Physicists come in two colors. One group tries to find a mathematics that will explain a set of reliable observations. The second group attempts first to imagine the physical event behind the observation. Werner Heisenberg belongs to the first group, Paul Ehrenfest to the second. Psychologists can be assigned to two comparable groups. Some try to model, logically or mathematically, the mechanisms behind the learning of associations. A larger group broods on the psychological and biological processes that occur when an association is formed. The history of all the sciences suggests that the collection of evidence in the service of proving a conjecture is most useful when the synthetic notion originates in prior, trustworthy observations. This strategy can be dangerous in immature sciences, like psychology, that have a meager store of reliable facts.