The Physicists’ Philosophy of Physics

P. J. E. Peebles1
Joseph Henry Laboratories
Princeton University, Princeton, NJ, USA
arXiv:2401.16506v2 [[Link]-ph] 9 Aug 2024

Abstract
Research in physics operates by an implicit community philosophy. I offer
a definition I think physicists would accept, by and large, and explanations
by examples from physics and comparisons to what philosophers, sociolo-
gists, and historians as well as physicists conclude from what they observe
physicists doing.

1 Introduction
Physicists follow operating procedures that have been close enough to standard
for long enough to constitute a working philosophy. I do not have in mind the
methods of discovery of specific physical theories and the tests that establish
some as canonical physics, both of which which have evolved a lot, but rather our
canonical practice. Experience can force evolution of what is canonical, but it
has been stable long enough to merit a careful definition of the physicists’ way of
thinking: the physicists’ homegrown philosophy of physics.
Natural scientists seek empirically tested regularities that characterize what is
observed. In fundamental physics, the subject of this essay, the regularities are
described by mathematical theories that are meant to be usefully close to what is
assumed to be the foundational basis for how the physical world operates. That it
can succeed is an assumption that is implicit in what physicists do and encouraged
by experience, but seldom stated. I offer in Section 2 a statement of this assumption
and what I take to be the other three basic elements of the physicists’ philosophy
of physics. These assumptions need interpretations; it is the subject of Section 3.
Research in physics has arrived at good approximations to what is assumed to
be the fundamental nature of mind-independent reality. Section 4 is a review
evidence of this from the comparisons of what our theories predict and what is
measured or observed. It fits the assumptions in Section 2 so far. Physics has
a sociology, though physicists seldom think about it. I offer thoughts about this
along with assessments of what sociologists say about physics in Section 5. The
1E-mail: pjep@[Link]

1
mixed opinions about the future of research in fundamental physics are reviewed
in Section 6. Section 7 presents concluding remarks.
Scattered notes: The topic of this essay is the philosophy of research aimed
at advancing the physics that is the closest approximation we have to physical
reality, assuming it exists. The other physical sciences that have theories capable
of predictions that can be compared to experiments, and the natural sciences that
deal with observations, have similar philosophies though of course adapted to be
suited to the working conditions. These categories of science cannot be ranked
by intrinsic merit, but fundamental physics has gone particularly far because it
deals with particularly simple analyzable situations. I do not mean to trespass
on thinking about natural science by philosophers and sociologists, or add to the
literature on ontology by my comments about the physicists’ (usually only implicit)
concept of reality. My purpose is to present what I take to be the framework of
ideas implicit in the culture of research into fundamental physical reality. The four
starting assumptions to be stated next (in Sec. 2) are meant to be succinct statements
of this philosophy. They might be of interest to philosophers and sociologists who
judge what people are doing in science, or maybe not. But I think is useful to have
a careful statement of the philosophy of physicists as we find ourselves practicing
it.

2 Starting Assumptions
The starting idea of the natural sciences is that the world operates by rules and
regularities that can be discovered by measurements: experiments, observations,
descriptions. In fundamental physics the idea can be narrowed to four starting
assumptions.
A: The world operates by rules and the logic of their application that can be
discovered in successive approximations.
B: A useful approximation to the rules, a theory, yields predictions that are
reliably computed or described and found to agree with reliable measurements or
observations, within the uncertainties of the predictions and measurements.
C: Fundamental physics is growing more complete by advances in the variety
and precision of independent empirical measurements that check and agree with
predictions, and by occasional unifications that demote well-tested fundamental
physical theories to good approximations to still better more broadly tested theories.
D: Research in fundamental physical science is advancing toward a unified

2
mind-independent physical reality, though not one that need ever be completely
known.
In correspondence received about the draft of this essay on arXiv and else-
where, these statements have been termed “axioms,” or “beliefs,” or “faiths.” All
make sense, but to me the connotation of confidence feels too strong. I prefer
“assumptions,” or “hypotheses,” because the statements are subject to adjustment
if the empirical evidence requires it. I use the word “know” only in the informal
sense, that “it’s pretty clear.” Correspondents also have wondered why the as-
sumptions make no mention of how specific theories were discovered and came
to be accepted or rejected, and why there is scant attention to this in the following
discussion. It is because my topic is the framework, or guidelines, for curiosity-
driven scientific research implicitly defined by what we find ourselves doing that
works, with particular attention to what has happened in fundamental physics on
very small and very large scales.

3 Interpreting the Assumptions

The four assumptions in Section 2 require interpretations: judgements of intent,
policies for ambiguities, and allowances for sensible exceptions, all informed by
experience. I draw lessons about this from the history of physics since Maxwell
unified the theories of electricity and magnetism. More is to be learned from the
history of how natural science grew up to the time of Newton, but only occasional
thoughts about earlier advances are offered here.

3.1 Rules and logic

The starting thought in Assumption A in Section 2 is that the world operates by
rules we can discover. This bold idea is implicit in scientists’ expectation that their
research is adding to what is known about the world. We have no guarantee that
this is so, of course, but it is encouraged by experience.
The topic of research might be birds. There are distinct observational differ-
ences among the species of birds, with the interesting exceptions of crossbreeds,
and I am told that genetic testing by and large confirms these visual classifications,
with rare exceptions that are big news to bird watchers. It reenforces the ornithol-
ogists’ implicit assumption that they are discovering aspects of the rules by which
the world operates. The Ptolemaic system produces remarkably accurate predic-

3
tions of the angular positions of the planets. This also amounts to useful rules for
how the world operates, but the system says nothing about how an apple falls to
the ground. The Newtonian theory takes care of falling apples, the precession of
Foucault’s pendulum, the flow of water in tides and air in tornados, and so much
more: a far more unified approximation to reality. The equations Maxwell wrote
down in the mid-1800s are tested and prove to be sound by everyday experience.
That is not what you would expect in a universe that does not operate by fixed rules.
For that matter, how could life form and how could the fittest survive if Nature
kept changing what is required to be fit? All the many considerations of this sort
are commonplace evidence that Nature is operating by rules we can discover and
rely on.
The rules expressed as mathematical laws in fundamental physics, the theories,
are supposed to approximate the fundamental basis for how our world operates.
The demonstration that this is so cannot be compared to a mathematician’s proof
of a theorem. It is instead the accumulation of checks of theory by observation,
some precise, some qualitative but important, in situations that are simple enough
to allow tests. In a situation that is too complicated to apply tests we don’t know
for sure the theory is a good approximation; it is a loose end that awaits further
work. I comment on this in Section 6.
The mathematical nature of the rules was not always so important; Faraday
made great discoveries about the properties of electricity and magnetism without
the use of mathematics much beyond arithmetic. But then Maxwell added the
hypothetical displacement current to arrive at a mathematical theory that unified
two subjects, electricity and magnetism, to get the theory of electromagnetism
that has given us so many practical applications. But we cannot know the future
of fundamental physics; maybe it will involve something beyond the mathematics
now known; so it seems best to use the more generic term “rule” rather than
“theory” in Assumption A.
Quantum physics taught us a cautionary lesson about the logic mentioned
in Assumption A. An electron acts like a particle in many ways, but in standard
quantum physics an electron can be placed in a pure state of two spatially separated
wave packets, two entangled states, as in a double slit experiment. It seems logical
to insist that the electron, being a particle, really is in just one of the two wave
packets. The physicist John Stewart Bell argued for this logic, and for versions of
David Bohm’s theory of an invisible pilot wave that allows the electron to act as if
it were in a quantum entangled state of the two wave packets yet really be present
in only one of the packets. Normal logic demands it. Bell gave us a celebrated test
to distinguish this from the logic of quantum physics. In the essay, Bertlmann’s

4
socks and the nature of reality, Bell (1981) explained how to test between pilot
waves and the standard quantum physics. Bell mentioned that the physicist Alain
Aspect “is engaged in an experiment” applying the test. John Francis Clauser had
led earlier experimental tests. The results of these and many other experiments
consistently reject pilot waves and agree with quantum physics. We must adjust
our logic of how the world operates, for now at least.
Bell (1976) was troubled by the absence of a
sharply defined boundary between what is to be treated as microscopic
and what as macroscopic, and this introduces a basic vagueness into
fundamental physical theory. But this vagueness, because of the
immense difference of scale between the atomic level where quan-
tum concepts are essential and the macroscopic level where classical
concepts are adequate, is quantitatively insignificant in any situation
hitherto envisaged. So, it is quite acceptable to many people.
Bell’s distinction between microscopic and macroscopic is important but not al-
ways that pronounced. The electrons in the coherent state of a superconductor with
macroscopic dimensions, a meter around, are entangled in Cooper pairs that do not
have definite positions. The standard arguments for the reconciliation of this with
classical physics are reasonable but not totally convincing. We must fall back on
Assumption B, that the reconciliation is a useful approximation because it works.
(In a little more detail, the density matrix for a classical world can be placed in
diagonal form: Schroedinger’s cat is either alive or dead. The density matrix for
our quantum word has off-diagonal elements representing entangled states, and
predicts expectation values and correlations of observables. A macroscopic object
is entangled an immense variety of ways with observables in the system and the
external world that produce myriad phase shifts that “surely” average out to zero
in the off-diagonal elements in the macroscopic limit. Thus a molecule passing
through the double slit experiment is in a coherent state in two wave packets that
pass through the slits and then merges into the same molecule that leaves a macro-
scopic signature at the detector. This is sensible, consistent with practice, but hard
to prove.)
We cannot prove that research in fundamental physics is leading to an ever
better approximation to some definite law of physics to be discovered; we must
state it as Assumptions C and D. In the philosophy literature this amounts to the
Duhem-Quine thesis that it is impossible to prove that our canonical physics is the
only way to fit the available empirical evidence and maybe as well evidence we
have not thought to seek. This is clear enough; we cannot judge the empirical basis

5
for a theory we don’t know, and we cannot disprove its existence. The American
philosopher of science Norwood Russell Hanson (1961, p. 36) offered another
way to think of the Duhem-Quine thesis.
Given the same world, it might have been construed differently. We
might have spoken of it, thought of it, perceived it differently. Perhaps
facts are somehow moulded by the logical forms of the fact-stating
language. Perhaps these provide a ‘mould’ in terms of which the world
coagulates for us in definite ways.
We must live with this. No matter how tight and accurate the agreement of theory
and practice it remains conceivable that a different theory would do as well or even
better. We can only aim to add tests of predictions that if successful make this
unlikely thought even more unlikely, and add to the weight of evidence that we are
approaching what we term reality.
Thoughts along the lines considered here are expressed in the books Doubt
and Certainty by the physicists Tony Rothman and George Sudarshan (1998) and
Dreams of a final theory by the physicist Steven Weinberg (1992).

3.2 Predictions and falsifications

Assumption B in Section 2 express a central concept of science: a theory is to be
judged by the degree of consistency of its predictions with the empirical evidence.
If a prediction, something that was not anticipated, is observed, it is taken as
evidence that the theory might be a useful approximation to the nature of the
world. Why else the consistency? Maybe only an accident? Checking that calls
for more predictions to be tested.
Karl Raimund Popper’s contribution to the philosophy of science is informally
honored by physicists by the use of Popper’s word, “falsifiable,” as in a list of
advantages of a proposed theory that ends with “ and it is falsifiable." Popper’s
(1959, p. 10) more nuanced argument includes this:
singular statements — which we may call ‘predictions’ — are deduced
from the theory . . . those are selected which are not derivable from
the current theory, and more especially those which the current theory
contradicts. Next we seek a decision as regards these (and other)
derived statements by comparing them with the results of practical
applications and experiments . . . if the conclusions [predictions] have
been falsified, then their falsification also falsifies the theory from
which they were logically deduced.

6
The concept of empirical tests of predictions is essential to Assumptions B and C.
People can be wonderfully imaginative in constructing stories, or theories, that
account for what is thought to be appropriate. It is encouraging if a theory can be
devised to fit many known observations, but that might be only a clever contrivance.
That is why we rely on tests of predictions: implications that were not considered
in constructing the theory. The greater the number and variety of successful
predictions the better the evidence for the theory, and the more persuasive the
case that the theory is a useful approximation to this aspect of reality. (But bear
in mind that physical theories tend to have parameters that can be adjusted to
fit the data. An adjustment of a parameter within the bounds allowed by other
tests that improves the fit is encouraging but not a demanding check. The tests of
non-negotiable predictions that remain after adjustments of such free parameters
are crucial.)
Another way to put the situation is the philosophers’ “miracle argument” (Put-
nam 1975, p. 73), that

The positive argument for realism is that it is the only philosophy that
doesn’t make the success of science a miracle.

I take Putnam’s “success” to be the many confirmations of scientific predictions

in controlled experiments and observations as well as in practical applications.
These successful applications are predictions, not likely to have been imagined
by the people who devised the theories. Consider the many applications of elec-
tromagnetism and quantum physics. Maybe some successful predictions are only
accidental, and maybe others are successful because many theories were examined
and one at last found that happens to fit the measurements. But such thoughts are
difficult to maintain for the numerous successful predictions of canonical physics
reviewed in Section 4.
The American philosopher Charles Sanders Peirce made perceptive comments
about this and other aspects of physics, much of it published in the magazine
Popular Science Monthly. (The magazine was renamed Popular Science. I and
I imagine many other youths read it and some became physicists in part because
of that magazine.) The Canadian philosopher Cheryl Misak (2016), in Cam-
bridge Pragmatism From Peirce and James to Ramsey and Wittgenstein, gives an
informative account of Peirce and like thinkers.
Peirce (1878, p. 299) presented a good illustration of Assumptions C and D:

all the followers of science are fully persuaded that the processes
of investigation, if only pushed far enough, will give one certain

7
 solution to every question to which they can be applied. One man may
investigate the velocity of light by studying the transits of Venus and
the aberration of the stars; another by the oppositions of Mars and the
eclipses of Jupiter’s satellites; a third by the method of Fizeau . . . [and
another] may follow the different methods of comparing the measures
of statical and dynamical electricity. They may at first obtain different
results, but, as each perfects his method and his processes, the results
will move steadily together toward a destined centre.

The first two of the measurements Peirce mentioned are from astronomical ob-
servations that use trigonometry and Newtonian mechanics to relate the measured
radius of the Earth to distances to planets, moons, and meteorites. The next is
from laboratory measurements of the speed of light, and the last in this abbreviated
quotation is from the theories of electricity and magnetism, which scientists were
starting to realize predict the speed of electromagnetic radiation, as in light. These
are quite different ways to measure the speed of light. We can put it that one of
these measurements produces a prediction, and the others are quite different meth-
ods of testing the prediction. The consistency is notable. Why should the motion
of light coming to us from the planet Jupiter be the same as the speed of light on
Earth and the speed derived from experiments on the properties of electricity and
magnetism? The physicists’ answer is that this is a demonstration of the unity of
physics, evidence that the speed of light is a universal physical constant.
Peirce was confident that all would grow ever closer to agreement as the
measurements improved; it was Peirce’s understanding of reality. More examples
are reviewed in Section 4. They agree with the intuitive feeling of natural scientists
that their research is establishing aspects of reality.
The four assumptions in Section 2 do not include Popper’s falsifiability. To see
why consider the classical theory of electromagnetism that Maxwell put together
in the mid-1800s. By the 1930s Maxwell’s theory had passed many tests by
its applications in the laboratory and to a broad range of technology, things like
transoceanic telegraph cables and electric streetcars. The theory nevertheless was
known to fail when applied to an atom. One could say that this falsified Maxwell’s
equations, but a better way to put it is that the classical theory is a demonstrably
useful approximation that requires improvement. That proves to be quantum
electrodynamics, QED. It contains the classical theory as a limiting case,2 as a
viable theory must, and it yields new predictions that pass experimental tests,
2I take it here and in all this essay that the conventional quantum measurement theory mentioned
in footnote 2 in Sec. 3.1 is an adequate approximation.

