't happen until much later. The dependent measure (look time) is pretty standard in development psych studies with infants. They also sometimes use rate of pacifier sucking as a measure. Tricky stuff, because infants can't verbally report, so researchers take a little bit of a leap of faith that this action actually measures interest. And I agree with you, Evo, the hand could certainly be a confound. There's no way to say how they controlled for that without reading the study.