THEORY EVALUATION: EITHER/ OR-ISM VS. PRAGMATISM

to have even that question in his “official” epistemic reper­toire. The important epistemic question pertains to proof or demonstration.

With regard to epistemic evaluations, Mill, Newton, and Whewell are “either/or” theorists.

For Whewell, just change “each assumption” in the pre­vious sentence to “the set of assumptions.” That is, either provide evidence (satisfying consilience and coherence) suf­ficient to believe the set of assumptions; or recognize that you are speculating (in which case, for Whewell, you may have done something “of service to science,” but you have not provided “proof”). To be sure, for Whewell, by contrast with Mill and Newton, you do not and cannot provide evi­dence sufficient to believe each assumption—only evidence sufficient to believe the theory as a whole. But with regard to method,” but not the “inductive” condition. Whewell says that if a theory in its early development stage explains and predicts phenomena but does­n't yet satisfy consilience and coherence, it can still be of value.

this disjunction, both sides agree: either you come up with evidence sufficient to believe the whole theory or you have a speculation.

There is a better alternative, I believe. It rejects the idea of focusing exclusively on epistemic evaluations, and even within the latter, focusing exclusively or primarily on either/or evaluations (“proved or unproved,” “possible or impossible”). Theory evaluation is much broader than that, a claim to which I now turn.

In chapter 1, I argued that a speculative theory such as Maxwell's kinetic-molecular theory of gases can be evaluated from different perspectives. Which perspective(s) to use in an evaluation is a pragmatic matter that depends on the knowledge, aims, and interests of the speculator and the evaluator. One such perspective, a completely historical one, is non-epistemic. From such a perspective, Maxwell's theory gets high marks, since it set the stage for mathematically rig­orous molecular investigations and it introduced statistical considerations in generating the Maxwell distribution law. As the historians of science Garber, Brush, and Everitt note, when Maxwell died in 1879, his contemporaries regarded his work on kinetic theory as more important than his work on electricity and magnetism (a view later reversed during the twentieth century.) But, these historians add, “we should take these nineteenth-century assessments seriously if we are to judge the role of kinetic theory and Maxwell in the devel­opment of modern physics.”^[129]

Even if we confine our evaluations to epistemic ones, in practice such evaluations are often more complex than those whose principal evaluative categories are “proved,” “unproved,” and “possible.” '1 his is particularly so when the theory is a “work in progress” and is evaluated as such. It is not my claim that the evaluations in question are always “correct” or justified. In general, evaluations, whether ep- istemic or non-epistemic, can be questioned and are sub­ject to defense. My claim is that evaluations—epistemic or otherwise—are in general much more nuanced and informa­tive than those offered by the holists and particularists I have been considering.

As I noted in chapter 1, in his 1875 paper, “Evidence for a Dynamical '1 heory of Gases,” Maxwell presents a largely epi­stemic evaluation of a theory he himself calls a “physical spec­ulation.” He assesses his theory, and the various assumptions within it, by referring to the epistemic reasons he has for them at the time in question. Some of these reasons, he claims, are strong enough to rise to the level of evidence sufficient to believe some assumptions (e.g., the existence of experiments which he thinks establish that heat is a form of motion of “parts too small to be observed separately”). Some epistemic reasons he offers for certain assumptions are weaker and plausible, but not sufficiently compelling to believe those assumptions (e.g., weak inductive reasons arguing from the success of mechanical theories in astronomy and electricity to their success in molecular-kinetic theory, and hence to the claim that molecules satisfy mechanical Newtonian laws^[130]). And some reasons are “explanatory” ones showing how, if true, the assumptions could explain various phenomena. The latter by themselves don't count as evidence for the truth of the assumptions of the theory. But in showing how, if true, the theory could explain various phenomena, Maxwell is showing that the theory is epistemically possible—i.e., pos­sible given what is known.

