The Evolution of the Physicist’s Picture of Nature

Paul Dirac

The axiomatic basis of theoretical physics cannot be extracted from experience but must be freely
invented - A. Einstein

Article by Paul Dirac from the May 1963 issue of Scien- What appears to our consciousness is really a three-
tific American.? ? dimensional section of the four-dimensional picture. We
must take a three-dimensional section to give us what
In this article I should like to discuss the development appears to our consciousness at one time; at a later time
of general physical theory: how it developed in the past we shall have a different three-dimensional section. The
and how one may expect it to develop in the future. One task of the physicist consists largely of relating events in
can look on this continual development as a process of one of these sections to events in another section referring
evolution, a process that has been going on for several to a later time. Thus the picture with fourdimensional
centuries. symmetry does not give us the whole situation. This
becomes particularly important when one takes into ac-
The first main step in this process of evolution was count the developments that have been brought about
brought about by Newton. Before Newton, people looked by quantum theory. Quantum theory has taught us that
on the world as being essentially two-dimensional-the two we have to take the process of observation into account,
dimensions in which one can walk about-and the up-and- and observations usually require us to bring in the three-
down dimension seemed to be something essentially dif- dimensional sections of the four-dimensional picture of
ferent. Newton showed how one can look on the up-and- the universe.
down direction as being symmetrical with the other two
directions, by bringing in gravitational forces and show- The special theory of relativity, which Einstein in-
ing how they take their place in physical theory. One troduced, requires us to put all the laws of physics
can say that Newton enabled us to pass from a picture into a form that displays four-dimensional symmetry.
with two-dimensional symmetry to a picture with three- But when we use these laws to get results about ob-
dimensional symmetry. servations, we have to bring in something additional
to the four-dimensional symmetry, namely the three-
Einstein made another step in the same direction, dimensional sections that describe our consciousness of
showing how one can pass from a picture with three- the universe at a certain time.
dimensional symmetry to a picture with fourdimensional Einstein made another most important contribution
symmetry. Einstein brought in time and showed how to the development of our physical picture: he put for-
it plays a role that is in many ways symmetrical with ward the general theory of relativity, which requires us
the three space dimensions. However, this symmetry is to suppose that the space of physics is curved. Before
not quite perfect. With Einstein?s picture one is led to this physicists had always worked with a flat space, the
think of the world from a four-dimensional point of view, three-dimensional flat space of Newton which was then
but the four dimensions are not completely symmetrical. extended to the fourdimensional flat space of special rela-
There are some directions in the four-dimensional picture tivity. General relativity made a really important contri-
that are different from others: directions that are called bution to the evolution of our physical picture by requir-
null directions, along which a ray of light can move; hence ing us to go over to curved space. The general require-
the four-dimensional picture is not completely symmet- ments of this theory mean that all the laws of physics
rical. Still, there is a great deal of symmetry among the can be formulated in curved four-dimensional space, and
four dimensions. The only lack of symmetry, so far as that they show symmetry among the four dimensions.
concerns the equations of physics, is in the appearance But again, when we want to bring in observations, as
of a minus sign in the equations with respect to the time we must if we look at things from the point of view of
dimension as compared with the three space dimensions quantum theory, we have to refer to a section of this
[see top equation in diagram]. four-dimensional space. With the four-dimensional space
curved, any section that we make in it also has to be
We have, then, the development from the three- curved, because in general we cannot give a meaning to
dimensional picture of the world to the four-dimensional a flat section in a curved space. This leads us to a pic-
picture. The reader will probably not be happy with ture in which we have to take curved threedimensional
this situation, because the world still appears three- sections in the curved fourdimensional space and discuss
dimensional to his consciousness. How can one bring observations in these sections.
