Philo's Cosmological Analogy as an Epistemological Critique in Hume's Dialogues — Epoche B2
An Argument from Experience: Philo's Alternative Cosmogonies in Hume's Dialogues The claim under examination Hume's Dialogues Concerning Natural Religion [1] , published posthumously in 1779, is often summarised as showing that the design argument rests on a weak analogy. Against that summary a stronger reading is sometimes offered: that Philo's alternative cosmogonies — the world as an animal, the world as a vegetable, the world as a chance settling of matter — amount to an attack on empiricist epistemology itself. This research note argues that the stronger reading is wrong, and wrong in an instructive way. Philo's alternatives are not launched from outside empiricism. They are an application of Cleanthes's own rule for arguing from experience, run correctly, and the reason they damage Cleanthes is that he does not run it correctly himself. Getting this right matters for reading Part XII, where Philo apparently concedes the argument. On the reading defended here that passage is not a reversal but the argument's natural terminus. Part II: the rule Cleanthes proposes Cleanthes states the design argument as an argument a posteriori. The world is a machine subdivided into lesser machines, all adjusted to one another with an accuracy that resembles, though it exceeds, human contrivance; like effects prove like causes; therefore the cause of the world resembles a human mind. Everything here is offered as ordinary empirical inference, which is why Cleanthes thinks it superior to the a priori arguments of the schools. Philo agrees to the method. He argues only about how the method applies. His two objections in Part II both concern the conditions under which an inference from experience is warranted. The first is that causal inference rests on observed constant conjunction between two species of objects. We have observed houses arising from architects many times. We have not observed universes arising from anything, because we have seen one universe and its origin is not among our observations. The second objection concerns degree: wherever you depart, in the least, from the similarity of the cases, you diminish proportionably the evidence (Philo, Part II) This is not scepticism about experience. It is a constraint drawn from experience about how much an analogical inference can carry. A strong resemblance licenses a strong conclusion; a faint one licenses a faint conclusion. Keep this rule in view, because Philo obeys it for the rest of the book. Part V: running the inference forward Part V is the decisive evidence that Philo is not attacking empiricism. He accepts Cleanthes's principle that like effects prove like causes, and then insists on all of it. If the world is an artefact and we infer its maker from its qualities, then we must infer a maker proportioned to what we actually find. The world is finite, so the evidence supports finite power. Great works are usually produced by many hands, so polytheism is at least as well supported as monotheism. Artefacts are often the botched early attempts of an apprentice, or the discarded work of a deity now dead. Philo does not assert any of these. He points out that Cleanthes's rule cannot exclude them, and that the theologically satisfying conclusion is reached only by stopping the inference early. That is the shape of the whole critique. Cleanthes's argument is empirical in form and selective in practice. Part IV: where the explanation is allowed to stop Before the alternative cosmogonies, Part IV supplies the argument that makes them necessary. Cleanthes's inference explains the arrangement of the material world by an ideal world in the mind of a designer. Philo asks what explains the arrangement of that . A mind is a system of ideas standing in relations to one another, and if arrangement in matter demands a cause outside itself, arrangement in ideas demands one just as much. Either the regress continues, or it is stopped somewhere — and if it may be stopped at a self-ordering mind, Philo cannot see why it may not be stopped one step earlier, at a self-ordering material world. Cleanthes replies that philosophers are not obliged to explain their explanations: an astronomer who identifies a cause need not also account for the cause. Philo's rejoinder is that this defence is exactly what he wants, since it works equally for his side. If we may rest with an unexplained ordering mind, we may rest with unexplained order in matter. The two hypotheses are then on the same footing, and the whole weight falls back on which of them experience actually supports. That is the question the next parts take up. Parts VI to VIII: other principles of order The alternative cosmogonies belong to the same programme. In Part VI Philo develops the ancient hypothesis of the world as an animal with the deity as its soul. In Part VII he presses further: The world plainly resembles more an animal or a vegetable, than it does a watch or a knitting-loom (Philo, Part VII) The point is not that plants are a better model of the cosmos. It is that generation and vegetation are principles that we do observe producing intricate order, and that they produce it without any mind arranging the parts. An acorn produces an oak without deliberating. If our experience is the evidence base, then our evidence base contains at least four principles of order — reason, instinct, generation and vegetation — and reason is the one we have observed operating in the smallest corner of it. Cleanthes has selected one, and the selection is not itself supported by the experience it claims to rest on. Part VIII adds the Epicurean hypothesis, on which a finite quantity of matter over unlimited time falls eventually into a self-sustaining arrangement that would look designed from within. Philo concedes readily that each of these hypotheses is poorly supported. That concession is the argument, not a weakness in it. His conclusion is comparative: the observed order of the world underdetermines the choice