Crafting the Quantum

Arnold Sommerfeld and the Practice of Theory, 18901926

Suman Seth

The MIT Press

Cambridge, Massachusetts
London, England
 2010 Massachusetts Institute of Technology

All rights reserved. No part of this book may be reproduced in any form by any electronic or
mechanical means (including photocopying, recording, or information storage and retrieval)
without permission in writing from the publisher.

For information on special quantity discounts, email special_sales@mitpress.mit.edu.

Set in Stone sans and Stone serif by Toppan Best-set Premedia Limited. Printed and bound in the
United States of America.

Library of Congress Cataloging-in-Publication Data

Seth, Suman, 1974

Crafting the quantum : Arnold Sommerfeld and the practice of theory, 18901926 / Suman Seth.
p. cm. (Transformations : studies in the history of science and technology)
Includes bibliographical references and index.
ISBN 978-0-262-01373-4 (hardcover : alk. paper)
1. Sommerfeld, Arnold, 18681951. 2. Quantum theory. 3. Physics. I. Title.
QC16.S76S48 2010
530.09'04dc22
2009022212

10 9 8 7 6 5 4 3 2 1
1 The Physics of Problems: Elements of the Sommerfeld Style,
18901910

In 1906, Sommerfeld was called to Munich to fill the chair in theoretical physics. The
position had been vacant for a dozen years, ever since Ludwig Boltzmann had left it
to return to Vienna. The high standards required by the Munich faculty and the
paucity of practitioners in theoretical physics led to an almost comical situation in
the intervening years, as the job was repeatedly offered to the Austrian in an attempt
to lure him back. Failing both in this and a further attempt to win Hendrik Antoon
Lorentz for the position, the search moved on to younger men. Of the three candi-
dates (Emil Wiechert, Emil Cohn, and Sommerfeld), only Cohn held a position in
physics. When Wiechert declined the position, the ministry offered it to Sommerfeld,
who had come highly recommended by both Lorentz and Boltzmann.1 Sommerfelds
work on Rntgen-ray (i.e., X-ray) diffraction and the electron theory had probably
also attracted the attention of Wilhelm Rntgen, Munichs professor of experimental
physics. Rntgen signaled his approval, and the 38-year-old Sommerfeld jumped at
the opportunity to occupy a full professorship at the prestigious university.
The opportunity, however, brought with it a major challenge. In Munich I had for
the first time, wrote Sommerfeld in an autobiographical sketch, to give lectures on
the different areas of theoretical physics and special lectures about current questions.
From the beginning I plugged away atand wouldnt let any trouble divert me
fromthe founding through Seminar and Colloquia activities of a nursery [Pflanz-
sttte] for theoretical physics in Munich. These lectures,2 written in Sommerfelds
hand and delivered during a critically formative period in the development of theo-
retical physics, provided a means for Sommerfeld to educate a new generation of stu-
dents and researchers in his methods. They also offered a means for him to develop
these methods and to master the relevant material himself. These early lectures
(19061910), Sommerfelds published writings, and his reports on his students dis-
sertations to the Munich philosophical faculty constitute the basis for this chapter,
which is an attempt to describe the Sommerfeld style of theoretical physics. This is
the first of three chapters devoted to an examination of the Munich school in the
years before 1918. Chapter 2 takes up the question of the transference of this style to
14 Chapter 1

a generation of students and researchers via an examination of the day-to-day practices

of pedagogy. In chapter 3, Sommerfelds correspondence, particularly letters written
to him from his students during the First World War, is used to delineate the subse-
quent applications of Sommerfelds teachings during the years of the conflict.
Sommerfeld worked closely with his students and was known as an excellent teacher.
He fused teaching and research, incorporating his most recent work into lectures.
Concentrating on Sommerfelds lectures thus provides a means of bridging the gap
between pedagogy and research practice and a means of better understanding Som-
merfelds approach to the problems of physics and the state of the field as he saw it.3
The casual prose of the lectures is in stark contrast to more guarded comments in
Sommerfelds published writings. For example, in his lectures on heat radiation, first
delivered in 1907, one can clearly see an attempt to master Max Plancks Vorlesungen
ber die Theorie der Wrmestrahlung, published the year before.4 At the same time,
Sommerfeld was (as he was not in his publications) openly skeptical toward Plancks
black-body work, preferring the theories of Lorentz and James Jeans.5 The lectures thus
provide insight into both the early years of one of the most important sites for theo-
retical physics in the early twentieth century and Sommerfelds own approach to the
problems of contemporary physics. Sommerfelds case also, and more generally, pro-
vides a particularly telling example of one of the central arguments of this book: that
theoretical physics at the turn of the twentieth century cannot be understood as a
distillation of theory from physics, but rather must be seen as having been actively
constructed from multiple and varied parts. Far from being merely a subset of an
existing discipline, the subject that emerged in Munich was a blend of at least three
components: mathematics, physics, and engineering. Drawing from his experience in
each of these three fields, Sommerfeld selected and modified components that would
make up the theoretical physics of the Sommerfeld School.6
After 1906, having previously held positions teaching first mathematics and then
technical mechanics, Sommerfeld quite consciously refashioned himself into a theo-
retical physicist. Problems previously deemed mathematical were now reformulated
to emphasize a new, more physical perspective; technical applications were blended
with mathematical and physical methods that may well have seemed alien to Som-
merfelds former colleagues in engineering. Not merely an incorporation of fields dis-
tinct from physics, however, the process also involved the selection and emphasis of
specific areas within physics itselfin particular, those parts of the field that were in
accord with the electromagnetic view of nature, a worldview of which Sommerfeld
was an ardent supporter.
Central, indeed perhaps essential, to Sommerfelds work in these eclectic fields was
his and his students emphasis on the solution of specific problems in areas such as
wireless telegraphy, the wearing on ball bearings, the turbulent flow of water in the
channels of the Isar, and black-body radiation. It would be remarkably fitting that the
The Physics of Problems 15

Festschrift prepared for him by his students on the occasion of his sixtieth birthday
should be titled simply Problems of Modern Physics.7 Eschewing an axiomatic and gen-
eralized approach, Sommerfeld sought out, both in his teaching and his research,
issues of contemporary interest that he would then attempt to understand in theoreti-
cal detail. And it would be these problemsboth in terms of topic and their forms of
solutionthat would provide coherencetechnical unityfor the wide-ranging
work of the Sommerfeld School.

Mathematics: The Kind of Notion We Call Heat

Der Verstand schpft seine Gesetze (a priori) nicht aus der Natur, sondern schreibt sie dieser vor.
Kant

The words abovemeaning The understanding draws its laws (a priori) not from
Nature, but rather prescribes them to herare to be found at the beginning of a draft
of what is probably Sommerfelds first scientific work.8 They should not, perhaps, be
taken as evidence of any particular philosophical commitment. Sommerfeld was born
and grew up in Knigsberg, where his father was a practicing physician.9 At the local
university, he had walked the same halls as Germanys most famous philosopher. Yet,
insofar as Kant was important to Sommerfelds development in this period, it was
probably only through his dictum that in every specific natural theory only so much
actual science can be found as there is mathematics within it.10
The seven-page paper never takes up philosophy again, but a parenthetical remark
at the beginning lays out Sommerfelds main aim: The leading thought in my work
is to simplify the problem of heat conduction by establishing a characteristic func-
tion.11 While his approach here was principally mathematical, the inspiration for
the work could be found in an existing physical problem, one set as a prize question,
worth 300 Marks, by the Physikalisch-konomische Gesellschaft (Knigsberg Physical-
Economical Society), of which Sommerfelds father was a member.12 In the local
Botanic Garden, Franz Neumann, a co-founder of the mathematical-physical seminar
at Knigsberg, had established a meteorological station that measured the tempera-
ture below the surface. The taskset by three of Neumanns students and four other
members of a commission established by the society in 1889was to analyze the
data the station produced in its measurements of temperature at different depths.
The Society, read the question, would like as comprehensive as possible a theoreti-
cal evaluation of the geothermal measurements made at Knigsberg, especially to
understand the thermal conductivity of the earth and the causes of it. . . . 13
Sommerfeld set to work on solving the problem in a long essay, clearly intended to
be presented to the society, and approached the institute for theoretical physics
and Neumanns eventual successor there, Paul Volkmann (also a member of the prize
16 Chapter 1

commission), for aid in evaluating the terms that arose in his efforts to reduce an
arbitrary curve to the sum of a trigonometric series.14 Sommerfeld and Wiechert,
Volkmanns assistant, built an integrating machine to deal with the calculations,
although this met with limited success in its operation as a result of what Sommerfeld
described only as an insufficient practical understanding of the apparatus.15 In the
end, owing to a significant error made in assumptions about appropriate boundary
conditions, Sommerfeld was forced to withdraw the solution. As it stood, however,
the paper he had prepared was a good example of what Olesko has described as the
theoretical physics of the Knigsberg School.16 Sommerfeld, who had attended Volk-
manns lectures, had clearly noted his teachers enthusiasm for the problem set forth
by the Physical-Economic Society, and provided a solution that would conform to his
expectations. In the first of two sections, Sommerfeld set out the mathematical theory
for the ideal case of heat conduction in terms of Fourier series and then in terms of
Fourier integrals. A short discussion of the operation of the integrating machine fol-
lowed. In the second section, Sommerfeld dealt with the modifications to the theory
that had to be considered in the real world: corrections for nonperiodic temperature
functions, for inhomogeneous surfaces, and for non-level ground (the station stood
at the foot of a small hill). Finally, he considered the non-ideal character of the mea-
suring instruments, offering a theoretical treatment of the air thermometer. The
project shows a striking similarity to part of the Knigsberg paradigm for theoretical
physics: essentially the mathematical analysis of experiment. The only important
element missing was a numerical error analysis of the results.
Yet, while he could do the problem Knigsberg-style, Sommerfeld did not revel
in it. His dissertation on arbitrary functions in mathematical physics, which he later
claimed to have conceived and written out in a few weeks, made use of his earlier
work on Fourier series and integrals, but largely without mention of the physics of
heat conduction.17 Consistently, over the course of his life, Sommerfeld would refer
to himself in this period as a mathematician. His papers, even when dealing with
possible topics in physics, would often emphasize that they were mathematical treat-
ments, such as his 1894 work Zur mathematischen Theorie der Beugungserscheinungen
(On the Mathematical Theory of Diffraction Phenomena) or his Habilitationsschrift,
Die Mathematische Theorie der Diffraction (1896) (The Mathematical Theory of Diffrac-
tion), which he bragged would wake physicists up to the flaws in their analysis. In a
letter to his mother in 1894, Sommerfeld referred to all that physicists had done for
the mathematical theory of optics as humbug and meaningless words.18 The same
year he refused the offer of an assistantship with the theoretical physicist Woldemar
Voigt on the ground that he would have to work there on matters which I do not
wholeheartedly consider as my mission. That mission, especially given that Som-
merfeld was happy to take up an assistantship with the Gttingen mathematician
Felix Klein, was clearly mathematics.19
The Physics of Problems 17

