Sie sind auf Seite 1von 52

Eur. Phys. J.

H
DOI: 10.1140/epjh/e2014-50026-9

THE EUROPEAN
PHYSICAL JOURNAL H

The story of the Higgs boson: the origin


of mass in early particle physics
Arianna Borrellia
Technische Universit
at Berlin, Institute for Philosophy and History of Literature, Science
and Technology, H 72 Strae des 17. Juni 135, 10623 Berlin, Gemany
Received 22 April 2014 / Received in nal form 27 October 2014
Published online 9 January 2015
c EDP Sciences, Springer-Verlag 2015


1 Introduction
On July 4th 2012 scientists at CERN announced that the ATLAS and CMS collaborations working at the Large Hadron Collider had observed a particle whose
characteristics were compatible with those of the long-sought Higgs boson, the only
elementary particle of the Standard Model on whose existence doubts still lingered
(CERN press release Jul 4th 2012). More comprehensive data analyses later conrmed
that the particle discovered fully matched the prole of the Higgs boson: a scalar (i.e.
spin 0) particle whose interactions t the prescriptions of the Standard Model (CERN
press release Mar 14th 2013). The scientic relevance of this result resides not only in
the further conrmation of the validity of the Standard Model, but also, and indeed
especially, in the fact that the Higgs boson has a special status among elementary
particles, a status often expressed colloquially by saying that it gives mass to all
others particles.
How did the Higgs boson attain such a special status? Today, the connection
between the Higgs particle and the origin of mass passes through a mathematical
construct usually referred to as the Higgs mechanism1 . As we shall discuss in detail
later on, this mechanism is today characterized as the spontaneous breakdown of a
local gauge symmetry and is seen as having emerged in the early 1960s from the
combined eorts of a broad network of theorists (Karaca 2013). The perception of
this mathematical structure as a key to understanding the origin of mass has dominated particle research since decades, as expressed in the motivations for awarding research prizes to some of its many alleged creators2: the 2004 Wolf prize in physics for
a

e-mail: ari@weatherglass.de
The term Higgs mechanism is used today in textbooks of particle physics discussing
the subject (e.g. Peskin and Schroeder 1995: 690716). The papers most frequently quoted as
original references in this context are (Englert and Brout 1964; Guralnik, Hagen and Kibble
1964; Higgs 1964a, 1964b).
2
As we shall see more in detail in the course of the paper, beside the six authors whose
names are usually associated with the Higgs mechanism (i.e. Brout, Englert, Guralnik,
Hagen, Higgs and Kibble) other theorists have oered important contribution to its comingto-be, among them Philip Anderson, Murray Gell-Mann, Maurice Levy, Abdus Salam, Julian
Schwinger, Gerhard t Hooft and John Ward. Not all of them have raised a claim to be
1

The European Physical Journal H

pioneering work that has led to the insight of mass generation, whenever a local gauge
symmetry is realized asymmetrically in the world of sub-atomic particles (recipients:
Francois Englert, Peter Higgs, Robert Brout) (Wolf Prize in Physics 2004); the 2010
J.J. Sakurai Prize for Theoretical Particle Physics for elucidation of the properties
of spontaneous symmetry breaking in four-dimensional relativistic gauge theory and
of the mechanism for the consistent generation of vector boson masses (recipients:
Carl R. Hagen, Francois Englert, Gerald Guralnik, Peter Higgs, Robert Brout, Tony
Kibble) (Sakurai Prize 2010). Finally, the 2013 Nobel Prize for Physics for the theoretical discovery of a mechanism that contributes to our understanding of the origin
of mass of subatomic particles, and which recently was conrmed through the discovery of the predicted fundamental particles, by the ATLAS and CMS experiments at
CERNs Large Hadron Collider was awarded to Francois Englert and Peter Higgs
(Nobel Prize, 2013), and there is no doubt that Robert Brout would have shared it
with them, had he not already passed away in 2011. Despite the appearances, I shall
argue in this paper that the idea that the origin of mass might be explained by, or at
least connected to, a particle having characteristics similar to those of todays Higgs
boson had been around at least since the late 1950s. The earliest reection on the origin of mass within modern particle physics combined verbal statements with sketchy
mathematical arguments which would later be embedded in more complex schemes,
and the present contribution aims at reconstructing some of the key episodes in these
developments up to and including the emergence of spontaneous symmetry breaking
and of the Higgs mechanism. I shall present the reections of more and less prominent authors motivated by a mathematically and physically vague, but intellectually
stimulating conviction that the physical origin of mass could be linked to a particle
having the same quantum numbers as the vacuum, as Abdus Salam wrote in 1961
(Salam, 1961: 428).
In telling my story of the Higgs boson I have more than one aim. First of all, I
wish to set a key issue of todays high energy physics in its broader historical context.
In the second place I would like to show how, although mathematical structures are
indispensable for constructing physical theories, an equally fundamental component
of those theories are statements expressed in words. These verbal constructs are not
necessarily simplied translations of rigorous mathematical arguments, but can be
independent elements of theories, although scientists often tend to regard them as
representing mathematical structures which have not yet been spelled out in detail
(Borrelli 2012, 2014). Finally, the present study should provide an example highlighting the collective dimension of theoretical research in high-energy physics. Other than
experimental activity, theoretical research in particle physics is often seen as an individual eort whose results can be uniquely ascribed to single authors. However,
although this may be true at the level of a specic formula or a given numerical
prediction, when looking at large-scale processes of concept formation like the one
at issue in this case one has to take into account how authors creatively build upon
each others work, taking over and transforming approaches and ideas. Indeed, since
the late 1950s, due to the growing frequency of visiting scholarships and conferences,
the increasingly fast and broad circulation of preprints, and the tendency to build
up occasional, short-lived collaborations, theoretical research in particle physics was
taking more and more the form of a collective enterprise in which it makes no sense
to ask who had which ideas rst. It is therefore hardly surprising that priority disputes have become increasingly frequent, too, when some results which appeared relatively unimportant at the time of their rst formulation later came to be regarded as
ground-breaking developments. In fact, a paradigmatic example of this phenomenon
counted among the discoverers of the Higgs mechanism, but for each of them a plausible
argument could be made.

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

are the discussions which preceded and followed the Nobel prize awarded in 2013 for
the theoretical discovery of the mechanism of mass generation, and at the end of
this paper I shall make a short detour to oer an overview on how priority issues in
particle physics may become highly charged political matters.
In the next section I shall set the stage for the historical analysis by summarizing
those developments in particle physics of the 1940s and 50s which provided both the
basis and the motivation for reections on the origin of mass. In the following sections
I will sketch the work of some of the historical actors reecting on mass generation
and symmetry breaking between the late 1950s and the early 1960s. The overview
will include, but not be limited to, the authors usually quoted as having participated
to the construction of spontaneous symmetry breaking and of the Higgs mechanism.
I have tried to represent the aims and arguments of the historical actors without imposing on them an interpretation in terms of later notions, while at the same time
pointing out how some elements found in their reections were taken up and expanded
later on and a posteriori may seem to invite an interpretation in terms of the Higgs
mechanism. I will begin by discussing how, in the late 1950s, a number of authors
employed the properties of the quantum vacuum to argue that the observed variety of
particle masses and properties, which had been formalized in terms of approximate
symmetries, might be explained by means of some fundamental, symmetric model
containing only a few elements (Sects. 37). As we shall see, a scalar boson in many
ways similar to the Higgs one often played a role in those schemes. Section 8 will deal
with Salam and Wards attempt at unifying particle interactions using the principle of
local gauge symmetry with a scalar boson breaking the symmetry and giving mass to
gauge vector bosons. The turn of events initiated by Nambu with the development of
an analogy between high-energy physics and superconductivity is discussed in Sections
911. Section 12 presents a paper by Marshall Baker and Sheldon Glashow (1962) in
which the term spontaneous symmetry breakdown was introduced to characterize a third way between symmetry and non-symmetry, subsuming under this name
many of the schemes brought forward in the previous years. Section 13 oers a very
sketchy overview on the further development of physical and mathematical notions of
spontaneous symmetry breaking in the early 1960s, outlining the broader context
in which discussions on the breakdown of local gauge symmetries took place. Finally,
in Sections 14 and 15 I summarize the well known events of the early 1960s, when a
number of theorists tackled the question of whether and how it would be possible to
use a scalar eld with non-vanishing vacuum expectation value to spontaneously
break a local gauge invariance and generate gauge vector boson masses without at
the same time requiring the introduction of an undesirable massless scalar particle
known as the Goldstone boson. These works are today regarded as the theoretical
discovery of a mechanism of mass generation which is embedded in the WeinbergSalam model of electroweak interaction, whose formulation I discuss in Section 16.
That section concludes the historical overview on the emergence of theories of mass
generation, and the following chapter turns the attention to the discussions in the
physics community before and after the 2013 Physics Nobel prize was awarded to
Englert and Higgs, as an example of the clash between the wish to recognize individual scientic merit, the increasingly collective character of theoretical research and
the high political relevance of the Nobel prize. A few concluding remarks are found
in Section 18.

2 Particles, masses and symmetries in the early 1950s


The notion of mass is among the most complex of modern science, and no denition exsists which equally ts all theoretical and experimental elds making use

The European Physical Journal H

of the term. With the emergence of the theory of relativity in the early 20th century,
the equivalence of mass and energy was established, but in nonrelativistic quantum
physics rest mass continued to be regarded as an intrinsic property of elementary
particles. This view appeared at the time unproblematic, since electrons and protons,
the only known elementary particles, were stable, a fact that was mostly taken for
granted. Beside electrons and protons stood photons, the quanta of light, initially
regarded by most scientists as entities of a dierent kind than the particles composing matter, since they had neither mass nor a rest frame and could be created and
annihilated. Indeed, up to the early 1920s most scientists found the notion of photons
problematic (Hendry 1980). However, the 1930s brought experimental evidence for the
existence of new particles which were massive like electrons and protons, but could be
created and annihilated like photons: rst came the neutron (1932), then the and
the mesons (today muon and pion) (Brown and Hoddeson 1983; Brown et al. 1989).
Equally innovative was Paul Diracs relativistic theory for electrons (1928) which led
to postulate the existence of its antiparticle (positron) and to the idea that the two
would mutually annihilate into photons. The positrons existence was established experimentally in 1932 by Carl Anderson (Anderson 1952, 1953). In retrospect, these
events may appear (and are often presented as) experimental discoveries of new
particle, but the history of muon and pion shows how complex the process actually
was. In 1937 three independent experiments established that cosmic rays contained a
component whose characteristics were similar to those of the electron, except that it
had a mass intermediate between electron and proton (hence the name meson or
mesotron) (Neddermeyer and Anderson 1937, Street and Stevenson 1937, Nishina,
Takeuchi and Ichimiya 1937, Monaldi 2005). At rst it remained open whether the
mesotrons were one or more new particles, or possibly a single particle with variable mass, but in the late 1930s the mesotron was interpreted as the particle proposed
in 1935 by the Japanese theorist Hideki Yukawa (19071981) as mediator of nuclear
interactions (Yukawa 1935). This interpretation was a boost for the study both of
the meson and of nuclear interaction, but in the early 1940s it became clear that not
all observed properties of mesotrons matched those of the Yukawa meson, and in
1947 experimental evidence led to the conclusion that there were in fact two mesons
and that only one of them (the pion) could be identied with the Yukawa particle
(Conversi, Pancini and Piccioni 1947).
After the end of the Second World War investment in particle research led to
the discovery of an even greater variety of new phenomena, and physicist tentatively
made sense of them by assuming that they involved a further score of new particles
which not only had a broad range of masses, but also possessed a number of hitherto unknown intrinsic properties (internal quantum numbers), such as isospin and
strangeness (Borrelli 2014). The particles interacted through electromagnetic and nuclear forces, as well as by means of forces having intermediate or weak strength
and which did not conserve some or all internal quantum numbers. To the astonishment of many scientists in 1956 it was even established that some interactions are not
invariant under mirror transformation, a symmetry which had so far been regarded
as a fundamental characteristic of nature3 .
To make sense of the complex landscape, particles were tentatively ordered into
smaller groups (mutliplets) whose members shared some basic features (among
them usually mass), but diered in other aspects. An example were the three pions,
3

The hypothesis of parity non-conservation was formulated by Tsung Dao Lee and Chen
Ning Yang (1956). The rst experimental evidence was sought and found by the group led
by Chien-Shiung Wu (Wu et al. 1957) and soon further conrmed by other experiments
(Garwin, Ledermann and Weinreich 1957) and (Friedman and Telegdi 1957). For a survey
on the subject with further references see (Pais 1986: 523533).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

which have spin zero, almost identical masses, but three dierent electric charges: positive, negative and zero. While today this method of classifying particles may appear
almost immediately derivable from experience, its emergence was the result of a very
complex process of concept formation due to a close interplay of theory and experiment (Borrelli 2014). At rst, mass remained a main characteristic used to identify
particles and the terminology used today still bears traces of this early phase: beside
meson (intermediate mass), one had heavy baryons and light leptons, named
accodring to the Greek words for heavy and light. Today these terms have a meaning
which does not refer anymore to mass values, reecting a change of attitude that took
place in the course of the 1950s, when classifying particles primarily according to mass
appeared increasingly problematic. There were both practical and theoretical reasons
for this development. Until the early 1930s the mass distinction between light electrons and heavy atomic nuclei (i.e. protons and neutrons) was clear-cut and also
corresponded to a dierence in behavior, since heavy particles appeared to interact
much more strongly than light ones. Thus, mass appeared as a primary feature of particle identity which determined other properties. From the 1940s onward, however,
more and more particles with intermediate masses and ambiguous interaction properties were found, so that no clear-cut distinction obtained any more between light and
weakly interacting particles on the one side and heavy and strongly interacting ones
on the other. Mass slowly lost its status as a fundamental particle property and, as
mentioned above, at times even the possibility of particles having dierent possible
mass states was considered. Moreover, it was neither in practice nor in principle
possible to measure the mass of the new, unstable particles with arbitrary precision:
in practice due to their short lifetime, in principle because of Heisenbergs uncertainty
relation between mass/energy and time, according to which only innitely stable particles could have a sharply dened mass value. As a consequence, other properties
(spin, internal quantum numbers) became equally or even more important than mass
for identifying and classifying particles4 . In short, mass appeared increasingly often
as a contingent property of elementary particles, raising the question of why certain
particles had those, and not other mass values. As we shall see presently, a main motivation behind these reections on the origin of mass was the wish to nd a theoretical
scheme capable of expressing both masses and other particle characteristics in terms
of a smaller number of physical properties.
While experimental results were rapidly transforming the landscape of particle
physics, pivotal theoretical developments took place with the emergence of quantum
eld theory. Quantum eld theory as we know it today was developed around 1950
to account for quantum-relativistic interactions of electromagnetic kind (quantum
electrodynamics) (Schweber 1994). During the rst half of the 20th century several
quantum theories of electromagnetic interactions had been formulated, but all were affected by a fatal aw: when trying to use them to perform perturbative computations,
formally diverging (and therefore mathematically meaningless) expressions appeared.
The solution found for this problem in the late 1940s was a method (renormalization)
to formally subtract the divergent contributions and obtain nite and empirically
correct predictions for a broad range of quantum electrodynamics phenomena. The
formal subtraction of divergent terms was eected at a relatively high price, though:
the mass and charge parameters appearing in the mathematical expression dening
the dynamics of the system (the bare Lagrangian) could not anymore be set equal
to their observed values. Instead, certain combinations of parameters from the bare
4
Although mass remains today one among the possible characteristics identifying particles,
it is known since the late 1950s that particle states exist which have no denite mass value,
such as the long- and short-lived components of K-mesons. On the contrary, no particle
states have ever been observed which have no denite value of electric charge or spin.

The European Physical Journal H

Lagrangian and of divergent terms were assumed to have a nite value equal to the
real (i.e. measured) masses and charges. Thus, the measurable rest masses and charges
of particles entered quantum-eld-theoretical computations by way of an experimentally successful, but mathematically nonrigorous procedure which some regarded as
representing hitherto unknown physical eects, but others simply took as a warning
that the theory was awed and that its results should be taken with much caution.
For its physical interpretation the new quantum eld theory employed notions
developed before the war and originally applied to the case of photons: individual
particles were conceived as excitations of a quantum eld, and their production
and decay were regarded as the transfer of energy and other properties (e.g. spin,
charge) from one quantum eld to another (Pais 1986: 32435). A particularly important notion was that of the state in which the quantum eld had minimum energy,
i.e. the state where no excitations/particles were present. This state was (and still
is) usually referred to as the vacuum (state), although already in the 1930s it was
clear that it could not be physically understood as a classical vacuum. In the same
way in which quantum particles do not always have a denite value of position or
energy, quantum elds were regarded as not necessarily having zero value in their
vacuum/ground state, but rather as characterized by non-observable vacuum uctuation whose quantum average was zero. The origin of the divergent terms eliminated
by renormalization could be formally linked to the existence of those vacuum uctuations. In the new quantum eld theory the vacuum state was also important because
all particle states were represented by particle creation and annihilation operators acting on it. Since the 1950s the formal properties of vacuum state, vacuum
uctuations and renormalization have been the subject of increasingly rened mathematical analyses, but at that time, despite the empirical successes, the whole construct
still appeared quite problematic. In that context, many hypotheses on how to extend
quantum eld theory beyond the case of electromagnetic phenomena were tentatively
expressed in words or in sketchy mathematical arguments, some of which only found
a systematic mathematical formulation later on. This is the kind of approach that we
shall meet in the works to be discussed in the following sections.

