HISTORICAL REASONS FOR REJECTING ATOMISM
We need to consider reasons that certain scientists rejected atomic-molecular theory until the first decade of the twentieth century. These reasons were
32.For example, atomic theories, as proposed, were purely mechanical theories that should entail reversible processes, but the latter are incompatible with observed thermodynamic phenomena.
33.Whether Maxwell and Boltzmann were in fact doing this is questionable. For a discussion of Maxwell’s reasoning, see my Particles and Waves and chapter 14 in this book. known to Perrin, who responded to them. Critics of atomic theory included Duhem, Ernst Mach, Friedrich Wilhelm Ostwald, Henri Poincare, and Max Planck. Grounds cited for rejecting the theory were in part scientific.32 But they also included important philosophical or methodological reasons. One was the claim that physicists such as James C. Maxwell and Ludwig Boltzmann, who proposed atomic-molecular theories, were employing an illegitimate “method of hypothesis.” From unproved hypotheses about atoms and molecules, they were deducing observable consequences and claiming that from the truth of the latter, one could infer the truth of the hypotheses themselves.33 While Mach, for example, did not repudiate the use of hypotheses, he did reject the idea that one could infer their truth or probability from true observed consequences. He regarded the hypotheses of the atomic theory as provisional “mental artifices” for summarizing known observational facts and facilitating observational predictions.[137] Once such hypotheses have served their organizational and predictive purposes, they are to be discarded, not accepted as true or probable. Indeed, Mach championed an antirealist “sensationalist” view about observable matter, which he regarded as complexes of sensations.[138]
A related philosophical claim made by some of the critics of atomic theory (whether or not they were Machian reductionists) was that hypotheses about unobservables can never be established as true or probable by empirical means; such hypotheses can never be known to be true.
For example, Poincare[139] speaks of the atomic hypothesis as “indifferent,” meaning that although it may be useful, it can never be empirically established or shown to be more probable than rival hypotheses that assert the continuity of matter (pp. 152-153). For Duhem,[140] [141] as noted earlier, science can know only “sensible appearances and can discover nothing beyond them” (p. 304). Unlike Poincare and Mach, Duhem does not even regard atomic hypotheses as useful devices for summarizing and predicting observable phenomena.Like some of the critics of the atomic theory, and no doubt because of their criticism, Perrin offers his own general philosophical/methodolog- ical reflections on how to proceed with scientific investigations. He distinguishes two scientific methods, which he calls the inductive method and the intuitive method.38 The former, associated with the critics of atomic theory previously noted, begins with what is observed and reasons only to statements about “objects that can be observed and to experiments that can be performed” (p. vii). The second method infers the existence of an unobservable mechanism:
In studying a machine, we do not confine ourselves only to the consideration of its visible parts, which have objective reality for us only as far as we can dismount the machine. We certainly observe these visible pieces as clearly as we can, but at the same time we seek to divine the hidden gears and parts that explain its apparent motions.
To divine in this way the existence and properties of objects that still lie outide our ken, to explain the complications of the visible in terms of invisible simplicity, is the function of intuitive intelligence which, thanks to men such as Dalton and Boltzmann, has given us the doctrine of Atoms. This book aims at giving an exposition of that doctrine. (p. vii)
Perrin claims that in the times in which he is writing, the method of intuition has gone ahead of induction rejuvenating the doctrine of energy by the incorporation of statistical results borrowed from atomists (p.
viii).In these introductory passages, Perrin does not describe the intuitive method in detail other than to say that it is a method for inferring an invisible underlying reality from visible things and events in such a way that the former explains the latter. I suggest that we take his own arguments for molecules from experiments on Brownian motion as representing his use of the “intuitive method.” If so, then that method is not simply a matter of speculating about an invisible realm. Nor is it a Machian provisional “mental artifice” for representing the facts, to be discarded once it has served its purpose. Nor is it a method simply for inferring the truth of claims about what can be observed (that is what he calls the inductive method). Nor, by contrast with Mach, does Perrin say or imply that assertions about ordinary matter are to be understood as claims about sensations, or (with later logical positivists) that assertions about an invisible realm of molecules are to be understood as claims about a visible realm such as that of Brownian motion, or that when he infers that molecules are real and that certain claims about them are true, by “real” and “true” he means “useful” or “saves the phenomena.”
Finally, Perrin's intuitive method is not the method of hypothesis, which infers the truth or probability of a hypothesis simply from the fact that the hypothesis explains and predicts observable phenomena. To be sure, Perrin's full argument involves citing explanatory chemical reasons from combinations of elements and compounds. But in addition, it involves eliminative-causal reasoning of a sort indicated in section 2. It involves an appeal to similar determinations of Avogadro's number N from a variety of experiments on phenomena other than Brownian motion. And perhaps most important, it involves a calculation of N from Perrin's own experiments on Brownian motion and a demonstration that this number remains constant (proposition H' in section 2). The former arguments, according to Perrin, show that the molecular hypothesis is highly probable. The latter sustains and even increases this probability on the basis of new, precise, experimental results.
Now we are in a position to respond to Salmon's challenge.
7.