Learning and growth
Karl Popper
It has to be refutable
Karl Popper (1902–1994 · Vienna and London) — It has to be refutable.
What actually happened?
In 1919, before any of the philosophy, he was volunteering in Vienna's child guidance clinics. He brought Alfred Adler a case that did not look to him like an Adlerian one, and Adler explained it in terms of inferiority feelings without difficulty, and without having seen the child. Popper, uneasy, asked how he could be so certain. Because of my thousandfold experience, Adler said. Popper could not stop himself: and with this new case, I suppose, your experience has become thousand-and-one-fold. What struck him was not that Adler was wrong. It was that nothing could have counted against him.
Explaining everything explains nothing
If any result at all can be absorbed by a theory, the theory carries no information. The way to check is to ask what observation would make its holder abandon it. No answer means no test, and no test means no reason to lean on it.
Sentiment drove the rise. And the fall? Sentiment again. It accounts for every outcome and therefore accounts for none.
Write the refuting condition before the result
People are remarkably able to read any outcome as support. The only defence is to fix the criterion in advance. Standards set afterwards give you a position that cannot lose, and therefore cannot teach.
Before shipping, write down the number that would mean the assumption was wrong, and the date on which you will go and look at it.
Refuted is not the same as discarded
The usual misreading is that a theory shown wrong is finished. His demand was the opposite: that it be exposed to serious attempts to break it. Good ideas mostly get stronger by surviving those attempts.
Once a quarter, go looking for the strongest evidence against your main position. Failing to find it is the only real endorsement.
Collect counter-examples, not confirmations
The default is to gather cases that agree, and the hundredth one adds close to nothing. A single case that disagrees is decisive. So the productive hour goes on finding the exception, not extending the list.
To find out whether people like the product, do not survey the people still using it. Talk to three who left.
How do I use it today?
Next time somebody hands you a confident claim, ask what would have to happen for it to be wrong. If nothing would, treat it as a story rather than a finding. And before your own next bet, write down the condition that would make you drop it, with a date attached.
Deep read
Read alongside
Further
Lines to keep
We cannot prove a theory true. We can only fail to refute it.
A claim that fits every outcome has told you nothing.
One counter-example outweighs a hundred supporting cases.