Yet even more than a preference for mathematics over other fields, the young Som-
merfeld seems to have had a distaste for the physics of Volkmann and others on the
prize commission. Upon hearing of the death of Heinrich Hertz, for example, he wrote
to his mother, asking her whether she had read about it yet: It is Awful! The man
began his brilliant experimental investigations five years ago. Half of all physicists at
the moment are following in his footsteps and are working on Hertzian oscillations.
There are few discoveries that can stand next to his electromagnetic light waves. If it
had to be a physicist that died, why couldnt it have been one of the useless Papes,
Volkmanns etc.20 In 1908, however, when he took up the problem of heat conduc-
tion again, Sommerfelds approach to physics (if not to particular physicists) had
changed completely. The title of his summer lecturesHeat Conduction, Diffusion
and Conduction of Electricity, together with their Molecular and Electron-Theoretical
Connectionsalready attested to an involvement with matters of profound interest
to physicists at the time. The electron theory in particular was one of the main areas
of research of Lorentz, one of the most respected members of the physical community,
and the faculty at Munich had, in their choice of Boltzmanns successor, made clear
their desire for someone expert in what Sommerfeld would term the burning ques-
tions of electrons.21 In fact, if any one topic could be said to have been the center of
path-breaking theoretical research it was this one, and Sommerfeld himself had already
contributed three significant articles to the field, published in the reports of the
Gttingen Science Society between 1904 and 1905.22
If the topic was explicitly physical, the approach was even more so. Commenting
on the topic of his Knigsberg lecture, heat conduction, Sommerfeld noted that it
was the source of the methods of mathematical physics: [T]he book by Fourier is
the original Organon of these methods. At the beginning of the 19th century the
problems of heat conduction were the order of the day. Apart from Fourier: Poisson,
Lam, Kelvin. Nowadays its totally out of fashion, because its physical result [is]
not great.23
Yet these mathematical methods would not be the sum total of the course. Because
the mathematical approach [Gesicht] is too one-sided, Sommerfeld wrote we will
give the lecture in addition a more physical orientation. This then has a more current
interest.24 Thus, Sommerfeld took up the topic in a way that Paul Ewald, one of his
students, described as characteristic. He penetrated quickly through the classical parts
of the subject and, after having laid this foundation, dwelt on the topical problems
requiring researchship waves and turbulence in Hydrodynamics, theory of relativity
in Electrodynamics, radiation theory, specific heat and energy quanta in Thermody-
namics.25 Recent research topics discussed in the lectures included not only electron
theory in general but also Nernsts osmotic theory and the Wiedemann-Franz
law, which postulated a proportionality between coefficients of heat and electrical
conduction.
18 Chapter 1

The most fundamental equations of early-nineteenth-century mathematical physics

were consciously redirected toward a more physicalalbeit theoretical physicalend.
Under the German academic principle of Lehrfreiheit (academic freedom), Sommerfeld
had no requirements other than to teach something called theoretical physics. No
particular subjects were prescribed, and no specific curriculum had to be worked
through. His mode of structuring his lectures can thus be seen as representative of his
own vision of the shape of the field. It was, clearly, a vision that made use of the
methods of mathematicians, but did not necessarily accept their mindset. From his
origins as a mathematician in Knigsberg, working on Fourier series and arbitrary
functions, Sommerfeld had begun, after the turn of the century and certainly after the
move to Munich, to refashion himself.26
Perhaps most telling of Sommerfelds explicit attempts to make such a change in
persona was his insistence on being provided with experimental facilities. Experimen-
tal ability was still considered a fundamental prerequisite for a fully trained physicist,
and Sommerfeld clearly felt that he lacked such ability. His previous chair had been
in technical mechanics (a fusion of mathematics and engineering), and it seems
reasonable to read his desire for an adequate laboratory as part of an attempt to shift
fields, and, just as importantly, to be seen as doing so.27 This latter aspect comes to
seem even more important in light of the fact that very little experimental work was
actually done at the institute. Sommerfeld himself did none, and while he began at
Munich emphasizing the importance of laboratory work to his students, compara-
tively few of them actually followed that path. When Ewald arrived, in 1908, only
one student was to be found in the fourth room of the building, that devoted to
experiment: Ludwig Hopf was attempting, without a great deal of success, to observe
and measure the onset of turbulence in an open trough as he varied the velocity
of flow and the viscosity.28 If a detailed engagement with experimental and observa-
tional data would remain a lasting characteristic of the Sommerfeld School, a dual
experimental-theoretical strategy would slowly drop away.29 By 1910, Sommerfeld had
given more than a dozen courses and had published a series of well-received papers
in physics. One might also imagine that his anxietythe worry expressed to his
mother in 1894 that he knew nothing about experiment and feared making a fool of
himself in physicshad receded.30 Sommerfelds shift from a mathematician to a
theoretical physicist was underway.

Technical Mechanics: Gyroscopes and Ship Waves

The year after completing his thesis in 1891, Sommerfeld sat for the state exam that
would qualify him as a teacher. After satisfying the requirements of military service,
he moved in 1893 to Gttingen, the site of (as he put it) mathematical high culture.
Family connections initially won him a position as assistant to Theodor Liebisch in
The Physics of Problems 19

the mineralogical institute, but his interests continued to lie in mathematics. In 1894
he became an assistant to Felix Klein, with whom he completed his Habilitation thesis
two years later. It was during this period that Sommerfeld came to an appreciation of
the value of connections between mathematics, physics, and engineering. He later
credited Klein with giving to my mathematical outlook that sense which is best suited
to applications.31
In 1897 Sommerfeld became professor of mathematics at the mining academy in
Clausthal, a position that seems to have involved teaching basic mathematics to
largely uninterested students. Kleins maneuverings eventually resulted in a better
offer, and in 1900 he took up the professorship for technical mechanics at the tech-
nische Hochschule at Aachen.32 As a student of Kleins, however, he was initially
greeted with suspicion: In 1900 I was called as a professor of technical mechanics to
the Aachen Hochschule. As a result, I was compelled for several years to apply the
main focus of my works to engineering problems. I had there the satisfaction, that
my Aachen colleagues and students, who first regarded the pure mathematician
with mistrust, soon recognized me as a useful member, not only in education, but
also in practical matters of engineering, so that I was consulted for expert reports, for
collaboration for the engineering society etc.33
That Sommerfeld was not exaggerating the degree of hostility which commonly
met non-engineers when hired into a technical college may be seen from the welcome
that greeted Otto Krigar-Menzel, a student of Helmholtzs, when he was hired to the
professorship of theoretical physics at the technische Hochschule at Berlin in 1904.34
Alois Riedler, who had previously held a chair at Aachen and who numbered among
the most prominent of the engineering professors on the faculty in Berlin, serving as
Rector in 1899, reacted rapidly and with aggression. There had been an urgent
need, he wrote to the Kultusminister (minister of education and arts) in July, for a
teacher who knows our needs and who gives to physics, which at the moment has
proven to be a scarcely fruitful activity, a new style and content.35 That need, Riedler
made clear, had not been satisfied through the hire of Krigar-Menzel. The teacher
and the teaching task stand in the most glaring opposition to the urgent needs of our
Hochschule. The Hochschule, he continued, is no playground for areas of science
that belong exclusively in universities, and can find with us as many listeners as
engineering lectures among theologians.36
A tension, apparent and even acknowledged, thus existed between Sommerfelds
early training as a mathematician and his need to fit into his new role in an engineer-
ing school. One can begin to see that there would be two reasons that Sommerfelds
Aachen years would be such a formative element of his physics of problems. Not
only did he arrive at the school at an early, impressionable stage of his career, but
alsoand more importantlyhe was compelled while there to adapt himself to his
(at times antagonistic) environment.
20 Chapter 1

Things had changed dramatically since 1870, when the Rheinisch-Westphlisch

Polytechnical School in Aachen celebrated a long-awaited opening as Prussian and
French forces faced one another across battlefields not far away. Then technical col-
leges had sought to imitate universities as closely as possible, and a mathematician
like Sommerfeld would have been welcomed as one who could be counted on to
provide the requisite Bildung (cultivation) for engineering students. Indeed, it was in
this period that Klein himself took up a position at the technische Hochschule in
Munich, his classes on analytic geometry drawing over two hundred students.37 By
the turn of the century, however, an increasingly muscular and praxis-minded engi-
neering community could set its own terms, consciously contrasting the aims of
technische Hochschulen with the classical university.
Around the middle of the nineteenth century, when the Verein Deutscher Inge-
nieure (Society of German Engineers) was founded, an enormous distinction still
divided the artisanal engineer from the Hochschule engineers who led the society.
Without the social status that in England, for example, had followed from an obvious
economic utility, and in the absence of an effective German bourgeoisie, the societys
leadership had sought a means of raising the public status of engineers by allying
them with an intellectual aristocracy, the so-called German Mandarins.38 The espoused
ideals of engineering became those of the mandarinate: pure over applied science,
general cultivation over specific training. Engineers would pursue the abstract ethics
of German Kultur in an institution for higher learning rather than grub after money
in a workshop or factory.
In the 1880s, however, entrepreneurial visions of engineering replaced professo-
rial ones. Under the leadership of Theodor Peters, who became the business manager
of the Verein Deutscher Ingenieure in 1882, the society emphasized far more practi-
cal issues, attracting to its Zeitschrift authors who wrote about concrete matters and
emphasized criteria such as application, cost effectiveness, fuel efficiency, material
strength and ease of construction.39 In terms of education, this change in outlook
meant radical alteration. At the level of secondary schooling, moving away from its
former support for the classical Gymnasium, the association began a vehement push
for the modern Realschulen.40 At the tertiary level, new community leaders argued for
a vision of technische Hochschulen as institutions equal to universities, but nonethe-
less completely independent of, and fundamentally different from them (see figures
1.11.3).41 This effort to keep the two institutions separate but equal would constitute
one of the principal differences between these engineers and Klein, who sought to
bring them ever closer.
After the late 1870s, the emphasis was on increased shop and laboratory training
for engineering students at the technische Hochschulen. In terms of detailed curricu-
lar change, the 1890s and later years saw the de-emphasis of calculus in favor
of less precise but much more pragmatic graphic methods.42 More generally, and
The Physics of Problems 21

