SPECULATION CONTROVERSIES REVISITED
meta-inductive, “only game in town,” or authoritarian. And both have in mind evidence that provides a good reason for believing a hypothesis, not just one that simply increases its probability.
Using A- or explanatory B-evidence in defining speculation will yield the kind of activity that Newton opposes and Whewell supports. Feyerabend, a champion of the “very liberal” view that encourages unbridled speculation (“proliferation”) in the absence of evidence, presents no particular view of evidence or of how to get evidence. I will construe his self-described “anarchist” view to encourage speculating where this at least includes speculation in the sense of (Scientific Spec).Understanding these conflicting claims using (Scientific Spec), how shall we respond to them? As for the “very conservative” view—the view which Newton puts by saying “hypotheses have no place in experimental philosophy”— although both Newton and Descartes promulgate this idea in their official principles, they violate it in their practice. In the Principia, Newton introduces two propositions he himself calls “hypotheses.” (Hypothesis 1 is “the center of the system of the world is at rest.”[42]) And at the end of his book Opticks, there are speculations about the particle nature of light that are clearly hypotheses in his sense. Here, after offering empirical proofs of various laws of geometrical optics, he raises the question of what light is and offers a speculative hypothesis: “Are not the rays of light very small bodies emitted from shining substances? For such bodies will pass through uniform mediums in right lines without bending into the shadow, which is the nature of rays of light.”[43] These are Newtonian speculations in the sense given by (Scientific Spec).
Descartes, as well, speculates in his Principles of Philosophy. After proving his general laws of motion (to his satisfaction)—laws regarding inertial motion and motion after impact—he introduces assumptions about the vortex motion of celestial bodies in the universe, saying “I will put forward everything that I am going to write [in what follows] just as a hypothesis.”[44] He then proceeds to “suppose that all the matter constituting the visible world was originally divided by God into unsurpassably equal particles of medium size...
that each turned round its own centre, so that they formed a fluid body, such as we take the heavens to be; and that many revolved together around various other points... and thus constituted as many different vortices as there now are stars in the world.” These also satisfy (Scientific Spec).Now, as these examples show, proof of a proposition or set of propositions about the behavior of some objects often leads to more questions about the nature of those objects and their behavior. It is natural for scientists to raise these questions and, when they do so, to think of possible answers. Refusing to do so is tantamount to stifling curiosity. How else could progress in science occur? Propositions don't usually come to the mind proved, and even when and if they do, additional questions will come to mind, some of which, at least, are accompanied with answers that are not proved. This happened with both Descartes and Newton, and is part of both everyday and scientific thinking. Indeed, humans could not survive without it. Despite pronouncements in their “official” methodologies to eschew speculations altogether, perhaps a more charitable interpretation of the positions of Newton and Descartes is this: Speculations are allowed, provided that they are clearly labeled as such, and provided they are not inferred to be true or probable on the grounds that, if true, they could explain, or help to explain, various known phenomena. If these conditions are met, why object to publicly communicating speculations, which in fact was done by both Newton and Descartes?
What about the “moderate” position (“Speculate, but verify”) typically taken by hypothetico- deductivists? There are several questions here. First, to raise a practical question, how long a period of time should be allowed between the speculation and the verification? Taking string theory as an example, and construing “verification” as providing (Newtonian or Whewellian) scientific evidence sufficient for belief, how long on the “moderate” view should string theorists be given to provide such verification of their theory—10 years, 20 years, 100 years? Is the view simply that until there is verification, the speculation is not to be believed? Or is it that until at least there is some reasonable prospect or plan for verification, the theory should not be seriously pursued (developed, worked out, taught in the classroom) by scientists or even regarded as “scientific”? Or, as a practical question, is the answer a subjective one, to be decided by the interests, finances, and patience of individual scientists?
More important, until it is tested, how, if at all, is a speculation to be evaluated? That is the main question I want to raise in the case of both the “moderate” view and Feyerabend's “very liberal” one (“Speculate like mad”).
Hypothetico-deductivists, as well as Feyerabendians, provide little information here. The “context of discovery,” where, according to hypothetico-deductivists, speculation is supposed to take place has few if any constraints placed upon it. Popper wants the speculations to be “bold.” Hypothetico-deductivists want the speculations at least to be potential explanations of the data or potential solutions to the problems prompting the speculation. Feyerabend is more liberal than this. For him, a speculation can be inconsistent with the data and can even reject the problems prompting the speculation. A speculation, as such, is not subject to any standards of evaluation.I reject all three views, not only because scientists do and must speculate but also because, as I will argue next, speculations can be evaluated in various ways other than by testing, and doing so is entirely appropriate. Giving “free rein to the imagination”—a favorite slogan of hypothetico- deductivists and Feyerabendians—sounds good, but isn't always the best policy.
12.