Einstein and the quantum theory*

A. Pais
Rockefeller Uniuersity, New York, ¹w York 10021

This is an account of Einstein's work and thoughts on the quantum theory. The following topics will be
discussed: The light-quantum hypothesis and its gradual evolution into the photon concept. Early history
of the photoelectric effect. The theoretical and experimental reasons why the resistance to the photon was
stronger and more protracted than for any other particle proposed to date. Einstein's position regarding
the Bohr —Kramers — Slater suggestion, the last bastion of resistance to the photon. Einstein's analysis of
fluctuations around thermal equilibrium and his proposal of a duality between particles and waves, in 1909
for electromagnetic radiation {the first time this duality was ever stated) and in January 1925 for matter
(prior to quantum mechanics and for reasons independent of those given earlier by de Broglie). His
demonstration that long-known specific heat anomalies are quantum effects. His role in the evolution of
the third law of thermodynamics. His new derivation of Planck's law in 1917 which also marks the
beginning of his concern with the failure of classical causality. His role as one of the founders of quantum
statistics and his discovery of the first example of a phase transition derived by using purely statistical
methods. His position as a critic of quantum mechanics. Initial doubts on the consistency of quantum
mechanics (1926—1930). His view maintained from 1930 until the end of his life: quantum mechanics is
logically consistent and quite successful but it is incomplete. His attitude toward success. His criterion of
objective reality. Differences in the roles relativity and quantum theory played in Einstein s life. His
vision regarding quantum theory in the context of a unified field theory. His last autobiographical sketch,
written a few months before his death, concluding with a statement about the quantum theory, a subject
to which {by his own account) he had given more thought than even to general relativity.
CONTENTS B. Bose 894
C. Einstein- 895
I. Einstein, the Quantum and Apartness 863 D. Postscript on Bose- Einstein condensation 897
A. Introduction VII. Einstein, as a Transitional Figure: The Birth of
B. Particle physics: The first fifty years 865 Wave Mechanics 897
C. The quantum theory: Lines of influence 866 A. From Einstein to de Broglie 897
II. The Light-Quantum 867 B. From de Broglie to Einstein 898
A. From Kirchhoff to Planck 867 C. From de Broglie and Einstein to Schroedinger 899
B. Einstein on Planck: 1905. The Rayleigh- VIII. Einstein's Response to the New Dynamics 899
Einstein —Jeans law 871 A. 1925—1933. The debate begins 899
C. The light-quantum hypothesis and the heuristic B. Einstein on objective reality 903
principle IX. A Time Capsule 905
D. Three remarks on Einstein's statistical phys- X. Particles, Fields and the Quantum Theory:
ics 873 Einstein's vision 906
E. Einstein on Planck: 1906 875 A. Some reminiscences 906
F. The photoelectric effect: The second coming B. Einstein, Newton, and success 907
of h 875 C. Relativity theory and quantum theory 908
1. 1887: Hertz 876 D. Einstein's vision 909
2. 1888: Hallwachs 876 XI. Epilogue 910
3. 1899: J. J. Thomson 876 Acknowledgments 911
4. 1902: Le nard 876 References 911
5. 1905: Einstein 876
6. 1915: Millikan: the Duane —Hunt limit 877
G. The fusion of particles and waves and Ein-
stein's destiny 877
IQ. Einstein and Specific Heats 878
A. Specific heats in the nineteenth century 878
B. Einstein 881 I. EINSTEIN, THE QUANTUM AND APARTNESS.
C. Nernst; Solvay I 882
IV. From the Light-Quantum to the Photon 883 Apart, adv. , 4. Away from others in
A. Reactions to the light-quantum hypothesis 883 action or function; separately,
1. Einstein's caution 883 independently, individually
2. Electromagnetism: Free fields and inter-
actions 884 Oxford English Dictionary
3. The impact of experiment 885
B. Spontaneous and induced radiative transitions 886 A. Introdoctiop
C. The completion of the particle picture 886
D. Earliest Unbehagen about chance 888
E. The Compton effect In 1948 I undertook to put together the Festschrift in
honor of Einstein's seventieth birthday. ' In a letter to
889
V. Interlude: The BKS proposal 890
VI. A Loss of Identity: The Birth of Quantum prospective contributors I wrote (1948): "It is planned
Statistics 893 that the first article of the volume shall be of a more
A. Boltzmann's axiom 894 personal nature and, written by a representative col-

ork supported in part by the U. S. Department of Energy

under contract grant No. EY-76-C-02-22328. 000. Rev. Mod. Phys. 21, No. 3, 1949.

Reviews of Modern Physics, Vol. 51, No. 4, October 1979 Copyright 1979 American Physical Society
 A. Pais: Einstein and the quantum theory

league, shall pay homage to Einstein on behalf of all phenomena all would find their place. Einstein pursued
contributors. " I then asked Robert Andrews Millikan this program from about 1920 (before the discovery of
(1868—1953) to do the honors, as the senior contributor. ' quantum mechanics!) until the end of his life. Numerous
He accepted and his article (1949) is written in his cus- discussions with him in his later years have helped me
tomary forthright manner. On that occasion he ex- gain a better understanding of his views. Some personal
pressed himself as follows on the equation E= A v —I' reminiscences of my encounters with Einstein are found
for the photoelectric effect. "I spent ten years of my in Sec. X.A.
life testing that 1905 equation of Einstein s and, con- But let me first return to the days of the old quantum
trary to all my expectations, I was compelled in 1915 theory. Einstein's contributions to it can be grouped
to assert its unambiguous verification in spite of its under the following headings.
unreasonableness since it seemed to violate everything (a) The lzglzf-quantum. In 1900 PLanck had discovered
we knew about the interference of light. " the blackbody radiation law without using light-quanta.
Physics had progressed, and Millikan had mellowed In 1905 Einstein discovered light-quanta without using
since the days of his 1915 paper on the photoeffect, as Planck's law. Section II is devoted to the light-quantum
is evidenced by what he had written (1916a) at that ear- hypothesis. The interplay between the ideas of Planck
lier time: "Einstein's photoelectric equation. . . . ap- and Einstein-is discussed. A brief history of the photo-
pears in every case to predict exactly the observed re- electric effect from 1887 to 1915 is given. This section
sults . . . Yet the semicorpuscular theory by which Ein- ends with Einstein's formulation, in 1909, of the parti-
stein arrived at his equation seems at present wholly cle-wave duality for ihe case of electromagnetic radia-
untenable"; and in his next paper (1916b) Millikan men- tion.
tioned "the bold, not to say the reckless, hypothesis of (b) Specific heats. Toward the end of the nineteenth
an electromagnetic light corpuscle. " Nor was Millikan century there existed evident conflicts between the data
at that time the only first-rate physicist to hold such on specific heats and their interpretation in terms of the
views, as will presently be recalled. Rather, the phys- -
equipartition theorem of classical statistical mechan-
ics community at large had received the light-quantum ics. In 1907 Einstein published the first paper on quan-
hypothesis with disbelief and with skepticism bordering tum effects in the solid state which showed the way out
on derision. As one of the architects of the pre-1925 of these paradoxes. This paper also played an impor-
quantum theory, the old quantum theory, Einstein had tant role in the final formulation of the third law of
quickly found both enthusiastic and powerful support for thermodynamics. These topics are discussed in Sec.
one of his two major contributions to this field: the III.
quantum theory of specific heat. (There is no reason (c) The PIzotozz. The light-quantum as originally de-
to believe that such support satisfied any particular need fined was a parcel of energy. The concept of the pho-
in him. ) By sharp contrast, from 1905-1923 he was a ton as a particle with definite energy and momentum
man apart in being the only one, or almost the only one, emerged only gradually. Einstein himself did not dis-
to take the light-quantum seriously. cuss photon momentum until 1917. Relativistic energy
If I had to characterize Einstein by one single word momentum conservation relations involving photons
I would choose "apartness. " This was forever one of were not written down till 1923. Einstein's role in
his deepest emotional needs. It was to serve him in his these developments is discussed in Sec. IV. At the be-
singleminded and singlehanded pursuits, most notably ginning of that section I continue the discussion of the
on his road to triumph from the special to the general reactions to the light-quantum hypothesis, of which I
theory of relativity. It was also to become a practical have already given a few samples. This section also
necessity for him, in order to protect his cherished contains an account of Einstein's discovery of theA and
privacy from a world hungry for legend and charisma. B coefficients and of his earliest concern with the break-
In all of Einstein's scientific career, this apartness was down of classical causality. The section concludes with
never more pronounced than in regard to the quantum remarks on the role of the Compton effect.
theory. This covers two disparate periods, the first The reader may wonder why the man who wrote down
one oi' which (1905—1923) I have just mentioned. During the relation F. = h v for light in 1905 and who propounded
the second period, from 1926 until the end of his life, the special theory of relativity in that same year would
he was the only one, or again nearly the only one, to not have stated sooner the relation p = Iz v/c. I shall
maintain a profoundly skeptical attitude to quantum me- comment on this question in Sec. X.C.
chanics. I shall discuss Einstein's position on quantum (d) Einstein's work on quantum statistics is treated
mechanics in Secs. VIII and X but cannot refrain from in Sec. VI, which also includes a discussion of Boltz-
stating at once that Einstein's skepticism should not be mann's axiom on identical distinguishable particles and
equated with a purely negative attitude. It is true that of Bose's contribution.
he was forever critical of quantum mechanics. But at (e) Einstein's role as a key transitional figure in the
the same time he had his own alternative program for a discovery of wave mechanics will be discussed in Sec.
synthetic theory in which particles, fields, and quantum VII.
I shall continue the outline of this paper in part (c) of
this section. First, however, I should like to take leave
of our main character for a brief while in order to give
an'overview of the singular role of the photon in the
It was decided later that L. de Broglie, M. von Laue, and history of the physics of particles and fields. In so do-
Ph. Frank should also write articles of a more personal nature. ing I shall i.nterrupt the historical sequence of events in

Rev. Mod. Phys. , Vot. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 865

order to make some comments from today's vantage models of the nucleus were proposed, and nuclear phys-
point. ics could start in earnest. The muon is still one of the
strangest animals in the particle zoo, yet its discovery
B. Particle physics: The f irst f ifty years was liberating too since it made possible an understand-
ing of certain anomalies in the absorption of cosmic
rays (.Prior to the discovery of the muon, theorists
Let us leave aside the photon for a while and ask how
had already speculated about the need for an extra parti-
physicists reacted to the experimental discovery or the
theoretical prediction (whichever came first) of other cle to explain these anomalies. )
new particles. ' To complete the particle list of the first half century
there are four more particles' which have entered phys-
The discovery in 1897 of the first particle, the elec- —
tron, was an unexpected experimental development
ics but in a different way: initially they were theo-
which brought to an end the ongoing debate: are cathode
retical proposals.
The first neutrino was proposed in order to save the
rays molecular torrents or aetherial disturbances~ The
law of energy conservation in beta radioactivity. The
answer came as a, complete surprise: They are neither
but rather a new form of matter. There were some
first meson (now called the pion) was proposed as the
conveyer of nuclear forces. Both suggestions were in-
initial reactions of disbelief. Joseph John Thomson
(1856-1940) once recalled (1936) the comment of a col- genious, daring, innovative, and successful —
but did
not demand a radical change of theory. Within months
league who was present at the first lecture Thomson .
after the public unveiling of the neutrino hypothesis the
gave on the new discovery: "I [J.J.j was told long after-
wards by a distinguished physicist who had been present
first theory of the weak interaction, which is still im-
men 8 ely Useful was proposed The meson hypothesis
at my lecture that hethought I had been 'pul1. ing their
legs'. " Nevertheless the existence of the electron was immediately led to considerable theoretical activity as
well.
widely accepted within the span of very few years. By
The neutrino hypothesis was generally assimilated
1900 it had become clear that beta. rays are electrons
long before this particle was actually observed. The
as well. The discoveries of the free electron and of the
interval between the proposal and the first observation
Zeeman effect (in 1896) combined made it evident that a
of the neutrino is even longer than the corresponding
universal atomic constituent had been discovered and
interval for the photon. The meson postulate found rapid
that the excitaiions of electrons in a.toms were somehow
the sources of atomic spectra.
—
experimental support from cosmic-ray data or so it
seemed. More than a decade passed before it became
The discovery of the electron was a discovery at the
outer experimental frontier. In the first instance this
clear that the bulk of these observations actually in-
volved muons instead of pions.
finding led to the abandonment of some earlier qualita-
Then there was the positron, "a new kind of particle,
tive concepts (of the indivisibility of the atom), but it
unknown to experimental physics, having the same mass
did not require, or at least not at once, a modification
and opposite charge to an electron" (Dirac, 1931). This
of the established corpus of theoretical physics.
During the next fifty years three other particles en-
particle was proposed in 1931, after a period of about
tered the scene in ways not so dissimilar from the case three years of considerable controversy over the mean-
ing of the negative energy solutions of the Dirac equa-
of the electron, namely via unexpected discoveries of
tion. During that period one participant (Weyl, 1930)
expressed fear for "a new crisis in quantum physics. "
an experimental nature at the outer frontier. They are:
the proton, (or, rather, the nucleus), the neutron'
—
and just half a century after the electron —
the muon,
The crisis was short-lived, however. The experimental
discovery of the positron in 1932 was a triumph for
the first of ihe electron's heavier brothers. As to the
theoretical physics. The positron theory belongs to the
acceptance of these particles, it took little time to rea-
most important advances of the nineteen thirties.
lize that their coming was in each instance liberating.
And then there was the photon, the first particle to
Within two years after Rutherford's nuclear model,
be predicted theoretically.
Niels Bohr (1885-1963) was able to make the first real
Never, either in the first half-century or in the years
theoretical predictions in atomic physics. Almost at
thereafter, has the idea of a new particle met for so
once after the discovery of the neutron, the first viable
long with such almost total resistance as the photon.
The light-quantum hypothesis was considered somewhat
of an aberration even by leading physicists who other-
No detailed references to the literature will be given, in
keeping with the brevity of my comments on this subject. wise held Einstein in the highest esteem. Its assimila-
4It is often said, and not without grounds, that the neutron tion came after a struggle more intense and prolonged
was actually anticipated. In fact, twelve years before its dis- than for any other particle ever postulated. Because
covery, in one of his Bakerian lectures (1920), Ernest Ruther- never, to this day, has the proposal of any particle but
ford (1871—1937) spoke of "the idea of the possible existence of the photon led to the creation of a new inner frontier.
an atom of mass one which has zero nuclear charge. "
Nor is The hypothesis seemed paradoxical: light was known
there any doubt that the neutron being in the air at the Caven-
dish was of profound importance (Chadwick, 1962) to its dis-
to consist of waves, hence it could not consist of parti-
coverer James Chadwick (1891—1974). Even so, not even a cles. Yet this paradox alone does not fully account for
Rutherford could have guessed that his 1920 neutron (then con- the resistance to Einstein's hypothesis. %e shall look
jectured to be a tightly bound proton-electron system) was so more closely at the situation in Sec. IV.A.
essentially different from the particle that would eventually go
by that name. ~It is too early to include the graviton.

Rev. Mod. Phys. , Vot. 51, No. 4, October '1979

 A. Pais: Einstein and the quantum theory

Kirchhof f
/g
Bunsen

Balmer
W. Wien

Bohr P lonck Einstein

](

Bose

'I(
Brog lie
Heisenberg

Dirac
v'

Matrix Quantum Nave

Mechanics Field Mechanics
Theory
FIG. 1. The quantum theory: Lines of influence.

C. The quantum theory: Lines of influence in the birth of wave mechanics, discussed in Sec. VII.
The h marking the arrow from Planck to Bohr serves
The skeleton diagram given in Fig. 1 is an attempt to as a reminder that not so much the details of Planck's
reduce the history of the quantum theory to its barest work on radiation as the very introduction by Planck
outlines. At the same time this figure will serve as a of his new universal constant h was decisive for Bohr's
guide to the rest of this paper. X —Y means: the work ideas about atomic stability. An account of Bohr's in-
of X was instrumental to an advance by Y. Arrows fluence on Heisenberg and of Heisenberg's and Schroe-
marked M and R indicate that the influence went via dinger's impact on Dirac is beyond the scope of the
the theory of matter and of radiation, respectively. present paper.
If Planck, Einstein, and Bohr are the fathers of the In the case of Einstein and Bohr it cannot be said that
quantum theory, then Gustav Robert Kirchhoff (1824- the work of one induced major advances in the work of
1887) is its grandfather. Since he was the founder of op- the other. Therefore the simplified diagram does not
tical spectra analysis fin 1860, together with Robert and should not contain links between them. Neverthe-
Bunsen (1811-1899)], an arrow leads from him and less, for forty years there were influences at work
Bunsen to Johann Jakob Balmer (1825-1898), the in- between Einstein and Bohr and these were in fact in-
ventor of the Balmer formula (1885). From Balmer tense, but they were on a different plane. In a spirit
we move to Bohr, the founder of atomic quantum dy- of friendly and heroic antagonism these two men argued
namics. Returning to Kirchhoff, as the discoverer of about questions of principle. Section V deals with
the universal character of blackbody radiation (Kirch- Bohr's resistance to Einstein's idea of the photon. This
hoff, 1860), we note that his influence goes via Wien was but a brief interlude. It ended with the detailed
to Planck. (See further Sec. II.A. ) experimental vindication of the photon concept to which
The arrow from %ien to Planck refers to the latter's Bohr fully subscribed from then on. Their far more
formulation of his blackbody radiation law (Sec. II.A), important debate on the foundations of quantum mechan-
and the triangle Mien-Planck-Einstein to the mutual ics began in 1927. On these issues the intellectual re-
influences which led to the light-quantum hypothesis sistance and opposition of one against the most basic
(Sec. II.B-II.E). views held by the other continued unabated until the end
The arrow from Bose to Einstein refers to Bose's of Einstein's life. At issue were the criteria by which
work on electromagnetic radiation and its impact on one should judge the completeness of the description of
Einstein's contributions to the quantum statistics of a the physical world. Their discussions have noi affected
material gas. See Sec. VI, where Einstein's influence the evolution of physical theory. Yet theirs will be re-
on Dirac is also briefly mentioned. membered as one of the great debates on scientific
The triangle Einstein-de Broglie-Schroedinger has principle between two dominant contemporary figures.
to do with the role of Einstein as the transitional figure The dialog between Bohr and Einstein had one positive

Rev. Mod. Phys. , Vol. 61, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 867

outcome: it forced Bohr to express the tenets of corn Qv. Let A„be its absorption coefficient for frequency v.
plementarity in increasingly precise language. This de- Kirchhoff's theorem (1860) states that E,/A, depends
bate will be one of the themes of Sec. VIII which deals only on v and the temperature T and is independent of
with Einstein's objections to quantum mechanics. any other characteristic of the body:
A point made earlier bears repeating here: Einstein's
own visions on physics issues were often in opposition E,//i „=J(v, T) .
to the mainstream, but they were never negative. So Kirchhoff called a body perfectly black if A, = 1. Thus
it was in the case of quantum mechanics. After 1930 J(v, T) is the emissive power of a black body. He also
he considered this theory to be consistent and success- gave an operational definition for a system, the
ful but incomplete. At the same time he had his own "Hohlraumstrahlung", which acts as a perfectly black
aspirations for a future theory of particles and fields. body: "Given a space enclosed by bodies of equal tem-
I shall try to make clear in Sec. X what these were. perature, through which no radiation can penetrate, then
I do not believe that Einstein presented valid argu- every bundle of radiation within this space is constituted,
ments for the incompleteness of quantum theory. But with respect to quality and intensity, as if it came from
neither do I think that the times are ripe to answer the a completely black body of the same temperature. "
question whether the quantum-mechanical description Kirchhoff (1860) challenged theorists and experimen-
is indeed complete, since to this day the physics of talists alike: "It is a highly important task to find this
particles and fields is a subject beset with many un- function [J]. Great difficulties stand in the way of its
resolved fundamental problems. Among these there experimental determination. Nevertheless there appear
is one which was most dear to Einstein and with which grounds for the hope that it can be determined by ex-
he (and all of us, to date) struggled in vain: the syn- periment, since undoubtedly it has a simple form as do
thesis of quantum physics with general relativity. Since all functions which do not depend on the properties of
we still have far to go, any assessment of Einstein's individual bodies and which one has become acquainted
views must necessarily be tentative. In order to stress with til. l now. "
this I have prefaced Sec. X on Einstein's vision with a Kirchhoff's emphasi. s on the experimental complexi-
very brief overview (Sec. IX) of the current status of ties turned out to be well justified. Even the simple
particle physics. property of J that it has one pronounced maximum which
moves to lower v with decreasing T was not firmly es-
I I. THE L IG HT-QUANTUM tablished experimentally until about twenty years later
(Kangro, 1976). The experimentalists had to cope with
A. From Kirchhoff to Planck
three main problems: (1) to construct manageable
In the last four months of 1859 there occurred a num- bodies with perfectly black properties; (2) to devise
ber of events which were to change the course of sci- radiation detectors with adequate sensitivity, and (3)
ence. to find ways of extending the measurements over large
On the twelfth of September, Urbain Jean Joseph Le frequency domains. Forty years of experimentation
Verrier (1811-1877) submitted to the French Academy had to go by before the data were sufficient to answer
the text of a letter to Hervh Faye (1814—1902) in which Kirchhoff's question.
he recorded that the perihelion of Mercury advances by Kirchhoff derived Eq. (1) by showing that its violation
thirty-eight seconds per century due to "some as yet would imply the possibility of a "perpetuum mobile" of
unknown action on which no light has been thrown, "' (Le the second kind. The novelty of his theorem was not
Verrier, 1859). The effect was to remain unexplained' un- so much its content as the precision and generality of
til the days of general relativity. On the twenty-fourth of its proof, based exclusively on the still-young science
November a book was published in London, entitled On the of thermodynamics. A quarter of a century passed be-
Origin of SPecies by Means of Natural Selection, or the fore the next theoretical advance in blackbody radiation
Preservation of favoured Races in the Struggle for Life, came about.
by Charles Robert Darwin (1809-1882). Meanwhile on the In 1879 Josef Stefan (1835-1893) conjectured on ex-
twentieth of October Gustav Kirchhoff (1859) from perimental grounds that the total energy radiated by a
Heidelberg submitted his observation that the dark D- hot body varies with the fourth power of the absolute
lines in the solar spectrum are darkened still further temperature, (Stefan, 1879). This statement is not true in
by the interposition of a sodium flame. As a result, its generality. The precise formulation was given in 1884
a few weeks later he proved a theorem and posed a when Ludwig Boltzmann (1844 —1906), [then a professor
challenge. The response to Kirchhoff's challenge led to of experimental physics in Graz (Austria)j, proved
the discovery of the quantum theory. —
theoretically that the strict T' law holds an«»y
Consider a body in thermal equilibrium with radiation. —
holds for bodies which are black, (Boltzmann, 1884). His
Let the radiation energy which the body absorbs be con- proof involved again thermodynamics, but combined this
verted to thermal energy only, not to any other energy time with a still younger branch of theoretical physics: The
form. Let E„dv denote the amount of energy emitted by electromagnetic theory of James Clerk Maxwell (1831-
the body per unit time per cm' in the frequency interval 1879).
For the case of Hohlraumstrahlung the radiation is
homogeneous, isotropic and unpolarized so that
".. . du a quelque action encore inconnue, 'cui theoriae lumen
nundum access crit'. " Z(v, T) = (c/8m) p(v, T) . (2)
VThe present value is 43 sec per century. p(v, T), the spectral density, is the energy density per

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

868 A. Pais: Einstein and the quantum theory

unit volume for frequency v. In this case the Stefan- Progress was slow in the next forty years, as demon-
Boltzmann law reads (V is the volume of the cavity) strated by a question raised by Samuel Pierpont Lang-
ley (1834 —1906) in a lecture given in 1885 before the
E(7') -=v
f P(v, Y')dL =ave'. AAAS meeting in Ann Arbor: "Does [the] ultimate wave-
length of 2. V p which our atmosphere transmits corres-
This law was the very first thermodynamical conse- pond to the lowest [frequency] which can be obtained
quence derived from Mmovell's theorem according to from any terrestrial sour ce ?"'(Langley, 1886). The
which the numerical value of the radiation pressure great advance came in the 1890s. The first sentence
equals one-third of the energy per unit volume. When in of the first paper in the first issue of the Physical Re-
1893 Wilhelm Wien (1864 —1928) proved his displacement view reads as follows: "Within a few years the study
law (Wien, 1893) of obscure radiation has been greatly advanced by sys-
tematic inquiry into the laws of dispersion of the in-
frared rays. " This was written in 1893, by Ernest
(4)
one had come as far as is possible on the basis of Fox Nichols (1869 —1924). At about that time new tech-
thermodynamics and general electromagnetic theory. niques were developed which culminated in the
[Proofs of Egs. (3), (4) are found in standard texts. ] "Reststrahlen" ("residual rays") method of Rubens and
Meanwhile, ever since the 1860s proposals for the Nichols (1897): one eliminates short wavelengths from
correct form of p had begun to appear. All these gues- a beam of radiation by subjecting it to numerous re-
ses may be forgotten except for one, Wien's exponen- flections on quartz or other surfaces. This procedure
tial law, proposed in 1896, (Wien, 18S6): leads to the isolation of the long wavelengths in the
3 -BV/T beam. These experimental developments are of funda-
p (5)
mental importance for our main subject, the quantum
Experimental techniques had sufficiently advanced by theory, since they were crucial to the discovery of the
then to put this formula to the test. This was done by blackbody radiation law.
Friedrich Paschen (1865-1947) from Hannover whose The paper by Rubens and Kurlbaum was presented to
measurements (very good ones) were made in the near- the Prussian Academy on October 25, 1900. Figure 2
infrared, A. = 1 —8p, (and T =400 —1600 K). He published shows some of the measured points they recorded and
his data in January 1897. His conclusion: "It would some theoretical curves with which they compared their
seem very difficult to find another function [of v and T, findings. One of these was the Wien curve, which did
Eq. (5)] which represents the data with as few con- not work. Neither did a second curve proposed by Lord
"
stants, (Pasc hen, 1897). For a brief period it appeared Rayleigh. (I return to Rayleigh's work in Sec. II.B.)
that Wien's law was the final answer. But then, in the year I shall leave aside the other two comparison curves
1900, this conclusion turned out to be premature and the which they drew and turn to the all important "fifth
correct response to Kirchhoff's challenge was found. Two formula, given by Herr M. Planck after our experiments
factors were decisive. One, a breakthrough in experimen- had already been concluded. . . . [and which] reproduces
tal techniques in the far infrared. The other, the per- our observations [from —188 to 1500 C] within the lim-
sistence and vision of Max Karl Ernst Ludwig Planck its of error" (Rubens and Kurlbaum, 1900).
(1858-1947). Kirchhoff had moved from Heidelberg to Berlin to take
It happened in Berlin. At the Physikalisch Technische the chair in theoretical physics. After his death this
Reichsanstalt, at that time probably the world's best position was offered to Boltz mann who declined. Then
equipped physics laboratory, two teams were inde- Heinrich Hertz (1857-1894) was approached; he also
pendently at work on blackbody radiation experiments. declined. The next candidate was Planck to whom the
The first of these, Otto Lummer (1860—1925) and offer of extraordinarius (associate professor) was made.
Ernst Pringsheim (1859-1917), had tackled the prob- Planck accepted and was soon promoted to full pro-
lem in an as yet unexplored wave length region, A. = 12 fessor. His new position brought him in close proximity
—18&, (and T = 300-1650 'K). In February 1900 they to the experimental developments outlined above. This
stated their conclusion: Wien's law fails in that re- nearness was to be one of the decisive factors in the
gion, 8 (Lummer and Pringsheim, 1900). The second team, destiny of this most unusual man.
consisting of Heinrich Rubens (1865-1922) and Ferdinand Planck most probably" discovered his law in the early
Kurlbaum (1857-1927), moved even farther into the in- evening of Sunday, October 7. Rubens and his mife had
frared: A, = 30 —60 p, , (and T = —200 —1500 C). They ar- called on the Plancks on the afternoon of that day. In
rived at the same conclusion (Rubens and Kurlbaum, 1SOO). the course of the conversation Rubens mentioned to
I need to say more about the latter results but I should
like to comment first on the role of experiment in the
discovery of the quantum theory. The Rubens-Kurlbaum These refer to observations at A, =51.2p. This wavelength
paper is a classic. The work of these authors as mell was isolated by multiple reQections off rock salt. The black-
as that of Paschen and of Lummer and Pringsheim was body radiation intensity is plotted as a function of T. (Recall
of a pioneering nature. By the middle of the nineteenth that after multiple reflection those specific frequencies pre-
century wavelengths had been measured up to A. - 1.5p, . dominantly survive which correspond to the ionic vibrations in
the crystal lattice chosen as reflector. )
~
Here I rely on the obituary of Rubens by Georg Hettner
(1922) (himself an experimental expert on blackbody radiation).
There had been earlier indications of deviations from Wien's Hettner's account differs slightly from the recollections which
law, but these were not well documented. Planck himself wrote (1958) in his late eighties.

Rev. IVIod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum iheory

——
Lip. llll. Restslrahlen
—E-"f(i)
von iSteinsr ilz.

herechnei noh %en .

TI&iesen
!
Lj&re Sayleijh
LI &m'mer J-aha ke(p=4 P.=, )p)
beoba ehlet ~ jjj+

f00 200 800 400~ 60!0!' 700 800 900 000 BOO l$00' f400'
1000' fSOO 0

Q 0 Q beobaehtet milden sehearzen Korpel n l, mu. m

+++ jj " dern sdavargen lQrpel.
OOO ~ j ~j j
~

S j
~ jj jj

FIG. 2. Sample of the Rubens-Kurlbaum data which led Planck to guess his radiation formula (Rubens and Kurlbaum, 1900). p is
plotted versus T for A. = 51, 2p, . ("berechnet nach" means "computed after", "beobachtet" means "observed". ) The curves marked
"Wien" and "Lord Rayleigh" refer to best fits to the Eqs. (5), (17), respectively. The curves marked "Thiesen" and "Lummer-
Jahnke" refer to theoretical proposals which are not discussed in the present paper. Planck's formula is not yet plotted.

Planck that he had found p(v, T) to be proportional to T the far infrared.

for small v. Planck went to work after the visitors had It would do grave i. njustice to Planck if I left the read-
left and found an interpolation between this result and er with the impression that Planck's discovery was ex-
Wien's law, Eq. (5). He communicated his formula by clusively the result of interpolating experimental data.
postcard to Rubens, that same evening, and stated it For years it had been his ambition to derive the correct
publicly (Planck, 1900a) in a discussion remark on radiation law from first principles. Thus the rapidity
October 19, following the presentation of a paper by of his response to Rubens' remark is less surprising
Kurlbaum. Expressed in notations introduced by Planck than the correctness of his answer. I shall discuss
two months later, he proposed that elsewhere Planck's earlier research (cf. also Klein,
1977) and shall not even describe here how he made his
am@ v' 1
P(v~ T) = c3 e gpykr
(6) guess. However, it is very important for what follows
to give a brief account of Planck's activities from
Equation (6) contains Wien's law of 1896: October to December 1900, the heroic period in his

p(v, T) =, e '
""' for „»1,
life. It is necessary to do so for an understanding not
only of Einstein's starting point in 1905 (Sec. II.B) but
also of the subsequent reactions to the light-quantum
which is indeed correct in the quantum regime A v/AT hypothesis (Sec. IV. A).
»1, a condition which is well satisfied in Paschen's ex- Even if Planck had stopped after October 19, he would
periment (189 I) mentioned earlier, (A v/AT = 15 for T forever be remembered as the discoverer of the radia-
= 1000 'K and A. = 1p, ). Strange as it may sound, the quan- tion law. It is a true measure of his greatness that he
tum theory was discovered only after classical devia- went further. He wanted to interpret Eq. (6). That made
tions from the quantum regime had been observed in him the discoverer of the quantum theory. I shall briefly

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

870 A. Pais: Einstein and the quantum theory

outline the three steps he took (Planck, 1900). fine P~~ to be the number of ways in which the I' indis-
(1) The electromagnetic step. This concerns a result tinguishable energy elements can be distributed over N
which Planck (1900c) had obtained some time earlier distinguishable oscillators. Example: for N = 2, I' = 3
Consider a Linear oscillator with mass m and charge e the partitions are (3&, 0); (2e, s); (e, 2c); (0, 3c). In
in interaction with a monochromatic periodic electric general
field (with frequency cu) in the direction of its motion. —1+ P)!
(N
The equation of motion is (as Planck was the first to (12)
show)
P! (N —1)!
mx+fx —, 3c
2
x —eEcos2m ut.
Insert this in S~ = kin W„, use P/N
ply the Stirling approximation.
= U/c, S„=NS and ap-
This gives

I.et denote the frequency of the free oscillator, f/rn

v S=k — ——
1+ U ln 1+ U —U ln—
E c c
U
=(2mv)'. Consider in particular the case in which the
radiation damping due to the x term is very small, that It follows from Eqs. (4), (10), and T dS =dU that S is a
is, y«v where y= 8m'e' v'/3mc'. Then one ma. y approxi- function of U/v only. Therefore
mate'x" by —(2mv)'x. The solution of Eq. (8) can be
written" as x=Ccos(2m&et —o). One can readily solve (14)
for C and n. The energy. E of the oscillator equals Thus one recovers Eq. (11). And that is how the quan-
m(2m v)'C'/2 and one finds tum theory was born. This derivation was first pre-
] sented (1900b) on December 14, 1900.
E- 2m 4m (v —(u)'+ y'
(9) From the point of view of physics in 1900 the logic of
Planck's eLectromagnetic and thermodynamic steps was
Next let the electric field consist of an incoherent iso- impeccable, but his statistical step was wild. The lat-
tropic superposition of frequencies in thermal equili- ter was clearly designed to argue backwards from Eqs.
brium at temperature T. In that case the equilibrium en- (12) —(14) to Eq. (11). In 1931 Planck referred to it as
ergy U of the oscillator is obtained by replacing the "an act of desperation. . . I had to obtain a positive re-
electric field energy density &'/2 in Eq. (9) by sult, under any circumstances and at whatever cost"
47tp(&u, T)du/3 and integrating over cu: (Hermann, 1969, p. 32). Actually there were two des-
4me' perate acts rather than one. First there was his un-
p((u, T)d~
heard-of step of attaching physical significance to finite
3m J 47k (v —&d) + p
"energy elements" [Eq. (14)]. Secondly there was his
Since y is very small, the response of the oscillator is equally unheard-of counting procedure given by Eq. (12).
maximal if &u = v. Thus we may replace p(&u, T) by p(v, T) In Planck's opinion (1901a) "the electromagnetic theory
and extend the integration from -~ to +~. This yields of radiation does not provide us with any starting point

p(v, T) =,e' U(v, T), (10)

whatever to speak of such a probability [W„] in a defi-
nite sense". This statement is of course incorrect.
As will be di. scussed in Sec. II.B, the classicaL equi-
an equation to which we shall often refer in what fol- partition theorem could have given Planck a quite defi-
lows. nite method to determine all thermodynamical equili-
(2) The thermodynamic step. Planck concluded from —
brium quantities he was interested in but would not
Eq. (10) that it suffices to determine U in order to find have given him the answer he desired to derive. Planck
p. (There is a. lot more to be said about this seemingly correctly referred to Eq (12) as. a hypothesis. "Ex-
innocent statement; see Sec. II.B). Working backward perience wiLL prove whether this hypothesis is realized
from Eqs. (6) and (10) he found U. Next he determined in nature" (190la). In his papers Planck alluded to the
the entropy S of the linear oscillator by integrating inspiration he had derived from Boltzmann's prob-
T dS = QU, where T is to be taken as a function of U abilistic methods. But in Boltzmann's case the question
(for fixed v). This yields was to determine the most probable way in which a fixed
number of distinguishable gas molecules with fixed total
S= k 1+ Ln 1+ — Ln energy are distributed over cells in phase space.
(Boltzmann's method is reviewed in Sec. VI.A below. )
Equation (6) follows if one can derive Eq. (11). The correspondI. ng combinatorial problem had nothing
(3) The statistical step. I should rather say: what
Planck held to be a statistical step. Consider a large uishable objects, the energy elements. "
to do with Planck's counting of partitions of indisting-
(Planck's pro-
number N of linear oscillators, all with frequency v. cedure is actually a precursor of Bose-Einstein count-
Let U~=NU and S~=NS be the total energy and entropy ing. ) The only justification for Planck's two desperate
of the system respectively. Put S„=ALn~~, where &„ acts was that they led to the answer he wanted. His
is the thermodynamic probability. Now comes the reasoning was mad, but his madness has that divine
quantum postulate. quality which only the greatest transitional figures can
U» is supposed to be made up out of finite energy bring to science. It cast Planck, conservative by in-
elements c: U~=&c, where I' is a large number. De-

~2Planck may have found the formulae he used in Boltzmann's

For more details see a review by Pauli (1964) pp. 602 —607. work, however (Klein, 1977).

