Al-Ghazali's Incoherence of the Philosophers: Testing Philosophical Proofs

Bring a piece of cotton into contact with a flame. If it burns, what have you learned? You have observed contact followed by combustion. Have you also established that the cotton could not have failed to burn? That extra claim concerns necessity: what must happen, rather than what did happen. Al-Ghazali makes this difference the subject of the seventeenth discussion of The Incoherence of the Philosophers.

The example belongs to a religious dispute about miracles. It also exposes a difficulty that matters wherever an argument claims more than its evidence establishes. Observing a sequence, explaining it through a causal theory, and proving that no alternative is possible are different achievements. Al-Ghazali asks whether the philosophers he criticizes have supplied the last of these.

His book is an adversarial investigation of philosophical arguments about God, the world, and the soul. Its central question is which conclusions those arguments actually demonstrate. Reading it carefully requires distinguishing a failed proof from a false conclusion, and a possible alternative from an established explanation.

Al-Ghazali identifies his principal targets in the first preface: al-Farabi and Ibn Sina, known in Latin as Avicenna. They represent the philosophical tradition he proposes to examine. The title does not mean that every person who reasons about difficult questions holds one shared doctrine. Nor should a reader assume that every position attributed to the philosophers is an uncontroversial description of Aristotle himself.

The distinction matters because much of the book takes the form of objections and replies. Sometimes al-Ghazali reconstructs an opponent's argument. Sometimes he offers a reply without adopting every premise of that reply as his own settled teaching. In the third preface, he explicitly presents his purpose as attacking the positions under discussion rather than defending a complete system of his own.

That method can make the text frustrating. A reader may want to ask, after each refutation, what al-Ghazali positively believes. The book's immediate task is often narrower: to show that his opponent has not established the claimed conclusion. A critic can expose an unsupported step without supplying a complete replacement account of reality.

The introduction also distinguishes different kinds of disagreement. Some concern terminology. Others concern matters that do not threaten religious belief. Still others concern doctrines al-Ghazali regards as incompatible with religion. He therefore does not approach all philosophical subjects as equally suspect. The distinctions determine which disputes the book will pursue.

In this setting, a demonstration is an argument offered as yielding certain knowledge, rather than merely persuading someone who already accepts a disputed opinion. The important question is whether the premises and the inference justify that confidence. Calling an argument philosophical does not answer it.

Al-Ghazali objects to borrowing the authority of one discipline for another. In the first preface, he contrasts the agreement associated with arithmetic with philosophical disagreement about metaphysics. Success in a clear mathematical argument does not automatically make a claim about the origin of the world equally certain. Each argument needs its own support.

His discussion of eclipses in the second preface makes the point concrete. He refuses to make religious belief depend on rejecting the astronomical explanation of an eclipse. Opposing a well-supported explanation on religious grounds, he argues, can damage religion by making its defender appear ignorant. The text gives us a reason to resist the familiar picture of this book as a blanket refusal of rational inquiry.

That passage does not turn al-Ghazali into a modern scientist. It identifies a distinction within his own argument: astronomical findings and disputed metaphysical conclusions do not all stand or fall together. His fourth preface similarly accepts the need for logic while denying that philosophers possess a monopoly on it.

Consider a hypothetical dispute about a building. One person measures its rooms accurately and then claims that the measurements prove what its owner intended. The measurements may be excellent, while the further inference remains unsupported. This is our illustration, not a reported episode from al-Ghazali. It shows why respecting a method's achievements can coexist with challenging an extension of its authority.

The first discussion examines arguments for the world's eternity. The dispute concerns whether the world began to exist, rather than merely whether it depends on God. A world without a temporal beginning could still be described as dependent on a divine cause. The disagreement therefore cannot be resolved simply by observing that both sides use the word “creation.”

One argument al-Ghazali considers asks how something temporal could proceed from an eternal cause. If the cause is complete and unchanging, why would its effect appear at one point rather than another? If something new determines the effect, the explanation seems to require a further change. If nothing new determines it, the proposed beginning seems unexplained.

Al-Ghazali replies by defending the possibility of an eternal will that determines a world with a beginning. The will need not acquire a new decision through deliberation. The philosophers have to establish that this proposed relationship is impossible, rather than rule it out by assuming that a willing agent must operate like a necessary natural cause.

