Being right, being early, and being unbearable
In 1847 Ignaz Semmelweis introduced chlorinated hand washing in a Vienna obstetric clinic and maternal mortality fell dramatically. Clinic registers support the effect. The familiar story of one obvious truth rejected by fools compresses a harder episode involving an unknown mechanism, delayed publication and Semmelweis’s increasingly hostile advocacy.
Semmelweis compared two maternity divisions: doctors and students moved from autopsies to deliveries in one, while midwives staffed the other. After a colleague died from a wound sustained during an autopsy, he proposed that “cadaverous matter” was carried on hands and required chlorinated washing. Germ theory did not yet supply a shared mechanism. He reported strong results but was slow to publish a complete account and later accused opponents of murder, making an already difficult claim harder to absorb.
The case is not a licence for every rejected outsider. It asks what evidence can achieve before a field has the concepts to interpret it—and how presentation, replication and mechanism affect whether correct evidence becomes usable knowledge.
Semmelweis was sitting on close to an ideal comparison: two wards, alternating admission, one variable that plausibly differed. He reported the result but never designed a further test, and he refused for years to publish in a form his opponents could examine. The tragedy is not only that he was disbelieved; it is that he made being believed harder than it needed to be.
“Semmelweis reflex” now names the reflexive rejection of a finding that contradicts a paradigm. The term is most often invoked by people who are simply wrong, which has made it nearly useless. A parable about a rare event became a licence for a common one.
The parable is used to license any rejected claim whatsoever.
The asymmetry that matters is not that Semmelweis was rejected — most rejected claims are rejected correctly. It is that he had a controlled comparison and a large measured effect. Invocations of his name almost never come with either.
Nothing about the history; the useful question is about the inference drawn from it, not the facts.
What would you have to show to distinguish yourself from someone who is merely being rejected?
Return to todayA sealed edition.
Reading an old edition does not reopen today's door. The archive is a shelf, not a feed.