Figure 1.1
The four faculties of the university (theology, philosophy, medicine, law), represented by four
spheres, balanced by the six sections of the technische Hochschule (architecture, civil engineer-
ing, mechanical engineering, shipbuilding, chemistry and mining, general). An imperial eagle
perches above the scale. Source: Die Hundertjahrfeier der kgl. T. H. zu Berlin, 1899 (1900).

controversially, prominent engineers, including Alois Riedler, argued against what

they portrayed as an excessive reliance on theory over practiceon Wissen (knowl-
edge) over Knnen (Capability)and began to agitate for a demotion of mathematics
and other theoretical subjects to the status of auxiliary sciences [Hilfswissen-
schaften].43 Just how vehement this agitation was can be seen from the lines below,
from a tract written in 1895. Having acknowledged that engineering requires the aid
of sciences to climb the heights of the knowledge of reality, Riedler asserted that

the ruling Theory remains below in the comfortable valley: The terrible preparatory education
forces it to do so. Down there in the valley, it pursues all kinds of Gymnastics, [yet] knows not
itself the efforts and dangers of the mountains and also deceives its disciples about them. The
disciples probably storm the heights sometimes, far away from the creating world, with neither
aim nor purpose. But the learned, unfruitful theory, when it raises itself to bold flight, then flies
from the sight of the real world, up over the clouds to Abel and Riemann, where the Theta func-
tions disappear, where the special concept Dimension is replaced by the general concept
manifold and then can be gymnastically performed [geturnt] in a world of four and more
manifolds.44
22 Chapter 1

Figure 1.2
The university (represented by four texts, labeled according to the four faculties) clasps hands
with the technische Hochschule (represented by a hand using a compass). Note, further, that
the V shape subtended by the quills in the ink-pot on the left opens to the sky, whereas the V
formed by the compass offers an inverse image, opening to the Earth. Source: Die Hundertjahrfeier
der kgl. T. H. zu Berlin, 1899 (1900).

In view of how closely this mocked mathematics resembled some of the work of the
Gttingen school, it should be taken as a mark of Kleins power that he was able to
place Sommerfeld in a technische Hochschule at all.
A great deal was at stake for young theoreticians in the debate over the importance
of the mathematical sciences to engineers. Riedler and others were arguing for a sub-
stantial reductionin schools that had long been seen as stepping stones to better
jobs in universitiesin both the numbers and significance of such young scholars. In
this environment, Sommerfelds 1903 speech The Scientific Aims and Results of
Modern Technical Mechanics takes on a particular significance. One can read it both
as a defense of the technical sciences, now that they have developed from their own
innate power a confident and self-sufficient position, and as a claim for the relevance
of theory to such sciences.45 Sommerfeld, in other wordsas befitted a man in his
positionwas seeking to play two roles, both that of theoretician and of engineer.
The Physics of Problems 23

Figure 1.3
The six sections of the technische Hochschule. Note that the cherub representing the general
section is the only one not using his hands. Instead, he stares (rather perplexed) at an integral.
Source: Die Hundertjahrfeier der kgl. T. H. zu Berlin, 1899 (1900).
24 Chapter 1

Sommerfeld and Klein

Sommerfeld presented his paper on modern technical mechanics as part of a panel
organized by Klein, and there are strong similarities between the positions of the two
men. Kleins program sought to bring Hochschulen and universities closer together.
This program, however, did not advocate equality for engineers; rather, it sought to
subordinate all fields to the central control of a mathematics that would function as
a specific power pervading the whole.46 Thus, his plan for a Gttingen institute for
physical technical researchposited as an educational counterpart to the physikalisch-
technische Reichsanstaltwould also have a complementary relationship to the tech-
nische Hochschulen. The latter would provide an engineering education for the
Front-Line Officers, while Kleins smaller, more exclusive institute would supply a
specific education in the exact sciences for future General Staff Officers.47 It was not
difficult to work out the implied power relations between the two, and one can see
why engineers might fight so strongly to keep the two institutions separate while
guaranteeing their equal status.48
It is perhaps not surprising that Sommerfeld, Kleins student and supporter, was met
with distrust when he took up his position at Aachen. And although Sommerfeld
claimed to win over many of his doubters, there is a strong sense in which he remained
far more of a member of the general staff than a front-line officer. He certainly worked
on engineering topics, examining the problems involved in railway braking, studying
the operation of dynamos, and, in what he saw as his most important project, working
on the hydrodynamic theory of lubrication.49 Yet even in the latter the aim of the
task was less that of providing an effective solution to a practical problemSommer-
feld admitted that his theory doesnt correctly treat in all parts the real occurrences
of bearing frictionand more that of demonstrating the general applicability of the
mathematics. His purpose, he claimed, was merely to show how far one can come
with the pure hydrodynamic theory.50 The enjoyment of the project, he would write
later, came from helping the power of mathematical-physical thought to a victory
in its exact treatment of an apparently inapproachable subject.51 The papers on this
and other technical topics should be seen as displays of mathematical virtuosity and
proofs that engineering was part of the whole that mathematics pervaded.
And yet it is also necessary to distinguish between Klein and Sommerfeld. Klein, it
has been argued, was driven to his attempts to bring universities and technische
Hochschulen together by the fear that an increasingly powerful engineering move-
ment would sideline university mathematics.52 He thus pushed for a mathematics that
would be more useful for the technical sciences, irritating both engineers (who were
increasingly independent) and neohumanist mathematicians (who resented the
importation of Americanismus and the loss of their pure ideals).53 But, as the com-
ments about the Gttingen institute imply, Kleins position was not completely cen-
tered between these two poles. While his rhetoric was as much about mathematicians
The Physics of Problems 25

learning from engineers and vice versa, this was not the way that things actually
played out. Thus, in spite of his support for the efforts of those who worked on tech-
nical mechanics, it was not such a compromise between engineering and mathematics
that emerged in 1898 as a result of Kleins maneuverings over teaching reform, but
the far less industrial applied mathematics. An alliance with the physical sciences
was pursued largely in order to strengthen the position of mathematicians, and a
separation between mathematicians and natural scientists could still occur, Klein
noted to Paul Gordon in 1890, if we are strong enough or feel overly restricted.54
The ultimate aim was the salvation of mathematics, and Kleins own neohumanist
predilections meant that he leaned more strongly toward a solution that would keep
mathematics strong and independent. Quite simply, he advocated mathematical
engineering rather than engineering mathematics.55
For Sommerfeld, on the other hand, with his chair in technical mechanics, the
applications of mathematics were avowedly industrial. In his 1903 paper, Sommer-
feld spoke on his own field, discussing problems of machine construction and of
gyroscopic compasses on ships. The audience, of course, was different here, and Klein
had organized the session, so one can assume he was in full accord with the idea
that mathematicians should be told to wake up to the need to enter the industrial
age. Yet Sommerfelds active contact with the problems of technology was, as the
topics of his papers demonstrate, more than an occasional tactic. His aim may have
been, like Kleins, to sell mathematics to the world, but Sommerfeld also clearly
believed that mathematicians and physicists had something equivalent to buy. Since
Sommerfeld was the outsider at Aachen, it was largely a one-way trade. But once he
had (much to Kleins chagrin) given up that post in favor of Munich, the situation
changed. Like his interest in mathematical physics, Sommerfelds enthusiasm for
technical mechanics became an integral part of his theoretical physics.
One need not portray Klein as a Machiavellian figure, politicking in support of his
own field, in order to acknowledge the local and contextual differences between him
and Sommerfeld. Those differences are clear in a book they wrote together: the four-
volume, 966-page ber die Theorie des Kreisels (On the Theory of the Gyroscope).56 The
idea of working on such a topic came to Klein during the annual meeting of the Verein
Deutscher Ingenieure in 1895. Here the thought came to me, he wrote later, of
connecting theoretical reflections that were familiar to me with the needs of physical
and technical understanding via a detailed lecture over a specific mechanical problem,
like gyroscopic theory.57 A two-hour lecture followed, after which Klein approached
Sommerfeld about writing up and publishing a more detailed but still small treatment
of the subject, much as had been done with the lectures that same semester on number
theory.58
The timing of the book, with the first volumes appearing at the height of Kleins
battles, is crucial for understanding its purpose. Whereas the second volume was a
26 Chapter 1

