The argument made against homeopathy usually goes like this:
1. There's nothing in homeopathic medicines
2. If there's nothing in them, they can't work
3. Therefore, homeopathic medicines are placebos
It's a strange argument, because placebos work, by definition, and yet there's nothing in a placebo either. So there's at least one example of a kind of medicine with nothing in it that works. So why not two or more? I'm not saying that homeopathy is not a placebo. I'm saying that this argument does not establish it, because the second premise of the argument is false. Whether homeopathy is placebo can only be established through experimental trial. If trials show that homeopathy gives results statistically different than placebo, then homeopathy is not a placebo. But whether homeopathy is not a placebo can't be established through a priori argument. This is simply because we don't understand what placebos are or how they work. We don't know them when we see them, so we can only determine what is a placebo through experiment. And that is why the a priori argument against homeopathy is just so much hot air.