3 Julian Schwingers -meson


Julian Schwinger (19181994) was one of the major actors in the development of
quantum electrodynamics and in the early 1950s he devoted much eort to extending
the quantum-eld-theoretical approach to the newly discovered phenomena. Paying
close attention to experimental results, he was among the rst ones to systematically try and describe all observed particles and interactions on the basis of a single
Lagrangian containing the smallest possible number of quantum elds. He did so in a
seminal paper published in 1957 which is often regarded as the starting point for the
development of the unication of electromagnetic and weak forces (Schwinger 1957)5 .
Schwinger introduced a small number of quantum elds carrying degrees of freedom
corresponding not only to space, time and spin, but also to the new internal quantum numbers, such as isospin. Other authors, among them Abraham Pais, Murray
Gell-Mann and many Japanese theorists, had already made use of this method to
make hypotheses on how specic processes might be formalized, but Schwinger employed rened group-theoretical considerations in an attempt to match, at least in
5

For a discussion of this paper and of the concerns from which it emerged, which cannot
be dealt with here, see (Mehra and Milton 2000: 415428).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

principle, all observed phenomena6 . His basic idea was that mass dierences within
particle multiplets should be shown to be the result of the breakdown of symmetries
of a unied bare Lagrangian in which all particles of a multiplet had equal mass. The
origin of the breakdown were eects due to vacuum uctuation, i.e. to renormalization, which he often described with the term dynamical. In other words: the bare
Lagrangian was symmetric and only contained very simple mass terms, while the real
Lagrangian leading to observable phenomena turned out to be much more complex,
displaying a hierarchy of symmetries and masses. Schwinger was not able to give any
mathematical details as to how the dynamical eects could be computed, and in
the paper he only described the symmetry structure of the bare Lagrangian and of
the terms that would eventually emerge in the renormalized one due to dynamical
eects. Yet his discussion was plausible enough to motivate the research of many later
authors, some of whom will be discussed in the next sections.
Having sketched Schwingers general strategy, let us take a closer look at his model.
At the beginning of the paper he described his aim in these words:
This note is an account of some developments in an eort to nd a description of the present stock of elementary particles within the framework
of the theory of quantized elds. [. . . ] We shall attempt to describe the massive, strongly interacting particles by means of elds with the smallest spin
appropriate to the statistics, 0 (B.E. [Bose Einstein]) and 1/2 (F.D. [Fermi
Dirac]). [. . . ] If the spin value is limited, the origin of the diversity of known
particles must be sought in internal degrees of freedom. We suppose that the
various intrinsic degrees of freedom are dynamically exhibited by specic interactions, each with its characteristic symmetry properties, and that the nal
eect of interactions with successively lower symmetry is to produce a spectrum of physically distinct particles from initially degenerate states. Thus we
attempt to relate the observed masses to the same couplings responsible for
the production and interaction of these particles (Schwinger 1957: 407).
Schwinger noted that [a]ssociated to each symmetry transformation T [of the
Lagrangian] is a physical quantity T [. . . ] which is conserved if the transformation
is a dynamical invariance property (Schwinger 1957, 408). A dynamical invariance
property meant for him an invariance valid not only at level of the bare Lagrangian,
but also when taking into account renormalization eects. Schwinger could never underpin his verbal statements with a full computation of dynamical eects, so that
these statements constituted an independent element of his theory which gave unity to
fragmentary mathematical arguments. In particular, Schwinger claimed that mass had
a dynamical origin (Schwinger 1957, 415) and a key assumption of his model was
that the bare Lagrangian should contain only dimensionless parameters like electric
charge, and that masses would emerge as products of an unknown physical agency:
In seeking possible forms for the interaction term in L [the Lagrangian] we
shall be guided by the heuristic principle that the coupling between elds is described by simple algebraic functions of the eld operators in which only dimensionless constants appear. This principle expresses the attitude that present
theories, which are based on the innite divisibility of the space-time manifold,
contain no intrinsic standard of length. The mass constants of the individual
elds are regarded as phenomenological manifestations of the unknown physical agency that produces the failure of conventional space-time description and
establishes the absolute scale of lenght and mass. On this view, the coupling
6

For a discussion on how authors used various methods, among them group theory, to
formalize the observations usually subsumed around 1950 under the name of conservation
of heavy particles see (Borrelli 2014).

The European Physical Journal H

terms employed within the present formalism should not embody a unit length
that nds its dynamical origin outside of the domain of physical experience to
which the theory is applicable (Schwinger 1957: 411).
The idea of the possible existence of a fundamental unit of length in nature (analogous
to the maximum velocity c and minimum quantum of action h) had been suggested by
Werner Heisenberg in the late 1930s as a possible solution to the divergence problem
of quantum eld theory, since the divergent terms became nite if space-time had
a discrete instead of a continuous structure (Schweber 1986: 100101). Since masses
were necessary parameters of particle interactions and had a physical dimension, they
could in some way be regarded as xing the characteristic length of a given interaction.
This was the rationale behind Schwingers choice of heuristic principle.
The rst step in building the bare Lagrangian was the choice of elds that would
appear in it. As already stated, Schwinger considered only fermion elds with spin
1/2 ( ) and boson elds with spin 0 (), though later on he also introduced weakly
interacting spin 1 (i.e. vector) bosons. Beside space-time dependence, each eld displayed indices representing internal degrees of freedom to be associated with the
dierent values of isospin and, for the baryons, nucleonic charge (todays baryon
number) attributed to the newly observed particles (Schwinger 1957: 412). Building
upon and group-theoretically formalizing already existing schemes of particle classication, Schwinger introduced in his Lagrangian a number of elds among which, for
our present purposes, the following are relevant:
(1/2) :
(1) :

a fermion isospin doublet (isospin +1/2 and 1/2) representing nucleons


(proton, neutron)
a boson isospin triplet (isospin 1, 0, 1) representing pions ( + 0 )7 .

Like today, the fermion eld had nucleonic charge (baryon number) equal to 1, and the
boson equal to 0. Most of the elds Schwinger introduced corresponded to observed
particles, but he also considered one that had so far no correspondence in reality
(Schwinger 1957: 410):
(0) : a boson isospin scalar (isospin 0) representing an as-yet-unobserved meson.
The eld (0) was particularly important since it had charge, spin and isospin equal to
zero and so, as we shall see, had in a certain sense the same properties as the vacuum
state. Out of these (and other) elds Schwinger constructed a bare Lagrangian in
which the usual terms describing the dynamics of the free fermion and boson elds
appeared. In Schwingers notation, which is not the one commonly used today, the
free fermion Lagrangian had the form (Schwinger 1957: 411):
L(1/2) = 1/2 [ (1/i) + m0 ]
where represented the Dirac matrix 0 , and the dot indicates symmetrical and
antisymmetrical multiplication, respectively as appropriate to the statistics of the
eld (Schwinger 1957: 411)8 . The dynamics of boson elds had the form:


L(0) = 1/2 + 20 2 .
7
Schwinger also introduced additional fermion isospin doublets representing the and
baryons, as well as: (1) , a fermion isospin triplet representing the baryons ( + 0 ); (0) ,
a fermion isospin scalar representing the 0 ; boson isospin doublets (1/2) representing the
K and K mesons (Schwinger 1957: 410).
8
Today, the term is usually written in one as = 0 and the details of the way in
which fermion elds are multiplied are not represented by a dot, but rather by indicating the
transposition of the eld variables. Here as in the following, I shall take over and comment on
the dierent notations employed by historical actors, as it is important to grasp how many

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

Here, too, a notation dierent from todays standards was employed, with standing
for and the dot again indicating the appropriate kind of eld index multiplication.
As we see, the free Lagrangian contained mass terms both for fermions (m0 )
and bosons (20 2 ), but Schwinger asked whether one should set the parameters m0
and/or 0 to zero and regard masses as arising only dynamically, in accordance to
the heuristic principle quoted above. As to possible interaction terms, he stated:
For interacting spin 0 and spin 1/2 elds only two types of coupling terms are
admitted by this [heuristic] principle, and (Schwinger 1957: 411). He
then employed group-theoretical considerations and experimental results to argue
that the interactions between nucleons ((1/2) ) and the ((0) ) and ((1) ) mesons
had the form:
L = g (0) 1/2(1/2)(1/2)
L = g (1) 1/2(1/2) 5 (1/2) ,
where g was the coupling constant for strong interactions, was again 0 , and ,
were matrices acting respectively on nucleonic charge and isospin indices (Schwinger
1957: 415). The matrix 5 was required in the coupling between nucleons and pions
because of the experimentally established pseudoscalar nature of pions. Having written down the general form of the bare Lagrangian, Schwinger went back to the issue
of whether bare mass terms for fermions and/or bosons should be present. To answer
this question he considered an innitesimal transformation T of the elds which would
leave the bare Lagrangian invariant if no fermion mass term was present (Schwinger
1957: 416). For nucleon, meson and pion elds, the transformation T , respectively
had the form:
(1/2) (1 + 1/25 )(1/2)
(0) (0) + (1)
(1) (1) (0)
where was an arbitrary innitesimal quantity9 . Although the fermion mass spoiled
the invariance, Schwinger noted that the fermion mass term m0 and the
fermion-boson interaction g (0) had a similar structure and could be combined
into a single term (m0 + g (0) ) . At this point, Schwinger stated: Hence if
m0 = 0, it is (0) m0 /g that transforms with (1) as a four-vector and the addition
of the constant to (0) does not upset the invariance of L if 0 = 0 (Schwinger 1957:
416). In other words, the transformation T was to be regarded as acting not on ((0) ,
(1) ), but rather on ((0) m0 /g , (1) ), i.e., writing explicitly what Schwinger only
hinted at:
((0) m0 /g ) ((0) m0 /g ) + (1)
(1) (1) ((0) m0 /g ).
As can be veried by direct computation, with this new denition of the action
of T the changes in the fermion mass term and in the nucleon-meson interaction
non-trivial mathematical details are today black-boxed in a convenient set of symbols
which theorists working in the 1950s still had to develop and standardize. Thus, working
out the correct form of a Lagrangian with some desired invariance properties was much
less straightforward that it may appear when using todays optimized notation.
9
Readers may recognize the transformation as a type of chiral symmetry acting dierently
on left- and right-handed components of the fermion eld and therefore broken by fermion
mass terms, but not by boson ones.

10

The European Physical Journal H

compensate, leaving the Lagrangian invariant as long as the boson mass 0 is set
to zero10 . Thus, Schwinger chose a bare Lagrangian where the fermion mass term
m0 was present, the scalar eld was shifted by the constant m0 /g , and the
boson masses were set to zero. The latter choice was no problem since, as Schwinger
pointed out:
[the quartic coupling of (0) and (1) elds] will produce eective mass
terms for each eld through the action of the vacuum uctuations of the other
elds. [. . . ] The 2 2 term that generates masses also disturbs the four dimensional symmetry and leads to dierent observed masses for the -meson and
the -meson. If the former becomes suciently heavy [. . . ] the requirement for
practical unobservability of this scalar particle will be satised (Schwinger
1957: 416417).
Later on, Schwinger described how the scalar eld might generate mass also for nonstrongly interacting fermions like the -meson (muon):
As a eld which is scalar under all operations [. . . ] (0) has a nonvanishing expectation value in the vacuum [. . . ] and thus a suitable -meson mass
constant might emerge from g (0) (Schwinger 1957: 423).
Once again, no mathematical details were given to describe the generation of mass:
only the remark that the eld (0) , being, like the vacuum, invariant (scalar) under
all operations, could have not only vacuum uctuations like all other elds, but also
a non-vanishing vacuum expectation value which here was equal to m0 /g . The
vacuum expectation value of the scalar eld could be used also to formulate a unied
theory of electromagnetic and weak interactions, an issue discussed in the second part
of Schwingers paper, where the author also attempted to explain the lower symmetry
(e.g. isospin and parity non-conservation) exhibited by weak interactions (Schwinger
1957: 422)11 . He suggested that all processes that could not be interpreted as strong
or electromagnetic interactions should be explained by introducing two charged vector
(i.e. spin 1) bosons (Z-elds) similar to the photon, but massive one of the earliest
systematic discussions of this idea (Schwinger 1957: 42425). Extending the notion of
isospin to non-strongly interacting particles, Schwinger assumed that the photon and
the two massive vector bosons formed an isospin-triplet, but explained:
we again use the -eld to remove three-dimensional internal symmetries
and produce masses for charged particles (i.e. the charged vector bosons). At
the same time, the coupling augments the mass of the -particle (Schwinger
1957: 424425).
It is important to note that the symmetry removed by the -eld was an isospindependent phase transformation which did not depend on space-time variables, i.e.
a global gauge symmetry and not a local one. Nonetheless, the structure of the
dynamical term for the Z-triplet and of the interaction between the Z-elds and the
-eld were very similar to those in todays model of electroweak unication and
indeed, as we shall see presently, provided the starting point for reections on local
10

Explicit computation also shows that, if boson mass terms are present, the application
of T to the Lagrangian gives rise to linear terms in the boson eld. While Schwinger avoided
such terms, as we shall see they were present in the version of the model proposed by GellMann and Levy.
11
Schwinger did not speak of weak interactions, but rather of physical processes that
exhibit a very long time scale (Schwinger 1957: 422). The notion that all such phneomena
might be explained by one new kind of interactions, today known als weak forces, was not
yet fully established at the time (Pontecorvo 1989).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

11

gauge invariant interactions (Schwinger 1957: 424). For example, the interaction term
between Z-elds and the scalar had the form (Schwinger 1957: 424):
2
1/22(0) 1/2Z t3 2 Z
LZ = gZ

(1)

with t3 a matrix acting on isospin variables. In conclusion, Schwingers theory provided a template for dynamically deriving both the variety of particle masses and
the dierent degrees of symmetry of particle interactions from an (almost) massless,
highly symmetric Lagrangian. The key ingredient in this scheme were the special
properties of a eld which was scalar with respect to all transformations and thus
might have a nonvanishing expectation value in the vacuum state. Thanks to this
vacuum expectation value both the scalar itself and the other elds could acquire
mass, while the symmetry of the bare Lagrangian was lowered. However, the main
steps of the dynamical origin of mass and of symmetry loss were described only verbally, and Schwinger was well aware that his work only represented a sketchy template
for further research, concluding the paper with the words:
What has been presented here is an attempt to elaborate a complete dynamical theory of the elementary particles from a few general concepts. Such
a connected series of speculations can be of value if it provides a convenient
frame of reference in seeking a more coherent account of natural phenomena
(Schwinger 1957: 433).