8
some to considerable accuracy, as an established theory should. This is reviewed
in Section 4.
But QED also is an approximation, a limiting situation in the electroweak
theory that is part of the standard model of particle physics that the community
agrees calls for improvement by discovery of another step toward the unification
of science. This is Assumptions C and D.

3.3 Physical reality

Physicists replace the rich array of philosophical and cultural ideas about reality
with the operational approach in Assumption D in Section 2. Our physical world is
assumed to be mind-independent and capable of being probed to reveal the rules,
the fundamental theories, by which the world operates. It must be in successive
approximations because tests have limited accuracy and probe only parts of the
world. Hermann Bondi (1977) put it that “science is only possible because one can
say something without knowing everything.” And we hope new approximations
can be found and empirically shown to be better. Other scientists must have stated
similar thoughts, I suppose many times, but I have not found other references.
We do not know there is a physical reality to be found, of course. Peirce (1877,
pp. 11, 12) wrote about this in a Popular Science Monthly article:

Such is the method of science. Its fundamental hypothesis, restated in

more familiar language, is this: There are real things, whose characters
are entirely independent of our opinions about them; those realities
affect our senses according to regular laws, and, though our sensations
are as different as are our relations to the objects, yet, by taking
advantage of the laws of perception, we can ascertain by reasoning
how things really are . . . It may be asked how I know that there are any
realities . . . If investigation cannot be regarded as proving that there
are real things, it at least does not lead to a contrary conclusion.

This is sensible by physicists’ way of thinking. Assumption D is the expectation

that reality is to be discovered, but of course we can’t know that for sure, we
can only say that our experience so far is encouraging. Percy Bridgman (1927),
the 1949 Nobel laureate honored for research on the behavior of matter at high
pressure, offered a similar thought.

It is of course the merest truism that all our experimental knowledge

and our understanding of nature is impossible and non-existent apart

9
 from our own mental processes, so that strictly speaking no aspect
of psychology or epistemology is without pertinence . . . We shall
accept as significant our common sense judgment that there is a world
external to us, and shall limit as far as possible our inquiry to the
behaviour and interpretation of this “external” world.

Research in natural science has advanced a lot since these remarks but they remain
good expressions of implicit and sometimes explicit thinking in the natural science
community. We have no evidence against the four assumptions in Section 2, and
considerable evidence for them from the successes of science.
An interesting example of an alternative way of thinking about reality came
from Bruno Latour’s experience during two years embedded in the Salk Institute
for Biological Studies. Latour knew nothing about natural science; he was in
effect an anthropologist observing a previously unknown culture. It left him with
mixed feelings about the reality of biological molecules. In the book about this
remarkable experience Latour and Woolgar (1986) reported that (p. 183)

We are not arguing that [the biophysical molecule] somatostatin does

not exist, nor that it does not work, but that it cannot jump out of the
very network of social practice which makes possible its existence.

So is the biophysical molecule somatostatin real?

To keep it simple let us consider first atomic helium. We have an excellent
case for its reality. Helium emission lines are observed in plasma around the
sun and in the plasma around massive stars in galaxies near and far. Its effect
as a low mass density noble gas is required to account for the mass density as a
function of radius in the theory of the structure of the sun, with predictions that
are well tested by heliosismology (Aerts, Christensen-Dalsgaard, and Kurtz 2010
Chapter 1). The same low mass density is tangible in a helium-filled balloon you
hold by a string. The phase diagrams of condensed helium at low temperature in
its two isotopes illustrate remarkable aspects of quantum physics. The evidence
reviewed in Section 3.4 is that the energy levels of an isolated helium atom follow
from quantum physics (as illustrated in eq. [5]). This adds to the evidence that
helium atoms exist and quantum physics provides an excellent approximation to
their structure. Much the same evidence applies to the next most complicated atom,
lithium (eq. [6]). All of this argues that the standard physical pictures of atomic
helium and lithium are good approximations to reality. But quantum physics is not
a useful tool for the study of the therapeutic effect of lithium. From the point of
view of physicists lithium in a living body is real, along with all the other atoms

10
that are detected by nuclear magnetic resonance measurements, but analysis of
the effect lithium has on how people feel is far too complex for computation from
first principles. Physicists must leave this emergent phenomenon in more capable
hands.
Latour observed studies of the biological molecule somatostatin. In its less
massive form (fourteen amino acids) it consists of 219 atoms with 870 electrons.
A computation from first principles of the properties of this molecule in isolation
(assuming it can exist in isolation as a close to pure state), as has been done for
atomic helium and lithium, is unthinkable. It would have to be done in successive
approximations by successively improved models. The evidence reviewed in
Section 3.4 is that the properties of this molecule follow from quantum physics,
but a thorough check for somatostatin is a serious challenge. I return to this point
in Section 6.
The rules and logic mentioned in Assumption A are supposed to apply whether
or not we are present to attempt to discover them. Peirce put it that “real things
. . . are entirely independent of our opinions about them.” We have a good case
for real things. The galaxies of stars certainly look real, and the evidence is that
the older stars in our galaxy were present and evolving well before people or any
other kind of life capable of it could have been looking at them. It seems equally
sensible to accept observations of the spectra of radiation from distant galaxies
have the familiar absorption lines of the chemical elements, but redshifted, to mean
the chemical elements existed before there was life on Earth. But of course we
cannot hope to send a robot to a planet in a distant galaxy to conduct chemistry
experiments and report back. A term for this situation is “objective reality,” but
on the advice of David Hogg (NYU) I use “mind-independent reality.”
Let us pause to consider that fascinating things surely are happening on the
surfaces of the immense numbers of planets around stars in the immense numbers
of observable galaxies, things that are real in the mind of a scientist, and surely
objectively real, but we will never observe. These are examples of loose ends in
science.

3.4 The place of physics in the natural sciences

What is the place for fundamental physics among the many lines of enquiry in
the natural sciences? All seek accurate and useful accounts of Nature by research
shaped by the operating conditions, some qualitative, some informative descrip-
tions, some tightly constraining mathematical theories. Fundamental physics is
the reductionist search for a mind-independent reality defined and constrained by

11
the four assumptions in Section 2. It aims for theories that yield reliable quantita-
tive predictions that can be compared to reproducible and accurate measurements.
This seriously limits the reach of fundamental physics but it aids the approach to
what we are assuming is mind-independent reality.
What is the relation to chemistry? Baird, Scerri, and McIntyre (2006) put it
that
Chemistry does sit right next to physics, with all its lovely unifying
and foundational theory. Squinting our eyes up tight, it is possible
to see chemistry as complicated applied physics. Even in denial, we
say we are materialists, but the material world of our denial is the
foundational world of physical theory, and so chemistry—in principle
anyway—must be reducible to physics. But this has never been much
more than an article of faith.
Physicists tend to take it without question that chemistry could be derived from
standard quantum theory if only chemistry were not so complicated. We have
some justification. Molecular hydrogen is simple enough for a clean close to first
principles quantum computation of its binding energy: the work required to pull
apart a hydrogen molecule into two hydrogen atoms at rest. We see an example
of an entangled quantum state in the ground level of an ammonia molecule, where
the nitrogen atom is on both sides of the triangle of the three hydrogen atoms.
This is contrary to usual logic but a consequence of the quantum physics discussed
in Section 3.1. In the first excited level the nitrogen atom in effect oscillates
between positions on either side of the triangle. The energy released by the
transition from this inversion oscillation to the ground level was used to power the
microwave radiation in early generations of atomic clocks. Steps to the application
of physics to chemistry have gone still further with the application of advances
in computation and data storage, but Earman and Roberts (1999) put the broader
situation correctly:
The concept of a law of nature seems to us to be an important one for
understanding what physics is up to, but it is a misguided egalitari-
anism that insists that what goes for physics goes for all the sciences.

Philip Warren Anderson (2011, pp. 144, 201) put it more explicitly:
The important lessons to be drawn are two: 1) totally new physics can
emerge when systems get large enough to break the symmetries of the

12
 underlying laws; 2) by construction, if you like, those emergent prop-
erties can be completely unexpected and intellectually independent of
the underlying laws, and have no referent in them . . . I think almost all
the things worth studying are irreducibly complex [and that] requires
research which I think is as fundamental in its nature as any other.

Anderson was a good physicist with a good point to make. As Earman and Roberts
wrote, Anderson’s subject, condensed matter, is as intellectually interesting and
challenging as elementary particle physics. What is more, research in condensed
matter physics costs far less and contributes far more to the world economy. And
there are far more irreducibly complex things, from condensed matter to botany to
people, than phenomena that are useful probes of foundational reality.
Anderson’s point suggests the thought that, just as unexpected properties of
condensed matter physics such as superconductivity are emergent from quantum
physics, quantum physics might be an effective theory emergent from something
even more fundamental. There is a difference: condensed matter physicists and
botanists must take account of the atoms and electrons and electromagnetic field
that are actors in the best approximation to fundamental theory so far, while there
is little manifest evidence of actors from some deeper level. There are hints from
theory. The gauge/gravity duality Juan Maldacena (1999) conjectured has led to
thoughts such as that expressed by Nima Arkani-Hamed (2012) that

Attempts to make sense of quantum mechanics and gravity at the

smallest distance scales lead inexorably to the conclusion that space-
time is an approximate notion that must emerge from more primitive
building blocks.

Particle physicists are quite aware of the challenge of unifying our micro- and
macro-physics, and the possibility that this might lead to a deeper level of funda-
mental physics to be explored and established.
I became interested in the topic of this essay too late to discuss it with Phil
Anderson, who was a colleague at Princeton and an excellent physicist. I expect
Phil would have insisted that I have not given proper credit to the study of complex
systems. We agree on the science but differ on priorities. My essay is on the phi-
losophy of fundamental physics summarized in the four assumptions in Section 2.
The philosophy of the science of complex systems is in no way inferior to that of
fundamental physics, but it belongs in a different essay.

13
3.5 The anthropic principle
The starting assumptions in this essay do not mention mathematics, the language of
physics. That is because there are other ways to probe the world; consider Faraday’s
great discoveries of the properties of electricity and magnetism in descriptive
terms. The physics community is divided about another way to probe the universe,
the anthropic principle.
Brandon Carter (1974) introduced his thinking about this with the remark that
Copernicus taught us the very sound lesson that we must not assume
gratuitously that we occupy a privileged central position in the Uni-
verse. Unfortunately there has been a strong (not always subconscious)
tendency to extend this to a most questionable dogma to the effect that
our situation cannot be privileged in any sense. This dogma (which in
its most extreme form led to the ‘perfect cosmological principle’ on
which the steady state theory was based) is clearly untenable, as was
pointed out by Dicke . . .
Dicke (1961) drew attention to the consistency condition that the time elapsed
since the early stages of expansion of the universe had to have been long enough
to have allowed stars to produce the chemical elements we need, but not so long
that stars suitable for supporting our existence would have exhausted their supplies
of nuclear fuel and faded away. The evolution ages of the oldest stars and the
radioactive decay age of the solar system are consistent with these conditions, as
is the time elapsed to a reasonable present mean mass density in an expanding
relativistic universe. It would have been a serious problem otherwise.
Weinberg (1989) discussed another consistency condition. The cosmic mean
mass density expected from quantum physics, if represented by Einstein’s cosmo-
logical constant, Λ, is far larger than what is allowed by the standard relativistic
cosmology. Nima Arkani-Hamed (2012) put it that
This is the largest disagreement between a “back of the envelope”
estimate and reality in the history of physics—all the more disturbing
in a subject accustomed to twelve-decimal place agreements between
theory and experiment.
What is more, we have good evidence that we flourish just as the rate of expansion
of the universe is making the transition from slowing due to the attraction of
gravity to increasing due to the effect of a positive value of Λ. Why should that
be? Weinberg offered an anthropic explanation. If Λ had been positive and much

14
larger than observed the universe would have been expanding too rapidly to have
allowed the formation of the galaxy of stars within which the chemical elements we
need were produced by recycling of matter through stars. If Λ had been negative
and large the universe would have collapsed before natural evolution could have
produced observers such as us. So Weinberg invited us to imagine a statistical
ensemble of universes, a multiverse, with Λ different in different universes and
typically large, whether positive or negative, as might be expected from the large
value suggested by quantum physics. Then we would have expected to have
flourished in a universe in the multiverse with the largest absolute value of Λ
allowed by our existence. This is at least roughly what is observed.
A multiverse of universes is expected in some versions of the cosmological
inflation picture of what our universe was doing before the very early stages of
expansion. A test of sorts would be to compare the value of Λ derived from
cosmology to the range of values expected in those universes in a multiverse that
have physics capable of supporting life of a kind that would take an interest in the
value of Λ. But we do not have an adequate theory of the properties of universes
in a multiverse, or of the kinds of physical theories that allow the formation of
entities that take an interest in the value of Λ.
If research continues to fail to reconcile the value of Λ from quantum physics
with the value from cosmology we can anticipate formation of two camps. Physi-
cists on one side will insist on working even harder to avoid the anthropic argument,
risking a lot of time spent looking for an improved theory with no guarantee of
success. The other will accept the idea of a multiverse and the mixed feelings
about it. On the positive side Polchinski (2019), a respected authority, wrote that
there is no rational argument that a multiverse does not exist, or even
that it is unlikely.
The multiverse offers a way to account for the value of Λ from cosmology, but at
the cost of a troubling departure from the test of physics by its predictive power. It
is a resolution of a puzzling situation in physics with no empirical test. My faith
in human ingenuity is such that I would not be surprised if a more elegant theory
were found to resolve the problem with Λ. Maybe it’s too much to hope that the
theory would be falsifiable.
My former colleague John Archibald Wheeler liked new ideas. He encouraged
Carter’s thinking about the anthropic principle and offered his own adventurous
“participatory anthropic principle,” that we are entangled with the rest of the
universe. This is standard quantum physics, but I cannot see how it could be
relevant to Schrödinger’s cat in the quick and dead states because the entanglement

15
of the two states of the cat with all that is around us is so broad that I expect
entanglement averages out to zero, leaving a diagonal density matrix, the cat in
one of two classical states. I expect entanglement across the universe is even more
suppressed. But progress demands adventurous ideas that might aid discovery
of the fundamental reality postulated in the starting assumptions presented in
Section 2, or maybe demonstrate that the assumptions are inadequate in some
sense. It is a great adventure to explore adventurous ideas or seek other new ways
of thinking, but usually best to keep your day job.

