The Regress Problem: Why Justification Has to Stop Somewhere

Ask someone why they believe something, and they will give a reason. Ask why they believe that reason, and they will give another. Keep asking, and the conversation either goes on forever, circles back on itself, or hits a wall. This is the epistemic regress problem, and how you resolve it shapes your entire theory of knowledge.

What the regress actually shows

The problem is not merely a debating trick. It exposes a genuine structural tension in the concept of justification. If a belief B₁ is justified only by appeal to another belief B₂, then B₁ inherits whatever justificatory status B₂ has. But B₂ itself needs justification from B₃, and so on. Three outcomes are logically possible: the chain goes on infinitely (infinitism), the chain loops back on itself (coherentism), or it terminates at beliefs that are justified without needing support from further beliefs (foundationalism). Each option is coherent. Each carries costs that its proponents have spent centuries trying to minimize.

Foundationalism and its price

The most intuitive response is to insist that some beliefs are basic—justified not by other beliefs but by something outside the inferential chain altogether. Classical foundationalists like Descartes pointed to clear and distinct ideas; later empiricists pointed to direct sensory experience. Contemporary modest foundationalism weakens the claim considerably: basic beliefs need not be infallible or even certain, just sufficiently credible on their own to anchor the rest.

The difficulty is explaining what makes a basic belief basic without appealing to another belief. If the answer is "experience," one has to say what counts as the right kind of experience, which seems to require further commitments. If the answer is "self-evidence," critics note that self-evidence has historically endorsed some spectacularly wrong conclusions—moral intuitions about slavery, perceptual intuitions about a geocentric sky. The foundationalist owes an account of which stops on the regress are legitimate and why.

Coherentism and the isolation objection

Coherentists dissolve the regress by denying that justification is a one-directional chain at all. What justifies a belief is its coherence with the web of other beliefs a person holds—its consistency with them, its explanatory fit, its mutual support. No belief is foundational; all beliefs justify one another simultaneously, the way pieces of a puzzle mutually determine their neighbors.

This is elegant, but it faces the isolation objection: a wildly false system of beliefs can be internally coherent. A detailed, consistent conspiracy theory coheres with itself perfectly. Coherentism has difficulty explaining why the web of belief should track the world rather than just track itself. Proponents reply that perception and memory play a privileged role in shaping the web, but this concession starts to look like smuggled foundationalism.

Infinitism: taking the regress seriously

Infinitism, associated most clearly with the contemporary philosopher Peter Klein, proposes that the regress is not vicious—it is simply what justification requires. To be justified in believing something is to have available an endless supply of non-repeating reasons, each supporting the last. No reason need actually be traversed; what matters is that the chain is there to be explored.

Critics find this unsatisfying. Human cognition is finite. If justification requires an actual infinity of available reasons, it seems to demand something psychologically impossible. Klein responds that justification is about potential, not performance—but that distinction is contested. The debate turns on whether justification is a property of a belief at a moment or of an agent's long-run epistemic situation.

What the regress problem means for skepticism and ordinary inquiry

The regress problem is not merely academic. It connects directly to global skepticism: if no solution is fully satisfying, perhaps no belief is genuinely justified. Descartes used exactly this lever to strip away almost all his commitments before rebuilding on the cogito. Contemporary skeptics note that the three responses each preserve justification only by redefining it.

The more productive response, favored by philosophers like Wittgenstein in On Certainty, is to accept that certain commitments function as hinges—not believed on evidence but presupposed in the practice of inquiry itself. "The earth has existed for many years" is not a hypothesis tested against experience; it is part of what makes testing possible. This does not refute skepticism logically, but it reframes the question: demanding justification for hinge propositions may not be a coherent request, because the apparatus of justification presupposes them.

For skeptics of religious and supernatural claims, the regress problem cuts in an interesting direction. Appeals to revelation, scripture, or religious experience as epistemic foundations are foundationalist moves. The skeptic's challenge is not simply to deny those foundations but to ask what makes them foundational rather than arbitrary stopping points—the same question foundationalism must answer for all purported basic beliefs. The regress problem does not privilege atheism, but it does demand that every claimed stopping point justify its own privilege.