Rev. IVlod. Phys. , VoI. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 871

clination, into the role of a reluctant revolutionary. fication. "

Toward the end of this paper (Sec. X.D) I
Deeply rooted in nineteenth-century thinking and preju- shall mention the last sentence published by Einstein
dice, he made the first conceptual break that has made on scientific matters, also written in March, exactly
twentieth-century physics look so discontinuously dif- one half-century later. It also deals with the quantum
ferent from the preceding era. Although there have been theory. It has one thing in common with the opening
other major innovations in physics since December sentence mentioned above. They both express Einstein's
"
1900, the world has not seen since a figure like Planck.
Einstein venerated him.
view that the quantum theory is provisional in nature.
The persistence of this opinion of Einstein is one of the
In all the history of physics there has been no period main themes of the present paper. Whatever one may
of transition as abrupt, as unanticipated, and over as think of the status of the quantum theory in 1955, in
wide a front as the decade 1895—1905. The discovery 1905 this opinion was, of course, entirely justified.
of radioactivity, of the electron, of the quantum theory,
and of the special relativity theory are its four main
Einstein completed his first scientific paper (on capil-
themes. When special relativity appeared, it was at lary phenomena) on December 13, 1900, one day before
once "all there. "There never was an "old" theory of
Planck (1900b) presented the first paper on the quantum
special relativity. It took more than a quarter of a theory. He had finished his academic studies at the
ETH in Zurich in the preceding August. The next five
century for the quantum theory to be "all there" (at
least in its nonrelativistic aspects). It is remarkable years brought many changes in his life. To begin with
he had trouble finding a job. Then, in June 1902, he
that the old quantum theory would originate from a ques-
tion as complex as blackbody radiation. From 1859-
started work at the Patent Office in Bern and in that
1926 this problem remained at the frontier of theoreti- same month submitted his first paper (Einstein, 1902)
on the statistical foundations of thermodynamics. Ein-
cal physics, first in thermodynamics, then in electro-
stein had marriage plans to which his parents were
magnetism, then in the old quantum theory, and finally
strongly opposed, Only on his deathbed did his father
in quantum statistics.
[Hermann Einstein (184V —October 1902)] give his con-
From the experimental point of view, blackbody radia-
tion had in essence been understood by 1900. The sub-
"
sent. The marriage" took place in January 1903. His
sequent years only saw refinements of the early results.
first son was born in May 1904. By that time two more
The quality of the work by the pioneers can best be il- papers (1903, 1904) on the foundations of statistical
lustrated by the following numbers. In 1901 Planck mechanics had been submitted. Then came 1905, the
year in which Einstein received his doctor's degree in
"
(1901b) obtained from the available data the value h
=6.55&& 10 ergsec for his constant. The modern value Zurich" and in a series of papers revolutionized phys-
is 6.63' 10 ". For the Boltzmann constant he found k ics. For the purpose of the present account I shall of
course focus on the light-quantum paper (1905a) sub-
=1.34&&10 "erg K '; the present best value is 1.38
~10 ". Using his value for k he could determine
mitted March 17.
In his autobiographical notes, published in 1949, Ein-
Avogadro's number N from the relation R =Nk, where R
is the gas constant. Then from Faraday's law for univa- stein recalled that his concern with Planck's pioneering
work began shortly after 1900. "All my attempts. . .
lent ele tcr loytes:F = N ehe obtained (Planck, 1901a) the
value e =4.69& 10 ' esu. The present best value is
'
to adapt the theoretical foundations of physics to this
4. 80& 10 ' . At the time of Planck's determination of [new type of] knowledge failed completely. It was as if
the ground had been pulled out from under one, with no
e, J. J. Thomson (1899) had measured the charge of the
electron with the result e = 6. 5x 10 'o! Not until 1908, firm foundation to be seen anywhere . . . " (Einstein,
1949a). What was the meaning of Planck's derivation of
&&10 "
when the charge of the o. particle was found to be 9.3
(Rutherford and Geiger, 1908), was it realized Eq. (8)'? "The imperfections of [that derivation] re-
mained at first hidden, which was most fortunate for
how good Planck's value for e was.
the development of physics" (1949a). Einstein's 1905
Nothing further happened in quantum physics after
1901 until Einstein proposed the light-quantum hypo- paper opens with a section entitled "On a difficulty con-
thesis. cerning the theory of blackbody radiation" in which he
put these imperfections in sharp focus.
B. Einstein on Planck: 1905. The Rayleigh-Einstein-Jeans His very simple argument was based on two solid con-
law
The first sentence on the quantum theory published by I learned these and other details of Einstein's life from a
biographical sketch written in 1924 by his sister Maja %'inteler-
Einstein was written in the month of March, in the year Einstein {1881—1951). This biography is in the Einstein Ar-
1905. It is the title of his first paper on light-quanta chives i.n Princeton.
(1905a): "On a heuristic point of view concerning the 5To Mileva Marie (1876-1948). They had two sons, Hans
generation and conversion of light. "
Webster's Diction- Albert (1904—1973) and Eduard (1910—1965). (There were two
ary contains the following definition of the term heuris- sons from Hans Albert's first marriage and also one adopted
tic: "Providing aid and direction in the solution of a daughter. The younger son died at age five. ) Einstein and his
problem but otherwise unjustified or incapable of justi- wife separated in 1914 and divorced in February 1919. He re-
married in June 1919. His second wife, Eisa Einstein Lowen-
thal (1874—1936) was a widowed cousin of his. She had two
daughters, Ilse and Margot, by her first marriage.
~3In September 1918 Einstein proposed Planck for the Nobel ~8His thesis was on "A new determination of the dimensions
Prize for 1919, too late as it happened, since Planck received of rnolecules. "It was also published in Ann. Phys. (Leipz. ) (Ein-
it later in 19%8. stein, 1906a).

Rev. Mod. Phys. , VoI. 5't, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

sequences of classical theory. The first of these was off factor and proposed the overall radiation law
Planck's Eq. (10). Recall that the quantity U in that '"~
equation is the equilibrium energy of a one-dimensional p(v, T) = c, v'Te (17)
harmonic oscillator. Einstein's second ingredient was This expression became known as the Rayleigh law.
the equipartition law of classical statistical mechanics Already in 1900, Rubens and Kurlbaum (and also Lum-
according to which mer and Pringsheim) found this law wanting, as is seen
in Fig. 2.
U(v, T) = (R/N)T, Thus the experimentalists close to Planck were well
aware of Bayleigh's work. One wonders whether or not
where It,'is the gas constant, and N Avogadro's number.
Planck himself knew of this important paper which ap-
R/N= k, the Boltzmann constant. (For a number
peared half a year before he proposed his own law.
of years Einstein did not use the symbol k in his papers. )
From Eqs. (10) and (15) Einstein obtained
to Bayleigh's contribution. "
Whichever may be the case, in 1900 Planck did not refer

p(v,
v' A
T)= 87t~ T— (16)
(ii) Einstein gives the derivation of Eq. (16) discussed
previously. His paper is submitted March 17, 1905 and
appears June 9 of that year.
and went on to note that this classical relation is in (iii) In a letter to Nature (submitted May 6, published
disagreement with experiment and has the disastrous May 18, all in 1905) Rayleigh (1905a) returns to his v'T
consequence that a =~, where a is the Stefan-Boltzmann law and now computes c, . His answer for c, is off by a
constant given in Eq. (3). factor 8.
"If Planck had drawn this conclusion, he would prob- (iv) On June 7, 1905, James Hopwood Jeans (1877—
ably not have made his great discovery" Einstein said 1946) adds a postscript to a completed paper in which he
later (1949a). Planck had obtained Eq. (10) in 1897. At corrects Bayleigh's oversight. The paper appears a
that time the equipartition law had been known for al- month later (Jeans, 1905a). In July 1905 Rayleigh
most thirty years. (See also Sec. III.) During the 1890s (1905b) acknowledges Jeans' contribution.
Planck had made several errors in reasoning before he It follows from this chronology (not that it matters
arrived at his radiation law, but none as astounding and much) that the Rayleigh-Jeans law ought properly to
of as great a historical significance as his fortunate be called the Bayleigh-Einstein- Jeans law.
failure to be the first to derive Eq. (16). This omission The purpose of this digression about Eq. (16) is not
is, no doubt, related to Planck's decidedly negative merely to note who said what first. Of far greater in-
attitude (before 1900) to Boltzmann's ideas on statistical terest is the role which this equation played in the early
mechanics. reactions to the quantum theory. From 1900-1905
Equation (16), commonly known as the Rayleigh- Jeans Planck's radiation formula was generally considered to
law, has an interesting and rather hilarious history, as be neither more nor less than a successful representa-
may be seen from the following chronology of events. tion of the data (cf. Benz, 1975). Only in 1905 did it
(i) In June 1900 there appeared a brief paper (1900a) begin to dawn, and then only on a few, that a crisis in
by Lord Rayleigh (1842-1919). It contains for the first physics was at hand (Einstein, 1913). The failure of the
time the suggestion to apply to radiation "the Maxwell- Bayleigh-Einstein- Jeans law was the cause for this
Boltzmann doctrine of the partition of energy" (i. e. , the turn of events.
equipartition theorem). From this doctrine Rayleigh Rayleigh's position on the fa, ilure of Eq. (16) as a uni-
went on to derive the relation p = c, v'T but did not evalu- versal law was that "we must admit the failure of the
ate the constant cy It should be stressed that Rayleigh's law of equipartition in these extreme cases" [i.e. , at
derivation of this result had the distinct advantage of high frequencies] (1905a). Jeans (1905b) took a, different
dispensing with the material oscillators altogether ~" view: The equipartition law is correct but "the supposi-
Rayleigh also realized that this relation should be in- tion that the energy of the ether is in equilibrium with
terpreted as a limiting law: "The suggestion is then that of matter is utterly erroneous in the case of ether
that [p=c, v'T], rather than [Wien's law Eq. (5)], may
be the proper form when [T/v] is graf" (my italics). " vibrations of short wavelength under experimental con-
ditions". Thus Jeans considered Planck's constant h
In order to suppress the catastrophic high-frequency as a phenomenological parameter well suited as a help
behavior, he introduced next an ad hoc exponential cut- to fit data but devoid of fundamental significance. The
—
debate nonequilibrium versus failure of equipartition-
continued for a number of years (Hermann, 1969). The
~~Planck derived his radiation law in a circuitous way via the issue was still raised at the first Solvay Congress in
equilibrium properties of his material. oscillators. He did so 1911, but by then the nonequilibrium view no longer
because of his simultaneous concern with two questions: How aroused much interest.
is radiative equilibrium established~ What is the equilibrium
distribution~ The introduction of the material oscillators
would, Planck hoped, show the way to answer both questions.
Bayleigh wisely concentrated on the second question only. He Neither did Lorentz, who in 1903 gave still another deriva-
considered a cavity filled with "aetherial oscillators" assumed tion of the p2T law (Lorentz, 1903). The details need not con-
to be in equilibrium. This enabled him to apply equipartition cern us. It should be noted that Lorentz also gave the correct
directly to these radiation oscillators. answer for the constant c~. However, he did not derive the ex-
This same observation w'as also made independently by pression for c~ directly. Bather he found c& by appealing to the
Einstein in 1905. long-wavelength limit of Planck's law.

Rev. Mod. Phys. , Vot. 51, No. 4, October 1979

 , A. Pais: Einstein and the quantum theory 873

C. The light-quantum hypothesis and the heuristic sity)

principle.
sg 1
(20)
Let us return to Einstein. In 1905 his position was as p
follows. Equation (6) agrees with experiment but not Assume that Wien's law fEq. (5)] is applicable. Then
with existing theory. Equation (16) agrees with existing
theory but not with experiment. He therefore set out to
study blackbody radiation in a way "which is not based
(ln, — .
1) (21)
on a picture of the generation and propagation of radia-
—
tion" that is, by not making use of Planck's Eq. (10). Let the radiation be contained in a volume v.
S(v, v, T) = @v d v and E(v, v, T) = pod v are the total en-
But then something had to be found to replace that equa-
tion. For this purpose Einstein chose to reason "im tropy and energy in that volume in the interval v to
Anschluss an die Erfahrung, " "in connection with ex- v+Qv, respectively. In the Wien regime S follows
periment. "
The experimental information he made use trivially from Eq. (21). One finds in particular that
of was the validity of Wien's law fEq. (5)] in the region
of large v/T. He extracted the light-quantum postulate S(v, v, T) —S(v, v„T) =
Pv
ln — —ln-
v
=
NE /R Bv

from this law by drawing an analogy between radiation

in the Wien regime and a gas of (classical) noninteract- (22)
ing point particles, often called an ideal Boltzmann
Compare Eqs. (22) and (19) and we have Einstein's
gas. Specifically he made use of the volume dependence L, ight-quantum hypothesis. Monochromatic radiation
of the entropy for such a gas. This well known relation
of low density [(i.e. , ) within the domain of validity of
between entropy and volume can be derived in the fol-
the Wien radiation formula] behaves in thermodynamic
lowing way from the second law of thermodynamics and
the ideal gas law.
respect as if it consists of mutually independent energy
quanta of magnitude RPv/N" (P = h/k, R/N= k, RP v/N
Consider a system in equilibrium described by its = h v).
pressure P, volume v, and temperatur e T . According
It is astonishingly simple. Einstein phrased this hypo-
to the second law an infinitesimal reversible change of
thesis in the form of a theorem. It was a hypothesis
such a system obeys the relation
nevertheless, since it was based on Mien's law which
BU
itself still needed proof from first principles.
T 4f~=c„QT+ +p 6v ~ Einstein (1905a) made next a daring step which was
Bv
to gain him the Nobel prize in 1921.
c„ the specific heat, 8the entropy, and U the internal The heuristic pxinciple. "If, in regard to the volume
energy all are in general functions of v and T. From dependence of the entropy, monochromatic radiation
(of sufficiently low density) behaves as a discrete rned-
(s S/s v) (s S/s T) ium consisting of energy quanta of magnitude RPv/N,
BT Bv then it is suggestive to inquire whether also the laws of
the generation and conversion of light are constituted
and from Eq. (18) it follows that as if light were to consist of energy quanta of this kind. "
(~+»/») In other words: The light-quantum hypothesis is an
T assertion about a quantum property of free electromag-
netic radiation; the heuristic principle is a tentative
For an ideal Boltzmann gas this last relation reduces to
BU/Sv = 0 since in this case NPv =nRT, where n is the between light and matter. "
extension of these properties of light to the interaction

"
number of molecules in the gas. In turn SU/Sv =0 im-
plies that c„ is a function of T only, hence T dS(v, T)
I shall continue next with a discussion of certain mat-
ters of principle contained in Einstein's papers of 1904,
= c„(T) dT +nRT dv/Nv. Consider a finite reversible 1905, and 1906. In part F of this section I shall turn
change at constant T in which the gas molecules in the to the application of the heuristic principle to the photo
volume vo are confined to a subvolume v. Then the en- electric effect.
tropy change for an ideal Boltzmann gas is given by
D. Three remarks on Einstein's statistical physics
S(v, T) -S(v„T) = —ln—
v
(1) Einstein's contributions to physics fall under
0 three main headings: statistical physics, quantum
Einstein derived Eq. (19) by a different method on which theory, and relativity theory. His researches in sta-
I shall comment in the next subsection D. tistical physics and on the quantum theory are strongly
Now back to the radiation problem. Let Q(v, T)dv be the
interrelated. This will become progressively clear in
the further discussion of Einstein's work in the years
entropy density per unit volume in the frequency interval
between v and v+dv. Then (p is again the spectral den-
The need for a distinction between the light-quantum hypo-
thesis and the heuristic principle was emphasized to me by
Actually c„does not depend on T either for an ideal Boltz- R. Jost. I am grateful to him for an illuminating discussion on
mann gas, but we do not need this fact for the present reason- this subject. See also Jost's own contribution (1979) to the
ing. Einstein Centennial meeting held in Zurich, February 24, 1979.

Re~. Mod. Phys. , Vol. 51, No. 4, October 1979

874 A. Pais: Einstein and the quantum theory,

1905-1925. Moreover, when we come to Einstein's role few remarks if he had been aware of Gibbs's book.
as a transitional figure (Sec. VII) it will become evident However, in 1909 Einstein (1909b) gave a new deriva-
that wave mechanics is an offspring of statistical phys- tion of Eq. (24), this one all his own. It is characteris-
ics in a sense to be described. tic that his statistical argument would appear in a paper
As Klein has noted, the first intimations of this inter- principally devoted to the quantum theory.
play between statistical methods and quantum arguments Einstein argued as follows. Consider a large system
in Einstein's work are found already before 1905. For with volume V in equilibrium at temperature T. Divide
details on the early influences of thermodynamics on V into a small subvolume Vp and a rest V„V = V, + V„
Einstein's thought the reader is referred to Klein's Vp +& Vy The f ixed total energy is likewise divided,
papers (1967, 1979). Here I shall confine myself to a E Ep + E1 A ssume that also the entropy is additive:
few brief remarks on this subject.
S =Sp+$ (2&)
The first of these concerns the following question.
What led Einstein to combine %'ien's law with of all — Suppose that E„E, deviate by amounts aE„aE, from
things —
the volume dependence of thermodynamic quan- their respective equilibrium values. Then
tities? I have no firm answer, but it seems relevant BS ~ 1 B'S
to note that volume dependence also plays a role in the &S= BS
first published remarks by Einstein on radiation in 1

—,
p~

thermal equilibrium. These date from 1904 and ap- 1 S

(aE,)'+. . .
B
+ (26)
peared in his early work on fluctuation phenomena 2 BE1
(1904) .
where the expressions in [ ] refer to equilibrium val-
Consider a system which may have various energies
ues. The first-or der terms cancel since AE p AEy
Ey E2 in ther mal equil ibrium with a large s e cond
(energy conservation) and [BS,/BE, ] = [BS,/BE, ] (equili-
~

system at temperature T. The equilibrium energy of brium). Furthermore [B'S,/BE,'] = —I/caT' and
the first system is given by
[B'S,/BE,'] = —1/c, T', where c„c„are the respective
-BE'i
E heat capacities at constant volume. cy++cp since V,
, P= I/uT. (23) » V, . Thus Eq. (26) becomes
-BE i
In 1904 Einstein deduced a formula for the mean square (aE, )'
energy fluctuation(e') =—((E —(E))') =(E') —(E)' of the zS = zS,P = —--
2c A@2
(27)
first system, namely
Next Einstein applied the relation S, = klnlVp to the sub-
B(E)
BP
~, B(E)
BT
(24) system and reinterpreted this equation to mean that 8'p
is the probability for the subsystem to have the entropy
He then introduced a criterion for fluctuations to be S, (at a given time). Hence
large:
W p
=8'
Q (28)
«)' (2&)
where W, is the equilibrium value of W, . Equations (27)
This relation is not satisfied by an ideal Boltzmann gas and (28) show that W, is Gaussian in AE, . Denote (as
under normal conditions since (E) =nkT/2 so that g before) the mean square deviation of this distribution by
=0(n '), independent of the volume. Einstein went on (s'). Then (s') = kc,T' which is again Eq. (24).
to note that $ can be of order unity for one kind of sys- This derivation is typical for Einstein's approach.
tem: blackbody radiation. In that case (E) =avT, [Eq. Instead of reasoning from the microscopic to the macro-
(3)], hence g =4k/avT'. T is proportional to the inverse scopic (W —S) he often argued in the inverse direction
of X „, the wavelength at which the spectral distribution
reaches its maximum. Thus volume dependence is im-
(and did so to great advantage). In Sec. II.G I shall note
that this last derivation was briefly challenged at a later
',
portant: for given T, $ can get la, rge if A. „/v is la. rge, time.
hence if v is small. (2) The fact that Einstein devoted two and a half pages
In 1904 Einstein was mainly interested in Eq. (24) be- of his light-quantum paper to a derivation of Eq. (22)
cause of the new vistas it opened for experimental de- from a molecular statistical point of view illustrates
terminations of k (and therefore of Avogadro snu'mber again how strongly statistical and quantum theory are
N) in model-independent ways. These interests were interwovenin Einstein's work. It also demonstrates that
to branch in the following year, 1905, when the macro-
scopic fluctuations typical for Brownian motion (Ein- young discipline. "
in those days statistical mechanics was still a quite
Earlier I gave an essentially thermo-
stein, 1905b) gave him N while volume dependences led dynamic proof of Eq. (22) in order not to divert attention
him to light-quanta. from the arguments which led Einstein to the light-
Equation (24) plays an important role in Einstein's quantum. Let me now briefly sketch Einstein's own de-
papers of 1909, 1917, and 1925 on the quantum theory. r ivation.
When Einstein first derived this equation, he did not
know that Josiah Wiliard Gibbs (1839-1903) had done
so before him (Gibbs, 1902). Some years later Einstein In 1910, Lorentz gave an instructive comparison of the
wrote (1911a) that he would have reduced his early statistical methods of Boltzmann, Gibbs, and Einstein (see
published papers on the molecular theory of heat to a Lorentz, 1927).

Rev. Mod. Phys. , Vo(. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

Einstein started from Boltzmann's relation nection between Planek's quantization related to U and
Einstein's quantization related to p.
S= —
A
lnW + const . (29) Einstein's answer: This is indeed possible namely
by introducing a, new assumPtion: Equation (10) is also
according to which a reversible change from a state a valid in the quantum theoryl Thus Einstein proposed
to a state b satisfies to trust Eq. (10) even though its theoretical foundation
had become a mystery when quantum effects are im-
A —
N'
S —S = ln W' ' (30) portant. Einstein then re-examined the derivation of
Planck's law with the help of this new assumption. I
Let the system consist of subsystems 1, 2, . . . which do omit the details and only state his conclusion. "We
not interact and therefore are statistically independent. must consider the following theorem to be the basis of
Then Planck's radiation theory: the energy of a [Planck os-
W =W~W2. .. cillator] can only take on values which are integral mul-
tiples of k v; in emission and absorption the energy of
a a
W~ W2 a [Planck oscillator] changes by jumps which are multi-
"
ples of A. v. Thus already in 1906 Einstein had guessed
For the case of an ideal Boltzmann gas, the subsystems correctly the main properties of a quantum-mechanical
oscillator and its behavior in radiative transitions. We
may be taken to be the individual molecules. Let the
shall see later that Planek was not at all prepared to
gas in the states a and b have volume and temperature
accept at once Einstein's reasoning, in spite of the fact
(v, T) and (vo, T) respectively. Then W;(v)/R';(vo)
that it lent support to his own endeavors. As to Ein-
=u/vo for all i Th.e n molecules are statistically inde-
stein himself, his acceptance of Planck's Eq. (10), al-
pendent, so that
beit as a hypothesis, led to a major advance in his own
work: The quantum theory of specific heats, to be dis-
(32)
cussed in Sec. III.
Equations (30) and (31) again give Eq. (19)." F. The photoelectric effect: The second coming of h.
(3) Einstein s introduction of light-quanta in the Wien
regime is the first step towards the concept of radiation The most widely remembered part of Einstein's 1905
as a Bose gas of photons. From ihe analogy made be- paper on the quantum theory deals with his interpreta-
tween Eqs. (19) and (22) it follows that Einstein's energy tion of the photoelectric effect. The present discussion
quanta are statistically independent in the Wien region, of this subject is organized as follows. After a few gen-
see Eq. (32). The photon gas is described (for all fre- eral remarks I first sketch its history from 188V to
quencies) by Bose statistics. In that description Eq. 1905. Then I turn to Einstein's contribution. Finally I
(32) does not hold in general. I shall comment in Sec. outline the developments up to 191 by which time Ein-
VI. C on the equivalence of Boltzmann and Bose statis- stein's predictions were confirmed.
tics in the Wien regime. These days photoelectron spectroscopy is a giant field
of research with its own journals. Gases, liquids, and
E. Einstein on Planck: 1906. solids are being investigated. Applications range from
In 1906 Einstein returned once more to Planck's theory
solid state physics to biology. The field has split into
of 1900. Now he hgd much more positive things to say subdisciplines such as the spectroscopy in the ultra-
about Planck's radiation law. This change in attitude
violet (UPS) and in the X-ray region (XPS). In 1905,
was due to his realization that "Planck's theory makes
however, the subject was still in its infancy. We have
implicit use of the . . . light-quantum hypothesis" (Ein-
a detailed picture of the status of photoelectricity a few
months before Einstein finished his paper on light-
stein, 1906b). Einstein's reconsideration of Planck's
quanta: In December 1904 the first review article on
reasoning and of its relation to his own work can be
this topic was completed (von Schweidler19, 04). We
summarized in the following way:
infer from it that at that time photoelectricity was as
(1) Planck had used the p —U relation, Eq. (10), which much a frontier subject as were radioactivity, cathode
follows from classical mechanics and electrodynamics. ray physics, and (to a slightly lesser extent) the study
(2) Planck had introduced a quantization related to U, of Hertzian waves.
namely the prescription U =Ph v/N [see Eqs. (11)-(14)]. In 1905 the status of experimental techniques was still
(3) If one accepts step 2, which is alien to classical -
rudimentary in all these areas, yet in each of them,
theory, then one has no reason to trust Eq. (10) which initial discoveries of great importance had already been
is an orthodox consequence of classical theory. made. Not surprisingly, an experi. mentalist, mainly
(4) Einstein had introduced a quantization related to active in one of these areas, would also apply himself
p: the light-quantum hypothesis. In doing so he had to some of the others. Thus Heinrich Hertz (1857-1894),
not used the p —U relation (10). the first one to observe a photoelectric phenomenon, '
(5) The question arises if one can establish a con- made this discovery at about the same time he demon-
strated the electromagnetic nature of light. The high

Einstein noted that Eqs. (18) and (32) together with g =0 lnR'
yield the ideal gas law. 24I consider only the so-called external photoelectric effects.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

876 A. Pais: Einstein and the quantum theory

school teachers Julius Elster (1854 —1920) and Hans separately for cathode rays. Nom he saw his way clear
Geitel (1855-1923) pioneered the study of photoelectric to do this for photoelectrons. His second conclusion:
effects in vacuum and constructed the first phototubes "e is the same in magnitude as the charge carried by
(Elster and Geitel, 1890); they also performed funda- the hydrogen atom in the electrolysis of solutions. "
mental experiments in radioactivity. Pierre Curie Thomson's method for finding e is of major interest
(1859-1906) and one of his co-workers were the first to since it is one of the earliest applications of cloudcham-
discover that photoelectric effects can be induced by x ber techniques. His student Charles Thomson Bees
rays (Curie and Sagnac, 1900). J. J. Thomson is best Wilson (1869-1959) had discovered that charged parti-
remembered for his discovery of the electron in his cles can f or m nuclei f or condensation of super saturated
study of cathode rays (Thomson, 1897). Yet perhaps mater vapor. Thomson applied this method to deter-
his finest experimental contribution deals with the mine the number of charged particles by droplet count-
photoeff ect. ing. Their total charge was determined electrometrical-
Let us nom turn to the mork by the pioneers.
(1) 2887: Hertz. Five experimental observations e (6.8x 10 "
ly. In vi. em of these technical innovations his value for
esu) must be considered as very respect-
made within the span of one decade have largely shaped able.
the physics of the twentieth century. In order of ap- (4) 1902: Lenard. In 1902 Philip Lenard (1862 —1947)
pearance they are the discoveries of the photoelectric studied the photoeffect using a carbon arc light as a
effect, x rays, radioactivity, the Zeeman effect, and the source. He could vary the intensity of his light source
electron. The first three of these were made accidental- by a factor -1000. He made the crucial discovery that
ly. Hertz found the photoeffect when he became intrigued the electron energy showed "not the slightest depen-
by a side effect which he found in the course of his in- dence on the light intensity" (Lenard, 1902). What about the
vestigations on the electromagnetic wave nature of light variation of the photoelectron energy with the light fre-
(Hertz, 1887). At one point he was studying spark dis- quency? One increases with the other, that was all that
charges generated by potential differences between two was known in 1905 (von Schweidler, 1904).
metal surfaces. A primary spark coming from one surface (5) 1905: Einstein. On the basis of his heuristic prin-
generates a secondary spark on the other. Since the lat- ciple Einstein proposed the folloming "simplest picture"
ter mas harder to see Hertz built an enclosure around for the photoeffeci. A light-quantum gives all its ener-
it to eliminate stray light. He was struck by the fact gy to a single electron. And the energy transfer by one
that th is caused a shortening of the sec ondary spark. light-quantum is independent of the presence of other
He found next that this effect was due to that part of the light-quanta. He also noted that an electron ejected
enclosure which was interposed between the two sparks. from the interior of the body will in general suffer an
It was not an electrostatic effect since it made no quali-
tative difference whether the interposed surface was a
energy loss before it reaches the surface. Let E
be the electron energy for the case that this energy
„
conductor or an insulator. Hertz began to suspect that loss is zero. Then, Einstein proposed, we have the re-
it might be due to the light given off by the primary lation (in modern notation)
spark. In a delightful series of experiments he con-
firmed his guess: light can produce sparks. For ex- E „=hv —P, (33)
ample he increased the distance between the metal sur- where v is the frequency of the incident (monochromatic)
faces until sparks ceased to be produced. Then he il- radiation, and P is the work function, the energy needed
luminated the surfaces with a nearby electric arc lamp: to escape the surface. He pointed out that Eq. (33) ex-
the sparks reappeared. He also came to the (not quite plains Lenard's observation of the light intensity inde-
correct) conclusion that ".
. . If the observed pheno- pendence of the electron energy.
menon is indeed an action of light, then it is only one of Equation (33) represents the second coming of h. This
ultraviolet light. " equation made very strong predictions. First, E should
(2) 1888: Hallzoachs. Stimulated by Hertz's work, vary linearly with v. Secondly, the slope of the (E, v)
Wilhelm Hallwachs (1859—1922) showed next that irradi- plot is a universal constant, independent of the nature
ation with ultraviolet light causes uncharged metallic of the irradiated material. Thirdly, the value of the
bodies toacquireapositivecharge (Hallwachs, 1888). slope was predicted to be I'lanck's constant determined
The earliest speculations on the nature of the effect from the radiation law.
predate the discovery of the electron in 1897. It was Einstein gave several other applications of his heu-
suggested in 1889 that ultraviolet light might cause ristic principle. (1) The frequency of light in photo-
specks of metallic dust to leave the metal surface (Len- luminescence cannot exceed the frequency of the inci-
ard and Wolf, 1889). dent light (Stokes' rule). (2) In photoionization the en-
(3) &899: J; &. Thomson. Thomson (1899) was the
first to state that the photoeffect induced by ultraviolet is the incident light frequency. "
ergy' of the emitted electron cannot exceed h v, where v
These two statements
light consists of the emission of electrons. He began were made in 1905 (Einstein, 1905a). (3) In 1906 he
his photoelectric studies by measuring the e/m of the discussed the application to the inverse photoeffect
particles produced by light, using the same method he (Volta effect) (Einstein, 1906b). (4) In 1909 he treated
had applied to cathode rays two years earlier (the parti-
cle beams move through crossed electric and magnetic
fields). His conclusion: "The value of m/e in the case In 1912 Einstein (1912a, b, c) noted that the heuristic princi-
of ultraviolet light. . . . is the same as for cathode ple could be applied not only to photoionization but also in a
"
rays. In 1897 he had been unable to determine m or e quite similar way to photochemical processes.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 877

the generation of secondary cathode rays by X-rays of Geneva). In September he attendedhis first physics
(Einstein, 1909a). (5) In 1911 he used the principle to conference (which took place in Salzburg). In October
predict the high-frequency limit in Bremsstrahlung he moved to Zurich, to start work as Extraordinarius
(Einstein, 1912d, p. 443). (associate professor) at the University. In those days
(6) 1915: Millikan; tive Duane-Hunt limit. In 1909 he was already much involved in the problem of how to
a second review paper on the photoeffect appeared generalize the special theory of relativity. Yet his in-
(Ladenburg, 1909). We learn from it that experiments tense preoccupation with the quantum problems con-
were in progress to find the frequency dependence of tinued. In 190S he had written to a friend: "I am in-
E „. but that no definite conclusions could be drawn as cessantly, busy with the question of the constitution of
yet. Among the results obtained during the next few radiation. . . . This quantum question is so uncommonly
years those of Arthur Llewellyn Hughes (1883-1978), important and difficult that it should concern everyone"
J. J. Thomson's last student, are of particular interest. (Einstein, 1908). I now turn to the discussion of two
Hughes found a linear R —v relation and a value for the profound papers on the quantum theory which Einstein
slope parameter which varied from 4. 9-5.7x 10 de-", published in 1909. The first one (1909b) was completed
pending on the nature of the irradiated material (Hughes, in January. The second one (1909a) was presented to
1912). These and other results were critically reviewed in the Salzburg conference in October.
1913and technical reservations about Hughes' results were In 1905 Einstein had used the %ien law although it had
expressed (Pohl and Pringsheim, 1913). However, soon not yet acquired a firm theoretical foundation. In 1909
thereafter Jeans stated in his important survey of the he did the same with Planck's law. In earlier days he
theory of radiation (Jeans, 1914) that "there is almost had never mentioned doubts about the experimental
general agreement" that Eq. (33) holds true. Opinions vajidity of Planck's expression for p(v, T) But e had h.