more collaborative effort, the first volume was almost entirely based on Kleins ideas,
and we can use it as a means of better understanding his position.59 Klein had long
argued for a more intuitive means of teaching mathematics, one that would have an
appeal beyond the ranks of pure mathematicians and would establish not only a
knowledge of mechanics, but so to say, a feeling for it.60 At the same time, he wanted
to push the idea that mathematics was useful beyond its own boundaries and that it
had a role to play in practice as well as in theory. ber die Theorie des Kreisels would
thus have two aimsone for each of its audiences. Gyroscopic motion was a well-
known but rather poorly understood problem in mechanicsone, moreover, that had
applications in many neighboring fields, including astronomy and physics. The gyro-
scope thus offered a bridge to the natural sciences. At the same time, an appropriate
analysis of gyroscopic motion could provide an example of a general approach to
mechanics for pure mathematicians. Klein deplored what he referred to as the too
abstract and formal direction that mechanics had taken in Germany, a tack that hin-
dered a direct understanding: The student who probably learns general mechanical
principles analytically by heart, for that reason never grasps their actual mechanical
meaning in a lively enough way and appears, when positioned before a specific
problem, clumsy in its solution.61
The solutions the text offered, which it drew largely from the pattern of English
textbooks, would be manifold. First, it eschewed the approach of Lagrange and those
who followed him, and emphasized the importance of geometrical representation.
Second, it emphasized the importance of understanding the causes of motion and
made these comprehensible through the introduction of vector notation. Finally,
emphasis was placed not on understanding the mechanics of the problem through
the equations, but rather on allowing the analytical formulation, as a final conse-
quence, to appear of its own accord from a fundamental understanding of the mechan-
ical relationships.62
The book that Klein had planned was thus principally one that would elucidate
complicated mathematical concepts by applying them to a more easily grasped, more
intuitive mechanical object. Through a work originally intended to span at most two
volumes, the reader would be introduced to some of the mathematics of Euler,
Lagrange, and Jacobi, and to elliptical functions and integrals. The place of these topics
in areas outside of mathematics would be emphasized to appeal to a wider audience.
And yet, although Klein was to talk in 1922 of the needs of physical and technical
understanding that the gyroscope book was supposed to satisfy, the book that he
planned in 1895 did not deal with any technical problems. The applications men-
tioned were to astronomy and physics, and even these were largely dealt with in the
volume that Sommerfeld completed while in Aachen. The technical applications of
ber die Theorie des Kreisels were all undertaken by Sommerfeld.63
The Physics of Problems 27

The opening words of Klein and Sommerfelds joint introduction to the fourth
volume (1910) laid out the division of labor that had developed between the inception
and the completion of their work:

When Felix Klein held a two-hour lecture in the winter semester of 1895/96 On the Gyroscope
he intended it in the first instance to emphasize the immediate grasp of mechanical problems
that was extended particularly in England in opposition to the more abstract coloring [Frbung]
of the German schools and, on the other hand, to render fruitful the methods of Riemannian
function theory that were principally built up in Germany. The consideration of applications
and physical reality was at the time certainly indicated and heartily supported, but was not yet
carried out to its full extent.

In the extensive publication stemming from the pen of A. Sommerfeld the interest in applications
took over more and more, particularly after his taking up of chairs in technical mechanics and
later in physics.64

The result of the increasing emphasis on astronomical, geophysical, and technical

applications, the authors continued, was a forced reworking of the kind of mathemat-
ics being utilized within the text. Thus, although the original aim of the text under
Klein had been to produce intuitive instruction methods for particular mathematical
techniques, Sommerfeld had introduced a change in mathematical content to suit an
emphasis on scientific and engineering applications.65

Hydrodynamics: Theory in Practice

One can track, through the volumes of ber die Theorie des Kreisels, the changes in
Sommerfelds intellectual interests: mathematics when he arrived in Gttingen, applied
mathematics after his time as assistant to Klein, increasingly industrial problems while
a professor for technical mechanics at Aachen. In the same period, however, Som-
merfeld had been engaged on another extensive project: the editorship of the physics
section of Die Enzyklopdie der mathematischen Wissenschaften mit Einschluss ihrer
Anwendungen.66 This latter, after his failed attempt to pass the job on to his colleagues,
became Sommerfelds letter of introduction to leading members of the physics com-
munity, including, in particular, Boltzmann and Lorentz. Between these and the
contacts made during the work on gyroscopic theory (including Carl Cranz, the Pope
of Ballistics), Sommerfeld came to be connected to a majority of mathematics-minded
professors in the German-speaking world.67
This multitude of interests came together after Sommerfelds call to Munich in 1906,
as can be seen in his work on hydrodynamics. A subject Sommerfeld had studied as
a mathematician in 1900, and had called upon in his examination of the theory of
lubrication while at Aachen, now became the topic of a lecture series in the summer
semester of 1908.68 In 1909, after a five-year hiatus, he returned to the question of
28 Chapter 1

hydrodynamics in his published work. His paper A Contribution to the Hydrody-

namical Explanation of Turbulent Fluid Motion displayed the multiple resources
upon which its author was able to draw: published as part of the proceedings of the
fourth international mathematical congress in Rome, its second reference was to a
book on theoretical physics, and its first section offered a literature review of works
that dealt with the relationship between theory and praxis.69 The paper began as
follows: The cleft between technical hydraulics on the one side and theoretical hydro-
dynamics on the other, between the investigations of large-scale properties on the one
hand and laboratory investigations of capillarity on the other, has often been noted
and bemoaned.70 The gap was first bridged, Sommerfeld claimed, by Osborne Reyn-
olds, who introduced a theoretical means of examining the phenomenon of turbu-
lence. He did not, however, supply an exact method of calculating the value of the
parameter that bears his name (the Reynolds number). This was the role that Lorentz
played, simplifying the case of Couettes flow (that between two cylinders) to the
geometrically easier case of the flow between two parallel plates, one moving at a fixed
velocity with the velocity gradient approximated as linear.71
Sommerfelds work, which took up the problem at this point, restricted itself to
examining the mathematically easier case of the onset of turbulencethe border
between turbulent and non-turbulent flowand did not consider the character of the
flow above the critical value of Reynoldss parameter. This problem, as well as a com-
plete discussion of the equation he arrived at which appears to me to form the actual
content of the problem of turbulence, he reserved for another time.72 The problem
chosen, that of turbulence, was portrayed as standing on the border between technical
and theoretical studies, one foot on either side. Technical hydraulics and theoretical
hydrodynamics come together in a subject that would form a part of the lectures on
theoretical physics. The fusion would pervade the work of the Sommerfeld School
more generally. Sommerfelds research topics often became dissertation projects for
his students; thus he passed on the problem of turbulence to Ludwig Hopf in an
attempt yet again to unite the worlds of theory and engineering.
Hopfs partially experimental work was touched upon above, but little mention was
made there of the importance of technical applications within it. At first glance, studies
of sugar water in troughs may seem obscure, but the point was to extend Sommerfelds
own studies. The open channel in which the water flowed would be a different case
and one more often seen in the real-world systems, such as rivers or canalsfrom that
usually studied, where the fluid was enclosed on all sides. Sommerfeld hoped that the
different set-up would allow the observation of different phenomena, such as the effect
of capillarity, as the water shifted from laminar to turbulent flow.73 Hopfs project
would thus, quite literally, combine what Sommerfeld saw as large-scale properties,
and phenomena studied in the laboratory, the sets of observations that Sommerfeld
associated, respectively, with technical hydraulics and theoretical hydrodynamics.
The Physics of Problems 29

In almost perfect mimicry of his supervisors paper, Hopfs dissertation began with
the placement of hydrodynamics within physics, and continued with a discussion of
the theory/praxis divide:

In perhaps no other branch of physics do theory and praxis remain as foreign to each other as
in Hydrodynamics; and it is particularly peculiar and regrettable, given that both have grown
into very extensive and important areas. Alone, the hydrodynamic theory on the one hand
faced insurmountable mathematical difficulties in its treatment of phenomena in real fluids; on
the other hand it opened up an easily accessible and praiseworthy [dankbar] area by going
around these difficulties via idealized frictionless and eddyless fluids. Engineering [Technik] natu-
rally developed completely independently of these theories and created its own mathematical
fundamentals, with all the pros and cons of empirical formulas. Right at the moment the need
for a metamorphosis of these conditions has become many times more keen, and a great deal of
attention has been turned to real fluid [dynamical] problems.74

Of these latter, that of fluid flow, Hopf claimed, is in fact the most urgent and most
interesting of hydrodynamical problems, precisely because it can be completely and
satisfactorily solved.75 The laboratory examination of these real-world flows would
constitute half of Hopfs dissertation.
Sommerfeld described the second half, On Ship Waves, as purely theoretical
when he reported on the project to the philosophical faculty.76 That was certainly true
in the sense that there were no experimental observations in that section. Yet even
here, Hopf made clear the dual role that his investigations served. The shape of waves
which accompany a ship moving on tranquil water, and the corresponding resistance,
against which the ship exerts continual effort, are not only of great significance practi-
cally, but also offer the theoretician several interesting problems.77 As Sommerfeld
had advised, the theoretician was coming into active contact with the problems of
technologyso active, in fact, that Hopfs investigation later found outside financial
support from the Gesellschaft Mittlere Isar, a group that wanted to regulate the flow
of Munichs largest river.78 Purely theoretical it may have been in one sense, but
Hopfs dissertation found a significant role to play in practice.

