PREFACE

Every problem has a solution, and every solution has a problem. That's why there is speculation. It’s easier to do and has fewer problems. That’s also why, according to some scientists, philosophers speculate while scientists test and prove.

This is not the only view about speculation that has been advocated by great thinkers. A contrasting one is that spec­ulation is crucial when you are trying to come up with an explanation of a group of observed phenomena. For the ac­tivity to be scientific, however, it must be followed, as soon as possible, by an attempt to test and prove the speculation by experiment and observation. Proponents of this view say that the speculating stage should be subject to no rules or constraints whatever. Scientists should be given free license to invent even the wildest speculations. Rules and constraints enter in the testing stage.

A third view is much more liberal than the other two. It agrees with the second view in saying that speculation is crucial and is not subject to rules and constraints, but it disagrees with views one and two about testing. It says that speculation is crucial in science especially in the absence of testing and proof—indeed, when no tests have been made or even planned. It is an important way of finding fault with currently accepted theories, and it can lead to new ways of thinking that may turn out to be fruitful and even right.

What shall we make of these contrasting positions? To answer this question and others to be raised, we first need to decide what a speculation is.

I will start with scientific speculations—ones made by scientists about specific constituents of the physical world and the laws governing them, e.g., nineteenth-century speculations about the existence of light waves and an ether in which light is waving, and twentieth- and twenty-first- century speculations about the existence of strings vibrating in 10-dimensional spacetime. Later, I will argue in some detail that Newton's law of gravity, despite his vehement claims to the contrary, was indeed a speculation. I will also consider much broader speculations made by scientists or philosophers, or both, about the physical world and methods to be used in finding out about that world. These include the claim that nature is simple and that simplicity is an epistemic virtue (claims made by Newton and Einstein in support of theories they propose); that scientific theories can only be confirmed “holistically” and not by establishing individual propositions within them (Whewell, Duhem, Quine); and that there is and must be a “Theory of Everything”—a theory that can explain all phenomena by reference to fundamental laws governing the universe and fundamental objects in that universe (various physicists, especially string theorists, and various philosophers, especially those who preach a strong form of reductionism). These are all speculations, in a sense I will give to that term.

I have contrasted speculation with proof. But “proof” is too strong a term here because it suggests certainty. If speculations are understood simply as claims that have not been proved, too many scientific claims would be speculations. What scientists typically supply is evidence. And what they want, if they can get it, is evidence that provides a good reason to believe a claim they are making. If such evidence is lacking for a claim, then the latter might well be considered a speculation. At least that is the general idea of speculation I want to develop, make precise, and de­fend. To do so, I will need to talk about evidence itself, a con­cept I have examined in detail in other works.¹ “Evidence,” I argue, has several different senses, each of which can be defined by reference to a basic concept I call “potential evi­dence,” which I define using an objective epistemic concept of probability and a concept of “correct explanation,” which I also define.^{[1] [2] However, my purpose in the present work is not to develop or defend these concepts further but, rather, to use them to help us understand the idea of speculation, to show how speculations are to be evaluated as speculations, and to evaluate various “grand” and “less grand” speculations, including the ones mentioned above, that have been made by scientists and philosophers. For readers not familiar with my concepts and definitions of evidence and explanation, I will explain them briefly when they are introduced.}

William Whewell regarded speculation as crucial to sci­ence. The “tendencies of our speculative nature” lead the greatest scientists to produce the most important ideas.

James Clerk Maxwell, one of physics' greatest speculators, held strong pragmatic views about how and when to speculate and how to evaluate speculations as such. He put them into practice when theorizing about electricity and molecules. In this book, I will invoke some of Maxwell's speculations in physics, particularly those about molecules, and his phil­osophical views about speculating. Like him, I will take a pragmatic approach. (For what this means, and how such an approach works in the case of speculations, the reader is invited to keep reading.) Unlike Maxwell, Isaac Newton held a non-pragmatic opinion about speculation in his “Rules for the Study of Natural Philosophy” and “General Scholium” in Book 3 of the Principia. Briefly expressed: Prove, never spec­ulate! His practice, however, is somewhat different, as I will argue, even with respect to his greatest accomplishment, the law of gravity. Newton and Maxwell both engaged in specu­lation, and both defended important positions on the topic that are well worth examining. Their views are among those that will be considered in determining what attitude we should take toward speculation. Unlike the suggestion at the beginning, I will argue that speculation is an essential part of science, not just philosophy, and it is not easy to do, or at least to do well. But, by contrast with those scientists and philosophers who are in favor of speculation and hold the second and third views noted earlier, it is not, nor should it be, done freely and without constraints.