What is an Ad Hominem Fallacy?

Start by reading this out loud to your kid.

Bernadette the beaver slapped her tail on the water. "Listen up! Big storm coming. If we don't build the dam higher, the whole pond floods by morning!" The ducks rolled their eyes. "Oh, sure," quacked the loudest one. "A beaver says we need more dam. You only want that because you LOVE building dams. You'd build a dam out of your own breakfast if we let you." The other animals laughed and paddled off to bed. That night it rained, and rained, and rained, and by morning every last duck woke up floating around a very soggy living room.

Ask your kid whether the ducks made a good call. You're listening for something like: it's true Bernadette loves building dams, but that doesn't tell you whether the flood is actually coming. The ducks needed to look at the storm, not at Bernadette's hobby. Loving dams and being right about the flood can both be true at once.

This is the trickiest kind of ad hominem, the one about bias. The ducks weren't completely silly to notice that Bernadette benefits from more dam-building. That's a fair reason to double-check her. The mistake was stopping there. Someone having a reason to say something is a cue to look harder at the evidence, not an excuse to ignore it. Bernadette had a bias and she was right. The ducks skipped the part where they actually looked at the sky.

In the 1840s, Ignaz Semmelweis, a physician at the Vienna General Hospital, noticed that the maternity ward run by doctors and medical students lost about one mother in ten to childbed fever, while the ward run by midwives lost far fewer. The doctors, unlike the midwives, often arrived at deliveries straight from performing autopsies. Semmelweis had them disinfect their hands in a chlorinated lime solution, and mortality in his ward dropped roughly tenfold. He reproduced the result after moving to a hospital in Budapest.

The evidence was strong and repeatable, and the profession largely rejected it. Some of the resistance was legitimate: there was no accepted theory of why it worked, since germ theory was still decades away. But much of it was ad hominem. Accepting the finding meant conceding that doctors' own hands had been killing patients, and rather than engage that possibility, colleagues went after the man — abrasive, eccentric, foreign, unscientific. Those descriptions were largely accurate, and all of them were beside the point: the mortality figures did not depend on whether Semmelweis was easy to work with.

He was forced out, deteriorated, and died in an asylum in 1865, at forty-seven, of an infection. Germ theory arrived shortly after and confirmed he had been right. The case shows the fallacy's cost precisely. For years the profession answered the question "is Semmelweis insufferable?" — which was real, and irrelevant — while the question that mattered, "is he correct?", was answerable straight from his own mortality tables, and went unexamined.