4 Gregor Wentzels spurion


Schwingers -eld was part of a very rened mathematical scheme for the unication
of particle interactions, but the notion that a particle having the same properties as the
quantum vacuum could be used to formalize and thus unify newly observed processes
had already been proposed earlier, albeit in a much more limited context, by Gregor
Wentzel (18981978). Wentzel was a German scientist who had contributed to the
development of quantum eld theory before the war and was at the time professor at
the University of Chicago (Schweber 1986: 346347). At the sixth annual Rochester
Conference on High Energy Nuclear Physics held at Rochester University in April
1956 Wentzel made a short contribution discussing how to account for a selection rule
observed in some slow processes (Wentzel 1956: VIII-1517).
To understand Wentzels argument one has to take into account the stand of theoretical particle physics at the time: in absence of a comprehensive theory, particle
decay and production were often formalized using Feynman diagrams, a visual and
computational tool introduced by Richard Feynman (19181988) in his version of
quantum electrodynamcs and further developed by Freeman Dyson (Kaiser 2005). In
Feynman diagrams, lines represented particles and the vertices where lines met stood
for processes in which particles were created or annihilated. In QED the lines and
vertices of the diagrams were fully determined by the Lagrangian, but for other particles interactions no Lagrangian yet existed. Nonetheless, physicists at times employed
a bottom-up approach where they tentatively introduced diagrams to help formalize processes in analogy to QED. The (at the time still hypothetical) selection rule
discussed by Wentzel stated that in a vertex due to a slow interaction the value
of isospin I either remained constant or changed by 1/2. Wentzel noted that the
evidence for the validity of this selection rule in slow transitions is, at best, rather
fragmentary (Wentzel 1956: VIII-15). If true, though, the rule raised a problem: in
absence of a Lagrangian, one would need some other formal means of imposing the
empirically established rule in the Feynman diagrams associated to various processes.
For example, electric charge conservation was insured by the fact that in the diagrams

12

The European Physical Journal H

Fig. 1. G. Wentzel: Vertex with spurion, from (Wentzel 1956: VIII-16).

only vertices where the charges added to zero appeared. Isospin conservation could
of course be guaranteed in a similar way, but what about the condition that isospin
should either be conserved or change by 1/2 and not more (e.g., 1 or 3/2)? Wentzels
suggestion to keep track of the selection rule was the following:
Let us, for merely formal purposes, imagine that in every vertex (open
circles in the gure) a spurious particle is emitted which carries away an I-spin
1/2 but carries no charge, or spin, or energy-momentum. It would be represented by an isospinor (corresponding to zero charge) with constant amplitude
in space-time. In the above diagram, the weak vertex would then look as follows (Fig. 1):
(S for spurion), and it should be of such nature as to conserve I-spin. Then,
for each simple or complicated diagram, I-spin is conserved for the system including the spurion; or, discounting the latter, we have, obviously, |I| = 1/2
(Wentzel 1956: VIII-16).
In other words, two kinds of vertices were allowed in Feynman diagrams: those where
isospin was conserved, and those where a nonobservable spurion originated from
the vertex carrying with it 1/2 isospin unit, representing a change of isospin equal
to 1/2. Wentzels spurion was conceived as a spurious, i.e. ctive, particle whose
only purpose was to formalize the isospin selection rule for slow processes, allowing to
derive further predictions in case the rule actually obtained. The spurion diered from
Schwingers -meson in many respects: the -meson was a real, albeit unobserved,
particle and it carried no isospin. It also had mass and did not disappear into vacuum.
Yet the two did share some formal properties, since the spurion, although endowed
with isospin, was a scalar with respect to all other transformations and was represented
by a eld with constant amplitude in space-time, i.e. with a non-vanishing vacuum
expectation value, just like the -eld (Wentzel 1956: VIII-16). Wentzels spurion
had in its early form no connection with mass generation, and was not referred to
by Schwinger, but, as we shall see later on, it eventually became entangled in the
discussions on the origin of mass. The wide reception of a sketchy idea like the spurion,
which Wentzel only published in the short proceedings contribution, may appear
surprising, but it is evidence of the key role of international conferences as a means
of communication complementing papers and preprints in (early) particle physics. In
these conferences, new ideas were often tested for the rst time or promoted in front
of a broad audience, giving rise to impromptu collaborations or leading scientists to
reect on them on their way home. Paradigmatic in this respect were the Rochester
conferences on high energy nuclear physics (later on on high-energy physics) which
Robert Marshak initiated at Rochester University in 1950 with the stated aim of
promoting an international networking involving also scientists from the far East and
from the Soviet block (Marshak 1989, Polkinghorn 1989).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

13

5 Werner Heisenbergs non-symmetric vacuum


In the early 1950s Werner Heisenberg (19011976), in collaborations with other authors, developed a scheme for a quantum eld theory accounting for all particle masses
and properties . Heisenbergs approach was unique in that he explicitly rejected the
possibility of performing perturbative computations with his theory and chose for it a
Lagrangian containing nonlinear couplings of spin 1/2 fermions which did not admit
of a perturbative treatment12 . Abandoning perturbative computations was a quite
problematic step, since they constituted the methodological core of quantum electrodynamics, but on the other hand it was clear at the time that nuclear, i.e. strong
interactions, could only be represented in terms of a very large coupling constant, so
that they would not to be amenable to perturbative treatment anyway. In this sense,
Heisenbergs theory made a virtue of necessity, attempting to exploit the peculiar
mathematical features which were assumed to be associated with nonperturbative effects. Although Heisenberg and his collaborators were in the end never able to deliver
more than qualitative estimates of observable phenomena, their verbal statements
and fragmentary mathematical arguments were seminal in establishing the direction
for further research on the dynamical origin of masses and particle properties and in
particular on the role of the vacuum state in this context. A further key assumption
of Heisenbergs approach was the rejection of the distinction between elementary
and composite particles (Heisenberg 1957). His nonlinear Lagrangian was not to
be thought of as representing a fundamental level of microphysics, but only as a tool
to compute the S-matrix, a mathematical construct from which observable prediction on particle production, decay and scattering could be obtained. Thus, not the
equations of the system, but only their solutions were amenable to a physical interpretation in terms of particles. As we shall see later on this approach would be taken
up and further expanded around 1960 in form of the bootstrap theory (Rechenberg
1989).
Heisenbergs idea was that all approximate symmetries of masses and particle
interactions could be derived from a highly symmetric Lagrangian of nonlinear fermion
interactions. The symmetry reduction (Symmetrieverminderung) was due to the
fact that, because of non-perturbative eects, no unique, symmetric ground state
(vacuum) solutions for the system equations existed, so that a number of vacua
could be dened having dierent symmetry properties and quantum numbers. This
situation was not surprising, according to Heisenberg: indeed, also in classical eld
theory and quantum mechanics it often happened that the ground state of a system
did not possess all symmetries of the systems equations and was therefore not unique:
It is in no way clear from the beginning that there must be a state vacuum possessing all symmetry properties of the initial equations. [. . . ] For
example the Dirac equation for the hydrogen atom has full rotational symmetry, but has no rotationally symmetric solution. [. . . ] If it turns out to be
impossible to construct a symmetrical vacuum state, this fact can only be
intuitively interpreted in the sense that the asymmetrical ground level is not
properly a vacuum, but rather a world state which constitutes the basis for
the existence of elementary particles13 .
12
Beside the two papers quoted below (Heisenberg 1957, D
urr et al., 1959), Heisenberg
and his collaborators published many others on the subject: an overview with references on
them can be found in Heisenbergs collected works (D
urr 1993).
13
Es ist keineswegs vom vornerein sicher, da es auch einen Zustand Vakuum geben
mu, der alle Symmetrieeigenschaften der Ausgangsgleichung besitzt. [. . . ] Zum Beispiel
besitzt die Dirac-Gleichung des Wasserstoatoms die volle Rotationssymmetrie, es gibt aber
keine rotationssymmetrische L
osung. [. . . ] Wenn es sich als unm
oglich erweist, einen voll

14

The European Physical Journal H

Since elementary particles in quantum eld theory were formally represented by particle operators applied to the vacuum state, it appeared plausible that dierences
between particle states might be represented not only by properties of the operators,
but also by those of the vacuum, so that the same operator could give rise to different particles. For example, Heisenberg suggested that the vacuum/world could be
endowed with a practically innite value of isospin, so that the [particle] states
neutron and proton would receive slightly dierent masses. They would in a sense
be the components of a doublet state nucleon + world in which the additional
isospin positions itself either parallel or antiparallel to that of the world (D
urr
et al. 1959: 446447)14. Moreover, a part of the vacuums isospin could be detached
and attached to a particle, while the isospin of a decaying particle could reattach
to the vacuum, giving rise to an apparent non-conservation (D
urr et al. 1959: 447).
The vacuum/world functioned therefore as a sort of isospin reservoire and took active
part in particle phenomena, although it could never be directly observed. The authors
noted: this idea is clearly closely related to the notion of spurion introduced by
Wentzel15 . The spurion was here taken to represent the fraction of isospin which
was taken away or added to the innite reservoir, transforming the vacuum into one
of its degenerate equivalents and resulting in a symmetry-reduction of the observed
processes. In later papers by Heisenberg and his collaborators, the spurion would
become increasingly important and be described as a change from one vacuum to
another vacuum (Heisenberg in D
urr et al. 1968: 20). Statements on the asymmetrical vacuum/world generating observed asymmetry and mass dierences were central
to Heisenbergs theory, but he was never able to fully underpin the verbal claims
with mathematical models, although he explored many of them with nonperturbative
methods. Thus, the vacuum/world story was not an interpretation of some mathematical structure, but rather an independent element of a hybrid knowledge construct
which would later inspire many authors.

6 Tony Skyrmes non-linear theory of strong interactions


Schwingers and Heisenbergs methods to obtain asymmetry from symmetry were
quite dierent: Schwinger employed a scalar eld with nite vacuum expectation value,
while Heisenberg assumed that, due to non-perturbative eects, the vacuum state itself would be degenerate and therefore asymmetric. Nonetheless, the two approaches
would be later combined, and among the rst authors to quote both of them was the
British theorist Tony Skyrme (19221987). Skyrme referred to Heisenberg (1957) in
attempting to formulate A non-linear theory of strong interactions (Skyrme 1958),
but chose to consider non-linear couplings between bosons instead of fermions. As
in Heisenbergs case, the non-linear nature of the Lagrangian could be exploited to
generate asymmetry form symmetry, but at the same time prevented from treating
the model perturbatively and therefore from establishing a simple relationship between the elds appearing in the Lagrangian and the particles observed in reality.
symmetrischen Zustand Vakuum zu konstruieren, so kann dies anschaulich wohl nur so
gedeutet werden, da es sich bei dem unsymmetrischen Grundzustand nicht eigentlich um
ein Vakuum, sondern um einen Zustand Welt handelt, der die Grundlage f
ur die Existenz
der Elementarteilchen bildet (Heisenberg et al. 1959: 446).
14
[es wird verst
andlich] da die Zust
ande Neutron und Proton eine etwas verschiedene
Masse erhalten. Sie w
aren gewissermaen die beiden Doublettenkomponenten eines Zutandes
Nukleon + Welt, bei dem sich der hinzukommende Isospin parallel oder antiparallel zu
dem der Welt stellen kann (D
urr et al. 1959: 446447).
15
diese Vorstellung ist oenbar eng verwandt mit dem von G. Wentzel eingef
ugten Begri
des spurions (D
urr et al. 1959: 447).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

15

While Heisenberg had started outright with a non-linear Lagrangian, though, Skyrme
quoted and followed Schwinger, taking as a starting point the usual theory of nucleons
interacting with a symmetric pseudo-scalar pion eld through a pseudo-scalar coupling containing an interaction term between nucleons and pions (g ) as well
as mass terms for nucleons (M0 ) and pions ((1/2)k 2 2i , with i (i = 1, 2, 3))
(Skyrme 1958: 264)16 . Like Schwinger had done, he then introduced a scalar isospin
singlet 4 and stated:
The rst proposed change is that the bare mass M0 [of nucleons] be
replaced by g4 , where the new meson eld is generally similar to the i [. . . ]
The second proposal is that these four elds should not be independent but
be constrained by a relation 2 (x) = Q2 (2) at all points [of space-time] x,
where Q is a fundamental constant of the dimensions of a mass or reciprocal
length. [. . . ] The third proposal concerns the meson mass term. The natural
generalization (1/2)k 22 has no signicance if the elds are constrained
by the condition (2). One possibility would be to argue that the meson mass
originates from the nucleon coupling. The alternative chosen here is to add a
fourth-order term +(1/4) 2 4 (Skyrme 1958: 264).
Thus, while Schwinger had simply stated verbally the possibility that a fundamental length would dynamically generate mass, Skyrme actually tried to write down
a model showing how the fundamental length mathematically came into play and
evenutally generated mass and broke symmetry through non-linear eects. He followed Schwinger in stating that the mass of the meson eld is generated by a 4
term (Skyrme 1958: 260) and added: This interactions destroys the invariance under the continuous groups of rotations in four dimensions (Skyrme 1958: 261). With
these changes, the initial Lagrangian took the form (Skyrme 1958: 265):


4 ( /x + g(4 + i5 i i )).17
L = 1/2
(d /dx )2 + 1/4 2
Although this Lagrangian was still linear, due to the constraint 2 (x) = Q2 the
four meson elds were not mutually independent and this fact would eventually
lead to non-linearities. To explore this issue, Skyrme introduced a simplied, twodimensional model in which he expressed the two components of in terms of a singe
angle variable (Skyrme 1958: 270)
3 = Q sin
4 = Q cos .
The resulting rewritten Lagrangian depended on sin and was therefore non-linear.
Discussing the quantization of the classical solutions to this simplied non-linear
eld theory, Skyrme suggested that dierent kinds of travelling and standing waves
might, when quantized, correspond to various observed particles (Skyrme 1958: 262).
In this context, non-linear fermion terms similar to Heisenbergs had to be added,
too (Skyrme 1958: 2623, 276277). In a subsequent paper Skyrme discussed a more
realistic, four-dimensional model where he dropped the strong constraint 2 (x) =
Q2 and instead required that 2 should have a non-vanishing vacuum expectation
16

Skyrmes notation is almost the same as Schwingers, but instead of the dot to represent the way in which elds are multiplicated, he used the symbol to indicate the
transposed and conjugate fermion led, as is usual today.
17
Here is once again the Dirac matrix 0 and the i are isopin matrices as in Schwingers
paper.

16

The European Physical Journal H

value Q2 which he again linked to the existence of a cuto, i.e. a minimal lenght, due
to unspecied non-linearities (Skyrme 1959: 236). As in the previous paper, this
condition led to expressing the meson eld in terms of angle variables whose dynamics
was determined by a non-linear Lagrangian. Only one set of meson elds was
introduced, but its dierent kinds of quantized waves were used to represent dierent
particles: K-mesons and pions (Skyrme 1959: 236). Skyrme proposed a hydrodynamic
analogy for interpreting how meson states would relate to the vacuum (Skyrme 1959:
236237):
The relation of a particle to vacuum is imagined as analogous to that
of uctuations in a uid endowed with density and direction at each point
in space. K-mesons correspond to waves involving density uctuations, while
-mesons are likened to waves of directional oscillations (Skyrme 1959: 236).
K and pion masses emerged dynamically: The phenomenological mass of the Keld has been replaced by a self-mass arising from the Fermion interactions [. . . ] the
current of the nuclear [i.e. fermion] eld can create [. . . ] a self-mass for the pion elds
(Skyrme 1959: 238, 241). We may thus conclude that, while Skyrmes mathematical
procedures were quite dierent from those of Schwinger and Heisenberg, his verbal
statements and the physical picture which he painted were very similar to the ones
Schwinger and Heisenberg proposed, with unspecied non-linearities giving rise to
a nonzero vacuum expectation value of scalar elds, which in turn was linked to a
cuto. These eects were formalized in a non-linear Lagrangian where no meson
masses were present, but which allowed their emergence through interactions. Like
Heisenberg, Skyrme connected his model to Wentzels spurion: The structure of the
model suggests a natural way for the introduction of the spurion, describing weak
interactions the violate strangeness (Skyrme 1959: 236) and, at the end of the paper,
he even referred to local gauge invariance, which we shall meet again presently: It is
apparent from the formalism of this model that the pion elds [. . . ] are closely related
to the concept of a gauging eld (Yang and Mills 1954) for locally variable rotations
in isobaric spin space (Skyrme 1959: 245). The mathematical model developed by
Skyrme in these and in later papers was quite dierent from those by Schwinger or
Heisenberg, and today his work is usually seen as part of the history of what would
later be called skyrmions, a subject which cannot be discussed here further (Skyrme
1988, Suryak 1992). As we have seen, though, Skyrme developed his ideas motivated
by the same issues prompting other authors to discuss the origin of mass, and like
them he employed the notion of a non-perturbative, dynamical transition from a
fundamental, symmetric, massless Lagrangian to the phenomenology of massive, nonsymmetric particles. Moreover, as we shall presently see, Skyrmes 1959 paper was
referred to by Jerey Goldstone in his seminal paper introducing what would later be
termed the Goldstone theorem (Goldstone 1961). Thus, although Skyrmes work
may a posteriori seem to have little to do with the Higgs boson, it has its place in the
story of the Higgs as a constitutive part of the knowledge which contributed to the
emergence of the Higgs mechanism. The same is true for the paper by Gell-Mann
and Levy to be discussed in the following section.

