Reappraising Al-Ghazālī's Skeptical Turn in Causality — Epoche C2
In the Seventeenth Discussion of the Tahāfut al-falāsifa , completed in 1095, al-Ghazālī denies that fire is the agent of the burning of cotton. His claim is not that fire and burning fail to occur together — he grants that they always do — but that observation establishes only that the burning occurs with the contact and not by it, and that the agent which creates blackness in the cotton, separates its parts and reduces it to ash is God. That passage is standardly read as a defence of Ashʿarite occasionalism: the doctrine that God is the sole efficient cause of every event, created things being no more than the occasions on which He acts. The reading is not wrong, but it is incomplete, and this essay argues that the argument's more durable content is epistemological — a claim about what observation can and cannot certify — and that al-Ghazālī's own way of closing the gap he opens has a recognisably externalist shape. Two things the essay in its earlier form asserted are corrected along the way: that al-Ghazālī anticipates Hume's problem of induction, and that he traced our causal beliefs to habit and custom in Hume's sense. Neither is accurate, and the reasons why are instructive. The argument, stated as it stands in the text The target is the position al-Ghazālī attributes to the philosophers, and to Avicenna above all: that the connection between a cause and its effect is necessary, that the effect follows from the cause by the cause's own nature, and that scientific knowledge consists in grasping such connections. On this view fire burns by virtue of what fire is; nothing but the removal of fire or of cotton could prevent it. Al-Ghazālī's counter-argument has two steps, and they are of different kinds. The first is logical. Between any two distinct existents, he says, there is no entailment: the affirmation of one does not entail the affirmation of the other, and the denial of one does not entail the denial of the other. Contact with fire and combustion in cotton are distinct existents, each with its own definition, neither included in the other. So the proposition that fire is in contact with cotton and the cotton does not burn is not a contradiction — unlike, say, the proposition that a body is a triangle and has four sides. Since it is not a contradiction, it is possible, and God's power extends to everything possible. The second step is evidential, and it is the one the epistemological reading rests on. Grant that the two events are always found together. What does observation deliver? It delivers the co-occurrence. It does not deliver the further item the philosophers require, namely that the first event produces the second and that no other agent does. The Arabic distinction is between the burning occurring maʿahu , with it, and bihi , by it, and al-Ghazālī's point is that no amount of watching discriminates between those two descriptions of the same sequence. Efficacy is not among the things the senses report. It is worth being exact about what this does and does not show. It does not show that fire lacks the power to burn. It shows that the philosophers' claim to have demonstrated the necessity of the connection is unsupported by the evidence they cite, and that a rival hypothesis — that the sequence is maintained by an agent whose action is regular but not necessitated — is consistent with every observation ever made. The argument is a demonstration of underdetermination, not of occasionalism. Two theories, not one That last point has consequences for how the Discussion should be read as a whole, and it is here that the standard occasionalist interpretation runs into difficulty. Michael Marmura argued in 1981 that the Seventeenth Discussion in fact contains two distinct causal theories, presented in succession and not reconciled. The first is the occasionalist one: God creates each event directly, and there are no efficacious secondary causes. The second is different. On it, God has created in things their natures and dispositions — fire does have a capacity to burn — but the exercise of that capacity depends at every moment on God's continued willing, so that He may withhold the effect without any contradiction arising. Miracles are secured on either theory. Only the first eliminates natural causal powers. Al-Ghazālī does not tell us which he holds, and the resulting question has divided the field. Richard Frank argued in Creation and the Cosmic System (1992) that al-Ghazālī's real cosmology is substantially Avicennan: a world of genuine secondary causes operating within an order established and sustained by God, and thus at odds with the Ashʿarite school in whose name he wrote. Marmura held that al-Ghazālī remained an occasionalist and that the second theory is offered dialectically, to show that even a concession to the philosophers leaves the possibility of miracles intact. Frank Griffel, in Al-Ghazālī's Philosophical Theology (2009), argues that the non-committal presentation is deliberate: al-Ghazālī wished to leave open a position on which the regularities studied by the natural sciences are real and investigable while God retains the power to suspend them, and he had reasons of his own for not choosing between the two theories in a polemical work. The epistemological argument is untouched by this dispute, which is one measure of its independence. Whether the fire has a nature that burns or is merely the occasion of God's burning, observation cannot tell us which, and the philosophers' claim to demonstrative knowledge of necessary connection fails either way. That is why the sceptical content of the Discussion survives interpretive disagreements that its metaphysical content does not. Why "the problem of induction" is the wrong label The essay in its earlier form said that al-Ghazālī raises the Humean problem of induction six and a half centuries before Hume — the interval from 1095 to the Enquiry concerning Human Understanding of 1748 is 653 years — and that his acco