Jones and Harris built the cleanest possible test in 1967. Participants read an essay defending Fidel Castro, and were told outright that the writer had been assigned that position, not chosen it. No choice was involved; the writer might privately think the opposite. Participants still concluded the writer probably held pro-Castro views.

That's the fundamental attribution error in its starkest form: even when the situational explanation is handed to you explicitly, character wins. Lee Ross gave it the name in 1977 and called it bedrock for social psychology, and it does explain an enormous amount of ordinary unfairness — the rude waiter, the careless colleague, the driver who cut you off, all of whom you observed for seconds and diagnosed for life.

Two caveats keep this accurate, and both are usually dropped. First, the word fundamental overstates it: comparisons between American and Indian adults found the situational default was the norm in the second group, and the divergence grew with age — which looks much more like something taught than something built in. Second, this is one half of a pair, and the other half broke. The familiar extension — I flip to situational explanations when it's me — is the actor-observer bias, and a meta-analysis of 173 studies found the overall effect close to zero.