Richard Feynman
The first person not to fool is you
And you are the easiest one
Caltech commencement address, 1974; Appendix F to the Challenger report, 1986
He put this ahead of every method in science. This page is about why the first line of defence points inward.
What actually happened?
At the 1974 Caltech commencement he gave the graduates one rule and called it the first principle: you must not fool yourself, and you are the easiest person to fool. Twelve years later he sat on the Challenger commission, asked for a glass of ice water at a televised hearing, clamped a piece of the O-ring rubber, dropped it in for a few seconds and took it out. It did not spring back. NASA had hundreds of pages arguing the launch was sound. His appendix said reality must take precedence over public relations.
The bias sits wherever your interest sits
Fooling yourself takes no dishonesty, only a stake. Wherever you would prefer a conclusion, your efficiency at finding supporting evidence rises sharply. So the check is not how honest you feel. It is whether the answer pays you.
Positive evidence is unusually easy to find about the project you sponsored. That is the moment to hand it to somebody whose bonus does not depend on it.
Write the disconfirming column first
The order changes the content. Write down what would show you are wrong before you write the support, and both columns come out more honest. Reverse it and you have already chosen a side by the time the second column starts.
Put the falsifying evidence in the first column of the review template. Same person, same week, same facts, different conclusion.
When the report and the world disagree, change the report
Challenger was not a shortage of engineering. It was an organisation that preferred a document. This holds anywhere there is a reporting line: if the reflex when numbers look bad is to adjust the wording, the problem has simply been forwarded.
The metric missed. Ask whether you are revising the conclusion or the phrasing. Revised phrasing clears this review and is still there at the next one.
How do I use it today?
Where you are: assessing something you have a stake in.
Ask first: does this conclusion pay me? If it does, how hard have I actually looked for the evidence against it?
Where it goes wrong: turning it into never concluding anything, or aiming the principle exclusively at other people.
Lines to keep
The first principle is that you must not fool yourself.
And you are the easiest person to fool.
Reality must take precedence over public relations, for Nature cannot be fooled.
Same situation, other people are asking
Everything I find supports me. Should that worry me?I want this to be true, and I'm the one checking it.The numbers came out badly. Do I fix them or fix my view?What would have to happen for me to admit I was wrong?I might be fooling myself · all 6 questions →If this one named what you are going through,
send it to someone who needs it, or keep it somewhere you will find it again.