Last updated: November 3, 2026

The Münchhausen trilemma is an argument that no claim can ever be justified with complete certainty. If you try to prove something, you need reasons, and those reasons need reasons of their own. In the end you have only three options, and none of them is satisfying: an infinite regress of reasons, a circle in which the conclusion is used to support itself, or a point where you simply stop and accept something without proof. The German philosopher Hans Albert gave the problem its name in 1968. In older form it is known as Agrippa’s trilemma, after an ancient skeptic.

That sounds like a recipe for despair about knowledge. It is not, or at least Albert did not think so. The trilemma does not show that we know nothing. It shows that “proven beyond any possible doubt” is the wrong standard, and it forces us to choose a better one. This post explains the three horns, where the name comes from, what philosophers have said in reply and why the problem turns up every time someone in an argument keeps asking “but why?”.

Wait, It’s All a Münchhausen Trilemma?

The meme above borrows a famous template. One astronaut looks at the Earth, which is labelled “Philosophy”, and asks in shock: wait, it’s all a Münchhausen trilemma? Behind him, a second astronaut raises a pistol: always has been. The joke is that the realisation feels like a dangerous secret, when in fact it sits underneath every argument anyone has ever made. Every chain of reasons ends in one of three places, and philosophers have known that for about two thousand years.

The name comes from a fictional character. Baron Münchhausen is the hero of a collection of tall tales first published in English by Rudolf Erich Raspe in 1785 and expanded in German by Gottfried August Bürger in 1786, loosely based on a real German officer, Hieronymus Karl Friedrich von Münchhausen, who was known for telling outrageous stories about his adventures. In one of them the baron is stuck with his horse in a swamp and pulls both of them out by his own hair. That is exactly what a final justification would have to do: lift itself out of the swamp without anything to stand on.

Three Ways to Fail

Imagine any claim you care about, say that the Earth goes around the sun. Someone asks how you know. You give a reason: astronomical observations. How do you know the observations are reliable? Because of the theory of the instruments. How do you know that theory is true? And so on. Albert’s point is that this questioning can end in only three ways.

The infinite regress is the most obvious failure. If justification never ends, then no belief is ever actually justified, because every step is waiting for a step that never arrives. There is a widely told anecdote, attributed to many different scientists, about a lecture on astronomy after which an old lady insists that the world rests on the back of a giant turtle. And what does the turtle stand on? Another turtle, of course: it is turtles all the way down. That is the regress in its purest form.

The circle looks better from a distance because it does end, but it ends where it started. Circular reasoning is a classic fallacy for a reason: if your conclusion appears among your premises, you have not shown anything, you have only repeated yourself with extra steps. Big circles are harder to spot than small ones, which is why they survive in long arguments. C.M. Black lists it among the 5 frustrating fallacies we all use.

The third option, stopping somewhere, is what we actually do all the time. Mathematics starts from axioms, law from a constitution, science from observations and methods it does not question in the middle of an experiment. Albert called this breaking off the procedure at an arbitrary point, and he saw it as a form of dogmatism: the stopping point is declared certain, self evident or sacred, and further questions are forbidden. The trouble is not that we stop. The trouble is pretending that the stopping point has been proven.

Try it with a belief that feels completely safe: the sun will rise tomorrow. Why do you think so? Because it always has. But why does “always has” support “will”? Because nature is regular. And how do you know nature is regular? Because it always has been. You have just walked into the circle that David Hume described in the eighteenth century, the problem of induction. You can escape the circle by stopping, saying that the regularity of nature is simply a principle we have to assume, or you can keep asking forever. What you cannot do is give a reason that does not itself need a reason. It is the same structure that philosophers answer with questions instead of answers, a habit we explain in What Is Philosophy, and Why Do Philosophers Answer Questions With Questions?.

Older Than the Baron: Agrippa and the Skeptics

Albert did not discover the problem. The ancient Pyrrhonian skeptics already used it as a weapon. Sextus Empiricus, writing around the second century AD, reports in his Outlines of Pyrrhonism five “modes” of argument that he attributes to the later skeptics and that tradition credits to a philosopher called Agrippa, about whom almost nothing else is known. The five are disagreement, infinite regress, relativity, hypothesis and circularity. Three of them, regress, hypothesis and circle, are exactly Albert’s three horns, which is why the problem is also called Agrippa’s trilemma.

For the Pyrrhonists, the conclusion was practical. If no claim can be fully justified, the wise response is to suspend judgment, and that suspension was supposed to bring peace of mind. Much later, in the nineteenth century, the German philosopher Jakob Friedrich Fries described a similar trilemma between dogmatism, infinite regress and what he called psychologism, grounding knowledge in perception, and Karl Popper discussed “Fries’s trilemma” in his Logic of Scientific Discovery (1934). The structure keeps getting rediscovered because it is built into the very idea of giving reasons.

Hans Albert and Critical Rationalism

Hans Albert (1921 to 2023), a German philosopher and sociologist who lived to 101, set out the trilemma in his Treatise on Critical Reason (Traktat über kritische Vernunft, 1968). He was a close ally of Karl Popper, and his conclusion was Popper’s. If certain justification is impossible, we should stop looking for it. Instead of asking “can you prove it?”, we should ask “has it survived serious attempts to refute it?”.

