This research utilized longitudinal and cross-sectional methods to investig
ate the relation between the development of a representational theory of mi
nd and children's growing ability to search their own minds for appropriate
problem solutions. In the first experiment, 59 preschool children were giv
en 3 false-belief tasks and a divergent-thinking task Those children who pa
ssed false-belief tasks produced significantly mole items, as well as more
original items, in response to divergent-thinking questions than those chil
dren who failed. This significant association persisted even when chronolog
ical age and ver bal and nonverbal general ability were partialed out. In a
second study, 20 children who failed the false-belief task in the first ex
periment were retested 3 months later. Again, those who now passed the fals
e-belief tasks were significantly better at the divergent-thinking task tha
n those who continued to Sail. The associations between measures of diverge
nt thinking and understanding false beliefs remained significant when contr
olling for the covariates. Earlier divergent-thinking scores did not predic
t false-belief understanding three months later. Instead children who passe
d false-belief tasks on the second measure improved significantly in relati
on to their own earlier performance and improved significantly more than ch
ildren who continued to fail. False-belief task performance was significant
ly, correlated to the amount of intraindividual improvement in divergent th
inking even when age and verbal and nonverbal skills were partialed out. Th
ese findings suggest that developments in common underlying skills are resp
onsible for the improvement in understanding other minds and searching one'
s own. Changes in representational and executive skills are discussed as po
tential causes of the improvement.