this appearance into the four-dimensional picture that During the past few years people have been trying to
Einstein requires the physicist to have? apply quantum ideas to gravitation as well as to the other
 2

phenomena of physics, and this has led to a rather un- different points of view. Heisenberg worked keeping close
expected development, namely that when one looks at to the experimental evidence about spectra that was be-
gravitational theory from the point of view of the sec- ing amassed at that time, and he found out how the ex-
tions, one finds that there are some degrees of freedom perimental information could be fitted into a scheme that
that drop out of the theory. The gravitational field is a is now known as matrix mechanics. All the experimental
tensor field with 10 components. One finds that six of data of spectroscopy fitted beautifully into the scheme of
the components are adequate for describing everything of matrix mechanics, and this led to quite a different picture
physical importance and the other four can be dropped of the atomic world. Schrodinger worked from a more
out of the equations. One cannot, however, pick out the mathematical point of view, trying to find a beautiful
six important components from the complete set of 10 in theory for describing atomic events, and was helped by
any way that does not destroy the four-dimensional sym- De Broglie?s ideas of waves associated with particles. He
metry. Thus if one insists on preserving four-dimensional was able to extend De Broglie?s ideas and to get a very
symmetry in the equations, one cannot adapt the theory beautiful equation, known as Schrodinger?s wave equa-
of gravitation to a discussion of measurements in the way tion, for describing atomic processes. Schrodinger got
quantum theory requires without being forced to a more this equation by pure thought, looking for some beautiful
complicated description than is needed bv the physical generalization of De Broglie?s ideas, and not by keeping
situation. This result has led me to doubt how funda- close to the experimental development of the subject in
mental the four-dimensional requirement in physics is. the way Heisenberg did.
A few decades ago it seemed quite certain that one had I might tell you the story I heard from Schrodinger
to express the whole of physics in fourdimensional form. of how, when he first got the idea for this equation, he
But now it seems that four-dimensional symmetry is not immediately applied it to the behavior of the electron in
of such overriding importance, since the description of the hydrogen atom, and then he got results that did not
nature sometimes gets simplified when one departs from agree with experiment. The disagreement arose because
it. at that time it was not known that the electron has a
spin. That, of course, was a great disappointment to
Now I should like to proceed to the developments that Schrodinger, and it caused him to abandon the work for
have been brought about by quantum theory. Quantum some months. Then he noticed that if he applied the the-
theory is the discussion of very small things, and it has ory in a more approximate way, not taking into ac count
formed the main subject of physics for the past 60 years. the refinements required by relativity, to this rough ap-
During this period physicists have been amassing quite proximation his work was in agreement with observation.
a lot of experimental information and developing a the- He published his first paper with only this rough approx-
ory to correspond to it, and this combination of theory imation, and in that way Schrodinger?s wave equation
and experiment has led to important developments in the was presented to the world. Afterward, of course, when
physicist?s picture of the world. people found out how to take into account correctly the
The quantum first made its appearance when Planck spin of the electron, the discrepancy between the results
discovered the need to suppose that the energy of elec- of applying Schrodinger?s relativistic equation and the
tromagnetic waves can exist only in multiples of a certain experiments was completely cleared up.
unit, depending on the frequency of the waves, in order I think there is a moral to this story, namely that it
to explain the law of black-body radiation. Then Ein- is more important to have beauty in one?s equations
stein discovered the same unit of energy occurring in the than to have them fit experiment. If Schrodinger had
photoelectric effect. In this early work on quantum the- been more confident of his work, he could have published
ory one simply had to accept the unit of energy without it some months earlier, and he could have published a
being able to incorporate it into a physical picture. more accurate equation. That equation is now known as
the Klein-Gordon equation, although it was really dis-
The first new picture that appeared was Bohr?s picture covered by Schrodinger, and in fact was discovered by
of the atom. It was a picture in which we had electrons Schrodinger before he discovered his nonrelativistic treat-
moving about in certain well-defined orbits and occasion- ment of the hydrogen atom. It seems that if one is work-
ally making a jump from one orbit to another. We could ing from the point of view of getting beauty in one?s
not picture how the jump took place. We just had to equations, and if one has really a sound insight, one is
accept it as a kind of discontinuity. Bohr?s picture of the on a sure line of progress. If there is not complete agree-
atom worked only for special examples, essentially when ment between the results of one?s work and experiment,
there was only one electron that was of importance for one should not allow oneself to be too discouraged, be-
the problem under consideration. Thus the picture was cause the discrepancy may well be due to minor features
an incomplete and primitive one. that are not properly taken into account and that will
The big advance in the quantum theory came in 1925, get cleared up with further developments of the theory.