7 Murray Gell-Mann and Maurice L


evys -model
By the end of the 1950s some progress had been made in formalizing particle interactions, albeit no fully satisfying theory existed yet. Following the discovery of
parity non-conservation in weak interactions, a model for them had been developed
based on a Lagrangian containing two terms having opposite mirror transformation

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

17

properties: a vector (negative parity) and an axial (positive parity) current18 . This
V-A model for weak interactions displayed an interesting relationship to strong
forces: Assuming that the model was correct, experimental data implied that the
vector part of the weak interaction was almost equal for strongly and non-strongly interacting particles, with the small dierence probably due to electromagnetic eects.
This result in turn suggested that strong interactions, whatever their form might be,
left invariant (i.e. conserved) the weak vector current. For the axial part of weak interactions, no such invariance obtained, yet the relative dierence between weak axial
forces involving strongly and non-strongly interacting particles was only about 25%
and so appeared smaller than expected, leading to the hypothesis that strong interactions partially conserved weak axial current (PCAC hypothesis). If this hypothesis
was correct, the V-A theory of weak interactions could provide indications on the form
of a quantum eld theory for strong interactions. Reecting along these lines in 1958
John Polkinghorne employed Schwingers model of strong and weak interactions to
oer a theoretical explanation for the PCAC hypothesis (Polkinghorne 1958a, 1958b).
He wrote:
Feynman and Gell-Mann [1958] have recently proposed a universal Fermiinteraction using V and A couplings. They point out that the eective coupling
constant calculated from -decay agrees with that calculated from the decay
of 14 O to a high degree and that this is rather surprising as the strong interactions present in the latter case would be expected to renormalize the eective
coupling to a dierent value. They suggest that there may be a mechanism
at work that prevents this, analogous to the mechanism in electrodynamics
that ensures that all particles have the same renormalized charge. In fact this
requires that the coupling should be through a current that is conserved by the
strong interactions and they point out that such a current is available in the
vector coupling [. . . ] There remains the question of the axial vector current.
In this note we wish to point out that a recent suggestion of Schwinger
[Schwinger 1957] enables one to construct an axial vector current that is conserved to the extent that one neglects the medium strong19 K-meson interactions (Polkinghorne 1958a: 179).
Polkinghorne quoted Schwinger (1957) and employed that approach to construct a
partially conserved axial current. Like Schwinger, he used the interaction term between the meson and strongly interacting fermions (baryons) to produce mass terms
for the latter by shifting the -eld by m0 /g:
It is possible to reintroduce the baryon bare mass by using the similarity between the bare mass term and the baryon- coupling [Schwinger 1957].
A theory consistent with the conservation law [i.e. the partial conservation
of the axial current] is obtained by replacing everywhere by ( + m0 /g)
(Polkinghorne 1958b: 1709).
In 1960 Murray Gell-Mann and Maurice Levy referred to and built upon Schwingers
and Polkinghornes work to formulate a model of strong interactions which allowed
to derive not only the PCAC hypothesis, but also a recently formulated, empirically
plausible conjecture: the Goldberger-Treiman relation linking the weak and strong
interaction couplings of the pion (Gell-Mann and Levy 1960: 717719). This relation
18

The V-A theory was introduced by (Feynman and Gell-Mann 1958, Sudarshan and Marshak 1958, Sakurai 1958). For a historical overview of the developments with further references see (Pais 1986: 533538, Pickering 1984: 6871).
19
As noted above, in the 1950s it was still unclear whether, apart from electromagnetism
and gravity, only strong and weak interactions existed, or whether also medium-strong
ones obtained.

18

The European Physical Journal H

had been formulated by Marvin Goldberger and Sam Treiman (1958a, 1958b) on the
basis of dispersion theory and its original form was (Goldberger and Treiman 1958b:
358):

F (m2 ) = 0.13 2GmgA /(2 2 ).


Here F (m2 ) was a quantity whose value could be determined from the pion lifetime,
the parameter G depended on the strong coupling constant between nucleons and
pions, m was the nucleon mass and gA the weak axial coupling of the pion20 . The
relation was physically very interesting, as it connected strong and weak coupling
constants, and was (and still is) experimentally valid to a high degree of accuracy.
The original derivation by Golberger and Treiman, though, did not appear convincing
to most theorists and in the following years various authors proposed alternatives to
it (Treiman 1989). Gell-Mann and Levy (1960) showed how the relation could be
obtained as a consequence of the partial conservation of the weak axial current. Their
idea was to construct a Lagrangian for strong interactions of nucleons and pions which
possessed an approximate symmetry. The approximately conserved current associated
to that symmetry would have the right characteristics to be regarded as the weak axial
current. Once a form for the strong-interaction Lagrangian and the relevant weak axial
current had been proposed, Gell-Mann and Levy showed how one could derive from
them also the Goldberger-Treiman relation.
In their paper the authors wrote down three possible Lagrangians and one of
the them (the -model) built upon Schwingers model and contained a scalar eld
associated to a hypothetical meson with no charge and no isospin: we introduce
(following Schwinger [1957]) a new scalar meson , with isotopic spin zero. It has
strong interactions, and thus might easily have escaped observation if it is much
heavier than (Gell-Mann and Levy 1960: 717). Like Schwinger, they wrote down
a Lagrangian in which strong interactions between nucleons (eld N) were mediated
by an isospin triplet (the three pions) and an isospin scalar (the meson).
The bare Lagrangian had the same structure as Schwingers one, and in particular
contained a nucleon mass term m0 N N and the fermion-boson interaction g0 N N
(Gell-Mann and Levy 1960: 717). Again following Schwinger (and Polkinghorne) the
eld associated to the meson was shifted to  = ( m0 /g), giving rise to a
Lagrangian in which the nucleon mass term had disappeared. However, other than
Schwinger, Gell-Mann and Levy had introduced a boson mass term 20 /2( 2 + 2 ) in
the initial Lagrangian, and this term, when written as a function of  , gave rise to a
linear term. The new Lagrangian had the form:
L2 = N [ g0 (  + i 5 )]N 20 /2( 2 + 2 ) 20 /2f0 
+ [other derivative and quartic terms for the bosons] (Gell-Mann and Levy 1960:
717)21 .
The reason why the two authors introduced 0 is that they were not interested in
preserving symmetries, but rather in breaking them to obtain a partially conserved
current. The linear term in  broke a symmetry similar to Schwingers T , i.e. a
transformation having chiral character, and the current associated to it had exactly
the right form to be a candidate for the weak axial weak current (Gell-Mann and Levy
1960: 718). From these premises Gell-Mann and Levy were able to derive both PCAC
20

Today, the relation is usually written in the form: f = mgA /gNN where f is the
pion decay constant and gNN the coupling of the strong interaction vertex where a pion, a
nucleon and an antinucleon meet (Treiman 1989: 385).
21
As in todays use, the term indicates here a scalar product between the space-time
derivative and the four Dirac matrices, while the are isospin matrices, as in Schwingers
model.

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

19

and the Goldberger-Treiman relation. Although the fermion mass had disappeared
from the bare Lagrangian, the eld gave rise to nucleon mass terms dynamically, as
was the case in Schwingers theory, but Gell-Mann and Levy were not very enthusiastic
about this curious property and indeed regarded the whole construct as articial:
The model, although it has some agreeable features that we have mentioned, is quite articial. A new particle is postulated, for which there is no
experimental evidence. [. . . ] The fact that the coupling is responsible for the
nucleon mass is a curious property of the model. Unless we can explain all
masses, or at least all baryon masses, in a similar way, it is not very satisfactory (Gell-Mann and Levy 1960: 719).
As we see, Gell-Mann and Levy did not combine their mathematical construction with
verbal statements about the vacuum state, the dynamical origin of mass, or the dynamical breaking of symmetry: they were interested rather in explaining phenomenological relations than in speculating about the origin of mass and asymmetry. Yet their
-model (still unter that name) is found today in most quantum-eld-theory manuals
as a paradigmatic example of spontaneous symmetry breaking (Peskin and Schroeder
1995: 348350 Itzykson and Zuber 1984: 540549). It seems therefore strange that
some historians have denied that Gell-Mann and Levys model should be seen as part
of the history of spontaneous symmetry breaking and of the Higgs mechanism (Brown
and Cao 1991: 232233), and it is my opinion that this may be due to the fact that,
although the mathematical formalism seen today as that of spontaneous symmetry
breaking was present in Gell-mann and L`evys work, the verbal narrative associated
to it was not a further indication of how important verbal statements are in the
construction and transmission of physical concepts.
The work of Gell-Mann and Levy was of key importance for the development of
notions of broken symmetry. Schwingers idea that the many approximate symmetries
characterizing particle phenomena could be explained in terms of a symmetrical fundamental theory was extremely suggestive, but also vague in both mathematical and
physical sense. Gell-Mann and Levy had now provided a concrete example of how a
connection between Schwingers Lagrangian and the partial symmetry between axial
and vector weak interactions could be established. Moreover, the symmetry breaking
was connected to the appearance of mass terms. Finally, there was a more technical
mathematical aspect of the -model which would play a key role later on: the -model
was renormalizable thanks to a divergence cancellation due to so-called tadpole diagrams where a meson decayed into a nucleon and an antinucleon which then
mutually annihilated letting the meson disappear into the vacuum (Gell-Mann and
Levy 1960: 719). Such processes are only possible for scalars with nonvanishing vacuum expectation value and formally corresponds to Feynman diagrams of the sort in
which spurions appeared.

8 Abdus Salam and John Wards unied local gauge theory


Among the rst authors to pick up and extend Schwingers approach to unication of electromagnetic and weak interactions were Abdus Salam (19261996) and
John Ward (19242000), who did so in a series of papers published between 1959
and 1961 (Salam and Ward 1959, 1960, 1961, Salam 1961, Fraser 2008: 203232). In
his 1957 paper Schwinger had proposed to unify weak and electromagnetic interactions by assuming that the photon was one component of an isospin triplet whose
other two elements were massive, charged boson mediating weak interactions. In the
bare Lagrangian there was full symmetry between the components of the triplets,

20

The European Physical Journal H

but dynamical eects due to the nonvanishing vacuum expectation value of the eld generated the masses of the weak vector boson. Those eects also lowered the
symmetry of weak interactions, allowing for parity non-conservation. In Schwingers
scheme the bare Lagrangian was invariant with respect to (innitesimal) rotations
of the fermion elds in isospin space (a global gauge transformation). However, both
Chen Ning Yang and Robert Mills (19271999) in a paper and Robert Shaw in his unpublished dissertation had proposed a Lagrangian symmetric with respect to local
rotations of fermion elds in isospin space, i.e. rotations whose parameters depended
on space-time variables (Yang and Mills 1954, Shaw 1955, Pickering 1984: 160165).
Such a symmetry could be implemented in a fermion Lagrangian only by adding to
it a triplet of vector elds which compensated with their own change for the eect
of the space-time-dependent transformation of fermion elds, just like the photon
did in electromagnetism. In Salam and Wards model Schwingers three vector bosons
became the elds which guaranteed local gauge invariance for the unied interactions:
Following some ideas rst advanced by Schwinger [Schwinger 1957], we
show that parity-conserving weak and electromagnetic interactions taken together form a unit which exhibits full rotational symmetry in Q [i.e. charge]22
space. This full symmetry is broken on the one hand by parity-violating weak
interactions [. . . ] and on the other hand by strong interactions. We go further
in respect of the rotational invariance of electro-magnetic and (parity conserving) weak interactions in Q space; we wish to make the hypothesis that the
orientation of all three charge axes can be chosen arbitrarily at all space time
points. Just as it is necessary to introduce the electromagnetic eld if it is
assumed that the orientation of axis in the conventional space [. . . ] is arbitrary
at all space time points, our three-.dimensional gauge invariance makes it
necessary to introduce in Q space a triplet of elds consisting of two charged
vector Bose elds in addition to the electromagneitc eld (Salam and Ward
1959: 569).
However, the symmetry was only valid as long as the gauge vector bosons remained
massless lke the photon. In their rst paper Salam and Ward admitted that introducing masses in their scheme was a problem and deferred its treatment to a future
paper:
In a sense then the existence of mass is incompatible with full rotational
symmetry in charge space; the fact that all known charged particles have mass
is the expression of the fact that one particular axis is preferentially chosen in
chage space. [. . . ] The major problem that remains is the problem of the mass
of the A elds [i.e. the charged vector bosons]. [. . . ] We propose to come back
to this problem in a subsequent paper (Salam and Ward 1959: 573574).
In a follow-up paper the two authors reiterated the importance of local gauge invariance as a guiding principle for writing fundamental interactions of elds and
stated that they would now reconsider the problem already addressed in the earlier
paper (Salam and Ward 1961: 165166). Their rst step was to follow Schwinger in
introducing a scalar iso-scalar particle to propose a unied local gauge invariant
scheme for strong and weak interactions. They introduced six vector bosons eld, out
of which the three particles mediating strong interactions (v-bosons) and the three
mediating weak ones (u-bosons) would arise (Salam and Ward 1961: 166167). The
22

Salam and Wards charge space was dened in terms of isospin space. To this aim they
followed Schwinger in assigning isospin values also to non-strongly interacting elds.

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

21

interaction Lagrangian had the form (Salam and Ward 1961: 167):
Lint = [( /x /x ) + N+ 4 N]u
+ [/x /x + N+ 4 5 N]v
[+ terms quadratic in u and v].23
The authors wrote: We wish to identify the -containing part of the above Lagrangian
as represeinting strong interactions and the remaining as representing weak interactions. [. . . ] The strength of the strong coupling comes about if we assume that the
vacuum expectation value of (0 ) does not equal zero but (gs /2MN ) (1/g ) and
added in a footnote: Notice that the term 2 (X 2 + Y 2 ) [X, Y = gauge vector boson elds] in [the Lagrangian] could give mass terms for u and v particles (Salam
and Ward 1961: 167). This unied model had serious aws in its compatibility to
experiments, such as that it conserved parity, and Salam and Ward where aware
that it might not be renormalizable because of the mass problem (Frazer 2008: 214).
Nonetheless, at the Conference on strong interaction which took place in Berkeley
on December 27th29th 1960, Salam presented an overivew of the model, showing
how it could in principle oer solutions to various problems of particle physics. Salam
strove to give credibility to his and Wards model by showing how the hypothetical
meson could be identied with an observed particle: the K10 meson (Salam 1961).
This identication had already been proposed not only by Skyrme, but also by Salam
and Ward in a short paper where they had suggested that an observed candidate
for Wentzels spurion could be the recently observed K10 meson (Salam and Ward,
1960). In his 1961 paper Salam combined this suggestion with the model of unied
interactions previously developed, and also employed the scalar eld to give rise to
the I = 1/2 rule discussed by Wentzel and Heisenberg. First of all, he noted the
special properties of the K10 meson:
We [i.e. Salam and Ward] started with the observation that of all the
known particles of physics, there is one which carries the same quantum numbers as the vacuum (except for energy and momentum). This is the K10 particle
[. . . ] with zero spin, zero charge, zero baryonic number, and P C = +1. Thus,
there is no intrinsic reason for the eld corresponding to this particle to have
a zero expectation value. [. . . ] We would like to take the attitude that the K10
eld consists of two parts: the classical substratum K10  [i.e. a constant]
0(q)
and the quantized part K1 responsible for the particle aspects of K10 . Thus,
0(q)
K10 = K10  + K1 . [. . . ] A neutral particle of zero spin, zero I spin [isospin],
and even parity has long been a favorite with theoreticians (Schwinger called
this object ). [. . . ] Ward and I have tried to combine this type of thinking
with a gauge theory of elementary interactions (Salam 1961: 428429).
Salam went on to expound the model formulated earlier on, where the -eld was
now identied with that of the K meson. Its vacuum expectation value gave rise to
the symmetry dierences among interactions, as well as to the masses for the new
vector bosons and was formalized by means of tadpole diagrams. At the level of verbal statements and fragmentary mathematical arguments, Salam and Wards model
contained two key element that would characterize later explanations of the origin of
mass: a broken local gauge invariance and a scalar eld with non-vanishing vacuum
expectation value. Soon these notion would combine with Heisenbergs approach, and
23

Here 4 stands for 0 and the symbols , N, u and v stand for triplets or doublets
in charge space representing repsectively pions, nucleons and the vector bosons mediating
strong and electro-weak interactions. The symbol indicates vector product.

22

The European Physical Journal H

the link was made in the work of Yoichiro Nambu and Giovanni Jona-Lasinio, usually
referred to as the origin of the notion of spontaneous symmetry breaking (Brown
and Cao 1991, Pickering 1984: 168171).

