An Internalist Re-foundation of Ash'arite Epistemology — Epoche C2
In al-Munqidh min al-Ḍalāl ("Deliverance from Error"), written around 1108 in the last years of his life, al-Ghazālī (d. 505/1111) fixes a standard for certain knowledge before he applies the word to anything. Certain knowledge — al-ʿilm al-yaqīnī — is knowledge in which the object is laid open so completely that no doubt clings to it and no possibility of error accompanies it: so completely that if a man were to turn a stone into gold and a rod into a serpent in front of you, the performance would not disturb your conviction that ten is more than three. The clause about the conjurer does technical work rather than rhetorical work. It says that certainty is a property the knower can check from where he stands — not merely that nothing will in fact dislodge the belief, but that the knower can see that nothing could. Measured against that standard, al-Ghazālī's attack on taqlīd — assent to a doctrine on the authority of another, without grounds the assenter has examined — is not the opening of a sceptical campaign but a specification of what a theology has to supply. This essay argues that the attack is best read as an internalist re-founding of the epistemology of the Ashʿarite school in which he was trained: internalist in the accessibilist sense that what justifies an assent must be something the assenter can determine to obtain by reflection; a re-founding rather than a demolition, because the doctrines whose grounds he demands are the school's own. Two complications have to be met rather than smoothed over. The first is that al-Ghazālī elsewhere prescribes taqlīd , and forbids the laity the rational theology his own teachers had made obligatory. The second is a matter of scholarly record: the historian whose name is most often attached to the "from within" reading, Richard Frank, argued the reverse, and an earlier version of this essay cited him for a thesis he wrote a book to deny. That citation is corrected below and the argument rebuilt on the scholars who do hold the position. The standard, and the two collapses that produce it Before the criterion can be used it is worth seeing how the Munqidh reaches it, because the route determines how much of it is available to anyone else. Al-Ghazālī places the collapse in his youth — he dates the loosening of the bond of taqlīd to before his twentieth year — and not in the famous crisis of 488/1095, when he abandoned his chair at the Niẓāmiyya college of Baghdad. Keeping the two episodes apart matters, because the later crisis was about sincerity, reputation and the fear of judgement, not about the reliability of cognition. The epistemic collapse comes first, and it has two stages. First the senses go. The eye reports a shadow at rest which, watched longer, is found to have been moving throughout; it reports a star the size of a coin, where astronomical argument shows the star to exceed the earth in size. In each case a verdict of sense is overturned, and the judge that overturns it is reason. Then the manoeuvre is turned on the judge. If sense was confident and was convicted by a faculty above it, what secures reason's own axioms against a faculty above reason, whose arrival would show the necessary truths to be as ill-founded as the shadow's rest? Al-Ghazālī refuses to let his confidence in the axioms count as an answer, on the ground that exactly that confidence was available to sense before its conviction. He offers the sleeper as the model: a man sees things in a dream and assents to them without reserve, and on waking finds the whole set groundless — so how does he know that waking does not stand to some further state as dreaming stands to waking? By his own account he was for nearly two months a sceptic in fact, though not in speech or in doctrine, and no argument delivered him. What delivered him, he writes, was not the composing of a proof or the ordering of an argument but a light which God cast into his breast, and that light he calls the key to most cognitions. Here the comparison with Descartes, which this essay made in its earlier form, needs care rather than deletion. The parallel is real at the level of procedure — a deliberate suspension used to find whatever resists it — and it fails at the level of the exit. Descartes's meditator recovers by producing something he can inspect. Al-Ghazālī recovers by receiving something he did not produce and cannot instruct anyone else to produce. That is a genuine embarrassment for the word "internalist" and should be stated, not hidden. It is contained, though, once one notices what accessibilism is a thesis about: it constrains what makes an assent justified, not what causes the assent. The light restores the primary necessities — the ḍarūriyyāt , propositions a mind cannot consider without assenting — and those are self-presenting, in that whether one has them is settled by attending to them. The criterion of certainty therefore survives intact and internalist. What does not survive is any suggestion that al-Ghazālī offers a method his reader could carry out. He offers a test and no procedure for passing it. Why the target is taqlīd and not tradition With the standard fixed, the question is what it convicts, and the answer is narrower than "tradition". Al-Ghazālī's complaint about the muqallid , the man who assents on authority, is never that what he assents to is false; al-Ghazālī holds the creed to be true. The complaint is that the muqallid possesses a fact about his own believing which, once noticed, he cannot answer. At the decisive point of the Munqidh the observation is carried by a prophetic report: every infant is born in the native constitution, and it is his parents who make him a Jew, a Christian or a Zoroastrian. The report is not being used devotionally. It supplies a complete causal explanation of the young al-Ghazālī's own creed that makes no reference to the creed's truth, and that would have produced a different creed, held with the same conviction, in a child born el