For some assumptions, no epistemic reasons at all are given, only ones such as simplicity of calculation (e.g., that the coordinates of molecular velocity vectors are inde­pendent). In addition, there are what he calls “unconquered difficulties.” The latter include consequences of the theory that are incompatible with observations (e.g., specific heat ratios of gases), as well as the inability of the theory to ex­plain various properties of gases (transparency, electrical conductivity). Maxwell writes:

But while I think it is right to point out the hitherto uncon­quered difficulties of this molecular theory, I must not forget to remind you of the numerous facts which it satisfactorily explains. We have already mentioned the gaseous laws, as they are called, which express the relations between volume, pressure, and temperature, and Gay-Lussac’s very important law of equivalent volume.

molecules exist, then, given that mechanical theories have been successful in astronomy and electricity, it is reasonable to believe that they are subject to mechanical laws. The success of mechanical theories does not, of course, provide a reason to believe that molecules do exist, or that they satisfy the particular mechanical assumptions introduced pertaining to their shapes, motions, and forces exerted.

of these inquiries is a more distinct conception of thermal phenomena.^[131]

From an epistemic perspective, the theory is doing rea­sonably well so far, Maxwell is saying, even though there are problems with it and further development will be required. The epistemic reasons offered vary considerably in their strength. Even if Maxwell is right in claiming that some are sufficiently strong to rise to the level of “experimental proof,” this is not so for the theory as a whole or for each assump­tion of the theory. But some of the reasons, while not strong enough, or of the right sort, to count as evidence for, or “proof” of, unconditional (or even conditional) assumptions of the theory, do carry some epistemic weight. Maxwell ob­viously regarded the theory as a whole as unproved, since he classified it as a speculation. But “proved,” “not proved” (“possible” and “not possible”) are not the only informative categories of epistemic evaluation for the theory, or for parts of it, even if the theory as a whole and some of its parts are “not proved” or “not possible.”

Other examples of such more varied and nuanced ep- istemic evaluation are ones given of the wave and par­ticle theories of light by John Herschel in 1827, by Baden Powell in 1833, and by Humphrey Lloyd in 1835.^[132] 'lhese involve showing how each theory explains various known optical phenomena, giving the advantages and disadvantages of each.

I am not claiming that all epistemic theory evaluations have to be “Maxwellian” (or “Powellian”), only that they needn't all be assessments from a “proved/not proved,” or “sufficient/insufficient evidence,” or “possible/not possible” perspective. '1 hey can be ones that indicate what level of sup­port the theory and its individual assumptions have, whether or not this rises to the level of “proof,” and why this does or doesn't rise to that level. Evidence may be presented for some but not all individual hypotheses, and defenses of evi­dential claims may be offered that don't appeal to the entire set of beliefs in a scientist's epistemic situation. Moreover, individual hypotheses of a theory may be evaluated from the viewpoint of the theorizer and his epistemic situation, or from the viewpoint of the evaluator and his epistemic situation. If we offer an epistemic evaluation of the mid­nineteenth-century wave theory from our perspective today, we will obtain a very different result from those of Herschel, Powell, and Lloyd.

Accordingly, evidential particularists are correct in claiming that individual hypotheses can be evaluated episte- mically, and holists are correct in claiming that the theory as a whole can be evaluated epistemically. But holists are mistaken in their claim that only the theory as a whole, not particular parts, can be evaluated.

Finally, no matter whether the evaluation is detailed and nuanced or more global, scientists can and do use assumptions of a theory to explain, predict, and calculate, even if the theory as a whole is not “proved,” even if some or many of the assumptions are not either—indeed, even if no support for any of the assumptions has been offered. Scientists do not, and should not, wait for evidence that provides “ho­listic” proof of the assumptions or particularist proof of each assumption in order to proceed to provide explanations, predictions, and calculations. That would stifle inquiry. It doesn't follow, of course, that the explanations, predictions, and calculations offered are correct or fully warranted. They, like the assumptions of the theory, are subject to a range of evaluations.