Introduction
David Hume published A Treatise of Human Nature in 1739. He was twenty-seven years old. He expected it to cause a sensation. It fell, as he later wrote, dead-born from the press. Nobody noticed. He revised his ideas, published them again in a more accessible form, and spent the rest of his life as one of the most widely read and socially sought-after philosophers in Edinburgh, a man famous for his good humor, his appetite for conversation, and his ability to hold a deeply unsettling idea without being unsettled by it.
The idea, briefly stated, is this: we cannot prove that the future will resemble the past. We assume it constantly. All of science depends on it. When we say that the sun will rise tomorrow, or that fire will burn, or that a drug that worked in trials will work in practice, we are making an assumption about the uniformity of nature that cannot itself be demonstrated from experience. Every proof we might offer would be circular, assuming the very thing it is trying to establish.
Hume called this the problem of induction. It has not been solved. And yet here you are, reading this, and somewhere outside a kettle is boiling.
I. The Gap Between What We See and What We Conclude
The argument begins simply. We learn from experience. We touch a hot stove, we pull back, we learn not to touch hot stoves. We observe the sun rising, day after day, and we form the expectation that it will rise again. This is the basic mechanism of empirical knowledge: repeated observation generates expectation.
But Hume noticed something uncomfortable about this mechanism. The expectation is not in the observations. It is something we add. We see the sun rise a thousand times, and from those thousand observations we derive no logical necessity that it will rise again. All we can say with strict accuracy is that it has risen, not that it must. The connection between past and future is a habit of mind, not a law of nature that experience can verify.
What makes this more than a philosophical puzzle is what it implies about the nature of causation. When we say that one thing causes another, we are saying something that goes beyond what we have ever directly observed. We have seen events follow one another. We have never seen the necessity that connects them. The causal connection, Hume argued, is something the mind imposes on experience, not something it discovers there.
This is a genuinely strange conclusion. It means that the bedrock of scientific reasoning, the assumption that the world is regular, that like conditions will produce like effects, rests on something we cannot prove and cannot observe. It rests, Hume said, on custom and habit. On the way the mind is built.
II. What to Do with a Problem You Cannot Solve
What is remarkable about Hume is not the argument. It is his response to it. He does not collapse into skepticism. He does not conclude that nothing can be known. He concludes something more nuanced and more useful: that we should hold our beliefs with the degree of confidence that the evidence actually warrants, and no more.
This sounds obvious. It is not. Most of us, most of the time, hold our beliefs with far more certainty than the evidence supports. We take the regularities of experience as guarantees. We mistake familiarity for proof. We confuse the fact that something has always been so with the claim that it must always be so.
Hume's problem of induction is, in this sense, not just a puzzle for philosophers. It is a practical discipline. The person who has genuinely absorbed it holds their conclusions more lightly. They have stopped asking for a floor that is not there, and learned to stand well without it. Kant, reading Hume, said it woke him from his dogmatic slumber. The fact that the greatest systematic philosopher of the modern era spent his most productive decades responding to Hume's small, cheerful problem suggests that the problem was real.
Closing Reflection
Hume walked home from his studies and ate his dinner without apparent distress. He had found that human knowledge rests on foundations that cannot be fully justified, and he had made his peace with that. Not by pretending the problem was not there, but by concluding that a life lived with calibrated uncertainty is not a lesser life. It is simply a more honest one.
The kettle boils. We cannot prove it had to. We put our hand above the steam anyway, and draw it back, and learn something real. Perhaps that is enough.