Well, for starters, the 1200-page _Standard Dictionary of Folklore, Mythology, and Legend_ (rev. ed. 1972) offers this:

"Folk song comprises the poetry and music of groups whose literature is perpetuated not by writing and print, but through oral tradition."

This seems to imply that folksongs are peculiar to people who are entirely or at least functionally illiterate.

In a complex post-medieval society, how do we know that's true? It becomes "true" only if we're willing to accept it. Are we?

When literacy was extremely rare, there were many songs like that. But how many are there now? Or a hundred years ago, or three hundred? How can we know in a population of millions whether most singers of a particular song can barely read or write? And why exactly is that criterion so important?

I'm not ridiculing the definition, just suggesting that it may not work for everyone. And if it and others don't work for *nearly* everyone interested in "folksong," we're back at Square One.

The same reference work, by the way, gives no less than twenty definitions of "folklore" by various authorities. (The article on folksong was, interestingly, written by just one authority, George Herzog.)

Bert's point about collectors being influenced by their own tastes in creating a definition is correct. Which means even a consensus definition will have very fuzzy edges, allowing the question, "But is it *really* a folksong?"

Those of us who prefer a fairly narrow definition (whatever it might be) are derided as wet-blanket "pedants" and "folk police" by everybody else. And we deride them in turn for being crudely "uninformed" and "undiscerning."

The label, if you ask me, is usually of less interest than the song and how it fits (or doesn't fit) into folk or popular or refined culture.