The argument then turns to what it means for will to distinguish among alternatives. The opponent demands an explanation of why one possibility is selected rather than another. Al-Ghazali treats the specification of an alternative as characteristic of will. The disagreement concerns what counts as an explanation of choice, and whether divine action must follow the explanatory pattern the opponent demands.

It would be misleading to reduce this exchange to a human deciding to wait until tomorrow. Such an example already presupposes a surrounding time and a person whose thoughts change. The divine case is precisely about whether an unchanging will can determine temporal existence. Familiar human choices might suggest a question, but they cannot settle it.

Al-Ghazali's immediate purpose is to defeat the proposed demonstration of eternity. That result would not, by itself, prove temporal creation. The distinction between undermining one argument and establishing the contrary conclusion must stay visible throughout the discussion. His religious commitment to creation supplies a further context; it should not be mistaken for the conclusion of every individual objection.

The thirteenth discussion concerns divine knowledge of changing particulars. A particular is an individual thing or event, such as this eclipse occurring at this time. A universal description concerns a kind or general pattern. The philosophical difficulty arises when knowledge of a changing event seems to require a change in the knower.

Al-Ghazali reconstructs the issue through an eclipse. Before it occurs, someone can know that it will occur. During it, that person can know that it is occurring. Afterwards, the person knows that it occurred. Human knowledge changes as the event moves from future to present to past. How can God know such an event without undergoing that sequence of changes?

The position he attributes to Ibn Sina needs care. He reports both the claim that God knows particulars in a universal manner and the claim that nothing is hidden from divine knowledge. Merely saying that Ibn Sina's God knows generalities and is ignorant of individuals suppresses the qualification that makes the disagreement philosophically difficult.

Al-Ghazali challenges the adequacy and consequences of the proposed account. His concern is whether it truly preserves knowledge of particular persons and their changing actions. A religious account in which God knows obedience, disobedience, and prophecy must explain that knowledge, rather than replace it with a general knowledge of human types.

The issue is not whether a powerful observer could gather more information. Both sides are discussing knowledge unlike ordinary observation. Their dispute concerns the relation between eternal knowing and temporal objects. The reader has to examine whether an argument has transferred assumptions about our knowledge to God's, and whether the proposed alternative avoids the same difficulty.

This is a useful point at which to slow down. “God knows everything” and “God does not change” are statements a thinker might accept together. Their compatibility still requires explanation when what is known changes. Al-Ghazali's critique targets the offered explanation; repeating either statement does not resolve the tension.

The seventeenth discussion asks whether departures from the ordinary course of nature are impossible. Al-Ghazali lists familiar connections, including drinking and the satisfaction of thirst, eating and the satisfaction of hunger, and contact with fire and burning. He denies that the connections have the necessity his opponents claim for them.

His cotton example separates two propositions. Burning occurs with contact. Burning occurs through the fire as an agent whose nature makes the result unavoidable under the stipulated conditions. Observation of the first does not, he argues, establish the second. The existence of an observed association leaves an explanatory question open.

One answer he develops attributes the production of the burning to God, directly or through mediation. On this account, the fire is not an independent agent that constrains divine action. God can create the contact without creating the burning. The ordinary sequence reflects divine action rather than an indissoluble connection between created things.

This answer is associated with occasionalism: the view that events ordinarily called natural causes are occasions for divine production of their effects. The term helps identify an argument in the discussion. It becomes misleading if it is used to conceal the other possibilities al-Ghazali considers.

Later in the same discussion, he offers a second answer. He grants that fire acts uniformly on similarly situated cotton, while allowing that God could alter a relevant property of the fire or of the body exposed to it. The proposed miracle would then involve a difference in the conditions under which the natural interaction occurs.

These answers have different explanatory structures. One directly challenges the necessity of the connection. The other grants a regular natural operation while disputing whether all relevant conditions must remain unchanged. The text's argumentative strategy does not require us to collapse them into one exclusive theory of how every event happens.