3.6 Philosophies of physics

Physicists do not tend to see much use for philosophy. Weinberg (1992) expresses
the feeling in the chapter Against Philosophy in his book, Dreams of a Final
Theory:

Physicists do of course carry around with them a working philoso-

phy. For most of us, it is a rough-and-ready realism [but] we should
not expect [philosophy] to provide today’s scientists with any useful
guidance about how to go about their work or what they are likely to
find.

This makes sense; just let physicists get to work. But learned people, nonscientists
and scientists, can offer thoughts that help us better understand what we are doing,
maybe particularly so when disagreements force us to think about why we disagree.
I admire Ernst Mach’s (1902) book, The Science of Mechanics, for its infor-
mative discussions and elegant demonstrations of classical mechanics. Mach’s
demonstrations still serve as valuable teaching tools; I used them in my introduc-
tory physics lecture demonstrations. Mach was a good physicist and we should
pay attention to his opinions.
To Mach the science of mechanics is an economical way to state empirical
results. Mach wrote that the

atomic theory plays a part in physics similar to that of certain auxiliary

concepts in mathematics; it is a mathematical model for facilitating
the mental reproduction of facts. Although we represent vibrations
by the harmonic formula, the phenomena of cooling by exponentials,
falls by squares of times, etc., no one will fancy that vibrations in
themselves have anything to do with the circular functions, or the
motion of falling bodies with squares. It has simply been observed

16
 that the relations between the quantities investigated were similar to
certain relations obtaining between familiar mathematical functions,
and these more familiar ideas are employed as an easy means of
supplementing experience.

Mach’s negative thinking about atoms could not last. Rutherford, Boltzmann,
Einstein and others were using the model of atoms to arrive at predictions that
were encouragingly similar to observations
Mach’s auxiliary concepts still are essential parts of physics. Quantum opera-
tors on state vectors in an abstract space are used to compute wonderfully precise
and successful predictions. The dark matter of physical cosmology is not directly
detected at the time of writing, and for all we know will prove to be observable
only by the effect of its gravity. If so dark matter might remain classified as another
of Mach’s auxiliary concepts, but the web of indirect evidence from the effects of
its gravity is tight enough that dark matter has a place in standard and accepted
physics.
Mach understood what we would term the predictive unifying power of New-
tonian mechanics, as we see in these remarks:

The riddle of the tides, the connection of which with the moon had long
before been guessed, was suddenly explained [by Newton’s theory] as
due to the acceleration of the mobile masses of terrestrial water by
the moon . . . The trade-winds, the deviation of the oceanic currents
and of rivers, Foucault’s pendulum experiment, and the like, may also
be treated as examples of the laws of areas [conservation of angular
momentum].

In present-day thinking it is impressive that the compact formulation of Newtonian

physics accommodates this broad list of phenomena and more: Mach could have
mentioned the motions of the planets, their moons, and falling apples. The broad
success of Newtonian physics offers an excellent case for the assumptions in
Section 2: Nature does seem to operate by rules we can discover. But Mach saw
the situation differently. He wrote that

It is the object of science to replace, or save, experiences, by the repro-

duction and anticipation of facts in thought. Memory is handier than
experience, and often answers the same purpose. This economical
office of science, which fills its whole life, is apparent at first glance;
and with its full recognition all mysticism in science disappears.

17
Mach’s “anticipation of facts” can be read to mean “successful predictions,” but
he does not seem to have been interested in this way of putting it. Mach’s thinking
about the nature of the physics he understood so well is a mystery that has given
his philosophy of positivism a bad name.
The physicists’ philosophy summarized in the four assumptions in Section 2 is
an approximation to the philosophers’ realism. Anjan Chakravartty (2017) wrote
that
Scientific realism is a positive epistemic attitude toward the content of
our best theories and models, recommending belief in both observable
and unobservable aspects of the world described by the sciences. . . .
It is perhaps only a slight exaggeration to say that scientific realism is
characterized differently by every author who discusses it.
This is a reasonable description of physicists’ ways of thinking, though I would
rather not mention “belief.” Physicists can adjust to physical evidence. Cheryl
Misak (2016) considers Charles Sanders Peirce to be a pragmatist who starts
with this realism. Since Peirce demonstrated admirable understanding of physics
(discussed in Secs. 3.2 and 3.3) I take it that physicists have a pragmatist philosophy.
It fits the fact that empirical tests of physical theories cannot check all eventualities
to all accuracy, meaning physical theories cannot be empirically established as
mathematical theorems. We must instead rely on pragmatic judgements of how
well predictions fit evidence and what that signifies. David Hogg (2009) puts it
that the acceptance of an advance in physics is a plausibility argument. This is
accurate, but as Hogg points out many aspects of physics have become plausible
enough to inspire confidence, though never absolute belief, in the thought that we
are approaching reality.
The discovery of the uncertainty principle in quantum physics was a surprise
that still exercises philosophers and physicists (e.g. Mermin 2019), for good reason
(as discussed in Sec. 3.1). How could a particle not have a definition position? But
although quantum measurement theory moves the search for the nature of reality
farther from our common experience, and makes it more interesting, it fits the
fundamental starting assumptions in Section 2. That is, the physicists’ working
philosophy as I state it in Assumptions A to D is pragmatic, not disturbed by
quantum physics.
Niels Bohr (1925) presented a review of the many phenomena that are sug-
gestive of quantum aspects of physics. It is an excellent illustration of the rich
empirically-driven side of the invention of quantum theory. In this paper Bohr’s
early statement of his correspondence principle is that the

18
 demonstration of the asymptotic agreement between spectrum and
motion gave rise to the formulation of the “correspondence princi-
ple,” . . . [which] expresses the tendency to utilise in the systematic
development of the quantum theory every feature of the classical the-
ories in a rational transcription appropriate to the fundamental contrast
between the postulates and the classical theories.

This is an outline of a prescription for the algebra of quantum observables and the
always serious condition that the predictions of the quantum theory agree with the
classical theory in conditions where the classical version is known to be accurate.
(Bohr’s later thoughts about the complementary roles of observables that do not
commute need not be considered here.)

3.7 Thomas Kuhn

Why do scientists accept Assumption A in Section 2, that what is observed by
eye and instruments that probe the world on scales large and small operates by
rules we can discover? Thomas Kuhn’s thinking about this is interesting because
he was trained as a physicist at Harvard University but became skeptical of the
physicists’ philosophy and expressed his thoughts in the influential book, The
Structure of Scientific Revolutions (Kuhn 1962, 1970a). Evidence of its influence
includes Kaiser’s (2016) report that the book had “Cumulative sales [that] exceed
one million copies, and at least sixteen foreign-language translations have been
published,” and the sociologist Andrew Abbott’s (2016, p. 168) report of some
17,000 citations in the Web of Science in the humanities and social sciences in
publications up to 2012, a mean rate of about one citation a day. Paul Hoyningen-
Huene (1993) presents a close analysis of The Structure of Scientific Revolutions in
the book Reconstructing Scientific Revolutions. Kuhn liked the analysis; he wrote
in the forward that “I recommend it warmly,”
Kuhn’s book, hereinafter SSR (the second edition), is particularly interesting
to physicists because of Kuhn first-hand experience with physics. He studied
physics when he was an undergraduate at Harvard University, graduated in 1943,
then spent several years on war research tracking German use of metallic chaff
to avoid radar detections, and then returned to Harvard and completed a doctoral
dissertation (Kuhn 1949) on The Cohesive Energy of Monovalent Metals as a
Function of Their Atomic Quantum Defects. His advisor was Nobel laureate John
H. Van Vleck. The thesis was followed by a paper with Van Vleck as co author, and
then Kuhn’s (1950) single-author paper on An Application of the W.K.B. Method

19
to the Cohesive Energy of Monovalent Metals, published in The Physical Review,
a well-respected journal. Here is a trained physicist whose thinking is widely
influential and a good foil for conventional thinking in the physics community.
Thomas Kuhn (1977) recalled how his thinking was shaped by his attempt to
understand Aristotle’s ideas about what we now term mechanics, writing that

Aristotle had been an acute and naturalistic observer . . . How could

his characteristic talents have failed him so when applied to motion?
How could he have said about it so many apparently absurd things?
And, above all, why had his views been taken so seriously for so long a
time by so many of his successors? The more I read, the more puzzled
I became.

In an interview by Skúli Sigurdsson (1990) Kuhn recalled that

What Aristotle could be saying baffled me at first, until—and I re-

member the point vividly—I suddenly broke in and found a way to
understand it, a way which made Aristotle’s philosophy make sense.
[It] first got me onto the idea of gestalt switches and changes in concep-
tual frameworks, which was to show up in the Structure of Scientific
Revolutions in 1962.

Kuhn (1977) explained that

Position itself was a quality in Aristotle’s physics, and a body that

changed its position therefore remained the same body only in the
problematic sense that the child is the individual it becomes. In
a universe where qualities were primary, motion was necessarily a
change-of-state rather than a state.

Thus Kuhn concuded (in SSR p. 120) that

Aristotle and Galileo both saw pendulums, but they differed in their
interpretations of what they both had seen.

Kuhn’s interest in Aristotle’s way of thinking led him to a concept of physics

that differs from the physics community in ways that are worth considering. Kuhn
asked (in SSR, pp. 166, 171) two good questions:

1. Why should progress also be the apparently universal concomitant

of scientific revolutions?

20
 2. Does it really help to imagine that there is some one full, objec-
tive, true account of nature and that the proper measure of scientific
achievement is the extent to which it brings us closer to that ultimate
goal?

They are direct challenges to the assumptions in Section 2. Indeed, natural sci-
entists have not been issued a guarantee that the world acts on the basis of mind-
independent rules that can be discovered in approximations that are converging
toward ever better theory. Natural scientists rarely question these assumptions,
in part because our instructors and their instructors seldom questioned them, but
certainly also because we have been encouraged by generations of advances in
physics that continue to yield new predictions that agree with improving observa-
tions. This is what would be expected if we were approaching an ultimate goal,
our Assumption D.
How did Kuhn’s experience in physics affect his thinking about these two
questions? Galison (2016) and Kaiser (2016) report that inspections of Kuhn’s
notebooks do not offer much evidence of what Kuhn thought of his experience
as a physics undergraduate and graduate student. So it is important that we have
Kuhn’s recollections ranging from his childhood to experiences at Harvard to
life after SSR in “an edited transcript of a tape-recorded three-day discussion—
essentially an extended interview” in Athens on October 19-21, 1995, by Aristides
Baltas, Kostas Gavroglu, and Vassiliki Kindi, in the book Kuhn (2000). These
authors combine expertise in physics, history, and philosophy.
In the interview Kuhn (pp. 272, 273) recalled that, after completion of a
reduced three-year undergraduate degree in physics at Harvard, employment in
war research, and returning to Harvard,

I was finding it [physics] fairly dull, the work was not interesting . . . I
couldn’t go back and sit still for that undergraduate chicken-shit and go
on from there. So, I decided I’m going to take my degree in physics.
But it also was clear, and becoming increasingly clear, that I was not
being very much fulfilled by my graduate physics teaching.

This feeling is to be contrasted with Kuhn’s (p. 276) more positive recollection of
what was leading up to SSR:

I used to think . . . I could read texts, get into the heads of the people
who wrote them, better than anybody else in the world. I loved doing
that.

21
Kuhn expressed similar thoughts in an interview by Skúli Sigurdsson (1990):

I wondered whether a physics career was what I really wanted. I was

very conscious of the narrowing, the specialization required, and . . .
I was beginning to look for alternatives. No one of those seemed
more attractive than the rest, until all of a sudden I was asked to assist
President [James B.] Conant in teaching an experimental General
Education course on the history of science, through readings of case
histories. It sounded like a pretty good idea; it would be a good
experience, a chance to work with the President of Harvard, and also
my first exposure to history of science. So I grabbed the opportunity
and found it fascinating.

Kuhn’s story seems familiar. I have observed that some of the best theoretically-
inclined undergraduate students dislike the approximation methods used in fun-
damental physics: why don’t we just go straight to the applications of the funda-
mentals? But we can’t do that; we must rely on approximations. Maybe Kuhn
distrusted the clever methods of approximation of quantum many-body physics re-
quired by his PhD dissertation on the physics of condensed matter. The condensed
matter physicist Phil Anderson (2011, p. 38) put it that

This process of “model building,” essentially that of discarding all

but the essentials and focusing on a model simple enough to do the
job but not too hard to see all the way through, is possibly the least
understood—and often the most dangerous—of all the functions of a
theoretical physicist.

Anderson was discussing the models used in the first successful theory of super-
conductivity. Kuhn’s doctoral dissertation employed quantum model-building that
was less abstract but still employed clever, one might say adventurous, approxi-
mations to quantum physics. Kuhn might have had his experience in condensed
matter in mind in the statement in SSR (p. 179) that

normal puzzle-solving research [is] possible . . . as consequences of

the acquisition of the sort of paradigm that identifies challenging
puzzles, supplies clues to their solution, and guarantees that the truly
clever practitioner will succeed.

This is a reasonable description of research in condensed matter physics if the

paradigm Kuhn mentioned is the invention by “clever practitioners” of ways to

22
approximate what is expected from quantum physics and arrive at reliable and
useful approximations that account for observations. But the approximations in
condensed matter physics are bold and Kuhn could have wondered whether they
were only contrived to get the wanted answers. Kuhn seems not to have accepted
that the successful predictions were useful approximations to quantum physics, or
to have appreciated the power of the physicists’ treasured concept of unification.
The puzzle of superconductivity could not have been resolved within classical
Newtonian mechanics and Maxwell’s electromagnetism. Quantum physics solved
the puzzle, left the classical theories as useful limiting cases, and explained why
the light emitted from a heated gas of the atoms of a particular chemical element
is concentrated at discrete wavelengths that are quite specific to the element. This
unification, which is a central goal of science, is discussed further in Sections 4
and 5.5.
Kuhn was naturally disposed to question the physicists’ notion of reality. Kuhn
(SSR p. 206) wrote that

There is, I think, no theory-independent way to reconstruct phrases

like ‘really there’; the notion of a match between the ontology of a
theory and its “real” counterpart in nature now seems to me illusive
in principle.

Wikipedia ([Link] accessed 28 Decem-

ber 2023) informs us that, in

metaphysics, ontology is the philosophical study of being. It investi-

gates what types of entities exist, how they are grouped into categories,
and how they are related to one another on the most fundamental level
(and whether there even is a fundamental level).

Fundamental quantum physics does not seem to have a secure ontology. The quan-
tum field operators in an abstract multidimensional space are not likely candidates
for it; they are better described as examples of Mach’s auxiliary concepts. They
enable wonderfully precise agreements of predictions and measurements, but are
they “really there?” It this question meaningful? Physicists tend to ignore both
questions.
Kuhn was interested in the thinking of people doing physics, while I am
interested in Kuhn’s thinking from a physicists’ point of view. I am led to offer a
personal thought. Among my early memories is of looking into an older sister’s
schoolbook and seeing an illustration of a compound pulley (technically a block

23
and tackle). I thought that was neat and still do. Kuhn was born to write SSR. I
was born to write a close examination of SSR from a physicist’s point of view. It
is not surprising that we differ on many issues and that I align my positions to the
physics I know and love.

