Reinterpreting Philo's Critique of the Cosmological Argument in Hume's Dialogues — Epoche C1
Which argument, and who makes the objection In Part IX of David Hume's Dialogues Concerning Natural Religion , published posthumously in 1779, the character Demea sets out an argument that the chain of causes in the world must terminate in a being who exists necessarily — and the objections that demolish it are put not into Philo's mouth but into Cleanthes'. That misattribution is worth clearing up at the start, because the essay's thesis depends on distinguishing what Philo actually attacks from what he is credited with attacking, and the distinction turns out to sharpen rather than weaken the claim that his critique is more radical than a hunt for logical slips. Two arguments must be kept apart, since a compressed account of the Dialogues can suggest they are variants of one another on the ground that both make design inferences. They do not. The teleological argument, or argument from design, reasons from the observed adaptation of means to ends in nature — the eye to seeing, the parts of an animal to its survival — to an intelligent designer, and it is an argument from experience by analogy with human contrivance. The cosmological argument reasons from the mere existence or contingency of the world, without appeal to any observed orderliness, to a first cause or a necessarily existing being; it makes no design inference at all. Cleanthes, in the Dialogues , is the advocate of the first; Demea, in Part IX, offers the second in the form given it by Samuel Clarke in his Boyle Lectures of 1704, published in 1705. Philo, the sceptic, opposes both, but not in the same way and not to the same degree. Part IX, and what Cleanthes says there Taking the cosmological argument first, since the attribution needs correcting before anything is built on it: Demea argues that whatever exists must have a cause, that an infinite regress of causes is itself something requiring explanation, and that only a being carrying the reason of its existence within itself can end the series. The reply comes from Cleanthes, and it has three moves. First, nothing is demonstrable unless its contrary implies a contradiction, and whatever we conceive as existing we can equally conceive as not existing, so there is no being whose non-existence is contradictory — which removes the sense in which any being could be necessary. Second, if the notion of necessary existence were intelligible, the material universe itself might satisfy it, and the argument would deliver no God. Third, and most quoted, once the cause of each member of a series has been given, it is unreasonable to ask further for the cause of the whole: the whole is nothing over and above its members, and to demand a further cause is to treat a collection as though it were an additional item. Philo's contribution in Part IX is a short addendum, and it is a different sort of point. He observes that the products of nine, when their digits are added, always yield nine or a multiple of nine — a regularity that looks like design to someone ignorant of arithmetic but is bare necessity. The suggestion is that the order of the universe might likewise flow from some unknown necessity in matter, so that the appearance of contrivance is no evidence of a contriver. This is not an objection to the cosmological argument's logic; it is the offer of a rival hypothesis of the same explanatory adequacy. The compressed claim that Philo convicts the cosmological argument of violating the principle of sufficient reason by exempting its own terminus therefore misdescribes both the speaker and the objection. Cleanthes' third move is not that the argument violates that principle but that it applies it where it has no work to do. What Philo does attack, and the internal reading of it Philo's sustained assault is on Cleanthes' design argument, and the familiar catalogue of his objections is what invites the reading this essay disputes. In Part II he insists that the argument is an argument from experience and must be judged as one: we know that houses have builders because we have repeatedly seen houses built, and we have seen nothing of the kind for worlds. In Part V he presses the principle that a cause may be credited with no more than the effect requires, and extracts from it a series of embarrassments — that the designer might be one of many, working as shipwrights and architects do in teams; that the designer might be an infant deity whose first attempt this is, or a superannuated one long since dead; that many worlds may have been botched before this one. In Part VII he observes that the universe resembles an animal or a vegetable rather more than it does a watch, so that generation or vegetation would be as apt a model of its origin as contrivance. Read as a list, these look like internal criticisms: they concede the form of the inference and quarrel with its strength or with the attributes it can deliver. On that reading Philo is a careful auditor of an argument he broadly accepts the shape of, and a proponent might hope to answer him by tightening the analogy or moderating the conclusion. This is the reading the essay contends is insufficient, and the ground for that contention is in Hume's positive philosophy rather than in the Dialogues at all. The external reading: Hume's account of causal inference The bridge is Hume's analysis of what a causal inference is, set out in the Enquiry Concerning Human Understanding of 1748. On that analysis, no inference from a cause to an effect or back again is available a priori : reason cannot discover, in the idea of one object, the idea of another it must produce. What licenses such an inference is experience of constant conjunction — the repeated observation of pairs of resembling objects standing in the relation, which builds the habit on which the inference runs. The strength of any particular causal inference is a function of how closely the present case resembles the pairs already observed. Apply that standard to the design argument