More on theory and fact (medicine and ritual)
There was an interesting discussion on the BBC's Today Programme (click for audio) over some recent Lancet-published study doubting the efficiency of homeopathic treatments vs. a study conducted at the Bristol homeopathic hospital showing that about 70% of patients report some improvement in how they feel.
Since many educated, and otherwise critical, people of my acquintance also subscribe to what I always thought was the nonsense of homeopathy (Skrabanek's and McCormick's Follies and Fallacies in Medicine (Amazon.com) made a very convincing case against most 'alternative' treatments) I have been recently thinking quite a bit about how to reconcile the two views.
1. It seems to be a fact that people report improvements in their condition
2. However, any such possible result by 'trace elements' (roughly a molecule in a bucket of sugar water) flies in the face of all our folk and theoretical knowledge of chemical causation (also a fact)
Maybe the problem really is in the our trying to square a particular fact with a particular theory and having to reject them (either depending on the theory). Of course, we could have a completely different theory (call it 'complementary', if you will) for each set of facts.
Of course, conventional science tries to cover the 'effects' of homeopathy (and other such treatments) under the umbrella of placebo (see here for an interesting discussion and here for a definition). But that is not very satisfactory to the 'complementary' crowd.
In these contexts, I am always reminded (and I'm open to being shown its irrelevance) of Claude Levi-Strauss' point that (in magic as in science) it is not as important whether there is a causal link between the touch of a bird's beak and a broken bone but whether we can establish a symbolic connection. (I'm paraphrasing from memory - if I track down the quote, I'll post it here).