1 All page references in this paper are to this work, unless footnoted otherwise. Mackie, 's most important previous papers on causation include: “Causes and Conditions”, American Philosophical Quarterly, 2 (1965),Google Scholar 245-64, reprinted in Sosa, E. (ed.), Causation and Conditionals, (Oxford: Oxford University Press, 1974);Google Scholar “Counterfactuals and Causal laws”, in Butler, R. J. (ed.), Analytic Philosophy, First Series (Oxford: Blackwell, 1962)Google Scholar, reprinted in Beauchamp, Tom L. (ed.), Philosophical Problems of Causation, (Encino, California: Dickenson, 1973);Google Scholar “The Direction of Causation”, Philosophical Review 55, (1966), 441-66. Mackie's views in the present work represent a substantial shift from many of the claims of these articles.

2 In effect, then, we shall not directly treat any of the matters discussed in chapters 4, 5, 6, 9, 10, and 11 of The Cement of the Universe., For many concerns, they are of course no less important than the issues in chapters 1, 2, 3, 7, and 8. But it is mainly within these latter chapters that the philosophically most controversial and far-reaching issues are discussed. Chapter 10 is also an important and controversial one. We do not treat it here both for reasons of space and because it is somewhat tangential to our defence of Hume. But see Rosenberg, Alexander “Concrete Occurrences vs. Explanatory Facts: Mackie on the Extensionality of Causal Statements”, Philosophical Studies, 31 (1977): 133-40.CrossRefGoogle Scholar

3 J. A. Robinson, “Hume's Two Definitions of ‘Cause’ ”,and D. C. Stove, “Hume, Probability and Induction”, both in V. C. Chappell (ed.), Hume (New York: Doubleday, 1966), 129-47 and 187-213, respectively. Robinson's claims have been criticised by, for example: Lesher, James “Hume's Analysis of ‘Cause’ and the ‘Two Definitions’ Dispute”, Journal of the History of Philosophy, 11 (1973), 387-92;CrossRefGoogle Scholar Gotterbarn, Donald “Hume's Two Lights on Cause”, Philosophical Quarterly, 21 (1971), 168-71;CrossRefGoogle Scholar McRae, Robert “Hume on Meaning”, Dialogue, 8 (1969), 486-91;CrossRefGoogle Scholar Beauchamp, Tom L. “Hume's Two Theories of Causation”,Archivfür Geschichte der Philosophie, 55 (1973), 281-300.Google Scholar

4 It is worth noting in this connection that Mackie's view on the “two definitions“ dispute reflects some inconsistency. In chapter 1 he suggests that the definitions of the Treatise, are repeated with “minor changes” in the Enquiry, (p. 3). But in chapter 2 it is claimed that in the Enquiry, the first definition is interpreted by Hume in a crucially different way (p. 30). Whether this change is major or minor has been the source of debate, and it is unclear on which side Mackie stands.

5 Space prevents the further elaboration of this interpretation here. One of us has, however, provided an account of it: Beauchamp, Tom L. and Mappes, Thomas “Is Hume Really a Sceptic About Induction?”, American Philosophical Quarterly, 12(1975), 119-29.Google Scholar In what follows textual citations from Hume refer to page numbers in Hume's Treatise, ed. Selby-Bigge, (Oxford: Oxford University Press, 1965),Google Scholar Enquiries, seconded. Selby-Bigge, (Oxford: Oxford University Press, 1962),Google Scholar and Keynes, J. M. and Straffa, P. (eds.), 36 An Abstract of, a Treatise of Human Nature, (Cambridge: Cambridge University Press, 1938).Google Scholar

6 Cf. the interpretation and refutation of the regularity theory in Dretske, Fred and Snyder, Aaron “Causal Irregularity”, Philosophy of Science, 39 (1972) 69–71,CrossRefGoogle Scholar and in their sequel, “Causality and Sufficiency: Reply to Beauchamp”, Philosophy of Science, 40 (1973) 288-291.

7 Davidson, who was the first to our knowledge to advance explicitly this detailed view of the implicit generality of singular causal statements, describes it as Hume's view. Cf. “Causal Relations”, journal of Philosophy, 64 (1967), 691-703, and especially section Ill. Hume's discussions of “secret causes” seem to us to bear the same message. Cf. Treatise, p. 132, Enquiries, sections 47 and 54.

