Does complete foreknowledge make divine testing incoherent?

The argument concerns a specific meaning of "test." In an examiner-centered epistemic sense, a test is an activity intended to determine how a subject will perform or what the subject will choose. If God infallibly knows every future choice, then the event cannot update God's beliefs or supply information that God previously lacked. But "test" can also mean a trial, an opportunity to exercise a capacity, a process that makes conduct manifest, or a condition of fair judgment. Complete foreknowledge conflicts with the first purpose, not automatically with all of these others.

This is a serious philosophical challenge, although it is often presented in a weaker and less careful form online. The point is not simply that God "already knows the answer," as though divine knowledge were a guess made sufficiently early. The stronger claim concerns the purpose of the test. If a test is essentially an attempt to acquire knowledge of a person's future choice, then infallible foreknowledge removes epistemic risk: the result cannot surprise God and cannot resolve a divine uncertainty. That conclusion challenges a particular account of testing; it does not yet show that every non-investigative purpose is pointless.

The strongest form of the argument

The argument can be stated as follows:

  • In ordinary cases, testing is at least partly an information-gathering activity.
  • Complete divine foreknowledge entails that God already knows every future human decision infallibly.
  • What is already known infallibly cannot be discovered through a later event.
  • Therefore, a divine test cannot be an information-gathering exercise for God.
  • If the essential or sole purpose of divine testing is to give God information about a future choice that God already knows infallibly, then the test cannot fulfill that purpose. If the test has another purpose, this argument alone has not shown it incoherent.

This argument does not depend on assuming that God causes human decisions. Foreknowledge and causation are different concepts. A critic can grant, for the sake of argument, that a person chooses freely while maintaining that God knows infallibly what that choice will be. Some philosophers argue that infallible foreknowledge is itself difficult to reconcile with libertarian freedom or the ability to do otherwise, but that is a separate dispute. Even if that dispute is set aside, the epistemic question remains: what could the test add to a body of knowledge that is already complete?

Nor is the argument restricted to a particular picture of time. Augustine and Boethius offered influential accounts on which God's knowledge is not ordinary prediction from an earlier temporal moment. Aquinas similarly held that God knows temporal events through an eternal, non-temporal act of knowledge. These views avoid treating divine knowledge as a prediction that is later corrected by evidence. On such accounts, God's knowing is not an inquiry conducted from an earlier moment but an eternal relation to temporal events. That gives the theist a principled reason to say that God was never conducting an investigation. It does not, however, make an investigation possible: if the purpose is to learn the result, the result is already known whether God's knowledge is described temporally or timelessly. The timeless theist can therefore preserve the event only by assigning testing a purpose other than divine information-gathering.

The strongest theistic reply

The best reply is that a test need not be conducted for the examiner's information. Human tests often have several purposes, and information-gathering is only one of them. A driving examination may reveal competence to an official, but it also gives the learner an opportunity to exercise skill, makes a public record possible, and provides grounds for a decision that is visibly connected to performance. A court may already have strong evidence about what happened while still allowing testimony and cross-examination so that judgment proceeds through publicly accessible procedures. In these cases, the test is not pointless merely because one participant is not personally surprised by the result.

Applied to divine testing, the claim would be that God tests people in order to produce or reveal something within the created order rather than to improve divine knowledge. A test may give a person an opportunity to form character, exercise an intention, turn away from wrongdoing, or make a latent disposition into an actual deed. It may also disclose a person's conduct to other persons, establish a basis for social judgment, or make the justice of a later reward or punishment intelligible to the creature being judged. The test would then be performative, developmental, judicial, or relational rather than investigative for God.

This answer has substantial force. It is a mistake to assume that every test has only one purpose. A teacher may know that a diligent student will pass and still administer an examination because the student must actually perform, receive feedback, and complete a fair course of assessment. Likewise, if God knows what a person will freely do, the person's doing it can still matter. Knowledge of an action is not the action itself, and foreknowledge does not make an event unreal. The distinction between knowing that someone will act and bringing about the act is especially important here.