This position is called critical rationalism. Every belief, including our most basic ones, stays open to criticism. We hold theories as provisionally as possible and replace them when they fail. Science, on this view, does not proceed by proving theories true but by eliminating the ones that turn out false. The Münchhausen trilemma, for Albert, is not a catastrophe for reason. It is the end of a fantasy, the fantasy of a final foundation, and with it the end of the excuse that some claims are beyond question. The same anti dogmatic spirit runs through Descartes, although Descartes hoped to find an unshakeable foundation, as we explain in Cartesian Doubt Explained.

Not everyone accepted this. In the 1970s Albert had a long public dispute with Karl-Otto Apel, who argued that some principles can be ultimately justified after all. Anyone who seriously argues against them, Apel claimed, already presupposes them, for example the rules of rational argument themselves, so denying them involves a performative self contradiction. Albert replied that this simply moves the dogmatic stop to a more elegant place. Whether a transcendental argument like Apel’s escapes the trilemma is still debated.

Axioms, Constitutions and Piles in the Swamp

The third horn, the stopping point, is everywhere once you look for it, and some of the most impressive human achievements are built on one. Euclid’s geometry starts from a handful of axioms and postulates that are not proven but assumed. For two thousand years the fifth one, the parallel postulate, looked so obvious that mathematicians tried to derive it from the others. In the nineteenth century, mathematicians including Lobachevsky, Bolyai and Riemann showed that you can replace it and still get consistent geometries. The stopping point turned out to be a choice, and a different choice later proved useful in physics.

Law works the same way. Every legal rule is valid because a higher rule authorises it, a statute because of the constitution, the constitution because of how it was adopted. The Austrian legal theorist Hans Kelsen argued that this chain has to end in a basic norm that is not itself enacted by anyone but presupposed, roughly that the historically first constitution ought to be obeyed. It is a Münchhausen stop made explicit, and Kelsen did not pretend otherwise.

Science gave the trilemma its most hopeful image, and it fittingly returns to the swamp. In The Logic of Scientific Discovery Karl Popper wrote that science does not rest on solid bedrock. It is like a building erected on piles that are driven down into a swamp, not to any natural base, but just deep enough to carry the structure for now. When the building needs to grow, the piles are driven deeper. That is not a failure of science. It is how it manages to keep building without ever pretending to have found the bottom.

Four Ways Philosophers Have Answered

Beyond critical rationalism, epistemologists have developed several responses, each of which embraces one of the horns and tries to show that it is not as bad as it looks. Foundationalists accept a stopping point but argue that it is not arbitrary: some beliefs, like the awareness of one’s own experiences or simple logical truths, are justified without needing other beliefs, so they can serve as foundations. Coherentists accept a kind of circle but argue that justification comes from the way a whole web of beliefs hangs together, not from a straight chain. The Austrian philosopher Otto Neurath captured this in 1932 with a famous image: we are like sailors who have to rebuild their ship on the open sea, without ever being able to bring it into dry dock and start from scratch.

A smaller group, the infinitists, led by Peter Klein, argue that an infinite chain of available reasons is not a problem at all, because justification is a matter of always being able to give another reason, not of finishing. And then there is Wittgenstein, who in On Certainty observed that at the foundation of well founded belief lies belief that is not founded. Some things, he thought, are not conclusions we reached but the hinges on which our questions turn. Doubting them makes no sense, not because they have been proven but because doubt itself needs them. You can find more of Wittgenstein’s view of language and limits in “The Limits of My Language Mean the Limits of My World”.

None of these answers has won, which is one reason the trilemma is still taught. But notice that none of them claims what the trilemma rules out, a justification that is both complete and self supporting. They all accept that the swamp is real and argue about the best way to stay afloat.

Why It Matters Outside the Seminar

The trilemma is a useful tool for spotting bad arguments. When someone presents a claim as beyond question, it is fair to ask where their chain of reasons stops and why there. Is the stopping point something that can be examined and tested, like a measurement anyone can repeat? Or is it a holy book, a charismatic leader or “everyone knows”? The trilemma does not tell you who is right. It tells you where to look.

It also explains why online arguments so often go nowhere. Two people with different stopping points can exchange reasons forever without either moving, because each keeps defending their own foundation with arguments that only work if you already accept it. Recognising that is the first step toward a better conversation: compare the foundations openly instead of pretending one of them has been proven. Our post on conclusion shopping and C.M. Black’s guide to critical thinking in the age of misinformation offer practical help.

There is a gentler lesson as well. If nobody can prove their deepest commitments, then intellectual humility is not a weakness but simply honesty. That does not mean every view is equally good. Views that survive criticism, make testable predictions and correct their mistakes are better than views that forbid questions. But it does mean that “I might be wrong” belongs in every serious belief system, including your own.

What a Careful Critic Would Add

Two cautions are worth adding, because the trilemma is often oversold. First, it is about justification in the strongest sense, a proof that would rule out all doubt. It does not show that our beliefs are unjustified in the ordinary sense, that we have no good reasons for them or that science is just another opinion. Second, some philosophers argue that it rests on a questionable picture of knowledge as a chain of reasons that must reach a final anchor. If knowledge works differently, for example if a belief can be justified because it was produced by a reliable process like normal vision, then the trilemma may be asking the wrong question. That idea leads into the debate about what knowledge even is, which C.M. Black approaches through Socrates in Do You Really Know What You Know, You Know? and which the Gettier problem complicates in The Gettier Problem and the Filter Bubble.

The baron’s trick is impossible, and that is fine. Nobody has ever pulled themselves out of a swamp by their own hair. People get out of swamps by grabbing whatever solid ground they can find, testing whether it holds and moving on when it does not. That is a decent description of how knowledge actually grows. If you want a shirt that admits as much, there is always How Philosophy Works.