4 Empirical Tests of Physical Theories

The subject for this section is evidence for the four assumptions in Section 2, that
scientific research can yield persuasive evidence of useful approximations to a
unique unified fundamental physics. The central idea, which is considered in Sec-
tion 3.2, is that if a theory produces successful predictions within the uncertainties
of theory and practice it is evidence that the theory is a useful approximation to
what we take to be fundamental reality. Physicists tend to be proud of how well
this has succeeded, maybe even arrogant: Don’t bother thinking about making a
machine that violates local energy conservation. But there is a reason, the remark-
ably tight consistency of theory and practice when both can be reliably established,
which covers a considerable variety of experience. Here are examples.

4.1 Tests of quantum electrodynamics

When Thomas Kuhn was writing SSR, the foil for this essay, the tests of the
quantum theory of electrodynamics, QED, were celebrated for their precision. The
following illustrations of the state of this physics around this time is taken from the
Brodsky and Drell (1970) review, The Present Status of Quantum Electrodynamics.
The magnetic dipole moments of the electron, muon, and tau particles are
written as 𝑔 = 2(1 + 𝑎), where 𝑔 = 2 follows from Dirac’s equation and 𝑎 is the
effect of the quantum interaction of the particle with the electromagnetic field and
the other field operators of standard particle theory. Brodsky and Drell concluded
that the best estimates of the theory and observation of the quantum contribution
to the magnetic dipole moment of the electron, 𝑒 − , in 1970 was

predicted: 𝑎 − = 0.001 159 663,

measured: 𝑎 − = 0.001 159 646. (1)

For the positron Broadsky and Drell found

measured: 𝑎 + = 0.001 160. (2)

24
The theoretical values of the electron and positron moments are the same. For the
positive and negative muons, the more massive relatives of the electron, Broadsky
and Drell gave

predicted: 𝑎 𝜇 = 0.001 165 87,

measured: 𝑎 𝜇 = 0.001 166. (3)

I have slightly spoiled the precision of the theory by simplifying to the fixed value
of the fine-structure constant presented in Broadsky and Drell.
I have not found evidence that Kuhn was aware of this line of research, and if
so how it affected his thinking discussed in Section 3.7. It certainly led physicists
then as now to conclude that equations (1) to (4) with related measurements make
an excellent case that this branch of physics, QED, is a good approximation to
reality. To repeat, why else would theory and practice be so close? The position
taken in a branch of sociology is discussed in Section 5.3.
Recent measurements of the electron magnetic moment and related physical
constants are so precise that a statement of the comparison of theory and ob-
servation requires more discussion than is useful here. (Bear in mind that it is
meaningless to think of the numerical value of the speed of light 𝑐 as a universal
constant because the value depends on the choice of units. The use of a centimeter
or an inch does not matter for most of the results discussed in this section, as
long as the unit is stated. But the standard inch and centimeter are not defined
well enough for the precision measurements of the magnetic dipole moment of
the electron. Instead measurements are reduced to dimensionless quantities such
as the fine structure constant 𝑒 2 /ℏ𝑐 that do not depend on standards of length,
time, or mass. Arbitrary values of remaining free parameters enable statements in
familiar units.) I instead offer the concluding statement by Fan, Myers, Sukra, and
Gabrielse (2023):

The most precise prediction of the SM [standard model of particle

physics] agrees with the most precise determination of a property
of an elementary particle [the quantum contribution to the magnetic
dipole moment of the electron with related measurements] to about 1
part in 1012 .

The value of 𝑎 for a muon is a little different from the electron because the greater
muon mass increases the effect of the quantum interactions with the other quantum
fields. The present prediction from Aoyama et al. (2020) and measurement from

25
Aguillard et al. (2024) are

predicted: 𝑎 𝜇 = 116 591 810.(43) × 10−11 ,

measured: 𝑎 𝜇 = 116 592 059.(22) × 10−11 . (4)

The numbers in parentheses are the uncertainties in the last two digits.
Theory and measurement for the muon differ by three or four standard devia-
tions. This might be a significant difference, and if so it might be an indication
that the muon interacts with a field not yet detected and entered in the standard
model for particle physics. Or maybe it indicates a flaw of some sort in standard
quantum physics, but if so the flaw is tiny and not likely to be seriously considered
until sources of error within the parameters of standard physics, including the
possibility of a new field, have been thoroughly explored.
The measurement of the magnetic moment of the tau lepton, the still more
massive relative of the electron, is really difficult, and not yet a demanding check
of quantum physics and the standard model for particle physics.
Fan et al. (2023) state a motivation for these spectacularly accurate measure-
ments: the “quest to find physics beyond the Standard Model of Particle Physics.”
Physics is not complete and clues to how to improve it might be found in discrep-
ancies between theories and measurements of 𝑎 ± for the electron, muon, and tau
leptons. This is a common way of thinking in all branches of science. A second
motivation is the natural human impulse to aim for the best possible measurement.
A third, or maybe an unintended consequence, is the demonstration of how close
theory and practice can be in physical science.
There were and are other precision tests of QED, but this illustrates the situation.
Quantum Chromodynamics, QCD, the theory of the strong interaction, grew as
a parallel to QED; it now passes the rich variety of tests reviewed by Campbell,
Huston, and Krauss (2018) in The Black Book of Quantum Chromodynamics: A
Primer for the LHC Era. The remarkable precision and consistency of theory and
observation deeply impressed physicists in the 1960s and still does.

4.2 The quantum physics of atoms

Another illustration of the powerful empirical support for Assumptions B and C
in Sec. 2) comes from the application of quantum theory to the structures of
atoms. Computation of the structure of a helium atom its ground level requires
serious numerical computation to take account of the two electrons bound to the
more massive atomic nucleus, but important results were obtained using 1950s-era

26
computers. Kuhn could have been aware of the celebrated paper, Pekeris (1959),
which reported a precision computation of the ionization energy of atomic helium:
the energy required to pull an electron from a helium atom in its ground level. The
result compared to the measurement was

predicted: 198 310.687cm−1 ,

measured: 198 310.82 ± 0.15cm−1 . (5)

The three electrons in the next heavier element, lithium, make a precision com-
putation of its structure much more difficult, but it has been done more recently.
King (1999) reported the lithium ionization potential

predicted: 43 487.14(2)cm−1 ,
measured: 43 487.150(5)cm−1 . (6)

It is natural to push on to similar computations of properties of berylium, boron,

and on, with increasing difficulty. I discuss the importance of this research in
Section 6.
These precision tests of energy levels finesse the challenge of unifying classical
and quantum physics (the off-diagonal terms in the density matrix mentioned in
footnote 2 in Sec. 3.1) by focusing on good approximations to energy eigenstates,
where the meaning of energy agrees well enough with classical physics. This
standard physics yields the remarkably precise tests of consistency of theory and
measurement in equations (5) and (6). It is what would be expected if quantum
physics is a useful approximation to reality, and it is an important illustration of
what physicists have in mind in Assumptions B and C.

4.3 Measurements of the fundamental physical constants

The check of consistency of results from different kinds of measurements of a
physical constant, such as the speed 𝑐 of light, is a test of the theories used to
interpret the measurements and a test of the idea that 𝑐 and the other physical
constants have universal meaning. Peirce’s (1878) explanation of this point is
discussed in Section 3.2. Since then a great deal of effort has been devoted to
obtaining the best possible measurements of 𝑐 and the other fundamental constants.
They are important as standards, it is a natural impulse to make the measurements
as good as possible, and the results are important evidence of the reality of the
speed of light and the unity of physics.

27
 The physicist Raymond Thayer Birge (1929) reported three different measure-
ments of 𝑐 (in units of 1010 cm s−1 ):

𝑐 = 2.99796 from laboratory timing of reflected light pulses,

𝑐 = 2.9970 from wavelengths and frequencies of standing waves,
𝑐 = 2.9971 from the ratio of electrostatic to magnetostatic units. (7)

The treatment of measurement uncertainties was not well organized in 1929;

uncertainties in these three quantities likely are in the last digit quoted. Peirce
(1878) had mentioned results from the first and last of these methods along with
several astronomical measurements that could not be made accurate enough for
Birge’s purpose but remain important historical evidence that 𝑐 in the solar system
agrees with terrestrial measurements, as would be expected of a measure of reality.
Birge wrote that

The decision as to the most probable value, at a particular time, of any

given constant, necessarily demands a certain amount of judgment . . .

We have no fixed and final evidence in natural science. Of course Birge had to make
judgement calls: which of the measurements in equation (7) is most trustworthy?
Reputations of the authors are influential, though who knows how fair or unfair
that is in each particular case? But let us bear in mind that the close agreement
of the three numbers in equation (7), which were obtained in quite different ways
by different groups, is serious evidence that they were obtained using theories that
are close to the physics of reality.
Birge reanalyzed data from Millikan’s (1917) report of measurements of the
electric charges on water and oil drops, and concluded that the electron charge is

𝑒 = −(4.768 ± 0.005) × 10−10 ) absolute electrostatics units, (8)

with the statement that “This, the writer believes, is the most reliable value that
can be deduced from Millikan’s oil-drop work.” Millikan (1917) had reported
𝑒 = −4.774 ± 0.005, close enough to Birge’s value.
In the book, Scientific Knowledge: A Sociological Analysis, Barnes, Bloor, and
Henry (1996, pp. 33 - 45) reviewed evidence found in Millikan’s surviving note-
books of charge measurements that Millikan discarded as faulty, maybe because of
problems with experimental setups, perhaps because some measurements simply
seemed questionable. It led some to wonder whether Millikan rejected measure-
ments that indicated charges less than the theoretical minimum, the charge 𝑒 of

28
a single electron. That is not something you would expect of such a capable and
experienced physicist, and debate on this issue became known in some circles
as the “battle of the electron.” It is notoriously difficult to control unconscious
bias against unexpected results such as this even with serious attempts to control
it. Maybe Millikan, despite the great care of a good physicist, overlooked a frac-
tionally charged particle, maybe a quark if theory somehow allowed it. But all
this is unlikely because Birge had a check of consistency from a different method,
electrolysis.
Electrolysis gives the mass transported by the transport of charge that could be
accurately measured: current times time. That with molecular weights and lattice
spacings in crystals from X-ray diffraction gives a measure of 𝑒. Birge quoted
results from this approach by two groups:

𝑒 = −4.776 and 𝑒 = −4.794, (9)

in the units in equation (8). Birge placed greater trust in the first value. Both differ
from Millikan’s result by more than Millikan’s stated uncertainty, it is thought due
to an error in Millikan’s viscosity of air. But they are close. Common experience
is that actual errors tend to exceed estimates because subtle sources of uncertainty
can be difficult to find.
Birge’s values of 𝑐 (in eq. [7]) and the values of 𝑒 (in eqs. [8] and [9]) are
consistent within sensible allowance for systematic errors. Physicists take the fact
of consistency of results derived from quite different ways to probe the world to
be evidence that the world operates by rules that can be discovered. This is our
Assumption A. It is simple enough.
Richard Cohen and Jesse DuMond led a celebrated a series of papers in the
mid-1940s to the mid-1960s on the values of the fundamental constants. DuMond
and Cohen (1953) reported improved measurements

chiefly made possible by the intense development since World War

II of microwave and atomic beam techniques for the study of pro-
ton resonance in magnetic fields, the fine structure of energy levels
in hydrogen and deuterium, the magnetic moments, spin gyromag-
netic ratios, nuclear magnetic resonance frequencies, and cyclotron
frequencies of fundamental particles such as protons and electrons,
etc.

Precision measurements by applications of this variety of technology to a variety of

phenomena placed tight constraints on combinations of values of the fundamental

29
parameters, including 𝑐 and 𝑒, always provided standard physics is a sufficiently
accurate approximation to physical reality. DuMond and Cohen (1953) reported
the values

𝑐 = 299792.9 ± 0.8km sec−1 ,

𝑒 = −(4.80288 ± 0.00021) × 10−10 esu. (10)

The nominally most precise value for 𝑐 from Birge’s 1929 list in equation (7)
differs from DuMond and Cohen by one part in 105 . This is impressive; 𝑐
certainly appears to be a physically real constant. Birge’s estimate of 𝑒 (eq. 8)
from Millikan’s notebooks differs from the DuMond and Cohen value by 0.7%.
The “battle of the electron” was worthwhile; near endless checks are worthwhile;
but no flaw appears in Millikan’s experimental result. The measurements of 𝑒
were not as precise as the measurements of 𝑐, but the charge measurements yield
independent evidence of mind-independent reality.
To summarize, we have the sensible and reasonable degree of agreement of
the speed of light that Peirce (1878) reported, the speed of light and the electron
charge Birge (1929) found, the DuMond and Cohen (1953) reports, and the recent
CODATA (Committee on data of the international science council) recommended
values of the fundamental constants of physics and chemistry presented in Table 31
in Tiesinga, Mohr, Newell, and Taylor (2021). There are apparent anomalies,
notably the three or four standard deviations difference between the prediction
and measurement of the magnetic dipole moment of the muon. It is judged to be
likely significant and a hint to an addition to the particle physics model, always an
interesting possibility. But the greater point is the consistency of results from the
long tradition of precision measurements of 𝑐, 𝑒, and the other physical constants.
The physics community treasures this evidence of the unity of physical science.

4.4 Detection of the neutrino

Kuhn, our foil for physics’ conventional ways of thinking, was aware of great
developments in post-WW II particle physics. He wrote (in SSR pp. 26, 27) of

the gigantic scintillation counter designed to demonstrate the existence

of the neutrino—these pieces of special apparatus and many others
like them illustrate the immense effort and ingenuity that have been
required to bring nature and theory into closer and closer agreement.3
[Footnote 3 is a list of references to experiments, including that of

30
 Reines and Cowan to be discussed here.] That attempt to demonstrate
agreement is a second type of normal experimental work, and it is
even more obviously dependent than the first upon a paradigm. The
existence of the paradigm sets the problem to be solved; often the
paradigm theory is implicated directly in the design of apparatus able
to solve the problem. Without the Principia, for example, measure-
ments made with the Atwood machine would have meant nothing at
all.
Indeed, without the theory detection of the neutrino would have been puzzling,
and of course the experiment would be exceedingly unlikely to have been done.
But physicists had a successful theory of nuclear reactions involving the creation
and annihilation of electrons that required hypothetical particles, the neutrino and
antineutrino. The direct detection of the antineutrino in the 1950s was an important
addition to the case that the theory is a useful approximation to reality. Instead
of a precise measurement this is the demonstration of existence of a hypothetical
particle.
Wolfgang Ernst Pauli introduced the idea of neutrinos to save local conservation
of energy and momentum in nuclear reactions, and Enrico Fermi introduced the
four-fermion theory of the interaction of a neutrino with an electron, a proton, and
a neutron, as in the reactions
n → p + 𝜈¯ + e− , (11)
p + 𝜈¯ → n + e+ . (12)
The proton and neutron are n and p, e+ is a positron, e− an electron, and 𝜈¯ an
antineutrino. The first line describes the decay of a neutron into a proton with
the creation of an electron that conserves electric charge and an antineutrino that
conserves lepton number. In the second line the annihilation of an antineutrino
by a proton with the creation of a positron and the conversion of the proton to a
neutron follows by a rotation of the in and out states in the first line. It is the same
quantum theory.
Fermi’s theory with refinements had served well in experiments in nuclear
physics, but in 1950 the neutrino in these reactions was a postulate. Now dark
matter in the standard cosmology is a postulate required by a successful theory but
not detected at the time of writing. A serious difference is that the 1950 theory
predicted that neutrinos could be detected, by a difficult but feasible experiment.
The first direct detections in the 1950s (Cowan, Reines, Harrison, Kruse,
and McGuire 1956 and references therein) were of antineutrinos from a nuclear

31
reactor, where fissions of neutron-rich heavy atomic nuclei released neutrons that
decayed by equation (11) and produced the antineutrinos that were streaming out
of the reactor. A tiny fraction of them were annihilated in a detector near the
nuclear reactor by the reaction in equation (12), where the proton was in an atomic
nucleus. The 𝜈¯ annihilation would be accompanied by the creation of a positron
that would be annihilated with an electron in the detector to produce gamma rays
that could be detected. Some neutrons could be captured by an atomic nucleus
with the production of more gamma rays. The expected rate of 𝜈¯ annihilations in
the detector was computed from standard quantum physics using the laboratory
measurement of the neutron half-life in the reaction (11).
Reines and Cowan designed their experiment to detect the flux of antineutrinos
predicted by the theory. This is what Kuhn charged (in the quotation at the start
of this section). But it is good physics, a test of a prediction. If antineutrinos
had not been detected at the expected rate, within the uncertainties of the neutron
half-life and what was happening in the reactor, the falsification would have been a
serious challenge to standard physics. Why would the reactions in equations (11)
and (12) have successfully fit nuclear reaction measurements involving electrons
and positrons but failed direct neutrino detection? But the flux of antineutrinos
was detected.
The physics community welcomed the Reines and Cowan detection, though
not with the enthusiasm it merited because it was expected. The detection ranks
in importance with the precision measurements reviewed in Sections 4.1 to 4.3
because it was a confirmation of a prediction that added to the weight to the
evidence of the existence of neutrinos and the reliability of quantum theory. The
Nobel Prize Committee recognized this some 40 years later.

