What to Do If You Want to Defend a Theory You Cannot Prove: a Method of “physical Speculation”

In 1875 the theoretical physicist James Clerk Maxwell published a paper in Nature entitled “On the Dynamical Evidence of the Molecular Constitution of Bodies.” In it he argues for the existence of molecules and for various claims of the molecular theory that he and others had been developing, including that molecules satisfy dynamical principles of classical physics. He does so without any experimental proof for his fundamental claims. Since he regards this as contrary to a prominent methodological view that the defense of a theory requires experimental proof, at the outset he announces that he will employ a different scientific method. It is designed for developing and defending theories that postulate objects that, at the time, cannot be observed, and that make claims about such objects that, at the time, cannot be demonstrated to be true by observation and experiment. Two questions are of special interest to Maxwell. First, can you use the method to develop and defend a theory about unobservables in a way that can make it possible to be justified in believing the theory (or at least the set of its central and distinctive assumptions) to be true, without being able to experimentally prove that it is true? Second, can you do so in a way that is sufficiently precise and complete to answer a range of questions about the unobservables postulated, even if you have no epistemic warrant for some of these answers? Maxwell gives an affirmative answer to both questions. He refers to his method as a “method of physical speculation.” He takes it to be different from an inductive method of a sort espoused by Newton and Mill, which he regards as too demanding for his purposes, and from “the method of hypothesis” (or hypothetico-deductivism), which he deems too weak.