with the discovery of quantum mechanics. This advance That is how quantum mechanics was discovered. It
was brought about independently by two men, Heisen- led to a drastic change in the physicist?s picture of the
berg first and Schrodinger soon afterward, working from world, perhaps the biggest that has yet taken place. This
 3

change comes from our having to give up the determin- ory? These Class One difficulties do not really worry the
istic picture we had always taken for granted. We are physicist. If the physicist knows how to calculate results
led to a theory that does not predict with certainty what and compare them with experiment, he is quite happy
is going to happen in the future but gives us informa- if the results agree with his experiments, and that is all
tion only about the probability of occurrence of various he needs. It is only the philosopher, wanting to have a
events. This giving up of determinacy has been a very satisfying description of nature, who is bothered by Class
controversial subject, and some people do not like it at One difficulties.
all. Einstein in particular never liked it. There are, in addition to the Class One difficulties, the
Although Einstein was one of the great contributors to Class Two difficulties, which stem from the fact that the
the development of quantum mechanics, he still was al- present laws of quantum theory are not always adequate
ways rather hostile to the form that quantum mechanics to give any results. If one pushes the laws to extreme
evolved into during his lifetime and that it still retains. conditions?to phenomena involving very high energies or
very small distances?one sometimes gets results that are
The hostility some people have to the giving up of the ambiguous or not really sensible at all. Then it is clear
deterministic picture can be centered on a much discussed that one has reached the limits of application of the the-
paper by Einstein, Podolsky and Rosen dealing with the ory and that some further development is needed. The
difficulty one has in forming a consistent picture that still Class Two difficulties are important even for the physi-
gives results according to the rules of quantum mechan- cist, because they put a limitation on how far he can use
ics. The rules of quantum mechanics are quite definite. the rules of quantum theory to get results comparable
People know how to calculate results and how to com- with experiment.
pare the results of their calculations with experiment. I should like to say a little more about the Class One
Everyone is agreed on the formalism. It works so well difficulties. I feel that one should not be bothered with
that nobody can afford to disagree with it. But still the them too much, because they are difficulties that refer
picture that we are to set up behind this formalism is a to the present stage in the development of our physical
subject of controversy. picture and are almost certain to change with future de-
I should like to suggest that one not worry too much velopment. There is one strong reason, I think, why one
about this controversy. I feel very strongly that the stage can be quite confident that these difficulties will change.
physics has reached at the present day is not the final There are some fundamental constants in nature: the
stage. It is just one stage in the evolution of our picture charge on the electron (designated e), Planck?s constant
of nature, and we should expect this process of evolution divided by 2 pi (designated h-bar) and the velocity of
to continue in the future, as biological evolution continues light (c). From these fundamental constants one can con-
into the future. The present stage of physical theory is struct a number that has no dimensions: the number h̄c e2
merely a steppingstone toward the better stages we shall That number is found by experiment to have the value
have in the future. One can be quite sure that there will 137, or something very close to 137. Now, there is no
be better stages simply because of the difficulties that known reason why it should have this value rather than
occur in the physics of today. some other number. Various people have put forward
I should now like to dwell a bit on the difficulties in ideas about it, but there is no accepted theory. Still,
the physics of the present day. The reader who is not one can be fairly sure that someday physicists will solve
an expert in the subject might get the idea that because the problem and explain why the number has this value.
of all these difficulties physical theory is in pretty poor There will be a physics in the future that works when
h̄c
shape and that the quantum theory is not much good. e2 has the value 137 and that will not work when it has
I should like to correct this impression by saying that any other value.
quantum theory is an extremely good theory. It gives
wonderful agreement with observation over a wide range The physics of the future, of course, cannot have the
of phenomena. There is no doubt that it is a good the- three quantities h̄ e and c all as fundamental quantities.
ory, and the only reason physicists talk so much about Only two of them can be fundamental, and the third must
the difficulties in it is that it is precisely the difficulties be derived from those two. It is almost certain that c will
that are interesting. The successes of the theory are all be one of the two fundamental ones. The velocity of light,
taken for granted. One does not get anywhere simply c, is so important in the four-dimensional picture, and it
by going over the successes again and again, whereas by plays such a fundamental role in the special theory of
talking over the difficulties people can hope to make some relativity, correlating our units of space and time, that it
progress. has to be fundamental. Then we are faced with the fact
that of the two quantities h-bar and e, one will be funda-
The difficulties in quantum theory are of two kinds. mental and one will be derived. If h-bar is fundamental,
I might call them Class One difficulties and Class Two e will have to be explained in some way in terms of the
difficulties. Class One difficulties are the difficulties I square root of h-bar, and it seems most unlikely that any
have already mentioned: How can one form a consistent fundamental theory can give e in terms of a square root,
picture behind the rules for the present quantum the- since square roots do not occur in basic equations. It is
 4

much more likely that e will be the fundamental quantity results. This method is known as the renormalization
and that h-bar will be explained in terms of c2 . Then method.
