The Feynman Technique, Examined

Richard Feynman did not, as far as the historical record shows, ever describe a single technique by that name. The method that has come to bear his name was distilled from his attitudes toward learning by writers and educators after his death, and it has acquired the somewhat mystical reputation that often attaches to things attributed to dead physicists. Stripped down, the technique is simple.

Step one: pick a concept you want to understand. Step two: try to explain it, in your own words, in plain language, as if to a person who knows nothing about the subject. Step three: notice where the explanation breaks down. Step four: go back to the source material, learn the parts that exposed your ignorance, and try the explanation again.

The reason this works is not that explaining to a beginner is magically clarifying. It is that the act of writing or speaking forces a particular kind of compression that silent reading does not. When you read a sentence, your brain produces a feeling of comprehension that is loosely connected to actual comprehension. The feeling is generated by recognition — these words are familiar, these concepts fit a pattern I have seen before — and recognition is cheap. Generation is expensive.

When you try to explain, you have to generate. You have to produce the words, in the right order, that capture the idea. You discover, immediately, that there are gaps. You thought you understood why entropy is a measure of disorder; you cannot explain why this is the right intuition or in what cases it is misleading. You thought you understood compound interest; you cannot say what the difference between continuously and annually compounded means in plain language. You thought you understood quantum tunneling; you can recite the words but not produce them in an order that would convince a curious teenager.

These gaps are the point. They are not evidence that the technique failed. They are the diagnostic. The gaps tell you exactly where to study next.

The technique has a seductive failure mode. It is possible to write an explanation that sounds plausible to you, because you wrote it, and feel that you have understood. The test is whether someone who does not already know the material can follow your explanation and use it. Without that external check, the explanation collapses into self-flattery.

There is a related principle that Feynman did articulate himself, in his teaching. He warned his students against the difference between knowing the name of something and knowing the thing. A child can learn to identify a bird as a wren — knowing that the name is wren — without learning anything about how the bird actually lives. The verbal label produces a feeling of familiarity that masks the absence of understanding. This pattern runs through education at every level. Adults can recite the names of philosophical positions, mathematical theorems, and chemical processes without being able to use any of them.

The Feynman technique, taken seriously, is a defense against this. It is a way of constantly reminding yourself that recognition is not understanding, that fluency is not mastery, that being able to nod when someone uses a term is not the same as being able to use it yourself. The technique is hard because the natural human bias is toward feeling competent. The technique is useful because the cost of being unable to do something is much higher than the cost of finding out, early, that you cannot.