Think Theory Testing, Not Realism

Yes, but that’s irrelevant. Realistic assumptions depend on context. It is not realistic to assume people act in their own best interest when formulating theo-ries of dysfunctional behavior. The same assumption is usually realistic when formulating theories about market behavior. Can We Test Assumptions as We Test Predictions? Yes, but do not reject theories based on tests of assumptions. For example, do not reject the assump-tion that matter is composed of atoms because the assumption fails for plasmas. Do not reject the bio-logical theory of cells because the assumptions fail for viruses. Do not abandon your clock because time depends on the observer’s speed and location in a gravitational field. Can the Same Prediction be Generated by Com-pletely Different Assumptions? Yes, but different assumptions also generate conflict-ing implications. Test competing theories on the con-flicting implications, not the common implications. Can Unrealistic Assumptions Generate Dangerous Theories? Yes, but so can realistic assumptions. For example, Aristotle’s theory of five elements, Aristotle’s theory of gravity, the emission theory of vision, caloric the-ory, Humoalism and many other theories all seemed realistic in their day, but were wrong and shown wrong by predictive testing. It is dangerous to evaluate theories on arbitrary notions about the realism of assumptions rather than evaluating them based on the objective testing of predictions