Physics: The Electromagnetic Worldview79

It should by now be clear that the theoretical physics espoused by Sommerfeld was
considerably more than simply a subset of physics. Sommerfeld drew heavily from his
knowledge of both mathematics and technical mechanics in constructing his own
brand of theory. Theoretical physics in Munich was an interdisciplinary creation. Yet
it should also be clear that Sommerfelds was a selective borrowing. Although he col-
laborated on two texts with the mathematician Felix Klein, one on number theory
and the other on gyroscopes, it was only the latter that became a resource for disserta-
tion topics in Munich. Although he published papers during his Aachen period on
30 Chapter 1

both railroad brakes and lubrication (in fact related topics), it was the hydrodynamic
theory utilized in the latter that passed into his lectures on theoretical physics. Along
with this selectivity came a process of modification. If the (traditionally mathematical)
problem of the diffusion of heat was to become a topic in theoretical physics, then it
was to be altered to fit into its new surrounding. If technical hydraulics was to be part
of the Munich curriculum then this would be in combination with more overtly theo-
retical considerationsthe analysis of wear on ball bearings would be replaced by the
still practical but less entirely industrial problem of turbulence in open channels.
Sommerfelds papers and lectures dealt, not with general principles, but with specific
problems. Mathematical resources and theoretical underpinning were marshaled to
get to a point where these problems could be dealt with by Sommerfeld in his research
and by his students in their dissertations. The aim was not, that is, to create a unified,
general physics and a community that corresponded to it, but to studyand to create
students who would studyparticular problems, drawing topics from a wide range of
sources. The wideness of this range contrasts with historians concentrations on what
have come to be called the twin pillars of modern theoretical physicsthe quantum
and relativity theories. Yet Sommerfeld studied these subjects too. With the nature of
theoretical physics still under construction, it was not merely the outside disciplines
that were selected from and modified in the building up of the Sommerfeld style of
theoretical physics. Sommerfelds approach to the third strandphysicswas also one
of partial inclusion.
By 1911, the year he presented a paper on the quantum of action at the Solvay
Conference, Sommerfeld vocally espoused the necessity of some form of a quantum
hypothesis. In his earlier lectures, however, his reservations concerning the validity
of Plancks position were far more apparent. While Sommerfelds resistance to the
theory has been noted before, usually in passing, there has been no detailed analysis
of why this should have been so. What his lectures make clear is that Sommerfelds
attachment to the electromagnetic worldview led him to favor the so-called Lorentz-
Jeans formula for black-body radiation, despite its known failure to accord with experi-
ment as well as Plancks did. This conclusion, as well as illustrating elements of the
Sommerfeld style, has deep implications for our understanding of the conversion
of several leading physicists to the quantum theory. In Black-Body Theory and the
Quantum Discontinuity, Thomas Kuhn claimed that a lecture that Lorentz gave in Rome
in 1908 marked a turning point in the history of the early quantum theory and led
to a growing acceptance of the idea of a quantum discontinuity. While I grant the
importance of the Rome lecture, it is argued here that the acceptance of discontinuity
followed what was, in fact, a more profound realization. In 1906, Sommerfeld still
assumed that electromagnetic theory was untroubled by the problems that plagued
mechanics. It was to be expected, he implied, that a mechanical description of the
electromagnetic ether should produce inconsistency in the form of the incorrect
The Physics of Problems 31

blackbody curve. Lorentzs lecture of 1908 was the first statement by one of its leading
proponents that the electromagnetic worldview must fail for the case of radiation.
What was at stake for a significant part of the theoretical physics community was, far
more critically than the question of discontinuity, the question of whether the elec-
tromagnetic worldview could incorporate Plancks results, or whether the universaliz-
ing dream of the electromagneticists had to be abandoned. As Sommerfelds Solvay
paper would show, it would be the latter view that prevailed.

Relativity and the Electromagnetic Worldview

In the winter of his first year at Munich, Sommerfeld began a series of lectures on
Maxwells Theory and Electron Theory, a topic described in a letter that December
to Lorentz as the burning questions of electrons.80 Sommerfeld introduced his stu-
dents almost immediately to the current problems plaguing the subject area at hand.
After a short historical overview of the topic, he noted:

Of course all this is only valid for the negative electron and the apparent mass bound to it. About
the positive electron and the matter apparently inseparably bound to it, we know nothing. Also
there are still serious difficulties to overcome regarding electro-optical phenomena, which should,
according to the electron theory, show the influence of the earths movement. Lorentz recently
said, in reply to my question as to how the electrons were doing: badly. The difficulties are
not overcome in the face of Kaufmanns newest experiments. Therefore Planck was also
pessimistic.81

The reference was to the experiments of Walter Kaufmann, who had attempted to
distinguish between the two most prominent electron theories at the time: that of
Max Abraham, a former student of Max Plancks, who assumed a rigid, spherical elec-
tron, and the so-called Lorentz-Einstein theory, which assumed a deformable elec-
tron.82 Planck, in his report on Kaufmanns results at the 78th meeting of the
Gesellschaft Deutscher Naturforscher und rzte (GDNA, Society of German scientists
and physicians) in September 1906, called the two possibilities respectively the
sphere and the relative theories and concluded that one still couldnt work out
which of them was right. Therefore, he wrote,

no option remains but to assume that some essential gap is still contained in the theoretical
interpretation of the measured values, which first has to be filled before the measurements can
be utilized for a definitive decision between the sphere theory and the relative theory. One could
think here of various possibilities, but I dont want to discuss these further, because to me the
physical foundations [Grundlagen] of each theory appear too uncertain.83

Sommerfeld commented on Plancks pessimism in the discussion that followed. A

strong supporter of Abrahams theory, Sommerfeld disliked in equal measure, as he
wrote in a letter to Lorentz, both the latters deformable electron and Einsteins
deformed time.84 The 38-year-old Sommerfeld was clearly one of the younger
32 Chapter 1

contributors to the conversation in Stuttgart, and his suggested explanation for the
difference in opinions caused some merriment:

Sommerfeld (Mnchen): I would not, for the time being, like to ally myself with the pessimistic
standpoint of Mr. Planck. In the extraordinary difficulties of measurement the deviations could
perhaps yet have their ground in unknown sources of error. In the question of principles formu-
lated by Mr. Planck, I would suspect that the men under forty years of age will prefer the elec-
trodynamic postulate, those over forty the mechanical-relativistic postulate. I give preference to
the electrodynamic. (laughter)85

In an article written in 1970, the historian Russell McCormmach explained Som-

merfelds hostility to the mechanical-relativistic postulate as deriving from the
younger physicists devotion to an electromagnetic view of nature.86 This worldview
encompassed three related positions: a distaste for, and mistrust of, mechanical model-
ing, especially as applied to microscopic phenomena; a belief that the only physical
realities were electromagnetic in nature; and a programmatic commitment toward a
concentration of effort on problems whose solution promised to secure a universal
physics based solely on electromagnetic laws and concepts. 87 This last notion of an
electromagnetic program is crucial. Many, if not most physicists at the turn of the
century made some use of electromagnetic concepts in their work. Yet, as Sommer-
felds comments make clear, even when working on the subject of the electron itself,
not all methods and approaches were equally palatable. Preference, from the propo-
nent of the electromagnetic worldview, would go to those theories that used only
electromagnetic properties (or those assumed electromagnetic in nature, such as the
electrons mass),88 eschewing mechanical concepts like deformability.89 It was thus not
merely the problems chosen, but also the modes of solution deemed acceptable, that
marked Sommerfeld out from his older colleagues.
In common with many physicists of the generation that completed their university
studies in the last two decades of the nineteenth century, Sommerfeld had seen what
was portrayed as the gradual failure of the mechanical worldview, which held that
all physical phenomena could be explained in terms of the equations and concepts
of mechanics. His generation had also witnessed both Hertzs discovery of electro-
magnetic waves and the later successes of Lorentzs electron theorycrowned with
the discovery of the electron in the last years of the century. In 1901, Wilhelm Wien
(37 years old at the time) discussed the possibility of reducing all of mechanics to
electromagnetic theory. A return to mechanical explanation thus seemed to be a
throwback. For the younger physicists, wrote McCormmach, the electromag-
netic concepts clearly pointed to the future of physics.90
One should not, however, confuse support of the electromagnetic program with a
form of theoretical determinism. Many of the problems studied in Munich can cer-
tainly be understood as part of the furtherance of the program of the electromagnetic
The Physics of Problems 33

worldview, understood not merely as support for microphysical studies, such as those
of the electron theory, but also macrophysical ones, like those of wireless telegraphy.
Nonetheless, the aim of Sommerfelds lectures does not seem to have been to establish
the electromagnetic worldview as the sole philosophy of the Munich school, a hege-
mony like that supposedly put forward by Niels Bohr in Copenhagen in the 1920s
with respect to the correspondence and complementarity principles.91 This would have
been almost impossible in any case, for as a program, the electromagnetic worldview
was as much a promise for the future as a claim about the present. As Sommerfeld
noted in his lectures: Certainly, the electromagn[etic] foundation of mechanics is the
music of the future [Zukunftsmusik]. But I am convinced that matters will proceed here
just as with the music that received the name Zukunftsmusik thirty years ago.92
The worldview functioned as one standpoint from which Sommerfeld and his stu-
dents could understand and critique the work of others, and the lectures are not (even
where it would have been possible) written entirely from a single perspective. The
electromagnetic worldview provided the glasses through which Sommerfeld exam-
ined the world at this time, but one does not get the sense that he insisted that
everyone else use the same prescription. He was selective, but neither exclusive nor
dogmatic. With this proviso regarding the flexibility of the electromagnetic program
in mindat least in Sommerfelds handsthe next section explores the program in
action, for it was not merely the relativity theory that was viewed through the lenses
of the electromagnetic worldview. The quantum theorymore particularly, the theory
of black-body radiationwas judged according to its fit with the requirements of the
electromagnetic program as well.