formula(24)
were divided, yet evidently experimentalists were be- never expressed himself more firmly in favor of ac-
ginning to close in on the Einstein relation. cepting this law than in his talk at Salzburg: "One can-
In the meantime, in his laboratory at the University not think of refusing [to accept] Planck's theory. " In
of Chicago, Millikan had already been at work on this the next sentence he gave a new reason for this con-
problem for several years. He used visible light (a set viction. ' Geiger and Rutherford's value for the electric
of lines in the mercury spectrum); various alkali met- charge had been published and Planck's value for e had
been "brilliantly confirmed. " (See the discussion at
fluctuation
als served as targets. (These are photosensitive up to
-0.6p). On April 24, 1914 and again on April 24, 1915 the end of Sec. II.A).
he reported on the progress of his results at meetings In his January paper Einstein gave the derivation
of the American Physical Society (Millikan, 1914, 1915). mentioned in Sec. II. D of the and
A long paper (Millikan, 1916b) published in 1916 gives applied this result to energy fluctuations of blackbody
the details of the experiments and a summary of his radiation in a frequency interval between v and v+dv.
beautiful results: Eq. (33) holds very well and" Planck's In order to understand how this refinement is made,
k has been photoelectrically determined with a pre- consider a small subvolume v of a cavity filled with
cision of about 0. 5' and is found to have the value thermal radiation. Enclose v with a wall which prevents
h = 6. 57x 10 all frequencies but those in dv from leaving e while
Also the Volta effect confirmed the heuristic princi- those in dv can freely leave and enter v. %e may then
ple. This evidence came from x-ray experiments per- apply Eq. (24) with (E) replaced by pvdv. (e2) is
formed in 1915 at Harvard by William Duane (1872- now a function of v and T, and we have
1935) and his assistant Franklin Hunt. (Duane was one
of the first biophysicists in America, his interest in (s '(v, T)) = kT'vd v(a p/a T) . (34)
x rays was due largely to the role they play in cancer
therapy. ) Working with an x-ray tube which is operated This equation expresses the energy fluctuations in terms
at a constant Potential V they found that the x-ray fre- of the spectral function p in a way which is independent
quencies produced have a sharp upper limit v given by of the detailed form of p. Consider now the following
eV= hv, as was predicted by Einstein in 1906. This three cases.
limiting frequency is now called the Duane-Hunt limit. (1) p is given by the Rayleigh-Einstein- Jeans law Eq.
They also obtained the respectable value &= 6.39x 10 ". (16). Then
In Sec. I, I have already mentioned some of Millikan's 3
reactions to these developments Du. ane and Hunt (1915) (s'(v, T)) = , p vdv. (35)
did not quote Einstein at all in their paper. In Sec. IV.A
I shall discuss further the impact of these discoveries (2) p is given by the Wien law Eq. (5). Then
on the acceptance of Einstein's ideas.
(c'(v, T)) = h vpv dv .
G. The fusion of particles and waves and Einstein's destiny
(3) p is given by the Planck law Eq. (6). Then
1909 was another eventful year in Einstein's life. He
had started his academic career as a Privatdozent in
Bern and gave his first lectures there in the winter term (37)
1908-1909. His subject: the theory of radiation. On
July 6 he left the Patent Office. He received his first (I need not apologize for having used the same symbol p
honorary degree in that same month (from the University in the last three equations even though p is a different

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

function of v and T in each of them. ") change be fully explained on the grounds of his scien-
In his discussion of Eq. (37) Einstein stressed that tific philosophy? It i. s too early to tell, but I doubt it.
"the current theory of radiation is incompatible with
this result. " By "current theory" he meant, of course,
Exactly half a century had passed since Kirchhoff
stated that there had to be a blackbody radiation law.
the classical wave theory of light. Indeed, the classical
The law had been found. A small number of physicists
theory would only give the second term in Eq. (37), the had realized that its implications were momentous. A
"wave term" [compare Eqs. (37) and (35)]. About the
proof of the law did not yet exist.
first term of Eq. (37) Einstein had this to say: "If it At this point I need to interrupt the account of the his-
alone were present, it would result in fluctuations [to
tory of radiation theory in order to describe Einstein's
be expected] if radiation were to consist of indepen-
contributions to the theory of specific heat. I shall re-
dently moving pointlike quanta with energy h v. " In turn to radiation in Sec. IV, which contains another im-
other words, compare Eqs. (36) and (37). The former
portant result found in Einstein's 1909 papers. As a
corresponds to Mien's law which in turn holds in the postscript to the present section I add a brief remark
regime in which Einstein had introduced the light-quan- on Einstein's energy fluctuation-formula.
tum postulate.
Equations (35) —(37) were derived by a statistical
Observe the appearance of a new element in this last
reasoning. One should also be able to derive them in a
statement by Einstein. The word "pointlike" occurs.
directly dynamical way. Einstein himself had given
Although he did not use the term, he now was clearly
qualitative arguments for the case of Eq. (35). He noted
thinking of quanta as particles. His own way of re-
that the fluctuations come about by interference between
ferring to the particle aspect of light was to call it "the
point of view of the Newtonian emission theory.
" Eq. waves with frequencies within and outside the dv inter-
val. A few years later Lorentz (1916) gave the detailed
(37) suggests (loosely speaking) that the particle and calculation, obtaining Eq. (35) from classical electro-
wave aspects of radiation occur side by side. This is
magnetic theory. However, dif f iculties arose when it
one of the arguments which led Einstein in 1909 to was attempted to derive the Planck case Eq. (37) dy-
summarize his view on the status of the radiation namically. These were noted in 1919 by Leonard Salo-
theory in the following way. '7 mon Ornstein (1880—1941) and Frits Zernike (1888-
"I already attempted earlier to shove that our current
1966), two Dutch experts on statistical physics. The
foundations of the radiation theory have to be aban
problem was further elaborated (1925) by Paul Ehren-
doned. . . it is my opinion that the next Phase in the fest (1880—1933). It was known at that time that one
development of theoretical physics will bring us a theory
can obtain Planck's expression for p by introducing the
of light which can be interpreted as a kind of fusion of quantum prescription" that the electromagnetic field
the wave and the emission theory . . . [the] wave struc
oscillators could only have energies nh v. However,
ture and [the] quantum structure . . . are not to be con-
the mentioned authors found that the same prescription
sidered as mutually incompatible. . . it seems to follow
applied to the fluctuation formula gave the wrong ans-
from the Jeans law [Eq. 26] that we wi ll have to mod-
wer. The source of the trouble seemed to lie in Ein-
ify our current theories, not to abandon them com-
pletely. "
stein's entropy additivity assumption, Eq. (25). Accord
ing to Uhlenbeck (private communication) these dis-
This fusion now goes by the name of complementarity.
crepancies were for some years considered to be a
The reference to the Jeans law we would now call pn serious problem. In their joint 1925 paper, Born,
application of the correspondence principle. Heisenberg, and Jordan refer to it as a fundamental
The extraordinary significance for twentieth century
difficulty (Born, Heisenberg, and Jordan 1925). In that
physics of Einstein's summation hardly needs to be same paper it was shown, however, that the new quantum
stressed. I also see it as highly meaningful in relation mechanics applied to a set of noninteracting oscillators
to the destiny of Einstein the scientist if not of Einstein
does give the Einstein answer. The noncommutatvity of
the man. In 1909, at age 30, he was prepared for a
coordinates and momenta plays a role in this derviation.
fusion theory. He was alone in this. Planck certainly
Again according to Uhlenbeck (private communication)
did not support this vision. Bohr had yet to arrive on
the elimination of this difficulty was considered as one
the scene. Yet when the fusion theory arr ived, Einstein
of the early successes of quantum mechanics. (It is not
could not accept the duality of particles and waves in-
necessary for our purposes to discuss subsequent im-
herent in that theory, quantum mechanics, as being
provements of the Heisenberg-Born-Jordan treatment. ")
fundamental and irrevocable. It may have distressed
him that one statement he made in 1909 needed re-
vision: moving light-quanta with energy h v are not II I. EINSTEIN AND SPECIF lC HEATS
pointlike. In the later parts of this paper I shall have A. Specific heats in the nineteenth century.
to make a number of comments on the scientific reasons
which changed Einstein's apartness from a figure far By the end of the first decade of the twentieth century
ahead of his time to a figure on the sidelines. Can this
three major quantum theoretical discoveries had been

The elementary derivation due to Debye in 1910 is found in

Equations (35) and (36) do not explicitly occur in Einstein's Sec. VII. C.
own paper. 2
The reader interested in these further developments is re-
2~In the following quotation l combine statements made in the ferred to a paper by Gonzalez and Wergeland (1973) which also
January and in the October paper. The italics are mine. contains additional references to this subject.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

Pierre Louis Dulong (1V85-1838) and Alexis Therhse

0 O Petit (1791-1820) made ari unexpected discovery during
W ~ aP the researches in thermometry on which they had been
~~ ~

/ jointly engaged for a number of years. For a dozen

/
/'
metals and for sulphur (all at room temperature) they
8/ found that c, the specific heat Per gram atom (briefly
/
referred to as the specific heat hereafter) had prac-
/
tically the same value, approximately 6 cal/mole deg
s (Petit and Dulong, 1819). They did of course not regard
I
CY
this as a mere coincidence: "One is allowed to infer
e/I
Op
i [froin these data] the following law: The atoms of all
/
0/ simple bodies [elements] have exactly the same heat
capacity. " They did not restrict this statement to ele-
/
/
0 I I

01 0, 2 03 0, 4 05 06 0,7 08 08 10 ments in solid form. Bather they believed initially that

FIG. 3. The first published graph dealing with the quantum improved experiments might show their law to hold also
theory of the solid state: Einstein's expression for the specific for gases. By 1830 it was clear, however, that the rule
heat of solids [given in Eq. (42)] plotted versus Pv+T. The could, at best, apply to solids only (Fox, 1968).
little circles are Weber's experimental data for diamond.
Much remained to be learned about atomic weights in
Einstein's best fit to Weber's measurements corresponds to
hvtk =—1300 K. those early days of modern chemistry. In fact in sev-
eral instances Dulong and Petit (correctly) halved the
value of atomic weights, obtained earlier by other means,
made. They concern the blackbody radiation law, the in order to bring their data into line with their law (Fox,
light-quantum postulate, and the quantum theory of the 1968). For many years their rule continued to be an
specific heat of solids. All three arose from statistical important tool for atomic weight determinations.
considerations. There are, however, striking differ- It became clear rather soon, however, that even for
ences in the time intervals between these theoretical solid elements the Dulong-Petit rule is not as general
advances and their respective experimental justification. as its propounders had thought. Amedeo Avogadro
Planck formulated his radiation law in a most uncom- (1VV6-1856) was one of the first to remark on deviations
monly short time after learning about the experiments
in the far infrared which complemented earlier results
in the case of carbon, but his measurements Were not
very precise (Avogadro, 1833, especially pp. 96-98). "
at higher frequencies. It was quite a different story Matters got more serious in 1840 when two Swiss physi-
with the light-quantum. Einstein's hypothesis was many cists, Auguste de la Rive (1801-1873) and Franqois
years ahead of its decisive experimental tests. As me Marcet (1803-1883) reported on studies of carbon. ln
shall see next, the story is quite different again in the parti. cular they had obtained "not without difficulty and
case of specific heats. Einstein's first paper on the expense" an amount of diamond powder sufficient to
subject (190Va), submitted in November 1906, contains the experiment with and for which they found c = 1.4. At
qualitatively correct explanation of an anomaly that had almost the same time diamond was also studied (Reg-
been observed as early as 1840: the lorn value of the specific nault, 1841, especially pp. 202-205) by Henri Victor
heat of diamond at room temperature. Einstein showed Regnault (1810-1878) who more than any other physicist
that this can be understood as a quantum effect. His contributed to the experimental investigations of specif-
paper contains one graph: the specific heat of diamond ic heats in the nineteenth century. His value: c =- 1.8.
as a function of temperature, reproduced in Fig. 3. Regnault's coriclusion about carbon was unequivocal:
The dashed line is Einstein's theoretical curve. The It is "a complete exception among the simple bodies:
little circles refer to data obtained in 1875. Figure 3 it does not satisfy the general law which [relates] spe-
is the first published graph in the history of the quan- cific heats and atomic weights. " During the next twenty
tum theory of the solid state. It also represents one of years he continued his studies of specific heats and
only two instances I know in which Einstein published found many more deviations from the general law,
a graph for the purpose of comparing theory with ex- though none as large as for diamond.
periment. (The other example will be mentioned later We nom move to the 1870's when Heinrich Friedrich
in this section. ) Weber (1843-1912), then in Berlin, made the next ad-
In order to recognize an anomaly one needs a theory vance (Weber, 1872). He began by reanalyzing the data,
or a rule or at least a prejudice. As I just mentioned, of de la Rive and Marcet and of Regnault and came to
peculiarities in specific heats were diagnosed more than
half a century before Einstein explained them. It mas
also known well before 1906 that specific heats of gases Zo be precise, these and other measurements on solids to
exhibited even more curious properties. In what sense be mentioned hereafter refer to c& at atmospheric pressure.
did one so early consider diamond so exceptional? And Later on, a comparison will be made with theoretical values
what about other substances'? For a perspective on for c„. This requires a tiny correction to go from cp to c& .
Einstein's contributions it is necessary to sketch the This correction will be ignored (Lewis, 1907).
3~In 1833 Avogadro obtained c= 3 for carbon at room tempera-
answer to these questions. I therefore begi. n with a ture. This value is too high. Since it was accidentally just
short account of specific heats in the nineteenth cen- half the Dulong-Petit value, Avogadro incorrectly conjectured
tury. "that one must reduce the atom [i.e. , the atomic weight] of
This story begins in 1819 when two young Frenchmen, sulphur and metals in general by [a factor] one half. "

Rev. Mod. Phys. , Vol. 51, No. 4, October 1S7S

880 A. Pais: Einstein and the quantum theory

the correct conclusion that the different values for the hydrogen in 1898. In 1905 he reported on the first spe-
specific heat of diamond found by these authors was not cific heat measurements in the newly opened temper-
due to systematic errors. However, the Swiss value ature region. It will come as no surprise that diamond
referred to a temperature average from 3 —14 C while was among the first substances he chose to study. For
Hegnault's value was an average from 8 —98 C. Weber this case he found the very low average value c = 0.05
noted that both experiments could be ri. ght if the spe- in the interval T =20 -85'K. "An almost endless field
cific heat of carbon were to vary with temperature t
of research in the determination of specific heats is now
Tiny temperature variations of specific heats had long opened, " Dewar remarked in this paper (Dewar, 1905). His
been known for substances such as water (see Neu- work is included in a detailed compilation by Wigand
mann, 1831). In contrast, Weber raised the issue of (1907) of the literature on the specific heats of solid
a very strong T dependence —
a new and bold idea. His elements which appeared in the same issue of the An-
measurements for twelve different temperatures be- nalen der Physik as Einstein's first paper on the quan-
tween0 -200 Cconfirmedhis conjecture: for diamond tum theory of specific heats. We are therefore up to
c varied by a factor three over this range. He wanted date in regard to the experimental developments pre-
to continue his observations, but it was March and there ceding Einstein's work.
was no more snow for his ice calorimeter. Heannounced The early theoretical considerations began in 1871
that he would go on with his measurements "as soon as with Boltzmann. At that time only the simplest appli-
meteorological circumstances permit. " The next time cation of the equipartition theorem was known: the av-
we hear from Weber is in 1875 when he presented his erage kinetic energy equals kT/2 for each degree of
beautiful specific heat measurements for boron, siiicon, freedom. In 1871 Boltzmann (1871) showed that the av-
graphite and diamond, from -100' to 1000 C (Weber, erage kinetic energy equals the average potential en-
1875). For the case of diamond c varied by a factor 15 ergy for a system of particles, each one of which os-
between these limits. cillates under the influence of external harmonic forces.
Already in 1872 Weber had made a conjecture which he In 1876 he applied these results to a three-dimensional
confirmed in 1875: at high T one gets close to the Du- lattice (Boltzmann 1876). This gave him an average en-
long-Petit value. In Weber's words (1875): "The three ergy 3RT =6 cal/mol. Hence c„, the specific heat at
curious exceptions [C, B, Si] to the Dulong-Petit law constant volume, equals 6 cal/mol deg. Thus after half
which were till now a cause for despair have been elimi- a century the Dulong-Petit value had found a theoretical
nated: the Dulong-Petit law for the specific heats of justificati6n. As Boltzmann himself put it, his result
solid elements has become an unexceptional rigorous was in good agreement with experiment "for all simple
law. " This is of course not quite true but it was dis- solids with the exception of carbon, boron' and silicon. "
Weber's points of 1875. "
tinct progress. The experimental points in Fig. 3 are Boltzmann went on to speculate that these anomalies
might be a consequence of a loss of degrees of freedom
[Weber later movedfrom Germany to Zurich andbecame due to a "sticking together" at low temperatures of at-
one of Einstein's physics professors during the latter's oms at neighboring lattice points. 'This suggestion was
student days at the ETH (1896-1900). Einstein's note- elaborated by others (Richarz, 1893) and is mentioned
books of Weber's lectures are preserved. They do not by Wigand (1907) in his 1906 review as the best expla-
indicate that as a student Einstein heard of Weber's re- nation of this effect. I mention this incorrect speculation
sults. ] only in order to bring out one important point: Before
In 1872, not only Weber but also a second physicist Einstein's paper of 1906 it was not realized that the di-
made the conjecture that the Dulong-Petit value c = 6 amond anomaly was to be understood in terms of the
would be reached by carbon at high temperatures: failure (or, rather, the inapplicability) of the classical
James Dewar (1842 —1923). His road to the carbon equipartition theorem. Einstein was the first one to
problem was altogether different: for reasons having to state this fact clearly.
do with solar temper atur es Dewar became inter ested By sharp contrast, it was well appreciated that the
in the boiling point of carbon. This led him to do high- equipartition theorem was in trouble when applied to
temperature experiments from which he concluded the specific heat of gases. This was a matter of grave
(Dewar, 1872) that the mean specific heat of carbon between concern to the nineteenth century masters. Even though
0 and 2000 C equals about 5, and that "the true specific this is a topic which does not directly bear on Einstein's
heat [per gram] at 2000 must be at lea, st 0, 5 so that at work in 1906, I believe it to be useful to complete the
this temperature carbon would agree with the law of nineteenth century picture with a brief explanation why
Dulong and Petit. "" gases caused so much more aggravation.
Dewar's most important contribution to our subject The reasons were clearly stated by Maxwell (1965) in
deals with very low temperatures. He had liquefied a lecture given in 1875. "The spectroscope tells us that
some molecules can execute a great many different kinds
of vibrations. They must therefore be systems of avery
By the end of the nineteenth century it was clear that the considerable degree of complexity, having far more
decrease of c with temperature occurs far more generally than six variables [the number characteristic for arigid
than just for C, B, and Si (Behn, 1893).
33There followed a controversy about priorities between
.
body]. . every additional variable increases the specific
Weber and Dewar, but only a very mild one by nineteenth cen- .
heat. . every additional degree of complexity which we
tury standards. In any event, there is no question but that the
issues were settled only by Weber's detailed measurements
in 1875. The good professor wrote "bromine" but meant "boron. "

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 881

attribute to the molecule can only increase the difficulty of principle. Otherwise this paper is much like his
of reconciling the observed with the calculated value of other innovative articles: succinctly directed to the
the specific heat. I have now put before you what I con- heart of the matter.
sider the greatest difficulty yet encountered by the mo- Earlier in 1906 Einstein had come to accept Planck's
lecular theory. " relation (10) between p and U as a new physical as-
Maxwell's conundrum was the mystery of the missing sumption (see Sec. II.E). We have seen in Sec. II.A
vibrations. The following oversimplified picture suffices that Planck had obtained the expression
to make clear what troubled him. Consider a molecule
made up of n structureless atoms. There are 3n de- (38)
grees of freedom, three for translations, at most three for
rotations, the rest are vibrations. The kinetic energy as- by introducing a prescription which modified Boltzmann's
sociated with each degree of freedom contributes kT/2 to way of counting states. Einstein's specific heat paper
c„. In addition there is a positive contribution from the begins with a new prescription for arriving at the same
potential energy. Maxwell was saying that this would result. He wrote U in the form"
almost always lead to specific heats which are 'too large.
One consequence of Maxwell's lecture was that it fo- U(v, T) = f Ee se~'r~(E, v) dE
(39)
cused attention on monatomic gases and in 1876 the f e ~'r ~(E, v) dE
equipartition theorem scored an important success: it The exponential factor denotes the statistical probability
was found that c~/c„— = 5/3 for mercury vapor, in ac- for the energy E.
The weight factor ~ contains the dy-
cordance with c„=3R/2 and the ideal gas rule e~ —c„=A namical information about the density of states between
(Kundt and Warburg, 1876). Also, it had been known E and E+dE. For the case in hand (linear oscillators)
since the days of Regnault" that several diatomic mol- &s is trivial in the classical theory: a(E, v) =1. This
ecules (including hydrogen) have a c„close to 5A/2. It yields the equipartition result U= kT. Einstein proposed
was not yet recognized by Maxwell that this is the value a new form for (d. Let c =hv. Then shall be different
prescribed by the equipartition theorem for a rigid from zero only when nc ~E ~nc+&, n =0, 1, 2, . . .
dumbbell molecule, this observation was first made by "where ~ is infinitely small compared to G, " and such
Boltzmann (1876). This theorem was therefore very that
helpful, yet, on the whole, the specific heat of gases r rid++
remained a murky subject. ~ dE=A, for all n, (40)
Things were getting worse. Already before 1900 "ng
instances were found in which c„depends (weakly) on where the value of the constant A is irrelevant. Math-
temperature (Wiillner, 1896), in flagrant contradiction ematically this is the forerunner of the 5 functions To-
to classical concepts. No wonder these results troubled day we would write
Boltzmann: molecules in dilute gases hardly stick to-
gether) In 1898 he suggested a rather desperate way
out: lack of thermal equilibrium (Boltzmann, 1912).
(u(E, v) = Q n(E' —nbv) .
In 1900 Rayleigh (1900b) remarked that "What would From Eqs. (39} and (40) we recover Eq. (38). This new
appear to be wanted is some escape from the destruc- formulation is important because for the first time the
tive simplicity of the general conclusion [derived from statistical and the dynamical aspects of the problem
equipartition]. " Later that year Kelvin (1901), quoting are clearly separated. "Degrees of freedom must be
Rayleigh in a lecture before the Royal Institution, added weighed and not counted, " as Sommerfeld (1968) put it
his own comment: "Such an escape [from equipartition] later.
would mean that in the beginning of the twentieth century In commenting on his new derivation of Eq. (38) Ein-
[we would] lose sight of a cloud which has obscured the stein remarked: "I believe we should not content our-
brilliance of the molecular theory of heat and lightduring selves with this result. " If we must modify the theory
the last quarter of the nineteenth century. " of periodically vibrating structures. in order to account
Such was the state of affairs when Einstein took on the for the properties of radiation, are we then not obliged
specific heat problem. to do the same for other problems in the molecular
theory of heat, he asked. "In my opinion the answer
B. Einstein cannot be in doubt. If Planck's theory of radiationgoes
Until 1906 Planck's quantum had played a role only to the heart of the matter, then we must also expect
in the rather isolated problem of blackbody radiation. to find contradictions between the present [i.e. , clas-
Einstein's work on specific heats (1907a} is above all sical] kinetic theory and experiment in other areas of
important because it made clear for the first time that
—
the theory of heat contradictions which can be re-
quantum concepts have a far more general applicability. solved by following this new path. In my opinion this
His 1906 paper is also unusual because here we meet expectation is actually realized. "
an Einstein who is quite prepared to use a model he Then Einstein turned to the specific heat of solids.
knows to be approximate in order to bring home a point He introduced the following model of a three-dimen-
sional crystal lattice. The atoms on the lattice points
oscillate independently„ isotropically, harmonically, and
35A detailed review of the specific heats of gases from the
days of Lavoisier until 1896 is found in Wullner's textbook
(i896). 3~I do not always use the notations of the original paper.

Rev. Mod. Phys. , Vol. 5't, No. 4, October 1979

S82 A. Pais: Einstein and the quantum theory

sitions. "
with a single frequency
He emphasized
v around their equilibrium po-
that one should of course not
Einstein's equation (42): The exponential drop of c„
-
as T 0, predicted by that equation, is far too steep.
expect rigorous answers because of all these approxi- Einstein did become aware of this discrepancy in 1911
mations. when the much improved measurements by Nernst
The first generalization. Einstein applied Eq. (39) to (1911a) showed that Eq. (42) fails at low T. Nernst cor-
his three-dimensional oscillators: In thermal equilib- rectly ascribed the disagreement to the incorrectness of
riurn the total energy of a, gram atom of oscillators the assumption that the lattice vibrations are mono-
equals 3&U(v, T), where U is given by Eq (3. 8) and & is chromatic. Einstein himself (1911c) explored some
Avogadro's number. Hence modifications of this assumption. The correct T de-
pendence at low temperatures was first obtained by
$'exp 5 Peter Debye (1884 —1966): for nonmetallic substances
(exp ( —1)' '
c„-0 as T3 (Debye 1912). Einstein had ended his active
Einstein's specific heat formula. research on the specific heats of solids by the time the
work of Debye and the more exact treatment of lattice
The second generalization. 'For reasons of no par-
vibrations by Max Born (1882 —1970) and Theodore von
ticular interest to us now, Einstein initially believed
that his oscillating lattice points were electrically
Karman (1881—1963) appeared (Born and von Karman,
charged ions. A few months later he published a cor- 1912, 1913).
rection to his paper (1907b) in which he observed that These further developments need therefore not be dis-
this was an unnecessary assumption. (In Planck's case
cussed here.
the linear oscillators had of course to be charged! ) However, in 1913 Einstein returned once again to
Einstein's correction freed the quantum rules (inpassing, specific heats, this time to consider the case of gases.
one might say) from any specific dependence on electro- This came about as the result of important experimental
magnetis m. advances on this subject which had begun in 1912 with a
Let us next consider a few properties of Einstein's key discovery by Arnold Eucken (1884 —1950). It had
long been known by then that c„=5 for molecular hydro-
specific heat formula.
gen at room temperature. Eucken (1912) showed that
(1) It yields the Dulong-Petit rule in the high-tem-
this value decreased with decreasing T and that c, = 3
perature limit.
at T = 60 K. As is well known today, this effect is due
(2) It is the first recorded example of a specific heat
formula with the property to the freezing of the two rotational degrees of freedom
of this molecule at these low temperatures. In 1913
c„(T)-0 as T-0. (43) Einstein correctly surmised that the effect was related
As we shall see in the next subsection, Eq. (43) played to the behavior of these rotations and attempted to give
an important role in the ultimate formulation of Nernst's
a quantitative theory. In a paper on this subject we find
the second instance of curve fitting by Einstein (Ein-
heat theorem.
stein and Stern, 1913). However, this time he was
(3) It is a one-parameter formula. The only freedom
wrong. His answer depended in an essential way on the
is the choice of the frequency" v or, equivalently, the
"Einstein temperature" T~, the value of T for which incorrect assumption that rotational degrees of freedom
have a zero point energy.
f =1. As was mentioned before, Einstein compared his
In 1925 Einstein was to turn his attention one last
formula with Weber's points for diamond. Einstein's fit
time to gases at very low temperatures, as we shall see
can be expressed in temperature units by T~ =1300 K
for which "the points lie indeed almost on the curve. " in Sec. VI. C.
This high value of T~ makes clear why a light and hard l"
C. Nernst; Solvay
substance like diamond exhibits quantum effects at room
temperature. (By contrast Ts =70'K for lead. ) "As the temperature tends to absolute zero the en-
(4) By his own account, Einstein took Weber's data tropy of a system tends to a universal constant which
from the Landolt-Bornstein. tables. He must. have used is independent of chemical or physical composition or
the 1905 edition (Landolt and Bornstein, 1905) which of other parameters on which the entropy may depend.
would be readily available in the Patent Office. These "
The constant can be taken to be zero. This modern
tables do not yet contain the earlier-mentioned results general formulation of the third law of thermodynamics
by Dewar in 1905. Apparently Einstein was not aware (barring a few exceptional situations) implies that spe-
of these data in 1906 (although they were noted in that cific heats tend to zero as T-0 (see Huang, 1963). The
year by German physicists (Wigand, 1907)). Perhaps earliest and most primitive version of the "heat the-
this was fortunate. In any case, Dewar's value c„ orem" was presented in 1905, before Einstein had written
=0.05 for diamond refers to an average over the range his first paper on specific heats. The final form of the
g = 0. 02 — 0.07. This value is much too large to be ac- third law was arrived at and accepted only after de-
commodated (simultaneously with Weber's points) by cades of controversy and confusion. For the present

37Volume changes due to heating and contributions to the The preparation of this subsection was much facilitated by
specific heat due to the motions of electrons within the atoms my access to an article by Klein (1965) and a book by Hermann
are neglected, Einstein notes. (1969).
In a later paper, Einstein (1911b}attempted to relate this Simon (1956) has given an exce11ent historical survey of this
frequency to the compressibility of the material. development.

Rev. tVlod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 883

account it is important to notethe inQuenceof Einstein's well say a rule with most curious, indeed grotesque
work on this evolution. properties. However. . . it has borne such rich fruits
On December 23, 1905 Hermann Walther Nernst in the hands of Planck and Einstein that there is now a
(1864 —1941) read a paper at the 0'ottingen Academy scientific obligation to take a stand in its regard and to
entitled "On the computation of chemical equilibria "
subject it to experimental test. He went on to compare
from thermal measurements" in which he proposed a, Planck with Dalton and Newton (Nernst, 1911b). Also
new hypothesis for the thermal behavior of liquids and in 1911, Nernst tried his hand at a needed modification
solids at absolute zero (Nernst, 1906). For our pur- of Einstein's Eq. (42) (Nernst and I.indemann, 1911).
poses the 1905 hypothesis is of particular interest in- Nernst was a man of parts, a gifted scientist, a man
sofar as it applies to a chemically homogeneous sub- with a sense for practical applications, a stimulating
stance. For this case the hypothesis states in essence influence on his students and an able organizer. Many
that the entropy difference between two modifications of people disliked him. But he commanded respect "so
such a, substance (for example graphite and diamond in long as his egocentric weakness did not enter the pic-
the case of carbon) tends to zero as T-0.
This does ture" (Einstein, 1942a). He now saw the need for a
not guarantee that the entropy of a given substance tends conference on the highest level which should deal with
to zero, (it does not exclude the possibility that the the quantum problems. His combined talents as well
entropy becomes singular for T -0), and therefore it as his business relations enabled him to realize this
does not exclude a nonzero specific heat at zero tem- plan. He found the industrialist Ernest Solvay willing
peratures. In fact, in 1906 Nernst (1906a, b) assumed to underwrite the conference. He planned the scientific
that all specific heats tend to 1.5 cal/deg at T = 0. How- program in consultation with Planck and Lorentz. On
ever, he noted that he had no proof of this statement October 29, 1911 the first Solvay Conference convened.
because of the absence of sufficient low temperature Einstein was given the honor of being the final speaker.
data. He stressed (],906a) that it was a "most urgent The title of his talk: "The current status of the specific
task" to acquire these. Nernst" s formidable energies heat problem. " He gave a beautiful review of this sub-
matched his strong determination. He and his col- —
ject and used the occasion to express his opinion on
laborators embarked on a major program for measuring the quantum theory of electromagnetic radiation as well.
specific heats at low temperatures. It covered the same His contributions to the latter topic are without doubt
temperature domain already studied by Dewar but the more profound than his work on specific heats. Yet his
precision was much increased and more substances work on the quantum theory of solids had a far greater
were examined. One of these was diamond, obviously. immediate impact and considerably enlarged the au-
By 1910 Nernst (1910) was ready to announce his first dience of those willing to take quantum physics seri-
results. From his curves "one gains the clear impres- ously.
sion that the specific heats become zero or at least take Throughout the period discussed in the foregoing, the
on very small values at very low temperatures. This third law was considered to apply only to solids and
is in qualitative agreement with the theory developed by liquids. Only in 1914 did Nernst dare to extend his the-
Herr Einstein. . . " orem to hold for gases as well. Eucken's results on the
Thus the order of events was as follows. Late in 1905 specific heat of molecular hydrogen were a main mo-
Nernst stated a primitive version of the third law. In tivation for taking this bold step (Nernst, 1914). Unlike
1906 Einstein gave the first example of a theory which the case for solids, Nernst could not point to a con-
implies that e„-0 as T —0 for solids. In 1910 Nernst vincing theoretical model of a gas with the property
noted the compatibility of Einstein's result with "the c„-0 as &-0. So it was to remain until 1925 when the
heat theorem developed by me. " But it was actually first model of this kind was found. Its discoverer:
Planck who, later in 1910, took a step which "not only Einstein. (See Sec. VI. C. )
in form but also in content goes a bit beyond [the for-
mulation given by] Nernst himself. ." In Planck's for- IV. FROM THE LIGHT-QUANTUM TO THE PHOTON
mulation the specific heat of solids and liquids does go to
zero as T —0 (Planck, 1911). It should be stressed that nei- A. Reactions to the light-quantum hypothesis
ther Nernst nor Planck gave a proof of the third law. In Sec. II we followed the development of Einstein's work
The status of this law was apparently somewhat con- on electromagnetic radiation from 1905 to 1909. The
fused, as is clear from Einstein's remark in 1914 that present section deals with the continuation and con-
"All attempts to derive Nernst's theorem theoretically clusion of his work on this problem. It contains one
in a thermodynamical way with the help of the experi- more result which he obtained in 1909. Then it moves
mental fact that the specific heat vanishes at T =o must on to his contributions in 1916 and 1917. It ends with
be considered to have failed. " Einstein (1914) went on the discovery of the Compton effect in 1923.
to remark that the quantum theory is indispensable for Einstein moved around a good deal during this period.
an understanding of this theorem. In March 1911 he left Zurich for Prague, starting his
Nernst's reference to Einstein in his paper of 1910 first appointment as full professor. In August 1912 he
was the first occasion on which he acknowledged the returned to Zurich to occupy a similar position, this
quantum theory in his publications. His newly aroused time at the ETH. In December 1913 he accepted ap-
interest in the quantum theory was, however, thoroughly pointments in Berlin to a special chair at the Prussian
pragmatic. In an address (on the occasion of the birth- Academy, as director of an Institute for Physics, to
day of the emperor) he said: "At this time the quantum be founded by the Kaiser-W11helm Gesellschaft (this
theory is essentially a computational rule, one may institute was established in 1917); and as a professor

