My understanding of Van Ronk's repetition of John Hurt's story is that MJH played his song exactly the way he would normally play it, that it was too long to "fit" on a single record, and that therefore the recording engineer "sped it up" (i.e., slowed the recording process so that fewer inches of "groove" would be used, causing playback to be quicker ~ and higher-pitched as well).

Whichever song was affected ~ I haven't gone back to Dave's book to look it up ~ I seriously doubt the change in speed/pitch could have been anywhere nearly as high as 20%. A 5% to 10% speed-up would make a complicated arrangement quite a bit harder to play/imitate, while not making the pitch so absurdly high that the recording technician's intervention would be obvious.

Now, keep in mind that I'm discussing what happened with a John Hurt recording, not any of Mr. Johnson's, but the analogy should be pertinent.