Watch one billiard ball strike another, and the second ball moves. It feels obvious that the first ball caused the second one's motion. The eighteenth-century philosopher David Hume pointed out something genuinely unsettling about that obvious-seeming judgement: careful attention to what's actually perceived in the moment shows only two events, the collision and the subsequent motion, occurring one after another. The causal connection supposedly linking them — the actual "making happen" — is never itself something you see, hear, or otherwise directly perceive.
Constant conjunction is not the same thing as necessary connection
Hume's argument distinguishes between what he called constant conjunction — the repeated observation that one type of event is reliably followed by another type of event — and necessary connection, the stronger claim that the first event genuinely brings about, or necessitates, the second. We build an expectation that fire will produce heat, or that one billiard ball striking another will set it moving, from repeatedly observing that pattern hold. But Hume's point is that the repeated observation itself only ever delivers constant conjunction, sequence after sequence of one thing following another; it never delivers direct perceptual access to a necessary connection forcing the second event to occur, however strong the resulting expectation feels from the inside.
Causation as a habit of the mind, not a feature directly perceived in the world
Hume's conclusion was that our idea of necessary causal connection arises from a psychological habit — repeated experience of constant conjunction produces, in the observing mind, a felt expectation that the same pattern will continue, and that felt expectation is what we're actually calling "causation," rather than any additional feature of the external events themselves that perception directly delivers. This doesn't mean Hume thought causal reasoning was useless or that we should stop relying on it in everyday life; he explicitly argued we can't actually help relying on it. His point was more precisely epistemological: the strong, seemingly self-evident sense that event A necessarily brings about event B outruns what direct observation alone can actually justify, a gap between how confident causal reasoning feels and how much observation alone can support that confidence.
What we're still unsure about
Hume's argument that direct perception delivers only constant conjunction, not necessary connection, is a well-documented and highly influential position in the history of philosophy, still discussed extensively in contemporary metaphysics and epistemology. What remains genuinely, actively disputed among philosophers is how to respond to it — whether some other account of causation can restore a stronger, mind-independent notion of necessary connection, whether Hume's more deflationary, habit-based account is simply correct, or whether the whole question needs reframing in light of later developments in physics and probability theory that Hume himself couldn't have anticipated. There's no settled consensus resolving the problem Hume raised, only a long-running set of competing responses to it.
This sits inside The Nature of Time & Causation, one of seven topics in Metaphysics, one of five domains in Philosophy, one of seventeen subjects the app can quiz you on.