9 Yoichiro Nambu: symmetry loss and recovery in superconductivity


Before going on to consider the work of Nambu and Jona Lasinio, let us look back
at the various proposals for generating mass discussed up to now. In all of them
both the vacuum state and some form of symmetry loss played a role: Schwinger,
and following him Gell-Mann and Levy as well as Salam and Ward, introduced a
scalar eld with nonvanishing vacuum expectation value which, through dynamical
eects due to renormalization, led from a symmetric, (almost) massless Lagrangian
to a non-symmetric, massive one. In these models the vacuum state was unique and
unproblematic. In Heisenbergs model, on the other hand, there was no scalar particle
with nite vacuum expectation value, and the vacuum state itself led to the loss
of the symmetry of the bare Lagrangian, and so to the emergence of the variety of
particle properties and masses. Heisenberg suggested that the spurions proposed by
Wentzel, when they disappeared into the vacuum, could be interpreted as representing
transitions from one asymmetric vacuum to another. However, Heisenberg did not
regard the spurions as the origin of a symmetry loss and of mass, as both results were
due to non-perturbative eects. Finally, Skyrme used unique mathematical methods to
argue that, starting from a massless, symmetric Lagrangian, non-linear (and therefore
non-perturbative) eects could give rise to a nonzero vacuum expectation value of a
scalar eld and, through that, to a description of the observed massive, asymmetric
particles. Despite the dierences, in all these cases mass generation was linked to
symmetry loss in the transition from the bare to the real Lagrangian and today
we are tempted to regard these models as examples of a spontaneously broken or
hidden symmetry, although this way of looking at things was developed only later
on building upon the results discussed until now. Nonetheless, there is little doubt
that most authors discussed were struggling to interpret the approximate symmetries
in terms of which particle phenomena had come to be described by conceiving a third
way between an exact symmetry and a broken one. What must be noted is that at
the time it remained largely open whether the symmetry loss should in itself be seen
as a physical phenomenon, or rather as a formal step. As we shall see presently, it was
the connection to solid state physics, and more precisely to superconductivity, that
gave the whole discussion a more concrete character.
As seen in the previous chapters, perturbative contributions resummed at all orders and constitutively non-perturbative eects were essential to the approaches of
Heisenberg, Schwinger and the other authors discussed above. Methods to compute
such eects were largely lacking in particle theory, but some had been developed in
solid state physics and were imported back into high energy theory in a process of
cross-fertilization which had started already in the 1930s. The formal analogy between
high energy and condensed matter physics has a prominent role in todays particle
research and in particular the mass generation through the Higgs mechanism is often
regarded as formally analogous to phase transitions such as those of ferromagnetism or
of superconductivity, which are also described as instances of spontaneous symmetry
breaking (Peskin and Schroeder 1995: 292294). In fact, it is not possible to discuss
reections on the origin of mass in particle physics without making a detour through
condensed matter physics, looking at how the interplay between the two elds led to
the emergence of the notion of spontaneous symmetry breaking. Since that notion
remains problematic still today, no exhaustive treatment of the issue is possible in

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

23

the present context, and I shall only report the reections of the historical actors as
far as they are relevant for the theme of this essay.
In the 1950s it was common to regard the ground state of the electron gas
in metals as formally analogous to the quantum-eld-theoretical vacuum, expressing
the states of the electrons in metals in terms of electrons and holes, i.e. of electrons excited above the ground state and of the corresponding empty places left in it
(Hoddeson et al. 1992). In 19561957 the method of Feynman diagrams, which had
been developed for quantum electrodynamics, was adapted to solid state physics to
allow a treatment of many-body problems also in cases in which the interactions could
not be regarded as small perturbations (Kaiser 2005: 237244). Although Feynman
diagrams were in themselves perturbative, in this case they provided a means of approximate non-perturbative computation in terms of electrons and holes. This
formalism proved essential to explain superconductivity, a phenomenon known since
the early 20th century in which a material, when brought to a very low temperature,
suddenly loses practically all resistance against electric current, becoming a superconductor (Hoddeson et al. 1992). From the physical point of view there was in the
1950s little doubt that superconductivity was an eect due to electromagnetic interaction, but it had proven very dicult to formulate on the basis of electromagnetism
a model which could account for superconducting behaviour. The phenomenon was
only sketchily formalized by assuming that the superconducting state was characterized by a nite energy gap separating the ground state from the lowest excited
state of electrons. As long as electric current could not provide electrons with enough
energy to bridge the gap, the current lost no energy to them and could propagate
practically without resistance. In 1957 John Bardeen (19081991), Leon Cooper and
John Schrieer formulated a theory from which the energy gap and other features of
superconductivity could be derived, but the BCS-theory was at rst met with scepticism because it did not possess the local gauge invariance of electromagnetism which
guaranteed to charge conservation. Soon, various authors showed how to derive the
BCS-results in a gauge-invariant way, among them Nikolai Bogoliubov (19091992),
who employed Feynman diagrams methods to this aim (Bogoliubov 1958). Bogoliubov
showed that the superconducting ground state could only be derived by using nonperturbative methods and could be expressed in terms of quasi-particles, i.e. linear
combinations of (negative) electrons and (positive) holes, underscoring the apparent
non-gauge invariant and non-charge conserving character of the theory. However, he
also showed that the physical properties of the superconducting system had to be
computed by taking into account not only quasi-particles, but also their collective
excitations again a non-perturbative eect whose contributions restored gauge
invariance. Thus, in the late 1950s the theory of superconductivity was linked to a
problem of apparent symmetry loss, which was however regarded as existing only at
the formal level, since no one believed that the fundamental symmetry of electromagnetism was actually lost.
In 1960 the Japanese theorist Yoichiro Nambu, who had been a professor at the
University of Chicago since 1958, oered a new physical interpretation of the apparent loss of local gauge invariance linked to superconductivity (Nambu 1960a). He
summarized Bogoliubovs result and stated:
If such collective states [i.e. collective excitations] are essential to the
gauge-invariant character of the theory, then one might argue that the former is a necessary consequence of the latter. But this point has not been clear
so far (Nambu 1960a: 648).
To explore this questions Nambu reformulated Bogoliubovs model in terms of selfconsistency relations: the wave function of quasi-particles in the superconducting state
was regarded as the solution of an equation in which a single quasi-particle interacted

24

The European Physical Journal H

with the mean eld of all others. That mean eld was in turn derived by assuming that
all quasi-particles were described by the same wave function as the rst one. Nambu
showed how the self-consistency equation could formally be derived by starting from
the Hamiltonian H describing the relevant electromagnetic interactions and rewriting
it in terms of the mean elds (Nambu 1960: 649). The Hamiltonian H was as usual
the sum of free electron Hamiltonian (H0 ) and the interaction terms (Hint ), but
Nambu considered also the Hamiltonian describing mean eld interactions (Hs )
and followed Bogoliubov in writing:

H = H + Hint = (H0 + Hs ) + (Hint Hs ) = H0 + Hint
.

The new way of separating the Hamiltonian allowed Nambu to look at old com
putations in a new light. The term Hint
contained non-perturbative eects linked

to the mean eld, but an expansion considering the whole Hint
as a perturbation
was formally possible, provided the relevant Feynman diagrams were computed using
the non-perturbative techniques of solid-state-physics. This was the path chosen by
Nambu: Combining perturbative techniques from quantum electrodynamics with nonperturbative diagrammatic methods he was able to reproduce Bogoliubovs results in
a form which allowed to explore the overall symmetry properties of the theory, where
specic contributions (energy gap, collective excitations) could be non-symmetric,
but nonetheless combined into a symmetric whole (Nambu 1960: 657658). In particular, he could argue that the superconducting, non-perturbative solutions to the
self-consistency equations implied the existence both of an energy gap and of bound
states (i.e. collective excitations) of quasi-particles, because only the combination of
those two elements guaranteed local gauge invariance in physical processes. Nambus
analysis delivered no new phenomenological insights, but purported to oer a deeper
understanding of the physical connection between local gauge invariance and collective
excitations:
The existence of the bound states is a logical consequence of the existence
of the special self-energy [i.e. the energy gap of the superconducting state]
and the gauge invariance, which are seemingly contradictory to each other
(Nambu 1960a, 650).
In other words: the combined existence of local gauge invariance and of the noninvariant superconducting state with the energy gap logically led to the emergence of
non-invariant collective excitations to ensure that in the end the invariance would be
only apparently lost. Although his work is usually referred to as the rst explicit discussion of spontaneous symmetry breaking, Nambu never used the terms symmetrybreaking or symmetry reduction and never even suggested that a breakdown of
the invariance really took place. In his treatment quasi-particles and their collective excitations were presented not as two terms in a mathematical computation,
but as two distinct physical consequences of local gauge invariance which could only
be computed nonperturbatively and were both non-invariant, yet combined to guarantee invariance. Thus, Nambu had framed formal results concerning the theory of
superconductivity in terms of apparent loss and recovery of a physically signicant
invariance, constructing a hybrid theory out of verbal statements, experimental results and noncompelling mathematical arguments. This mixture, as he soon realized,
could be extremely appealing for particle physicists.

10 Yoichiro Nambu and Giovanni Jona-Lasinios model of mass


generation
Already in 1960 in a short contribution to Physical review letters Nambu had applied
his approach to particle physics (Nambu 1960b). In 1961, together with Giovanni

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

25

Jona-Lasinio, he extended and rened his reections on apparent loss and recovery
of invariance in the subnuclear domain. Starting point was a purely formal similarity
between Bogoliubovs equations for quasi-particles on the one side and the Dirac
equation for fermions on the other. Bogoliubovs equations for electrons and holes
with momentum p read (Nambu and Jona-Lasinio 1961: 345):

Ep+ = p p+ + p

Ep
= p p
+ p+

where E was the energy operator, the s described electrons or holes, p was the
kinetic contribution to the Fermi energy and the energy gap of the superconductor24 . The two latter values were linked to the energy Ep of an electron or hole with
momentum p by the relation:
1/2

.
Ep = 2p + 2
The Dirac equations had the form (Nambu and Jona-Lasinio 1961: 345):
E1 = p1 + m2
E2 = p2 + m1 ,
where 1 and 2 stood for the left- and right-handed components of the Dirac eld,
were spin matrices, and m was the fermion mass, related to the energy and momentum
of the fermions by the usually relativistic formula:
1/2

Ep = p2 + m2
.
Although the physical signicance of the two formulas was completely dierent, the
structural similarity was evident, and in particular the Dirac mass m occupied exactly
the same place as the energy gap . The authors accordingly suggested that, in the
same way in which the existence of a nonzero energy gap of superconductors was
linked to an apparent loss of local gauge invariance, so the mass m of Dirac particles
might be connected to the apparent loss a specic symmetry (chiral symmetry)
which only existed for m = 0:
As the energy gap in a superconductor is created by the interaction,
let us assume that the mass of a Dirac particle is also due to some interaction between massless bare fermions. A quasi-particle in a superconductor is
a mixture of bare electrons with opposite charges (a particle and a hole) but
with the same spin; correspondingly a massive Dirac particle is a mixture of
bare fermions with opposite chiralities. [. . . ] if a Dirac particle is actually a
quasi-particle [. . . ] then there must also exist collective excitations of bound
quasi-particle pairs (Nambu and Jona-Lasinio 1961: 346).
In other words, Nambu and Jona-Lasinio assumed that some as-yet-undetected fundamental massless fermions existed, whose interactions gave rise to massive Dirac
particles. They identied the Dirac particles with nucleons and their collective excitation with pions, and noted that, although their theory had the characteristic
features of a compound-particle model, it was unique in that unlike most of the existing theories, dynamical treatment of the interactions makes up an essential part of
24
Because of Paulis exclusion principle, no two fermions can occupy the same state, so
that the minimum energy state of a system of N fermions correspond to a situation where
all the N lowest energy states are lled and the higher ones are empty. The Fermi energy is
the energy of the highest level occupied.

26

The European Physical Journal H

the theory (Nambu and Jona-Lasinio 1961: 345). Nambu and Jona-Lasinio used the
term dynamical much in the same sense as Schwinger had done, although they did
not refer to his work. However, Schwinger had not given any mathematical procedure
for computing how mass arised dynamically, but had only shown through general considerations that mass terms could arise from a massless Lagrangian. Instead, Nambu
and Jona Lasinio could provide approximate computations to show how the properties of mass and interaction terms were produced dynamically by nonperturbative
eects. They performed these computations following the analogy to superconductivity and using the diagrammatic methods developed in solid state physics which were
known to deliver correct predictions in that case. As hypothetical underlying symmetric theory they employed a model similar to Heisenbergs nonlinear spinor theory,
saying [a]lthough this [model] looks similar to Heisenbergs theory, the dynamical
treatment will be quite dierent and more amenable to qualitative understanding
(Nambu and Jona-Lasinio 1961: 347). The qualitative understanding was a combination of verbal statements and mathematical arguments supporting the notion of
apparent symmetry loss and equally apparent recovery in which nucleons replaced
quasi-particles, and chiral symmetry the local gauge invariance. The key idea of the
work was to exploit the techniques used by Nambu in his previous paper to explore
non-perturbative eects in particle theory. In formal analogy to the case of superconductivity, the authors took the Lagrangian for Dirac elds L = L0 + Li (free dynamics
plus interaction terms) and rewrote it as (Nambu and Jona-Lasinio 1961: 348)25 .
L = L0 + Li = (L0 + Ls ) + (Li Ls ) = L0 Li
where Ls was the Lagrangian representing the self-interactions of the Dirac eld,
which formally corresponded to the mean eld interactions of Nambus model. At
this point, the methods used for determining the properties of (perturbative and nonperturbative) quasi-particle states could be employed to investigate those of (perturbative and non-perturbative) nucleon (and more in general particle) states emerging
from the hypotheical fundamental particle Lagrangian.
Although from the formal point of view this analogy was rather straightforward,
it is important to note that, physically, it implied a radical twist: nucleons were real
particles, while quasi-particles were formal constructs that had to be regarded as
strictly non-observable precisely because of their asymmetry, i.e. of their having no
denite electric charge. Electrons in solids, on the other hand, were observable physical
entities, but in the analogy they corresponded to Heisenbergs hypothetical nonlinear
spinor elds. Yet the analogy was very powerful, as it provided an empirically-rooted
exemplar to conceive the two levels of particle theory suggested by both Heisenberg
and Schwinger, and even oered approximate methods to compute how the properties
of the two levels were connected. Here, too, the vacuum state played an important
role. Following the formalism developed earlier on by Nambu, the authors wrote down
approximate self-consistency equations for the system, whose solutions dened all
possible particle states. The equations had two kinds of solutions: those that could
be found using perturbative approximations, and those that could only be computed
non-perturbatively. Accordingly, two vacuum states existed: a perturbative and a
nonperturbative one, and the two innite sets of solutions were represented by creation
operators acting respectively upon one or the other of the two vacua. No actual
computations of the solutions were possible, but the argument was plausible enough
when combined with the analogy to solid state physics. In condensed matter theory,
25
In the particle case, Lagrangian formalism had to be used instead of the Hamiltonian one,
since relativistic invariance was necessary. The procedures could be adapted, but the authors
were well-aware of the risks of making analogies between non-relativistic and relativisitc
systems (Nambu and Jona-Lasinio 1961: 346).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

27

only nonperturbative solutions were superconducting, and a similar situation would


obtain for particles, where nonperturbative solutions corresponded to a massive world:
The two worlds based on (0) and (m) [i.e. the perturbative/massless and
nonperturbative/massive vacuum] are physically distinct and outside of each
other. No interaction or measurement, in the usual sense, can bridge them in
nite steps (Nambu and Jona-Lasinio 1961: 350).
Here the authors presented a partly verbal and partly mathematical discussion of
degenerate vacua which might appear similar to Heisenbergs, but was actually quite
dierent from it. While Heisenbergs vacuum/world acted like an innite reservoir
of quantum numbers which participated in particle phenomena by constantly transforming itself, Nambu and Jona-Lasinios vacua were passive, relegated in dierent,
non-communicating worlds and could not physically transmute into each other by
means of spurions. Moreover, Nambu and Jona-Lasinios fundamental symmetry did
not manifest itself in approximately symmetric form, such as particle quasi-multiplets,
or in processes violating an otherwise valid conservation law (e.g. (isospin) = 1/2 decays), but rather as phenomena in which the symmetry was not even approximately
present. Despite these dierences, though, the work by Nambu and Jona-Lasinio was
soon combined with previous reections on symmetry reduction and mass generation
and, within a couple of years, the idea of nding nonperturbative superconductor
solutions in particle physics came to be seen as a most promising answer to the questions raised on symmetry reduction since the 1950s. As we shall see presently, it was
only at that stage that this approach came to be described as a symmetry breaking,
and, later on, as a spontaneous symmetry breaking.

11 Jerey Goldstones massless bosons and the black-boxing


of nonperturbative eects
Among the rst to react to Nambus work was the British theorist Jerey Goldstone
(1961), who learnt about the scheme very early, from the preprint of Nambus physical
review letter (Nambu 1960b). Goldstone, who had contributed to develop diagrammatic methods for non-relativistic systems (Goldstone 1957), took the decisive step
of studying the superconducting solutions of particle physics by black-boxing the
non-perturbative eects in a scalar eld with a non-vanishing vacuum expectation
value which represented the collective excitations: the models considered here all
have a boson eld in them from the beginning. It would be more desirable to construct bosons out of fermions (Goldstone 1961: 154). As we have seen, the notion
of generating mass and symmetry breaking with the help of the vacuum expectation
value of a scalar eld was not a new idea at the time, yet Goldstone did not quote any
of the authors who had discussed it previously. Nonetheless, he described the dynamics of the scalar eld using a Lagrangian containing the same couplings as Schwingers
one (Goldstone 1961: 155):
L = (i /x m) + 1/2(/x /x 20 2 ) g0 5 1/240 4 .
As we shall see, Goldstone concluded his paper by referring to Skyrmes 1959 work, so
possibly it was that text that prompted him to connect nonperturbative eects to the
nonzero vacuum expectation value of a symmetry-breaking scalar eld. In his 1959
paper Skyrme did not quote Schwingers paper from 1957, but only referred to his
own previous publication (Skyrme 1958), and this fact may suggest that Goldstone
was at the time not aware of the similarities between his model and Schwingers one.

28

The European Physical Journal H

Fig. 2. The two-well potential.

