This should be a book (and no doubt some of you will think it's close to book-length as it is), but I don't have time. I hope people will accept the summaries here as coming from someone who has spent thirty years studying ballad scholarship every day, and knows more than the odd comment people hear at folk concerts!

(Note that the above is not targeted at Nick Dow, even though I quote him below.)

Nick Dow wrote (regarding Communal Composition): It's an American term coined a few years ago.

More than a few years ago. It was Francis Gummere who came up with the Community Composition idea -- people sitting around in a circle and, I dunno, each one adding a word or something. The fact that no one had ever seen it happen didn't stop him.

The only way Gummere got away with it was that Gummere and his followers basically didn't look at anything outside the Child Ballad canon, where it was generally impossible to trace the full history. If they had looked at folk songs collectively, the utter absurdity of the idea would instantly have been obvious.

It's worth reminding ourselves that Child's work was called The English and Scottish Popular Ballads, not ...Traditional Ballads. The two are not the same! Anyone who has studied, for instance, the Percy Folio (one of Child's leading sources), and all the chopped-down romances it contains, would know that Child did not consider traditional status to be a necessary condition for inclusion in is opus. It was probably a sufficient condition, but there are plenty of Child Ballads that almost certainly were never traditional.

The fact that no sane person still suffers from the Communal Composition delusion does not mean that oral recomposition does not exist. Oral transmission can -- and does -- improve songs dramatically -- both adding and subtracting words. But there is generally a "skeleton" around which this happens. In Britain this often was based on broadsides; in America, where the tradition is more recent, pop songs often play a part.

Tunes of course often float, which is why I eventually included a "Same Tune" field in the Ballad Index, to allow people to track this. The lists for "Yankee Doodle" and "John Brown's Body" and "Vilikens" are astounding.... It is quite obvious that many writers deliberately used old tunes to make it easier for others to sing their songs.

John Baxter, who I believe posts here sometimes, is responsible for the Folk Song and Music Hall site. I'm discussing with the Ballad Index Board how we can incorporate links to that site. We don't have a ood answer yet, but it's a goal.

Now to the topics that certain people think are controversial, although they are mostly not controversial to experts: There are three terms which are quite distinct but are often used interchangeably:
TRADITIONAL
PUBLIC DOMAIN
AUTHOR UNKNOWN ("Anonymous")

Any song can meet, or fail to meet, any of the three criteria. So there are actually eight possible statuses. For instance:
"Barbara Allen" is Traditional, Public Domain, Author Unknown
Tom Paxton's "The Last Thing on My Mind" is Traditional (been collected in Newfoundland), not public domain, author known
A Lady Gaga song is almost certainly not traditional, not public domain, author known

Taking the last one first, truly anonymous songs are very rare, although there are some. (Often political pieces or the like, or slurs on various people.) Ella May Wiggins, to cite an example I already used, was murdered for her songs, and robbed and threatened before that, and her pre-teen daughter was raped, but she admitted to her compositions despite the cost to her! So truly anonymous songs are rare. But there are obviously many, many songs whose authors we don't know.

Public Domain is also straightforward: The rules differ from nation to nation, but anything more than a century old is surely public domain. But a song can have a known author and still be public domain!

That leaves "Traditional," which is the one term which has a slightly fuzzy definition. There are two parts often cited:
1. Songs which have been handed down from person to person, usually orally
2. Songs which people consider their own, so that they have the right to modify them
Note that having an unknown author is not part of either of these.

Both criteria have problems. It is my opinion that #2 in isolation is hopeless, because it describes anything anyone ever sings! We have a word for that; the word is a "song." Calling it "traditional" is utterly un-helpful. Meeting criterion #2 is perhaps a hint of something or other, but ultimately the definition has to revolve around #1.

But then what do you do about all the thousands of hymns (not spirituals) that parents learned in church and taught to their children, but the children may remember in part because of hearing it in church?

Or take "The Wreck of the Old 97" (a song which, notably, uses a Henry Clay Work tune!). It is traditional, it is now public domain, and it the author is unknown. But if you check the field collections, the majority include the line "It was on that grade that he lost his AVERAGE" rather than "It was on that grade that he lost his AIR-BRAKES." In other words, most of the informants learned it from Vernon Dalhart's recording, not from oral tradition. Are those collections traditional?

Or if you look at Vance Randolph's Ozark collection, it includes a great many songs that just happen to have been released on 78s a year or two before Randolph collected them. For example, there are a couple of dozen songs in his collection ("Amber Tresses Tied in Blue" is the first one alphabetically) that appear to be straight from the Carter Family.

And what do you do about songs that someone learned from somebody else at a jam session somewhere?

Traditional? Ugh. It's hard to figure out, sometimes. I don't have answers. On the other hand, taking the example of Beatles songs, I generally would not consider them traditional. Many of them meet criterion #2 above. But they don't meet criterion #1; people learned them from Beatles records, not other people. Which is why we have criterion #1, even if it is hard to apply.

In the Ballad Index, I follow the rule "When in doubt, index." But the criterion I am trying to follow is #1.

Incidentally, if anyone wants insight into how wrong-headed Frank Gummere was, just consider his text of "The Gest of Robyn Hode" [Child 117]. (A piece that which is public domain, with an unknown author, though very likely was not traditional -- it may not even have been sung.) It's important to understand that Child had seven texts of this, all printed (probably as chapbooks). Child had three complete copies (b, f, and g), one substantial but incomplete copy (a), and three fragments (c, d, e). We have now recovered more pages of e, so it too constituted a substantial if incomplete copy. This lets us know that there are two families of texts: a was a very bad reprint of e, which was printed by Richard Pynson; and f was copied from b, which was printed by Wynken de Worde, and g was copied from f. Put another way, f is a corrupted copy of b, and g a corrupted copy of f.

When Gummere printed his text of the "Gest," it is clear that he mostly just copied Child's text; he did not edit from scratch. He made only a handful of changes. Almost all of these were based on g, a doubly-corrupted copy of b! One can argue for either the Pynson or the de Worde tradition (Child used mostly Pynson but sometimes followed de Worde; my own edition splits the difference), but since all other texts are printed from their two, one should never use any text except a, b, or e. Yet Gummere used g!