The axiom and corollary appear to have originated at a conference held in 1984 and are named after Stuart Ross Taylor, at the time affiliated with the Research School of Earth Sciences, Australian National University, Canberra.
Stuart Ross Taylor, "The Origin of the Moon: Geochemical Considerations", In W.R. Hartmann, R.J. William, and G. J Taylor (eds.), Proceedings of the Conference Origin of the moon, Kona, HI, October 13-16, 1984. Houston: Lunar and Planetary Institute 1986, pp. 125-143 (scan online). At the end of the paper one finds:
Hypotheses in science should be testable. A test to distinguish between the two viable hypotheses of lunar origin, large impactor and double planet, can be made in principle by dating the volatile depletion event.
This statement may have triggered some discussion on the issue of testability, because a prefatory page in the conference record that I found described in Robert Malcuit, Geoforming Mars: How could nature have made Mars more like Earth?, Springer 2021, p. 142, supposedly reads as follows:
The Page Before the PREFACE (Unlabeled Page VI) Has a Few Interesting Comments by at Least Two of the Participants (in Approximately the Same Format as the Page in the Book)
TAYLORS'S AXIOM
The best models for lunar origin are the testable ones.
TAYLOR's COROLLARY
The testable models for lunar origin are wrong. — S. Ross Taylor paraphrased by Sean Solomon, at the Conference on the Origin of the Moon, Kona, 1984.
So far I have been unable to find a scan of the relevant page of the conference proceedings online to confirm this description, however I found a blog post "The Genesis of the Moon" that provides matching information. Google provides merely snippet views (1, 2, 3) of page VI, just enough to confirm that this page holds the relevant information.