(Am. J. of Phys., Vol. 58, No 3, March 1990, pp. 209-211.)

Freeman J. Dyson tells us of a few programs that
were tried by some of the greats of physics: Yukawa,
Born, Heisenberg, and Feynman, intended to
accomplish "radical reform in physics." Here Dyson
tells us about Feynman's "failed" program. Failed at least
in Feynman's mind, but not in Dyson's mind! I agree
with Dyson.

Dyson tells us that Feynman was on war path against
the dogmas of his day about the suitability of quantum
mechanics. He wanted to find a new starting point
to end up with "new physics."

Feynman's starting point is the following three axioms:

  mx''_j = F_j (x,x',t),                               (1)

with commutation relations

  [x_j, x_k] = 0,                                        (2)
  [x_j, x'_k] = i\hbar \delta_{jk}               (3)

(where the prime indicates differentiation with respect
to time) from which he proves Maxwell's equations. For
the details see the article.

In typical predictable style, as Dyson reports it,
the modern, recently educated physicist cannot
comprehend what the importance of this proof is
anyway. It is too long and against the modern
dogma of employing Lagrangians from the start.
Dyson explains that the importance can best be
appreciated from the historical perspective! In
this Dyson follows Einstein in maintaining that so
much gets lost when physics is presented in a
non-historical perspective.

Dyson tells us the story (210-211):

          Feynman in 1948 was not alone in trying to
     build theories outside the framework of
     conventional physics. At the time many of the
     greatest physicists, including Yukawa, Born,
     and Heisenberg, were pursuing programs for
     the radical reform of physics. All these radical
     programs, including Feynman's, failed. But
     Feynman was the only one who thoroughly
     tested his program before rushing into print.
     His proof of the Maxwell equations was a
     demonstration that his program had failed. The
     proof showed him that his assumptions (1), (2),
     and (3) were not leading to new physics. The
     road that he had been exploring had been a
     dead end. From Feynman's point of view, the
     proof was a failure, not a success. That is why
     he was not interested in publishing it.
          I venture to disagree with Feynman now, as
     I often did while he was alive. I still believe that
     his proof is worth publishing. It is not only a historic
     relic of a failed program. It also raises some new
     questions. The Maxwell equations are relativistically
     invariant, while the Newtonian assumptions (1), (2),
     and (3), which Feynman used for his proof, are
     nonrelativistic [meaning that they are Galilean
     relativistic].  The proof begins with assumptions
     invariant under Galilean transformations and ends
     up with equations invariant under Lorentz transformations.
     How could this have happened? .....

Ok, now I'll sum up.

This short article gives us a lot to think about.
Once again we see that the conventional way physics
gets done is a "program." We also see that in attempting
to diverge from that program one must institute a different
program of research. What is a program of research?
Well, it is the set of rules and constraints that have no
physical content and that are accepted without necessity
from one's knowledge of the external world. Apparently,
one of the cardinal rules that Feynman included in his
program was that he would not employ Lagrangians in
his basic set of axioms, as an example of what I mean.

We see, once again, that in building a program we
have the right to establish any consistent set of
axioms we wish to start with. There can be no "true"
set of axioms in Nature, at least we could never
prove such a thing. Theories are peculiar to the
human mind, as far as anyone can prove. It is
meaningless to speak of "truth" without having a
means to verify it.

We see, contrary to the complaints of modern-day
cranks, that many top physicists do try to break out
of the confines of establishment physics. That a
major breakout has not occurred since it was founded
in the first three decades of the 20th century is not
for lack of good physicists trying to do so. (It remains
to be seen how much of a breakout is given by David
Bohm's program or by David Hestenes's program to reform
qunatum mechanics by adopting the view of local variables.