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

884 A. Pais: Einstein and the quantum theory

at the University of Berlin with the right but not the verified consequences of the wave theory" (Einstein,
obligation to teach. He moved there in April 1914 and 1912d, p. 443).
gave his inaugural address less than one month before This statement seems to have created the belief in
the outbreak of the First World War. several quarters that Einstein was ready to retract. In
In Sec. I, I have already touched on the reactions of 1912 Arnold Sommerfeld (1912) wrote: "Einstein drew
the physics community to Einstein's light-quantum pos- the most far-reaching consequences from Planck's dis-
tulate. (See the end of Sec. I.B.) As a prelude to the covery [of the quantum of action] and transferred the
discussion of the further evolution of this concept I be- quantum properties of emission and absorption phenom-
gin with a more detailed analysis of the reasons why ena to the structure of light energy in space without,
the light-quantum was so strongly resisted. To set the as I believe, maintaining today his original point of
tone I mention what Planck, Nernst, Rubens, and War- view [of 1905] in all its audacity. " Referring to the
burg wrote in 1913 when they proposed Einstein for light-quanta, Millikan (1913) stated in 1913 that Ein-
membership in the Prussian Academy. Their recom- stein "gave. . . up, I believe, some two years ago; "
mendation expressed the highest praise for his achieve- and in 1916 he wrote: "Despite. . . the apparently com-
ments. It concludes as follows. "In sum one can say plete success of the Einstein equation [for the photo-
that there is hardly one among the great problems, in effect) the physical theory of which it was designed to
which modern physics is so rich, to which Einstein has be the symbolic expression is found so untenable that
not made a remarkable contribution. That he may Einstein himself, I believe, no longer holds t, o it"
sometimes have missed the target in his speculations, (Millikan, 1916b).
as for example in his hypothesis of light-quanta, cannot It is my impression that the resistance to the light-
really be held too much against him for it is not pos- quantum idea was so strong that one almost hopefully
sible to introduce really new ideas even in the most mistook Einstein's caution for vacillation. However,
exact sciences without sometimes taking a risk" judging from his papers and letters there is no evidence
(Kirsten and Korber, 1975). that at any time he retracted any of his statements
(1) Einstein's caution. Einstein and Michele Angelo made in 1905.
Besso (1873 —1955) had met in Ziirich in about 1897 (2) Electromagnetism: fyee fields and interactions.
and pursued their friendship by correspondence from Einstein's 1905 paper on light-quanta is the second of
1903 until the year of their death. Einstein's letters the revolutionary papers on the old quantum theory. 'The
provide a rich source of his insights into physics and first one was of course PLanck's paper (1900b) of
people. His struggles with the quantum theory in gen- December 1900. Both papers contained proposals which
eral and with the light-quantum hypothesis in particular flouted classical concepts. Yet the resistance to Planck's
is a recurring theme. In1951hewrote (195la) to Besso: — —
ideas while certainly not absent was much less pro-
"Die ganzen 50 Jahre bewusster Grubelei haben mich der nounced and vehement than in the case of Einstein.
Antwort der Frage Was sind Lichtquanten' nicht naher Why?
geb racht. "" First a general remark on the old quantum theory.
Throughout his scientific career quantum physics Its main discoveries concerned quantum rules for sta-
remained a crisis phenomenon to Einstein. His views tionary states of matter and of pure radiation. By and
on the nature of the crisis would change, but the crisis large no comparable breakthroughs occurred in regard
would not go away. This led him to approach quantum to the most difficult of all questions concerning electro-
problems with great caution in his writings —
a caution magnetic phenomena: the interaction between matter
already evident in the way the title of his first paper on and radiation. There, advances became possible only
the quantum theory (1905a) was phrased. In the earliest af ter the advent of quantum f ield theo ry when the con-
years following his light-quantum proposal Einstein had cepts of particle creation and annihilation were for-
good reasons to regard it as provisional. He could only mulated. Since then, progress on the interaction prob-
formulate it clearly in the domain hvt'hT» 1, where lems has been enormous. Yet even today this is not by
Wien's blackbody radiation law holds. Also, he hadused any means a problem area on which the books are
this law as an experimental fact without explaining it. closed.
Above all, it was obvious to him from the start that As we have seen in Sec. II, when Planck introduced
grave tensions existed between his principle and the the quantum in order to describe the spectral properties
wave picture of electromagnetic radiation —
tensions of pure radiation he did so by a procedure of quanti-
which, in. his own mind, were resolved neither then nor zation applied to matter, namely to his material os-
later. A man as perfectly honest as Einstein had no cillators. He was unaware of the fact that his proposal
choice but to emphasize the provisional nature of his implied the need for a revision of the classical radi-
hypothesis. He did this very clearly in 1911, at the ation field itself. His reasoning alleged to involve only
first Solvay congress, where he said: "I insist on the a modification of the interaction between matter and
provisional character of this concept [Light-quanta] which radiation. This did not seem too outlandish, since the
does not seem reconcilable with the experimentally interaction problem was full of obscurities in any event.
By contrast, when Einstein proposed the light-quantum
he had dared to tamper with the Maxwell equations for
"All these 50 years of pondering have not brought me closer free fields which were believed (with good reason) to be
to answering the question 'What are light-quanta'. " The col- much better understood. Therefore it seemed less
lected Einstein-Besso correspondence will be referred to as repugnant to accept Planck's extravaganzas than Ein-
(Speziali, 1972) in what follows. stein's.

Rev. Mod. Phys. , Voa. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 885

This difference in assessment of the two theoretical his derivation — as, among others, Einstein did in 1905.
issues, one raised by Planck, one by Einstein, is quite But at the same time neither Einstein nor any one else
evident in the writings of the leading theorists of the denied the fact that Planck's highly nontrivial universal
day. Planck himself had grave reservations about light- curve admirably fitted the data. Somehow he had to be
quanta. In 1907 he wrote to Einstein" "I am not seeking doing something right.
for the meaning of the quantum of action [light-quanta] Bohr's paper (1913a) of April 1913 about the hydrogen
in the vacuum but rather in places where absorption and atom was revolutionary and certainly not at once gen-
emission occur and [I] assume that what happens in the erally accepted. But. there was no denying that his ex-.
vacuum is rigorously described by Maxwell's equations" pression 2m'e m/h'c for the Rydberg constant of hy-
(Planck, 1907). A remark by Planck (1909) at a. phys- drogen was remarkably accurate (to within 6%, in 1913).
ics meeting in 1909 vividly illustrates his and others' When, in October 1913, Bohr (1913b) was able to give
predilections for "leaving alone" the radiation field and an elementary derivation of the ratio of the Rydberg
for seeking the resolution of the quantum paradoxes in constants for singly ionized helium and hydrogen, in
the interactions: "I believe one should first try to agreement with experiment to five significant figures,
move the whole difficulty of the quantum theory to the it became even more clear that Bohr's ideas had a
domain of the interaction between matter and radi- great deal to do with the realworld. Whentoldof the he-
ation. " In that same year Hendrik Antoon Lorentz lium/hydrogen ratio, Einstein is reported to have said
(1865 —1928) expressed his belief in "Planck's hypoth of Bohr's work, "Then it is one of the greatest dis-
esis of the energy elements" but also his strong res- coveries" (de Hevesy, 1913).
ervations regarding "light-quanta which retain their Einstein himself had little to show by comparison.
individuality in propagation" (Lorentz, 1909). To be sure, he had mentioned a number of experi-
Thus by the end of the first. decade of the twentieth mental consequences of his hypothesis in his 1905 paper.
century many leading theorists wereprepared to accept But he had no curves to fit, no precise numbers to show.
the fact that the quantum theory was here to stay. How- He had noted that in the photoelectric effect the electron
ever, the Maxwell theory of the free radiation field, energy & is constant for fixed light frequency v. This
pure and simple, neither provided room for modifi- explained Lenard's results. But Lenard's measurements
cation (it seemed) nor a place to hide one's ignorance, were not so precise as to prevent menlike J. J. Thomson
xn contrast wxth the less transparent sztuatxonconcernxng and Sommerfeld from giving alternative theories of the
the interaction between matter and radiation. This po- photoeffect of a kind in which Lenard's law does not
sition did not change much until the nineteen twenties rigorously apply (see Stuewer, 1975, Chap. 2). Ein-
and remained one of the deepest roots of resistance to stein's photoelectric equation E = h v —& predicts a linear
Einstein's ideas. relation between E and v. At the time Einstein proposed
(3) The impact of experiment. The first three rev- his heuristic principle no one knew how & depended on
olutionary papers on the old quantum theory were those v beyond the fact that one increases with the other. Un-
by Planck (1900b), Einstein (1905a) and Bohr (1913a). like Bohr and Planck, Einstein had to wait a decade be-
All three contained proposals which flouted classical fore he saw one of his predictions vindicated, the linear
concepts. Yet the resistance to the ideas of Planck and E-v relation, as was discussed in Sec. II. F. One im-
— —
Bohr while certainly not absent was much less pro- mediate and salutary effect of these experimental dis-
nounced and vehement than in the case of Einstein. coveries was that alternative theories of the photoeffect
Why? The answer: because of the impact of experi- vanished from the scene.
ment. Yet Einstein's apartness did not end even then.
Physicists, good physicists, enjoy scientific spec- I have already mentioned that Millikan relished his
ulation in private, but tend to frown upon it when done result on the photo effect but declared that, even so, the
in public. They;are conservative revolutionaries, re- light quantum theory "seems untenable" (Millikan, 1916a)
sisting innovation as long as possible at all intellectual In 1918, Ernest Rutherford (1871—1937) commented as
cost, but embracing it when the evidence is incontro- follows on the Duane-Hunt results: "There is at present
vertible —
or, if they do not, physics tends to pass them no physical e~lanation possible of this remarkable con-
by. nection between energy and frequency" (Rutherford,
I often argued with Einstein about reliance on ex- 1918). One can go on. The fact of the matter is that,
perimental evidence for confirmation of fundamental even after Einstein's photoelectric law was accepted,
new ideas. In Sec. X, I shall have more to say about almost no one but Einstein himself would have anything
Einstein's position on this issue. Meanwhile, I shall to do with light-quanta.
discuss next the influence of experimental developments This went on until the early 1920s, as is best illus-
on the acceptance of the ideas of the three men just trated by quoting the citation (Arrhenius, 1965) for
mentioned. Einstein's Nobel prize in 1921: ' "To Albert Einstein for
First, Planck. His proximity to the first-rate ex- his services to theoretical physics and especially for
periments on blackbody radiation being performed at "
his discovery of the photoelectric effect. This is not
the Physikalisch Technische Reichsanstalt in Berlin only an historic understatement but also an accurate
was beyond doubt a crucial factor in his discovery of
1900 (though it would be very wrong to say that this
was the only decisive factor). In the first instance, ex- Einstein could not attend the festivities since he was in
periment also set the pace for the acceptance of the Japan at that time. He showed his indebtedness one year later
Planck formula. One could (and did and should) doubt by going to Goteborg and giving an address on relativity theory.

Rev. IVlod. Phys. , Vol. 51, No. 4, October 1979

886 A. Pais: Einstein and the quantum theory

reflection on the consensus in the physics community. could be computed directly, if we were to possess an
To summarize: the enormous resistance to light- electrodynamics and mechanics modified in the sense
quanta found its roots in the particle-wave paradoxes. of the quantum hypothesis. " That, of course, was not
The resistance was enhanced because the light-quan- yet the case. He therefore continued his argument in
tum idea seemed to overthrow that part. of electro- the following way. For fixed E —E„and T — we should™
magnetic theory which wa, s believed to be best under- get the Rayleigh —Einstein- Jeans law [Eq. (16)]. This
stood: the theory of the free field. Moreover, ex- implies that
perimental support was long in coming and, even
after the photoelectric effect predictions were veri- B„p„=B „p (46)
fied, light-quanta were still largely considered un- whence
acceptable. Einstein's own emphasis on the pro-
visional nature of the light-quantum hypothesis tended (49)
to strengthen the reservations held by other physicists.
where n „=A„/B— „„.
Then Einstein (1916b) concluded
B. Spontaneous and induced radiative transitions his derivation by appealing to the universality of p and to
Wien's displacement law Eq. (4): "That o.' „and E
In 1916 Einstein (1916a) wrote to Besso: "A splendid —E„cannot depend on particular properties of the mo-
light has dawned on me about the absorption and emis- lecule but only on the active frequency v follows from
sion of radiation. " He had obtained a deep insight into the fact that p must be a universal function of v and T.
the meaning of his heuristic principle and this had led Further it follows from Wien's displacement law that.
him to a new derivation of Planck's radiation law. His & „and E —E„are proportional to the third and first
reasoning is contained in each one of three papers, two power of v respectively. Thus one ha. s
of which appeared in 1916 (1916b, 1916c), the third one
early in 1917 (1917a). His method is based on general
E —E„=hv (5o)
hypotheses about the interaction between radiation and where h denotes a constant. "
matter. No special assumptions are made about in- The content of Eq. (50) is far more profound than a
trinsic properties of the objects which interact with the definition of the symbol v (and h. ). It is a compatibility
radiation. These objects "will be called molecules in condition. Its physical content is this: In order for
what follows" (Einstein, 1916b). (It is completely in- Eqs. (45) and (46) to lead to Planck's law it is necessary
essential to Einstein's arguments that these molecules that the transitions m -n are accompanied by a single
could be Planck oscillators! ) monochromatic radiation quantum. By this remarkable
Einstein considered a system consisting of a gas of reasoning Einstein therefore established a bridge be-
his molecules in interaction with electromagnetic ra- tween blackbody radiation and Bohr's theory of spectra.
diation. The entire system is in thermal equilibrium. About the assumptions he made in the above deri-
Denote by E the energy levels of a molecule and by vation Einstein wrote (1916b): "The simplicity of the
N the equilibrium number of molecules in the level hypotheses make it seem probable to me that these will
E . Then become the basis of the future theoretical description. "
e-Rm /A& Once again he was right.
tn p tn (44)
Two of the three papers under discussion(1916c, 1917a)
where P is a weight factor. Consider a pair of levels contained another result which Einstein himself con-
E, E„,E E„. Einstein's hypothesis is that the total sidered far more important than his derivation of the
number d~ of transition in the gas per time interval dt
is given by This will be our next topic. "
radiation law: light-quanta, carry a momentum hv/c.

dW „=N (pB „+A „)dt, for m -n, (45)

C. The completion of the particle picture
dW„=N„pB„dt, for n —m . (46)
A photon is a state of the electromagnetic field with
The & coefficient corresponds to spontaneous transi- the following properties.
tions m -n which occur with a probability which is in- (1) It has a definite frequency v and a definite wave
dependent of the spectral density p of the radiation pre- vector k.
sent. The & terms refer to induced emission and ab- (2) Its energy &
sorption. In Eqs. (45) and (46), p is a function of v and
T where "we will assume that a molecule can go from
the state E„ to the state E by absorption of radiation and its momentum p
with a definite frequency v, and [similarly] for emis-
sion" (Einstein, 1916b). Microscopic reversibility (52)
implies that dpi' „=dW„„. Using Eq. (44) we therefore
have
A „p = p(B„p e~+m @rr~~r~ B p ) (47) I can only mention in passing another paper on the quantum
theory published by Einstein (1917b) in 1917. It deals with the
(Note that the second term on the right hand side cor-
restrictions imposed by the old quantum theory on allowed
responds to induced emission. Thus if there were no phase space orbits. Einstein examined the topological char-
induced emission we would obtain Wien's law. ) Ein- acterization of these orbits. (He never returned to this sub-
stein (1916b) remarked that "the constants A and B ject.)

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 887

satisfy the dispersion law was derived by Einstein (1909a, 1909b) in 1909:
E=c Ip I
{s'(v, T)) = [phv+ c'p'/8mv']v dv. (37)
characteristic for a particle of zero rest mass. 4' Recall that {c') is the energy fluctuation in the interval
(3) It has spin one and (like all massless particles dv referring to a subvolume v of a cavity filled with
with nonzero spin) two states of polarization. The radiation in thermal equilibrium; and that Eq. (37)
single particle states are uniquely specified (Wigner, holds if p is given by Planck's formula (6). For our
1939) by the properties (1)—(3). purposes it is important to note a second fluctuation
The number of photons is in general not conserved in formula which is found in the same 1909 papers. This
particle reactions and decays. I shall return to the non- one deals with momentum fluctuations and is pertinent
conservation of photon number in the section on quan- to the question of photon momentum. Einstein con-
tum statistics (Sec. VI) but would like to note here an sidered the case of a plane mirror with mass m and
ironical twist of history. The term "photon" first ap- f
area placed inside the cavity. The mirror moves
peared in the title of a paper written in 1926. ' The perpendicular to its own plane and has a velocity v at
title. "The conservation of photons. " The author: the time t. During a small time interval from t to t+& its
distinguished physical chemist Gilbert Newton Lewis momentum changes from mv to mv —Pv7 + &. The second
(1875 —1946) from Berkeley. The subject: a specu- term describes the drag force due to the radiation pressure
lation that light consists of "a new kind of atom. . . . (P is the corresponding friction constant). This force
uncreatable and indestructible [for which] I. . . . propose would eventually bring the mirror to rest where it
the name photon" (Lewis, 1926). This idea was soon not for the momentum Quctuation term & induced by
forgotten but the new name almost immediately be- the Quctuations of the radiation pressure. In thermal
came part of the language. In October 1927 the fifth equilibrium the mean square momentum m {v') should
Solvay conference was held. Its subject was "Electrons remain unchanged over the interval &. Hence" {&')
et Photons. " =2mPr{v'). The equipartition law applied to the kinetic
energy of the mirror implies that rn(v') =kT. Hence
When Einstein introduced light-quanta in 1905, these
were energy quanta satisfying Eq. (51). There was no {&')= 2P rkT . (51)
mention in that paper of Eqs. (52) and (53). In other
Einstein computed P in terms of p for the case in which
words, the full-fledged particle concept embodied in
the mirror is fully transparent for all frequencies ex-
the term photon was not there all at once. For this
cept those between v and v+d&, which it reflects per-
reason I make the distinction between light-quantum
("E =hv only" ) and photon in this section. The dis- fectly. Using Planck's expression for p he found that
symmetry between energy and momentum in the 1905 1 g 3p2
c phv~,
(b') —
8m'' fvdv.
=
paper is of course intimately connected with the origins .

of the light-quantum postulate in equilibrium statis-

The parallels between Eqs. (37) and (52) are striking.
tical mechanics. In the statistical mechanics of equi- The respective first terms dominate if hv/kT» 1, the
librium systems one derives important relations be-
regime in which p is approximated by Wien's exponen-
tween the overall energy and other macroscopic vari-
tial law. The first term in Eq. (37) extends Einstein's
ables. The overall momentum plays a trivial and sub- "energy quantum postulate" of 1905 to energy fluctua-
sidiary role. These distinctions between energy and tions. One might expect that the first term in Eq. (52)
momentum are much less pronounced when one con-
would lead Einstein to state, in 1909, the "momentum
siders Quctuations around the equilibrium state. It quantum postulate". Monochromatic radiation of low
was via the analysis of statistical fluctuations of black-
density behaves in regard to pressure Quctuations as
body radiation that Einstein eventually came to asso- if it consists of mutually independent momentum quanta
ciate a definite momentum with a light-quantum. This of magnitude hv/c. It is unthinkable to me that Ein-
happened in 1917. Before I describe what he did I
stein did not think so. But he did not quite say so.
should again draw the attention of the reader to the re-
This is what he did say (1909a): "If the radiation
markable fact that it took the father of special rela-
were to consist of very few extended complexes with
tivity theory twelve years to write down the relation
energy h& which move independently through space and
p=hv/c side by side with E=hv. I shall have more to which are independently reQected —
a picture which rep-
say about this in Sec. X.C. resents the roughest visualization of the light-quantum
Now, once again, to the fluctuations. To begin with I hypothesis —
then as a consequence of fluctuations in the
copy a formula (discussed earlier in Sec. II.G) which radiation pressure such momenta would act on our plate
as are represented by the first term only of our for-
mula [(Eq. 52)]." He did not refer explicitly to mo-
There have been occasional speculations that the photon mentum quanta nor to the relativistic connection be-
might have a tiny nonzero mass. Direct experimental informa- tween E=hv and p =hv/c. Yet a particle concept (the
tion on the photon mass is therefore a matter of interest. The photon) was clearly on his mind since he went on to
best determinations of this mass come from astronomical ob- conjecture (1909a) that "the electromagnetic fields of
servations. The present upper bound (Davis, Goldhaber, and
Nieto, 1975) is 8. 10 g. In what follows the photon mass is
taken to be strictly zero.
4~For ease I shall often use the term "photon" also in refer- 4~Terms 0(*2) are dropped. (v~) =0 since v and A are uncor-
ring to times prior to 1926. related.
I

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

888 A. Pais: Einstein and the quantum theory

light are linked to singular points similar to the oc- same answer for the quantities on the 1hs and rhs of
currence of electrostatic fields in the theory of elec- Eq. (51) Einstein (1917a) had to invoke a condition of
"
trons. It seems fair to paraphrase this statement as directedness: "If a bundle of radiation causes a mo-
follows: light-quanta may well be particles in the same lecule to emit or absorb an energy amount hv then a
sense that electrons are particles. The association be- momentum hv/c is transferred to the molecule, di-
tween the particle concept and a high degree of spatial rected along the bundle for absorption and oppositely
localization is typical for that period. It is of course to the bundle for [induced] emission. " (The question of
not correct in general. spontaneous emission is discussed below. ) Thus Ein-
The photon momentum made its explicit appearance stein found that consistency with the Planck distri-
in that same year, 1909. Johannes Stark (1874 —1957) bution [and Eqs. (45), (46)] requires that the radiation
had attended the Salzburg meeting at which Einstein be fully directed. (One often called thisXadelstvahlung. )
(1909a) had discussed the radiative fluctuations. A few And so with the help of his trusted and beloved fluctu-
months later Stark (1909) stated that according to the ation methods Einstein once again produced a major
light-quantum hypothesis, "the total electromagnetic insight, the association of momentum quanta with en-
momentum emitted by an accelerated electron is dif- ergy quanta. Indeed if we leave aside the question of
ferent from zero and. . . . in absolute magnitude is given spin ' we may say that Einstein abstracted not only the
by hv/c. "
As an example he mentioned bremsstrahlung light-quantu~ but also the ~owe general photon con-
for which he wrote down the equation cept entirely from statistical mechanical considerations.
hv
~z Vz+~2V2 —~z Vz+~2V2+ c 2 C, (53) D. Earliest Unbehagen about chance

the first ocasion on record in which the

Einstein prefaced his statement about photon momen-
photon enters
tum which I just quoted with the remark that this con-
explicitly into the law of momentum conservation for
clusion can be considered "als ziemlich sicher
an elementary process.
Einstein himself did not explicitly introduce photon erwiesen, "' ("as fairly certainly proven"). If he had
some lingering reservations, this was mainly because
momentum until 1916, in the course of his studies on
he had derived some of his equations (1917a) on the
thermal equilibrium between electromagnetic radiation
basis of "the quantum theory [which is] incompatible
and a molecular gas (Einstein, 1916c, 1917a). In ad-
dition to his new discussion of Planck's law Einstein
with the Maxwell theory of the electromagnetic field. "
Moreover, his momentum condition was a sufficient,
raised the following problem. In equilibrium the mo-
not a necessary condition, as was emphasized by Wolf-
leeules have a Maxwell distribution for the translational
velocities. How is this distribution maintained in time gang Pauli (1900-1958) in a review article (Pauli,
considering the fact that the molecules are subject to 1964, p. 680) completed in 1924: "From Einstein's
the influence of radiation pressure? In other words,
considerations it could. . . not be seen with complete
certainty that his assumptions were the only ones which
statistical equilibrium. "
what is the Brownian motion of molecules in the pres-
ence of radiation? guarantee thermodynamic —
Technically the following issue arises. If a molecule Nevertheless his 1917 results led Einstein to drop his
emits or absorbs an amount c of radiative energy all of caution and reticence about light-quanta. They had be-
which moves in the same direction, then it experiences come real to him. In a letter to Besso about the needle
rays he wrote (Einstein, 1916d): "Damit sind die I icht-
a recoil of magnitude c/c. There is no recoil if the ra-
diation is not directed at all, as for a spherical wave. quanten so gut wie gesichert. "" And, in a phrase con-
tained in another letter about two years later (Einstein,
Question: what can one say about the degree of direc-
tedness of the emitted or absorbed radiation for the 1918), ".. . I do not doubt anymore the reality of ra-
diation quanta, although I still stand quite alone in this
Einstein began the dis-
conviction", he underlined the word "Realitat. "
system under consideration?
cussion of this question in the same way he had treated
the mirror problem in 1909. Instead of the mirror he On the other hand, at about the same time that Ein-
now considered molecules which all move in the same
stein lost any remaining doubts about the existence of
direction. Then there is again a drag force —I'v& and light-quanta we also encounter the first expressions of
a fluctuation term &. Equipartition gives again his Unbehagen, his discomfort with the theoretical im-
plications of the new quantum concepts in regard to
m(v') =kT and one arrives once more at Eq. (51). Next
comes the issue of compatibility. With the help of Eqs.
"
"Zufall, "chance".
(45) and (46) Einstein could compute separately expres-
This earliest unease stemmed from the conclusion
concerning spontaneous emission which Einstein had
sions for (&') as well as for I' in terms of the &'s,
B's, and p, where p is now given by Planck's law. I " been forced to draw from his consistency condition
(51): The needle ray picture applies not only to induced
shall not reproduce the details of these calculations but
do note the crux of the matter. In order to obtain the
processes (as was mentioned above) but also to spon-
taneous emission. That is, in a spontaneous radiative
transition the molecule suffers a recoil hv/c. However,
47In 1910, Einstein had made a related calculation, together
with Ludwig Hopf (1884—1939) (Einstein and Hopf, 1910). At
that time he used the classical electromagnetic theory to com- 4 Some comments on the photon spin will be found in Sec. VI.
pote (~) andP. This east Eq. (51) into a differential equation 4 "With that the existence of] the light-quanta is practically
for p. Its solution is Eq. (16). certain. " t

Rev. Mod. Phys. , Vol. 5't, No. 4, October 't 979

 A. Pais: Einstein and the quantum theory

the recoil direction cannot be predicted. Einstein E. The Compton effect

(1917a) stressed (quite correctly of course) that it is
"a weakness of the theory. . . that it leaves time and We now come to the denouement of the photon story.
direction of elementary processes to chance. What " Since, after 1917, Einstein firmly believed that light-
decides when the photon is spontaneously emitted? What quanta were here to stay, it is not surprising that he
decides in which direction it shall go? would look for new ways in which the existence of pho-
These questions were not new. They also apply to tons might lead to observable deviations from the clas-
another class of emission processes, the spontaneity sical picture. In this he did not succeed. At one point,
of which had puzzled physicists since the turn of the in 1921, he thought he had found a new quantum cri-
century: radioactive trans formati. ons. A spontaneous terion (Einstein, 1921), but it soon turned out to be a
emission coefficient was in fact first introduced by false lead (Einstein, 1922; Klein, 1970). In fact, after
Rutherford (1SOO) in 1900 when he wrote down" the 1917 nothing particularly memorable happened in regard
equation dX = -@Mt for the decrease of the number & to light-quanta until capital progress was achieved"
of radioactive thorium emanation atoms in the time in- when Arthur Compton (1892-1S62) (Compton, 1923) and
terval dt. Einstein himself (1916b) drew attention to Debye (1923) independently wrote down the relativistic
this similarity: "It speaks in favor of the theory that kinematics for the scattering of a photon on an electron
the statistical law assumed for [spontaneous] emission at rest:
is nothing but the Rutherford law of radioactive decay. "
I have written elsewhere (Pais, 1977) about the ways hk =p+hk', (54)
physicists responded to this baffling lifetime problem. kl+~. =I, Ik I+(, & +~', ) (55)
I should now add that Einstein was the first to realize
that the probability for spontaneous emission is a non- Why were these elementary equations not published five
classical quantity. No one before Einstein in 1917 saw or even ten years earlier, as well they could have been'?
Even those opposed to quantized radiation might have
as clearly the depth of the conceptual crisis generated
found these relations to their liking since (independent
by the occurrence of spontaneous processes with a well-
of any quantum dynamics) they yield at once significant
defined (partial) lifetime. He expressed this inprophetic
differences from the classical theories of the scattering
terms: "The properties of elementary processes re- of light by matter" and, therefore, provide simple
quired by [Eq. (51)] make it seem almost inevitable to
formulate a truly quantized theory of radiation" (Ein-
tests of the photon idea.
I have no entirely satisfactory answer to this ques-
stein, 1917a).
tion. In particular it is not clear to my why Einstein
Immediately following his comment on "chance", to
himself did not consider these relations. However,
which I have just referred, Einstein (1917a) continued:
"Nevertheless I have full confidence in the route which there are two obvious contributing factors. First, be-
has been taken. "
If Einstein was confident at that time
cause photons were rejected out of hand by the vast
about the route, he also felt strongly that it would be a
majority of physicists, few may have felt compelled
to ask for tests of an idea they did not. believe to begin
long one. The chance character of spontaneous pro-
with. Secondly, it was only in about 1922 that strong
cesses meant that something was amiss with classical evidence became available for deviations from the clas-
causality. That would forever deeply trouble him. As
early as March 1917 Einstein (1917c) had written on
this subject to Besso: "I feel that the real joke which
sical picture. This last circumstance impelled both
Compton and Debye to pursue the quantum alternative. "
Debye (1923), incidentally, mentioned his indebtedness
the eternal inventor of enigmas has presented us with
has absolutely not been understood as yet. It is be- " "
to Einstein's work on needle radiation. Compton in his
lieved by nearly all of us that the joke was understood paper does not mention Einstein at all.
The same paper in which Compton discussed Eqs.
soon after 1925 when it became possible to calculate
Einstein's A „y and& „from firstprinciples. As Ishall (54) and (55) also contains the result of a crucial ex-
discuss below, Einstein eventually accepted these prin- periment. These equations imply that the wavelength
ciples but never considered them to be first principles. difference && between the final and the initial photon is
Einstein's attitude throughout the rest of his life was:
the joke has not been understood as yet. One further
example may show how from 1917 on Einstein could not 5~Einstein (1924a) attached great importance to an advance in
make his peace with the quantum theory. In 1920 he another direction which took place in the intervening years:
the effect discovered by Otto Stern (1888—1969) and Walther
wrote as follows to Born. "That business about causality Gerlach, (1889-1979). Together with Ehrenfest he made a pre-
causes me a lot of trouble too. Can the quantum ab- mature attempt at its interpretation (Einstein and Ehrenfest,
sorption and emission of light ever be understood in the 1922).
sense of the complete causality requirement, or would 5
For details on these classical theories see Stuewer's fine
a statistical residue remain? I must admit that there monograph (1975) on the Compton effect.
I lack the courage of a conviction. But I would be very Nor is it an accident that these two men came forth with the
unhappy to renounce complete causality" (Einstein, 1920a). photon kinematics at about the same time. In his paper Debye
(1923) acknowledges a 1922 report by Compton in which the
evidence against the classical theory was reviewed. A com-
plete chronology of these developments in 1922 and 1923 is
Here a development began which, two years later, culmin- found in Stuewer (1975), p. 235.
ated in the transformation theory for radioactive substances 54For a detailed account of the evolution of Compton's thinking
(Rutherford and Soddy, 1902). see Stuewer (1975), Chapter 6.

