The hypothesis is tested that during the preschool period a particular form of reasoning is applied to theory of mind and a set of problems that do not require the understanding of mental states. Three experiments each provided a different piece of support for this hypothesis. Experiment 1 found similar age-related changes between three standard theory-of-mind tasks (false belief, appearance-reality and representational change) and two specially designed nonmental state tasks (rule-based card sorting and physical causality). Experiment 2 demonstrated that 3-year-olds' poor performance on card sorting occurred across different pairs of dimensions (color-size, shape-number, size-number, color-shape) just as it does across different pairs of perspectives in the theory-of-mind tasks (self-other, looks like-is, before-after), and that sorting performance was related to theory of mind with age partialled. Experiment 3 showed that theory of mind was related to a specific complexity of sorting-sorting by one of two dimensions and not sorting by one dimension or two dimensions simultaneously. The results indicate that advances in theory of mind, card sorting, and causality between 3 and 5 years of age depend on being able to switch judgments across conditions and that reasoning by embedded rules could account for these changes.