A Method of Testing Classification Relatives

If in one of the natural sciences a scientist is to study a complex phenomenon which is determined by many elementary causes he usually approaches his problem in three different ways. He first studies the phenomenon in a purely empirical manner trying to describe his measurements by means of a mathematical formula or graph. He then tries to develop a theory regarding the action of the many small causes which in the aggregate produce the phenomenon under investigation, and finally he tests independently his hypothesis regarding the elementary causes. Thus, in thermodynamics the specific heats of gases, as determined by experiments, are first described as a function of the molecular composition of these gases, then a hypothesis is developed relative to the action of the individual molecules (e.g. the kinetic theory of gases) and finally the assumptions regarding the behavior of the individual molecules are tested by the physicist through independent experiments.