Holism vs. Particularism

Evidential holism is the view that individual hypotheses do not receive evidential support, only entire theoretical systems do. Evidential particularism is the view that within a theoretical system each hypothesis can and must receive evidential support for the system to be established. Both views are examined, the former as expressed by Whewell, the latter as expressed by Mill. Although each view is different, both adopt an “either/or” attitude toward theory evaluation: either you prove the whole theory, or it remains a speculation. In this chapter, the holist viewpoint is seriously challenged and a form of particularism is defended. However, the “either/or” attitude of both views is rejected in favor of a more nuanced, contextually based method of theory evaluation.

The Complex Story of Simplicity

Both Newton and Einstein claim that nature is simple and that simplicity is an epistemic guide to truth. This chapter examines various arguments for these claims, including that, historically, simpler theories have been more empirically successful than complex ones; that the epistemic value of simplicity can be demonstrated by appeal to Bayes’s probability theorem; and that simpler strategies for changing one’s beliefs in the light of new evidence can be shown to be more successful than complex ones. It is concluded that none of these arguments shows that nature is simple or that simplicity is an epistemic virtue.

Summing Up

This chapter provides a summary of the results of this investigation into the meaning, legitimacy, and importance of speculating within and about science.

The Ultimate Speculation

Some scientists and philosophers claim that a “Theory of Everything” (TOE) exists and that scientists should find it. Such a theory will explain everything on the basis of fundamental laws and fundamental constituents of the universe to which everything else is reducible. This chapter clarifies what a TOE is supposed to be and do, and examines arguments in favor of the idea that there is a TOE and it should be found. These arguments involve claims from the history of science, claims about what science presupposes, claims about unification in science, and others. In response, this chapter shows not only that these arguments fail to establish their claims, but also that they fail to establish the desirability of a TOE. Contrary to what TOE enthusiasts insist, the intelligibility of the world does not depend on finding a TOE and using it to explain what scientists want and need to explain. Intelligibility is a local and contextual matter.

Non-Epistemic Simplicity

This chapter examines five claims about simplicity: (i) that theories are underdetermined by evidence, and so must be selected on the basis of simplicity; (ii) that to do science you must presuppose that nature is simple; (iii) that it is the aim of science to present simple theories; (iv) that simplicity, like beauty, is a virtue worth having for its own sake; and (v) that simplicity is primarily a pragmatic virtue. Objections are raised either refuting or seriously weakening the first four claims. The fifth claim, the pragmatic one, is defended and illustrated by showing how James Clerk Maxwell employs simplicity pragmatically in his molecular theory of gases. It is also shown that, despite what Newton claims when he invokes epistemic simplicity in his argument for universal gravity, simplicity does no epistemic work for him. Despite what he claimed, his law was a speculation.

Scientific Speculation

The chapter presents a definition of speculation in terms of concepts of evidence the author has introduced in other works. This definition is used to examine three popular views about scientific speculation: never speculate (Newton), speculate freely but test (Whewell), and speculate freely even without testing (Feyerabend). All three views are rejected, and in their place a pragmatic position, suggested by the work of James Clerk Maxwell, is defended. According to the latter, whether and how a scientist should speculate depends on the aims of the investigation and on standards, both epistemic and non-epistemic, different from those utilized for experimental proof.