^{page 319 note 1} Ian Barbour and George Schlesinger are but two that come to mind. Cf. Barbour's, Issues in Science and Religion (New Jersey: Prentice-Hall, 1966)Google Scholar, and Myths, Models, and Paradigms: A Comparative Study in Science and Religion (New York: Harper & Row, 1974);Google Scholar and Schlesinger's, ‘The confirmation of scientific and theistic hypotheses’, Religious Studies XIII (1977), 17–28CrossRefGoogle Scholar, and Religion and Scientific Method (Boston: D. Reidel, 1977).

^{page 319 note 2} Swinburne, Richard, The Existence of God (Oxford: Clarendon Press, 1979), p. 3.Google Scholar (Further page references will be given in the body of this paper.)

^{page no 320 1} Swinburne sums all of this up in terms of Bayes' Theorem, ‘which follows directly from the axioms of the mathematical calculus of probability, for the truth of which there are good independent grounds’ (64). He claims, however, that this adds rigour but not substance to his position and is not essential; I will take him at his word and keep to plain (?) English. (Another philosopher who views the structure of scientific confirmation as given by Bayes' theorem is Salmon, Wesley C., in The Foundations of Scientific Inference [Pittsburgh: University of Pittsburgh Press, 1967], pp. 115–32.CrossRefGoogle Scholar Salmon's use of Bayes' theorem is less sophisticated than Swinburne's but arrives at a similar result.)

^{page 321 note 1} That, for example, the properties of elementary particles areinexplicable physically or scientifically makes them brute facts of scientific explanation; for Swinburne, though, a ‘brute fact’ is one without any explanation, scientific or personal. He recognizes (76) that there will be brute facts in any explanation whose terminus is not a self-caused or logically necessary being (or state of affairs), but sees an agent's intentions as a more natural stopping place for explanation.

^{page 322 note 1} I take the phrase ‘in rational parity’ to include A's and B's being in competition with one another, seeking to explain the same range of phenomena.

^{page 323 note 1} The existence of conscious agents confirms theism via (E) because, Swinburne thinks, agents are not likely to have emerged out of a purely materialistic universe; it confirms theism via (D) because, while the existence of God does not entail that there be moral agents, their existence is likely given God's (presumed) character.

^{page 323 note 2} Popper, Karl, The Logic of Scientific Discovery (London: Hutchinson, 1959); cf. appendices *vii, *viii.Google Scholar

^{page 323 note 3} Swinburne, Richard, An Introduction to Confirmation Theory (London: Methuen, 1973), p. 46Google Scholar – as if that showed something about simplicity and not about scientists!

^{page 323 note 4} Goodman, Nelson, Fact, Fiction, and Forecast, 2n ed. (New York: Bobbs-Merrill, 1965).Google Scholar

^{page 324 note 1} Swinburne, , An Introduction to Confirmation Theory, p. 113; cf pp. 118–21.Google Scholar

^{page 324 note 2} It is true that this only pushes the problem back to the confirmation of those other cases, but that is just to say that ‘fit with background knowledge’ is only one criterion out of several.

^{page 324 note 3} If eating Babcock Hall ice cream is the infrequent, but necessary, cause of one's tongue turning blue, then while from seeing me eat some ice cream you could not predict that my tongue would turn blue, you could correctly explain the latter (if it occurred) by the former. Cf. Railton, Peter, ‘A deductivenomological model of probabilistic explanation’, Philosophy of Science XLV (1978), 206–26.CrossRefGoogle Scholar

^{page 325 note 1} Mackie, J. L., ‘The relevance criterion of confirmation’, British Journal for the Philosophy of Science XX ( 1969), 27–40.CrossRefGoogle Scholar

^{page 326 note 1} Achinstein, Peter, ‘Concepts of science’, Mind LXXXVII ( 1978), 22–45CrossRefGoogle Scholar, groups Swinburne with those who treat e as evidence for h if e increases the probability of h, regardless of any explanatory connection between the two (he references p. 3 of An Introduction to Confirmation Theory.) Surely with respect to The Existence of God (and, perhaps, later chapters of the former book), that is an inaccurate classification.

^{page 326 note 2} Which says, roughly, that whatever confirms a hypothesis h also confirms every hypothesis entailing h. Cf. §8 of Hempel's, Carl G. ‘Studies in the logic of confirmation’, reprinted in Aspects of Scientific Explanation (New York: Free Press, 1965).Google Scholar

^{page 326 note 3} Says Popper (p. 266): ‘We regard incompatibility as falsification of the theory. But compatibility one must not make us attribute to the theory a positive degree of corroboration: the mere fact that a theory has not yet been falsified can obviously not be regarded as sufficient. For nothing is easier than to construct any number of theoretical systems which are compatible with any given system of accepted basic statements’.

^{page 326 note 4} Salmon, , p. 26.Google Scholar

^{page 327 note 1} Popper, §§, 82–3. Salmon, p. 120, quotes a passage from Popper's Conjectures and Refutations (p. 112) that is remarkable in its similarity to his and Swinbume's programme: ‘we try to select for our tests those crucial cases in which we should expect the theory to fail if it is not true’, and a theory ‘is the better corroborated the less expected, or less probable, the result of the experiment has been’.

^{page 327 note 2} Point 9 of *ix, pp. 399–401; point to relativizes this with respect to background knowledge. (On p. 263n, Popper refers the reader to the above with respect to ‘the formal analogies between Bayer's theorem on probability and certain theories on degree of corroboration’.)

^{page 327 note 3} Mackie, , p. 32.Google Scholar

^{page 327 note 4} Mackie, , pp. 39–40.Google Scholar Cf. Popper's ‘tenthdesideratum’ of an adequate theory of confirmation, which says (in part) that if ˜ h is disconfirmed by e, then e confirms h (p. 406).

^{page 328 note 1} Cf. Yandell, Keith E., ‘Religious experience and rational appraisal’, Religious Studies X (1974), 173–87CrossRefGoogle Scholar, for criteria (with respect to the appraisal both of religious conceptual systems and of religious experience) which seem not to overlap with those discussed here. I am grateful to Professor Yandell for his comments on earlier versions of this paper.

^{page 329 note 1} Popper, , p. 59.Google Scholar