Treating the Weapon
The man known as Paracelsus (see also Wikipedia’s entry) was unpopular and vilified in his day. But roughly thirty years after his death, his beliefs were taken up by those who wished to overturn Galenic physics, and so he ended up becoming highly influential and seriously affecting the development of medicine.
Many of his ideas are now recognizable as being worthwhile, although many are absurd, obvious superstitions. It’s worth crediting Parcelsus for what he got right, given that he represented a serious break with the worst elements of medieval thought, but his views were far from a modern understanding.
One of his more peculiar ideas was that injuries caused by weapons were best treated by ministering to the weapon responsible – or, if that were unavailable, to a stick coated in the blood of the injured person. The weapon (or stick) was rubbed with various ointments, wrapped in clean linens, and put away in a warm, dry place. To keep the weapon from spoiling.
What is extraordinary about this belief was how long it was adhered to. I recall reading about how advocates of this method kept pestering the rebels in the American Revolution to treat gunshot injuries this way – particularly how one man claimed he had done a careful comparison of standard treatments and Paracelsus’, and found that weapon treatment was superior.
Thing is, he was right. What he didn’t realize was that treating the weapon was absolutely useless. It had no effect. It did nothing – except displace the normal treatments. The standard responses to such injuries were actually worse than doing nothing.
This highlights the importance of always keeping a control group when evaluating a method of doing something. If you don’t know what doing nothing looks like, you can’t judge what effect doing anything else has – including whatever it is that you want to test.
There’s a saying that resulted from this:
Treat the wound, not the weapon.
there should be another:
Test your assumptions, not your hypotheses.
Testing your hypotheses is important, but it can’t be done until your basic assumptions are validated.