Teaching Plancks Lectures

In 1906 Planck published his now-famous book Vorlesungen ber die Theorie der
Wrmestrahlung, both a summary of his earlier work and a continuation and re-
examination of it.93 Sommerfeld appears to have gone over the text with a fine-toothed
comb, adopting both Plancks summary of previous approaches to the problem of
radiation and, for the most part, his units.94 After beginning with the rapidly written
claim that radiation is a focus of modern research, Sommerfeld divided previous
approaches, much as Planck had, into three types: thermodynamic, electrodynamic,
and statistical methods.95 In the thermodynamic category he placed the work of Kirch-
hoff, Stefan and Boltzmann, and Wien; in the electrodynamic, that of Helmholtz,
Maxwell, Rayleigh and Jeans, and Lorentz; whereas the statistical principally dealt with
the methods of Boltzmann, Gibbs, and Planck. An outline of the structure of Som-
merfelds lectures is reproduced here as table 1.96 The column on the left represents
the proposed structure of the course, as laid out in Sommerfelds first lecture. The
column on the right provides the topic headings for the course actually delivered. The
last three sections of the proposed course were compressed into a single one.
34 Chapter 1

The difference between Sommerfelds discussion of previous treatments and his

analysis of Plancks own contribution jumps to the eye. Whereas his summary of
earlier research (sections 18) was in some cases as detailed as Plancks, his discussion
of Plancks theory in section 9 is remarkable concise. Sommerfeld achieved this by
excising almost entirely the discussion of the production of radiation by Hertzian
resonators, a topic that took up nearly one-third of Plancks text. Instead, within half
a page of expressing the energy of a Hertzian dipole in terms of its total energy U, the
electromagnetic moment f, and constants K and L,

U = K 2 + 12 L 2, (1)

Sommerfeld merely restated the relation Planck derived between the total energy of a
resonator and its average energy u at frequency :

c3
U= u. (2)
8 2
The Physics of Problems 35

A parenthetical note after the equation promised that a proof would follow, perhaps
as an exercise, since no such proof appeared in the lecture notes themselves. Following
Planck closely from this point on, Sommerfeld eventually arrived at Plancks equation,
relating energy to frequency for a black body at a particular temperature T:
h
U = . (3)
h / kT
(e 1)
Having obtained Plancks formula, Sommerfeld immediately launched into a section
labeled critical remarks. Sommerfeld appears to have paid close attention to com-
ments made by Paul Ehrenfest in the Physikalische Zeitschrift, the year Plancks book
appeared.97 Planck had introduced the resonators into radiation theory in part in order
to obtain a parallel to the Maxwell-Boltzmann distribution in kinetic theory. Just as
interaction between molecules brought about the Maxwell-Boltzmann distribution as
an equilibrium distribution of velocities, the interaction of resonators was supposed
to ensure that an initially arbitrary distribution of energies in a black body would
result in an equilibrated radiation. Ehrenfest quashed that possibility by showing that
the resonators could not do what was required. Since they emitted and absorbed
energy at characteristic frequencies, only resonators at the same frequency interacted,
producing an equilibrium distribution of intensity and polarization for each color.
For resonators at different frequencies, however, no interaction was possible, so any
arbitrary frequency distribution would persist. Ehrenfest wrote:
1) The frequency distribution of the radiation introduced into the model [described
by Planck] will not be influenced by the presence of arbitrarily many Planck resona-
tors, but will be permanently preserved.
2) A stationary radiation state will [nevertheless] result from emission and absorp-
tion by the oscillators in that the intensity and polarization of all rays of each color
will be simultaneously equilibrated in magnitude and direction.

In short: radiation enclosed in Plancks model may in the course of time become arbitrarily dis-
ordered, but it certainly does not become blacker.For the discussion to come the following
formulation is especially suitable: Resonators within the reflecting cavity produce the same effect
as an empty reflecting cavity with a single diffusely reflecting spot on its wall.98

Planck had made similar remarks at the end of his lectures, realizing, in his books
conclusion, that much of his analysis to this end had been fruitless.99
It is clear that Sommerfeld drew his inspiration from Ehrenfests critique. His first
objection, under the heading The Role of the Resonators, reads:

The resonators only operate like a Reagent, strips of blotting paper, not like a catalyst [Ferment],
coal dust. The non-black radiation remains non-black. The resonators can only increase the dis-
order of directions, not the color distribution. Because the resonator only works in the region
(, d) to which it is allotted [abgestimmt]. The resonator does nothing more than a diffusing
mirror. (Cf. 6 Jeans).100
36 Chapter 1

Another comment referred to the dissimilarity between the methods of Boltzmann

and Planck. Whereas Boltzmann had proven that the entropy, S, was a maximum for
the Maxwell-Boltzmann distribution (that is to say, that the equilibrium distribution
was the most probable one), Planck had skipped this step. Sommerfeld noted, appar-
ently again following Ehrenfest, that the substitution for this unfortunately missing
consideration was the auxiliary assumption [Hilfsannahme] that we now know as
Plancks hypothesis, = h. It was only with this hypothesis that Planck was able to
get to a result that provided the requisite dependence of the total energy on both
temperature and frequency.101
Yet, while Ehrenfest was prepared to take the close accordance of Plancks formula
with experimental data as proof that there was some validity to his analysis, Sommer-
feld was less enthusiastic. In fact, although Ehrenfest rejected the resonator approach,
he did not reject the recourse to combinatorics. Rather, he explained the fundamentally
different assumptions that led to the different results of Boltzmann (his former teacher)
and Planck. For Ehrenfest, Plancks hypothesis was an additional (if peculiar) constraint
that led to an experimentally verifiable result. Ehrenfest was willing to accept a version
of Plancks thermodynamical/statistical approach as long as the appeal to resonators
was abandoned. For Sommerfeld, on the other hand, the failure of Plancks resonators
seems to have appeared emblematic of the problem with Plancks method in general,
and Sommerfeld treated the auxiliary assumption as little more than something
that allowed Planck to get to the desired result. I think it is very possible, he wrote
in the lecture, that Plancks formula is only a good approximation.102
As an approximation, Plancks equation was not alone. Sommerfeld described the
result of the Englishman James Jeans as an approximation as well. Jeans had assumed
that energy could be distributed equally among the eigenvalues of vibrations within
a cube of side length L. Doing so, however, resulted in a curve that was not in accor-
dance with the experimental data of researchers like Plancks friend at the Berlin
technische Hochschule, Heinrich Rubens.103 Sommerfeld explicitly compared the
assumptions implicit in Jeanss derivation to those of Planck in section 6 of his lectures
(the section to which he pointed at the end of his first critical remark):

The most interesting question is now this: Why do we only obtain an approximate
formula?

1. The assumption of the equipartition of energy is not generally valid for the
Aether, it is derived mechanically. It is, so to speak, [mere] chance that it is still valid
for long waves. Long thereby means nothing: Size depends on L, L drops out.
2. Standp. of Planck. The quantity h is the quantum of action of energy. The energy
can not be divided arbitrarily. If the smallest amount of energy were h = 0, then
Plancks formula would also reduce to that of Jeans.104
The Physics of Problems 37

And yet, although both approaches were seen as only partially successful, and
although Plancks formula seemed to fit the data better, Sommerfeld did not accept
Plancks derivation. In deciding which theory to reject, the impotence of Plancks
resonators outweighed the failure of the Rayleigh-Jeans equation to match available
experimental results. In effect, Sommerfeld bracketed off the question of which
expression was more correct in terms of its relation to experimental data. The
choice between Plancks and Jeanss formulas was, rather, reframed as a choice
between two distinct methods. In fact, Jeanss result, as Jeans himself had derived
it, did not receive Sommerfelds support. The italicized lines in the quotation
above suggest that Sommerfeld was skeptical of the very basis of Jeanss derivation:
the Englishman had assigned a mechanical property (the equipartition of energy)
to what wasfor a proponent of the electromagnetic worldviewa fundamentally
non-mechanical ether. Section 7 of Sommerfelds lectures, however, was titled
Lorentzs derivation of the same [i.e. Jeanss] limit formula from the electron
theory. Methodologically , then, Jeanss expression was (or could be shown to
be) a result following from the electron theory, and that spoke strongly in
its favor.
Lorentzs derivation thus provided, in Sommerfelds eyes, a positive endorsement
of Jeanss formula. On the other hand, Sommerfeld saw significant difficulties in
accepting Plancks approach to the theory of radiation. He laid these out in a series
of General Comments toward the beginning of the lectures, offering a critique of
the three possible methods for approaching the problem of radiation. Of the first,
thermodynamics, he noted that it was at once the most secure but the least satisfy-
ing possibility. In opposition to Energetics, he wrote, one demands an understand-
ing of Mechanism or Electrodynamism. The kinetic theory, Sommerfeld claimed, had
eliminated thermodynamics by explaining its laws in terms of statistical mechanics.
Along similar lines, The program offered by Planck of radiation th[eory] should offer:
to explain thermod[ynamics] electro-statistically.105
Planck, however, while utilizing the statistical techniques of the kinetic theory, had
come out firmly on the side of thermodynamics in his analysis of heat radiation. Dis-
cussing the calculation of radiation intensity in his lectures, he noted that it was in
no way determined, so that in a case where according to the laws of thermodynam-
ics and according to all experience a single valued result is to be expected, pure elec-
trodynamics leaves [one] completely in the lurch, and one in fact ends up with
infinitely many solutions. Mechanics served no better: The temporal course of a
thermodynamic process cannot be calculated on the mechanical heat theory or the
electrodynamic theory of heat radiation under the [same] initial and boundary condi-
tions that completely suffice in thermodynamics for the single-valued determination
of the process.106
38 Chapter 1

For Sommerfeld, the fact that Planck did not seek to explain radiation solely in
electro-statistical terms spoke against his methods: Plancks theory is therefore not
ideal; the theories of Jeans and Lorentz are better in principle. 107 Here, then, was the
programmatic aim of the electromagnetic view of nature in operationprogrammatic
because, as noted earlier, Sommerfeld had specific objections to Lorentzs particular
version of the electron theory, preferring Abrahams. Nonetheless, he clearly deemed
either better than one that did not seek to reduce all other explanatory means to
electrodynamics. Jeanss result, as derived through the electron theory, was to be pre-
ferred over any result following from a system of thought that might seek to deny the
unificatory capacities of electromagnetism. No doubt, like Lorentz himself, Sommer-
feld hoped that a more complete electromagnetic theory would result in an expression
in better accordance with experience and experiment. Until then, however, an
approximation derived along correct programmatic lines trumped one derived in a
manner deemed not ideal.
The continuation of Sommerfelds General Remarks shows him waxing lyrical
over the total explanatory possibilities of electrodynamics, which creates here as well
the highest unity:

Heat (radiated) is light, therefore electr[icity?]; but heat is, on the other hand, molecular motion.
How it should It must convert electr[ical] action into inertial action; as it does so, the theory shows
the apparent degree to which kinetic energy actually should be electromagn[etic] energy of the
charged matter. Therefore in short: From the ident[ity] of light Leslie Prevost Rumford 18th Cent. and heat,
the id[entity] of light and electr[icity] Maxwell Hertz end of the 19th Cent. and the id[entity] of heat and
molecular mechanics Clausius Maxwell Boltzmann 19th Cent follows necessarily the id[entity] of molecular
mech[anics] and electrodynamics (20th Century).108

