In a surprising number of cases, some of the most famous logicians either: - learned logic from an inconsistent system, or - developed an inconsistent system and didn't notice for some time Examples: - Frege's system was inconsistent, Russell (seems to have?) learned logic from it. - Church's first system was inconsistent. Kleene clearly stated that he learned logic from it. - Haskell Curry's combinatory logic was inconsistent. According to our standard informal reasoning about inconsistent formal systems, this should be impossible. In practice, inconsistent formal systems have surprised us not by their uselessness, but by their usefulness. Not by their triviality, but by their similarity to consistent systems. Of all facts in logic, no other is more important and less explored.