Let's take an example here. A Laws Ballad. Pretty high traditional credibility, surely, right?

I'm thinking of Zeb Tourney's Girl [Laws E18]

There are at least nine different traditional collections.

Good traditional song, right?

Except... it was copyrighted in 1925 by Marjorie Lamkin and Maggie Andrews (=Carson J. Robison). It was recorded in 1926 by Vernon Dalhart. A stemmatic analysis affirms that most of the field collections were directly learned from the Dalhart recording.

So was the song traditional because Laws did not do enough research to discover the author?

We can offer several other examples from Dalhart recordings -- e.g. Floyd Collins [Laws G22] is by Andy Jenkins.

Can we de-traditionalize a song when the author is discovered? Are all "traditional" songs simply Schroedinger's Ballads?

I can't see how ignorance of collectors should affect the traditionality of a song in any way.