Like Nambu had done, Goldstone employed diagrammatic methods from solidstate physics to show that, for 20 < 0, his model admitted at least one superconductor solution, i.e. a nonperturbative solution containing massive scalars and fermions,
which was present also when no mass term had been introduced in the bare Lagrangian
(Goldstone 1961: 159161). As already noted, Goldstone could reach this conclusion
by assuming that the scalar eld had non-vanishing vacuum expectation value, but
made no hypotheses as to how it would arise. He also said nothing on whether the
scalar eld was a composite of some more fundamental ones. At the end of the paper
he tentatively oered a rationale for the existence of the vacuum expectation value
using the example of a scalar eld in two dimensions subject to a two-well potential,
i.e. a potential function having a maximum in 0 and two minima placed symmetrically to the left and to the right of the zero point (Fig. 2) (Goldstone 1961: 162). This
model is today the paradigmatic example for spontaneous symmetry breaking, with
a Lagrangian of the form (Goldstone 1961: 162)26 :
L = 1/2(/x /x 20 2 ) 1/240 4

(2)

where 20 < 0 and 0 > 0. The minima of the potential corresponded to 2 = 2 =


620 /0 , i.e. = +/. Goldstone noted that, when a classical eld with a potential
of this kind was quantized, either one or the other minimum had to be chosen as
the vacuum state upon which particle states would be constructed. The choice of
one vacuum or the other formally corresponded to shifting the scalar eld by the
constant + or , rewriting the Lagrangian accordingly. Choosing one or the other
vacuum led to two dierent Lagrangians and therefore to completely distinct sets of
particle states which did not communicate with each other, as had been the case in
Nambu and Jona-Lasinios work:
the theory has two vacuum states, with a complete set of particle states
built on each vacuum, but [. . . ] there is a superselection rule between these
two sets so that it is only necessary to consider one of them (Goldstone 1961:
163).
26

See for example (Itzykson and Zuber 1984: 521523, Peskin and Schroeder 1995: 348
349).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

29

Fig. 3. The Mexican hat potential (Rupert Millard, Mexican hat potential polar, http://
commons.wikimedia.org/wiki/File:Mexican hat potential polar.svg, last accessed on August
24th 2014).

This argument was non-compelling, since Goldstone only showed how an unstable
classical system could be quantized in dierent way, and not how a nonzero vacuum
expectation value would arise in the rst place in a quantum system. However, for the
further development of the subject the two-well potential and its three-dimensional
version, the Mexican hat potential (Fig. 3) would provide a rhetorically and intuitively poweful means of exemplifying how a symmetry might break down without
external intervention. In the case of superconductivity, of course, the symmetry breaking followed from nonperturbative eects linked to the uctuations of electromagnetic
interactions, so that the physicists choice of one vacuum instead of the other stood
for a natural phenomenon. In the particle case, though, the situation was less clear
and, through the entanglement of dierent mathematical formalisms (perturbative
and nonperturbative) and physical frameworks (particles and superconductors), ambiguities in the physical status of the various mathematical components of the models
emerged. At the end of the paper, Goldstone added a further intuitive, physical picture
to his symmetry-breaking scheme, which reminded of the one Skyrme had suggested:
A simple picture can be made for this theory by thinking of the two
dimensionnal vector 1 at each point of space. In the vacuum state the vectors
have magnitude and are all lined up (apart from the quantum uctuations).
The massive particles 1 correspond to oscillations in the direction of . The
massles particles 2 correspond to spin-wave excitations in which only the
direction of 1 makes innitesimal oscillations. The mass must be zero, because
when all the (x) rotate in phase there is no gain in energy because of the
symmetry (Goldstone 1961: 163).
Since massless particles like the one described here had never been observed, they
constituted a serious problem for those who wanted to employ the model for phenomenological purposes:
A method for losing symmetry is of course highly desirable in elementary
particle theory but these theories will not do this without introducing nonexistent massless bosons [. . . ] Skyrme (1959) has hoped that one set of elds
could have excitations both of the usual type and of the spin-wave type,

30

The European Physical Journal H

thus for example obtaining the -mesons as collective oscillations of the four
K-meson elds, but this does not seem to be possible in this type of theory.
Thus if any use is to be made of these solutions something more complicated
than the simple models considered in this paper will be necessary (Goldstone
1961: 164).
This nal statement was the rst formulation of the so-called Goldstone theorem according to which a symmetry reduction of the superconductor kind would always be
accompanied by the appearance of massless bosons (Goldstone bosons). Goldstones
paper started a discussion among theorist on whether his theorem should be regarded
as valid always, never or only under certain conditions (e.g. explicit Lorentz invariance). Unsurprisingly, one of the rst authors to address the issue was Salam himself.
Already when presenting his and Wards model at the Berkeley conference Salam had
been aware of its possible connection to the proposal by Nambu and Jona-Lasinio,
as can be read in the conference proceedings. At the end of his talk, Gerald Feinberg
(193392) from Columbia University asked from the audience:
With regard to the K 0 you have indicated that the vacuum expectation
value comes from the weak interaction. It may be an interesting alternative
to follow a line of argument based on an idea introduced by Nambu, in which
one takes the strong interaction you have written there, and considers the
possibility that a vacuum state for the interaction is not invariant under the
strangeness transformation (which is certainly in principle possible) (Salam
1961: 430).
Salam replied: This is a type of idea which we have also considered, and it might lead
to a computation of the weak coupling constant (Salam 1961: 430). When Goldstones
paper appeared, Salam realised its signicance for the Salam/Ward model, since the
Goldstone boson problem would have put an end to their hopes of giving mass to
gauge vector boson by means of the vacuum expectation value of a scalar eld. In
1962 Salam teamed up with Goldstone and with Steven Weinberg to explore more in
detail Goldstones hypothesis in a paper entitled Broken symmetries (Goldstone,
Salam and Weinberg, 1962). In this paper they oered three proofs of Goldstones
theorem, yet recognized that they were all non-rigorous and therefore non-compelling
(Goldstone, Salam and Weinberg, 1962: 969). The paper referred to broken symmetries and to the Salam-Ward model, setting the ideas of Nambu and Jona-Lasinio in
direct connection with an already established object of research: the scalar -meson,
its vacuum expectation value and the generation of vector boson masses. Through
this connection Goldstones theorem was framed as a problem standing in the way of
a dynamical explanation for particle masses, and especially for those of gauge vector
bosons. In the following years, many theorists would attempt to solve the problem
by questioning both the validity and the premises of Goldstones theorem. Before we
turn to these well-known works, it is necessary to look into a paper which is today
less renowned, but constituted a seminal contribution to the debate on the origin of
mass and other particle properties.

12 Marshall Baker and Sheldon Glashows spontaneous


symmetry breakdown
The Higgs mechanism is today usually presented as an instance of spontaneous symmetry breaking, a mathematical construct which may be employed to model a number
of physical phenomena, such as ferro-magnetism or superconductivity. From the mathematical point of view spontaneous symmetry breaking is all but trivial and it is only

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

31

relatively recently that a deeper understanding of its possible deniton and properties
in classical and quantum systems has been achieved (Strocchi 2008). Moreover, even
if a rigorous mathematical denition is given, it still is not fully clear whether that
denition is applicable to the case of the breakdown of local gauge symmetry. From
the physical point of view, the situation is also unclear: in the case of ferromagnetism
or superconductivity we are obviously dealing with a physical transition from one
state to the other, but no observable process exists in which particles become massive. It is not my intention to discuss here these complex issues, but I mention them to
underscore how, even today, both the mathematical and the physical characterization
of the Higgs mechanism as spontaneous symmetry breaking remains problematic. In
the 1960s the situation was even more complex, since no rigorous mathematical denition of a third option between a symmetrical and an asymmetrical system existed
yet, while speculations on how observed approximate symmetries might arise from
a deeper, symmetric level of nature were quite fragmentary. However, it was in this
context that both the general mathematical notion of spontaneous symmetry breaking and the Higgs mechanism as an (alleged) spontaneous breakdown of a local gauge
symmetry arose. My dicussion will of course focus only on the latter developments,
but to understand their signicance it is necessary to rst sketch the development of
the more general notion of spontaneous symmetry breaking.
The origin of spontaneous symmetry breakdown is usually seen in Nambus work,
but, as we saw above, Nambu never conceived his reections as partaining to a symmetry breakdown, but rather as examples of how a symmetry might manifest itself
in an apparently non-symmetrical way. Moreover, his and Jona-Lasinios proposal for
mass generation was in physical terms nothing but a composite particle model. The
main role of Nambus model was to provide a physical and mathematical template
upon which bolder reections would built. Equally important for further developments were the ideas expounded by Marshall Baker and Sheldon Glashow in a paper
in which they introduced the term spontaneous symmetry breakdown (Baker and
Glashow 1962). For them, spontaneous symmetry breaking was primarily a physical
notion which could be grasped and formalized in terms of nonperturbative mathematical eects analogous to those treated by Nambu or Heisenberg. Baker and Glashow
had a very ambitious agenda:
Should not the complexities of the phenomena of elementary particle
physics likewise arise from a simple fundamental theory? Such a possibility was discussed by Heisenberg and co-workers [. . . ] Since nonperturbative
solutions to nonlinear equations do not in general possess the symmetry of the
equations themselves, it is conceivable that the eld equations may be highly
symmetric expressions, while their solutions may reect the asymmetries of
nature. This is the philosophy we adopt in this paper. This idea has been
developed along dierent lines in the work of Nambu and Jona-lasinio and
Goldstone upon which this work is based (Baker and Glashow 1962: 2462).
Accordingly, the authors set out in search of nonperturbative symmetry-violating
solutions to a symmetric eld theory (Baker and Glashow 1962: 2462) and stated:
We propose that a nonperturbative behaviour characterizes all the interactions to which elementary particles are subject. Mass is completely dynamical;
mass splittings and approximate symmetries result from nonsymmetric solutions to a fully symmetric Lagrangian theory (Baker and Glashow 1962:
24622463).
Baker and Glashow did not aim at proving this bold statement, but only
at demonstrating its plausibility by writing equations regulating the behaviour of the system and then showing that nonperturbative solutions existed

32

The European Physical Journal H

(Baker and Glashow 1962: 2463). Having done so, they went one step further and
asked why the nonperturbative, nonsymmetric solutions should be physically realized
instead of the perturbative, symmetric ones. Their tentative answer was: Perhaps
the symmetric solutions may be eliminated by some kind of stability criterion, but
we have not answered the questions (Baker and Glashow 1962: 2466). Stability was
thus the mathematical characteristic verbally associated with nonperturbative solutions, but Baker and Glashow described these solutions also as resulting from a
spontaneous breakdown of symmetry (Baker and Glashow 1962: 2466, 2470). The
word spontaneous expressed here the fact that the equations of the system, as was
the case in Nambus scheme, took the form of self-consistency conditions and that
therefore in some sense they determined themselves. This verbal component to the
theory was essential for its rhetoric strength and would eventually be the only element
of the construct to prove successful, as it expressed well the vague, but fascinating
idea of a symmetry which broke down of its own free will (whatever that might be
taken to mean in physical sense) and not due to external causes. Using simple models
Baker and Glashow argued that the stable, asymmetric solutions of state equations
were those which satised equations where the bare masses and coupling constants
were set to zero, so that observed masses and coupling constants were purely due to
nonperturbative, dynamical eects:
The requirement [. . . ] imposes the condition that the allowable interactions are just those that can arise spontaneously without the introduction of
the arbitrary parameters f and g [i.e. coupling constants]. This treatment of
the vertex function is analogous to our previous discussion of the self-energy
function. There, we used the condition m0 = 0 as a requirement that all masses
arise self-consistently [. . . ] we have shown the possibility that the fundamental
interactions can generate themselves from a bootstrap mechanism in a theory
where the bare coupling constants vanish (Baker and Glashow 1962: 2470).
These statements on spontaneity and self-generation cannot be seen as non-technical
characterization of well-dened mathematical properties, since no such properties were
described in the paper, but only shown to be possible through very simple, nonphysical examples. The characterization of a process as spontaneous emergence of
physically signicant numerical values of masses and coupling constants was thus an
independent component of the theory. The reference to the bootstrap is here of
great importance, as it indicates one of the main sources of inspiration for Baker and
Glashows work: Georey Chews bootstrap theory27 .
The premises of Chews approach were quite dierent from those of Baker, Glashow
and other quantum-eld-theorists, since Chew rejected the hypothesis of a two-level
structure of nature in which some particles were elementary and others were composite: all particle states, he claimed, were to be understood as composites of each other.
Particle phenomena should be modelled in terms of a formal construct (S-matrix)
comprising all scattering and decay amplitudes and determined by nonperturbative
self-consistency conditions in which no arbitrary parameters appeared. Chew took
up and expanded Heisenbergs earlier reections on the S-matrix, but his ideas took
their strenght from the recent successes of computational techniques which could not
be derived from perturbative quantum eld theory. Although his idealized picture
of a bootstrap theory was never fully realized, in the early 1960s his ideas found
many followers. His work displayed a very powerful story according to which all functions and parameters describing particle phenomena generated themselves in an act
27

The earliest papers in which Chew and his collaborators formulated their idea were
(Chew et al., 1957a, 1957b) For a historical overview and further references see (Chew 1989,
Kaiser 2005: 306355, Pickering 1984: 7378), on which the following overview is based.

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

33

whose paradox character was made rhetorically appealing by calling it a bootstrap,


a term which was becoming popular in computer science in the late 1950s (OED
1989). Baker and Glashow took quantum eld theory as their conceptual framework,
made use of nonperturbative techniques and also exploited Chews story: the analogy
to Nambus interpretation of superconductivity prompted to speak of the passage
from hypothetical symmetric equations to hypothetical asymmetric solutions as representing a physical process in which elements of reality generated themselves. Thus,
particle states became the physical agency behind their own coming-to-be, and the
symmetry breakdown was spontaneous. The relationship between mathematical elements and story was in Baker and Glashows work particularly complex, since the
passage from an equation to one of its solutions is not itself a mathematical structure,
but rather a procedure performed by the persons making the computation. . . or, as in
this case, by nature.
Despite its lack of concrete mathematical contributions, Baker and Glashows paper represented a turning point in the development of the notion of spontaneous
symmetry breaking and of the Higgs mechanism. The two authors not only came up
with a tting name for a physical notion of not-quite-symmetry which many other
theorists had been variously envisioning, but also spoke out boldly their conviction
that the spontaneous symmetry breakdown constituted a key feature of particle phenomena. They gave a detailed, plausible and rhetorically impressing expression of the
notion that a relatively simple mathematical construction might express an unknown
physical agency, to use Schwingers words, which further studies might lead to know
better.