The reply is also compatible with several sophisticated theistic positions, although those positions should not be conflated. Classical theists can say that temporal events are genuinely significant even though God eternally knows them. Molinists typically affirm libertarian freedom and add God's knowledge of counterfactuals about what creatures would freely do in various circumstances; they use that knowledge to explain providential ordering without treating the choices as coercively caused. A compatibilist, by contrast, may hold that a person's action properly expresses that person's motives even if the action is determined and alternative possibilities are not required for responsibility. These theories disagree about freedom and providence, but none requires that God learn new information from the test.

Where the reply succeeds and where it does not

The reply succeeds against the claim that foreknowledge makes every possible meaning of divine testing incoherent. There are coherent purposes for a test that do not involve divine acquisition of information. A test can be an occasion for an action, a means of moral development, a public basis for judgment, or a way of relating to a person through real choices and consequences. If "test" is being used in one of these senses, complete foreknowledge is not by itself a contradiction. But logical coherence is only a modest result: the defender must still show that the proposed purpose is necessary, proportionate, and morally adequate, especially when the test involves serious suffering.

At the same time, the reply does not remove the original tension; it changes the meaning of the test. If an account of divine testing explicitly says that God tests people to find out what they will choose, then complete foreknowledge makes that account untenable. God cannot acquire knowledge of a proposition that God already knows infallibly. Calling the event a "test" does not solve this problem unless the term refers to another function.

There is also a further question about whether some alternative purposes are sufficient. Making a choice actual is not obviously the same as discovering it. If God already knows that a person will choose courage, the actual exercise of courage may benefit the person or others, but it does not benefit God by correcting ignorance. Similarly, publicly displaying a person's conduct may justify a judgment before other creatures, but an omniscient God would not need such a display as evidence. The defender must therefore explain whose knowledge, character, or relationship the test is intended to affect.

The courtroom analogy has limits for the same reason. Human courts require public evidence because judges are fallible and because legitimate procedures must be visible to the community. God, if omniscient, does not need evidence to reach a correct verdict. A divine test could still be required for the person's own moral formation or for the public order of creation, but the justification must be specified rather than borrowed uncritically from human institutions.

Does this undermine omniscience?

The argument may motivate revisions to classical omniscience, but it does not force them. Open theists, including Clark Pinnock and John Sanders, have argued that future free actions are not yet definite facts to be known and that God knows the future as partly open. On this view, God can genuinely respond to unfolding decisions. This preserves a more ordinary notion of testing, but it revises the traditional claim that God knows every future event infallibly. It also raises separate questions about providence, prophecy, and the nature of truth about future contingents.

Other philosophers retain complete foreknowledge and deny that divine testing is primarily investigative. Still others, following arguments associated with thinkers such as Luis de Molina, attempt to reconcile providence and libertarian freedom through God's knowledge of counterfactuals of creaturely freedom. These theories are complex and contested. The present argument does not decide among them, and it does not establish that one account of omniscience is correct simply because another has a difficulty.

What the argument establishes

The argument establishes a conditional point with real force: if divine testing means that God administers a test in order to learn which future choice a person will make, then that purpose conflicts with complete infallible foreknowledge. A test cannot supply information that an omniscient being already possesses.

It does not establish that theism is false, that omniscience is incoherent, or that every form of divine testing is meaningless. The strongest theistic response shows that tests can have non-investigative purposes, including moral formation, actualization of choice, relational engagement, and the creation of a public basis for judgment. These are candidate explanations, not automatic justifications. If a supposed test involves severe or apparently gratuitous suffering, saying that the suffering might develop character does not by itself answer the evidential problem of evil; that independent objection remains a serious challenge to theism and must be addressed on its own terms.

The fair conclusion is therefore limited but important. Complete foreknowledge rules out one ordinary meaning of divine testing: testing as an attempt by God to discover an unknown future decision. Theism must either abandon that meaning or explain why the test has another purpose. That is a genuine conceptual cost for some popular presentations of divine testing, but it is not, by itself, a refutation of theism or a proof that naturalism is true.