Al-Ghazali's Agent Causation in *Tahafut al-Falasifa* — Epoche C2
Where the argument sits in the book, and what it is for The claim that fire does not burn cotton — that the blackening, the coming apart of the fibres and their reduction to ash are created directly by God at the moment the two touch — is made in the Seventeenth Discussion of Abu Hamid al-Ghazali's Tahafut al-falasifa ( The Incoherence of the Philosophers ), completed at Baghdad around 1095, and not in the Eighteenth, as an earlier version of this essay stated. The correction earns its place. The Tahafut runs to twenty discussions: sixteen on metaphysics ( ilahiyyat ), then four on natural science ( tabiiyyat ). The Seventeenth opens the second group, and the Eighteenth to Twentieth concern the soul and bodily resurrection. Placing the causal argument at the head of the physical discussions tells you what it is for: al-Ghazali needs the possibility of prophetic miracle — a staff that becomes a serpent, a fire that does not consume Abraham, the dead restored to life — before he can argue for the resurrection of bodies, and the philosophers' doctrine that natural causes necessitate their effects stands in the way. The received account of what he does about that doctrine is that he is an occasionalist in the manner of the Ash'arite school of theology ( kalam ) in which he was trained: created things have no real causal powers, and the apparent cause is merely the customary occasion on which God creates the effect. That is not false, but taken as the whole of his position it makes the causal argument purely destructive. The contention of this essay is that the demolition serves a positive account of what it is to be a cause at all, on which the only genuine causes are agents that act by will, knowledge and power — an account stated most explicitly not in the Seventeenth Discussion, where readers look for it, but in the Third. The Seventeenth Discussion, step by step The discussion opens with a denial and a restriction of its scope, and the two are worth separating, because summaries routinely report the first without the second. The denial is that the connection ( iqtiran ) between what is habitually taken to be a cause and what is habitually taken to be its effect is necessary. The restriction is that divine power extends to the possible and not to the impossible in itself: he is not claiming that a thing might be black and white at once, or that a man might be in two places at once. The target is narrower, and it is the philosophers' own thesis — that a natural agent, given its nature and the removal of impediments, must produce its characteristic effect. His examples are four, and they are chosen to span the range of what the philosophers took the sciences to have established: the burning of cotton by fire; the quenching of thirst by water; the healing of a sick person by medicine; the severing of a head from a body by a sword. The argument for the denial is epistemic and consists of a single move. What is observed when fire meets cotton is that the burning occurs with the contact, not through it: observation registers succession and concomitance, and necessitation is not among the things it delivers. Since, al-Ghazali insists, there is no other route to the claim — no demonstration establishes it — the philosophers' necessity is an inference from constant conjunction that the conjunction does not license. The real agent of the burning, he says, is God, who creates the blackness in the cotton and the separation of its parts; nothing in what we see tells against this, since what we see is compatible with both accounts. The resemblance to Hume's treatment of the idea of necessary connection, some six and a half centuries later, is exact at this step, and the conclusions drawn are opposite: Hume relocates necessity into the mind's habit of transition, leaving nothing in the world to answer to it; al-Ghazali relocates it into the will of an agent, which is very much in the world. The obvious objection follows at once in the text, and al-Ghazali's answer is the part most often dropped from summaries. If nothing necessitates anything, why is the world not a chaos, in which a man who leaves his house might return to find his books turned into beasts? The reply is not that such things are impossible, but that God has created in us the knowledge that He will not do them, so that our expectations remain firm and well founded even though what they anticipate is not necessary. The uniformity rests on a settled divine practice rather than on a necessity resident in things. This is why reading the discussion as a denial of empirical regularity gets it backwards: what is denied is a modal claim about the regularities, not the regularities themselves. Then comes a turn that the standard picture of al-Ghazali as a plain occasionalist cannot accommodate, and the text announces it as a second approach — a different way of securing the same conclusion, framed as a concession to an opponent who will not give up natural powers. Grant, he says, that fire does have a nature by which it burns. Even so it does not follow that a prophet cannot stand unharmed in it, for God may create in the body a quality that prevents the burning — and this is not even an unfamiliar kind of event, since a man who coats himself with talc can sit in a furnace and be unhurt. Here the power in the fire is left standing and the miracle is secured by a change worked in the patient rather than by the suspension of the agent's power. Two theories, and a dispute about which is his That the second passage is a second causal theory, not an illustration of the first, is the case set out by Michael Marmura (1981), and it has organised the scholarship since. The two differ in their commitments. On the first, no created thing has any power, and the regularity of nature is entirely God's habit of creating in a constant pattern. On the second, created things do have natures with characteristic effects, and what God's freedom guarantees is the abil