Rev. Mod. Phys. , Vol. 53, No. 4, October 1979

890 A. Pais: Einstein and the quantum theory

given by (1894 —1952) and John Clarke Slater (1900—1976) sub-

~ ~ = (l /mc)(I —cose), mitted to the Philosophical Magazine an article (Bohr,
Kramers, and Slater, 1924) which contained drastic
where is the photon scattering angle. Compton found
this relation to be satisfied within the error. 'The " theoretical proposals concerning the interaction of light
with matter. It was written after Compton's discovery,
quality of the experiment is well demonstrated by the yet it rejected the photon. It was written also after Ein-
value he obtained for the Compton wavelength: h/mc stein and Bohr had met. This section on the BKS pro-
= 0.0242 A which is within less than one percent from posal serves a twofold purpose. It is a postscript to the
the modern value. (For the current state of the subject story of the photon and a prelude to the Bohr-Einstein
see Williams, 1977). Compton concluded: "The ex- dialog which will occupy us more fully in later sections.
perimental support of the theory indicates very con- I have already mentioned that Einstein was immedi-
vincingly that a radiation quantum carries Kith it di- ately and strongly impressed by Bohr's work of 1913.
rected momentum as well as energy. "" They d'id not yet know each other at that time. A num-
This discovery "created a sensation among the phys- ber of years was to pass before their first encounter;
icists of that time" (Allison, 1965). There were the meanwhile they followed each others' published work.
inevitable controversies surrounding a discoveryof such Also, Ehrenfest kept Einstein informed of the progress
major proportions. Nevertheless the photon idea was in Bohr's thinking. "Ehrenfest tells me many details
rapidly accepted. Sommerfeld incorporated the Co&u- from Niels Bohr's Gedankenkiiche, [thought kitchen];
pton effect in his new edition (1924) of &tombau und he must be a very first-rate mind, extremely critical
SPektxallinien with the comment: "It is probably the and far seeing, which never loses track of the grand
most- important discovery which could have been made . design" (Einstein, 1919). Einstein remained forever
in the current state of physics. " deeply respectful of Bohr's pioneering work. When he
What about Einstein's own response'P A year after was nearly seventy he wrote (Einstein, 1949a): "That
Compton's experiments he wrote a popular article for this insecure and contradictory foundation [of.physics
the Berliner Tageblatt (Einstein, 1924b) which ends in the teens] was sufficient to enable a man of Bohr's
as follows: "The positive result of the Compton ex- unique instinct and tact to discover the major laws of
periment proves that radiation behaves as if it con- the spectral lines and of the electron shells of the at-
sistedofdiscreteenergy projectiles, not only in regard oms together with their significance for chemistry ap-
to energy transfer but also in regard to Stosswixkung
"
—
peared to me like a miracle and appears to me as a
(momentum transfer). Here then, in projectile (that is, miracle even today. This is the highest fofm of musi-
particle) language, is the "momentum quantum postu- cality in the sphere of thought. "
"
late, phrased in close analogy to the energy quantum Einstein and Bohr finally met in the spring of 1920,
postulate of 1905. In both cases we encounter the in Berlin. At that time they both had already beenwidely
.
phraseology: "Radiation. . behaves. . . as if it con- recognized as men of destiny who would leave their in-
sists of. . . ." delible marks on the physics of the twentieth century.
Still, Einstein was not (and would never be) satisfied. The impact of their encounter was intense and went well
There was as yet no real theory. In the same article beyond only a meeting of minds. Shortly after his visit,
he also wrote'. "There are therefore now two theories Einstein (1920b) wrote to Bohr: "Not often in life has
—
of light, both indispensable, and as one must admit a human being caused me such joy by his mere pres-
today in spite of twenty years of tremendous effort on "
ence as you did. Bohr (1920) replied: "To meet you
the part of theoretical physicists —
without any logical and to talk with you was one of the greatest experiences
connection. " "
I have ever had. Some years later Einstein (1923a)
The years 1923-1924 mark the end of the first phase began a letter to Bohr as follows: "Lieber oder vielmehr
of Einstein's apartness in relation to the quantum the- "
geliebter Bohr. ("Dear or rather beloved Bohr. ") Once
ory. Yet there remained one important bastion of re- when I talked with Helen Dukas, Einstein's devoted sec-
sistance to the photon, centering around Niels Bohr. retary, about the strong tie between these two men'
she made the comment which is at the head of this sec-
V. INTERLUDE: THE BKS PROPOSAL tion. Its literal translation would distort its meaning.
Rendered more freely it means: "They were deeply
pie pgQQepg sech bepiss und
involved with one another in affectionate ways. "
innig gelieM.
Helen Dukas. All who have known Bohr will be struck by this char-
acterization which Einstein (1954a) gave of him much
In January 1924 Niels Bohr, Hendrik Anton Kramers, later: "He utters his opinions like one perpetually
groping and never like one who believes to be in the
possession of definite truth. " Bohr's style of writing
~K line x rays from a molybdenum anticathode were scattered makes clear for all to see how he groped and struggled.
off graphite. Compton stressed that one should only use light "Never express yourself more clearly than you think, "
elements as scatterers so that the electrons will indeed be he used to admonish himself and others. Bohr's arti-
quasifree. Scattered x rays at 45, 90', and135 were analyzed. cles are sometimes dense. Having myself assisted him
5~The work of Compton and Debye led Pauli (1923) to extend
Einstein's work of 1917 to the case of radiation in equilibrium
with free electrons. Einstein and Ehrenfest (1923) subsequent-
ly discussed the connection between Pauli's and Einstein's 5
This discussion took place in June 1978. Miss Dukas started
Stoss zahlansatz. to wol k for Einstein in April 1928

Rev. Mod. Phys. , Vol. 5't, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 891

on a number of occasions when he was attempting to put 1925). But how can there be conservation of an energy
his thoughts on paper, I know to what enormous lengths which consists of two parts, one changing discontinuously,
he went to find the most appropriate turn of phrase. I the other continuously'? The BKS answer (Bohr, Kram-
have no such first-hand information about the way Ein- ers, and Slater, 1924): "As regards the occurrence
stein wrote. But, again for all to see, there are his of transitions, which is the essential feature of the quan-
papers, translucent. The early Einstein papers are tum theory, we abandon. . . a direct application of the
brief, their content is simple, their language is sparse. principles of conservation of energy and momentum. "
They exude finality even when they deal with a subject E nergy and momentum cons er vation, they suggested,
in flux. For example, no statement made in the 1905 does not hold true for individual elementary processes
paper on light-quanta needs to be revised in the light of but should only hold statistically, as an average over
later developments. Whether he published in German many such processes.
or English, he initially wrote in German. He had a del- 'The idea of energy nonconservation had already been
icate musical sense of language and a keen insight into on Bohr's mind a few years prior to the time of the
people, as his description of Bohr illustrates. BKS proposal (Bohr, 1923, especially Sec. 4). How- "
Their meeting in 1920 took place some years before ever, it was not Bohr but Einstein who had first
they found themselves at scientific odds on profound — —
raised and rejected this possibility. In 1S10 Ein-
questions of principle in physics. They did not meet stein wrote to a friend (see Seelig, 1954, p. 137). "At
very often in later times. They did correspond but not present I have high hopes for solving the radiation problem
voluminously. I was together a few times with both of and that without light-quanta. I am enormously curious
them some thirty years after their first encounter, how it will work out. One must renounce the energy prin-
when their respective views on the foundations of quan- ciple in its present form. " A few days later he was
tum mechanics had long since become irreconcilable. disenchanted: "Once again the solution of the radia-
More about that later. Let me note here only that nei- tion problem is getting nowhere. The devil has played
ther the years nor later events ever diminished the a rotten trick on me" (Seelig, 1954, p. 137). He raised the
mutual esteem and affection in which they held one an. — issue one more time at the 1911 Solvay meeting, noting
other. that his formula for the energy fluctuations of black-
Let us now turn to the BKS proposal. body radiation could be interpreted in two ways: "One
As I have already stressed in Sec. IV.A, it was the can choose between the [quantum]:structure of radia-
position of most theoretical physicists during the first tion and the negation of an absolute validity of the en-
decades of the quantum era that the conventional con- ergy conservation law. " He rejected the second alter-
tinuous description of the free radiation field should be native. "Who would have the courage to make a decision
protected at all cost and that the quantum puzzles con- of this kind. . . . We will agree that the energy principle
cerning radiation should eventually be resolved by a should be retained" (Einstein, 1912d, pp. 429, 436).
revision of the properties of interaction between ra- But others were apparently not as convinced. In 1916
diation and matter. The BKS proposal represents the the suggestion of statistical energy conservation was
extreme example of this position. Its authors suggested taken up by Nernst (1916). ' Not later than January 1922
that radiative processes have highly unconventional Sommerfeld (1922) remarked that the "mildest cure"
properties "the cause of [which] we shall not seek in for reconciling the wave theory of light with quantum
any departure from the electrodynamic theory of light phenomena would be to relinquish energy conservation.
as regards the laws of propagation in free space, but (See Klein, 1970, for similar speculations by other
in the peculiarities of the interaction between the virtual physicists. ) Thus the BKS proposal must be regarded
field of radiation and the illuminated atoms" (Bohr, as an attempt to face the consequences of an idea that
Kramers, and Slater, 1924). Before describing these had been debated for quite some time.
properties I should point. out that the BKS paper rep- In order to understand Bohr's position in 1924 it is
resents a program rather than a detailed research re-
port. In contains no formalism whatsoever. This " important above all to realize that the correspondence
principle was to him the principal reliable bridge be-
program was not to be the right way out of the diffi- tween classical and quantum physics. But the corre-
-

culties of the old quantum theory, yet the paper had a spondence principle is of course no help inunderstanding
lasting impact in that (as we shall see) it stimulated light-quanta: the controversial issue of photons versus
important experimental developments. Let us discuss waves lies beyond this principle. To repeat, thephoton-
next the two main paradoxes which BKS addressed. wave duality was the earliest known instance of what was
The first Paradox Consider .an atom which emits later to be called a complementary situation. The BKS
radiation in a transition from a higher to a lower state. theory, with its rejection of photons and its insistence
BKS assume that in this process "energy [is] of two on the continuous picture of light at the price of non-
kinds, the continuously changing energy of the field and conservation, historically represents the last stand of
the discontinuously changing atomic energy" (Slater, the old quantum theory. For very good reasons this

5
The same is true for a sequel to this paper which Bohr A letter from Ehrenfest (1922) to Einstein shows that Bohr s
(1925a) wrote in 1925. Schroedipger (1924) and especially thoughts had gone in that direction at least as early as 1922.
Slater (1925) did make attempts to put the BKS ideas on a more ~
The title of Nernst's paper is (in translation): "On an at-
formal footing. See also Slater's own recollections (1967) of tempt to revert from quantum-mechanical considerations to the
that period. assumption of continuous energy changes. "
Rev. Mod. Phys. , Vol. 51, No. 4, October 1979
 A. Pais: Einstein and the quantum theory

proposal was characterized some years later (Heisen- take sides meant to choose between the two most re-
berg, 1929) by one of the principal architects of the vered physicists. Ideally, personal considerations of
quantum mechanics as representing the height of the this kind ought to play no role in matters scientific. But
crisis in the old quantum theory. Nor was nonconser- this ideal is not always fully realized. Pauli (1924) re-
vation of energy and momentum in individual processes flected on this in a letter concerning the BKS issue:
the only radical proposal made by BKS. "Even if it were psychologically possible for me to form
The second paradox. This one concerns a question a scientific opinion on the grounds of some sort of be-
which had troubled Einstein since 1917 (as we have lief in authority (which is not the case, however, as
seen): How does an electron know liken to emit radi- you know), this would be logically impossible (at least
ation in making a spontaneous transition? in this case) since here the opinions of two authorities
In its general form the BKS answer to this question are so very contradictory. "
was: there is no txuLy spontaneous emission. They Even the interaction between the two protagonists was
associated with an atom in a given state a "virtual ra- circumspect during that period. They did not corre-
diation field"' which contains all the possible transition spond on the BKS issue (Einstein, 1924c). Nor (asbest
frequencies to other stationary states and assumed that I know) were there personal meetings between them
"the transitions which in [the Einstein theory of 1917] in those days even though Bohr had told Pauli repeatedly
are designated as spontaneous are, on our view, in- how much he would like to know Einstein's opinion
duced [my italics] by the virtual field. " According to (Pauli, 1924). Werner Heisenberg (1901—1976) wrote
BKS, the spontaneous transition to a specific final state (1924) to Pauli that he had met Einstein in Gottingen and
is connected with the virtual field mechanism "by prob- that the latter had "a hundred objections. " Sometime
ability laws which are analogous to those which in Ein- later Pauli also met Einstein whereupon he sent Bohr a
stein's theory hold for induced transitions. " In this way detailed list of Einstein s criticisms (Pauli, 1924).
"the atom is under no necessity of knowing what transi- Einstein of course never cared for BKS. He hadgiven
tion it is going to make ahead of time" (Slater, 1925). a colloquium on this paper at which he had raised ob-
Thus spontaneous emission is ascribed to the action of jections. The idea (he wrote Ehrenfest (Einstein,
the virtual field, but this action is noncausal. I shall 1924d)) "is an old acquaintance of mine, which I do not
not discuss details of the BKS picture of induced emis- hold to be the real fellow however" (". . . den ich aber
sion and absorption and other radiative processes. f'ur Keinen reellen Kerl halte. ") At about that time he
Suffice it to say that all of these are supposed to be due drew up a list of nine objections (Einstein, 1924e) which
to virtual fields and that in all of these causality is I shall not reproduce here in detail. Samples: "what
abandoned. In a paper completed later in 1924 Slater should condition the virtual field which corresponds to
(1925) noted that the theory "has unattra. ctive fea- the return of a previously free electron to a Bohr orbit?
tures. . . . [but] it is difficult at the present stage to see (very questionable). . . Abandonment of causality as a
how [these are] to be avoided. " matter of principle should only be permitted in the most
But what about the Compton effect? The successfully extreme emergency. " The causality issue (which had
verified Eq. (56) rests on the conservation laws (54) and plagued him already for seven years by then) was clearly
the one to which he took exception most strongly. He
(55). However (BKS argued), these equations do hold in
the average and the experiment on &~ refers only to the confided to Born (Einstein, 1924f) that the thought was
average change of the wavelength. In fact at the time of unbearable to him that an electron could choose freely
the moment and direction in which to move. The cau-
the BKS proposal Neve did not exist any dA ect experi-
mental proof of energy-momentum conservation nor of sality question would continue to nag him long after
causaLity in any individual Process. This is one of the experiment revealed that the BKS answers to both par-
reasons why the objections to BKS (held by many, "per- adoxes were incorrect.
Me exPeximental ver'dict on causality. The BKS ideas
haps the majority" (Pauli, 1924), of physicists") were
stimulated Walther Bothe (1891—1957) and Hans Geiger
initially expressed in a somewhat muted fashion. Thus
(1882 —1945) (Bothe and Geiger, 1924) to develop counter
Pauli (1924) wrote to Bohr that he did not believe in his
coincidence techniques for the purpose of measuring
theory but that "one cannot prove anything logically and whether (as causality demands) the secondary photon
also the available data are not sufficient to decide for
or against your view. " All this was to change soon.
and the knocked-on elec tron are produced simultaneously
in the Compton effect. Their result (Bothe and Geiger,
There was a second reason, I believe, for the subdued 1925a, 1925b): these two particles are both created in
character of comments by others. The physics com- a. time interval ~ 1O 'sec. Withinthe limits of accuracy
munity was witness to a rare occurrence. Here were causality had been established and the randomness
the two leading authorities of the day locked in conflict. (demanded by BKS) of the relative creation times dis-
{The term conflict was used by Einstein himself. To ") proved. Since then this time interval has beennarrowed
down experimentally (Bay, Henri, and Mcl ennon, 1955)
6~Born, Schroedinger, and R. Ladenburg were among the sec.
physicists who initially believed that BKS might be a step in The experimental verdict on energy-momentum con-
the right direction.
On October 25, 1924, the Danish newspaper Politiken
servation. The validity of these conserva. tion laws in
carried a news item on the Bohr —Einstein controversy. This individual elementary processes was established for the
led the editor of a German newspaper to send a query to Ein- Compton effect by Compton and A. W. Simon. From
stein (Joel, 1924). Einstein (1924c) sent a brief reply acknow- cloud chamber observations on photoelectrons and
ledging that a conflict existed, but also stating that no written knock-on electrons they could verify (Compton and
exchanges between himself and Bohr had ensued. Simon, 1925) the validity of the relation

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

gV note: "In my opionion Bose's derivation of the Planck

tang =-I(1 + MC
~ tang 9 (57) formula constitutes an important advance. The method
used here also yields the quantum theory of the ideal
gas, as I shall discuss elsewhere in more detail. "
in individual events, where P, 9 are the scattering
angles of the electron and photon, respectively, and v
is the incident frequency. It is not the purpose of this Section to discuss the
And so the last resistance to the photon came to an history of quantum statistics but rather to describe
Einstein's contribution to the subject. Nevertheless
end. Einstein's views had been fully vindicated. The
I include a brief outline of Bose's work, in part (B).
"
experimental news was generally received with great
relief (see, e. g. , Pauli, 1925a.). Bohr (1925b) took There are numerous reasons for doing so. (1) It will
give us some insight into what made Einstein diverge
the outcome in good grace and proposed "to give our
revolutionary efforts as honourable a funeral as pos- temporarily from his main pursuits. (2) It will facil-
"
sible. He was now prepared for an even more drastic itate the account of Einstein's omn research on the mo-
resolution of the quantum paradoxes. In July 1925 he
lecular gas. This work is discussed in part (C), with
the exception of one major point which is reserved for
wrote (1925c): "One must be prepared for the fact that
the next section. Einstein's last encounter with fluc-
the required generalization of the classical electro-
dynamical theory demands a profound revolution in the
tuation questions. (8) It will be of help to explain Ein-
stein's ambivalence to Bose's work. In a letter to
concepts on which the description of nature has until
now been founded. " Ehrenfest, written in July, Einstein (1924g) did not
These remarks by Bohr end with references to de withdraw, but rather qualified, his praise of Bose's
paper. ' Bose s "derivation is elegant but the essence
Broglie's thesis and also to Einstein's work on the
remains obscure. "(4) It will help to make clear how
quantum gas (the subject of the next section): the pro-
novel the photon concept still was at that time and it will
found revolution had begun.
throw an. interesting sidelight on the question of photon
VI. A LOSS OF IDENTITY: THE BIRTH OF QUANTUM spin.
STAT I ST I GS Bose recalled, many years later, that he had not been
aware of the extent to mhich his paper defied classical
A. Boltzmann's axiom logic. "I had no idea that what I had done was really
This episode begins with a letter (Bose, 1924a) dated novel. . . I was not a statistician to the extent of really
June 1924, written by a young Bengali physicist from knowing that I was doing something which was really
the University of Dacca, now in Bangladesh. His name different from what Boltzmann would have done, from
was Satyendra Nath Bose (1894—1974). The five papers Boltzmann statistics. Instead of thinking of the light-
he had published by then were of no particular distinc- quantum just as a particle, I talked about these states.
tion. The subject of his letter was his sixth paper. He Somehow this was the same question which Einstein
had sent it to the Philosophical Magazine. A referee
asked when I met him: How had I arrived at this meth-
had rejected it (Blanpied, 1972). Bose's letter was
od of deriving Planck's formula?" (Mehra. , 1975). In
addressed to Einstein, then forty-five years old, and order to answer Einstein's question it is necessary
already recognized as a world figure by his colleagues (though it may not be sufficient) to understand what gave
and by the public at large. In this section, I shall des- Bose the idea that he was doing what Ludwig Boltzmann
would have done. This in turn demands a brief digres-
cribe whathappenedin the scientific lives of these two
men during the half-year following Einstein's receipt sion on classical statistics.
of Bose's letter. For Bose the consequences were mo- Suppose I show someone two identical balls lying on a
mentous. From virtually an unknown he became a table. Next I ask him to close his eyes and a few mo-
physicist whose name mill always be remembered. For ments later to open them again. I then ask him whether
Einstein this period was only an interlude. 6 Already or not I have meanwhile exchanged the two balls. He
he mas deeply engrossed in a project which he was not cannot tell, since the balls are identical. Yet I do know
to complete. ' his search for a unified field theory. Such the answer. If I have exchanged the balls then I have
is the scope of Einstein's oeuvre that his discoveries been able to follom the continuous motion which brought
in that half-year do not even rank among his five main the balls from the initial to the final configuration. This
contributions, yet they alone would have sufficed to simple example illustrates Boltzmann's first axiom of
remember Einstein forever as well. classical mechanics which says, in essence, that iden-
Bose's sixth paper deals with a new derivation of tical particles which cannot come infinitely close to
Planck's law. Along with his letter he had sent Einstein each other can be distinguished by their initial conditions
a copy of his manuscript, written in English, and asked and by the continuity of their motion. This assumption,
him to arrange for publication in the Zeitschrift fur Boltzmann (1897) stressed, "gives us the sole possi-
Physik, if he were to think the work of sufficient merit. bility of recognizing the same material point at different
Einstein acceded to Bose's request. He personally times. "As Schroedinger (1955) has emphasized, "No-
translated the paper into German and submitted it to the body before Boltzmann held it necessary to define what
Zeitschrift fur Physik. He also added this translator's one means by [the term] the same material point. "

+Pauli's description of BKS, written early in 1925, is found 5Such a lack of awareness is not uncommon in times of
in Pauli (1964) pp. 83 —86. transition. But it is not the general rule. Einstein's light-
+In 1925 Einstein said of his work on quantum statistics: quantum paper of 1905 is an example of a brilliant exception.
"That's only by the way" (Salaman, 1979). In October or November 1925.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

Thus we may speak classically of a gas with energy counting by first Bose, then Einstein.
R consisting of N identical distinguishable molecules. (b) 1925-6: Discovery of nonrelativistic quantum
I et there be n& particles with energy 8&, . . . n,. par ti- mechanics. It is not at once obvious (see Heisenberg,
cles with energy g, , so that 1926a) how one should supplement the new theory with
a fine-grained counting principle which leads to BE
n, , E= n;c, (58) statistics.
(c) 1926. This principle is discoveredby PaulAdrien
The number of states corresponding to this partition is Maurice Dirac (1902-): Equation (59) is to be replaced
given by by

u =Nt n;t u =1, independent of the n;; (65)

only the single microstate which is symmetric in the N
where the factors jg,. reflect the application of Boltz-
1
particles is allowed. Dirac notes that (65) leads to
mann's axiom. These states are the microstates of the Planck's law (Dirac, 1926). After more than a quarter
system and the description in terms of them is the so- of a century the search for the foundations of this law
called fine-grained description. For the limited purpose comes to an end.
of analyzing the macroscopic proper ties of the sys tern Equation (65) was of course not known when Bose and
one contracts this description to the so-called coarse- Einstein embarked on their explorations of a new sta-
grained des cription. tistics. Let us next turn to Bose's contribution.
Divide the one-particle phase space into cells
such that a particle in (d& has the mean en- B. Bose
ergy E'~. Partition the N particles such that there are
N„particles in co» The paper by Bose (1924b) is the fourth and last of the
revolutionary papers of the old quantum theory. [The
N= N~, (60) other three are, respectively, by Planck (1900b), Ein-
stein (1905a) and Bohr (1913a).j Bose's arguments
(61) divest Planck's law of all supererogatory elements of
electromagnetic theory and base its derivation on the
The set (N„, &„) defines a coarse-grained state. For bare essentials. It is the thermal equilibrium law for
the special case of an ideal gas the relative probabil- particles with the following properties. they are mass-
ity of this state is given by less; they have two states of polarization; the number of
Ng particles is not conserved; and the particles obey a new
(62) statistics. In Bose's paper two new ideas enter physics
almost stealt, hily. One, the concept of a particle with
Equation (62) rests on two independent ingredients: two states of polarization, mildly puzzled Bose. The
(1) Boltzmann's distinguishability axiom and (2) sta- other is the non- conservation of photons. I do not know
tistical independence; that is, for an ideal classical whether Bose even noticed this fact. It is not explicitly
gas the individual molecules have no a priori preference mentioned in his paper.
for any particular region in phase space. Boltzmann's Bose's letter to Einstein (Bose, 1924a) begins as fol-
principle states that in thermodynamic equilibrium the lows: "Respected Sir, I have ventured to send you the
entropy is given by accompanying article for your perusal. I am anxious
to know what you will think of it. You will see that I
S=jpln W ~+C, have ventured to deduce the coefficient 8wv /c in
where C is an arbitrary additive constant. ; that is, C Planck's law independent of the classical electrody-
does not depend on the n;. W ~ is the maximum of « namics. . . . " Einstein's letter to Ehrenfest (Einstein,
as a function of the N~ subject to the constraints (60), 1924g) contains the phrase: "The Indian Bose has given
(61). The classical Boltzmann distribution then follows a beautiful derivation of Planck's law including the con-
from the extremal conditions stant [i. e. 8vp /c ]." Neither letter mentions the other
parts of Planck's formula. Why this emphasis on 8~v /
Q &N~( incr~ —ln N~ +X +P E~) =0, (64) C?
In deriving Planck's l.aw one needs to know the num-
and P =kT follows from BS/BE =T '. All these elemen- ber of states Z' in the frequency interval between v'
tary facts have been recalled in order to accentuate the and v'+dv'. It was customary to compute Z' by counting
differences between classical and quantum statistics. the number of standing craves in a cavity with volume
Both logically and historically classical statistics V which gives
developed via the sequence: Z' =8m(v')2Vdv'/c (66)
Pine-grained counting -coarse-grained counting.
Bose was so pleased because he had found a new deri-
This is of course the logic of quantum statistics as well, vation of this expression for &' which enabled him to
but its historical development went the reverse way, give a new meaning to this quantity in terms of Par-
from coarse-grained to fine-grained. For the oldest ticle language. His derivation res ts on the replacement:
.

quantum statistics, the Bose-Einstein (BE) statistics, counting wave frequencies -counting cells in one-par-
the historical order of events was as follows. ticle phase space. He proceeded as follows. Integrate
(a) 1924-5: Introduction of a new coarse grained the one-particle phase space element dxdp over V and

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

over all momenta between p' and p'+dp'. Supply a then derived Planck's law for E(v, T) by standard ma-
further factor two to count polarizations. Thisproduces nipulations —
and therewith concluded his paper without
the quantity 8m V(p') dp' which equals h Z' by virtue of further comments.
the relation p'=hv'/c. Hence Z' is the number of cells Bose (1924b) considered his Ansatz (71) to be "evi-
of size h3 which is contained in the particle phase space dent. " I venture to guess that to him the cell counting
region considered. How innocent it looks, yet how new (71) was the perfect analog of Boltzmann's particle
it was. Recall that the kinematics of the Compton ef- counting (62) and that his cell constraint: hold Z' fixed,
fect had only been written down about a year and a half was similarly the analog of Boltzmann's particle con-
earlier. Here was a new application of P =hv/c! straint: hold N fixed. Likewise the two Lagrange param-
Before I turn to the rest of Bose's derivation I should eters in (72) are his analogs of the parameters in (64).
like to digress briefly on the subject of photon spin. Bose's replacement of fixed& by fixed &' already implies
When Bose (1924b) introduced his polarization factor thatN is not conserved. The final irony is that the con-
of two he noted that "it seems required" to do so. This straint of fixed Z'is irrelevant: If one drops this con-
slight hesitation is understandable. Who in 1924 had straint thenone mustdrop X' in Eq. (72). Even so, it is
ever heard of a particle with two states of polarization? easily checked that one still finds Planck's law This is in
~

For some time this remained a rather obscure issue. accordance with the now familiar fact that Planck's law
After the discovery of the electron spin, Ehrenfest. follows from Bose statistics with E held fixed as the
(1926) asked Einstein "to tell me how the analogous hy- only constraint. In summary, Bose's derivation intro-
pothesis is to be stated for light-corpuscles, in a rel- duces three new features:
ativistically correct way. "
As is well known this is a
delicate problem since there exists, of course, no rest (1) Photon number nonconservation.
frame definition of spin in this instance. Moreover, (2) Bose's cell partition numbers p'„are defined by
gauge invariance renders ambiguous the separation into asking how many particles are in a cell, not which par-
orbital and intrinsic angular momentum (see, e. g. , ticles are in a cell. Boltzmann's axiom is gone.
Jauch and Rohrlich, 1955). It is not surprising that in (3) The Ansatz (71) implies statistical independence of
1926 the question of photon spin seemed quite confusing cells. Statistical independence of particles is gone.
to Einstein. In fact he went as far as to say that he was The astounding fact is that Bose was correct on all
"inclined to doubt whether the angular momentum law three counts. (In his paper he commented on none of
can be maintained in the quantum theory. At any rate them. ) I believe there has been no such successful shot
its significance is much less deep than the momentum in the dark since Planck introduced the quantum in 1900.
law" (Einstein, 1926a). I believe this is an interesting
comment on the state of the art some fifty years ago
and that otherwise not too much should be made of it. C. Einstein
Let us return to Bose. His new interpretation of Z'
was in terms of "number of cells, " not "number of par- As long as Einstein lived he never ceased to struggle
ticles". This must have led him to follow Boltzmann's with quantum physics. Insofar as his constructive con-
counting but to replace everywhere "particles" by tributions to this subject are concerned, these came
"cells, " a procedure which he neither did, nor could to an end with a connected triple of papers, the first
—
justify but which gave the right answer. He partitioned published in September 1924, the last two in early 1925.
Z' into numbers p„', where p'„ is defined as the number In the true Einstein style, their conclusions are once
of cells which contain r quanta with frequency v'. Let again reached by statistical methods, as was the case
there be ¹ photons in all with this frequency and let for all his important earlier contributions to the quan-
E be the total energy. Then tum theory. The best-known result is his derivation of
the Bose-Einstein (BE) condensation phenomenon. I
Zs pr s (67) shall discuss this topic next and shall leave for the sub-
sequent section another result contained in these papers
(68) which is perhaps not as widely remembered although it
is even more profound.
E=g N'@v' (69) First, a postscript to Einstein's light-quantum paper
of 1905. Its logic can be schematically represented in
while the following way.

(70) Einstein 1905: Wien's law

—Light- quanta.
is the total number of photons. Next Bose introduced Gas analogy
a new coarse-grained counting:
An issue raised at the end of Sec. II. D should be dealt
ZS )
(71) with now. We know that BE is the correct statistics
'S" POLYP~~ ~
0

when radiation is treated as a photon gas. Then how

He then maximized R' as a function of the p'„, holding could Einstein correctly have conjectured the exis-
Z' and E fixed so that tence of light-quanta using Boltzmann statistics? An-
swer: According to BE statistics, the most probable
I!1;If+in@;+v+—rhea'I=o, value Q;) of the n, for photons is given by (n, )
Q
S, r
S
(72)
=(exp(hv, /kT) —I) . This implies (n&) «1 in the Wien

Rev. Mod. Phys. , Vot. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

regime jgv,. » kT. Therefore up to an irrelevant factor with X =h /2wmkT and v= V/N. He then discusses the
¹ Equations (59) and (65) coincide in the Wien limit. regions «1 where the equation of state [obtained by
This asymptotic relation in the Wien region fully jus- eliminating A between the two equations (77)] shows
tifies, ex post facto, Einstein's extraordinary step for- perturbative deviations from the classical ideal gas.
ward in 1905 I All this is good physics, though unusually straightfor-
ward for a man like Einstein.
Bose"s reasoning in 1924 went as follows: In his second paper (1925a), the most important one
Bose 1924: Photons of the three, Einstein begins with the v —T relation at
A=1
) —Planck's law,
Quantum statistic~& (1)]"' ' (78)
2m[v, @,
and in 1924-5 Einstein came full circle: and asks what happens if T drops below To (for given
Einstein 1924 —5: Bose statistics vo). His answer:
"I maintain that in this case a number of molecules
!

-The quantum gas . steadily growing with incr'easing density goes over in the
Photon analog first quantum state (which has zero kinetic energy) while
the remaining molecules distribute themselves accord-
It was inevitable, one might say, that he would do so. ing to the parameter value A =1. . . a separation is
"If it is justified to conceive of radiation as a quantum
effected; one part condenses, the rest remains a 'sat-
gas, then the analogy between the quantum gas and a urated ideal gas'" (Einstein, 1925a). Einstein had come
molecular gas must be a complete one" (Einstein,
upon the first purely statistically derived example of a
1925a). phase transition which is now called Bose-Einstein con-
In his 1924 paper, Einstein (1924h) adopted Bose's
densation. I defer a few comments on this phenomenon
counting formula (71), but with two modifications. He
to the postscript at the end of this section and turn next
needed of course the Z' appropriate for nonrelativistic to other important facets of the three Einstein papers.
particles with mass m.' (1) Einstein on statistical dependence. After the
a'Z'=2~V(2m)'"(E')'" dE' 2mE'=(p')' papers by Bose (1924b) and the first one by Einstein
(73)
(1924h) had come out, Ehrenfest and others objected
Secondly (and unlike Bose! ) he needed the constraint tha, t (so we read in Einstein's second paper (1925a)) that "the
N is held fixed. This is done by adding a term quanta and molecules respectively are not treated as sta-
tistically independent, a fact which is not particularly
(74) emphasized in our papers" (i. e. , Bose, 1924b, and
inside the curly brackets of Eq. (72). 68 One of the con- Einstein, 1924h). Einstein replied (1925a): "This
sequences of the so-modified Eq. (72) is that the Lag- [objection] is entirely correct. " He went on to stress
range multiplier (- InA) is determined by that the differences between the Boltzmann and the BE
counting "express indirectly a certain hypothesis on a
mutual influence of the molecules which for the time
being is of a quite mysterious nature. " With this re-
Hence, Einstein noted, the "degeneracy parameter" mark Einstein came to the very threshold of the quan-
A must satisfy tum mechanics of identical particle systems. The mys-
terious influence is of course the correlation induced by
(75) the requirement of totally symmetric wave functions.
In his first paper (1924h) Einstein discussed the (2) Einstein on indistingniskability. In order to illus-
regime in which A does not reach the critical value one. trate further the differences between the new and the
He proceeds to the continuous limit in which the sum old counting of macrostates, Einstein (1925a) cast W
in Eq. (75) is replaced by an integral over phase space in a form alternative to Eq. (71). He counted the num-
and finds69 ber of ways in which N' indistinguishable particles in the
dE'-interval can be partitioned over the Z' cells. This
43m(+) p @5i2 &) (77) yields
3 'yT 3
. (N' + Z' —1)!
N't (Z'- 1)!
Einstein's Eq. (79) rather than Bose's Eq. (71) is the
one now used in all textbooks.
6~The%1 is irrelevant since it only affects C in Eq. (63). The (3) Einstein on the third lan& of thermodynamics. As
constant C is interesting nevertheless. For example, its value was noted at the end of Sec. III.C, in 1914 Nernst in-
bears on the possibility of defining S in such a way that it be- troduced the hypothesis that the third law of thermo-
comes an extensive thermodynamic variable. The interesting
dynamics applies to gases. It was also mentioned that
history of these normalization questions has been discussed in
detail by M. Klein (1958). no sensible model of a gas with that property was avail-
~:
A —exp( — p/k T), p is the chemical potential. Einstein of able at thai time. In 1925 Einstein made his last. con-
course never introduced the superfluous A, ~ in these curly tribution to thermodynamics by pointing out that the BE
brackets. In Eqs. (74) —(79) I deviate from Einstein's notations. gas does satisfy the third law. (A Boltzmann gas does
All technical steps are now found in standard textbooks. not do so, Einstein remarked. ) Indeed, since all par-

Rev. Mod. Phys. , Vol. 5't, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

ticles go into the zero energy state as T-O, we have tion. ]Uhlenbeck and Einstein were both right, how-
in this limit N =N, all other iV'=0. Hence W-I and ever. The point is that a sharp phase transition can
S-0 as T-0. It was as important to him that a mo- only occur in the so-called thermodynamic limit
lecular BE gas yield Nernst's law as that a BE photon N-~, V-~, v fixed. This view emerged in a morning-
gas yield Planck's law. long debate which took place during the van der %Vaals
(4) &instein and nonconsexvation of photons. After Centenary Conference in November 1937. The issue
1917 Einstein ceased to write scientific articles on was: does the partition function contain the information
questions related to radiation. ~0 The only mention of necessary to describe a sharp phase transition? The
radiation in the 1924/5 papers is that "the statistical transition implies the existence of analytically distinct
method of Herr Bose and myself is by no means beyond parts of isotherms. It was not clear how this could
doubt. , but seems only a posteriori justified by its suc- come about. The debate was inconclusive and Kramers,
cess for the case of'radiation" (Einstein, 1925b). the chairman, put the question to a vote. Vhlenbeck
There can be no doubt that he must have noted the recalls that the ayes and nays were about evenly divided.
nonconservation of photons. In his language this is im- However, Kramers' suggestion to go to the thermody-
plemented by putting A =1 in Eq. (74). Yet I have not namic limit eventually was realized to be the correct
found any reference to nonconservation, either in his answer. Shortly afterward, Uhlenbeck withdrew his
scientific writings or in the correspondence I have seen. objections to Einstein's result, in a joint paper with his
I cannot state with certainty why he chose to be silent gifted student, the late Boris Kahn (a Nazi victim)
on this and all further issues regarding photons. How- (Kahn and Uhlenbeck, 1938).
ever I do believe it is a fair guess that Einstein felt he (3) Until 1928 the BE condensation had "the reputation
would have nothing fundamental to say about photons of having only a purely imaginary character" (London,
until such time as he could find his own way of dealing 1938a). Recall that the HeI-HeII phase transition was
with the lack of causality he had noted in 1917. not discovered until 1928, by %'illem Hendrik Keesom
Such a time never came. (1876-1956) (see Keesom, 1942). In 1938 Fritz London
(1900—1954) proposed to interpret this He-transition as
l3. Postscripta on Bose-Einstein condensation a BE condensation. Experimentally, the transition point
lies at 2. 19'K. It is most encouraging that Eq. (78)
(1) In 1925 Einstein (1925a) mentioned hydrogen, he- gives T=3. 1'K (London, 1938b). It is generally be-
lium, and the electron gas as the best possible can- lieved but not proved that the difference between these
didates to observe his condensation phenomenon. In
two values is due to the neglect of intermolecular forces
1925 these were, of course, sensible proposals. Recall
in the theoretical derivations.
that the Fermi Dirac statistics was not discovered
until 1926 (Fermi, 1926; Dirac, 1926), following Pauli's
enunciation of the exclusion principle in 1925 (Pauli,
1925b). Even then it took some time until it was sorted Vll. ElNSTEIN AS A TRANSITIONAL FIGURE:
THE BIRTH OF WAVE MECHANICS
out when BE and FD statistics apply respectively. Re-
ferring to Dirac's paper, (1926) Pauli (1927) wrote in
December 1926: "%e shall take the point. of view also %e now leave the period of the old quantum theory and
—
advocated by Dirac, that the Fermi and not. the Ein- turn to the time of transition during which matter waves
stein-Bose statistics applies to the material gas. These " were being discussed by a tiny group of physicists at a
matters were cleared up by 1927. time when matter wave mechanics had not yet been dis-
(2) In his 1925 paper, Einstein (1925a) did not call the covered. This period begins in September 1923 with
condensation phenomenon a phase transition. According two brief communications (1923b, 1923c) by Louis de
to George Eugene Uhlenbeck (1900—) nobody realized in Broglie (1892 —) to the French Academy of Sciences. It
1925 that the existence of a phase transition was a "deep" ends in January 1926 with Erwin Schroedinger's (1887—
problem (private communication). In 1926 Uhlenbeck 1961)firstpaper(1962a) on wave mechanics. The main
(1927) himself raised an objection to Einstein s treat- purpose of this section is to stress Einstein's key role in
ment of the condensation problem. This critique was these developments, his influence on de Broglie, de
to lead to a more precise theoretical formulation of the Broglie's subsequent influence on him, and finally, the
conditions under which phase transitions can occur. influence of both on Schroedinger.
Uhlenbeck noted that the quantity N in Eq. (75) — ~ as Neither directly nor indirectly did Einstein contri-
A -1 ™.
(for fixed T), hence also N — Thus if A =1 it bute to an equally fundamental development which pre-
ceded Schroedinger's discovery of wave mechanics: the
is impossible to implement the constraint that N is a
fixed finite number. Therefore A = 1 can only be reached discovery of matrix mechanics by Heisenberg (1926b).
asymptotically and there is no two phase regime. Therefore I shall have no occasion in this article to
IUhlenbeck (1979) has recently described the communi- comment in any detail on Heisenberg's major achieve-
cations between Ehrenfest and Einstein on this ques- ments.

