IanC directed my attention to this thread in another one last week. I thought I'd add my 2 cents, even though the thread is old.

I can't altogether subscribe to the simpler definition of ballad IanC proposes because

(1) it does not exclude such genres as the epic. Indeed, by making the only necessary characterstics "narrative poem," it eliminates its own usefulness. Are "The Canterbury Tales" ballads? How about "Gilgamesh?" "Beowulf?" Kenneth Koch's "Ko?" I think not. The other characteristics are prefaced by "often" and "usually," suggesting they are not part of the definition per se but part of a description of some, or many, ballads.

(2) These associated characteristics are "often of folk origin and intended to be sung, consisting of simple stanzas and usually having a refrain." This is also a problem. It says that the poem is INTENDED to be sung, not that it WAS sung. This is a classic form of what scholars call the "intentional fallacy," defining a work by what we imagine to be the author's intention. It is a fallacy partly because we do not know the author's full intention, so scholars falling for this fallacy generally make it up for themselves and then claim it is the meaning of the work. Thus, although there is no evidence that the Auchinleck Manuscript's romance of Hind Horn is "intended to be sung" we can say "well, based on its stanzaic form, I think it was intended to be sung, so it is by definition a ballad." This is just as problematic with "The Two Sisters." The only evidence we have about the latter is that it WAS and IS sung; and about the former, we have no evidence either way.

So Child was operating in uncharted territory. He had to figure out what texts were in the genre he called "popular ballads" and what texts were not. Sure, he made up his definitional criteria. But they were based on the characteristics he observed in items he knew were or had been sung, either in English or in foreign versions. He even included some items (like "Geste of Robin Hood") that he KNEW were not ballads, because he thought they shed light on the properties of ballads.

One could, by the way, slap Child around either way: there were texts he apparently left out for ideological or moral reasons (Like "Froggie went a-courting," "The Crabfish" and "The Bitter Withy,") even though he must have known that they DID meet his own criteria for the popular ballad, others that he put in even though they did not. So he was human, and erred, and pandered a little to popular morality, and was a bit of a prude.

As to IanC's contention that he was not a collector, that depends on your point of view. He was not a field collector, but he did collect together and edit many texts from unpublished or hard-to find sources. Remember that those days were not these days, with Inter-Library Loan putting any book at any scholar's fingertips. It was hard work, as IanC recognizes, to do this. Finally, what few people have mentioned is his massive work of comparative scholarship directing readers to versions in other languages. To say "what's so special about Child?" is in my mind like knocking the scholars who put together the first dictionaries and encyclopedias. Sure, the information in those books was, by definition, common (or relatively common) knowledge. But the value of the books is undeniable. Where would a writer be without any dictionaries and encyclopedias? That's where a folksinger would be without any books like Child's.