"A habit mistaken for a law is still just a habit."
Drop a glass and something in you already knows what happens next: it falls, it breaks. You would bet anything it will happen the same way the next thousand times. But ask for proof that it must — that the falling isn’t just something that has always happened so far, but something guaranteed, necessary, built into reality — and the proof isn’t there. That gap between expecting and proving is what the Scottish philosopher David Hume found sitting underneath the whole idea of cause and effect. Decades later, reading Hume was what woke Immanuel Kant, in his own words, from a “dogmatic slumber,” and sent him looking for a way to rebuild human knowledge on steadier ground.
Hume first laid out the problem in A Treatise of Human Nature (1739–40), a book so ignored on publication that he later joked it “fell dead-born from the press,” and restated it more sharply in An Enquiry Concerning Human Understanding (1748). His method was to take the idea everyone relies on without examining it — causation — and ask where it actually comes from. Every idea, on the empiricist principle Hume worked from, has to trace back to some direct experience, so what experience gives us the idea of “necessary connection,” the sense that a cause doesn’t just come before its effect but forces it to happen? Watch one billiard ball strike another, as closely as you like, and name what you actually perceive: the first ball approaching, contact, the second ball moving. What you never perceive, no matter how carefully you look, is the necessity itself — some invisible “must” passing from one ball to the other. All experience ever hands you is one event followed by another, what Hume called a “constant conjunction.” Watching a thousand collisions instead of one doesn’t change that; it just supplies a thousand more instances of the same pattern, with nothing in any of them guaranteeing the next one will match. Trusting that it will requires a leap the evidence itself doesn’t contain, and that leap can’t be justified by logic — it isn’t contradictory to imagine the sun not rising tomorrow — or by appealing to the fact that the future has resembled the past until now, since that argument just assumes the very kind of reasoning in question. This is the problem of induction, and Hume’s own conclusion was blunt: the confidence behind cause and effect doesn’t come from reason discovering a necessary truth. It comes from custom, or habit — the mind forming an expectation after enough repetition, an expectation that feels like certainty without ever being proven.
Kant read Hume and, by his own account in the 1783 Prolegomena to Any Future Metaphysics, it interrupted his “dogmatic slumber” and gave his philosophy a new direction. He spent close to a decade answering, publishing the Critique of Pure Reason in 1781 and revising it in 1787. He agreed with much of Hume’s diagnosis — the claim “every event has a cause” can’t be proven true just by unpacking the words, the way “all bachelors are unmarried” can. But where Hume concluded this meant the causal principle must be pure habit, learned entirely from experience, Kant proposed a third category Hume had missed: “synthetic a priori” knowledge, genuinely informative about the world yet knowable independent of any particular experience, the way mathematics is. His move was to argue that the mind isn’t a clean window passively letting the world stream in as it already is. It’s more like a pair of glasses fused to your face since birth, an active structure that organizes experience before it even reaches awareness. Among the built-in “categories of understanding” the mind supplies, rather than reads off the world, is causation itself. You don’t perceive cause and effect the way you perceive the color red; you perceive events already filtered through the category of causation, because that category is one of the rules a mind uses to turn raw sensation into an orderly experience at all. Kant argued this in the notoriously difficult “Second Analogy” of the Critique, where he tried to show that without applying causation, a mind couldn’t even distinguish looking around a house, whose parts can be seen in any order, from watching a ship sail downstream, a forced sequence.
This is also where Kant drew his distinction between phenomena — things as they appear to a mind structured the way ours is — and noumena, things as they might be “in themselves,” permanently beyond human knowledge as a matter of principle, not a technological limit. So when Kant said causation is objectively necessary, he meant necessary within the world of phenomena, the only world any human mind can experience, not a free-floating fact about a universe apart from anyone experiencing it. He called this position transcendental idealism: cause and effect really is necessary, but that necessity has its source partly in the structure of the experiencing mind, not purely in an external world sitting there ready-made.
What makes the exchange worth dwelling on is not just that Kant answered Hume, but what the answer actually secured. Kant’s case is a strong one: that the general principle “every event has some cause” is necessary for experiencing an orderly world at all, not a mere habit picked up along the way. A mind with built-in categories, causation among them, is a condition for coherent experience, rather than a conclusion drawn from experience afterward — and that is a different, sturdier kind of claim than anything Hume’s habit-formation story could deliver.
It is worth being clear about what Kant’s answer does not settle. What he did not fully secure is whether any particular causal law — that heat follows sunlight, that this chemical triggers that reaction — is itself necessary rather than, as Hume said, learned from repeated observation and generalized by habit. On particular causal laws, the ones science actually spends its time discovering, Kant essentially agreed with Hume: we get those from experience and induction, with no ironclad guarantee attached. So the honest summary is that Kant answered a deep version of Hume’s problem, why causation as a general framework structures experience at all, without fully dissolving Hume’s original worry about whether any single inductive inference, about the glass, about tomorrow’s sunrise, carries genuine necessity. That narrower problem of induction is still, in a meaningful sense, unresolved more than two centuries later, revisited by nearly every major theorist of scientific knowledge since.