Hume and the Problem of Induction

David Hume published his masterpiece at twenty-eight and it failed completely. It fell dead-born from the press, as he put it later, with characteristic good humour. What it contained was the most corrosive set of arguments anyone had produced about the foundations of ordinary knowledge, and the corrosion is still working.

Start with the one that has caused the most trouble. Every prediction you make relies on the assumption that the future will resemble the past. The bridge held yesterday, so it will hold today. The medicine worked in trials, so it will work in patients. Ask what justifies that assumption and you find nothing.

It cannot be reason, because there is no contradiction in supposing nature changes course tomorrow. You can imagine it perfectly clearly, which you cannot do for a square circle. And it cannot be experience, because any argument from experience (it has always held so far, therefore it will hold) already uses the principle it is trying to establish. The argument is circular, and the circle is not a small technical flaw but the whole structure.

Hume's conclusion is not that we should stop predicting. It is that the thing doing the work was never reason. It is custom: the habit of expectation built up by repetition, which operates below the level of argument and would keep operating even if you were fully convinced by everything in this chapter.

The same method dismantles causation. Watch one billiard ball strike another as many times as you like. What do you actually observe? Contact, then motion. Contact, then motion. You never at any point observe the necessity linking them: the force, the making-happen, the because. What supplies the feeling of connection is the mind's habit after repeated pairings, which we then project outward onto the world and mistake for something we found there.

Then he turns the instrument on the self. Look inward and try to catch the self doing the looking. Hume reports that he only ever stumbles on a particular perception (warmth, light, irritation, a thought) and never on the perceiver having them. If the self were a genuine impression it would be constant, and nothing in experience is constant. So the self is a bundle of perceptions in rapid succession, held together by memory and resemblance rather than by an underlying thing.

There is a moment in the Treatise that tells you what kind of thinker he was. His own copy principle says every simple idea derives from a prior impression. Then he raises the missing shade of blue: lay out every shade except one, in order, and a person who had never seen the missing shade could probably supply it from imagination. That is a direct counterexample to his own rule. He prints it, calls it unimportant, and moves on.

Finally, the paragraph that has generated the most secondary literature per word in philosophy. Reading moral arguments, he notices they proceed through statements of fact and then, without notice, switch to statements containing ought. Since ought expresses a new relation, the move needs explaining, and no set of purely descriptive premises seems to supply it. He flags it and moves on. Whether it proves that values can never be derived from facts, or only that people are sloppy, is still contested.

What saves all this from paralysis is Hume's own answer to it. Nature cures the doubt. You leave the study, play backgammon, dine with friends, and the arguments lose their grip entirely. His claim was never about how to live. It was about what reason can justify, and the answer turned out to be much less than anyone had assumed: a finding that behavioural science, several centuries later, has broadly confirmed.