"Lighter":   thanks for the comments re the "tricky" cases; your take on Broomfield is how I was seeing it; which is why it adds to the class to compare it.

I did not know the detail abt streets of loredo never existing in print. I've seen it argued (perhaps in the liner notes of the Smithsonian/Folkways album on UR variants?) that St James Infirmary is included as a descendant not only for the hospital name but also that the speaker is visiting a dead/dying lover, and then giving instructions for his funeral; and vaguely blaming a dissolute lifestyle for his downfall ("I've got those gambler's blues".) But to me it's hard to tell if all that is "convergent" (independently arising song on a universal theme). The hospital name seems like the smoking gun.

Point taken that changes made intentionally by broadside writers would have very different motives or causes or constraints.

Which gets around to an idea that folk texts could be more or less descended from one or more earlier text(s) that has/have been transmitted through more or less oral channels. Despite the lack of any purely oral case, it seems that characteristics of folk song texts are distinctive and related to their greater degree of oral transmission.

Getting back to my OP, however, on tunes I'm still pretty unclear what we can say. what the students notice most is
- that they are modal (i.e. they notice the funny sound and we track it down to that); but I don't know why this is more common in old/folk songs than later/elsewhere.

- that they are strophic (or as the kids say "boring" -- they've never heard a song without a bridge). this structural simplicity seems consistent with aiding memorization.

- I think they'll hear that different versions of one song like Barbara Allen have similar but not identical tunes; while other versions have seemingly unrelated tunes. That part seems analogous to the text topic but as we've seen, difficult to pin down.