4.5 The general theory of relativity

Einstein’s general theory of relativity that Kuhn encountered in the 1950s was a
good example of a social construction: acceptance driven by elegance. When
Kuhn was writing SSR the scant empirical support for Einstein’s theory (detailed
in Peebles 2022a, chapter 3) could lead a skeptic to say, “but Einstein just made
up this theory.” I do not know how well Kuhn knew the state of tests of general
relativity, but we have an indication that the empirical situation was even worse
than Kuhn thought.
Kuhn wrote in SSR (p. 155) that Einstein

seems not to have anticipated that general relativity would account

32
 with precision for the well-known anomaly in the motion of Mercury’s
perihelion, and he experienced a corresponding triumph when it did
so.

The triumph was limited. In a letter to Conrad Habicht in 1907 Einstein wrote that

At the moment I am working on a relativistic analysis of the law of

gravitation by means of which I hope to explain the still unexplained
secular changes in the perihelion of Mercury [and added at the bottom
of the page] so far, however, it does not seem to be going anywhere.

(The English translation is in the Collected Papers of Albert Einstein, Vol. 5, Doc.
69.)
We see that the explanation of the anomaly was not a pure prediction: Einstein
found what he was seeking. There is nothing wrong with this, of course, except
that it reduces the significance of the agreement of the theory with the observations
of the motion of the planet Mercury as a test of general relativity.
When Kuhn was writing SSR there were three standard tests of general rela-
tivity. One, the explanation of the anomalous motion of Mercury, was important
but we see that it is a dubious test. The second, the measurement of the grav-
itational redshift, was in a confused state. The white dwarf star Sirius B was
thought to offer the best test because its mass is comparable to that of the sun
and its radius is much smaller, making the predicted gravitational redshift much
larger than on the sun, and maybe more reliablly measured. St. John (1932) listed
predicted and measured redshifts both equal to 19, in units that are not needed.
This looked good. But Greenstein, Oke, and Shipman (1971) found the predicted
value 83 ± 3, because better estimates of the radius of Sirius B are smaller, and
they found measured value 89 ± 16, partly because Sirius B was farther from its
much more luminous companion Sirius A, making the light from this companion
less distracting. The Greenstein et al. prediction and observation are consistent,
and both are four times the consistent earlier estimates. There was no skulduggery;
just honest attempts at difficult measurements and a willingness to accept what
proved to be an erroneous estimate of the radius of Sirius B, I suppose at least in
part because the computed redshift seemed to agree with the measurement. That
left one test of general relativity, the gravitational deflection of light by the mass
of the sun. The deflection was detected, measured to a few tens of a percent by
observations through the 1930s. It disagreed with the estimate from Newtonian
gravity theory, which was important, and agreed with the prediction of the general

33
theory of relativity (Trumpler 1956), which was important but slender empirical
support for general relativity.
Janssen and Renn (2022) present a careful analysis of How Einstein Found his
Field Equations. For our purpose it is sufficient to note that Einstein concluded
that he had the right theory because of the compelling elegance of his 1915 field
equations. To express the situation in modern terms, general relativity is the co-
variant classical tensor field theory that preserves local conservations of energy
and momentum with the simplest acceptable Lagrangian density, just as classical
electromagnetism is the covariant classical vector field theory that preserves charge
with the simplest acceptable Lagrangian density. This is why general relativity
is presented with classical electromagnetism in the book, The classical theory of
fields. My copy, Landau and Lifshitz (1951), is the English translation of the
Russian 1948 edition. It is the second volume in a celebrated series, the Course of
Theoretical Physics, which is a standard and respected compendium of theoretical
physics. The presence of general relativity in this series is recognition that it was
(and remains) a canonical part of theoretical physics, along with classical elec-
tromagnetism. But in the 1950s classical electromagnetism had been thoroughly
tested by experiments and an enormous variety of practical applications while gen-
eral relativity had meagre support. Yet general relativity before 1960 was standard
and accepted by the authority of respected theoretical physicists for its compelling
mathematical elegance.
General relativity now passes demanding checks of predictions on a broad
range of scales: tests of the inverse square law of gravity in the laboratory down
to lengths of about 0.1 mm (Lee, Adelberger, Cook, et al. 2020); timing tests of
electromagnetic signals on the scales of the Earth, ∼ 107 cm, and the solar system,
∼ 1013 cm, reviewed by Will (2014, 2018); tests of general relativity and the search
for gravitational waves by precision measurements of periods of binary pulsars in
the Milky Way galaxy, at distances ∼ 1022 cm (Hulse and Taylor 1975; Will 2014,
2018; Agazie, Anumarlapudi, Archibald., et al., 2023); images of the shadow of
the massive black hole in the center of the galaxy M 87 that is ∼ 1025 cm from us
(Event Horizon Telescope Collaboration 2019); detection of gravitational waves
from merging black holes at distances ∼ 1027 cm (LIGO Scientific Collaboration
and Virgo Collaboration 2016); and the cosmological tests reviewed in Peebles
(2022a) that probe the universe at close to the largest detectible scale according to
standard theory.
The cosmological tests include precision measurements of the anisotropy spec-
tra of the thermal cosmic microwave background radiation (CMB) temperature and
polarization. The measurements show the patterns produced by oscillations of the

34
interacting baryonic matter and radiation ending at redshift 𝑧 ∼ 1000, when the
primeval plasma combined and released the radiation. This physical situation is
simple enough to allow secure predictions of the effect of the matter-radiation
coupling and decoupling, computed in reliable perturbation theory. And the mi-
crowave sky observed in space and favorable sites on Earth is clear enough to
allow precise and accurate measurements. The theory and measurement of the
power spectrum of the large-scale space distribution of the galaxies are less precise
because there are far fewer galaxies than CMB photons, and the more complicated
behavior of the baryons must be considered, but the oscillation signal is clearly
detected and consistent with the signal in the CMB. This is important because the
distributions of the CMB and the matter traced by galaxies evolved in different
ways and are observed by quite different probes of the universe: the valued check
of consistency from different methods and phenomena. The numerical values of
the parameters in the standard cosmology are constrained by these measurements
and other tests, typically to a few tens of percent. This is not very precise, but the
evidence of consistencies of constraints on these parameters is important because
it is derived from what was happening at quite different stages of expansion of the
universe, ranging from the present back to the time of formation of the isotopes of
hydrogen and helium when the mean distance between nucleons was nine orders
of magnitude smaller than it is now and the temperature nine orders of magnitude
larger.
There are hints of discrepancies of theoretical constraints on parameter values
from the cosmological tests. That does not seem surprising because what behaves
like dark matter and Einstein’s cosmological constant in the present theory are
modeled in simple ways that leave room for better theories. Also to be considered
is that Einstein’s field equation for general relativity and all the rest of standard
physics might be a little different from what is found locally in the application to
the immense scales of cosmology. There is ample time for this to happen. But
easier to imagine is that the hypothetical dark matter and cosmological constant
can be more realistically modeled. The present models seem far too simple, and
there is empirical evidence to guide the search for something better.
The standard relativistic theory of the large-scale structure of the universe is
incomplete. Notably, we do not have a tested theory of what the universe was do-
ing before it was expanding. The tests we do have of macroscopic physics are far
less thorough and precise than in the microscopic limit, and what is more general
relativity with cosmology on large scales is not consistent with quantum physics
with the standard model for particle physics on small scales. The conditions under
which they have been tested so far are different enough that this has caused no

35
problem with empirical tests, except for the puzzle of the quantum and cosmo-
logical vacuum energy densities. These are among the loose ends in fundamental
physics. Some will be resolved, maybe not all. But the evidence I have mentioned
is not likely to go away; we have a convincing case that relativistic cosmology isa
good approximation to reality, which must be considered in the search for a deeper
physical theory of reality.

5 Sociology
I turn now to thoughts about physics to be found in sociology. Some of it makes
good sense to me, some not.
Kuhn wrote in SSR (p. 8) that “many of my generalizations are about the
sociology or social psychology of scientists.” Consistent with this is Kaiser’s
(2016) report that, in the abundant correspondence Kuhn received about SSR,
28% were about about psychology, sociology, or philosophy, and only 11% were
about physics or chemistry. Physicists tended to be a little offended by Kuhn’s
suggestion that psychology and sociology play a role in the establishment of our
supposedly objective physical principles, though we are people and it must be
so. The question is the degree to which nominally well-established evidence and
physical theories have been influenced by society, and how far that has confused and
frustrated the attempts to find good approximations to physical reality, assuming
there is reality. I offer examples of the interaction of science and society that seem
relevant to physics.

5.1 Elegance
Einstein offered the elegant thought that (in the English translation by Sonja
Bargmann 1954)

The supreme task of the physicist is to arrive at those universal elemen-

tary laws from which the cosmos can be built up by pure deduction.

I cannot judge how serious Einstein was. Weinberg (1992), in his book Dreams
of a Final Theory, was careful to explain that if we had the final physical theory,
Einstein’s “universal elementary laws,” it would not improve weather forecasts;
far too complicated. But a variant is attempted: deduce better approximations to
the final theory by pure thought guided by known physics, with proper attention to
elegance. This can produce good results. Einstein accomplished it in his general

36
theory of relativity. As we have seen in Section 4.5 this theory was for many years
socially accepted, which does not establish it, but it now passes demanding tests
that convincingly argue it is a useful approximation to reality. And now the physics
of general relativity figures in the concept of a next level of reality, a universe of
universes, a multiverse. Some like this fascinating conjecture, some are not so
sure.
There are other arguments for and against elegance. The relativistic theory of
the expanding universe (Sec. 4.5) passes demanding tests, which is encouraging,
but there is not a tested theory of the universe before it was expanding. The
community favorite for the very early universe is cosmological inflation, which
some argue is an elegant addition to cosmology inspired by quantum physics. It
offers to pushes back in time the question of what the universe was doing earlier
than what follows from standard physics, but it leaves open the question: what
was happening before inflation could apply? There is the concept of eternal
inflation that would extend to the indefinite future, but it cannot have an eternal
past because old universes would intrude. Einstein’s cosmological constant is
required in the standard cosmology but it is exceedingly inelegant because the
only known way to reconcile its value with standard physics is the anthropic
principle (Sec. 3.5). Community feeling about that is mixed. The hypothetical
cold dark matter of the standard cosmology is inelegant because it was introduced
ad hoc to resolve a puzzle in cosmology, and it is detected so far only by its gravity.
Maybe a consolation is that Maxwell had to introduce the hypothetical vacuum
displacement current to find a unified theory of electricity and magnetism. Both
hypotheses, displacement current and dark matter, now pass demanding tests of
predictions
The Nobel laureate Paul A. M. Dirac is celebrated for his important contribu-
tions to quantum theory, and remembered too for his thoughts about the elegance
and beauty of successful physical theories. An example is the statement that “A
physical law must possess mathematical beauty” (often cited, and Dirac is said to
have stated during a visit to the University of Moscow in 1956). Kuhn (in SSR p.
155) made the point more broadly, remarking on the influence of

arguments, rarely made entirely explicit, that appeal to the individual’s

sense of the appropriate or the aesthetic—the new theory is said to be
“neater,” “more suitable,” or “simpler” than the old.

Nobel laureate Frank Wilczek (2015) described his feeling of the elegance of
fundamental physical theory in the book, A Beautiful Question: Finding Nature’s

37
Deep Design. Nobel laureate Steven Weinberg (1992) wrote in his book, Dreams
of a Final Theory, that

in this century, as we have seen in the cases of general relativity and

the electroweak theory, the consensus in favor of physical theories
has often been reached on the basis of aesthetic judgments before the
experimental evidence for these theories became really compelling.
I see in this the remarkable power of the physicist’s sense of beauty
acting in conjunction with and sometimes even in opposition to the
weight of experimental evidence.

Nobel laureate Roger Penrose, in an interview for a documentary, declared that

I think beauty is a clear guide to truth . . . And it’s certainly the case
that if you have two alternatives where you worry about which is true,
it’s a better bet to think that the one which is more beautiful is more
likely to be true. But this is always a very subtle issue. You might find
there’s a deeper reason that you hadn’t realised before which makes the
other one actually beautiful in a deep sense that you hadn’t appreciated
before.

(This remark is from [Link] © Ard

Louis and David Malone.)
Just as elegance led the community to accept general relativity well before its
great empirical successes some argue that superstring theory, which is an elegant
extension of special relativity and the standard model for particle physics, might
similarly be elevated to established theoretical physics on Richard Dawid’s (2013)
grounds of “non-empirical confirmation,” that is, by the consensus of experienced
and respected physicists. This philosophy worked for general relativity, which
was canonical physics before it passed demanding tests. Why not the same for
superstring theory?
Thoughts about beauty are shared by many, but Penrose’s cautionary remark
is appropriate. We have no definition of beauty, and we cannot rely on beauty or
elegance to assess physics because people are adaptable: we learn to like what is
successful and might come to consider it beautiful. What is more, judgements from
elegance can be wrong. The steady-state cosmology is elegant but it is convincingly
ruled out. Most agree that Einstein’s general theory of relativity would be more
elegant without Einstein’s cosmological constant, Λ, though the evidence requires
it. (The convincing evidence for the presence of Λ was established after Weinberg

38
wrote his Dreams book.) These failures explain why Assumptions A to D in
Section 2 do not mention beauty or elegance.
But we cannot ignore the elegance of physical theory. I feel it, I expect many
colleagues do too, and I am reminded of it when I hear exclamations such as
“isn’t that neat.” This must have something to teach us about the search for
physical reality. The only interesting thought I know is that maybe evolution
by natural selection produced conditioned appreciation of what works for life in
the physical realities encountered on Earth, woven into our inheritance from the
diverse experiences of many generations of species. And maybe conditioned also
is the feeling of beauty of the instruments designed to probe reality (Ivanova and
Murphy 2023).
The failures of assessments from elegance argue against reliance on non-
empirical confirmation, but if society remains stable enough there will come a
time when theory can no longer be empirically tested and will have to be judged
by elegance, logic, and precedents: non-empirical assessments. These assessment
might be helped by a better understanding of the connection of physical sciences
to our feelings of elegance.