13 The rise of spontaneous symmetry breaking in the early 1960s:


an outline
Baker and Glashows paper was quoted increasingly often in the following couple of
years and the term spontaneous symmetry breakdown in the course of the 1960s
came to be used to subsume both generic schemes where the symmetry breaking
involved multiple vacua and models in which the breakdown was specically due to
a scalar particle with nonvanishing vacuum expectation value. These theories could
be conceived as composite-particle models, as Nambu and Jona-Lasinio had done, or
understood in the spirit of the bootstrap theory as describing how particles generated
themselves. There is no historical study on the emergence of spontaneous symmetry
breaking beyond the specic issue of the Higgs mechanism and it is not possible to ll
here this gap, but in a rst approximation it may be said that, in the early 1960s, there
were two distinct groups working on broken symmetries in Baker and Glashows sense,
often, but not always, using the term spontaneous symmetry breaking. On the one
side were particle theorists who tried to develop phenomenologically plausible models
of how approximate symmetries would emerge from fundamental symmetries; on the
other side were mathematical physicists who attempted to give a rigorous denition
of spontaneous symmetry breaking which went beyond the mathematically trivial
statement that a symmetric equation can have non-symmetric solutions. Interestingly,
the latter group included a large number of Japanese theorists, as well as a number
of authors from Eastern Europe and the Soviet Union, but not many U.S. Americans
and Western Europeans, except for those working on axiomatic eld theory, who were
trying to formulate a mathematically rigorous version of quantum eld theory. Among
phenomenological papers prominently using the expression spontaneous symmetry
breaking in their title we may note, beside Glashows (1963) spontaneous breaking of
meson octet symmetry, a model explaining the muon-electron mass dierence through
spontaneous symmetry breaking (Arnowitt and Deser 1965), a theory of CP-violation

34

The European Physical Journal H

(Marx 1965) and a non-linear spinor model for meson masses (Byrne et al. 1965).
As to the physical mathematical contributions, the earliest one was probably due
to Mahiko Suzuki, who in 1963 published a paper on Foundation of spontaneous
breakdown of symmetry. At the beginning of the paper Suzuki summed up the state
of the art in these words:
The existing particles and resonances appear to exhibit themselves in various kinds of elaborate symmetries. Many physicists have proposed a variety of
symmetries to which they are to be assigned. In all symmetry schemes, however, the introduced symmetries must be broken subsequently, that is, eective
symmetry-breaking interactions need to be added, in order that they may be
realistic. [. . . ] What is the origin ot the symmetry-breaking interactions? Why
are the symmetries contaminated? An answer to these questions has recently
been proposed in the quite charming form, which is the so-called spontaneous
breakdown [Baker and Glashow 1962, Glashow 1963]. According to it, symmetries fail without any symmetry-breaking interaction. It predicts that the
physical world can be in lower symmetries though the total Hamiltonian enjoys higher symmetries. [. . . ] In the usual perturbation solutions symmetries
in Hamiltonian come into appearance at every step; they are never broken.
However, it is not necessarily the case in the nonperturbative case, where the
ground states need not have the original higher symmetries. There remains
room for the fact that the symmetries are apparently violated in the stable
physical world (Suzuki 1963: 627).
Suzuki here brought together the various verbal and phenomenological characterizations used by dierent authors: the approximate symmetries of particle multiplets,
the role of nonperturbative eects and of the ground state, the idea of broken,
even contaminated symmetries, but also of an apparent symmetry violation and
the stability of physically relevant solutions. Moreover, his discussion conated the
physical and mathematical aspects of the matter, as he referred both to the physical world and its total Hamiltonian, suggesting that the symmetries of the latter
may have physical relevance although they are not realized in observable phenomena.
This passage sums up well the ambiguous attitude of particle physicists at the time
(and possibly also today) with respect to what should be regarded as symmetries
of nature, namely the conviction that an observed approximate symmetry of masses
or other particle properties should be regarded as the result of a symmetry breakdown whose origin is to be sought. However, Suzuki made clear that he was not
interested in formulating models for specic spontaneous symmetry breakdowns, but
rather wanted to explore the general mathematical questions associated to it:
here we have no interest in the trivial mass splitting caused by particle
mixture. The discussion does not depend on the detail of the model (Suzuki
1963: 627).
In 1965 Ray Streater oered a discussion of how to dene spontaneous symmetry
breaking in axiomatic eld theory. He began the paper by stating:
The term spontaneous breakdown of symmetry (Baker and Glashow
1962; Glashow 1962; Suzuki 1963) has come to mean a eld theory whose
Lagrangian is invariant under a certain transformation of the elds, whereas
there exist solutions, i.e. realizations of the algebra of operators, that do not
possess the symmetry as a unitary transformation. Our problem is to formulate
this in the Wightman framework [. . . ] and also in the Haag-Araki framework
[of axiomatic eld theory] (Streater 1965: 510).
We note here how Streater did not speak of physical states, but of the realizations
of the algebra of operators. He then discussed at length how this condition might

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

35

be formulated in the two frameworks of axiomatic eld theory (Wightman and HaagAraki) and how the two formulations might be shown to be equivalent. We do not need
to go into this quite complex, technical discussion, but we must note the following
statements, which referred to the Goldstone theorem:
In fact, Goldstones theorem [. . . ] says that if a symmetry (of the
Lagrangian) leads to no conserved quantity, and if the eld is transformed
by the generator of the symmetry to a new eld with a dierent vacuum expectation value, then there must in theory be particles of zero mass. This result
is surprising because it is remarkable tht an algebraic property of A should lead
to the statement about the spectrum U (a) [i.e. the particles contained by the
theory], which is not in the algebra of quasi-local operators. Moreover, there
is in the literature a model (Schwinger 1963) of electrodynamics with gauge
invariance, but no zero mass particles. The explanation of this (Guralnik 1964)
is that in theories with indenite metric (such as Schwingers model) a pole in
the propagator does not imply that there is a zero mass particle in the theory
(Streater 1965: 511).
Streater was writing in 1965, after all the papers to be discussed in the following
sections had already been published, establishing the fact that local gauge invariance could be spontaneously broken without giving rise to Goldstone bosons. It is
interesting to note that Streater here credited the solution of the Goldstone problem
to Guralnik 1964, which in the bibliography is revealed to be a private communication (Streater 1965: 517). In his recollections on the subject, Guralnik indeed
explained how he had often discussed this issue with Streater in 1964, stating that
Streater told him at the time that he did not believe in spontaneous symmetry breaking, and that it was Guralnik himself who convinced him to change his mind on the
subject and devote some eort to formulating it in terms of axiomatic eld theory
(Guralnik 2009: 26132614). Finally, in view of later developments it may be noted
that already in 1965 the Hungarian physicists Gabor Domokos and Peter Suraniy
published a paper on the renormalizability of Lagrangians with spontaneous symmetry breaking, arguing that, if the symmetry Lagrangian was renormalizable, so was
also the spontaneously broken one (Domokos and Suranyi 1965 (in Russian), English
translation 1966).
The few sketchy remarks oered above cannot substitute an in-depth study of
the emergence of spontaneous symmetry breaking, but they hopefully suce for outlining the context in which specic discussions on the validity and applicability of
Goldstones theorem took place. The relevant papers and the technical arguments
brought forward in them have often been expounded in the historical and scientic
literature, in most detail in (Karaca 2013), and I shall not address neither those technical issues nor the question of priority, and will instead focus on showing how those
discussions were embedded in the broader discourse on the origin of mass developed
in the previous years and were motivated by the conviction that the relatively simple,
non-symmetric models studied nonetheless represented a physical process in which a
fundamental symmetry was broken by as-yet-undetermined, nonperturbative eects.

14 Goldstone bosons and local gauge invariance: nonperturbative


approaches
Although nonperturbative eects were central to the notion of spontaneous symmetry
breaking as understood in the early 1960a, only a few authors actually tried to employ
nonperturbative techniques for addressing the issue of the validity of the Goldstone
theorem in the case of local gauge invariance. Among them was Schwinger, who, in

36

The European Physical Journal H

two short papers published in 1962, argued that the gauge invariance of a vector eld
does not necessarily imply zero mass for an associated particle if the current vector
coupling is suciently strong (Schwinger 1962a: 397, 1962b). Using nonperturbative
techniques he could plausibly demonstrate that, if the couplings were strong enough
(and the perturbative expansion therefore invalid), masses for the vector boson might
arise dynamically, as he had already postulated in 1957. He only discussed simple,
non-physical models, but embedded his results in his broader narrative of dynamical
eects:
[T]he essential point [of the computation] is embodied in the view that the
observed physical world is the outcome of the dynamical play among underlying
primary elds, and the relationship between these fundamental elds and the
phenomenological particles can be comparatively remote, in contrast to the
immediate correlation that is commonly assumed (Schwinger 1962a: 398).
Schwinger did not refer to the Goldstone theorem and never spoke of symmetry breaking, but rather explained how a symmetry of the Lagrangian might have unexpected
phenomenological manifestations which deserved closer investigation. The following
year the solid-state physicists Philip Anderson showed how realization of Schwingers
ideas could be found in the theory of collective excitations of electron gas (plasmons) (Anderson 1963). Like Schwinger, Anderson, too, never spoke of broken
symmetries, but pointed out that a series of models of solid-state systems could be
regarded as exemplifying the violation of the Goldstone theorem: crystals, liquid helium, ferromagnetism and superconductivity, and that in all those cases a local gauge
invariance was involved whose vector bosons acquired mass:
I should like to close with one nal remark on the Goldstone theorem.
This theorem was initially conjectured, one presumes, because of the solidstate analogs, via the work of Nambu and of Anderson. The theorem states,
essentially, that if the Lagrangian possess a continuous symmetry group under
which the ground or vacuum state is not invariant, that state is, therefore,
degenerate with other ground states. This implies a zero-mass boson. Thus,
the solid crystal violates translational and rotational invariance, and possesses
phonons; liquid helium violates (in a certain sense only, of course) gauge invariance, and possesses longitudinal phonons; ferro-magnetism violates spin
rotation symmetry, and possesses spin waves; superconductivity violates gauge
invariance, and would have a zero-mass collective mode in the absence of longrange Coulombe forces.
It is noteworthy that in most of these cases, upon closer examination, the
Goldstone bosons do indeed become tangled up with Yang-Mills gauge bosons
and, thus, do not in any true sense really have zero mass [...] We conclude,
then, that the Goldstone zero-mass diculty is not a serious one, because we
can probably cancel it o against an equal Yang-Mills zero-mass problem
(Anderson 1963: 441442).
Anderson was the rst one to point out that the formal analogy between superconductivity and particle theory could be extended also to other solid-state systems
notably ferromagnetism, which would later become the poster child of spontaneous
symmetry breaking (Peskin and Schroeder 1995: 347). These exemplars were crucial
for supporting stories of dynamical interplay and self-generation in high-energy
physics despite the absence of mathematical models implementing them. Today, they
form an essential component of the construct spontaneous symmetry breaking, supporting the physicists belief in the unity of nature (Karaca 2013). Unlike Anderson
or Schwinger, though, most authors followed Goldstones lead, black-boxing nonperturbative eects in a scalar eld with a nonvanishing vacuum expectation value and

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

37

studying only the perturbative consequences of this nonperturbative premise in the


case of local gauge invariant models (Englert and Brout 1964; Guralnik et al. 1964,
Higgs 1964a, 1964b). This approach, although limited in scope, had the advantage
that one could employ perturbative quantum-eld-theoretical techniques. Moreover,
Goldstones scalar eld could be connected to the -eld, already known as a possible formal explanation for particle masses and in particular for the masses of vector
bosons of weak and strong interactions.

15 Goldstone bosons and local gauge invariance: perturbative


approaches
The six scientists usually quoted as the main contributors to the construction of
the Higgs mechanism (Robert Brout (19282011), Francois Englert, Gerald Guralnik,
Carl Richard Hagen, Peter Higgs, Tom Kibble) all worked perturbatively on already
broken symmetries, and focussed on the question of the validity of Goldstones theorem in the case of local gauge symmetries and on the possibility of giving masses to
gauge vector bosons. Yet they were ready to embed their models in the broader narratives usually accompanying nonperturbative, dynamical computations. These stories
oered a needed motivation to regard the perturbative computations focussing on the
Goldstone boson as investigations of the dynamical origin of mass. In this context,
the term broken symmetry was not just an arbitrary label, but served to establish
that what was being studied was not simply a Lagrangian with low symmetry, but
rather the result of a physical transition from higher to lower symmetry in whose
context particles acquired mass. Francois Englert and Robert Brout spoke of theories with degenerate vacuum (broken symmetry) in which vector mesons do [. . . ]
acquire mass (Englert and Brout 1964: 321), but added:
This example [i.e. the one discussed in the paper] should be considered
as a rather general phenomenological model. As such, we shall not study the
particular mechanism by which the symmetry is broken but simply assume
that such a mechanism exists (Englert and Brout 1964: 321).
Thus, Englert and Brout assumed the existence of a mechanism breaking local
gauge symmetry, but stated that it was not their object of research. Thus, symmetrybreaking became formally an act performed by the scientists themselves when they
introduced the nonvanishing vacuum expectation value of the scalar eld: We shall
break the symmetry by xing  = 0 in the vacuum (Englert and Brout 1964: 321).
At the end of the paper, though, they wrote: We would like to emphasize that here
the symmetry is broken through the gauge elds themselves (Englert and Brout
1964: 321). In a later paper Englert, Brout and Thiry (1966) discussed more in detail
the perturbative properties of a quantum eld theory of scalar elds coupled to vector
mesons, and stated:
The answer to the mass problem of vector mesons has been suggested by
Schwinger ( 1962a, 1962b), who points out that for suciently strong coupling,
such mesons acquire mass through the existence of a certain sum rule. Another
possibility is that broken symmetry, either self-imposed through the solution
of self-consistent equations or externally imposed, leads to the desired mass
[. . . ]. In this paper we explore further the structure of vector meson theory
in situations of broken symmetry. For the most part, we situate ourselves in
the context of a self-consistently broken symmetry in the manner previously
developed in (Englert and Brout 1964) (Englert, Brout and Thiry 1966: 245).
The self-consistent breakdown was explicitly opposed to an externally imposed
one, underscoring an assumed qualitative dierence between the two. However, the

38

The European Physical Journal H

dierence was not spelled out and in this paper, too, a phenomenological approach
was pursued, where only the consequences of the self-consistent breakdown, but not
the breakdown itself, were discussed: Englert and Brout started from a Goldstonetype Lagrangian where a vacuum expectation value for the scalar eld had been put
in by hand as a place-holder for nonperturbative eects.
A similar strategy appears in the publications of Peter Higgs. In his st paper
Higgs spoke of a symmetry broken by the non-vanishing of the vacuum expectation
value of [a scalar eld] (Higgs 1964a: 133) and only discussed the formal validity of
the Goldstone theorem. In the following publication, though, he referred to spontaneous breakdown of symmetry and discussed how, for local gauge symmetries,
the spin-one quanta of some of the gauge elds acquire mass (Higgs 1964b: 508).
In this paper he discussed symmetries in a classical Lagrangian, showing how, by
performing a specially chosen local gauge transformation, mass terms for gauge vector bosons and a scalar particle appeared, while no massless scalar were present. No
demonstration was given of how the massive Lagrangian would physically arise from
the massless one. Nonetheless, Higgs, too, embedded his computation in a story of
transitions from massless to massive particles which he described as a spontaneous
symmetry breakdown, corroborating this view by referring to the mass acquisition as a
phenomenon which was the relativistic analogous of the plasmon phenomenon as
argued by Anderson (Higgs 1964b: 508). Guralnik, Hagen and Kibble instead focussed
more strictly on the details of the disappearance of the massless Goldstone bosons
in the case of a local gauge symmetry (Guralink et al. 1964). They did not mention
nonperturbative eects, a mechanism of symmetry breaking or the spontaneity of the
process, but they, too, described the subject of their research as a symmetry breaking
which obtained because of the introduction of a nonvanishing vacuum expectation
value, implying that the non-symmetric model was linked to a more fundamental,
symmetric one:
In summary, we have established that it may be possible consistently to
break a symmetry by requiring that the vacuum expectation value of a eld
operator be nonvanishing without generating zero-mass particles (Guralnik
et al., 1964: 587).
In other words: The formal step of imposing a nonvanishing vacuum expectation
value for the scalar eld was verbally presented as the breaking of a symmetry. The
physical nature of the discussion was suggested also by the explicit analogy to superconductivity which, the authors claimed, appears to display a similar behaviour
(Guralnik et al. 1964: 586). In a later paper Higgs spoke of a relationship of ancestry between the symmetric and the non-symmetric Lagrangian (Higgs 1966: 145) and
connected his results to Baker and Glashows story of symmetric equations having
asymmetric solutions, to the degeneracy of the vacuum, and to the -eld employed
by Schwinger and by Salam and Ward:
The idea that the apparently approximate nature of the internal symmetries of elementary particle physics is the result of asymmetries in the stable
solutions of exactly symmetric dynamical equations, rather than an indication of asymmetry in the dynamical equations themselves, is an attractive one.
Within the framework of quantum eld theory such as a spontaneous breakdown of symmetry occurs if a Lagrangian, fully invariant under the internal
symmetry group, has such a structure that the physical vacuum is a member of
a set of (physically equivalent) states which transform according to a nontrivial representation of the group. [. . . ] That vacuum expectation values of scalar
elds, or vacuons, might play such a role in the breaking of symmetries was
rst noted by Schwinger and by Salam and Ward (Higgs 1966: 145).

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

39

This passage is a particularly eloquent exposition of the notion that, despite the
observed variety of particle phenomena, a fundamental, symmetric level of interactions
might in the future be grasped using techniques dierent from those of perturbative
quantum eld theory. The scalar eld (later known as the Higgs eld) was then seen
as only a temporary, approximate representation of those eects. In fact, the presence
of a massive scalar particle in the theory was at the time not regarded as a new
feature, since the -meson had already been present in many previous models, as we
saw, and various candidates for its identication had been proposed. In this sense,
it is a clear misinterpretation of historical sources to claim that Peter Higgs (or any
other author in 1964) was the rst one to predict the existence of a massive scalar
particle associated to the breakdown of local gauge symmetry. Yet, as we shall see in
Section 17, this misinterpretation is often found in discussions on the history of the
Higgs boson.
In conclusion, all authors investigating perturbatively the Goldstone theorem
started from a Lagrangian which was not local gauge invariant and claimed that
it was the result of a symmetry breaking whose outcome was formally represented by
the nite vacuum expectation value of a scalar eld. The conviction that the scalar
represented a spontaneous breakdown of a symmetry existing at some more fundamental physical level would later lead Salam and Weinberg to postulate that the
scalar eld would also be able to ensure the renormalizability of the Lagrangian whose
symmetry it broke.

