Imagery and Action • Complex Physical Imagery

In everyday activity, people can infer simple physical behaviors. With varying degrees of accuracy, people can infer behavior for situations as various as colliding jellyfish, a water balloon hitting a bat, and the flexing of a candy bar. The figure shows an example of a problem that people can solve through imagery aided by physical action that they cannot solve by verbal reflection. There are two glasses of equal height filled to equal levels of water. The only difference is their diameter. The question participants have to answer is whether the two glasses of water start to spill at the same or different angle of tilt,  (and if one spills sooner, which one). When asked to solve the problem verbally, about 10% of people give the correct answer. However, if participants tilt the glasses, they are invariably correct. This is true even when participants eyes are closed and the glasses have no  water (a line indicates the water level) and participants need to imagine the water.  It is also true for children as young as 5-years old.

In this work, we develop evidence that these inferences can take place through spatial imagery. At the same time, we gather evidence indicating the role of physical knowledge in determining the specific spatial solutions that subjects deploy. When people need to infer how one object responds to the action of another, there is a problem of coordination. In imagery, people need to coordinate the behaviors of two objects. This raises the question of how people integrate general spatial knowledge and processes with more domain-specific, or context-sensitive, knowledge. Several lines of work demonstrate people’s ability to commit physical imagery and have helped to develop a time-based, object-oriented computational model of physical imagery.