OK, next point.

Jim quoted Phillips Barry:
"Popular tradition, however, does not mean popular origin. In the case of our ballad, the underlying folklore is Irish de facto, but not de-jure: the ballad is of Oriental and literary origin, and has sunk to the level of the folk which has the keeping of folklore. To put it in a single phrase, memory not invention is the function of the folk".   

Jim then comments "Everything I have experienced in working with singers has confirmed this to be offensively inaccurate."

Blame Phillips Barry for a false claim but please don't attribute the same to anyone here. Steve Roud makes no such categorical claim in his book, nor has anyone made it in this thread. We agree with MacColl's summary that Jim has quoted: all sorts of people have made songs.

The collectors of a hundred-odd years ago were selective in their collecting, because they were specifically seeking certain attributes in the songs and mostly had little or no interest in the singers except as sources of the songs that they wanted. Collectors in recent years have taken much more interest in the singers, and that has generally meant collecting whatever the singers chose to offer. In both eras the singers themselves have been selective, both in which songs they have chosen to learn and in which of those they have offered to the collectors.

I don't think Jim has answered my question from yesterday, so I will repeat it.
Jim, you claim that Roud has departed from a previously widely accepted definition of folk song, but exactly what definition have you in mind, who used it and where did they state it? You can't mean the "1954" definition because that one is not restricted by origin.