Richard Feynman’s Lucid Dreams: The Experiments a Nobel Physicist Ran Inside His Own Sleep

Richard Feynman was one of the greatest physicists of the twentieth century, a Nobel laureate, and a man who took nothing on faith. So when he grew curious about what happens to the mind at the moment of falling asleep, he didn’t reach for other people’s theories — he ran the experiment on himself. And he found he could not only observe his own dream from the inside, but steer it.

What he actually did

It began with a student assignment. In a philosophy class at MIT, Feynman was told to write a theme on the ‘stream of consciousness,’ and it reminded him of a question his father had once posed: how does falling asleep actually feel — do your thoughts stop all at once, or slow down? To answer it, for four weeks he lay down every day with the shades drawn and watched exactly how he fell asleep. He noticed the thoughts didn’t switch off but lost their logical thread, growing more and more ‘cockeyed’ until they scattered into sleep.

Having trained that observation, one night he went further. In his own words: ‘I realized I was observing myself in the dream. I had gotten all the way down, into the sleep itself!’ Then he began running direct experiments. In one dream he was on a train and found he could turn around and walk back through the cars — ‘I could control the direction of my dream.’ And then the scientist’s honest caveat: ‘I could modify the direction of my dream, but not perfectly.’

He tested perception too. Seeing a red-haired girl in a dream, he tried to make out each individual hair — and saw them ‘as sharp as you want: perfect vision.’ In another dream he noticed a thumbtack in a doorframe: he looked at it and ran his finger down — and felt it; he looked away, the tack was gone, yet his finger still felt it. And when a knocking seemed to come ‘from outside’ the dream, Feynman deliberately woke to check the source — dead silence. The sound had been the dream’s own invention.

Why it works

This memoir narrative is interesting as self-observation, not a validated WILD protocol. A retrospective account cannot establish sleep stage or the causal efficacy of observation; in-dream checks are not identical to daily reality testing.

This is exactly what you can learn

Modern WILD, MILD and a journal are distinct methods. Do not label them Feynman’s exact protocol or promise the same control to everyone.

How to start

Start by protecting ordinary sleep and recording observations. Checks are optional and fallible; a five-hour alarm is no universally safe starting point.

How to walk this path systematically

A memoir story proves no private-coaching outcome. Fatigue is reason to stop induction, not push through for a famous example. Practice limitations matter more than reproducing another person’s experience.

MIT Libraries on Surely You’re Joking, Mr. Feynman! explains the book’s origin in recorded conversations with Ralph Leighton. This is catalog context, not the full original chapter; details and short quotations above are memoir retelling, not an independently recorded experiment.