5.2 The bandwagon effect

Scientists, being people, are subjects to fads, what Lakatos (1978, p. 91) termed the
bandwagon effect. The distinguished astrophysicist Geoffrey Burbidge was fond
of the term. The distinguished astronomer Arthur Wolfe, who knew Burbidge
well, wrote that (Wolfe 2010)
In my opinion, Geoff’s contrarian views reflected his distrust of the
bandwagon phenomenon, which he interpreted as the uncritical be-
havior of scientists following what he regarded as the wrong-headed
views of a few elite leaders.
It happens: General relativity was a canonical part of theoretical physics because
elite physicists endorsed it. I have heard Geoff’s particular complaint that the
community acceptance of the general relativity theory of the expansion of the
universe was not empirically well supported. He was right prior to the year 2000,
when general relativity was more popular than merited by the evidence. But soon
after that advances in observations made a persuasive case that this theory is a
good approximation to reality (Peebles 2022a, chapter 6; here Sec. 4.5).
Andrew Pickering (1984, p. 7) had thoughts similar to Burbidge’s about the
bandwagon for QCD:

39
 by interpreting quarks and so on as real entities the choice of quark
models and gauge theories is made to seem unproblematic: if quarks
really are the fundamental building blocks of the world, why should
anyone want to explore alternative theories?

Pickering was observing the confusion of creation of QCD, the analog of QED
for strongly interacting particles, when there still was reason to question the new
ideas of quarks and gluons. But QCD has since become a predictive theory that
passes the abundant tests reviewed by Campbell, Huston, and Krauss (2018). The
bandwagon pointed in productive directions for QCD and relativistic cosmology.
Bandwagons, like elegance, can head in wrong directions. In the 1990s the
community opinion was that the universe is expanding in effect at escape speed,
the kinetic energy of expansion exactly balancing the magnitude of the negative
gravitational potential energy of the material content of the universe. If so the
universe would be predicted to expand into the indefinite future at a rate that is
ever slowing by the ever diminishing attraction of gravity. In general relativity
it would require that Einstein’s cosmological constant, Λ, vanishes. This has the
appeal of a helpful conjecture, that a physical principle to be discovered forces
the quantum vacuum energy density to vanish, and with it the value of Λ. But I,
Neta Bahcall at Princeton, and a few others were complaining that the evidence is
that the cosmic mean mass density is too small for this picture, a challenge to the
bandwagon favoring a negligible value of Λ. By the end of the century improved
tests forced an abrupt change of thinking: the community learned that it had to live
with the inelegant Λ (from the evidence reviewed in Peebles, Page, and Partridge
2009).
An aspect of the bandwagon effect, expectation bias, is illustrated by the varia-
tion with time of published values of the measured speed of light. Reports tended to
be consistent with other recent previous reports within measurement uncertainties,
while the similar values reported at similar times could be inconsistent with later
more accurate measurements. Henrion and Fischhoff (1986) show data illustrating
this. It is largely an effect of the peer pressure of a sociological bandwagon, the
hesitation to present results inconsistent with what already seemed to be known.
It is seen also in the history of measurements of Hubble’s constant in cosmology,
here largely the results of discoveries of systematic errors but also, inevitably, peer
pressure.
So what is the lesson? Pay attention to fads; community endorsement can be
perceptive. But be wary; the community can be wrong.

40
5.3 Sociologies of physical science
One camp in the sociology of science draws lessons from observations of what
scientists are doing. The sociologist Robert Merton (1961) discussed an impor-
tant example in the article Singletons and Multiples in Scientific Discovery: A
Chapter in the Sociology of Science. Merton discussed a common experience in
science: when an interesting idea starts to circulate there is a good chance the
idea had already been proposed, independently, and escaped wide attention, or
else will be independently proposed if news does not travel fast enough. Merton
referred to the sociologists William F. Ogburn and Dorothy S. Thomas (1922),
who published a list of 148 examples “collected from histories of astronomy,
mathematics, chemistry, physics, electricity, physiology, biology, psychology and
practical mechanical invention.”
The phrase, “the times were right,” has been dismissed as trite but it fits
this phenomenon. Multiples can result from developments of technology that
enable scientific or practical advances that more than one person or group might
independently recognize and use. But also important is a sociological effect,
that half-formed ideas tend to move as gossip through the community by casual
remarks and behavior that can be evocative but not recognized until the gossip
reaches someone, or maybe more than one person, who is prepared to act on
whatever form of the thought came through. Maybe this has something to do with
the concept of acausal synchronicity in psychoanalysis. But the simple fact is that
people communicate hints of what they are thinking and doing in many ways, by
words, gestures, attitudes, and even what is not said. Thoughts float around and
occasionally promote action.
Another camp, the sociology of scientific knowledge, or SSK, gives more
attention to the inevitable social influences on what scientists are thinking and
doing, and in the strongest versions suggest that this social influence causes physics
to be a largely social construction. Consider examples of the range of thinking.
Bloor and MacKenzie (1997) expressed a soft version of SSK:
the goal of the sociology of knowledge, in our view, is the explanation
of belief, not its evaluation . . . For the historian or sociologist studying
nineteenth-century evolutionism, for example, both Darwinism and
anti-Darwinism stand equally in need of explanation.
Bloor (1991) added
But doesn’t the strong programme [of SSK] say that knowledge is
purely social? Isn’t that what the epithet ‘strong’ means? No. The

41
 strong programme says that the social component always is present
and always constitutive of knowledge. It does not say that it is the
only component, or that it is the component that must necessarily be
located as the trigger of any and every change: it can be a background
condition.

This makes sense. I have mentioned the non-empirical confirmation of general

relativity before 1960. I began working on relativistic cosmology in the early
1960s, even though I was dismayed by the scant observational evidence for this
subject along with the scant tests of the fundamental theory, general relativity.
It helped that I was close to alone in this research; I had little competition. I
was encouraged by the tolerant attitudes of colleagues, and the interest in what
I was doing by my former thesis adviser, Professor Robert (Bob) Henry Dicke,
who I considered my professor of continuing education. This was sociology. But
also important was the realization in 1965 that we are in a near uniform sea of
microwave radiation with a spectrum that was known to be at least close to thermal.
This was something new and interesting to study, and maybe related to cosmology.
Though Kuhn rejected the strong programme of SSK I place on the more
debatable side of this philosophy Kuhn’s (1992, p. 7) argument that scientific
research seeks

the facts from which scientific conclusions should be drawn, together

with the conclusions—the new laws or theories—which should be
based upon them. These two aspects of the negotiation—the factual
and the interpretive—are carried on concurrently, the conclusions
shaping the description of facts just as the facts shape the conclusions
drawn from them. . . Such a process is clearly circular, and it becomes
very difficult to see what role experiment can have in its outcome.

This circular reasoning is an important part of the healthy search for hints from
experiments and observations that guide considerations of options for revising or
devising theories in the search for better fits to data, and it is hoped to reality. More
senior physicists usually have greater influence in assessing these considerations:
sociology at work. But Kuhn seems to have missed the key point that is illustrated
by the experiment mentioned in SSR (pp. 26–27, and discussed here in Section 4.4)
to detect the neutrino. The theory guided the design of the experiment. As Kuhn
put it the “paradigm theory is implicated directly in the design of apparatus.”
This is because the experiment was designed to test the theory. A failure of

42
detection would have been a serious falsification. This is the thinking discussed in
Section 3.2.
In the book Scientific Knowledge: A Sociological Analysis, Barnes, Bloor, and
Henry (1996) present arguments along a strong SSK line of thought that is to
physicists distinctly odd. An example begins with the statement on page 107 of
Scientific Knowledge that
the different physical constants which appear in different problem so-
lutions serve to connect them together and make the data emerging
from the use of one relevant to the use and appraisal of others. Thus,
the velocity of light [𝑐] appears as a universal constant in many ex-
emplars [paradigms, or theories] in physics . . . [𝑐 must] have a single
value, and this requirement makes the development of the different ex-
emplars in which they appear mutually interdependent and mutually
constraining.13
The superscript 13 refers to the note on page 207 in Scientific Knowledge in which
the authors ask
what is the nature of the need to keep these constants [including
𝑐] fixed: what kind of a transgression is it to refuse to keep them
fixed? . . . A strictly sociological answer to the second question is
that to accept the need to keep certain constants and values fixed is to
align one’s practice with the practice of other scientists, and nothing
more. To share physical constants is to share conventions. To propose
alternative values for the constants is to challenge the practice of other
scientists.
I trust the authors of Scientific Knowledge did not realize how insulting their
“strictly sociological answer” is to the tradition of generations of physicists who
spent so much time and effort making precision measurements of 𝑐, 𝑒, and the
other fundamental constants and carefully compiling and comparing the results
with due attention to the inevitable measurement uncertainties. Examples of his
great effort are reviewed in Sections 4.1 and 4.3. The authors could have known
about this through the celebrated series of papers by Richard Cohen and Jesse
DuMond in the late-1940s through to the mid-1960s. If there were a way to “align
one’s practice” to differ from “the practice of other scientists” these generations of
careful and capable people would have run across it. It would have been big news.
Barnes et al. (1996) drew evidence for their argument—the scientific tradition
allows adjustable results—from the measurements by Robert Andrews Millikan

43
(1917 and earlier references) of the electric charges on drops of water and oil.
It drew critics and the “battle of the electron” mentioned in Section 4.3. Bad
measurements must be rejected, as Millikan did, but it requires great care because
rejection of good data or acceptance of bad data could bias results, maybe allowing
artificial consistency or inconsistency with other measurements. Experimentalists
take great care to control the problem. In double blind medical trials the subjects
and the scientists do not know who received the placebo until completion of the
trial. The Higgs particle was detected by two groups that used the same particle
accelerator beam line but different technologies applied to different methods of
detection and analyses. Both found significant detections with consistent Higgs
particle masses, important support for a celebrated result. Peirce (1878) discussed
another important check, the comparison of values of fundamental constants de-
rived from measurements of different phenomena by different people, all of whom
had to take great care with the rejection of faulty measurements. Barnes et al.
(1996) chose an appropriate topic for sociology; the “battle of the electron” is a
fair illustration of the way scientists can behave. But the authors did not present a
clear example of bad science; compare Birge’s (1929) reduction of Millikan’s data
for the unit of electric charge (eq. [8]) to the quite similar DuMond and Cohen
(1953) result (eq. [10]).
What did the authors of Scientific Knowledge mean by their “strictly sociologi-
cal” explanation of the consistency of measurements of the fundamental constants?
Surely they did not intend to offer a sociological answer that is manifestly incon-
sistent with thoroughly explored and established physics. I guess they might have
presented their “strictly sociological answer” as an example of incorrect sociology,
but why would they do that, and if they did why would they not explain it? I do not
imagine the authors intended to offer the absurd proposal that Cohen and DuMond
manipulated the data from the far more capable post-WW II technology to secure
agreement with the values of 𝑐 and 𝑒 that Birge compiled in the late 1920s, or
the values of 𝑐 Peirce mentioned in 1878 (Sec. 3.2). Many experimental groups
contributed results to the Cohen and DuMond compilations. Try to imagine them
all keeping what they report consistent with prewar measurements, thus concealing
an anomaly that would be a celebrated discovery.
The authors of Scientific Knowledge were aware of challenges to Millikan’s
measurements of the charge of the electron, but they cannot have looked much
further into a tradition that is a deeply important part of the science that is the
subject of their sociological considerations.
I refer yet again to Thomas Kuhn’s thinking because he was trained as a
physicist and offered interesting opinions about a subject that he must have known

44
reasonably well. Norton Wise (2016) recalls Kuhn’s emphatic disapproval of SSK
in the early 1970s. Kuhn (1992, pp. 8,9) later wrote that

the most extreme form of the movement, “the strong program,” has
been widely understood as claiming that power and interest are all
there are. Nature itself, whatever that may be, has seemed to have no
part in the development of beliefs about it. . . . I am among those who
have found the claims of the strong program absurd: an example of
deconstruction gone mad.

It is easy to imagine why Kuhn objected to this version of SSK; he knew examples of
solid progress of physical science, as in the detection of neutrinos. But Kuhn held
to the thought that physicists underestimate the influence of our social conditioning
on the lessons we draw from Nature. It is illustrated by what Galison (2016, p.
58) found in Kuhn’s notebooks:

objective observation is, in an important sense, a contradiction in

terms. Any particular set of observations . . . presupposes a predispo-
sition toward a conceptual scheme of a corresponding sort: the ‘facts’
of science already contain (in a psychological, not a metaphysical,
sense) a portion of the theory from which they will ultimately be
deduced.

Kuhn was right to argue for the influence of society on the natural sciences,
including fundamental physics; it is an inevitable part of how people behave.
But the essential thing is the confrontation of theory with the empirical evidence
reviewed in Section 4 that offers abundant checks that we have quite a useful
approximation to mind-independent reality.

5.4 Knowledge gained and knowledge lost

Thomas Kuhn argued (in SSR p. 102) that Newton’s Laws are not

a limiting case of Einstein’s. For in the passage to the limit it is not

only the forms of the laws that have changed. Simultaneously we have
had to alter the fundamental structural elements of which the universe
to which they apply is composed . . . The normal-scientific tradition
that emerges from a scientific revolution is not only incompatible but
often actually incommensurable with that which has gone before.

45
The philosopher Paul Feyerabend recalled that he and Kuhn independently found
the term “incommensurable,” taken from mathematics, to be useful in their inter-
pretations of the natures of the natural sciences. (Oberheim and Hoyningen-Huene
2018 review the rich varieties of thinking about The Incommensurability of Sci-
entific Theories.) Feyerabend and his mentor Karl Popper shared an interest in
quantum measurement theory, but I have not found evidence that Feyerabend was
exposed to the application of quantum theory to real physical systems that Kuhn
experienced. Feyerabend (1970) recalled vigorous debates with Kuhn in 1960-61
when both were at the University of California in Berkeley, along with cases of
common thinking. Kuhn (1970b, p. 20) asked why physicists think they have been
accumulating ever better approximations to reality, which is our Assumption C.

Is it not possible, or perhaps even likely, that contemporary scientists

know less of what there is to know about their world than the scientists
of the eighteenth century knew of theirs? Scientific theories, it must be
remembered, attach to nature only here and there. Are the interstices
between those points of attachment perhaps now larger and more
numerous than ever before?

