² For example, Moravcsik, J. “Aristotle on Adequate Explanation”, Synthese 28(1974), 3–18CrossRefGoogle Scholar and Hintikka, J., “On the Ingredients of an Aristotelian Science”, Nous 6 (1972), 55–69.CrossRefGoogle Scholar

³ See especially Brody, B., “Towards on Aristotelian Theory of Explanation”, Philosophy of Science 39 (1972), 20–31CrossRefGoogle Scholar, perhaps the first study to draw attention to the examples of asymmetries of explanation discussed by Aristotle. My account of how Aristotle solved the problem they pose is quite different from Brody's since he argues that essences and a causal relation in nature are indispensable parts of the solution. I do see essence as crucial to another aspect of Aristotle's theory of science; see below.

⁴ Brody, op. cit.; Boyd, Richard, Journal of Philosophy 73 (1976), 633–635CrossRefGoogle Scholar; and recent papers by Saul Kripke and Hilary Putnam – for bibliographical information and an insightful critique, see Mellor, Hugh, “Natural Kinds”, British Journal of Philosophy of Science 28 (1977), 299–312.CrossRefGoogle Scholar

⁵ See my “The Pragmatics of Explanation”, American Philosophical Quarterly 14 (1977), 143–150Google Scholar; I hope to develop this theory of explanation in a book, The Scientific Image (Oxford University Press, scheduled for 1980).Google Scholar

⁶ Quotes from the Posterior Analytics are from the translation by McKeon, R., Aristotle (Chicago: University of Chicago Press, 1973).Google Scholar

⁷ The terminology of “cause” or “reason” traditionally used to translate “aitia” may have misleading connotations today; Hope's translation of the Physics and Moravcik (op. cit.) use “explanatory factor” which seems more nearly apt.

⁸ Cf. Hintikka op. cit.

⁹ Aristotle asserts that these two are essentially the same question (P.A. 11, 2; 90^a, 15–20). Note in passing that he is not always equally sceptical about how quickly we can grasp the reason for something; in the immediately following passage (11, 90^a, 25–30) he says that if we were on the moon, watching an eclipse, “both fact and reason would be obvious simultaneously”. It would certainly be more obvious than it is from here.

¹⁰ The translation of the Physics I use is that of Hope, R., Aristotle's Physics (Lincoln, Neb.: University of Nebraska Press, 1961).Google Scholar

¹¹ Nor is this contrary to tradition. Note that the importance of the theory of questions at the beginning of Book II is given central place in the exposition of the whole Posterior Analytics in Wallace, W.A., Causality and Scientific Explanation, volume 1 (Ann Arbor: University of Michigan Press, 1972), 11–18Google Scholar. I mean to be audacious only in presenting the subject here in modern terms (specifically in terminology derived from Nuel Belnap's logic of questions).

¹² There is another place where Aristotle discusses causes, and says that they can be known independently of whether the effects occur, namely, where he describes the division of the sciences and marks them as more or less pure (P.A. 1, 13; 78^b; 79^a, 15). Thus related are, for instance, geometry, optics and the theory of the rainbow. The scientists engaged in the purer science “are in possession of the demonstrations giving the causes, and are often ignorant of the fact”; he says they know the reasoned fact. Thus our “knowing why the moon suffers eclipse” involves more than what he calls knowledge of the reasoned fact.

¹³ In the exhibited pattern, the proposition drawn from science implies the equivalence of the explained and that which explains. This would certainly be the case if the proposition was a definition; and it seems to be the typical case for Aristotle, who gives little attention to other cases, although some of his examples do not obviously involve an equivalence.

¹⁴ See Brody, op. cit. and note 3 above.

¹⁵ This point may be obscured because the other attribute in the example is complex: the light of the planets if emitted and received on earth is twinkly light. But this iffy property must be shared by any celestial body which is near to the earth, according to the example. I shall discuss this issue further below.

¹⁶ Cf. Physics II, 5; 196^b, 22 “Thus just as anything may be either essentially or accidentally, so explanatory events or factors may also be of either kind.”

¹⁷ See Code, Alan, “Aristotle's Response to Quine's Objections to Modal Logic”, Journal of Philosophical Logic 5 (1976), 159–186CrossRefGoogle Scholar. It should be clear from this example why Quine's mathematical bicyclist objection is not telling.

¹⁸ Cf. Thomason, R.H., “Species, determinants, and natural kinds”, Nous 3 (1969), 95–101.CrossRefGoogle Scholar

¹⁹ See Hintikka, op. cit.

²⁰ In saying this, I am to some extent speaking as devil's advocate. For now, note that if the arranger of the books were known to be a compulsive librarian, professionally addicted to the Dewey decimal system, then the question would be completely answered by giving the books' Dewey numbers. In short, the distinction between properties to be relied on for classification, and others – mere indicators or symptoms – which is an absolute distinction for Aristotle, may be relative to some aspects of context.

²¹ I do believe that such distinctions will be necessary in philosophy of language, and hence in logic, and indeed, in the logic of why-questions and of explanation. But they appear there as semantic distinctions, with (to my mind) no implications for ontology. See further my “Extention, Intension and Comprehension” in Munitz, M., Logic and Ontology (New York: New York University Press, 1973)Google Scholar, and the last chapter of The Scientific Image (note 5 above).