I've been a bit busy lately, with no time to chip in here, but can I take the discussion back to 'The Demon Lover', which seems a good example with which to examine the 'broadside origins' theory?

Steve wrote:
"Regarding the Scottish versions, it's difficult to know to what extent the interference of antiquarian poets created/re-created these. It's common knowledge that Scott was following Percy's precedent and most of the others were part of Scott's circle at one time or another, and as you know there is at least one amongst them whose published material is ludicrous (IMHO he hastily added)."

This is very contentious. For a start, there's an enormous difference – as far as the present discussion is concerned – between 'created' and 'recreated'. 'Created' suggests the Scots collectors invented the 'Scottish strain' of DL out of the ether, presumably with an old English broadside to hand for inspiration. 'Re-created' suggests that they amended or improved what was already present in oral tradition. There may be evidence for the latter, but none that I know of for the former. Scott's Minstrelsey text may well have been 'improved', and in any case relies on what may have been a suspect transcription by William Laidlaw. I presume your description 'ludicrous' applies to Peter Buchan, and FJ Child would be on your side there, although Buchan has his defenders. That leaves three remaining Scots versions in Child, one from Kinloch and two from Motherwell – one from his own Minstrelsey, and another (substantially different text) from his MS. I haven't heard Kinloch's probity questioned before and, as for Motherwell, I believe Scott directly discouraged him from 'improving' the ballads he printed.

To the five Scottish texts in Child we can add that found in Robert Scott's Glenbuchat MS from Aberdeenshire (coll. between 1808 and 1818), and the version collected 100 years later by Gavin Greig from Alexander Robb. While the latter is fragmentary, it nonetheless includes key 'Scottish strain' elements like the 'Banks of Italy' / 'see the fishes swim', the chronology of the voyage, the rising storm and the wreck of the ship. The Glenbuchat copy, a lengthy version, has much in common with the other Scottish texts, including a demonic transformation of the former lover. Taken to its extreme, Steve's analysis above seems to be suggesting that all six of the 19th century Scots texts of 'Demon Lover' are the result of some kind of conspiracy amongst "antiquarian poets". I find this hard to believe, not least since the history of the ballad in North American tradition contains many examples that hark back to the 'Scottish strain', and would most likely have their roots in the mass migration of the 18th century.

It seems pretty clear to me that there was indeed an 18th century oral tradition of this ballad in Scotland, and that this bore precious little debt to the Laurence Price broadside of 1657, or to the 18th century English printed copy. The only oral version that has any substantial resemblance to either broadside is that collected by Baring-Gould, and it's a bit surprising to find you, Steve, resting your argument on that particular source! If you want to convince me (one who is interested in your ideas) that "between a third and half" of the Child Ballads are "artifacts of the broadside songwriting industry", then I hope that 'Demon Lover' is not one that you're proposing to include in that category.