Feyerabend (1970, p. 219) recalled that Kuhn and he

agreed that new theories, while often better and more detailed than
their predecessors were not always rich enough to deal with all the
problems to which the predecessor had given a definite and precise
answer. The growth of knowledge or, more specifically, the replace-
ment of one comprehensive theory by another involves losses as well
as gains.

Kuhn (in SSR p. 6) put it that

the major turning points in scientific development associated with

the names of Copernicus, Newton, Lavoisier, and Einstein . . . Each
of them necessitated the community’s rejection of one time-honored
scientific theory in favor of another incompatible with it.

From the physicists’ point of view these arguments overlook the fact that
empirical knowledge does not go away with the introduction of a new physical
theory; the improved theory must take account of the empirical knowledge gained
that had been used to test the older theory. Einstein’s general theory of relativity
would not be interesting if its predictions did not agree with Newtonian gravity

46
when speeds are much less than the speed of light and gravitational potential
differences are small. These are the conditions under which the Newtonian theory
is thoroughly tested and found to be accurate in the applications to the motions of the
planets and their moons, apart from tiny relativistic corrections. Einstein certainly
understood this condition; he had the check of consistency with Newtonian physics
in his search for the field equation of the general theory of relativity. Useful
knowledge was not lost by the replacement of Newtonian gravity with this new
theory; it was enriched. For another example consider what happened when
Maxwell put aside the search for a mechanical model of the ether and put together
two laboratory-based theories, electricity and magnetism. It required introduction
of the hypothetical vacuum displacement current. The lost of the mechanical ether
was not regretted because it never produced viable predictions. The two older
theories of electricity and of magnetism were not lost; they became useful limiting
cases of the unified theory. Another revolution reduced classical electromagnetism
to a limiting case of the quantum field theory, QED. It is the subject of Volume 4
in the Landau and Lifshitz Course of Theoretical Physics. But Maxwell’s classical
theory is not forgotten; it is reviewed in beautiful detail in the second volume of
the Landau and Lifshitz series and still put to productive everyday use.
An engineer designing a transmission line for electromagnetic energy thinks
about electric and magnetic fields with definite values as functions of position and
time, an effective ontology for the purpose. A scientist designing an experiment
to test QED thinks about the quantum electromagnetic field operator, which might
be said to be another effective ontology. But the scientist must also think about
the classical magnetic and electric fields that direct the motions of particles, cause
detectors to operate, and communicate what is detected to data storage and analysis
units and from there to the scientist. The engineer might not know about QED, and
the QED scientist might not be aware of continental drift, regrettable consequences
of the breadth of research in the natural sciences. These are losses to individuals
but not to the natural science community. The demotion of the Earth from the
center of the world to an ordinary place in a universe that is evolving might be a
cultural loss to some, but the Vatican is comfortable with this change of apparent
ontology.
Science and society have lost useful knowledge. One is handwriting; others
range from how to avoid tsunamis to the proper care and use of resources from
fields and forests, the ground and the air. But I cannot think of useful knowledge
physical science has lost in the past few centuries.

47
5.5 The unity of science
Thomas Kuhn (1992, pp. 18-20) argued that

[Natural sciences] should be seen as a complex but unsystematic struc-

ture of distinct specialties or species, each responsible for a different
domain of phenomena, and each dedicated to changing current be-
liefs about their domain in ways that increase accuracy and the other
standard criteria I’ve mentioned. For that enterprise, I suggest, the
sciences, which must then be viewed as plural, can be seen to retain
a very considerable authority . . . what replaces the one big mind-
independent world about which scientists were once said to discover
the truth is the variety of niches within which the practitioners of these
various specialties practice their trade.

Kuhn’s statement does not make mention of the unity of physics that is so valued
by the physics community. Another good challenge is to explain why physicists
are so enthusiastic about unification. Physicists feel they have a good answer, but
as usual it requires discussion.
Niches in the natural sciences and their practical applications are real and
unavoidable. The broad variety of complicated situations requires specialization
of knowledge and crafts: niche science. But consider that a farmer with a large
investment in the growth of wheat has the benefit of generations of practical
experience, but at least as important is the essential help of plant scientists who
know how to breed varieties of wheat that are resistant to pathogens I used to know
as “rust.” Now plant scientists can genetically engineer wheat for resistance to
pathogens and climate changes. If you drive through wheat-growing country in
North America you will see signs in wheat fields that identify healthy varieties that
seed companies will continue to produce and sell until the inevitable evolution of
some pathogen enables it to attack this strain of wheat. Here is an illustration of the
fragmentation and unity of science. Farmers need the expertise of plant scientists
who need the expertise of molecular biophysicists who study the functions of the
large biophysical molecules that help determine genetics, with the physicists’ help
with models of large molecules based on quantum physics that could be done
better.
Physicists treasure the revolutionary advances in quantum and relativity physics
for the additions to the demonstrations of the unity of science, the successful ap-
plications of the laws of physics to ever broader ranges of phenomena. Recall
in quantum physics the demonstrations of quantum entanglement of photons sep-

48
arated by a thousand kilometers; entangled electrons in Cooper pairs that have
undefined positions within a superconducting flow a meter around; the properties
of atoms and simple molecules; the properties of atomic nuclei inside atoms;
the existence of positrons and other antiparticles; and the remarkably successful
theory of subatomic physics that has been thoroughly tested down to quarks and
gluons. Consider the precision tests of quantum physics reviewed in Section 4.1,
and the great variety of ways to establish tight constraints on values of the fun-
damental physical constants reviewed in Section 4.3. And bear in mind the many
practical applications of this new physics, from light bulbs with great efficiency
and durability to cell phones that can help find a missing hearing aid.
If science is converging to the unique mind-independent reality envisioned in
Assumption D in Section 2 then research in each niche that probes the world in its
own way will reach a useful approximation to this unique reality. Apparent incon-
sistencies among niche theories must be expected because each probe reaches a
different approximation to fit what is observed under particular conditions. Sorting
this out and checking whether the effects of mind and the limitations of theories and
experiments can be taken into account to allow demonstration of convergence of
niche theories to a unique reality is a serious challenge and wonderful opportunity
for research.
Kuhn in 1992 could have known many of these considerations when he wrote
that the sciences must “be viewed as plural.” The thought is correct—research
in science requires specialization—but incomplete: missing the unity of physics.
Why did Kuhn, an intelligent and thoughtful person, not see the unity that so fas-
cinates scientists? I must leave thoughts about this to those who better understand
human behavior.

6 How Will Physics End?

Assumption D in Section 2 is that advances of research in fundamental physics are
approaching ever better approximations to foundational reality. Popper (1959, p.
452) made two good points about this.

I see no reason to believe that the doctrine of the existence of ultimate

explanations is true, and many reasons to believe that it is false. The
more we learn about theories, or laws of nature, the less do they remind
us of Cartesian self-explanatory truisms or of essentialist definitions.
It is not truisms which science unveils. Rather, it is part of the greatness

49
 and the beauty of science that we can learn, through our own critical
investigations, that the world is utterly different from what we ever
imagined—until our imagination was fired by the refutation of our
earlier theories. There does not seem any reason to think that this
process will come to an end.

As Popper said, future advances in physics could be incremental, advances that

never reach a final theory: turtles all the way down. Or maybe a brilliant conceptual
advance will compel acceptance.
The physics we have now offers two directions for the future; I term them
the theory program and the empirical program. The latter is the route followed
in Section 4 continuing to broader confrontations of theory and practice. The
major effort in the theory program acts on the thought that our present well-tested
physical theory might be complete enough to allow physicists to fill in the blanks
well enough to arrive at the “self-explanatory truism” that Popper distrusted, but
now might be adequately supported by established physics. Let us consider first
this theory program.
Steven Weinberg (1992) wrote that

I do not mean that the final theory will be deduced from pure math-
ematics . . . It seems to me that our best hope is to identify the final
theory as one that is so rigid that it cannot be warped into some
slightly different theory without introducing logical absurdities like
infinite energies . . . The final theory may be centuries away and may
be totally different from anything we can now imagine. But suppose
for a moment that it was just around the corner. What can we guess
about this theory on the basis of what we already know? . . . The one
part of today’s physics that seems to me likely to survive unchanged
in a final theory is quantum mechanics. This is not only because
quantum mechanics is the basis of all of our present understanding of
matter and force and has passed extraordinarily stringent experimental
tests; more important is the fact that no one has been able to think of
any way to change quantum mechanics in any way that would preserve
its successes without leading to logical absurdities.

Arkani-Hamed’s (2012) thinking is that

while we may not have experimental data to tell us about physics near
the Planck scale [length scale ∼ 10−33 cm], we do have an ocean of

50
 “theoretical data” in the wonderful mathematical structures hidden in
quantum field theory and string theory. These structures beg for a
deeper explanation. The standard formulation of field theory hides
these amazing features as a direct consequence of its deference to
space-time locality. There must be a new way of thinking about
quantum field theories, in which space-time locality is not the star of
the show and these remarkable hidden structures are made manifest
. . . by removing spacetime from its primary place in our description
of standard physics, we may be in a better position to make the leap
to the next theory, where space-time finally ceases to exist.

Weinberg and Arkani-Hamed are capable physicists whose views carry weight.
There is the current problem that, although advances in research are real and
significant, they are converging in two directions: elementary particle theory on
small scales, relativistic cosmology on large scales. The two peacefully coexist,
for the most part, but there is the difference of vacuum energy densities in quantum
physics and cosmology, and the challenge of reconciling the unitary evolution of
quantum states with the allowed flow of information in general relativity. Polchin-
ski (2022) gives a readable account of the trials and tribulations in this and related
aspects of the main theory program of completion of fundamental physics.
The progress of research in M-theory, a successor to string theory that Polchin-
ski indicates might be the “full quantum theory,” leads to the thought that the final
theory will correctly predict the forms of our established physics but leave the
values of the physical constants of nature—the fine structure constant or parameter
𝛼 = 𝑒 2 /ℏ𝑐 and all the rest—to be the values that happen to obtain in our universe
in the multiverse. It would then be a task for molecular biologists to explain how
the measured values permitted formation of the observed nature of life, and for
physicists to explain whether these parameter values would obtain in a habitable
universe in the multiverse. And the awkward question would remain: is there a still
deeper theory? Might it be found by pursuing other directions of research in fun-
damental physics? Alternatively, this proposed final theory might be empirically
justified by the fact that no comparably viable alternative theory has been found
(Dawid, Hartmann, and Sprenger 2015). But how do we judge how thoroughly
“theory space” has been surveyed for other viable theories? Reflections of this
sort are in the book Why Trust a Theory? (Dardashti, Dawid, and Théault 2019).
The other program, empirical, discussed in Section 4, is the confrontation of
what is predicted to what is observed, with attention to the discovery of anomalies
that might show the way to better theory. The macroscopic side of fundamental

51
physics is less well tested than the small-scale side, with more approachable
questions whose resolution might teach us something of value. What are the
physical properties of cosmological dark matter and dark energy, and what are
their places in a better standard model of particle physics? The difference of
values of Λ in quantum and relativistic physics is a clear anomaly. The theory of
the angular distribution of the CMB (the thermal cosmic background radiation),
baryon oscillations and all, fits the precise measurements, an impressive result, but
it requires a value of the Hubble parameter that differs from the directly measured
result by 5 to 10 percent. The other parameters required to fit the CMB theory
and measurement are consistent with the independent constraints, but that could
change as the constraints improve and maybe reveal more anomalies. There also are
curious discrepancies with observations in the broadly successful computations
of galaxy formation based on the physics and initial conditions of the standard
cosmology (Peebles 2022b). These discrepancies might grow into informative
challenges. Another question: is canonical physics, Einstein’s field equation for
general relativity and all the rest, really the same in our solar system and in the
immense extrapolation to the scales of length and time of the observable universe?
An example of this last question is the search for possible changes in the value
of the dimensionless fine structure parameter 𝛼. In standard physics 𝛼 is constant.
The value derived from emission line ratios in spectra of extragalactic objects
enables the search for evidence of variation of this nominal constant over cosmic
ranges of distance and time (Murphy, Webb, and Flambaum 2007; Jiang, Pan,
Aguilar 2024). This approach has not yet yielded confirmed evidence of evolution
of 𝛼. Precision laboratory measurements yield a remarkably tight constraint on the
evolution of the local value of 𝛼 (Chiba 2011). And the laboratory measurements
of 𝛼 continue to be consistent with the value from applications of basic physics
discussed in Section 4.3 (Parker, Yu, Zhong, Estey, Müller 2018). Since current
theoretical ideas about fundamental physics do not offer the prospect of a definite
prediction of the numerical value of 𝛼, evidence that it is not a fixed number
would be big news for the theory program. This is one motivation for continuing
these lines of research, along with the natural desire to make the best possible
measurements.
The program of tests of quantum physics by its application to chemistry includes
the computations of structures of atoms discussed in Section 4.2. Even more
demanding computations by methods as close as possible to basic quantum physics,
relativistic corrections and all, applied to larger atomic numbers, might prove to be
a demanding test of quantum physics. Quantum physics gives a good account of
molecular hydrogen, and the fascinating difference of properties of this molecule

52
and hydrogen deuteride, the bound state of a hydrogen atom and its stable heavier
isotope. More massive molecules are a lot more complicated. Models such
as density functional theory, but refined to the closest feasible approximation to
quantum physics, might improve understanding of large biophysical molecules
such as the somatostatin that interested Bruno Latour (Sec. 3.3). The results
would interest molecular biologists.
The consensus in the physics community is that standard quantum physics
will pass the test of application to these complicated systems and more. Recall
Weinberg’s remark that “no one has been able to think of any way to change
quantum mechanics in any way that would preserve its successes without leading
to logical absurdities.” But we have been surprised before and might be surprised
again by empirical evidence that challenges quantum physics.
This empirical program replaces the challenge of assessing the significance
of non-empirical assessments in the theory program with the challenge of com-
putation of predictions of standard fundamental physics applied to complicated
situations to compare to measurements we might already have. This program
likely ends in exhaustion and the usual awkward question, could a greater effort
using even more powerful means of computation do even better? But something
might turn up.
The end of physics has been announced on occasion. I prefer Peirce’s later
version of his remark quoted in Section 3.2, that

all the followers of science are animated by a cheerful hope that the
processes of investigation, if only pushed far enough, will give one
certain solution to each question to which they apply it.

If Peirce’s cheerful hope continues to animate research in the natural sciences,

including the physics that has so persuasively established empirically real things
such as the speed of light, then science has a promising future with many intellectual
adventures to come.

