Grishka wrote: Back to Sumer. The theory is tempting, but I'm not convinced. A "d" on beat 4 without proper stepwise progress would be irregular.

There must be scholars who have examined the MS with all its strange stains in many places.

There are many books worth consulting. Both the original and revised versions of Chappell are worth looking at, and Carleton Brown, editor, English Lyrics of the XIIIth Century and Bruce Dickins & R. M. Wilson, editors, Early Middle English Texts are vital resources. I believe there is also a discussion in Karin Boklund-Lagopolou, I have a yong suster: Popular song and Middle English lyric. And there are others that I haven't seen; the Harley Lyrics are too early for me to have given them much study. I know much more about the other major collections of Middle English secular verse, such as the Sloane Lyrics, which are later and which contain a few items which were preserved into the modern era.

But that's not the issue, exactly. There is a difference between what is written and what is meant. Most of the references I've read about the song -- and if you consult the Ballad Index entry cited above, you'll know that I've read a lot! -- think that the scribe was struggling with the notation. I am not knowledgeable enough about the progression from neuming to plainsong notation to modern staff notation to pass judgment, but several have said that the notation on "Sumer" is too modern to be used in a manuscript as old as the Harley MS. Some think the scribe wrote the arrangement; others think he copied it. If he composed it, he might write it down wrong; if he copied it, he might transcribe it wrong. Did he read music well enough to see his error? I dunno.

And anyone who thinks manuscripts are accurate transcriptions of texts simply doesn't know manuscript studies. To take an extreme example, consider the Gospel of Matthew. The very first sentence -- which would be the very first words of any gospel codex, so the line for which scribes should be most alert, most prepared, and most likely to have memorized -- reads, in translation, "[A] book (of) ancestry of Jesus Christ son of David." The word "David" is spelled four different ways -- two different ways just in the two earliest manuscripts (out of about 2000 Greek copies): in effect, "Daveid," "David," "Dabid," and an abbreviation I am not going to try to reproduce given the way Mudcat butchers unicode. :-)

If you look at the Canterbury Tales, there is only one major variation in the first line (Averylle/Aprille), though that involves the two best manuscripts (Hengwrt and Ellesmere), but there are three in the second line (Add or omit And before The droghte; hath/had; add or omit "to").

My hand edition of the Iliad shows a variation ("soul" versus "head") as early as line 3, and much bigger variations thereafter.

I could cite more examples, but you're probably already bored. :-) The point is, in manuscript studies, especially when there is only one copy of something, you truly cannot assume that what you see is the original, even if you can be sure of what was written -- which you often can't. Trivial details sometimes hint at important conclusions -- but more often they're just scribal errors and not worth much attention.