Abstract

I examine theism as an inductive hypothesis by critiquing four aspects of Richard Swinburne's seminal work, The Existence of God : use of induction, Bayes's theorem, simplicity as a test for theories, and the role of religious experience. Chapter one is a summary of the issues and outline of my work. ;Arguing against fideism, chapter two justifies the attempt to support theism. It shows that an inductive case for theism is acceptable, and that an argument to the best explanation is a form of induction. ;Chapter three examines Bayes's theorem, which Swinburne uses to unify his case. The theorem calculates the probability of an outcome given its probability both before considering evidence and after . It considers the likelihood of an outcome with and without the hypothesis , and the likelihood of the hypothesis itself, considering everything we know which is relevant . The possibility and method for affixing numbers to the elements within the formula depends on the underlying theory of probability. After surveying the major theories, I conclude that use of Bayes's theorem is acceptable in a theistic case given certain theories of probability and given limited claims for the posterior probability. ;Swinburne bases the prior probability of theism on its simplicity. Chapter four examines recent theories of simplicity and adds explanatory power to it to form the composite criterion of parsimony. ;Chapter five shows that Swinburne's use of the principle of credulity should be replaced by an approach that regards the veridicality of a religious experience as just one explanation for it which must compete with other explanations. ;Chapter six shows that the traditional evidentialist scheme on which Swinburne bases his case does not adequately account for the way presuppositions influence theories. The contrary views, hard-perspectivism and the presuppositionism of the Dutch Reformed thinkers , make presuppositions too influential by making them incorrigible. I argue for an intermediate view in which presuppositions interact with evidence: they influence interpretation of evidence but can themselves be influenced by evidence. ;Without rejecting Swinburne's traditional two-step case which tests the truth of a very simplified hypothesis about God, I argue for a one-step case in which whole detailed hypotheses are compared with each other to see which is the best explanation for the world.