Thanks, Brian.

Grishka, you're right about dictionaries, encyclopedias, and "typical use by the language community."

But there's frequently more than one "typical use," and reference books have to decide which one(s) to address. Sometimes (as for the word "run")there are dozens.

And appeals to "typical use" (especially when there's more than one) can be a poor standard for the kind of specialized discussions we have here, where carelessness will cost you. (In "typical use" a whale is often a fish and a predicted meteorite shower is "weather.")

> A rather primitive method is to identify, say, five yes-no criteria, so that an object that satisfies three of them is regarded as a "60% folk song".

I doubt anyone has done this or would want to. Examples of non-material culture, like songs and performances, aren't analyzable in the manner of chemical compounds. It would be far more useful to say that one of Seeger's (or even Steeleye's) performances has certain "folklike or 'traditional' characteristics" and then describe them. (Seeger would generally score higher on the percentage scale, but I'm nor sure that fact by itself is of much interest,)

Is a pizza "really" a "pie"? How about in "typical use"? (I don't believe the public in any country has been surveyed on this, so we have to base "typical" use on our own experiences.) Does it matter?

Well, it might in a lawsuit. Otherwise, what's on the pizza is more interesting.