If Boltzmann had shown that thermodynamics reduced to mechanics, this last

identity showed that both thermodynamics and mechanics could be reduced to elec-
trodynamics. This conclusion, in turn, suggests a pointwise refutation of Plancks
introductory theses on the basis of the electrodynamic worldview. The responsive
sentences below indicate Sommerfelds position:
1) Heat diffuses [fortpflanzt sich] in two different ways, conduction and radiation.
1a) Heat diffuses in only one way, electrod[ynamic], in conduction the electr[ic]
fields of the charges are bound to the molecule, in radiation they spread out freely
in the Aether.
2) Heat rad[iation] is much more compl[icated] than heat conduction, because in
that case the state cannot be characterized through a vector.
2a) Heat rad[iation] is much easier than heat conduction, because the particularities
of the charge distribution (matter) dont play a part. In the Aether only the direction
and intensity of the radiation, in heat conduction the directions of movement of
the molecule as well.109
The Physics of Problems 39

One can read these responses as the principle (and the in principle) reasons that
Sommerfeld approached Wrmestrahlung with skepticism. Although Plancks approach
allowed a simpler calculation of certain fundamental constants (and here Sommerfeld
was thinking much more of k, Boltzmanns constant, than h), it went against the
worldview that Sommerfeld had adopted. Certainly, Plancks resonator approach had
its own, intrinsic difficulties, but, more generally, it suffered from its adherence to a
viewpoint that Sommerfeld and others were seeking to supersede with the physics of
the future. As with his response to relativity theory, Sommerfeld considered quantum
theory a step backward, presumably also the domain of men over 40, not the young-
bloods in whose camp he placed himself.

Black Bodies in an Electromagnetic World

Kuhns argument that Lorentzs lecture in Rome in 1908 marked the beginning of the
acceptance of the quantum discontinuity runs as follows:

During 1908 Lorentz produced a new and especially convincing derivation of the Rayleigh-Jeans
law. Shortly thereafter he was persuaded that his results required him embracing Plancks theory,
including discontinuity or some equivalent departure from tradition. Wien and Planck quickly
adopted similar positions, the former probably and the latter surely under Lorentzs influence.
By 1910 even Jeanss position on the subject had been shaken, and he publicly prepared the way
for retreat. These are the central events through which the energy quantum and discontinuity
came to challenge the physics profession.110

In the Rome paper Lorentz proved that the electron theory must lead to Jeanss
result. That is to say, there could no longer be any suggestion of his electromagnetic
approach avoiding the problems that followed from the equipartition theorem. Lorentz
stated that such had been his hope, after reading Jeanss papers. Now that hope was
officially dashed.111 Without at this point making a choice between them, Lorentz
then stated the difference between the Rayleigh-Jeans and the Planck case as baldly
as possible. Accepting Planck would bring theory in line with experiment, but we
can adopt it only by altering profoundly our fundamental conceptions of electromag-
netic phenomena. Accepting Jeans on the other hand, would oblige us to attribute
to chance the presently inexplicable agreement between observation and the laws of
Boltzmann and Wien.112 For experimentalists, the issue was now clear: Jeanss equa-
tion did not work at all. If the choice was between it and Plancks, then the latter had
to be accepted. In a paper published a few months after the Rome lecture, Lorentz
acknowledged that he had been convinced in the interim by the arguments of experi-
mentalists (including Wilhelm Wien, Otto Lummer, and Ernst Pringsheim) and had
abandoned any support for Jeanss equation.113 For the final step, Kuhn claimed,
Lorentzs great personal authority was responsible, to a great extent, for spreading
the gospel to the rest of the physics community.114
40 Chapter 1

But exactly what gospel was being spread? Participants in the discussion referred,
variously, to the Rayleigh-Jeans, the Jeans, and the Jeans-Lorentz formula. While
the two former do not necessarily carry with them the association of Jeanss result
with the electron theory, the latter definitely does. Kuhns inconsistent attention to
this fact elides the difference.115 Proponents of the electromagnetic worldview (includ-
ing Sommerfeld, Wien, and to a lesser extent Lorentz) may not have regarded the
choice between continuity and discontinuity as the central issue. Rather, the question
that came to challenge them, the question over which they struggled, was whether
the electron theory could produce a Planck-like formula. Once it was accepted that
this was impossible, discontinuity was adopted quite readily by this group.
Lorentz wrote to Wien early in June 1908, noting that he had been ceaselessly
racking his brains over the last few years over the question of deriving Plancks formula
(or something similar) from the electron theory. Contrasted to this language of con-
stant struggle, Lorentzs description of Plancks alternative solution, the introduction
of elementary quanta of energy, seems almost casual: In and of itself, I have nothing
against it; I concede at once that much speaks in its favor and that it is precisely with
such novel views that one makes progress. I would, therefore, be prepared to adopt
the hypothesis without reservation if I had not encountered a difficulty.116 Kuhn
highlighted this difficulty to explain Lorentzs hesitancy in accepting discontinuity,
but the problem Lorentz outlined was not that of discontinuity per se but rather that
of an asymmetry between the (continuous) absorption and emission of energy by
resonators in interaction with the ether, and discontinuous emission and absorption
otherwise. This specific question would continue to bother those who had accepted
the idea of a quantum discontinuity for some time, and would eventually lead Planck
to his so-called second and third theories, each of which posited different mechanisms
(one continuous, one discontinuous) for resonator emission and absorption. Lorentz
did not have a difficulty with discontinuity in and of itself. What counted was
whether the electromagnetic worldview could include it. I can only conclude, he
wrote in the Physikalische Zeitschrift, that a derivation of the radiation law from elec-
tron theory is scarcely possible without profound changes in its foundation. I must
therefore regard Plancks theory as the only tenable one.117
The first radical move for proponents of the electromagnetic worldview was not the
adoption of a new theoretical positionthat would come laterbut the forced aban-
donment of their old one. For Wien, it was not immediately obvious after the Rome
lecture that such an abandonment was even being posited, and his route toward
Plancks theory can be understood as the inverse of Lorentzs. If Lorentz tried to obtain
Plancks result by beginning with the electron theory, Wienafter dismissing Jeanss
result on experimental grounds early onbegan with Plancks energy elements and
then sought to understand them in electromagnetic terms. His original reaction to the
Rome lecture evinced a certain irritation with what he saw as Lorentzs rather poor
The Physics of Problems 41

rederivation of Jeanss result in Rome. The lecture which Lorentz gave in Rome, he
wrote to Sommerfeld:, has disappointed me greatly. That he presented nothing more
than the old Jeans theory without bringing in any sort of new viewpoint I find a little
poor. Besides, the question of whether one should regard the Jeans theory as discussable
lies in the region of experiment. His opinion is not discussable here because observa-
tions show enormous deviations from the Jeans formula in a range in which one can
easily control how far the radiation source deviates from a black body. Whats the point
in presenting these questions to the mathematicians, who can make no judgment on
precisely this point? It seems, in addition, a little peculiar to seek the advantage of the
Jeans formula, in spite of the fact that it corresponds with nothing, in the fact that it
can preserve the whole unlimited multiplicity of electron oscillations. And the spectral
lines? Lorentz has not shown himself to be a leader of science this time.118
Ditching Jeanss result on experimental grounds, however, was relatively unprob-
lematic compared with doing so for methodological reasons. It was not until he read
Lorentzs second paper that Wien realized, with some dismay, what giving up Jeanss
result meant in relation to electromagnetic theory. He wrote to Sommerfeld: Lorentz
has recognized his error over radiation theory and that Jeanss hypothesis is untenable.
Now, however, the situation is not so simple, since in fact it appears as if Maxwells
theory must be abandoned for the atom. Hence I have a problem to pose you again.
Namely, to check how far Lorentzs statistical mechanics and proof is founded on the
fact that a system obeying Maxwells equations (including electron theory) must also
obey the supposition of the equipartition of energy, from which Jeanss law is
deduced. Namely a restriction of the degrees of freedom, as required by Plancks energy
element, must also require an electromagnetic interpretation. Now it seems to me
almost as if such [an interpretation] would be impossible, as if precisely this restriction
requires additional forces (fixed connections and the like) that dont fit in with a
Maxwellian system. If thats really the case, one doesnt need to rack ones brains any
more about an interpretation of the energy element and a representation of spectral
series on an electromagnetic basis, but rather must seek to find an extension of
Maxwells equations within the atom.119
Standing almost as bookends, outlining first the problem and then the proposed
solution, are the statements it appears as if Maxwells theory must be abandoned for
the atom and we rather must seek to find an extension of Maxwells equations
within the atom. Between the two is an interpretation of both Jeanss and Plancks
derivations in electromagnetic terms. That is, Wien translated the question of equi-
partition and the question of the meaning of Plancks energy elements into the lan-
guage of electromagnetic theory. The contradictions that arose in so doing led him to
both echo and reject a comment written to him less than two weeks earlier by Lorentz:
One doesnt need to rack ones brains any more. The effort to save the electron
theory and the electromagnetic worldview in its entirety now seemed fruitless, and
42 Chapter 1

