Hume's Philo and the Explanatory Challenge of Fine-Tuning — Epoche B2
Chance, Design, or Necessity: What Philo Adds to the Fine-Tuning Debate The problem Contemporary physics describes the universe using a handful of numbers that the theories do not themselves fix: the strength of gravity relative to the other forces, the masses of the elementary particles, the value of the cosmological constant. It is widely reported that if several of these had been slightly different, no stars, no chemistry and no observers would have formed. From this, some philosophers and physicists infer design. The universe is exquisitely suited to life; such suitability is wildly improbable by chance; therefore it was arranged. David Hume is regularly invoked against this argument [1] , and regularly invoked for the wrong reason. The standard summary says that Hume showed we cannot argue inductively from a single case, so we cannot infer a cause of the universe. That objection is real, it is in the Dialogues Concerning Natural Religion (1779), and defenders of fine-tuning have an answer to it. This column argues that Hume's more useful contribution lies elsewhere in the same book, in a remark of Philo's at the close of Part IX, and that it identifies a premise the fine-tuning argument needs and does not have. A correction about the text The Dialogues are frequently cited loosely, and the parts do different jobs. The argument from design is presented by Cleanthes in Part II and attacked by Philo through Part VIII. Part IX is a different argument altogether: Demea abandons empirical reasoning and offers an a priori proof of a necessarily existent being, and it is Cleanthes , not Philo, who does most of the refuting. The famous inductive objection belongs to Part II. Philo grants Cleanthes the analogy and then restricts it: If we see a house, Cleanthes, we conclude, with the greatest certainty, that it had an architect or builder; because this is precisely that species of effect, which we have experienced to proceed from that species of cause. The point of the concession is the restriction that follows it. We infer builders from houses because we have seen houses built. We have not seen universes made, and the universe is not relevantly like a house; it is at least as much like an animal or a vegetable, and those come about by generation and growth. The inference to an architect is not licensed by experience because the required experience does not exist. Why that objection does not settle the modern debate Defenders of fine-tuning have a reply, and it is a good one. Their argument is not an inductive generalisation from observed cases of universe-making. It is a comparison of two hypotheses in the light of one piece of evidence: the life-permitting values are, they claim, much more to be expected if the universe was arranged by something with a purpose than if it was not. Richard Swinburne set out an argument of this shape in The Existence of God (1979 [2] ). Reasoning of this kind is used all the time in science and in courts, and it does not require that we have previously observed anything of the type in question. Philo's Part II objection, taken alone, does not touch it. Philo's other remark Part IX is where the more penetrating point appears, and it appears almost as an aside. Demea has argued that there must be a necessarily existent being. Cleanthes replies that no existential claim can be demonstrated a priori, since whatever we conceive as existing we can also conceive as not existing, so nothing whose non-existence is conceivable can be necessary. He adds a second thought, that for all we know the material universe itself might be the necessarily existent being — the very thing Demea's argument was supposed to reach beyond. Philo then adds an example. Take the multiples of nine. The digits of every one of them sum to nine or to a multiple of nine. Someone who noticed this pattern without understanding it might well ascribe it to chance or to design; it looks too regular to be an accident and too clever to be nothing. But a competent mathematician sees at once that it is neither. It follows necessarily from the structure of decimal notation. Philo's suggestion is that the order of the universe might stand in the same relation to some necessity we have not discovered: apparent contrivance, actual entailment. That is a small remark with a large consequence, and it is the one this column wants to press. The premise the argument needs The fine-tuning argument compares two hypotheses on the evidence that the constants are life-permitting. For the comparison to be informative, the values must be things that could have been otherwise. If the values are fixed by necessity — if a deeper theory entails them, or they are metaphysically necessary — then their being life-permitting is no more surprising than the digits of the multiples of nine summing to nine, and there is nothing for a designer to explain. So the argument requires a modal premise: these constants are contingent, and could have taken a wide range of other values. What supports it? The usual answer is that current theories treat the constants as free parameters, values that must be measured rather than derived. But this is a fact about our theories, not about the world. The history of physics is in part a history of parameters ceasing to be free. Quantities that had to be measured independently have repeatedly turned out to be determined by, or identical with, others once a unifying theory was found. A theory's leaving a number open is evidence about the theory's completeness. It is not evidence that the number could have been different. A second difficulty follows even if contingency is granted. To say that the life-permitting range is a tiny fraction of the possibilities, one needs a probability distribution over the possible values. If the range of possible values is unbounded, there is no uniform distribution over it that behaves as a probability should: assign equal weight to equal intervals over an infinite range a