8 Rosenberg, Alexander “Causation and Recipes : The Mixture as Before?”, Philosophical Studies., 24 (1973) 378-85.CrossRefGoogle Scholar

9 See below for the outlines of a Humean argument to this effect.

10 For a more extended account in this vein, cf. Rosenberg, Alexander “Propter Hoc, Ergo Post Hoc”, American Philosophical Quarterly, 12 (1975), 245-254,Google Scholar especially section V. The entire paper constitutes a defence of temporal priority as constituting causal priority.

11 For example, the event of the Titanic's striking an iceberg caused the event of the Titanic's sinking. But although no doubt the former event is causally prior to the latter, when described as “the event reported on page 69 of A Night to Remember”, it provides no explanation of the latter event.

12 Cf. the discussion of the so-called “sophisticated way” of filling out a counterfactual, pp. 5Sf.

13 Counterfactuals, (Cambridge: Harvard University Press, 1973).

14 “A Theory of Conditionals,” in Rescher, N. (ed.), Studies in Logical Theory (Oxford: Blackwell, 1968).Google Scholar

15 Lewis offers the following counter-example to transitivity: If Otto had gone to the party, then Ann would have gone. If Ann had gone, then Waldo would have gone. If Otto had gone, then Waldo would have gone. The fact is that Otto is Waldo's successful rival for Ann's affections. Waldo still tags around after Ann, but never runs the risk of meeting Otto. Otto was locked up at the time of the party, so that his going to it was a far-fetched supposition; but Ann almost did go. Then the premises are true and the conclusion is false. The fallacy of transitivity is treated on pp. 32ff in Counterfactuals.

16 Counterfactuals, p. 33.

17 Cf. Rosenberg, “Propter Hoc, Ergo Post Hoc”, section IV, in which this claim is substantiated.

18 Mackie's claim about continuous micro-processes underlying macro ones is just that underneath apparently discontinuous causal processes—like lighting a match—are motions of molecules that are simply Newtonian in character. In general, Mackie seems implicitly committed to the idea that the disaggregation of any macroscopic causal connection will reveal that this causal connection is simply the “summation” of rectilinear motions, and interactions that are persistent in the sense that they conserve things like momentum or energy, perhaps by transferring them from one molecule to another. If this is Mackie's point then the criticism offered in the next paragraph of this page is just as appropriate to such a claim about micro-persistence underlying macro-change (and thereby bestowing causal connectability on the latter) as it is to the claim that persistence is a priori expectable. Mackie's whole train of thought here suggests an appeal to “secret powers”, unobservable at the level of macroobjects, that result in causal connections. But, as Hume noted, the same sceptical questions may be raised about the connection between these secret powers and their effects that is raised about the connection between events at the macro-level. Following Hume, we conclude that causation is nothing more than constant conjunction no matter whether between bats and balls or photons and electrons—and all talk of persistences or quantities conserved only systematizes these connections on the basis of inductively justified theories.

19 Most of the literature in the Popper-Kneale debate, as well as Mackie's article, first published in Butler, R. J. (ed.), Analytical Philosophy, First Series (Oxford: Blackwell, 1966) pp. 65–80,Google Scholar are reprinted in Tom L. Beauchamp (ed.), op. cit.

20 We find perplexing this claim that Mackie's analysis of persistence provides “an empirical counterpart” to Kneale's rationalism. It seems to us that Mackie explicitly argues for the a priori, expectability of causal persistences, thus subscribing to Kneale's very position (d. pp. 219–21). It is only at the end of this discussion (p. 223)—almost as if he sees the closeness of his own conclusions to Kneale's—that Mackie retreats to the view that these persistences are not really a priori, after all, but are the “empirical counterparts” of Kneale's a priori persistences. But does this not mean that they are persistences whose existence represents an empirical discovery, and provide no intrinsic tie, in the objects, distinguishing causal and accidental sequencesllf so, then Mackie's argument will not refute Hume's claim that” anything may cause anything”. Yet this is the claim he must refute in order to defend the view that there is a tie that binds cause to effect.

21 Ayer, A.J. “What is a law of Nature?”, Revue lnternationale de Philosophie, 10 (1956), pp. 144-65.Google Scholar Reprinted in Tom L. Beauchamp (ed.), op. cit., pp. 77–95, and elsewhere.

22 Mackie, might, be willing to concede this point, an inference we draw from a brief statement he makes in “The Elusiveness of Causation—A Reply to Professor Flew”, Philosophical Books, 16 (1975), pp. 6–7.Google Scholar