> What is pretty evident in the study of Anglo-American ballads, and murder ballads in particular, is that the Americanized versions of particular songs tended to lose the ghosts <

> My own belief: this is merely a reflection of the American tendency to simplify ballads, sometimes to the point of losing the narrative. <

Simplifying ballads happened to a greater or lesser degree everywhere - I see no particular evidence that it's an 'American tendency'. Remember that singers were often being asked by song collectors to remember songs that they had learned as children, or at least not sung actively for many years.

Nor am I convinced that 'Americanized versions... tended to lose the ghosts'. Which ballads do you mean? If you're comparing the C18 and early C19 texts published by Child - mostly from Scots sources, some remembered only as recitations, and some quite possibly 'improved' by poets - with those collected from oral tradition in Appalachia or the Canadian Maritimes a hundred years or more later, you're not comparing like with like.

If you look at the versions of those ballads collected from Scots and English oral tradition in the 20th century (which would be the fairer comparison), you find the same kind of 'simplifying' processes ocurred there too. And, as I pointed out above, there are many examples of the supernatural in C20 American ballad variants.