Steve,
A good idea to give this topic a thread of its own. I hope we see plenty of discussion of ideas which put a bit of a bomb under the notion of folk song as some of us used to understand it. As I said on the other thread, I find the notion of broadside or stage origins plausible for at least a part of the folksong canon, but I think the story re ballads is more complicated. Hence the following discussion of 'The Demon Lover'.

Steve wrote:
>> Firstly the main thrust of David's article in FMJ Vol 5 Number 5 demonstrates that the 17thc broadside 'A warning for Married Women' ascribed to Laurence Price is undoubtedly the model for the later 18thc broadsides 'The Distressed Ship Carpenter' and that these 18thc/19thc broadsides are very likely the origin of all of the versions collected in oral tradition , particularly in England and North America. <<

Before discussing this, it might be worth mentioning for newcomers' benefit that our mutual friend David Atkinson is the present editor of the Folk Music Journal and a ballad scholar in his own right. And from here on I'll follow Child's practice and refer to the C17th Price broadside as 'A' and C18th 'Distressed Ship Carpenter' as 'B'.

The main thrust of DA's article in FMJ 5 is that the transgression committed by the woman in the ballad, that results in her punishment (by demonic intervention or simple disappearance), is not her adulterous elopement with her former lover, but the earlier breaking of her betrothal vows to this former lover. It's true that in discussing the "alteration of the ballad" with time DA might be implying a direct lineage from A, although elsewhere he states that A "does not necessarily represent the beginning of… The Demon Lover".   Personally I think that the similarities between A (1657) and B (printed in a songbook c1737) are quite limited: it's not just that the first 16 verses of A (a long-winded account of the betrothal and parting) are entirely absent from B (which, in classic ballad fashion, 'begins in the third act'), but B introduces the "Well met, well met" first verse well-known in English and North American (but not Scots) tradition, the bribe of golden slippers, three verses describing the woman's regret during the voyage, and her death by drowning. Only six of A's 32 verses have analogues in B.   The differences between A and any of the texts collected from singers are even more marked.

>> The Scottish versions are something different and I don't want to go into my views on those here and cloud the issue. <<

I'd be interested to hear your views on the Scottish versions, as I don't think you can discuss this ballad without them. Certainly apart from Buchan's 'James Herries' they have very little in common with A (although they do sometimes mention betrothal vows), and all kinds of differences from B, including the demonic nature of the former lover, the promise of seeing the 'Banks of Italy', the incremental repetitions ('They had not sailed a league, a league'), the Hills of Hell, the shipwreck, etc. The survival of some of these 'Scottish' elements in North American versions argues against the vibrant American 'Housecarpenter' strain (itself stabilized by a 19th C broadside) deriving directly from those early English printings.

I only know of two versions collected in England, and they don't tell us a lot: Marina Russell had only three verses – ones commonly found in most printed and oral versions – and the lengthy text collected by Baring-Gould is so uncannily similar to B (apart from an interpolated middle section) as to suggest a conscious rewrite.

Although any conclusion here is necessarily speculative, it seems to me that Heylin's contention that 'Demon Lover' existed, probably in more than one distinct strain, before its partial annexation for Price's broadside A, is as likely - if not more so - than to suppose that A is the original template for all that followed.

>> David does indeed present the possibility that some of these MIGHT have predecessors. In fact he uses the word 'might' rather a lot in the article, where 'might not' would be just as apt. Having spoken to David on several occasions since that article, I think I can safely say I think his view has changed somewhat on the relationship between broadsides and the oral tradition. If you read the article carefully what he actually is saying is that some of these early broadside ballads were based on earlier models or stories. Absolutely. Nearly all of Shakespeare's plays were such, but I don't think anyone would claim he didn't write them. <<

If DA uses the word 'might' it's probably because he – like me – trained originally as a scientist, and as such is reluctant to pronounce certainty without hard proof. However, my understanding is that several of the examples he quoted are verified MS versions, in stanza form, of ballads predating the earliest known broadsides. These were 'Riddles Wisely Expounded', 'Lord Thomas and Fair Annet', 'Hunting of the Cheviot', 'Sir Andrew Barton' and 'Fair Flower of Northumberland' – the last of which he cites as a probable case of a broadside printer (Deloney) appropriating a ballad from oral tradition. I accept your Shakespearian analogy – it had occurred to me as well – but those examples are not simply a matter of versifying old tales. I've also read of an account in 'Complaynte of Scotland', of a shepherds' ballad session in which the likes of 'Tamlene' and 'Hunting of the Cheviot' were apparently being sung in oral tradition in 1549.

Having shared many a small-hours ballad discussion with Dave, I can reassure you that he's been stressing the importance of broadside dissemination of ballads for a good many years. I think he'd still stand by that 'Demon Lover' article, though.

Let's have some more views, then!