^{page 175 note 1} I am grateful to Professor Peter Bertocci for his careful and insightful criticisms of an earlier version of this essay and for stimulating discussion of its central ideas, as well as to Professor Erazim Kohak for reading the essay and making helpful suggestions.

^{page 175 note 2} Hartshorne, Charles, A Natural Theology For Our Time (LaSalle, Illinois: Open Court, 1967), p. 67.Google Scholar Hereafter referred to as NTT.

^{page 176 note 1} See Popper, Karl, Conjectures and Refutations (New York: Harper and Row, 1963), pp. 52 ff.Google Scholar

^{page 176 note 2} Ibid. p. 67.

^{page 176 note 3} Ibid. pp. 50, 53.

^{page 176 note 1} Hartshorne, , NTT, pp. 52, 53.Google Scholar

^{page 177 note 2} Ibid. p. 53.

^{page 177 note 3} Ibid. p. 67.

^{page 177 note 4} Popper, Karl, Conjectures and Refutations, op. cit. pp. 34, 35.Google Scholar

^{page 178 note 1} However, the problems of establishing the limit for using ad hoc hypotheses to save an established theory, and the extent to which our experiences are pre-interpreted by the theories we hold, must be taken seriously by empirical falsificationists. Unfortunately Hartshorne fails to discuss these problems in recommending a falsificationist empiricism (see p. 179 of this essay).

^{page 178 note 2} It is Henry Veatch who recently suggested that the cosmological argument may be cast as a straightforward modus ponens argument moving from the observed to the unobservable, if God is regarded as having the same status as an unobservable entity in science. We shall have occasion to object to the narrowing of empiricism to sensory observation suggested by regarding God as an ‘unobservable entity’ later in this essay. But the logical form of the argument as stated by Veatch seems correct. See Veatch, Henry B., ‘A Case for Transempirical and Supernaturalistic Knowledge Claims’, in Mind, Matter, and Method: Essays in Philosophy and Science in Honor of Herbert Feigl, Feyerabend, Paul K. and Maxwell, Grover, eds. (Minneapolis, Minnesota: University of Minnesota Press, 1966), pp. 402–4.Google Scholar

^{page 178 note 3} Carl Kordig argues for this translation of Veatch's modus ponens argument into modus tollen form in ‘Falsifiability and the Cosmological Argument’, New Scholasticism, 46 (1972), 485–7.CrossRefGoogle Scholar Both Veatch's and Kordig's statements of the cosmological argument suggest that God's existence is at least one question of metaphysics which may be falsified by empirical evidence (although Veatch qualifies this), thus blurring Popper's demarcation between science and metaphysics. Even if Popper himself would reject this application of empirical falsificationism to a metaphysical question, his rejection is unavailable to Hartshorne, for Hartshorne himself originally recommended such an application if it could be shown possible.

^{page 179 note 1} Any use of ad hoc hypotheses at all seems to threaten transformation of falsificationism into inductivism, where the maxim ‘accept the theory which is so far unfalsified’ must become ‘accept the theory with the fewest number of falsifications’ (i.e., the greatest number of confirmations?). And overuse of ad hoc hypotheses insulates a theory from empirical falsification altogether. The admission of a role for ad hoc hypotheses thus threatens to turn Popper's falsificationist philosophy of science into Thomas Kuhn's theory of science as a series of a prioristic paradigms insulated from empirical critique. See the exchange between Popper, and Kuhn, in Criticism and the Growth of Knowledge, Lakatos, Imre and Musgrave, Alan, eds. (Cambridge University Press, 1970).Google Scholar

^{page 180 note 1} Brightman, Edgar Sheffield, Person and Reality: An Introduction to Metaphysics, Bertocci, Peter A., ed. (New York: Ronald Press, 1958), p. 23.Google Scholar

^{page 180 note 2} Bertocci, Peter A., The Person God Is (New York: Humanities Press, 1970), p. 68.Google Scholar

^{page 180 note 3} Although American pragmatism and personalism have greatly broadened and redefined ‘empiricism’, I prefer the term ‘experientialism’ in this context to that of ‘empiricism’ due to the latter's continuing connotation of sense-empiricism dating from British empiricism.

^{page 181 note 1} See Bertocci, Peter A., The Person God Is, op. cit. p. 157.Google Scholar

^{page 181 note 2} This move from contingent parts to a contingent whole is not a logical fallacy. See note 4 below.