An opponent raises an obvious difficulty. If radically different outcomes are possible, how can anyone confidently go about ordinary life? In al-Ghazali's example, someone might leave a book at home and return wondering whether it has become an animal. Once the usual connections are denied necessity, what protects ordinary expectations?

Al-Ghazali answers by distinguishing what is possible from what is expected or known to occur. In the theological account developed here, God can create knowledge that the extraordinary event will not occur. Repetition of the ordinary sequence also establishes expectation. The admission of an alternative possibility does not require a person to treat that alternative as imminent.

This reply belongs to his account of divine power and knowledge. It is more than the modest observation that unusual things are unlikely. The reader should ask whether it explains our confidence or assumes the reliability that the objection puts in question. Recognizing the question helps us assess his answer without pretending that he never saw the problem.

A hypothetical illustration can clarify the logical distinction without proving his theology. You might know that a locked door could be opened by someone with a key while expecting it to remain shut tonight. Recognizing a possibility does not require predicting its occurrence. The example concerns the relationship between possibility and expectation; it does not establish that natural connections resemble locks or that miracles occur.

The same restriction applies to the cotton argument. Showing that an alternative has not been ruled out is not evidence that someone has witnessed it. Al-Ghazali's defense of a miracle's possibility and the historical evidence for a reported miracle are separate questions. Keeping them apart prevents a philosophical objection from becoming an unsupported empirical claim.

The twentieth discussion addresses bodily resurrection. Al-Ghazali opposes an interpretation that leaves the afterlife entirely at the level of the soul and excludes the restoration of bodies. The issue concerns what human beings can become after death, and whether philosophical arguments have demonstrated that bodily restoration cannot happen.

His discussion returns to the scope of divine power and our knowledge of causes. Familiar processes of bodily formation do not, he argues, exhaust what is possible. A person who knows one way a body develops has not thereby proved that no other mode of formation can exist. The argument is directed against an asserted impossibility.

The conclusion of the book singles out three doctrines for the judgment of unbelief: the eternity of the world, the disputed denial of divine knowledge of particulars, and the denial of bodily resurrection. This is al-Ghazali's religious judgment. It should be reported as such, rather than presented as a neutral classification accepted by every participant in the dispute.

That conclusion also prevents a comforting modernization of the book. It is not simply an invitation to suspend judgment about everything. Al-Ghazali is defending religious commitments and drawing boundaries around acceptable teaching. His arguments about proof must be understood alongside that purpose, even by readers who reject his judgment.

The strongest question for al-Ghazali is whether his alternatives are genuinely coherent. A critic cannot defeat a demonstration merely by saying that something else might happen. The alternative must survive scrutiny. Does an eternal will specifying a temporal beginning explain enough? Does divine production of knowledge preserve warranted confidence? These are questions raised by the arguments we have examined.

A second question concerns the movement from philosophical criticism to religious acceptance. Suppose an argument against bodily resurrection fails. That failure does not make every religious account of resurrection true. Further commitments concerning revelation and its interpretation are needed. The Incoherence cannot be read as though the collapse of one proof automatically establishes an entire theology.

The book nevertheless teaches a demanding way to read an argument. Identify the exact conclusion. Separate it from the confidence with which it is asserted. Ask which premise excludes the proposed alternative, and whether that premise has been defended. Then ask what follows if the argument fails. Sometimes the proper result is a narrower conclusion rather than a reversal.

The cotton in contact with the flame now presents several questions at once. What happened? What produced it? Under which conditions would it recur? Could it have been otherwise? Al-Ghazali insists that an answer to the first question cannot simply be promoted into a final answer to all the others. Whether his own alternatives succeed remains part of the philosophical work the reader must do.

• Al-Ghazali, The Incoherence of the Philosophers, introduction and four prefaces: the targets, scope, and adversarial method. The Kamali text hosted by ghazali.org is the accessible version consulted here; its hosting page describes an Urdu intermediary.
• Discussion I: eternity, temporal origination, and will.
• Discussion XIII: divine knowledge and changing particulars.
• Discussion XVII: cotton, fire, ordinary expectations, and alternative accounts of miracles.
• Discussion XX and the conclusion: resurrection and the final religious judgments. Compare Michael E. Marmura's parallel English-Arabic edition when studying technical wording.