A. From Einstein to de Broglie

Except for a brief refutation of an objection to his work on
needle radiation (Einstein, 1925c). I found a notice by Ein-
stein (1930a} in 1930 announcing a new paper on radiation During the period which began with Einstein's work
fluctuations. This paper was never published, however. on needle rays (1917) and ended with Debye's and
~
Dirac (1977) has given a charming account of the time se- Compton's papers on the Compton effect (1923) there
quence of these discoveries. were a. few other theoreticians who did research on

Rev. Mod. Phys. , Vol. O'I, No. 4, October 1979

898 A. Pais: Einstein and the quantum theory

photon questions. Among those the only one~ whose Put Vp dv =n(v)hv and z' = D(v) (hv) . n(v) can be inter-
contribution lasted was de Broglie. preted as the average number of quanta in the energy
de Broglie had finished his studies before the First interval dv, and ~(v) as the mean square fluctuation of
World War. In 1919, after a long tour of duty with the this number. One can now write (37) as
French forces he joined the physics laboratory headed
by his brother Maurice (1875—1960), where x-ray (80)
photoeffects and x-ray spectroscopy were the main
topics of study. Thus he was much exposed to questions where Z(v) is the number of states per interval dv given
conerning the nature of electromagnetic radiation, a in Eq. (66). In his paper submitted on January 8, 1925,
subject on which he published several papers. In one Einstein (1925a) showed that Eq. (80) holds equally well
of these de Broglie (1923a) evaluated independently of for his quantum gas, as long as one defines v in the
Bose (and published before him) the density of radia- latter case by E=hv=P /2m and uses Eq. (73) instead
tion states in terms of particle (photon) language (see of Eq. (66) for the number of states.
Sec. VI. B). That was in October 1923-one month after When discussing radiation, in 1909, Einstein recog-
his enunciation of the epochal new principle that par- nized the second term of Eq. (37) as the familiar wave
ticle-wave duality should apply not only to radiation but term and the first one as the unfamiliar particle term.
also to matter. "After long reflection in solitude and When, in 1925, he revisited the fluctuation problem for
meditation I suddenly had the idea, during the year the case of the quantum gas, he noted a reversal of
1923, that the discovery made by Einstein in1905 should roles. The first term, at one time unfamiliar for ra-
be generalized in extending it to all material particles diation, was now the old fluctuation term for a Poisson
and notably to electrons" (de Broglie, 1963). distribution of (distinguishable) particles. What to do
He made the leap in his September 10, 1923 paper with the second term (which incorporates indistinguish-
(1923b): E =hv shall hold not only for photons but also ability effects of particles) for the gas case'? Since this
for electrons, to which he assigns a "fictitious asso- term was associated with waves in the case of radiation,
ciated wave. "In his September 24 paper (1923c) he Einstein (1925a) was led to "interpret it in a correspon-
indicated the direction in which one "should seek ex- ding way for the gas, by associating with the gas a ra-
perimental confirmations of our ideas": A stream of diative phenomenon. "
He added: "I pursue this inter-
electrons traversing an aperture whose dimensions are pretation further since I believe that here we have to do
small compared with the wavelength of the electron with more than a mere analogy. "
waves "should show diffraction phenomena. " &ut zehat uence the scares~
Other important, aspects of de Broglie's work are At that point Einstein turned to de Broglie's thesis
beyond the scope of this paper. The mentioned articles (see de Broglie, 1963), "a very notable publication. "
were extended to form his doctor's thesis (see de He suggested that a de Broglie-type wave field should
Broglie, 1963) which he defended on November 25, 1924. be associated with the gas and pointed out that this as-
Einstein received a copy of this thesis from one of de sumption enabled him to interpret the second term in
Broglie's examiners, Paul Langevin (1872 —1946). A Eq. (80). Just as de Broglie had done, he also noted
letter to Lorentz (in December) shows that Einstein that a molecular beam should show diffraction phenom-
(1924i) was impressed and also that he had found a new ena but added that the effect should be extremely small
application of de Broglie's ideas, ' "A younger brother for manageable apertures. Einstein also remarked that
of. . . de Broglie has undertaken a very interesting the de Broglie-wave field had to be a scalar. [The po-
attempt to interpret the Bohr-Sommerfeld quantum rules larization factor equals two for Eq. (66), as noted
(Paris Dissertation 1924). I believe it is a first feeble above, but it equals one for Eq. (73)!].
ray of light on this worst of our physics enigmas. I too It is another Einstein feat that he would be led to state
have found something which speaks for his construction. " the necessity of the existence of matter waves from the
analysis of fluctuations. One may wonder what the his-
B. From de Broglie to Einstein tory of twentieth century physics would have been like
had Einstein pushed the analogy still further. However,
In 1909 and again in 1917 Einstein had drawn major
with the achievement of an independent argument for the
conclusions about radiation from the study of fluctua--
It goes without particle-wave duality of matter, the twenty-year period
tions around thermal equilibrium.
of highest scientific creativity in Einstein s life, at a
saying that he mould again examine fluctuations when,
level probably never equalled, came to an end.
in 1924, he turned his attention to the molecular quan-
Postscript, summer of Z978. In the course of pre-
tum gas. paring this article I noticed a recollection by Pauli
In order to appreciate what he did this time it is help-
(1949) of a statement made by Einstein during a physics
ful to copy, one last time, the formula givenearlier,
meeting held in Innsbruck in 1924. According to Pauli,
for the mean square energy of electromagnetic radia- Einstein proposed in the course of that meeting "to
tion (see Sec. II. G):
search for interference and diffraction phenomena mith
molecular beams. "
On checking the dates of that
(37)
meeting I found them to be September 21-27. This in-
trigued me. Einstein came to the particle-wave duality
of matter via a route which was independent of the one

7 The
other ones I know of are Brillouin {1921),Wolfke (1921),
Bothe (1923), Bateman (1923), Ornstein and Zernike (1919). ~3In Eq. (80) I drop the index p occurring in Eq. (66).

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

taken by de Broglie. The latter defended his thesis in Now back to Schroedinger. By his own admission
November. If Pauli's memory is correct, then Einstein (Schroedinger, 1926b) he was not much taken with the
made his remark about two months prior to that time. new BE statistics. Instead, he suggested, why not
Could Einstein have come upon the wave properties of evade the new statistics by treating Einstein's molec-
matter independently of de Broglie? After all, Ein- ular gas according to the Debye method? That is, seky
stein had been thinking about the molecular gas since' not start f~oypz a wave Picture of the gas and super-
July. The questions arise: When did Einstein become impose on that a quantization condition ~ la Debye? Now
aware of de Broglie's work? In particular, when did comes the key sentence in the article (1926b):
Einstein receive de Broglie's thesis from Langevin? means nothing else but taking seriously the. de Broglie-
Clearly it would be most interesting to know what Pro- Einstein wave theory of moving particles. " And that is
fessor de Broglie mighthaveto say about these ques- just what Schroedinger did. It is not necessary to dis-
tions. Accordingly I wrote to him. He was kind enough cuss further details of this article, which was received
to reply. With his permission I quote from his answers. by the publisher on December 25, 1925.
de Broglie does not believe that Einstein was aware Schroedinger's next paper (1926c) was received on
of his three short publications (1923a, b, c) written in January 2V, 1926. It contains his equation for the hy-
1923. "Nevertheless, since Einstein would receive the drogen atom. Wave mechanics was born. In this new
Comptes Rendus and since he knew French very well, he paper Schroedinger (1926c, p. 373) acknowledged his
might have noticed my articles" (de Broglie, 1978a). de debt to de Broglie and Einstein: "I have recently shown
Broglie noted further that he had given Langevin the first [1926b) that the Einstein gas theory' can be founded on
typed copy of his thesis early in 1924. "Iam certain that the consideration of standing waves which obey the
Einsteinknew of my These since the spring of 1924" (de dispersion law of de Broglie. . . . The above consid-
Broglie, 1978 b). This is what happened. "When in erations about the atom could have been presented as
1923 I had written the text of the These de Doctorat a generalization of these considerations". In April 1926
which I wanted to present in order to obtain the Doctorat Schroedinger (1926d) again acknowledged the influence
es Sciences, I had three typed copies made. I handed of de Broglie and "brief but infinitely far seeing
one of these to. M. Langevin so that he might decide remarks by Einstein. "
whether this text could be accepted as a Thyrse. M.
Langevin, 'probablement un peu etonne par la nouveaute
de mes idees, "asked me to furnish him withasecond Vill. EINSTEIN'S RESPONSE TO THE NEVV DYNAlVIICS
typed copy of my These for transmittal to Einstein. It
was then that Einstein declared, after having read my Everyone familiar with modern physics knows that
work, that my ideas seemed quite interesting to him. Einstein's attitude regarding quantum mechanics was
This made Langevin decide to accept my work" (de one of skepticism. No biography of him fails to mention
Broglie, 1978a). his saying that God does not throw dice. He was indeed
Thus Einstein was not only one of the three fathers of given to such utterances (as 1 know from experience),
the quantum theory, but also the sole godfather of wave and stronger ones such as "It seems hard to look in
mechanics. God's cards. But I cannot for a moment believe that
he plays dice and makes use of 'telepathic'means (as the
C. From de Broglie and Einstein to Schroedinger current quantum theory alleges he does" (Einstein,
Late in 1925 Schroedinger completed an article (1926b) 1942b). However, remarks such as these should not
entitled "On Einstein's gas theory. "
It was his last create the impression that Einstein had abandoned active
paper prior to his discovery of wave mechanics. Its interest in quantum problems in favor of his quest for a
content is crucial to an understanding of the genesis of unified field theory. Far from it. In fact even in the
that discovery (Klein, 1964). search for a unified theory the quantum riddles were
In order to- follow Schroedinger's reasoning it is nec- very much on his mind, as I shall discuss later on in
essary to recall first a derivation of Planck's formula Sec. X. In the present section, I shall attempt to des-
given. by Debye (1910) in 1910. Consider a cavity filled cribe how Einstein s position concerning quantum mech-
with radiation oscillators in thermal equilibrium. The anics evolved in the course of time. To some extent
spectral density is 8vrv2s(v, T)lc3. s is the equilibrium this is reflected in his later scientific papers. It be-
energy of a radiation field oscillator with frequency v.
'
comes evident more fully in several of his more auto-
Debye introduced the quantum prescription that the biographical writings and in his correspondence. My
only admissible energies of the oscillator shall be own understanding of his views have been helped much
.
nkv, n =0, 1, 2, . . . In equilibrium the nth energy level by discussions with him.
is weighted with its Boltzmann factor. Hence c To begin with I shall turn to the period 1925-1933
=S~nhvy "/5 y", y =exp(-hv/hT). This yields Planck's during which he was much concerned with the question:
law is quantum mechanics consistent?

7@'Probably a bit astonished by the novelty of my ideas. " A. 1925-1933. The debate begins
7~This derivation differs from Planck's in that the latter
quantized material rather than radiation oscillators. It differs As I mentioned earlier, the three papers of 1924-5
from the photon gas derivation in that the energy nhv is inter- on the quantum gas were only a temporary digression
preted as the nth state of a single oscillator, not as a state of from Einstein's program begun several years earlier
n particles each with energy hv. to unify gravitation with electromagnetism. During the

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

900 A. Pais: Einstein and the quantum theory

very early days of quantum mechanics~6 we find him now a singular tension in us sluggish people. " The next
"working strenuously on the further development of a month he expressed again his conviction that the
theory on the connection between gravitation and elec- Heisenberg —Born approach was off the track. That was
tricity" (Einstein, 1925d). Yet the great importance in a letter in which he congratulated Schroedinger on
of the new developments in quantum theory was not lost his new advance (Einstein, 1926c). In view of the sci-
on him. Bose, who visited Berlin in November 1925, entific links between Einstein's and Schroedinger's work
has recalled (Mehra, 1975) that "Einstein was very it is not surprising that Einstein (1926d) would express
excited about the new quantum mechanics. He wanted real enthusias m about wave mechanics: "Schroedinge r
me to try to see what the statistics of light-quanta and has come out with a pair of wonderful papers on the
the transition probabilities of radiation would look like quantum rules. "It was the last time he would write
in the new theory. "~7 approvingly about quantum mechanics.
Einstein's deep interest in quantum mechanics must
There came a parting of ways.
have led him to write to Heisenberg rather soon after
the latter's paper (1926b) had been published. All the Nearly a year passed after Heisenberg's paper before
letters of Einstein to Heisenberg have beenlost. How- there was a first clarification of the conceptual basis
ever, a number of letters by Heisenberg to Einstein of quantum mechanics. It began with Max Born's
are extant. One of these (dated November 30, 1925) is (1882 —1970) observation (Born, 1926) in June 1926 that
clearly in response to an earlier lost letter by Einstein the absolute square of a Schroedinger wave function is
to Heisenberg in which Einstein appears to have com- to be interpreted as a probability density. Born's
mented on the new quantum mechanics. ~~ One remark brief and fundamental paper goes to the heart of the
by Heisenberg (1925) is of particular interest. "You are problem of determinism. Regarding atomic collisions
probably right. that our formulation of the quantum he wrote: "One does not get an answer to the ques tion
mechanics is more adapted to the Bohr-Kramers- "what is the state after collision" but only to the ques-
Slater attitude; but this [BKS theory] constitutes in fact tion "how probable is a given effect of the collision". . . .
one aspect of the radiation phenomena. The other is From the s tandpoint of our quantum mechanics there is
your light-quantum theory and we have the hope that the no quantity [Grosze] which causally fixes the effect
validity of the energy and momentum laws in our quan- of a collision in an individual event. Should we hope to
tum mechanics will one day make possible the connection discover such properties later. . . and determine
with your theory. " I find it remarkable that Einstein [them] in individual events. . . . . I myself am inclined
apparently sensed that there was some connection be- to renounce determinism in the atomic world. But that
tween the BKS theory and quantum mechanics. No is a philosophical question for which physical argu-
such connection exists of course. Nevertheless the ments alone do not set standards. "
BKS proposal contains s tatis tical features, as we have Born's paper had a mixed initial reception. Several
seen. Could Einstein have surmised as early as 1925 leading physicists found it hard if not impossible to
that some statistical element is inherent in the quantum- swallow the abandonment of causality in the classical
mechanical description'P &
sense, among them Schroedinger. More than once
During the following months, Einstein vacillated in his Bohr mentioned to me that Schroedinger told him he
reaction to the Heisenberg theory. In December 1925 might not have published his papers had he been able to
he expressed misgivings (Einstein, 1925e). But in. foresee what consequences they would unleash. Ein-
March 1926 he wrote to the Horns (Einstein, 1926b): stein's position in the years to follow can be summa-
"The Heisenberg-Born concepts leave us all br eathless, rized succinctly by saying that he took exception to every
and have made a deep impression on all theoretically single statement contained in the lines I have quoted
oriented people. Instead of a dull resignation, there is from Born. His earliest expressions of discomfort I
know of date from late 1926 when he wrote Born (Ein-
stein, 1926e): "Quantum mechanics is certainly im-
~68ecall that Heisenberg's first paper on this subject (1926b) posing. But an inner voice tells me that it is not yet
was completed in July 1925, Schroedinger's (1926e) in January the real thing. The theory says a lot, but does not re-
1926. ally bring us closer to the secret of the 'old one'. "
It was not for Bose but for Dirae to answer this question. "Einstein's verdict. . . . came as a hard blow" to
In his 1927 paper which founded quantum electrodynamics Born (Born, 1971, pg. 1). There are other instances as
Dirac (1927) gave the dynamical derivation of expressions for well in which E ins tein s reactions were expe rienced with
the Einstein A and B coefficients. a sense of loss, of being abandoned by a venerated lead-
According to Helen Du. kas. The letters by Heisenberg to
Einstein referred to are now in the Einstein Archives in er in ba. ttle. Thus Samuel Goudsmit (1902—1978) told
Princeton. me of a conversation which took place in mid-192V [to
7
Heisenberg (1925) begins by thanking Einstein for his letter the best of his recollection (Goudsmit, 1978)] between
and then proceeds with a rather lengthy discussion of the role Ehrenfest and himself. In tears, Ehrenfest said that
of the zero point energy in the new theory. This seems to be in he had to make a choice between Bohr's and Einstein's
response to a point raised by Einstein. It is not clear to me position and that he could not but agree with Bohr.
from Heisenberg's letter what Einstein had in mind. Needless to say, Einstein's reac. tions affected the older
Heisenberg remarked much later {1955)that "the attempt at generation more intensely than the younger.
interpretation by Bohr, Kramers and Slater nevertheless con-
tained some very important features of the later, correct inter-
pretation [of quantum mechanics]. "
I do not share this view
but shall not argue this issue beyond what has been said in Sehroedinger retained reservations on the interpretation of
Sec. III. quantum mechanics for the rest of his life (see Scott, 1967).

Rev. Mod. Phys. , Vol. 5], No. 4, October 1979

 A. Pais: Einstein and the quantum theory 901

Of the many important events in 1927, three are par- one saw that it did not work. . .
. Pauli and Heisenberg
ticularly significant for the present account. who were there did not react to these matters, "ach
March 2927. Heisenberg states the uncertainty prin- was, das stimmt schon, das stimmt schon" [ah well,
ciple. In June Heisenberg (1927a) writes a letter to it will be alright, it will be alright]. Bohr on the other
Einstein which begins as follows: "Many cordial thanks hand reflected on it with care and in the evening, at
for your kind letter; although I really do not know any- dinner, we were all together and he cleared up the mat-
thing new, I would nevertheless like to write once more ter in detail, "
why I believe that indeterminism, that is the nonvalidity Thus began the great debate between Bohr and Ein-
of rigorous causality, is necessary [his italics] and not stein. Both men refined and sharpened their positions
just consistently possible. " This letter is apparently in in the course of time. No agreement between them was
response to another lost letter by Einstein triggered, ever reached. Between 1925 and 1931 the only objection
most probably, by Heisenberg's work in March. I shall by Einstein which appeared in print in the scientific
return to Heisenberg's important letter in Sec. X. I literature is the one at 0he 1927 Solvay Conference
only mention its existence at this point to emphasize (Einstein, 1928). However, there exists a, masterful
that once again Einstein did not react to these new de- account of the Bohr-Einstein dialog during these years,
velopments as a passive bystander. In fact at just about published by Bohr (1949) in 1949. I have written else-
that time he was doing his own research on quantum where about the profound role which the discussions with
mechanics (his first, I believe). "Does Schroedinger's Einstein played in Bohr's life (Pais, 1967).
wave mechanics determine the motion of a system com- The record of the Solvay meeting contains only minor
pletely or only in the statistical sense?"" he asked. reactions to E instein's comments. Bohr's later article
Heisenberg had heard indirectly that Einstein "had written (1949) analyzed them in detail. I et us consider next
a paper in which you. . . . advocate the view that it the substance of E inste in's remarks.
should be possible after all to know the orbj. ts of par-
ticles more precisely than I would wish. " He asked for Einstein's opening phrase (1928) tells more about him
more information "especially because I myself have than does many a book: "Je dois m'excuser de n'avoir
thought so much about these questions and only came to pas approfondi la mecanique des quanta" ("I must apolo-
believe in the uncertainty relations after many pangs of gize for not having examined quantum mechanics in
conscience, though now I am entirely convinced" depth").
(Heisenberg, 1927b). Einstein eventually withdrew his
He then discusses an experiment in which a beam of
paper.
SePtember 26, 1927. At the Volta meeting in Corno electrons hits a (fixed) screen with an aperture in it.
Bohr (1928) enunciates for the first time the principle The transmitted electrons form a diffraction pattern
of complementarity: "The very nature of the quantum which is observed on a second screen. Question: does
theory. . . forces us to regard the space-time coor- quantum mechanics give a complete description of the
dination and the claim of causality, the union of which individuat electron events in this experiment P His ans-
characterises the classical theories, as complementary wer: This cannot be. For let A and E be two distinct
but exclusive features of the description, symbolising spots on the second screen. If I know that an individual
the idealisation of observation and definition respec- electron arrives at A then I know instantaneously that it
tively. " did not arrive at B. But this implies a peculiar instan-
October l'927. The fif th Solvay Conference convenes. taneous action at a distance between A and B contrary
All the founders of the quantum theory were there, from to the relativity postulate. Yet (Einstein notes) in the
Planck, Einstein, and Bohr to de Broglie, Heisenberg, Geiger-Bothe experiment on the Compton effect (Bothe
Schroedinger, and Dirac. During the sessions "Ein- and Geiger, 1925a, b) there is no limitation of principle
stein said hardly anything beyond presenting a very to the accuracy with which one can observe coincidences
simple objection to the probability interpretation. . . in individual processes, and that without appeal to ac-
Then he fell back into silence" (de Broglie, 1962). tion at a distance. This c ircumstance adds to the sense
However, the formal meetings were not the only place of incompleteness of the description for diffraction.
of discussion. All participants were housed in the same Quantum mechanics provides the following answer to
hotel and there, in the dining room, Einstein was much Einstein's query. It does apply to individual processes
livelier. Otto Stern has given this first-hand account but the uncertainty principle defines and delimits the
"Eins tein came down to breakfast and expressed his mis- optimal amount of information which is obtainable in a
givings about the new quantum theory, every time [he] given exPerimental arrangement. This delimitation dif-
had invented some beautiful experiment from which fers incomparably from the restrictions on information
inherent in the coarse-grained description of events in
classical statistical mechanics. There the restrictions
This is the title of a paper w'hich Einstein submitted at the are wisely self-imposed in order to obtain a useful ap-
May 5, 1927 meeting of the Prussian Academy in Berlin. proximation to a description in terms of an ideally
3The records show that this paper was in print when Einstein knowable complete specification of momenta and posi-
requested by telephone that it be withdrawn. The unpublished tions of individual particles. In quantum mechanics the
manuscript is in the Einstein Archives. I would like to thank
Dr. John Stachel for bringing this material to my attention. delimitations mentioned earlier are not self-imposed
+Einstein had been invited to this meeting but did not attend. but are renunciations of first principle (on the fine-
5In a discussion with R. Jost, taped on December 2, 1961. grained level, one might say). It is true that one would
I am very grateful to R. Jost for making available to me a need action at a distance if one were to insist on a fully
transcript of part of this discussion. causal description involving the localization of the elec-

Rev. Mod. Phys. , Vol. 51, No. 4, October 'f979

 A. Pais: Einstein and the quantum theory

tron at every stage of the experiment in hand. Quantum

mechanics denies that such a description is called for
and asserts that in this exPemment the final position
of an individual electron cannot be predicted with cer-
tainty. Quantum mechanics nevertheless makes a pre-
diction in this case, concerning the probability for an
electron to arrive at a given spot on the second screen.
The verification of this prediction demands of course
that the "one electron experiment" be repeated as often
as is necessary to obtain this probability distribution
with a desired accuracy.
Nor is there a conflict with Geiger-Bothe, since now
one refers to another experimental arrangement in
which localization in space-time is achieved, but this
time at the price of renouncing information on sharp
energy-momentum properties of the particles observed
in coincidence. From the point of view of quantum me-
chanics these renunciations are expressions of laws of
nature. They are also applications of the saying "Il
faut reculer pour mieux sauter" ("It is necessary to
take a step back in order to jump better"). As we shall
Q4E l~
see, what was and is an accepted renunciation to others
was an intolerable abdication in Einstein's eyes. On l
this score he was never prepared to give up anything.
I have dwelt at some length on this simple problem
since it contains the germ of Einstein's position which
he stated more explicitly in later years. Meanwhile
the debate in the corridors between Bohr and Einstein
continued during the sixth Solvay conference (on magne-
tism) in 1930. This time Einstein thought he had found FIG. 4. Bohr's drawing of Einstein's clock-in-the-box experi-
a counterexample to the uncertainty principle. The ment (Bohr, 1949). (Reproduced with the kind permission of
argument was ingenious. Consider a box with a hole in Professor A. Schilpp. )
its walls which can be opened or closed by a shutter time taken to readjust the pointer, and g is the gravita-
controlled by a clock inside the box. The box is filled tional acceleration. Thus Tg~m Aq ~ h. Next Bohr used
with radiation. Weigh the box. Set the shutter to open the red shift formula"': the uncertainty Aq
for a brief interval during which a single photon es- of the position of the clock in the gravitational field im-
capes. Weigh the box again, some time later. Then {in plies an uncertainty AT =c 2gT Dq in the determination
principle) one has found to arbitrary accuracy both the
photon energy and its time of passage, in conflict with
of T. Hence c'~~ AT =~ ~T» h. Thus the accuracy
with which the energy of the photonis measured restricts
the energy-time uncertainty principle. the precision with which its moment of escape can be
"It was quite a shock for Bohr . . . he did not see the determined, in accordance with the uncertainty relations
solution at once. During the whole evening he was ex- for energy and time. (Note that every one of the many
tremely unhappy, going from one to the other and trying details in Fig. 4 serves an experimental purpose: the
to persuade them that it couldn't be true, that it would heavy bolts fix the position of the scale along which the
be the end of physics if Einstein were right; but he pointer moves; the spring guarantees the mobility of
couldn't produce any refutation. I shall never forget the the box in the gravitational field; the weight attached to
vision of the two antagonists leaving the club [of the the box serves to readjust its position; and so on.
Fondation Universitaire]: Einstein a tall majestic fig- There was nothing fanciful in Bohr's insistence on such
ure, walking quietly, with a somewhat ironical smile, details. Rather he had them drawn in order to illus-
and Bohr trotting near him, very excited .. . The next . trate that, since the results of all physical measure-
morning came Bohr's triumph" (Rosenfeld, 1968). ments are expressed in classical language„ it is neces-
Bohr (1949) later illustrated his argument with the
sary to specify in detail the tools of measurement in that
help of the experimental arrangement reproduced in same language as well. )
Fig. 4. The initial weighing is performed by recording After this refutation by Bohr, Einstein ceased his
the position of the pointer attached to the box relative search for inconsistencies. By 1931 his position onquan-
to the scale attached to the fixed frame. The loss of tum mechanics had undergone a marked change.
weight due to the escape of the photon is compensated First of all, his next paper on quantum mechanics
by a load (hungunderneath the box) which returns the (Einstein, Tolman, and Podolsky, 1931), submitted in
pointer to its initial position with a latitude Aq. Cor- February 1931, shows that he had accepted Bohr's
respondingly, the weight measurement has an uncer-
tainty d ~. The added load imparts a momentum to the ~
Becall that the only ingredients for the derivation of this
box which we can measure with an a, ccuracy 4p delimited formula are the special relativistic tixne dilation and the equi-
by ~ Aq=k. Obviously ~& Tg ~m, where T is the valence principle.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 903

criticism. " It deals with a new variant of the clock-in By 1933 Einstein stated explicitly his conviction that
-the-box experiment. Experimental information about quantum mechanics does not contain logical contradic-
one particle is used to make predictions about a second tions. In his Spencer lecture (1933) he said of the
particle. This paper, a forerunner of the Einstein- Schroedinger wave functions: "These functions are only
Podolsky-Rosen article to be discussed below, need not supposed to determine in a mathematical way the prob-
be remembered for its conclusions. abilities of encountering those objects in a particular
A far more important expression of Einstein's opin- place or in a particular state of motion, if we make a.
ions is found in a letter he wrote the following Septem- measurement. This conception is logically unexception-
ber. In this letter, addressed to the Nobel committee able and has led to important successes. "
in Stockholm, Einstein (1931a) nominates Heisenberg It was in 1935 that Einstein stated for the first time
and Schroedlnger for the Nobel Prize. In his motivation h is own des ide rata in a precise f orm. This is the c r i-
he says about quantum mechanics: "Diese Lehre ent- terion of objective reality. He continued to subscribe
hilt nach meiner Uberzeugung ohne Zweifel ein Stuck to this for the rest of his life.
endgultiger Wahrheit. ""
Einstein himself was never 1935 was also the year in which Einstein moved into
greatly stirred by honors and distinctions. Even so, his his final home, 112 Mercer Street in Princeton. In
nominations do not only reveal extraordinarily clearly 1932 he had been appointed to a professorship at the
what his thoughts were but are also deeply moving as an Institute for Advanced Study. The original intent was
expression of his freedom of spirit and generosity of that each year he would divide his time evenly between
mind. They show that Einstein had come to accept that Princeton and Berlin. In December 1932 he left Ger-
quantum mechanics was not an aberration but rather a many for a visit to Caltech. He never set foot in Ger-
truly professional contribution to physics. many again, for well-known reasons. He and his family
Not that from then on he desisted from criticizing spent the summer of 1933 in Belgium. On October 17
quantum mechanics. He had recognized it to be part they arrived back in the United States and came to
of the truth but was and forever remained deeply con- Princeton that same day. Shortly thereafter they moved
vinced that it was not the whole truth. From 1931 on, to 2 Library P lace and f rom the re, in 1935, to Mercer
the issue for him was no longer the consistency of quan- Street. Except for one br ief trip to Bermuda, E inste in
tum mechanics but rather its completeness. never again left the United States.
During the last twenty-five years of his life Einstein
maintained that quantum mechanics was incomplete. B. Einstein on objective reality
He no longer believed that quantum mechanics was
wrong but that the common view of the physics communi- In his Corno address Bohr (1928) had remarked that
ty was wrong in ascribing to the postulates of quantum quantum mechanics, like relativity theory, demands
mechanics a degree of finality which he held to be naive refinements of our everyday perceptions of inanimate
and unjustified. The content and shape of his dissent natural phenomena: "We find ourselves here on the very
will gradually unfold in what is to follow. path taken by Einstein of adapting our modes of percep-
In November 1931 Einstein (1932) gave a. colloquium tion borrowed from the sensations to the gradually
ln Berlin On the uncel tainty relation. The report of deepening knowledge of the laws of Nature. " Already
this talk does not state that Einstein objected to Heisen- then, in 1927, he emphasized that we have to treat with
berg's relations. Rather it conveys a sense of his dis- extreme care our use of language in recording the re-
comfort about the freedom of choice to measure pre- sults of observations which involve quantum effects.
cisely either the color of a light ray or its time of ar- "The hindrances met with on this path originate above
rival. Casimir (1977) has writ'ten to me about a col- all in the fact that, so to say, every word in the lang-
loquium which Einstein gave in Leiden with Ehrenfest uage refers to our ordinary perception. " Bohr's deep
in the chair. To the best of Casimir's recollection this concern with the role of language in the appropriate
took place in the winter of 1931-32. In his talk Einstein interpretation of quantum mechanics never ceased.
discussed several aspects of the clock-in-the-box ex- In 1948 he put it as follows (Bohr, 1948): "Phrases of-
periments. In the subsequent discussion it was men- ten found in the physical literature, as 'disturbance of
tioned that no conflict with quantum mechanics exists. phenomena by observation' or 'creation of physical
Einstein reacted to this statement as follows: "Ich attributes of objects by measurements' represent a use
weiss es, widerspruchsfrei ist die Sache schon, aber of words like 'phenomena' and 'observation' as well
sie enthalt meines Erachtens doch cine gewisse Harte" as 'attribute' and 'measurement' which is hardly com-
("I know, this business is free of contradictions, yet in patible with common usage and practical definition and,
my view it contains a certain unreasonableness" ). therefore, is apt to cause confusion. As a more ap-
propriate way of expression, one may strongly advocate
limitation of the use of the word phenomenon to refer
The Gedankenexperiment in this paper involves a time exclusively to observations obtained under specified
measurement. The authors take care to arrange things so that circumstances, including an account of the whole ex-
"the rate of the clock. . . is not disturbed by the gravitational periment. " This usage of "phenomenon, " if not general-
effects involved in weighing the box. "
~The authors are "forced to conclude that there can be no ly accepted, is the one to which nearly all physicists
method for measuring the momentum of a particle without now adhere.
changing its value. " This statement is of course unacceptable. In contrast to the view that the concept of phenomenon
"According to my conviction this tenet contains without irrevocably includes the specifics of the experimental
doubt a part of the ultimate truth. " conditions of observation, Einstein held that one should