7 Concluding Thoughts
A theme of this essay is that research into how the world functions on the deepest
level we have been able to probe has yielded results that are solid, by and large,
and have been applied in technology that has revolutionized everyday life. But
let us not forget that these fundamental results are themselves interesting, if to
your taste. It is remarkable, and worth knowing, that we can make some sense

53
of the fundamental nature of physical reality. I wrote this essay with interested
non-physicists in mind. I hope some will look into it, though they might have to
skip over lapses into technicalities, and see the elegance of fundamental physics.
Paul Feyerabend’s (2010, p. viii) opinion—he had many—was that
science should be taught as one view among many and not as the one
and only road to truth and reality.
But students should have the opportunity to see something of the empirical ev-
idence we have about the fundamental nature of the world. This good science
fascinates many non-physicists. I expect Feyerabend meant “truth and beauty”
to include such things as the pleasure of attending a concert that is to your taste,
or taking a walk in the woods. These are important, but they belong in a differ-
ent essay. And I hope we keep the difference clear to all: the one a matter of
taste and choice, the other a matter of painstaking observations and experiments
that tightly constrain and establish the theory of what certainly looks like mind-
independent reality. The distinction is not absolute—there are many differences
of opinion among physicists—but empirical evidence, what is observed, is the
essential moderator for disputes in science. I hope the same might be so in society.
It is good to question authority in physics but it can go too far. I admire
Ernst Mach’s expositions that demonstrate the broad predictive power of classical
physics, but cannot understand his lack of appreciation of predictions that he
understood so well and are such elegant demonstrations of the unity of physics.
A theory that produces successful predictions cannot be all bad (with thanks to
W. C. Fields). I admire Thomas Kuhn’s independence of mind in exploring the
undoubtedly real influence of cultural norms on the development of physics, but
I am perplexed by his lack of appreciation of the unity of physics exemplified by
the broad range of its successful predictions. Why Mach and Kuhn were skeptical
about established opinions in the physics community is a puzzle to physicists, and
maybe a topic for sociologists to examine from the skeptic and establishment sides.
Physicists are quite capable of fixed opinions that can be taken up by enough
to be termed the bandwagons discussed in Section 5.2. An example is the opinion
in the 1990s that Einstein’s cosmological constant Λ surely is negligibly small.
One reason was (and still is) that there is no natural place for it in our fundamental
theory. Where is the unification? The anthropic principle (Sec. 3.5) offers a way
out, but a more elegant theory might be found and the question then might be
why we should trust a theory that cannot be tested. The question is heard about
other theories, and will be increasingly common. But that will not be the end of
empiricism in fundamental research. It will continue in the essentially unlimited

54
task of extending the applications of fundamental physics to complicated situations
in the encroachment of fundamental physics into the other natural sciences, always
to be compared to what is observed.
Finally, let us bear in mind that fundamental physics is exploring a tiny slice of
human knowledge and experience, and a dry slice at that. The approach has been
seriously productive, but where are the floods of words needed to describe the
subtle shades of meanings of the world’s immense varieties of phenomena, from
the simpler chemical procedures to the self-awareness of spectacularly complex
human beings? All this is present in the many layers of reality that rest on
what physicists think they are empirically establishing but cannot prove or use to
explain all the wonderful complexity of the world. Explorations of such things
will continue, and research in the physics described in this essay will continue to
fascinate for its concrete reasons to think that our world rests on a fundamental
physical basis that we can explore and come to understand, in approximations.

8 Acknowledgements
I am grateful to David Hogg for the many discussions that led me to write this
essay. Discussions with Tony Rothman and his careful readings of drafts of this
essay helped me clarify what I thought I meant to say and forced me to say it more
clearly in a more sensible order. Victor Albert, George Efstathiou, Will Happer,
Peter Lindenfeld, David Mermin, and Timur Tscherbul answered my questions
and helped straightened my thinking. The appearance of an earlier version of this
essay on arXiv yielded useful advice from Bryce Cyr, Jorge Ernesto Horvath, Peter
Morgan, Boud Roukema, Angelo Vulpiani, and James Wells.
I am not complaining that I have not received financial support for this essay,
and I am grateful to Princeton University for its valuable support by providing me
an office.

References
Abbott, A, 2016, in Kuhn’s Structure of Scientific Revolutions at Fifty, eds. R. J.
Richards and L. Daston

Aerts, C., Christensen-Dalsgaard, J., and Kurz, D. W., 2010, Asteroseismology.

Dordrecht: Springer. doi:10.1007/978-1-4020-5803-5

55
Agazie G., Anumarlapudi A., Archibald A. M., Arzoumanian Z., Baker P. T., Bécsy
B., Blecha L., et al., 2023, ApJL, 951, L8. doi:10.3847/2041-8213/acdac6

Aguillard D. P., Albahri T., Allspach D., Anisenkov A., Badgley

K., Baeßler S., Bailey I., et al., 2024, arXiv, arXiv:2402.15410.
doi:10.48550/arXiv.2402.15410

Anderson, P. W., 2011, More and Different. Singapore: World Scientific Publica-
tions

Aoyama T., Asmussen N., Benayoun M., Bijnens J., Blum T., Bruno M., Caprini
I., et al., 2020, PhR, 887, 1. doi:10.1016/[Link].2020.07.006

Arkani-Hamed, N., 2012, Daedalus, 141, 53

Baird, D., Scerri, E. R., and McIntyre, L. C., eds. Philosophy of chemistry:
synthesis of a new discipline. Dordrecht: Springer

Baltas, A., Gavroglu, K., and Kindi, V. 2000, in Kuhn (2000), pp. 255-323

Bargmann, S., 1954, Ideas and Opinions: Albert Einstein. New York: Crown
Publishers

Barnes, S. B., Bloor, D, and Henry, J., 1996. Scientific Knowledge: A Sociological
Analysis. London: Athlone Press

Bell, J. S., 1976, in Quantum Mechanics, Determinism, Causality, and Particles,

ed. M. Flato et al. Dordrecht: Reidel

Bell, J. S., 1981, Journal de Physique, Journal de Physique Colloques, 1981, 42

(C2), pp.C2-41-C2-62.

Birge R. T., 1929, RvMP, 1, 1. doi:10.1103/RevModPhys.1.1

Bloor, D., 1991, Knowledge and Social Imagery, Second Edition. Chicago: The
University of Chicago Press

Bloor, D., and MacKenzie, D.,1997, Nature, 387, 544

Bohr N., 1925, Nature, 116, 845. doi:10.1038/116845a0

Bondi, H. 1977. In The Encyclopedia of Ignorance, pp. 5-8. Eds. R. Duncan and
M. Weston-Smith. New York: Pocket Books

56
Bridgman, P. W. 1927, The Logic of Modern Physics. New York: MacMillan

Brodsky S. J.,and Drell S. D., 1970, ARNPS, 20, 147.

doi:10.1146/[Link].20.120170.001051

Campbell, J., Huston, J., and Krauss, F., 2018, The Black Book of Quantum
Chromodynamics: A Primer for the LHC Era. Oxford: Oxford University Press

Carter, B., 1974, IAUS 63: Confrontation of Cosmological Theories with Obser-
vational Data, 291

Chakravartty, A., 2017, Scientific Realism, in The Stanford Encyclopedia of

Philosophy (Summer 2017 Edition), ed. Edward N. Zalta. [Link]
[Link]/archives/sum2017/entries/scientific-realism

Chiba T., 2011, PThPh, 126, 993. doi:10.1143/PTP.126.993

Cowan C. L., Reines F., Harrison F. B., Kruse H. W., and McGuire A. D., 1956,
Science, 124, 103. doi:10.1126/science.124.3212.103

Dawid, R., 2013, String Theory and the Scientific Method. Cambridge: Cambridge
University Press

Dawid, R., Hartmann, H., and Sprenger, J., 2015, The British Journal for the
Philosophy of Science, 66, 213

Dardashti, R., Dawid, R., Thébault, K., 2019, Why Trust a Theory? Epistemology
of Fundamental Physics. Cambridge: Cambridge University Press

Dicke, R. H., 1961, Nature, 192, 440. doi:10.1038/192440a0

DuMond J. W., Cohen E. R., 1953, RvMP, 25, 691.

doi:10.1103/RevModPhys.25.691

Earman, J., and Roberts, J., 1999, Synthese, Vol. 118, No. 3, 428

Event Horizon Telescope Collaboration, 2019, ApJL, 875, L1. doi:10.3847/2041-

8213/ab0ec7

Fan X., Myers T. G., Sukra B. A. D., Gabrielse G., 2023, PhRvL, 130, 071801.
doi:10.1103/PhysRevLett.130.071801

57
Feyerabend, P., 1970, in Criticism and the Growth of Knowledge: Proceedings of
the International Colloquium in the Philosophy of Science, London, 1965, 197

Feyerabend, P., 2010, Against Method. London: Verso

Galison, P., 2016, in Kuhn’s Structure of Scientific Revolutions at Fifty, eds. R. J.

Richards and L. Daston

Greenstein J. L., Oke J. B., Shipman H. L., 1971, ApJ, 169, 563.
doi:10.1086/151174

Hanson, N. R., 1961, Patterns of Discovery. Cambridge at the University Press

Henrion, M., and Fischhoff B., 1986, AmJPh, 54, 791. doi:10.1119/1.14447

Hogg, D. W., 2009, arXiv:0910.3374. doi:10.48550/arXiv.0910.3374.

Hoyningen-Huene, P., 1993, Reconstructing Scientific Revolutions: Thomas S.

Kuhn’s Philosophy of Science, translated by A. T. Levine. Chicago: University
of Chicago Press

Hulse R. A., Taylor J. H., 1975, ApJL, 195, L51. doi:10.1086/181708

Ivanova, M., and Murphy, A., 2023, The Aesthetics of Scientific Experiments.
New York: Routledge

Janssen M., Renn J., 2022, How Einstein Found his Field Equations. Cham,
Switzerland: Birkhäuser. doi:10.1007/978-3-030-97955-3

Jiang L., Pan Z., Aguilar J. N., et al., 2024, ApJ, 968, 120. doi:10.3847/1538-
4357/ad47b4

Kaiser, D. I., 2016, in Kuhn’s Structure of Scientific Revolutions at Fifty, eds. R.

J. Richards and L. Daston, pp. 71-95

King F. W., 1999, AAMOP, 40, 57. doi:10.1016/S1049-250X(08)60111-1

Kuhn T. S., 1949, PhD Thesis, Harvard

Kuhn T. S., 1950, PhRv, 79, 515. doi:10.1103/PhysRev.79.515

Kuhn, T. S., 1962, The Structure of Scientific Revolutions. Chicago: The Univer-
sity of Chicago Press

58
Kuhn, T. S., 1970a., The Structure of Scientific Revolutions, second edition,
enlarged. Chicago: The University of Chicago Press

Kuhn, T. S., 1970b, in Criticism and the Growth of Knowledge, eds I. Lakatos and
A. Musfrave. Cambridge at the Cambridge University Press

Kuhn, T. S., 1977, The Essential Tension. Chicago: The University of Chicago
Press

Kuhn, T. S., 1992, The Trouble with the Historical Philosophy of Science. Cam-
bridge Mass.: Department of the History of Science, Harvard University

Kuhn, T. S. 2000, The Road Since Structure, eds. J. Conant and J. Haugeland.
Chicago: The University of Chicago Press

Lakatos, I., 1978, The methodology of scientific research programmes: Philo-

sophical Papers, Volume 1, eds. John Worrall and Gregory Currie. Cambridge:
Cambridge University Press

Landau, L., and Lifshitz, E., 1951, The Classical Theory of Fields, translated by
M. Hamermesh. Reading Mass.: Addison Wesley

Latour, B., and Woolgar, S., 1986, Laboratory life: The Construction of Scientific
Facts. Princeton N.J.: Princeton University Press

Lee, J. G., Adelberger, E. G., Cook, T. S., Fleischer, S. M., and Hecke,l B. R.,
2020, PhRvL, 124, 101101. doi:10.1103/PhysRevLett.124.101101

LIGO Scientific Collaboration and Virgo Collaboration, 2016, PhRvL, 116,

061102. doi:10.1103/PhysRevLett.116.061102

Mach, E., 1902, The Science of Mechanics: A Critical and Historical Account of
its Development. Second revised and enlarged edition of the English translation
by T. J. McCormack. Chicago: The Open Court Publishing Company

Maldacena J., 1999, IJTP, 38, 1113. doi:10.1023/A:1026654312961

Mermin, N. D., 2019, Rep. Prog. Phys. 82, 012002

Merton, R. K.,1961, Proceedings of the American Philosophical Society, 105, No.

5, 470

59
Millikan, R. A., 1917, Philosophical Magazine, Series 6, 34, 1

Misak, C., 2016, Cambridge Pragmatism: From Peirce and James to Ramsey and
Wittgenstein. Oxford: Oxford University Press

Murphy M. T., Webb J. K., Flambaum V. V., 2007, PhRvL, 99, 239001.
doi:10.1103/PhysRevLett.99.239001

Oberheim, E., and Hoyningen-Huene, P., 2018, The Incommensurability of Sci-

entific Theories, The Stanford Encyclopedia of Philosophy (Fall 2018 Edi-
tion), ed. E. N. Zalta. [Link]
entries/incommensurability

Ogburn, W. F. and Thomas, D. S., 1922, Political Science Quarterly, 37, 83

Parker R. H., Yu C., Zhong W., Estey B., Müller H., 2018, Sci, 360, 191.
doi:10.1126/science.aap7706

Peebles, P. J. E., 2022a, The Whole Truth. Princeton: Princeton University Press

Peebles P. J. E., 2022b, AnPhy, 447, 169159. doi:10.1016/[Link].2022.169159

Peebles P. J. E., Page L. A., Partridge R. B., 2009, Finding the Big Bang. Cam-
bridge: University Press

Peirce, C. S., 1877, Popular Science Monthly 12, 1

Peirce, C. S., 1878, Popular Science Monthly 12, 286

Pekeris, C. L., 1959, PhRv, 115, 1216. doi:10.1103/PhysRev.115.1216

Pickering, A., 1984, Constructing Quarks: A Sociological History of Particle

Physics. Chicago: The University of Chicago Press

Polchinski J., 2019, in Dardashti, Dawid, and Théault (2019) pp. 339–353

Polchinski J. G., 2022, A Journey across the Landscape of Strings,

Black Holes, and the Multiverse. Cambridge MA: The MIT Press.
doi:10.7551/mitpress/12277.001.0001

Popper, K. R., 1959, The Logic of Scientific Discovery. Oxford: Oxford University
Press

60
Putnam, H. 1975, Mathematics Matter and Method. Cambridge: Cambridge Uni-
versity Press

Rothman, T. Sudarshan, E. C. G. 1998, Doubt and Certainty. Reading Mass.: Helix

Books

Sigurdsson, S., 1990, Harvard Science Review, Volume III, NO. 1 Winter, 18

St. John C. E., 1932, PASP, 44, 277. doi:10.1086/124248

Tiesinga E., Mohr P. J., Newell D. B., Taylor B. N., 2021, RvMP, 93, 025010.
doi:10.1103/RevModPhys.93.025010

Trumpler, R. J. 1956. In Jubilee of Relativity Theory. Eds. André Mercier and

Michel Kervaire. Helvetica Physica Acta, Suppl. IV, pp. 106-113

Weinberg, S., 1989, RvMP, 61, 1. doi:10.1103/RevModPhys.61.1

Weinberg, S., 1992, Dreams of a Final Theory. New York: Pantheon

Wilczek, F., 2015, A Beautiful Question: Finding Nature’s Deep Design. New
York: Penguin Press

Will, C. M., 2014, LRR, 17, 4. doi:10.12942/lrr-2014-4

Will, C. M., 2018, Theory and Experiment in Gravitational Physics. Cambridge:

Cambridge University Press

Wise, M. N., 2016, in Kuhn’s Structure of Scientific Revolutions at Fifty, eds. R.

J. Richards and L. Daston, 31

Wolfe A. M., 2010, PASP, 122, 1255. doi:10.1086/657432

61