Abstract

The underdetermination thesis , the claim that every theory has a rival that is equally reasonable to believe given all the possible evidence, is central in the debate between Scientific Realists and Scientific Anti-Realists. We address central arguments on both sides and argue for the truth of a certain version of the thesis. ;Scientific Realists argue that UD is false because the class of facts relative to which the rivals are compared can change in unforeseeable ways because: the class of observable facts is unstable , and the dependence of the derivation of consequences about the observable on auxiliary theories. We show first that being observable should be distinguished from our belief that a certain fact is observable, and thus we should care about whether theories have rivals given what is actually observable. We also argue that it is sufficient to compare theories with respect to what they say about a small class of facts that are observable in our world, and are manipulable by observers. In order to avoid problems that arise due to the possible appeal to auxiliary theories we propose to examine UD with respect to total fundamental theories. As a result, the Scientific Realists' worries can be avoided when we examine the following version of UD: UD If the true total fundamental theory of our world is X, then it has a total fundamental rival that entails exactly the same consequences about pointer readings. ;We show that this is true when X is the Quantum Formalism: Bohm's Theory and the Many Worlds Interpretation are empirically equivalent in that sense. ;Scientific Anti-Realists argue that there are many ways to generate empirically equivalent rivals. We show that the main theories of confirmation provide grounds to dismiss the proposed rivals because they are considerably less well-confirmed. We argue, however, that it seems unreasonable to assume that the eventual "ideal" theory of confirmation will guarantee that a choice can be made between any pair of empirically equivalent rivals. In the Quantum Mechanics case, we examine one argument against the reasonability of the initial probability distribution in Bohm's Theory and show why it is wrong