there will be no square root in the basic equations. I
think one is on safe ground if one makes the guess that I shall merely explain the idea in words. We start
in the physical picture we shall have at some future stage out with a theory involving equations. In these equa-
e and c will be fundamental quantities and h-bar will be tions there occur certain parameters: the charge of the
derived. electron, e, the mass of the electron, m, and things of
a similar nature. One then finds that these quantities,
If h-bar is a derived quantity instead of a fundamen- which appear in the original equations, are not equal to
tal one, our whole set of ideas about uncertainty will the measured values of the charge and the mass of the
be altered: h-bar is the fundamental quantity that oc- electron. The measured values differ from these by cer-
curs in the Heisenberg uncertainty relation connecting tain correcting terms?Delta e, Delta m and so on?so that
the amount of uncertainty in a position and in a mo- the total charge is e + e and the total mass m + Delta
mentum. This uncertainty relation cannot play a fun- m. These changes in charge and mass are brought about
damental role in a theory in which h-bar itself is not a through the interaction of our elementary particle with
fundamental quantity. I think one can make a safe guess other things. Then one says that e + Delta e and m +
that uncertainty relations in their present form will not Delta m, being the observed things, are the important
survive in the physics of the future. things. The original e and m are just mathematical pa-
Of course there will not be a return to the determinism rameters; they are unobservable and therefore just tools
of classical physical theory. Evolution does not go back- one can discard when one has got far enough to bring in
ward. It will have to go forward. There will have to be the things that one can compare with observation. This
some new development that is quite unexpected, that we would be a quite correct way to proceed if Delta e and
cannot make a guess about, which will take us still fur- Delta m were small (or even if they were not so small but
ther from Classical ideas but which will alter completely finite) corrections. According to the actual theory, how-
the discussion of uncertainty relations. And when this ever, Delta e and Delta m are infinitely great. In spite of
new development occurs, people will find it all rather fu- that fact one can still use the formalism and get results
tile to have had so much of a discussion on the role of in terms of e + Delta e and m + Delta m, which one can
observation in the theory, because they will have then a interpret by saying that the original e and m have to be
much better point of view from which to look at things. minus infinity of a suitable amount to compensate for the
So I shall say that if we can find a way to describe the Delta e and Delta m that are infinitely great. One can
uncertainty relations and the indeterminacy of present use the theory to get results that can be compared with
quantum mechanics that is satisfying to our philosophi- experiment, in particular for electrodynamics. The sur-
cal ideas, we can count ourselves lucky. But if we cannot prising thing is that in the case of electrodynamics one
find such a way, it is nothing to be really disturbed about. gets results that are in extremely good agreement with
We simply have to take into account that we are at a experiment. The agreement applies to many significant
transitional stage and that perhaps it is quite impossible figures?the kind of accuracy that previously one had only
to get a satisfactory picture for this stage. in astronomy. It is because of this good agreement that
I have disposed of the Class One difficulties by say- physicists do attach some value to the renormalization
ing that they are really not so important, that if one can theory, in spite of its illogical character.
make progress with them one can count oneself lucky, and It seems to be quite impossible to put this theory on a
that if one cannot it is nothing to be genuinely disturbed mathematically sound basis. At one time physical theory
about. The Class Two difficulties are the really serious was all built on mathematics that was inherently sound.
ones. They arise primarily from the fact that when we I do not say that physicists always use sound mathemat-
apply our quantum theory to fields in the way we have ics; they often use unsound steps in their calculations.