Wien pointed quite calmly to the need for an intra-atomic extension of Maxwells
equations.
Sommerfelds reply, dated 20 June, was less pessimistic. He claimed that he did not
find Lorentzs electrostatistical derivation of Jeanss result conclusive.120 He promised
Wien that he would communicate his objections to Lorentz, and indeed he did so the
same day. Rather than accept Lorentzs calculations as a proof that the electromagnetic
worldview must fail in the face of Plancks result, Sommerfeld merely used the oppor-
tunity to emphasize what was at stake in such a question. At one time, he wrote,
when I lectured on the theory of radiation, I believed Jeanss paradox could be
overcome by saying that electrodynamics is not subject to mechanical laws. Your
present remarks seem to me to be an excellent foundation for the resolution of this
question.121
Fixing a date for Sommerfelds acceptance of the necessity of discontinuity is not
easy.122 In November 1908 he wrote to Lorentz urging him to ignore his earlier criti-
cisms, but did not explicitly retract his objections to Lorentzs theory in general.123 It
was, however, in the latter part of 1908 that Sommerfeld attended Minkowskis lec-
tures on relativity and was converted by them.124 This is critical, since Abrahams
rigid spherical electron theory, which Sommerfeld had originally favored over Lorentzs,
was not relativistically invariant. If Sommerfeld applied the relativity theory consis-
tently to the choice between competing electron theories, that is, he would have been
induced to accept Lorentzs some time after 1908. By late 1909, Sommerfeld would
make this point explicitly, in lectures that mark the first classes taught anywhere in
the world on relativity theory. In introductory comments, Sommerfeld noted that the
hypothesis of the rigid electron was dropped because it includes the hypothesis of
absolute space and that the deformable electron follows from the concept of relative
space-time, which experience demands.125 This, in conjunction with the removal of
his specific reservations about Lorentzs derivation, would imply that Sommerfeld
accepted Lorentzs conclusion that the electron theory and the electromagnetic world-
view were incapable of dealing alone with the theory of radiation.
Three quite different responses to the questions raised by Plancks black-body
theorem are sketched above. Yet it should also be clear that Lorentz, Sommerfeld,
and Wien held much in common.126 All three men conceived the problem of radia-
tion as one to be cast at first solely in electromagnetic terms. If, after repeated
efforts, that should prove impossible, the answer was not to abandon electrodynam-
ics in favor of some other extant approach, but to find a new way of extending
it. That is, electrodynamics provided the only standpoint from which one could
begin to construct the future steps required to come to a comprehension of the
puzzles introduced by black-body theory. And the question at hand was not the
problem of the quantumsuch a problem did not yet exist in such terms for
the majority of physicists. For electromagneticists, Plancks result was a problem for
The Physics of Problems 43

and of the electromagnetic worldview in general and Lorentzs electron theory in

particular. Only after they had acknowledged the reality and insurmountability of
the problem within present electromagnetic theoryafter June 1908did they focus
on discontinuity.
On the other hand, for those not committed to the electromagnetic worldview, the
issue of discontinuity was an important means of understanding Plancks result. Ein-
stein and Ehrenfest, who approached the issue from the perspective of Boltzmannian
statistical mechanics, were the first, Kuhn argues, to discover the quantum discon-
tinuity, some years before the Rome lecture. Jeans, on the other hand, initially denied
the force of experimentalists arguments, not conceding their validity until 1910. His
description of the choice on offer at the time does not include discussion of electron
theory, but does place the issue of discontinuityexpressed in terms of differential
equationsfront and center:

Plancks treatment of the radiation problem, introducing as it does the conception of an indivis-
ible atom of energy, and consequent discontinuity of motion, has led to the consideration of
types of physical processes which were until recently unthought of, and are to many still unthink-
able. The theory put forward by Planck would probably become acceptable to many if it could
be stated physically in terms of continuous motion, or mathematically in terms of differential
equations.127

For proponents of the electromagnetic worldview, the most important issue intro-
duced by black-body theory was the apparent failure of electron theory to incorporate
or duplicate Plancks more experimentally verified result. The acceptance of disconti-
nuity followed with comparatively little struggle after that blow to their shared world-
view had been assimilated. For those who were not wedded to the electromagnetic
picture, however, discontinuity became the most troubling thing about Plancks
energy elements. Thus, perhaps one should, if one is to adapt Kuhns religious lan-
guage, speak not only of converts to discontinuity, but also of lapsed or at least
disillusioned electromagneticists.

Conclusion

All those who have written on Arnold Sommerfeld in any detail have noted the
number and eclecticism of both the problems he studied and the methods of their
solution. This emphasis on specific questions and their specific solutions, the search
for a mechanism or a process rather than a generalizing postulate is what distinguishes
Sommerfelds physics of problems from Plancks physics of principles. Thermody-
namics, which provided the model for Plancks unifying methodology, was to Som-
merfeld the most secure, but the least satisfying approach to physics, for it failed to
provide the specificities of mechanism. Historians have, perhaps naturally, tended to
44 Chapter 1

fragment Sommerfelds various projects, attributing some to theoretical physics, others

to mathematics or technical mechanics. Doing so is, in some ways, an obvious way
of understanding a physics of problems, for the specificity of problem solving can
suggest a lack a coherence, an inability to be unified. Heretofore the discussion in this
work, has also consideredseparatelythe three elements that went into making up
theoretical physics in Munich: mathematics, technical mechanics, and physics. It
remains to be considered how these three elements formed a recognizable single style.
What, to phrase the question in its starkest form, was Sommerfelds theoretical physics
other than a single name given to a collection of disparate interests?
Perhaps not surprisingly, the problems themselves provide the answer. The prob-
lems themselves would often cross and hence blur the disciplinary boundaries that
composed theoretical physics in Munich, producing what Andrew Warwick, in his
discussion of mathematics training in nineteenth-century Cambridge, has termed a
technical unity.128 The problems accorded with Sommerfelds physical worldview
and thus dealt on a majority of occasionsin the early yearswith electromagnetic
theory. At the same time, they were genuine problems of current technological inter-
est, solved with mathematical prowess turned to physical ends. It was this quality of
interdisciplinary fusion within the problems studied in the Sommerfeld School that
brought a commonality of approach.
One may clearly discern this emphasis on both interdisciplinarity and technical
unity in the selection and solution of problems in Sommerfelds reports to the Munich
Philosophical Faculty on his students dissertation and habilitation projects. As noted
earlier, most of these topics flowed from his own research, and the range of titles
provides a good insight into the problems that Sommerfeld deemed significant. In
addition, in the short commentaries describing the work, he would pick out those
elements he deemed most important, so that even within the context of a given
problem, one can discern those aspects representative of Sommerfelds own
interests.
The extent of Sommerfelds pursuit of the electromagnetic view of nature appears
in the number of his students projects that deal with problems related to electromag-
netic theory. Of the ten theses supervised or co-supervised by Sommerfeld in Munich
between 1908 and 1911, eight discussed some aspect of electromagnetism, such as
wireless telegraphy, electrical conduction in gases, measurements of capacitance, or
the calculation of light pressure on spheres of arbitrary material. In many of the theses
the question was not merely one of sheer theoretical analysis, but was derived from
a practical problem. Hence, the project conducted by Hermann von Hoerschelmann
took up the (very topical) question of the Mode of Operation of the Bent Marconi
Sender in Wireless Telegraphy. The problem, which Sommerfeld called rather mys-
terious, lay in the discrepancy between the theoretical and the actual operation of a
Marconi radio station. A bent sender should provide a signal in a preferred direction,
The Physics of Problems 45

that given by the antenna wire. Some of those who used such senders, however, had
failed to detect such a preferred direction. Sommerfelds discussion reveals his close
interaction with those at the forefront of technological use and production, as he cites
Count Arco (one of the doyens of German telegraphy) and an unnamed marine
officer:

Even though this [theoretical] effect is called into doubt by several practical menCount Arco
told me that in Marconis opinion the antenna only conducted horizontally in this way because
otherwise he couldnt accommodate the large length of the wire, and a marine officer in the
radio commandos [Funkencommandos] wrote to me that he had experienced no directional effect
in the vicinity of such a Marconi stationnonetheless Marconis data has, given all previous
experiences, the greatest right to attention. Therefore, because Marconi at his distance station
[Fernstation] now uses the bent sender throughout and has invested significant capital in it, a
clarification of its mode of operation is an important theoretical task.129

Rather than merely a question of electromagnetic theory, it is the practical issue of

the operation of an existing radio stationan issue considered by those who made
use of such stationsthat provides the impetus for a theoretical investigation. Armin
Hermann has noted that Sommerfeld would often pursue physical questions that he
examined up to their technical application. Here the situation is reversed, as Praxis
provides the problem for Theorie. In a similar fashion, commenting on a project that
dealt with the spreading out of wireless telegraphic waves on the Earths surface, Som-
merfeld effectively chided previous, more mathematically inclined researchers (Poin-
car and Nicholson) for their failure to explain clearly the success of practitioners.
Wireless telegraphers had succeeded in overcoming the problem of the curvature of
the Earth in their attempts to send long-distance signals, and the project of Hermann
March was devoted to explaining this practical success theoretically.130 At the same
time, this problem-focused fusion of practical technology and electromagnetic theory
required the development of a sophisticated mathematical apparatus, and Sommerfeld
lauded his students work in developing the means of representing the electromagnetic
fields in terms of the integrals of spherical functions. This, he claimed, was important
for several problems of mathematical physics and also appears noteworthy to me from
a pure mathematical standpoint.131 The one problem, in other words, would fuse all
three elements of the Sommerfeld Style.
Specific problems could, on occasion, not merely require all three aspects of Som-
merfelds theoretical physics, they could recur in areas that corresponded to different
disciplinary contexts. In Hopfs project, for example, Sommerfeld noted the similarity
of one part of the solution with another well-known phenomenon. For ship waves
in water of finite depth, the angle subtended by the wake is a constant, a result
similar to that arrived at by Ernst Mach through his studies of the shock waves that
were produced by an object moving at supersonic speeds (e.g., a bullet) that under-
went a rapid deceleration (by hitting a wall, for example). Both of these, of course,
46 Chapter 1

were problems in mechanics (with technical applications), but Sommerfeld, follow-

ing Stokes, Lenard, and Wiechert, had already used such a model in the case of X-ray
production through the braking radiation of an electron. According to this theory,
Paul Ewald wrote, X-rays are the electrical analogue to the sound cracks which
travel forth in air from a target hit by shot.132 Yet later the model would be used
as Sommerfelds point of entry into studies on quantum theory.
The problems that characterized Sommerfelds theoretical physics were thus among
the agents that provided a form of unity for his eclecticism, both through the fact
that multiple elements were mobilized toward their solution and through the recur-
rence of particular problems and modes of modeling and solution in different disci-
plinary contexts. Technology, mathematics, and physics were planted together in
Sommerfelds nursery for theoretical physics. In spite of the diversity of its subject
matter, the eclectic physics that emerged there shared common roots, and grew to
bind its tendrils together.