^{page 181 note 3} In a forthcoming book, The Cosmological Argument (Princeton: Princeton University Press, 1975)Google Scholar, William L. Rowe argues for the thesis that all forms of the cosmological argument appeal to the necessity of some version or other of the principle of sufficient reason. Reichenbach, B. R. also argues this case in his The Cosmological Argument: A Reassessment (Springfield, Illinois: Charles C. Thomas, Publisher, 1972), pp. 51–3 and 67–70.Google Scholar

^{page 181 note 4} I believe that the entire literature accusing the cosmological argument of committing the fallacy of composition (eliminating the hypothesis of an infinite series of contingent events by arguing that the series in its sum would still be contingent) is another way of framing this question about the range of application of the principle of sufficient reason.

This may be shown as follows: I think it plain that even on the level of sense-examination, the fallacy of composition (or division) is not fallacious in every instance, meaning that we only know if composition is a fallacy by examining the properties of the parts and the properties of the whole. But further, if we are not bound by a narrow method, we need not limit our hypotheses about wholes and their parts to those wholes we can inspect (take apart and put together) but may frame experientially coherent hypotheses about these wholes. Thus, the fallacy of composition criticism ultimately comes down to a question regarding the range of applicability for the principle of experiential coherence or sufficient reason.

The criticism is levelled by Munitz, Milton K., The Mystery of Existence (New York: Appleton-Century-Crofts, 1965), pp. 116 ff.Google Scholar, and more recently by Prado, C. G., ‘The Third Way revisited’, New Scholasticism 45 (1971), 495–501.CrossRefGoogle Scholar This charge is well answered by Edwards, Rem B., ‘Another Visit to the “Third Way”’, New Scholasticism 47 (1973), 100–4CrossRefGoogle Scholar, as well as Bertocci, , op. cit. pp. 178–81.Google Scholar

^{page 182 note 1} See Taylor, Richard, Metaphysics (Englewood Cliffs, New Jersey: Prentice-Hall, 1963), p. 105.Google Scholar In Taylor's terms, although the principle of sufficient reason can be denied without fear of straight-forward refutation, it is a datum which ‘all men, whether they ever reflect upon it or not, seem more or less to presuppose’.

^{page 182 note 2} I use the phrase ‘necessary being’ as a shortened version of the more Hartshornian phrase ‘necessary aspect of the divine being’. This is not meant as an elimination of the contingent aspect of God's existence without argument, but a terminological simplification.

^{page 182 note 3} Hartshorne, , NTT, p. 51.Google Scholar

^{page 183 note 1} See Hartshorne, Charles, The Logic of Perfection (La Salle, Illinois: Open Court, 1963), p. 53.Google Scholar

^{page 184 note 1} For example, J. J. C. Smart likens the concept of a necessary being to that of a round square. See his ‘The Existence of God’, in New Essays in Philosophical Theology, Flew, and MacIntyre, , eds. (London: SCM, 1955), p. 38.Google Scholar

^{page 184 note 2} Hartshorne, , NTT, p. 62.Google Scholar

^{page 184 note 3} For support of this analysis of ontological necessity, see Reichenbach, B. R., The Cosmological Argument, op. cit. pp. 107 ff.Google Scholar

Further, this distinction in the meaning of necessity that I am defending is the same as the distinction which Alvin Plantinga makes between modality de dicto and modality de re. In his recent book, The Nature of Necessity (Oxford: Clarendon Press, 1974)Google Scholar, Plantinga traces the distinction to Aristotle's Prior Analytics and St Thomas's Summa Contra Gentiles, as well as to G. E. Moore and apparently Norman Malcolm of more recent times. As Plantinga has it, modality de dicto attributes necessary truth to a proposition, that is, predicates the modality of necessity of a proposition. For example, an assertion of modality de dicto is ‘Necessarily nine is composite’, where necessary truth is predicated of another dictum or proposition. However, we may also ascribe to some object the necessary or essential possession of a property. For example, St Thomas argues that the proposition ‘It is necessarily true that whatever is seen to be sitting is sitting’ is true taken de dicto. But if taken de re it is false: ‘Whatever is seen to be sitting has the property of sitting necessarily or essentially’. To have a property necessarily means that under no possible circumstances could the object fail to possess that property. The defence of the de dicto and de re distinction is difficult and must be closely argued. See especially pp. 9–13 and 27–43 of Plantinga's book. For our purposes, it is enough to point out that acceptance of de re necessity, or essential attribution of properties, is a concomitant of empirical arguments for God. Perhaps it is the chief merit of Hartshorne's logical critique to have brought this into the open.

^{page 186 note 1} At one point, Plantinga explains the distinction between modality de dicto and modality de re with this interesting turn of phrase. Cf. Plantinga, , op. cit. p. 27.Google Scholar

^{page 186 note 2} Hartshorne, , NTT, p. 79.Google Scholar

^{page 186 note 3} Ibid. p. 85.