How do you prove it correct?
In Jake's example of the deadly mushroom hunt, were you to partner with a shaman from a local culture it is highly likely that you would only eat mushrooms that had properties that he predicted. Scientifically minded non-botanists who were foolish enough to participate in such a procedure on their own would be subject to the same winnowing he described. Hell, even animals usually do better than that. How do they manage? As the Dutch said while fighting the Spanish: "It is not necessary to have hope in order to persevere."
So yes, the shaman may know which mushrooms to pick. In fact, if I had to choose between a local who's been mushroom hunting for years on end in the same area and a botanist who's never studied the particular area in question, I'd go with the local. Hands down, no contest. But that doesn't mean that the local's explanation of how you can tell which mushrooms are poisonous makes sense. It's perfectly possible to construct a memorable narrative that gets all the known facts right and yet is completely incapable of dealing with new facts - i.e. has no predictive power.
Local knowledge is, in fact, often used as a starting point for scientific enquiry - but usually the first part of the exercise is to strip out all the concrete facts and discard the fluffy narrative.
So going back to the acupuncture example, you'd start by investigating whether sticking needles into people at random points produces medically interesting effects. It's perfectly testable whether "traditional" acupuncture has stumbled upon a useful fact that sticking needles in people actually triggers a useful biological response.
The second thing to investigate is whether sticking needles into certain points produces effects that differ from sticking needles into semi-random points. Again, this is a perfectly testable proposition.
Chakras, however, are not useful at any point in this exercise, except insofar as they can serve as mnemonic tricks, similar to learning multiplication by doing times tables. The times tables aren't interesting in and of themselves, they're just a scaffolding you can use to memorise facts. By contrast, while scientific theories can serve as mental scaffolding, their greatest value is that they can be extrapolated to cover unknown situations.
Faced with an unknown mushroom, our shaman wouldn't know whether it was poisonous or not (he'd very probably know whether to eat it or not, but that's not quite the same thing :-p). The botanist wouldn't either, of course, it being an unknown mushroom. But the botanist would know how to run a chemical analysis on it and compare the results to a list of known chemical substances. For that matter, if he has a good chemist at his lab, he'd be able to extrapolate from the known toxicity of related chemicals to hazard an educated guess as to whether the mushroom is dangerous or not.
- Jake If you only spend 20 minutes of the rest of your life on economics, go spend them here.
Not necessarily, because this assumes that it's the needles and the sticking which are important.
Something else could always be going on. A more interesting test would be a combination of self reporting - not for definitive results, but for pointers - with a battery of tests for a wider spectrum of outcomes.