Rev. Mod. Phys. , Vol. 5't, No. 4, October f979

 904 A. Pais: Einstein and the quantum theory

seek for a deeper lying theoretical framework which ther a paradox nor any flaw of logic. It simply concludes
permits the description of phenomena independently of that objective reality is incompatible with the assump-
these conditions. That is what he meant by the term tion that quantum mechanics is complete. This con-
"objective reality. " After 1933 it was his almost soli- clusion has not affected subsequent developments in
tary position that quantum mechanics is logically con- physics and it is doubtful that it ever will.
sistent but that it is an incomplete manifestation of an "It is only the mutual exclusion of any two experi-
underlying theory in which an objectively real descrip- mental procedures, permitting the unambiguous defini-
tion is possible. tion of complementary physical quantities which pro-
In an article written in 1935 together with Boris vides room for new physical laws", Bohr (1935) wrote
Podolsky and Nathan Rosen (Einstein, Podolsky, and in his rebuttal. He did not believe that the Einstein-
Rosen, 1935), Einstein gave reasons for his position Podolsky-Rosen paper called for any change in the in-
by discussing an example, simple as always. This te rpretation of quantum mechanics. Most physic ists
paper "created a stir among physicists and has played (including the writer) agree with this opinion. I shall
a large role in philosophical discussions (Bohr, 1949, reserve for the next section a further comment on the
p. 232). "~ It contains the following definition. "If completeness of quantum theory.
without in any way disturbing a system, we can predict This concludes an account of Einstein's position. He
with certainty (i.e. , with a probability equal to unity) returned to his criterion for objective reality in a num-
the value of a physical quantity, then there exists an ber of later papers (1948a, 1949a, 1951b, 1953) in which
element of physical reality corresponding to this physi- he repeated the EPR argument on several occasions.
cal quantity. " The authors then consider the following These papers add nothing substantially new. In one of
problem. Two particles with respective momentum and them (1951b) he discussed a further example (I omit
position variables (P„q, ) and (P„q,) are in a state with the details) stimulated by the question whether the quan-
definite total momentum P =P ~+@, and definite relative dis- tum-mechanical notion of phenomenon should also apply
tance q = q, —q, . This of course is possible since I' and q to bodies of everyday size. The answer is of course in
commute. The particles are allowed to interact. Ob- the affirmative.
servations are made on particle 1 long after the inter- Bohr was, of course, not the only one to express op-
action has taken place. Measure P, and one knows P, position to objective reality; nor was Einstein the only
without having disturbed particle 2. Therefore (in their one critical of the complementarity interpretation. ' I
language) P, is an element of reality. Next measure q, have chosen to confine myself to the exchanges between
-
and one knows q, again without having disturbed particle E instein and Bohr because I believe that E instein's
2. Therefore q, is also an element of reality. There- views come out most clearly in juxtaposing them with
fore both P, and q, are elements of reality. But quantum Bohr "s. Moreover I am well acquainted with their
mechanics tells us that P, and q, cannotbe simultaneously thoughts on these issues because of discussions with
elements of reality because of the noncommutativity of each of them. Bohr was in Princeton when he put the
the momentum and position operators of a given parti- finishing touches to his 1949 article (1949) and we did
cle. Therefore quantum mechanics is incomplete. discuss these matters often at that time. Iit was during
The authors stress that they "would not arrive at our one of these discuss ions that E inste in sneaked in to
conclusion if one insisted that two. . . physical quanti- steal some tobacco (Pa. is, 196'7).] However, it needs
ties can be regarded as simultaneous elements of reali- to be stressed that other theoretical physicists and
ty only uhen they can be simultaneously measured o~ mathematicians have made important contributions to
predicted" (their italics). Then follows a remark which this area of problems. Experimentalists have actively
is the key to Einstein's philosophy and which I have participated as well. A number of experimental tests
italicized in part. of quantum mechanics in general and also of the pre-
"This [simultaneous predictability] makes the reality
of P, and q, depend upon the process of measurement made. "
dictions of specific alternative schemes have been
This has not led to surprises.
carried out on the first system which does not disturb The foregoing was a brief sketch of the substance of
the second system in any way. reasonable definition
¹

of reality could be expected to permit this. The only "

part of this article which will ultimately survive, I
believe, is this last phrase which so poignantly sum-
marizes Einstein's views on quantum mechanics in his In 1950 Einstein (1950) mentioned Schroedinger and von
Laue as the only ones who shared his views. There were of
la, ter yea. rs. The content of this paper has been referred
course many others who at that time (and later) had doubts
to on occasion as the Einstein-Podolsky-Rosen para- about the complementarity interpretation, but their views and
dox. It should be stressed that this paper contains nei- Einstein's did not necessarily coincide or overlap (see Ein-
stein, 1952). Note also that the term "hidden variable" does
not occur in any of Einstein's papers or letters, as far as I
know.
Typically, these tests deal with variants of the EPR arrange-
This stir reached the press. On May 4, 1935 the Net Fowk ment, such as long-range correlations between spins or polar-
Tinzes carried an article under the heading "Einstein attacks izations. I must admit to being insufficiently familiar with the
quantum theory, "
which also includes an interview with a extensive theoretical and experimental literature on these
physicist. Its May 7 issue contains a statement by Einstein topics. My main guides have been a book by Jammer {1974)
in which he deprecated this release, which did not have his and a review article by Pipkin (1978). Both contain extensive
autho rization. references to other literature.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 905

E instein's arguments. There is another question which IX: A T IME CAPFUL E

at least to this writer is far more fascinating, the one
of motivation. What drove Einstein to use methods Einstein's life ended. . . with a
which he himself (1954b) called "quite bizarre as seen demand on us f or synthe s is.
from the outside F' Why would he continue "to sing my
W. Pauli (1958)
solitary old little song" (Einstein, 1949b) for the rest
of his life? As I shall discuss in Sec. X the answer has When Einstein and others embarked on programs of
to do with a grand design for a synthetical physical unification, three particles (in the modern sense) were
theory which Einstein conceived early (before the dis- known to exist, the electron, the proton, and the photon,
covery of quantum mechanics). He failed to reach this and there were two fundamental interactions„electro-
synthe s is. magnetism and gravitation. At present the number of
So to date have we all. particles runs into the hundreds. A further reduction to
The phenomena to be explained by a synthetical theory more fundamental units appears inevitable. The number
of particles and fields have become enormously richer of fundamental interactions is now believed to be at
since the days when Einstein embarked on his unified least four. The grand unification of all four types of
theory. Theoretical progress has been very impressive forces, gravitational, electromagnetic, weak and strong
but an all embracing theory does not exist. The need interactions, is an active topic of current exploration.
for a new synthesis is felt ever more keenly as the It has not been achieved as yet.
phenomena grow in complexity. Relativistic quantum field theories (in the sense of
Therefore, in the year of E instein's centennial, any special relativity) are the principal tools for these ex-
assessment of Einstein's visions can only be made from plorations. Our confidence in the general field theoretic
a vantage point which is necessarily tentative. It may be approach rests first and foremost on the tremendous
useful to record ever so briefly what this vantage point success of quantum electrodynamics (QED). One num-
appears to be to at least one physicist. This will be ber, the g factor of the electron, may illustrate the lev-
done next, with good wishes for the second centennial. el of predictability which this theory has reached:

1159652375(261U&&10 '2, predicted by pure QED, "

1 159652 410(200) &&10 ", observed.
It is nevertheless indicated that this branch of field be said to represent a program of geometrization which
theory has to merge with the theory of other fields. bears resemblance to Einstein's attempts, although the
"If we could have presented Einstein with a synthesis manifold subject to geometrization is larger than he
of his general relativity and the quantum theory, then antic ipated.
the discussion with him would have been considerably In the search for the correct field theory, model theo-
easier" (Pauli, 1958). To date this synthesis is beset ries have been examined which reveal quite novel pos-
with conceptual and technical difficulties. The existence sibilities for the existence of extended structures (soli-
of singularities associated with gravitational collapse tons, instantons, monopoles). In the course of these
is considered by some as an indication for the incom- investigations, topological methods have entered this
.
pleteness of the general relativistic equations. It is not area of physics. More generally, it has become clear
known whether or not these singularities are smoothed in the past decade that quantum field theory is much
out by quantum effects. richer in structure than was appreciated earlier. The
The ultimate unification of weak and electromagnetic renormalizability of non-Abelian gauge fields with
interactions has not been achieved but a solid beachhead spontaneous symmetry breakdown, asymptotic freedom,
appears to have been established in terms of local non- and supersymmetry are eases in point.
Abelian gauge theories with spontaneous symmetry The proliferation of new particles has led to attempts
breakdown. As a result it is now widely believed that at a somewhat simplified underlying description. Ac-
weak interactions are mediated by massive vector mes- cording to the current picture the basic constituents of
ons. Current expectations are that such mesons will ' particles, the lep-
matter are: two families of spin- —,
be observed within another decade from now. tons and the quarks; a variety of spin-1 gauge bosons,
It is widely believed that strong interactions are also some massless, some massive; and (more tentatively)
mediated by local non-Abelian gauge fields. Their sym- some fundamental spin-zero particles. The only gauge
metry is supposed to be unbroken so that the corres-
ponding vector mesons are massless. The dynamics of
these "non-Abelian photons" is supposed to prohibit This section is meant to provide a brief record without any
their creation as single free particles. The technical attempt at further explanation or reference to literature. It
exploration of this theory is only in its very early can be skipped without loss of continuity.
~3In this prediction {which does not include small contributions
stages. from muons and hadrons) the best value of the fine structure
Since electromagnetism and gravitation are also as- constant n has been used as an input: ~ ~=137.035987{29).
sociated with local gauge fields it is commonly held that The principal source of uncertainty {Kinoshita, 1978) in the
the grand unification will eventually be achieved in predicted value of g —2 stems from the experimental uncer-
terms of a multicomponent field of this kind. This may tainties of n.

Rev. Mod. Phys. , Vol. 51, No. 4, October |979

906 A. Pais: Einstein and the quantum theory

boson observed so far is the photon. To date three kinds I would visit with him in his office or accompany him
of charged leptons have been detected. The quarks are (often together with Kurt Godel) at lunchtime on his walk
hypothetical constituents of the observed hadrons. To home to 112 Mercer Street. Less often I would visit
date at least f ive species of quarks are needed. The dy- him there. In all I saw him about once every few weeks.
namics of the strong interactions is supposed to pro- We always talked German, the language best suited to
hibit the creation of quarks as single free particles. grasp the nuances of what he had in mind as well as the
This prohibition, confinement, has not as yet been im- flavor of his personality. Only once did he visit my
plemented theoretically in a convincing way. No cri- apartment. The occasion was a meeting of the Institute
terion is known which would enable one to state how many for Advanced Study faculty for the purpose of drafting a
species of leptons and of quarks should exist. statement of our position in the 1954 Oppenheimer af-
Weak, electromagnetic, and strong interactions have fair. I shall not go into E instein's outspoken opinions
distinct intrinsic symmetry properties, but this hier- on world affairs and public policy.
archy of symmetries is not well understood theoretical- E instein's company was comfortable and comforting
ly. Perhaps the most puzzling are the noninvariance to those who knew him. Of course he knew well that he
under space reflexion at the weak level and the non- was a legendary figure in the eyes of the world. He ac-
invariance under time reversal at an even weaker level. cepted this as a fact of his life. There was nothing in
It adds to the puzzlement that the latter phenomenon his personality to promote such attitudes. Nor did he
has been observed so far in only a single instance, relish them. Privately he would express annoyance if
namely in the Ko —K' system. [These phenomena, were he felt that his position was misused. I recall the case
first observed after Einstein's death. I have often of Professor X who had been quoted by the newspapers
wondered what might have been his reactions to these as having found solutions to Einstein's generalized
discoveries, given his "conviction that pure mathemati- equations of gravitation. Einstein said to me "Der
cal construction enables us to discover the concepts and Mann ist ein Narr" and added that in his opinion'X could
the laws connecting them" (Einstein, 1933).J calculate but could not think. X had visited Einstein to
It is not known why electric charge is quantized but it discuss this work and Einstein, always courteous, said
is plausible that this will be explicable in the framework to him that if his results were true they would be im-
of a future gauge theory. portant. Einstein was chagrined that he had been quoted
In summary, at the time of the centenary of the death in the papers without this proviso. He said that he would
of James Clerk Maxwell (June 13, 1831-November 5, keep silent on the matter but would not receive X again.
1879) and the birth of Albert Einstein (March 14, 1879- According to Einstein the whole thing started because,
April 18, 1955) the evidence is overwhelming that the in his enthusiasm, X had told some colleagues who saw
theory of particles and fields is incomplete. Einstein's the value of it as publicity for their university.
earlier complaint (1956a) remains valid to this day: To those physicists who could follow his scientific
"The theories which have gradually been connected with thought and who knew him personally, the legendary
what has been observed have led to an unbearable ac- —
aspec t was neve r in the f oregr ound yet it was never
cumulation of independent assumptions. "
At the same wholly absent. I recall an occasion, in 1947, when I
time no experimental evidence or internal contradiction gave a talk at the Institute about the newly discovered
exists to indicate that the postulates of general relativity, & and p. mesons. Einstein walked in just after I had be-
of special relativity, or of quantum mechanics are in gun. I remember that I was speechless for the brief
mutual conflict or in need of revision or refinement. moment necessary to overcome a sense of the unreal.
We are therefore in no position to affirm or deny that I recall a similar moment during a symposium" held in
these postulates will forever remain unmodified. Frick Chemical laboratory in Princeton on March 19,
1949. The occasion was Einstein s seventieth birthday.
X. PARTICLES, FIELDS AND THE QUANTUM THEORY: Most of us were in our seats when Einstein entered the
E INST E IN'S V IS ION
Apart. . .. 6. Away from common use for Advanced Study in Princeton where I hoped to work with Pauli.
a special purpose. I received a letter fram Pauli saying he would support my ap-
plication. I was accepted at both places and went first for one
Oxford English Dictionary year to Copenhagen. When I finally arrived at Princeton in
September 1946 I found that Pauli had in the meantime gone to
A. Some reminiscences Zurich. Bohr also came to Princeton that same month. Both
The rest of this paper is based in part on what I of us attended the Princeton Bicentennial Conference (where
learned from discussions with Einstein. I should like
P. A. Schilpp approached Bohr for a contribution to the Ein-
stein biography). Shortly thereafter Bohr introduced me to
to mention first some reminiscences of my encounters Einstein.
With him. My stay at the Institute had lost much of its attraction be-
I knew Einstein from 1946 until the time of his death~. cause Pauli was no longer there. As I was contemplating re-
turning to Copenhagen the next year, Oppenheimer contacted
me to inform me that he had been approached for the director-
4In 1941, I received my Ph. D. in Utrecht with L. Rosenfeld. ship of the Institute and to ask me to join him in building up
Some time later I went into hiding in Amsterdam. Eventually physics there. I accepted. A year later I was appointed to a
I was caught and sent to the Gestapo prison there. Those who five-year membership at the Institute and in 1950 to Professor.
were not executed were released a few days before V. E. Day. I remained at the Institute until 1963.
Immediately after the war, I applied for a postdoctoral fellow- The speakers were J. B. Oppenheimer, I. I. Rabi, E. P.
ship at Niels Bohr's Institute as well as at the Institute for Wigner, H. P. Robertson, S. M. Clemence, and H. . Weyl.

Rev. Mod. Phys. , Vol. 5't, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 907

hall. Again there was this brief hush before we stood secure enough basis for relativistic generalizations
to greet him. (Einstein, 1934, 1946}. Relativistic quantum field theory
Nor do I believe that such reactions were typical only was repugnant to him (Born, 1944).9' Valentine Barg-
of those who were much younger than he. There were mann has told me that Einstein asked him to give a
a few occasions when Pauli and I were both together with private survey of quantum field theory, beginning with
him. Pauli, not known for an excess of awe, was just second quantization. Bargmann did so for about a
slightly different in Einstein's company. One could per- month. Thereafter Einstein's interest waned.
ceive his sense of reverence. I have also seen Bohr
An unconcern with the past is a privilege of youth. In
and Einstein together. Bohr too was affected in a some-
all the years I knew Einstein I never read a single one
what similar way, differences in scientific outlook not-
of his papers from the 1905-1925 period on the quantum
w ith s tanding.
theory. It is now clear to me that I might have asked
Whenever I met Einstein our conversations might
him some interesting questions, had I been less blessed
range far and wide but invariably the discussion would
with ignorance. I might then have known some interest-
turn to physics. Such discussions would touch only oc-
ing facts by now, but at a price. My discussions with
casionally on matters having to do with the period be-
Einstein were not historical interviews, they concerned
fore 1925 and then they would mainly concern relativity.
live physics. I am glad it never was otherwise.
I recall asking E instein once what influence Poincare's
workhadhadonhim. Einstein replied that he never had B. Einstein, Newton, and success
read Poincare. On the other hand Einstein held no one
in higher esteem than Lorentz. He once told me that "It seems to be clear. . . that the Born statistical in-
without Lorentz he would never have beeri able to take terpretation of the quantum theory is the only possible
"den Schritt" (" the step"), (Einstein always talked "
one, Einstein (1936) wrote in 1936. He has also called
about relativity in an impersonal way. ) I.orentz was to (1949a} "the statistical quantum theory . . . the most
Einstein the most well-rounded and harmonious per- successful physical theory of our period. " Then why
sonality he had met in his entire life. He had also a was he never convinced by it 7 I believe Einstein (1933)
great veneration for Planck. himself answered this indirectly in his 1933 Spencer
Our discussions, however, centered first and fore- —
lecture perhaps the clearest and most revealing ex-
most on the present and the future. When relativity was pression. of his mode of thinking. The key is to be found
the issue he would often talk af his efforts to unify gravi- in his remarks on Newton and classical mechanics.
tation and electromagnetism and of his hopes for the In this lecture Einstein notes that "Newton felt by no
next steps. His faith rarely wavered in the path he had means comfortable about the concept of absolute space,
chosen. Only once did he express a reservation to me . . . . of absolute rest . . . [andJ about the introduction of
when he said, in essence: I am not sure that differen- action at a distance. " Einstein then goes on to refer to
tial geometry is the framework for further progress, the success of. Newton's theory in these words: "The
but if it is, then I believe I am on the right track. (This enormous practical success of his theory may well have
remark must have been made sometime during his last prevented him and the physicists of the eighteenth and
few years. ) nineteenth centuries from recognizing the fictitious
The main topic of discussion was quantum physics, character of the principles of his system. " It is im-
however. Einstein never ceased to ponder the meaning portant to note that by "fictitious, " Einstein means free
of the quantum theory. Time and time again the argu- inventions of the human mind. Whereupon he compares
ment would turn to quantum mechanics and its inter- Newton'8 mechanics with his own work on general rela-
pretation. He was explicit in his opinion that the most tivity: "The fictitious character of the principles is
commonly held views on this subject could not be the made quite obvious by the fact that it is possible to ex-
last word, but he had also more subtle ways of expres- hibit two essentially different bases [Newtonian mechan-
sing his dissent. For example he would never refer to ics and general relativistic mechanics] each of which in
a wave function as "die Wellenfunktion" but would al- its consequences leads to a large measure of agreement
ways use a mathematical terminology: "die Psifunk- with experience. "
tion. " We often discussed his notions on objective Elsewhere Einstein (1949a) addressed Newton as fol-
reality. I recall that during one walk Einstein suddenly lows. "Newton forgive me: you found the only way
stopped, turned to me and asked whether I really be- which, in your age, was just about possible for a man
lieved that the moon exists only when I look at it. The with the highest power of thought and creativity. " Only
rest of this walk was devoted to a discussion of what a one man in history could have possibly written that line.
physicist should mean by the term "to exist. " In the Spencer lecture Einstein mentioned the success
I was never able to arouse much interest in Einstein not only of classical mechanics but also of the statistical
about the new particles. It was apparent that he felt interpretation of quantum theory. "This conception is
that the time was not ripe to worry about such things logically unexceptionable and has led to important suc-
and that these particles would eventually appear as so- cesses. " But, he added, "I still believe in the possibili-
lutions of the equations of a unified theory. In some
sense he may well prove to be right.
It was even more difficult to discuss quantum field W. 'Thirring (1977) has written to me of conversations with
theory with him. He was willing to admit that quantum Einstein in which "his objections became even stronger when
mechanics was successful on the nonrelativistic level. it concerned quantum field theory and he did not believe in any
However, he did not believe that this theory provided a of its consequences. "
Rev. IVlod. Phys. , Vol. 61, No. 4, October 'l979
908 A. Pais: Einstein and the quantum theory

ty of giving a model of reality which shall represent basis, in that it must be deducible as a limiting case
events themselves and not merely the probability of their from that basis, just as electrostatics is deducible
occurrence. " from the Maxwell equations of the electromagnetic field
From this lecture as well as from discussions with or as thermodynamics is deducible from statistical me-
him on the foundations of quantum physics I have gained chanics" (Einstein, 1936).
the following impression. Einstein tended to compare (2) One should not try to find the new theory by be-
the successes of classical mechanics with those of quan- ginning with quantum mechanics and trying to refine or
tum mechanics. In his view both were on a par, being reinterpret it. "I do not believe that quantum mechanics
successful but incomplete. For more than a decade will be the sta. rting point in the search for this ba. sis
Einstein had pondered the single question of hom to ex- just as one cannot arrive at the foundations of mechanics
tend to general motions the invariance under uniform from thermodynamics or statistical mechanics" (Ein-
translations. His resulting theory, general relativity, stein, 1936).
had led to only small deviations from Newton's theory. —
(3) Instead and this was Einstein's main point — one
(Instances where these deviations are large were dis- should start all aver again, as it were, and endeavor to
cussed only much later. ) He was likewise prepared to obtain the quantum theory as a byproduct of a general
undertake his own search for objective reality, fearless relativistic field theory. As an introduction to a further
of how long it would take. He was also prepared for the discussion of this last issue it is useful to comment first
survival of the practical successes of quantum mechan- on the profound differences between Einstein's attitude
ics, with perhaps only small modifications. It is quite to relativity and to the quantum theory.
plausible that the very success of his highest achieve-
ment, general relativity, was an added spur to Ein- C. Relativity theory and quantum theory
stein's apartness. Yet it should not be forgotten that
this trait characterized his entire oeuvre and style. Einstein's paper on light-quanta was submitted in
Einstein was not oblivious to others' reactions to his March 1905, his first two papers on relativity in June
own position. "I have become an obstinate heretic in (1905c) and September (1905d) of that year, respectively.
the eyes of my colleagues" he wrote (1949c) to one In a, letter (1905e) to a friend written early in 1905 he
friend and "in Princeton they consider me an old fool" promised him a copy of his March paper "aboutradiation
he sa, id to another (Born, 1971, p. 131). He knew, and and the energy properties of light [which] is very revo-
on occasion mould even say, that his road was a lonely lutionary. " In the same letter he also mentioned that a
one (Einstein, 1948b), yet he held fast. "Momentary draft of the June paper was ready and added that "the
success carries more power of conviction for most purely kinematic part of this work will surely interest
people than reflections on principle" (Einstein, 1949d). you. " It is significant that Einstein mould refer to his
Einstein was neither saintly nor humorless in defend- light-quantum paper but not to his relativity paper as a
ing his solitary position on the quantum theory. Qn revolutionary step.
occasion he could be acerbic. At one time he said that If a revolutionary act consists in overthrowing an ex-
Bohr thought very clearly, wrote obscurely and thought isting order, then to describe the light-quantum hypo-
of himself as a prophet (Shankland, 1963). Another thesis in those terms is altogether accurate. It is like-
time he referred to Bohr as a mystic (Einstein, 1939). wise fitting not to apply these terms to relativity theory
On the other hand, in a, letter to Bohr, Einstein (1949b) since it did not overthrow an existing order but rather
referred to his own position by quoting an old rhyme: brought immediate order to new domains.
"Uber die Reden des Kandidaten Jobses/Allgemeines Einstein was one of the freest spirits there ever was.
Schutteln des Kopses. '"' There were moments of lone- But he was not a revolutionary, as the overthrow of
liness. "I feel sure that you do not understand hom I existing order was never his prime motivation. It was
came by my lonely ways . . . " (Einstein, 1948c). Ein- his genius that made him ask the right question. It
stein may not have expressed all his feelings on these was his faith in himself that made him persevere until
matters. But that was his way. "The essential of the he had the answer. If he had a God it was the God of
being of a man of my type lies precisely in zvhat he Spinoza. He had to follow his own reasoning regardless
thinks and hose he thinks, not in what he does or suffers" of where it would lead him. He had a deep respect for
(E ins tein, 1949a) . . the traditions of physics. But if his own reasoning in-
The crux of Einstein's thinking on the quantum theory dicated answers which lay outside the conventional pat-
was not his negative position in regard to what others terns, he accepted these answers, not for the sake of
had done, but rather his deep faith in his own distinct contradiction but because it had to be. He had been
approach to the quantum problems. His beliefs may be free to ask the question. He had no choice but to accept
summarized as follows. the answer. This deep sense of destiny led him farther
(1) Quantum mechanics represents a major advance,
—
than anyone before him but not as far as finding his own
answer to the quantum theory.
yet it is only a limiting case of a theory which remains
to be discovered. "There is no doubt that quantum me- It is striking how, from the very beginning, Einstein
chanics has seized hold of a beautiful element of truth, kept his scientific writing on relativity theory separate
and that it will be a test stone for a future theoretical from that on quantum theory. This was evident already
in 1905. In his first relativity paper (1905c) Einstein
noted: "It is remarkable that the energy and frequency
73oughly: "there was a generaI shaking of heads concerning of a light complex vary with the state of motion of the
the words of candidate Jobs. " observer according to the same law'. ". Here was an

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory 909

obvious opportunity to refer to the relation E =Av of anyone else. He demanded that the theory shall be
his paper on light-quanta, finished a few months earlier. strictly causal, that it shall unify gravitation and elec-
But Einstein did not do that. Also in the September tromagnetism, that the particles of physics shall
paper (1905d) he referred to radiation but not to light- emerge as special solutions of the general field equa-
quanta. In his 1909 address at Salzburg (1909a) Einstein tions, and that the quantum postulates shall be a corise-
discussed his ideas both on relativity theory and on quence of the general field equations. Einstein had all
quantum theory but kept these two areas well separated. these criteria in mind when he wrote (1949a) in 1949:
As we have seen, in his 1917 paper (1917a) Einstein "Our problem is that of finding the field equations of
ascribed to light-quanta an energy E= hv and a moment- the total field. "
um P= hv/c. This paper concludes with the following
Already in 1923 he had been brooding on these ideas
remark. "Energy and momentum are most intimately
for a number of years. In 1920 he had written (1920c)
related; therefore a theory can only then be considered
to Born: "I do not seem able to give tangible form to
as justified if it has been shown that according to it the
my pet idea ["meine Lieblingsidee" J, which is to under-
momentum transferred by radiation to matter leads to
motions as required by thermodynamics. " Why is only
stand the structure of the quanta by redundancy in de-
thermodynamics mentioned; why not also relativity? termination, using differential equations. " It is the
I believe that the reason Einstein kept the quantum earliest reference to Einstein's strategy that I am aware
theory apart from relativity theory is that he considered
of. It would seem likely that ideas of this kind began
to stir in Einstein soon after 1917, when he had not only
the former to be provisional (as he said (1912d) already
completed the general theory of relativity but had also
in 1911) while, on the other hand, relativity to him was
confronted the lack of causality in spontaneous emis-
the revealed truth. Einstein's destiny reminds me in
sion (Einstein, 1917a). The early response of others to
more than one way of the destiny of Moses.
these attempts by Einstein has been recorded by Born
The road to general relativity theory had been local
classical field theory. The same road, he hoped, would (1971, p. 88}: "In those days [early in 1925] we all
thought that his objective . . . was attainable and also
very important. " Einstein himself (19241) felt that he
also lead to the implementation of objective reality.
had no choice: "The road may be quite wrong but it
D. F instein's vision must be tried. "
As I already intimated in the Introduction, it is es-
In 1923 E instein published an articl'e (1923b) entitled
"Does field theory offer possibilities for the solution of sential for the understanding of Einstein's thinking to
the quantum problem?" It begins with a reminder of the
realize that there were two sides to his attitude concern-
ing quantum physics. There was Einstein the critic,
successes achieved in electrodynamics and general rela-
never yielding in his dissent from complementarity,
tivity theory in regard to a causal description: events
and there was Einstein the visionary, forever trying to
are causally determined by differential equations com-
realize the program outlined above, which went well
bined with initial conditions on a spacelike surface.
beyond a mere reinterpretation of quantum mechanics
However, Einstein continued, this method cannot be
(Einstein, 1952). His vision predates quantum mechan-
applied to quantum problems without further ado. As
he put it, the discreteness of the Bohr orbits indicates
ics; it was certainly with him in 1920 and probably even
that initial conditions cannot be chosen freely. Then he
a few years earlier.
A detailed description of his efforts in this direction
asked: can one nevertheless implement these quantum
belongs to a history of unified field theory, a topic
constraints in a (causal) theory based on partial differ-
which cannot be dealt with here. In concluding this
ential equations '? His answer: "Quite certainly: we
must only 'overdetermine' the field variables by [ap- paper I shall confine myself to a few brief observations
propriate] equations. "
Next he states his program,
concerning Einstein's own attitude toward his program.
Einstein believed that the field equations would gen-
based on three requirements. (1) General covariance erate particles with nonzero spin as particle-like solu-
(2) The desired equations should at least be in accord-
ance with the gravitational and the Maxwell theory
The desired system of equations which overdetermines
(3). mann,
tions which are not spherically symmetrical (V. Barg-
private communication). Presumably he hoped
the fields should have static spherically symmetric so-
lutions which describe the electron and the proton. If
crete spin values. "
that his idea of overdetermination would lead to dis-
He also hoped that the future theory
would contain solutions which are not absolutely loca-
this overdetermination can be achieved then "we may
lized and which would carry quantized electric charge
hope that these equations co-determine the mechanical
behavior of the singular points (electrons) in such a way (Einstein, 1933}. [In 1925 Einstein noted (1925f) that if
the combined gravitational/electromagnetic field equa-
that also the initial conditions of the field and the singu-
lar points are subject to restrictive conditions. " He
tions have particle-like solutions with charge e and
mass m, then there should also be solutions" with
goes on to discuss a tentative example and concludes as
follows: "To me the main point of this communication
is the idea of overdetermination. "
This paper contains all the essential ingredients of I note in passing that in 1925 Einstein gave a helping hand
to Uhlenbeck and Goudsmit in the explanation of the origins of
the vision on particles, fields, and the quantum theory the spin-orbit coupling of electrons in atoms (Uhlenbeck, 1976).
which Einstein was to pursue for the rest of his life. Zhe proof involves the application of time reversal to the
To Einstein the concept of a unified field theory meant combined equations. In related context the existence of the
something different from what it meant and means to (+e, m) solutions was first noted by Pauli (1919) in 1919.

Rev. Mod. Phys. , Vol. 5't, No. 4, October 1979

910 A. Pais: Einstein and the quantum theory

(—e, m)! This led him to doubt for some time that the p. 180): "Our respective hobby horses have irretriev-
unification of gravitation and electromagnetism was ably run off in different directions . . . even I cannot ad-
consistently possible. ] here to [mine] with absolute confidence. " I have men-
E instein's correspondence shows that the unified field tioned before the reservations which Einstein expressed
theory and the quantum problems were very often simul- to me in the early fifties. ' To his dear friend Besso
taneously on his mind. Here are but a few examples. he wrote (1954c), in 1954: "I consider it quite possible
In 1925, while he was at work on a theory with a non- that physics cannot be based on the field concept, i.e. ,
symmetric metric he wrote (1925g) to a friend: "Now on continuous structures. In that case nothing remains
the question is whether this field theory is compatible of my entire castle in the air, gravitation theory in-
with the existence of atoms and quanta. "
He discussed cluded, [and of] the rest of modern physics. It is to be "
the same generalized theory in a letter (1942c) written doubted whether any physicist can be found who would
in 1942: "What I am doing now may seem a bit crazy not respectfully and gratefully submit that this judgment
to you. One must note, however, that the wave-particle is unreasonably harsh.
duality demands something unheard of. In 1949 (1949e): " Otto Stern has recalled a statement which Einstein
"I am convinced that the . . statistical [quantum]
. once made to him: "I have thought a hundred times as
theory . . . is superficial and that one must be backed much about the quantum problems as I have about gen-
by the principle of general relativity.
"
And in 1954 eral relativity theory" (Jost, 1977). He kept thinking
(1954b): "I must seem like an ostrich who forever about the quantum till the very end. Einstein (1956b)
buries his head in the relativistic sand in order not to wrote his last autobiographical sketch in Princeton, in
face the evil quanta. " March 1955, about a month before his death. Its final
Forever and in vain Einstein kept looking for hints sentences dea, l with the quantum theory. "It appears
which would help him to realize his vision of a quantum dubious whether a field theory can account for the ato-
theory derived from a unified field theory. This urge mistic structure of matter and radiation as well as of
explains his reference to the quantum theory at unex- quantum phenomena. Most physicists will reply with a
pected places. In 1930 he gave a lecture on unified fieM convinced 'No, ' since they believe that the quantum
theory, a report of which (cabled to the New York Times) problem has been solved in principle by other means.
contains the statement: "He emphasized that he is in However that may be, I.essing's comforting word stays
no way taking notice of the results of quantum calcula- with us: the aspiration to truth is more precious than
tions because he believes that by dealing with micro- its assured possession.
scopic phenomena these will come out by themselves" During the last years of his life Einstein was not well.
(Einstein, 1930b). A report in 1931 by Einstein (1931b) "We around him knew since about five years of the
on a five-dimensional theory which should unify gravi- sword of Damocles hanging over us. He knew it too and
tation and electromagnetism ends as follows. "This waited for it calmly and smilingly" (Dukas, 1955). His
theory does not yet contain the conclusions of the quan-
tum theory. " Two months after the E instein-Podolsky-
final illness was not long. On April 15 he entered
Princeton hospital. He refused to be operated on. "I
Rosen article, Einstein and Rosen (1935) completed want to go when I want —
I have done my, share~ lt is
another paper, this one dealing with singularity-free —
time to go I will do it elegantly" (Dukas, 1955). Ein-
solutions of the gravitational-electromagnetic field stein died in the early morning hours of April 18.
equations. One phrase in this paper, "One does not see
a priori whether the theory contains the quantum phen-
"
omena, illustrates once again the scope of the pro-
X I. EP I LOG
gram which was on Einstein's mind. I saw Einstein for the last time in December 1954.
Simplicity was the guide in Einstein's quest: "In my As he had not been well he had for some weeks been
opinion there is t&e correct path and. .. it is in our absent from the Institute where he normally spent a few
power to find it. Our experience up to date justifies hours each morning. Since I was about to take a term's
us in feeling sure that in Nature is actualized the ideal leave from Princeton, I called Helen Dukas and asked
of mathematical simplicity" (Einstein, 1933). Already her to be kind enough to give my best wishes to Pro-
in 1927 Heisenberg (1927a) stressed, in a letter to Ein- fessor Einstein. She suggested I might come to the
stein mentioned earlier, that Einstein's concept of sim- house for a brief visit and a cup of tea. I was of course
plicity and the simplicity inherent in quantum mechanics glad to accept. After I arrived, I went upstairs and
cannot both be upheld. "If I have understood correctly knocked at the door of Einstein's study. There was his
your point of view than you would gladly sacrifice the
"
gentle "come. As I entered he was seated in his arm
simplicity [of quantum mechanics] to the principle of chair, a blanket over his knees, a pad on the blanket.
[classical] causality. Perhaps we could comfort our- He was working. He put his pad aside at once and
selves [with the idea that] the dear Lord could go beyond greeted me. We spent a pleasant half hour or so; I do
[quantum mechanics] and maintain causality. I do not not recall what was discussed. Then I told him I should
really find it beautiful, however, to demand physically not stay any longer. We shook hands, and I said good-
more than a description of the connection between ex- bye. I walked to the door of the study, not more than
periments. " f our or f ive steps away. I turned around as I opened

As Einstein's life drew to a close, doubts about his

vision arose in his mind. V. Bargmann informs me that Einstein made similar re-
He wrote to Born, probably in 1949 (see Born, 1971, marks to him in the late thirties {see also Einstein, &94&).