to if we are to make it agree with special relativity, in- But previously when they did so it was simply because
terpreting it in terms of the three-dimensional sections of, one might say, laziness. They wanted to get results as
I have mentioned, we have equations that at first look quickly as possible without doing unnecessary work. It
all right. But when one tries to solve them, one finds was always possible for the pure mathematician to come
that they do not have any solutions. At this point we along and make the theory sound by bringing in further
ought to say that we do not have a theory. But physi- steps, and perhaps by introducing quite a lot of cumber-
cists are very ingenious about it, and they have found a some notation and other things that are desirable from
way to make progress in spite of this obstacle. They find a mathematical point of view in order to get everything
that when they try to solve the equations, the trouble is expressed rigorously but do not contribute to the physi-
that certain quantities that ought to be finite are actu- cal ideas. The earlier mathematics could always be made
ally infinite. One gets integrals that diverge instead of sound in that way, but in the renormalization theory we
converging to something definite. Physicists have found have a theory that has defied all the attempts of the
that there is a way to handle these infinities according mathematician to make it sound. I am inclined to sus-
to certain rules, which makes it possible to get definite pect that the renormalization theory is something that
 5

will not survive in the future, and that the remarkable that if you suppose it to be a fluid filling up the whole
agreement between its results and experiment should be of space, in any place it has a definite velocity, which de-
looked on as a fluke. stroys the four-dimensional symmetry required by Ein-
This is perhaps not altogether surprising, because stein?s special principle of relativity. Einstein?s special
there have been similar flukes in the past. In fact, Bohr?s relativity killed this idea of the ether.
electron-orbit theory was found to give very good agree-
ment with observation as long as one confined oneself to But with our present quantum theory we no longer
one-electron problems. I think people will now say that have to attach a definite velocity to any given physical
this agreement was a fluke, because the basic ideas of thing, because the velocity is subject to uncertainty re-
Bohr?s orbit theory have been superseded by something lations. The smaller the mass of the thing we are inter-
radically different. I believe the successes of the renor- ested in, the more important are the uncertainty rela-
malization theory will be on the same footing as the suc- tions. Now, the ether will certainly have very little mass,
cesses of the Bohr orbit theory applied to one-electron so that uncertainty relations for it will be extremely im-
problems. portant. The velocity of the ether at some particular
place should therefore not be pictured as definite, be-
The renormalization theory bas removed some of these cause it will be subject to uncertainty relations and so
Class Two difficulties, if one can accept the illogical char- may be anything over a wide range of values. In that
acter of discarding infinities, but it does not remove all way one can get over the difficulties of reconciling the
of them. There are a good many problems left over con- existence of an ether with the special theory of relativity.
cerning particles other than those that come into electro- There is one important change this will make in our
dynamics: the new particles?mesons of various kinds and picture of a vacuum. We would like to think of a vac-
neutrinos. There the theory is still in a primitive stage. It uum as a region in which we have complete symmetry
is fairly certain that there will have to be drastic changes between the four dimensions of space-time as required
in our fundamental ideas before these problems can be by special relativity. If there is an ether subject to un-
solved. certainty relations, it will not be possible to have this
One of the problems is the one I have already men- symmetry accurately. We can suppose that the velocity
tioned about accounting for the number 137. Other prob- of the ether is equally likely to be anything within a wide
lems are how to introduce the fundamental length to range of values that would give the symmetry only ap-
physics in some natural way, how to explain the ratios proximately. We cannot in any precise way proceed to
of the masses of the elementary particles and how to ex- the limit of allowing all values for the velocity between
plain their other properties. I believe separate ideas will plus and minus the velocity of light, which we would have
be needed to solve these distinct problems and that they to do in order to make the symmetry accurate. Thus the
will be solved one at a time through successive stages vacuum becomes a state that is unattainable. I do not
in the future evolution of physics.At this point I find think that this is a physical objection to the theory. It
myself in disagreement with most physicists. They are would mean that the vacuum is a state we can approach
inclined to think one master idea will be discovered that very closely. There is no limit as to how closely we can
will solve all these problems together. I think it is ask- approach it, but we can never attain it. I believe that
ing too much to hope that anyone will be able to solve would be quite satisfactory to the experimental physicist.
all these problems together. One should separate them It would, however, mean a departure from the notion of
one from another as much as possible and try to tackle the vacuum that we have in the quantum theory, where
them separately. And I believe the future development we start off with the vacuum state having exactly the
of physics will consist of solving them one at a time, and symmetry required by special relativity.
