It is undeniable that mathematicians, with a self-denying ordinance about coefficients, can thus attain remarkable criteria, and are able to anticipate definite results; but we need not seek to engraft their modes of expression on the real world of physics.

Geometrization of Physics, and Its Supposed Bias on the, Michelson-Morley Experiment, Nature, Volume 106, Number 2677, February 17, 1921 (p. 798)