Rev. Mod. Phys. , Vol. 51, No. 4, October 'I979

 A. Pais: Einstein and the quantum theory

the door. I saw him in his chair, his pad back on his Born, M. , and Th. von Karma, n, 1913, Phys. Z. 14, 15.
lap, a pencil in his hand, oblivious to his surroundings. Bose, S. N. , 1924a, letter to A. Einstein, June 4 (unpublished,
He was back at work. Einstein Archives).
Bose, S. N. , 1924b, Z. Phys. 26, 178.
Bothe, W. , 1923, Z. Phys. 20, 145.
AC K NONI L EDG IVl E NTS Bothe, W. , and H. Geiger, 1924, Z. Phys. 26, 44.
No one alive is more familiar with the circumstances Bothe, W. , and H. Geiger, 1925a, Naturwissenschaften 13, 440.
Bothe, W. , and H. Geiger, 1925b, Z. Phys. 32, 639.
of Einstein's life and with his collected correspondence Brillouin, L. , 1921, J. Phys. (Paris) 2, 142.
than Helen Dukas. I am deeply grateful for her friend- Casimir, H. B. G, , 1977, letter to A. Pals, December 31 ~

ship and generosity through the years which have helped Chadwick, J., 1962, Proceedings of the Tenth International
me most importantly in gathering key information con- Congress on the History of Science, Ithaca (Hermann, Paris),
tained in this article. I am also much indebted to Res Vol. 1, p. 159.
Jost, Martin Klein, and George Uhlenbeck for advice Compton, A. H. , 1923, Phys. Rev. 21, 483.
and guidance. Numerous discussions with Sam Trei- Compton, A. H. , and A. W. Simon, 1925, Phys. Rev. 26, 889.
man about the contents of this article have been invalu- Curie, P. , and G. Sagnac, 1900, C. R. Acad. Sci. (Paris) 130,
1013.
able. I also wish to thank the Princeton Physics Depart- Davis, L. , A. S. Goldhaber, and M. M. Nieto, 1975, Phys.
ment for its hospitality and the staff of Fine Hall I i- Rev. Lett. 35, 1402.
brary for continued help. Finally I want to thank the ex- de Broglie, L. , 1923a, C. R. Acad. Sci. (Paris) 177, 630.
ecutors of the Albert Einstein Estate and of the Pauli de Broglie, I.
, 1923b, C. R. Acad. Sci. (Paris) 1VV, 507.
Estate for permission to quote from unpublished docu- de Broglie, L. , 1923c, C. R. Acad. Sci. (Paris) 1VV, 548.
ments. de Broglie, I.
, 1962, Ke~ P'erspectives in Physics (Basic,
New York), p. 150.
de Broglie, L. , 1963, Preface to his reedited 1924 Ph. D.
Thesis, Recherches sur la Thdorie des Quanta (Masson et
Allison, S. K. , 1965, Biogr. Mem. Nat. Acad. Sci. 38, 81. Cie. , Paris), p. 4.
Arrhenius, S. , 1965 in ¹be/
Lectures in Physics (Elsevier, de Broglie, 1978a, letter to A. Pais, August 9.
New York), Vol. 1, pp. 478 — 481. de Broglie, 1978b, letter to A. Pais, September 26.
A vogadro, A. , 1833, A nn. Chim. Phys. 55, 80. Debye, P. , 1910, Ann. Phys. (Leipz. ) 33, 1427.
Balmer, J. J. , 1885, Ann. Phys. (Leipz. ) 25, 80. Debye, P. , 1912, Ann. Phys. (Leipz. ) 39, 789.
Bateman, H. , 1923, Philos, Mag. 46, 977. Debye, P. , 1923, Phys. Z. '24, 161.
Bay, Z. , V. P. Henri, and F. McLennon, 1955, Phys. Rev. 97, de Hevesy, 1913, letter to E. Rutherford, October 14, quoted
1710. in A. S. Eve, 1939, Rutherford (Cambridge University Press,
Behn, U. , 1893, Ann. Phys. (Leipz. ) 48, 708. Cambridge, England), p. 226.

Bohr, ¹,
Benz, U. , &975, Arnold Sommerfeld (%'issenschaftliche Ver-
lagsgesellschaft, Stuttgart), p. 74.
Blanpied, W. A. , 1972, Am. J. Phys. 40, 1212.
1913a, Philos. Mag. 26, 1.
Bohr, N. , 1913b, Nature (Lond. ) 92, 231.
Bohr, N. , 1920, letter to A. Einstein, June 24 (unpublished),
housed in the Albert Einstein Archives at the Institute for
de la Rive, A. , and F. Marcet, 1840, Ann. Chim. Phys. V5,
1 13o
Dewar,
Dewar,
J.,
1872, Philos. Mag. 44, 461.
J.,
1905, Proc. R. Soc. Lond. 76, 325;
Dirac, P. A. M. , 1926, Proc. R. Soc. A 112, 661.
Dirac, P. A. M. , 1927, Proc. R. Soc. A 114, 243.
Dirac, P. A. M. , 1931, Proc. R. Soc. A 133, 60.
Advanced Study, Princeton, N. J.). Dirac, P. A. M. , 1977, in History of Tseentieth Century Phys-
Bohr, N. , 1923, Z. Phys. 13, 117, especially Section 4. ics, Varenna Summer School (Academic, New York), pp.
Bohr, N. , 1925a, Z. Phys. 34, 142. 133—134.
Bohr, N. , 1925b, letter to R. H. Fowler, April 21, quoted in Duane, W. , and F. L. Hunt, 1915, Phys. Rev. 6, 166.
Stuewer (1975}, p. 301. Dukas, H. , 1955, letter to A. Pais, April 30.
Bohr, N. , 1925c, Z. Phys. 34, 142, Appendix. Ehrenfest, P. , 1922, letter to A. Einstein, January 17 (unpub-
Bohr, N. , 1928, Nature (Lond. ) 121, 580. lished, Einstein A rchives).
Bohr, N. , 1935, Phys. Rev. 48, 696. Ehrenfest, P. , 1925, Z. Phys. 34, 362.
Bohr, N. , 1948, Dialectica 2, 312. Ehrenfest, P, , 1926, letter to A. Einstein, April 7 (unpub-
Bohr, N. , 1949, in P. A. Schilpp, editor, Albert Einstein, lished, Einstein Archives).
Philosopher-Scientist (Library of Living Philosophers, Evan- Einstein, A. , 1902, Ann. Phys. (Leipz. ) 9, 417.
ston, Hl. ), p. 199. Einstein, A. , 1903, Ann. Phys. (Ieipz. ) 11, 170.
Bohr, N. , H. A. Kramers, and J. C. Slater, 1924, Philos, Einstein, A. , 1904, Ann. Phys. (Leipz. ) 14, 354.
Mag. 47, 785. Einstein, A. , 1905a, Ann. Phys. (Leipz. ) 1V, 132.
Boltzmann, L. , 1871, Wien. Ber. 63, 679. Einstein, A. , 1905b, Ann. Phys. (Leipz. ) 1V, 549.
Boltzmann, L. , 1876, Wien. Ber. 74, 553. Einstein, A. , 1905c, Ann. Phys. (Leipz. ) 1V, 891.
Boltzmann, L. , 1884, Ann. Phys. (I.eipz. ) 22, 31, 291. Einstein, A. , 1905d, Ann. Phys. (Leipz. ) 18, 639.
Boltzmann, L. , 1897, Vorlesungen uber die PrinciPe der Me- Einstein, A. , 1905e, letter to Conrad Habicht, undated but
chanic (Barth, Leipzig), Vol. 1, p. 9; reprinted, 1974, by probably written in March; reproduced in Seelig (1954), pp.
Wiss. Buchges. DRrmstadt. 88, 89.
Boltzmann, L. , 1912, Vorlesungen uber Gastheorie (Barth, Einstein, A. , 1906a, Ann. Phys. (Leipz. ) 19, 289.
Leipzig), Vol. 2, p. 131. Einstein, A. , 1906b, Ann. Phys. (Leipz. ) 20, 199.
Born, M. , 1926, Z. Phys. 3V, 863. Einstein, A. , 1907a, Ann. Phys. (Leipz. ) 22, 180.
Born, M. , 1944, letter to A. Einstein, October 10, in Born Einstein, A. , 1907b, Ann. Phys. (Leipz. ) 22, 800.
(1971), p. 155. Einstein, A. , 1908, letter to J. J. Laub, quoted in Seelig (1954),
Born, M. , 1971, editor, The Born-Einstein Letters (Walker, p. 103.
New York), p. 91. Einstein, A. , 1909a, Phys. Z. 10, 817.
Born, M. , W. Heisenberg, and P. Jordan, 1925, Z. Phys. 35, Einstein, A. , 1909b, Phys. Z. 10, 185.
557. Einstein, A. , 1911a, Ann. Phys. (Leipz. ) 34, 175.
Born, M. , @nd Th. von KCrmfn, 1912, Phys. Z. 13, 297. Einstein, A. , 1911b, Ann. Phys. (Leipz. ) 34, 170.

Rev. Mod. Phys. ', Vof. 53, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

Einstein, A. , 1911c, Ann. Phys. (Leipz. ) 35, 679. Einstein, A. , 1926c, letter to E. Schroedinger, April 26, re-
Einstein, A. , 1912a, Ann. Phys. (Leipz. ) 37, 832. produced in M. Klein, 1967, translator, Letters on Wave
Einstein, A. , 1912b, Ann. Phys. (Leipz. ) 38, 881. Mechanics (Philosophical Library, New York).
Einstein, A. , 1912c, Ann. Phys. (Leipz. ) 38, 888. Einstein, A. , 1926d, letter to M. Besso, May 1, in Speziali
Einstein, A. , 1912d, in La theoric du rayonnement et les (1972), p. 224.
quanta, Proceedings of the First Solvay Conference, edited Einstein, A. , 1926e, letter to M. Born, Dec. 4, in Born (1971),
by P, Langevin and M. de Broglie {Gauthier-Villars, Paris). p. 90.
Einstein, A. , 1913, Naturwissenschaften 1, 1077. Einstein, A. , 1926f, in Born (1971), p. 91.
Einstein, A. , 1914, Verh. Dtsch. Phys. Ges. 16, 820. Einstein, A. , 1928, in Elect~ons et Photons, the Proceedings
Einstein, A. , 1916a, letter to M. Besso, November 18, in of the Fifth Solvay Conference (Gauthier-Villars, Paris), p.
Speziali (1972), p. 78. 253.
Einstein, A. , 1916b, Verh. Dtsch. Phys. Ges. 18, 318. Einstein, A. , 1930a, Sitzungsber. Preuss. Akad. Wiss. Phys. —
Einstein, A. , 1916c, Mitt. Ph. Ges. Zurich 16, 47. Math. KI. , p. 543.
Einstein, A. , 1916d, letter to M. Besso, Sept. 6, in Speziali Einstein, A. , 1930b, Science 71, 608.
(1972), p. 82. Einstein, A. , 1931a, letter to the Nobel Committee in Stock-
Einstein, A. , 1917a, Phys. Z. 18, 121. holm, Sept. 30 (unpublished, Einstein Archives).
Einstein, A. , 1917b, Verh. Dtsch. Phys. Ges. 19, 82. Einstein, A. , 1931b, Science 74, 438.
Einstein, A. , 1917c, letter to M. Besso, March 9, in Speziali Einstein, A. , 1932, Z. Angew. Chem. 45, 23.
-

(1972), p. 103. Einstein, 1933, On the method of theoretical physics (O«ord

Einstein, A. . 1918. letter to M. Besso, July 29, in Speziali U. P. , New York), reprinted in Philos. Sci. 1, 162 (1934) ~

(1972), p. 130. Einstein, A. , 1934, letter to M. Born, March 22, in Born

Einstein, A. , 1919, postcard to M. Planck, October 23, in (1971), p. 121.
Seelig (1954), p. 193. Einstein, A. , 1936, J. Franklin Inst. 221, 313.
Einstein, A. , 1920a, letter to M. Born, Jan. 27, in Born (1971), Einstein, A. , 1939, letter to E. Schroediriger, August 9, in
p. 23o M. Klein, translator, 1967, Letters on Wave Mechanics
Einstein, A. , 1920b, letter to N. Bohr, May 2 (unpublished, (Philosophical Library, New York), p. 32.
Einstein Archives). Einstein, A. , 1941, letter to I.Infeld, March 6 (unpublished,
Einstein-, A. , 1920c, letter to M. Born, March 3, in Born Einstein Archives) .
(1971), p. 26. Einstein, A. , 1942a, Sci. Monthly 54, 195.
Einstein, A. , 1921, Sitzungsber. Preuss. Akad. Wiss. Phys. — Einstein, A. , 1942b, letter to C. Lanczos, March 21 (unpub-
Math. KI. , p. 882. lished, Einstein Archives).
Einstein, A. , 1922, Sitzungsber. Preuss. Akad. Wiss. Phys. — Einstein, A. , 1942c, letter to M. Besso, August, in Speziali
Math. KI. , p. 18. (19V2), p. 366.
Einstein, A. , 1923a, letter to N. Bohr, January 11. Einstein, A. , 1946, letter to A. Sommerfeld, December 14, in
Einstein, A. , 1923b, Sitzungsber. P re uss. A kad. Wis s. Phys. — A. Hermann, editor, Albert Einstein/Arnold Sommerfeld
Math. Kl. , p. 359. Briefzvechsel (Schwabe, Basel), p. 121.
Einstein, A. , 1924a, letter to M. Besso, May 24, in Speziali Einstein, A. , 1948a, Dialectica 2, 320.
(19V2), p. 201. Einstein, A. , 1948b, letter to M. Born, March 18, in Born
Einstein, A. , 1924b, Das Komptonsche Experiment, " Berliner (19V1), p. 163.
Tageblatt, April 20. Einstein, A. , 1948c, letter to M. Born, March 18, in Born
Einstein, A. , 1924c, letter to K. Joel, November 3 {unpub- {1971), p. 162.
lished, Einstein Archives). "
Einstein, A. , 1949a, "Autobiographical Notes, in P. A.
Einstein, A. , 1924d, letter to P. Ehrenfest, May 31 (unpub- Schilpp, editor, 1949, Albert Einstein: Philosopher-Scientist
lished, Einstein Archives). {Library of Living Philosophers, Evanston, Illinois).
Einstein, A. , 1924e, undated documentinthe EinsteinArchives, Einstein, A. , 1949b, letter to N. Bohr, April 4 (unpublished,
obviously written in 1924. Einstein A rchives) .
Einstein, A. , 1924f, letter to M. Born, April 29, in Born Einstein, A. , 1949c, letter to M. Besso, August 8, in Speziali
(1971), p. 82. (1972), p. 407.
Einstein, A. , 1924g, letter to P. Ehrenfest, July 12 (unpub- Einstein, A. , 1949d, letter to M. Besso, July 24, in Speziali
lished, Einstein Archives) . (19V2), p. 402.
Einstein, A. , 1924h, Sitzungsber. Preuss. Akad. Wiss. Phys. — Einstein, A. , 1949e, letter to M. Besso, August 16, in Speziali
Math. Kl. , p. 261. (1972), p. 409.
Einstein, A. , 1924i, letter to H. A. Lorentz, December 16 Einstein, A. , 1950, letter to E. Schroedinger, Dec. 22, in
(unpublished, Einstein Archives) . M. Klein, 1967, translator, Letters on Wave Mechanics
Einstein, A. , 1924j, letter to M. Besso, Jan. 5, in Speziali (Philosophical Library, New York), p. 36.
(1972), p. 197. Einstein, A. , 1951a, letter to M. Besso, Dec. 12, in Speziali
Einstein, A. , 1925a, Sitzungsber. Preuss. Akad. Wiss. Phys. — (1972), p. 453.
Math. KI. , p. 3. Einstein, A. , 1951b, in Scientific PaPers Presented to Max
Einstein, A. , 1925b, Sitzungsber. Preuss. Akad. Wiss. Phys. — Born (Hafner, New York), p. 33.
Math. Kl. , p. 18. Einstein, A. , 1952, letter to M. Born, May 12, in Born (1971),
Einstein, A. , 1925c, Z. Phys. 31, 784. p. 192.
Einstein, A. , 1925d, letter to R. A. Millikan, Sept. 1 (unpub- Einstein, A. , 1953, in Louis de Broglie, Physician et Penseur
lished, Einstein Archives). (Michel, Paris), p. 5.
Einstein, A. , 1925e, letter to M. Besso, Dec. 25, 1925, in Einstein, A. , 1954a, letter to B. Becker, March 20 (unpub-
Speziali (1972), p. 215. lished, Einstein Archives).
Einstein, A. , 1925f, Physica 5, 330. Einstein, A. , 1954b, letter toI. de Broglie, Feb. 8 (unpub-
Einstein, A. , 1925g, letter to M. Besso, July 28, in Speziali lished, Einstein Archives).
(1972), p. 209. Einstein, A. , 1954c, letter to M. Besso, August 10, in Speziali
Einstein, A. , 1926a, letter to P. Ehrenfest, April 12 (unpub- (1972), p. 525.
lished, Einstein Archives). Einstein, A. , 1956a, Lettres a Maurice Solovine (Gauthier-
Einstein, A. , 1926b, letter to Mrs. M. Born, March 7, in Villars, Paris), p. 130.
Born (1971), p. 88. Einstein, A. , 1956b, in He/le Zeit, Dunkle Zeit, edited by

Rev. leod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

C. Seelig (Europa, Zurich}. Klein, M. , 1970, Hist. Stud. Phys. Sci. 2, 1.

Einstein, A. , and P. Ehrenfest, 1922, Z. Phys. 11, 31. "
Klein, M. , 1977, "The Beginnings of the Quantum Theory, in
Einstein, A. , and P. Ehrenfest, 1923, Z. Phys. 19, 301. History of Twentieth Century Physics (Academic, New York).
Einstein, A. , and L. Hopf, 1910, Ann. Phys. (Leipz. ) 33, 1105. Klein, M. , 1979, in the Proceedings of the Einstein Centennial
Einstein, A. , B. Podolsky, and N. Rosen, 1935, Phys. Rev. 4V, Symposium, Princeton, N. J., H. Woolf, ed. March (Addison-
777. Wesley, Reading, Mass. ).
Einstein, A. , and N. Rosen, 1935, Phys. Rev. 48, 73. Kundt, A. , and E. Warburg, 1876, Ann. Phys. (Leipz. ) 157,
Einstein, A. , and O. Stern, 1913, Ann. Phys. (Leipz. ) 40, 551. 353.
Einstein, A. , R. Tolman, and B. Podolsky, 1931, Phys. Rev. Ladenburg, R. , 1909, Jahrb. Radioakt. Elektron. 17, 93, 273.
3V, 780. Landolt, H. , and R. Bornstein, 1905, Physikalisch Chemische
Elster, J., and H. Geitel, 1890, Ann. Phys. {Leipz.) 41, 166. Tabellen, 3rd edition (Springer, Berlin), p. 384.
Eucken, A. , 1912, Sitzungsber, Preuss. Akad. Wiss. Phys. — Langley, S. P. , 1886, Philos. Mag. 21, 394.
Math. Kl. , p. 141. Lenard, P. , 1902, Ann. Phys. (Leipz. ) 8, 149.
Fermi, E. , 1926, Z. Phys. 36, 902. Lenard, P. , and M. Wolf, 1889, Ann. Phys. (Leipz. ) 37, 443.
Fox, R. , 1968, Brit. J. Hist. Sci. 4, 1. Le Verrier, U. J. J., 1859, C. R. Acad. Sci. (Paris) 49, 379.
Gibbs, J. W. , 1902, Elementary Principles of Statistical Me- Lewis, G. N. , 1907, J.
Am. Chem. Soc. 29, 1165, 1516.
chanics (Yale University, New Haven, Conn. ), Chap. 7. Lewis, G. N. , 1926, Nature (Lond. ) 118, 874.
Gonzalez, J. J., and H. Wergeland, 1973, K. Nor. Vidensk. London, F. , 1938a, Nature (Lond. ) 141, 643.
Selsk. Skr. , No. 4. London, F. , 1938b, Phys. Rev. 54, 947.
Goudsmit, S., 1978, letter to A. Pais, Jan. 16. Lorentz, H. A. , 1903, in Collected Works, 1936 (Nyhoff, the
Hallwachs, W. , 1888, Ann. Phys. (Leipz. ) 33, 310. Hague), Vol. 3, p. 155.
Heisenberg, W. , 1924, letter to W. Pauli, June 8 (unpublished; Lorentz, H. A. , 1909, letter to W. Wien, April 12, quoted by
housed at Salle Pauli, CERN, Geneva). Hermann (1969), p. 68.
Heisenberg, W. , 1925, letter to A. Einstein, Nov. 30 (unpub- Lorentz, H. A. , 1916, I.es Theories Statistiques en Thermo-
lished, Einstein A rchives). dynamique (Teubner, Leipzig), p. 59.
Heisenberg, W. , 1926a, Z. Phys. 38, 411. Lorentz, H. A. , 1927, «ctures in Theoretical Physics
Heisenberg, W. , 1926b, Z. Phys. 33, 879. {McMillan, London), Vol. 2, p. 171.
Heisenberg, W. , 1927a, letter to-A. Einstein, June 10 (unpub- Lummer, O. , and E. Pringsheim, 1900, Verh. Dtsch. Phys.
lished, Einstein Archives). Ges. 2, 163.
Heisenberg, W. , 1927b, letter to A. Einstein, May 19 (unpub- Maxwell, J. C. , 1965, "The Scientific Papers of James Clerk
"
lished, Einstein A rchives). Maxwell, edited by W. P. Niven (Dover, New York), Vol. 2,
Heisenberg, W. , 1927c, Z. Phys. 43, 172. p. 418.
Heisenberg, W. , 1929, Naturwissenschaften 1V, 490. "
Mehra, J., 1975, "Satyendra Nath Bose, Biogr. Mem. Fellows
Heisenberg, W. , 1955, in Nziels Bohr and the Development of R. Soc. 21, 117.
Physics, edited by W. Pauli (McGraw-Hill, New York), p. 12. Millikan, R. , 1913, Science 37, 119.
Hermann, A. , 1969, Fruhgeschichte der Quantentheorie, 1899- Millikan, R. A. , 1914, Phys. Rev. . 4, 73.
1913 (Mosbach, Baden). Millikan, R. A. , 1915, Phys. Rev. 6, 55.
Hertz, H. , 1887, Ann. Phys. (Leipz. ) 31, 983. Millikan, R. A. , 1916a, Phys. Rev. 7, 18.
Hettner, G. , 1922, Naturwissenschaften 10, 1033. Millikan, R. A ., 1916b, Phys. Rev. 7, 355.
Huang, K., 1963, StatisticaEMechanics (Wiley, New York), p. 26. Millikan, R. A. , 1949, Rev. Mod. Phys. 21, 343.
Hughes, A. L. , 1912, Trans. R. Soc. 212, 205. Nernst, W. , 1906a, Sitzungsber. Preuss. Akad. Wiss. Phys. —
Jammer, M. , 1974, The Philosophy of Quantum Mechanics Math. KI. , p. 933.
(Wiley, New York). Nernst, W. , 1906b, Gott. Nachr. , p. 1.
Jauch, J. M. , and F. Rohrlich, 1955, The Theory of Photons Nernst, W. , 1910, Sitzungsber. Preuss. Akad. Wiss. Phys. —
and Electrons (Addison-Wesley, Reading, Mass. ), p. 40. Math. Kl. , p. 262.
Jeans, J. H. , 1905a, Philos. Mag. 10, 91. Nernst, W. , 1911a, Sitzungsber. Preuss. Akad. Wiss. Phys. —
Jeans, J. H. , 1905b, Nature (Lond. ) 72, 293. Math. Kl. , p. 306.
Jeans, J. H. , 1914, Report on Radiation and the Quantum Nernst, W. , 1911b, Sitzungsber. Preuss. Akad. Wiss. Phys. —
Theory (The Electrician, London), p. 59. Math. Kl. , p. 65.
Joel, K. , 1924, letter to A. Einstein, Oct. 28 (unpublished, Nernst, W. , 1914, Z. Elektrochem. 20, 397.
Einstein A rchives). Nernst, W. , 1916, Verh. Dtsch. Phys. Ges. 18, 83.
Jost, R. , 1977, letter to A. Pais, August 17. Nernst, W. , and F. Lindemann, 1911, Sitzungsber. Preuss.
Jost, R. , 1979, Vierteljahresschr. Naturforsch. Ges. Zurich Akad. Wiss. Phys. -Math. KI. , p. 494.
124, 5. Neumann, F. E. , 1831, Ann. Phys. (Leipz. ) 23, 32.
Kahn, B., and G. E. Uhlenbeck, 1938, Physica 4, 399. Nichols, E. F. , 1893, Phys. Rev. 1, 1.
Kangro, H. , 1976, History of Plane@'s Radiation Laze (Taylor Ornstein, L. S., and F. Zernike, 1919, Proc. K. Ned. Akad.
and Francis, London). Wet. Amsterdam 28, 280.
Keesom, W. H. , 1942, Helium (Elsevier, New York). Pais, A. , 1948, letter dated Dec. 9.
Kelvin, W. T. , 1901, Philos. Mag. 2, 1. Pais, A. , 1967, in Niels Bohr (North-Holland, Amsterdam), p.
Kinoshita, T. , 1978, in New Frontiers in High Energy Physics 215.
(Plenum, New York), p. 127. Pais, A. , 1977, Rev. Mod. Phys. 49, 925.
Kirchhoff, G. , 1859, Monatsber. Berlin, p. 662. Paschen, W. , 1897, Ann. Phys. (Leipz. ) 60, 662.
Kirchhoff, G. , 1860, Ann. Phys. und Chemic 109, 275. Pauli, W. , 1919, Phys. Z. 20, 457.
Kirchhoff, G. , and R. Bunsen, 1860, Ann. Phys. und Chem. 110, Pauli, W. , 1923, Z. Phys. 18, 272.
160. Pauli, W. , 1924, letter to N. Bohr, Oct. 2 (unpublished, housed
Kirsten, G. , and H. Korber, 1975, Physiker uber Physik'er at Salle Pauli, CERN, Geneva).
(Akademisches Verlag, Berlin), p. 201. Pauli, W. , 1925a, letter to H. A. Kramers, July 27 (unpub-
Klein, M. , 1958, Proc. Ned. Akad. Wct. Amsterdam 62, 41, lished, housed at Salle Pauli, CERN, Geneva).
51. Pauli, W. , 1925b, Z. Phys. 31, 625.
Klein, M. J., 1964, Nat. Philos. 3, 1. Pauli, W. , 1927, Z. Phys. 41, 81.
Klein, M. , 1965, Science 148, 173. Pauli, W. , 1949, in P. A. Schilpp, editor, 1949, Albert Ein-
Klein, M. , 1967, Science 157, 509. stein: PhilosoPher-Scientist (Library of Living Philosophers,

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979

 A. Pais: Einstein and the quantum theory

Evanston; Illinois), p. 156. Scott, W. T. , 1967, Erwin Schroedinger (University of Massa-

Pauli, W. , 1958, Neue Zuricher Zeitung, Jan. 12, reprinted chusetts, Amherst, Mass. ).
in Pauli (1964), Vol. 2, p. 1362. Seelig, C. , 1954, Albert Einstein (Europa, Zurich).
Pauli, W. , 1964, Collected Sci. Papers (Interscience Pub- Shankland, R. S. , 1963, Am. J. Phys. 31, 47.
lishers, N. Y.), Vol. l. Simon, F. , 1956, Yearbook Phys. Soc. London, p. 1.
Petit, A. T. , and P. L. Dulong, 1819, Ann. Chim. Phys. 10, Slater, J. C. , 1925, Phys. Rev. 25, 395.
395. Slater, J. C. , 1967, Int. J. Quantum Chem. 1s, 1.
Pipkin, F. M. , 1978, Adv. At. Mol. Phys. (Academic, New Sommerfeld, A. , 1912, Verb. Ges. Dtsch. Naturforsch. Aerzte
York), Vol. 14. 83, 31.
Planck, M. , 1900a, Verh. Dtsch. Phys. Ges. 2, 202. Sommerfeld, A. , 1922, Atombau und Spektrallinien (Vieweg,
Planck, M. , 1900b, Verh. Dtsch. Phys. Ges. 2, 237. Braunschweig), 3rd edition, p. 311.
Planck, M. , 1900c, Ann. Phys. (Leipz. ) 1, 69. Sommerfeld, A. , 1924, Atombau und Spektrallinien (Vieweg,
Planck, M. , 1901a, Ann. Phys. (Leipz. ) 4, 553. Braunschweig), 4th edition, p. VIII.
Planck, M. , 1901b, Ann. Phys. (Leipz. ) 4, 564. Sommerfeld, A. , 1968, Gesammelte Schriften (Vieweg, Braun-
Planck, M. , 1907, letter to A. Einstein (unpublished, Einstein schweig), Vol. 3, p. 10.
Ar chives) . Speziali, P. , 1972, editor, Albert Einstein Michele Besso Cor-
Planck, M. , 1909, Phys. Z. 10, 825. respondance 1903—1955 (Hermann, Paris).
Planck, M. , 1911, "Vorlesungen uber Thermodynamik" (Von Stark, J., 1909, Phys. Z. 10, 902. (Stark wrote n instead of
Veit, Leipzig), 3rd ed. , Introduction and Section 292. v. )
Planck, M. , 1958, "Wissenschaftliche Selbstbiographie, in" Stefan, J., 1879, Sitzungsber. Akad. Wiss. Wien, Math-
Max Planck, Physikalische Abhandlungen and Vortr'age, Naturwiss. Kl. 2. Abt. 79 391.
edited by M. von Laue (Vieweg, Braunschweig), Vol. 3, p. Stuewer, R. H. , 1975, The Compton Effect (Science History,
374. New York).
Pohl, R. , and L. Pringsheim, 1913, Philos. Mag. 26, 1017. Thirring, W. , 1977, letter to A. Pais, November 29.
Rayleigh, J. W. S., 1900a, Philos. Mag. 49, 539. Thomson, J. J., 1897, Philos. Mag. 44, 269.
Rayleigh, J. W. S., 1900b, Philos. Mag. 49, 98. Thomson, J. J., 1899, Philos. Mag. 48, 547.
Rayleigh, J. W. S., 1905a, Nature (Lond. ) 72, 54. Thomson, J. J., 1936, Recollections and Reflections (Bell,
Rayleigh, J. W. S., 1905b, Nature (Lond. ) 72, 243. London), p. 341.
Regnault, H. V. , 1841, Ann. Chim. Phys. 1, 129. Uhlenbeck, G. E. , 1927, Over statistische methoden in de
Richarz, F., 1893, Ann. Phys. . (Ieipz. ) 48, 708. theoric der quanta, Ph. D. Thesis (Nyhoff, The Hague).
Rosenfeld, I . , 1968, in "Fundamental Problems in Elementary Uhlenbeck, G. E. , 1976, Phys. Today29, June, p. 43.
Particle Physics, " Proceedings of the I"ouiteenth Solvay Con- Uhlenbeck, G. E. , 1979, in Proceedings of the Einstein Cen-
ference (Interscience, New York, 1968), p. 232. tennial Symposium, Princeton, N. J., edited by H. Woolf
Rubens, H. , and F. Kurlbaum, 1900, Sitzungsber. Preuss. (Addison-Wesley, Reading, Mass. ).
Akad. Wiss. Phys. -Math. Kl. , p. 929. von Schweidler, E. , 1904, Jahrb. Radioakt. Elektron. 1, 358.
Rubens, H. , and E. F. Nichols, 1897, Ann. Phys. (I.eipz. ) 60, Weber, H. F. , 1872, Ann. Phys. (Leipz. ) 147, 311.
418. Weber, H. F. , 1875, Ann. Phys. (Leipz. ) 154, 367, 553.
Rutherford, E. , 1900, Philos. Mag. 49, 1. Weyl, H. , 1930, The theory of groups and quantum mechanics
Rutherford, E. , 1918, J. Rontgen Soc. 14, 81. (Dover, New York), 2nd edition, pp. 263-264 and preface;
Rutherford, E. , 1920, Proc. R. Soc. A 97, 374. original edition published 1938 as Gruppentheorie und Quan-
Rutherford, E. , and H. Geiger, 1908, Proc. R. Soc. A 81, 162. tenmechanik.
Rutherford, E. , and F. Soddy, 1902, Philos. Mag. 4, 370, 569. Wien, W. , 1893, Sitzungsber. Dtsch. Akad. Wiss. Berl. Kl.
Salaman, E., 1979, Encounter, April, p. 19. Math. Phys. Tech. , p. 55.
Schroedinger, E. , 1924, Naturwissenschaften 36, 720. Wien, W. , 1896, Ann. Phys. (Leipz. ) 58, 662.
Schroedinger, E. , 1926a, Ann. Phys. (I.eipz. ) 79, 361. Wigand, A. , 1907, Ann. Phys. (Leipz. ) 22, 99.
Schroedinger, E., 1926b, Phys. Z. 27, 95. Wigner, E. P. , 1939, Ann. Math. 40, 149.
Schroedinger, E., 1926c, Ann. Phys. (Leipz. ) 79, 361. Williams, B., 1977, editor, Compton Scattering (McGraw-Hill,
Schroedinger, E. , 1926d, Ann. Pl&ys. (Leipz. ) 79, 734, footnote New York).
on p. 73.5. Wolfke, M. , 1921, Phys. Z. 22, 315.
Schroedinger, E. , 1955, letter to E. Broda, reproduced in Wullner, A. , 1896, Lehrbuch der Experimentalphysik (Teub-
E. Broda, Ludwig Boltzmann (Deuticke, Vienna), p. 65. ner, I eipzig), Vol. 2, p. 507.

Rev. Mod. Phys. , Vol. 51, No. 4, October 1979