that after any one of them has been solved there will still That is one idea for the development of physics in the
be a great mystery about how to attack further ones. future that would change our picture of the vacuum, but
I might perhaps discuss some ideas I have had about change it in a way that is not unacceptable to the ex-
how one can possibly attack some of these problems. perimental physicist. It has proved difficult to continue
None of these ideas has been worked out very far, and I with the theory, because one would need to set up math-
do not have much hope for any one of them. But I think ematically the uncertainty relations for the ether and so
they are worth mentioning briefly. far some satisfactory theory along these lines has not
One of these ideas is to introduce something corre- been discovered. If it could be developed satisfactorily, it
sponding to the luminiferous ether, which was so pop- would give rise to a new kind of field in physical theory,
ular among the physicists of the 19th century. I said which might help in explaining some of the elementary
earlier that physics does not evolve backward. When I particles.
talk about reintroducing the ether, I do not mean to go Another possible picture I should like to mention con-
back to the picture of the ether that one had in the 19th cerns the question of why all the electric charges that are
century, but I do mean to introduce a new picture of the observed in nature should be multiples of one elementary
ether that will conform to our present ideas of quantum unit, e. Why does one not have a continuous distribution
theory. The objection to the old idea of the ether was of charge occurring in nature? The picture I propose goes
 6

back to the idea of Faraday lines of force and involves a and puts it on the electron, thereby making the electron
development of this idea. The Faraday lines of force are interact with the electromagnetic field. This brings a
a way of picturing electric fields. If we have an electric perturbation into the equations and causes a change in
field in any region of space, then according to Faraday the mass of the electron, the Delta m, which is to be
we can draw a set of lines that have the direction of the added to the previous mass of the electron. The proce-
electric field. The closeness of the lines to one another dure is rather roundabout because it starts off with the
gives a measure of the strength of the field?they are close unphysical concept of the bare electron. Probably in the
where the field is strong and less close where the field is improved physical picture we shall have in the future the
weak. The Faraday lines of force give us a good picture bare electron will not exist at all.
of the electric field in classical theory. Now, that state of affairs is just what we have with the
When we go over to quantum theory, we bring a kind discrete lines of force. We can picture the lines of force
of discreteness into our basic picture. We can suppose as strings, and then the electron in the picture is the end
that the continuous distribution of Faraday lines of force of a string. The string itself is the Coulomb force around
that we have in the classical picture is replaced by just a the electron. A bare electron means an electron without
few discrete lines of force with no lines of force between the Coulomb force around it. That is inconceivable with
them. Now, the lines of force in the Faraday picture end this picture, just as it is inconceivable to think of the end
where there are charges. Therefore with these quantized of a piece of string without thinking of the string itself.
Faraday lines of force it would be reasonable to suppose This, I think, is the kind of way in which we should try to
the charge associated with each line, which has to lie at develop our physical picture?to bring in ideas that make
the end if the line of force has an end, is always the same inconceivable the things we do not want to have. Again
(apart from its sign), and is always just the electronic we have a picture that looks reasonable, but I have not
charge, ? e or + e. This leads us to a picture of discrete found the proper equations for developing it.
Faraday lines of force, each associated with a charge, ? e I might mention a third picture with which I have been
or + e. There is a direction attached to each line, so that dealing lately. It involves departing from the picture of
the ends of a line that has two ends are not the same, the electron as a point and thinking of it as a kind of
and there is a charge + e at one end and a charge ? e sphere with a finite size. Of course, it is really quite an
at the other. We may have lines of force extending to old idea to picture the electron as a sphere, but previ-
infinity, of course, and then there is no charge. ously one had the difficulty of discussing a sphere that is
If we suppose that these discrete Faraday lines of force subject to acceleration and to irregular motion. It will
are something basic in physics and lie at the bottom of get distorted, and how is one to deal with the distortions?
our picture of the electromagnetic field, we shall have an I propose that one should allow the electron to have, in
explanation of why charges always occur in multiples of e. general, an arbitrary shape and size. There will be some
This happens because if we have any particle with some shapes and sizes in which it has less energy than in oth-
lines of force ending on it, the number of these lines must ers, and it will tend to assume a spherical shape with a
be a whole number. In that way we get a picture that is certain size in which the electron has the least energy.
