People make simple physical inferences by acting on objects. They might, fo
r example, tilt a container to determine its chances of spilling. Five expe
riments examined whether people can also draw physical inferences by taking
simulated actions. The basic experimental task involved 2 glasses of diffe
rent widths but equal heights. People imagined that the glasses were filled
to the same level with water and answered whether they would spill at the
same or different angles. When asked explicitly, people were usually wrong;
but, when they closed their eyes and tilted each glass until the imagined
water reached the rim, they correctly tilted a narrow glass farther than a
wide one. These experiments dissociated simulated actions from both proposi
tional inferences and visual imagery. The results suggest a new emphasis on
the role of motor activity in drawing inferences and address issues relate
d to joining naive intuitions and explicit understandings.