qualitatively quite reasonable. This picture of the extended electron has been stimu-
lated by the discovery of the mu meson, or muon, one of
We suppose these lines of force can move about. Some the new particles of physics. The muon has the surpris-
of them, forming closed loops or simply extending from ing property of being almost identical with the electron
minus infinity to infinity, will correspond to electromag- except in one particular, namely, its mass is some 200
netic waves. Others will have ends, and the ends of these times greater than the mass of the electron. Apart from
lines will be the charges. We may have a line of force this disparity in mass the muon is remarkably similar to
sometimes breaking. When that happens, we have two the electron, having, to an extremely high degree of ac-
ends appearing, and there must be charges at the two curacy, the same spin and the same magnetic moment in
ends. This process?the breaking of a line of force?would proportion to its mass as the electron does. This leads
be the picture for the creation of an electron (e-) and a to the suggestion that the muon should be looked on as
positron (e+). It would be quite a reasonable picture, an excited electron. If the electron is a point, picturing
and if one could develop it, it would provide a theory how it can be excited becomes quite awkward. But if the
in which e appears as a basic quantity. I have not yet electron is the most stable state for an object of finite
found any reasonable system of equations of motion for size, the muon might just be the next most stable state
these lines of force, and so I just put forward the idea as in which the object undergoes a kind of oscillation. That
a possible physical picture we might have in the future. is an idea I have been working on recently. There are
There is one very attractive feature in this picture. It difficulties in the development of this idea, in particular
will quite alter the discussion of renormalization. The the difficulty of bringing in the correct spin.
renormalization we have in our present quantum electro- I have mentioned three possible ways in which one
dynamics comes from starting off with what people call might think of developing our physical picture. No doubt
a bare electron?an electron without a charge on it. At there will be others that other people will think of. One
a certain stage in the theory one brings in the charge hopes that sooner or later someone will find an idea that
 7

really fits and leads to a big development. I am rather ematical basis of quantum theory, trying to understand
pessimistic about it and am inclined to think none of the theory better and to make it more powerful and more
them will be good enough. The future evolution of ba- beautiful. If someone can hit on the right lines along
sic physics?that is to say, a development that will really which to make this development, it may lead to a future
solve one of the fundamental problems, such as bring- advance in which people will first discover the equations
ing in the fundamental length or calculating the ratio of and then, after examining them, gradually learn how to
the masses?may require some much more drastic change apply them. To some extent that corresponds with the
in our physical picture. This would mean that in our line of development that occurred with Schrodinger?s dis-
present attempts to think of a new physical picture we covery of his wave equation. Schrodinger discovered the
are setting our imaginations to work in terms of inade- equation simply by .looking for an equation with mathe-
quate physical concepts. If that is really the case, how matical beauty. When the equation was first discovered,
can we hope to make progress in the future? people saw that it fitted in certain ways, but the general
There is one other line along which one can still pro- principles according to which one should apply it were
ceed by theoretical means. It seems to be one of the fun- worked out only some two or three years later. It may
damental features of nature that fundamental physical well be that the next advance in physics will come about
laws are described in terms of a mathematical theory of along these lines: people first discovering the equations
great beauty and power, needing quite a high standard of and then needing a few years of development in order to
mathematics for one to understand it. You may wonder: find the physical ideas behind the equations. My own
Why is nature constructed along these lines? One can belief is that this is a more likely line of progress than
only answer that our present knowledge seems to show trying to guess at physical pictures.
that nature is so constructed. We simply have to accept Of course, it may be that even this line of progress will
it. One could perhaps describe the situation by saying fail, and then the only line left is the experimental one.
that God is a mathematician of a very high order, and Experimental physicists are continuing their work quite
He used very advanced mathematics in constructing the independently of theory, collecting a vast storehouse of
universe. Our feeble attempts at mathematics enable us information. Sooner or later there will be a new Heisen-
to understand a bit of the universe, and as we proceed to berg who will be able to pick out the important features
develop higher and higher mathematics we can hope to of this information and see how to use them in a way sim-
understand the universe better. ilar to that in which Heisenberg used the experimental
This view provides us with another way in which we knowledge of spectra to build his matrix mechanics. It
can hope to make advances in our theories. Just by is in evitable that physics will develop ultimately along
studying mathematics we can hope to make a guess at the these lines, but we may have to wait quite a long time
kind of mathematics that will come into the physics of the if people do not get bright ideas for developing the theo-
future. A good many people are working on the math- retical side.