16 Steven Weinberg and Abdus Salams hopes and their realization


By 1965 it had been established that the Goldstone theorem did not prevent theorists
from using spontaneous symmetry breaking to construct a physically viable theory of
particle interactions based on the local gauge principle, but containing gauge boson
masses. The motivation for constructing such a theory was not simply a desire to
connect observed phenomena to a hypothetical, fundamental, symmetric theory, as
had been the case in Schwingers or Heisenbergs original work. The reason why local
gauge theories appeared to some authors as a promising canditate to model particle
interaction was that there were good arguments to believe that they would be renormalizable like quantum electrodynamics28 . However, gauge boson masses, which were
necessary to t experimental data, spoiled the symmetry, and so it seemed probable
that they would also make the theory non-renormalizable. While most theorist at the
time regarded spontaneous symmetry breaking as a possible means to explain mass
dierences within particle multiplets or other observed approximate symmetries, in
19671968 Salam and Weinberg independently suggested that spontaneous symmetry breaking through a scalar might ensure the renormalizability of a unied gauge
theory of electromagnetic and weak interactions. Although they could write down a
phenomenologically plausible Lagrangian (the later Weinberg-Salam model), neither of them had a mathematical argument to oer in favour of the renormalizability
claim: Salam stated that the (Higgs) mechanism would ensure that the breakdown occurred more gently than a brutal addition and subtraction of mass terms (Salam
1968: 369), while Weinberg asked Is this model renormalizable? and answered that
the vector bosons get their mass from the spontaneous breaking of the symmetry,
not from a mass term put in at the beginning, thus expressing the hope that the
28

Early local gauge theories of particle interactions were proposed among others by Yang
and Mills (1954), Ronald Shaw (unpublisehd PhD thesis 1955), Sidney Bludman (1958),
Sheldon Glashow (1961) and, of course, by Salam and Ward in the papers discussed in
section 8 above. For an overview on the history of gauge theories in particle physics and the
question of their renormalizability see (Pickering 1984: 159180).

40

The European Physical Journal H

spontaneity of the symmetry breakdown would somehow guarantee renormalization


(Weinberg 1967: 1266). Both authors characterized the symmetry breaking through
the scalar eld negatively, i.e. as an alternative to explicit breaking, and employed anthropomorphic, emotion-laden terms (gently, brutal, spontaneous) to describe
the dierence which would hopefully lead to renormalizability. Clearly their conviction was based only the rhetorically powerful story of a third path to symmetry
attached to the scalar eld with nonzero vacuum expectation. They did not quote
the work on renormalization and spontaneous symmetry breaking by Domokos and
Suranyi (1965/1966) and there are no sources to denively establish whether they
were aware of it or not. Neither Salam nor Weinberg refer to it in their Nobel lectures, in which they provide a rather detailed account of the authors whose work
they used (Salam 1979, Weinberg 1979). However, it is not implausible that they had
run across the paper by Domokos and Suranyi, since it had already been quoted by
Western authors, and in particular their results featured prominently in a paper by
Brout on Group theory of the possible spontaneous breakdown of SU(3) published
in early 1967 (Brout 1967).
However, no theorist seemed to have much interest in nding a mathematical support for the narrative of renormalizability. For example, Englert, Brout and Thirys
paper (1966) contained some relevant results, but they did not pursue the matter further. At the 1967 Solvay Conference Englert made Weinberg aware of this work, but
Weinberg did not follow up on it (Englert in Brown et al. 1997, 495496; Weinberg,
Englert in D
urr et al. 1968, 1819). Beside the technical diculites, the lack of interest
was also due to the fact that the Weinberg-Salam model had a limited phenomenological scope, as it could explain neither the approximate symmetric patterns observed in
particle properties, nor the general features of weak interactions of strange particles,
as for example the selection rule (isospin) = 1/2 (which incidentally remains unexplained until today). The Weinberg-Salam model fell short of the high expectations
raised by narratives such as those of Schwinger, Heisenberg or Baker and Glashow,
and in the end remained as unnoticed by the research community as Benjamin Lees
proof of the renormalizability of Gell-Mann and Levys -model (Lee 1966). The situation changed only in 1971, when Gerhard t Hooft proved the renormalizability of
the Weinberg-Salam model by building upon the work by Lee and others (t Hooft
1997, Cao in Brown et al. 1997: 502503). Although t Hooft was at the time not
aware of Weinbergs and Salams papers, his work soon came to be regarded as a
mere technical conrmation of what Weinberg and Salam had allegedly known all
along, underscoring the primary epistemic role of verbal statements for the scientic
community: it was not the mathematical proof of renormalizability, but rather the
verbal claim of its existence that was regarded as the seminal moment when the unication of electromagnetic and weak interactions came to be. After t Hoofts initial
proof was made complete by other authors, interest in the Salam-Weinberg model
was awakened and experimental conrmation for its key predictions was found. In
1979 Glashow, Weinberg and Salam were awarded the Physics Nobel prize for their
contributions to the theory of the unied weak and electromagnetic interaction between elementary particles, including, inter alia, the prediction of the weak neutral
current, and it was only in 1999 that also t Hooft and Martin Veltmann received
the same honor for elucidating the quantum structure of electroweak interactions in
physics (Nobel prize 1979, 1999).

17 Whats in a name? Disputes on the Higgs boson and the Nobel


prize at the turn of the 21st century
The task of a historian of science is not to dene discoveries or assign priority, but
to try and oer a broad picture of how knowledge is constructed, and this is what

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

41

I have attempted to do in the previous pages for the notion of mass generation
with the help of a scalar boson. I have argued that the special status which the Higgs
boson has in todays physics as the particle that gives mass to all others cannot be
explained as the sum of single theoretical discoveries, but emerged from a broad,
collaborative network in which a large number of physicists pursuing similar, yet
not necessarily identical aims, appropriated and built upon each others work. The
collective character of theoretical research is rarely if ever acknowledged, but in the
case of the mass generation mechanism I believe it can be shown how the community
of high energy theorists perceived that the merit for its discovery had a shared,
essentially indivisible nature. This perception led to strong tensions when the time
came for a political decision on the awarding of the Nobel prize, which in the eyes
of the broader public is today a sort of quality seal for ground-breaking scientic
results and has a meaning for the further development of science which goes well
beyond the honor and nancial help bestowed on its recipients.
One of the earliest and most visible expressions of these tensions were the discussions on how the mechanism of mass generation and the relevant scalar boson should
be referred to, and in particular how correct of incorrect it was to name them only
after (Peter) Higgs, as had become common use by the late 1980s. Let us take a look
at the historical development of this issue. When the notion emerged in the early
1960, no special name was attached to it, not even in the papers by Weinberg (1967)
and Salam (1968) where it was used to introduce mass terms in the unied theory
of electromagnetic and weak interactions. When in the early 1970s the predictions
of the unied theory were experimentally conrmed, though, some authors started
referring to the mass-generation mechanism and to the relevant scalar boson with
the name of one or more theorists. Some authors, like Martin Veltman and Gerhard
t Hooft called it the Higgs-Kibble mechanism (t Hoof and Veltman 1972: 189)
and, indeed, Kibble was mentioned both by Salam and Weinberg as the one who,
in person or through his writings, had inspired their work on spontaneous symmetry
breaking (Close 2011: 286287, Salam 1979: 521). However, by the middle of the 1970s
the terms Higgs mechansim and Higgs boson had become a common shorthand
in research literature and the 1976 paper by John Ellis, Mary Gaillard and Dimitri
Nanopoulos suggesting an experimental search for the mass-generating boson had the
title a phenomenological prole of the Higgs boson (Ellis, Gaillard and Nonopoulos
1976). There is no specic reason why Higgs name and not that of other authors
became mainstream, and there are no indications that at the time this was seen as a
special recognition of Higgs contribution29 . In general one may say that, in the initial
decades of its existence, the issue of a proper attribution of credit for the construction of the mass generation mechanism was not specically addressed and was anyway
distinct from the question of its labelling in scientic and popular writing. However,
while in the 1970s the Higgs mechanism was regarded as one of many components of
the Standard Model, by the turn of the twenty-rst century the fact that the Higgs
boson had proved to be more elusive than originally expected and the absence of signs
of physics beyond the Standard Model led physicists to focus their attention on the
mechanism of mass generation and the relevant scalar boson. It was at that time that
the naming question came to be felt as increasingly delicate, and this was clearly due
to the fact that a name might be perceived as a problematic priority attribution.
Early discussions on the subject were often informal, but indications on how the
problem of a fair naming was felt as increasingly important can be found by looking at
29
A plausible story of how the boson got only Higgss name is given by Ian Sample in a
letter to Nature on the basis of recollections of involved physicists, where it appears that in
the early 1970s Benjamin Lee was the rst one to make prominent use of the term (Sample
2010b).

42

The European Physical Journal H

the history of the Wikipedia-entry on the Higgs mechanism (Higgs Mechanism Revision History 20042014). In its initial version (2004) the entry referred in its rst sentence to the Higgs mechanism, but in the following years the label of the mechanism
underwent a series of changes: the Higgs mechanism or Anderson-Higgs mechanism
(January 2006), the Higgs mechanism, Higgs-Kibble mechanism or Anderson-Higgs
mechanism (August 2007), the Higgs mechanism, also called the Brout-EnglertHiggs mechanism, Higgs-Kibble mechanism or Anderson-Higgs mechanism (October
2007), the Higgs mechanism, also called the Brout-Englert-Higgs mechanism, HiggsBrout-Englert-Guralnik-Hagen-Kibble mechanism or Anderson-Higgs mechanism
(July 2008). Interestingly, in the early phase there was no simple addition of names
to the label, but alternatives were proposed that reected dierent views on which
individuals had most contributed to the construction of the mechanism. Finally, in
October 2008 a rst paragraph cleanup removed all alternative names to a section
on History and naming. It is important to note that these changes were not the
result of uncontrolled ghts between individual Wikipedia-contributors, but were stable modications reecting a growing awareness of the importance of the name (and
priority) issue in the community. In general, there was at this time no attempt to
push the claim of specic discoverers, but rather interest in preventing the merit
being exclusively assigned to a small number of authors. In 2009 articles by Kibble
on the Englert-Brout-Higgs-Guralnik-Hagen-Kibble mechanism and on its history
were published in Scholarpedia, the scholarly, peer reviewed version of Wikipedia
(Kibble 2009a, 2009b). In the same year a Wikipedia contributor identifying herself
only as Mary at CERN initiated a new entry specically devoted to discussing the
three 1964 papers and this led to claims that she was trying to pinpoint the priority
attribution and requests that the article be removed (Carroll 2012: 240241, Talk
1964, PRL symmetry breaking papers 2009). At the same time, an increasing number
of theorists started referring to the mechanism in their papers with its most extended
label, and this trend went so far that in 2012, as the discovery of the Higgs boson
seemed (and indeed was) imminent, it prompted the publication of an editorial contribution to Nature whose single aim was to plead that, despite all priority issues,
physicists should stick to the name Higgs boson because it was a succesful brand:
Correct allocation of credit is important, and authoritative accounts of the
history of science are useful and enlightening, but both must be balanced with
sciences need for consistent conventions, brevity, and public communication
and outreach, especially when taxpayers expenditures, such as the US$ 6.5
billion for the Large Hadron Collider, are at stake. Renaming the Higgs boson
in the year when it is most likely to be found gets the balance wrong (Nature
Editorial 2012).
This article addressed a question at the core of the (science) political signicance of
the Higgs boson: the high costs of the LHC (and of its possible follow-up experiments)
and the ensuing necessity of conveying to the general public in the clearest possible
terms the deep scientic relevance of the results that the machine could deliver. The
Higgs boson appeared as a most successful brand to attain that aim. The start of
the LHC experiment had been accompanied by expectation of either discovery of the
Higgs boson or proof of its non-existence and, whatever the case, the perception of the
general public, shared and fostered by the physics community, was that such a discovery would be signicant enough to warrant the awarding the Nobel prize. In other
words, there were very strong political reasons making a Nobel prize for the Higgs
boson highly desirable, if not outright necessary, and in this sense the identity of the
scientists who might receive the prize in their hands was a secondary matter. The
problem, of course, was that the Physics Nobel prize may not be shared among more
than three persons, and this fact already ruled out the possibility of awarding it to

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

43

one of the experimental collaborations working at CERN, since the widely recognized
collective nature of experimental work in particle physics would not have allowed the
credit for any discovery to be assigned, even only symbolically, to any individual member of the experiment. Thus, only a Nobel prize for the theoretical discovery of the
mechanism appeared possible, although in that case, too, the exact motivation and
relevant attribution of credit promised to be a delicate matter. In 2008 seven names
were listed on Wikipedia as potential labels for the mechanism: Anderson, Brout,
Englert, Guralnik, Hagen, Higgs and Kibble, and this was already a small choice not
approved by all. Yet with the start of the LHC (and with it the possible discovery
of the Higgs boson) getting nearer, rst attempts were made to narrow down the
circle of possible discoverers to a number equal or less than three. Now the discussion
became both personal and political, with the technicalities involved in dening what
the priority claim should be about becoming only the occasional, contingent battleeld of a larger conict. In September 2008 the CERN Courier published an interview
with Englert, Brout and Higgs on the mechanism of mass generation, and this move
prompted a response by Guralnik, Hagen and Kibble, who complained that their own
work was being slighted and somehow suggested that an attempt to belittle it was
underway: For 40 years after the 1964 appearance of these three papers there was no
discernible pattern of preference among them, with the vast majority of researchers
in the eld mentioning all three. We strongly object to any downgrading of our contribution. (Cern Courier 2008a, 2008b, Close 2011: 140). In 2009 Guralnik wrote his
reminiscences on the History of the Guralnik, Hagen and Kibble development of the
theory of spontaneous symmetry breaking and gauge particles, explaining in detail
why the fact that the Guralnik-Hagen-Kibble paper quoted those by Higgs and by
Englert and Brout did not detract in any way from its originality(Guralnik 2009).
Guralniks remarks and the whole issue of the Nobel prize were taken up in the popular science book Massive by Ian Sample (Sample 2010a). Sample described the issue
of attribution of credit for the mass generation mechanism as a dicult question, but
still earned critique from physicist and science writer Frank Close for having given
Higgs too prominent a role and having perpetuated the shibboleth, that only the
three groups (Higgs, Brout & Englert, Guralnik, Hagen & Kibble) had contributed
to the construction of the Higgs boson, while the story was longer and more complex
(Close 2010). In particular, Close underscored how important the contribution by
Goldstone had been.
In 2011 Close published a history of the Standard Model up to and including the
search for the Higgs boson, which at the time was well under way. Again, he addressed
the priority question, pointing out the gap between the expectations of public and
press and the complex situation from the scientic point of view, and, while not using
the term collective, nonetheless noted how merit attribution in science often tended
to overlook the broader picture: [Nobel] winners were often those who were carrying
the baton over the nishing line after a long relay race (Close 2011: 139). He then
compared the emergence of the Higgs mechanism to a steeplechase, in which the hurdles had been erected by Goldstone, Salam and Weinberg, Schwinger and Anderson
had identied the means of clearing the obstacles, before several others reached the
tape in a photo-nish (Close 2011: 140). He also noted how the priority disputes surrounding the Higgs mechanism in the last few years had taken a sharper, personalized
tone and quoted the episode of the 2008 CERN Courier articles (Close 2011: 140).
Close expressed eloquently the misgivings of many physicists on this subject, but like
them stopped short of proposing that, due to the high number of scientists involved
in it, no Nobel prize should be awarded for the theoretical discovery of the mass
generation mechanism. Another physicist and science writer who discussed the Nobel
issue as problematic was Sean Carroll (Carroll 2012, 2013). In his popular science
book on the Higgs boson The particle at the end of the universe Carroll mentioned

44

The European Physical Journal H

the disputes in the physics community and regretted that, if the Higgs boson was
found, the rules prevented the Nobel prize being awarded to those who had the most
immediate claim to it: the experimental collaborations (Carroll 2012: 241, Carroll
2013). He wrote: Its probably past time for the Nobel foundation to think about
relaxing the tradition that collaborations cannot win any of the prizes in science.
Whoever gets that rule change implemented might deserve the Nobel Peace Prize
(Carroll 2012: 241). Making the Nobel prize rules more exible to accommodate experimental collaborations would be certainly a good idea, but what about theorists?
Would it be acceptable to the community that theoretical discoveries, too, should be
regarded as the results of a collective, if often uncoordinated eort, and not as the
sum of individual contributions which can in principle be evaluated each on its own?
Discussions among working physicists on the priority in the discovery of the Higgs
mechanism suggest that the notion of separable individual contributions is still very
much present, yet nonetheless perceived as problematic. This can be seen for example in the comments made on Peter Woits much-read high-energy-physics blog Not
Even Wrong in correspondence to the entry on The Anderson-Higgs mechanism
(Woit 2010). In that contribution, Woit underscored the role of Philip Anderson in
the development of the mechanism of mass generation, sparking a series of comments
in which various, more or less known, possible discoverers of the Higgs mechanism
and their merits were discussed, with commentators apparently trying to bring up as
many names as possible, noting that many more theorists than usually assumed had
written on the subject.
At the same time, on the political front, it was increasingly clear that, if a Nobel
prize for the Higgs boson had to come, Brout, Englert and Higgs appeared as the
smallest team for which a case could be made that might appear plausible to the
majority of the physics community. Yet the resistance was stronger than some would
have expected, as discussed in a short article in Nature by science journalist Zeeya
Merali (Merali 2010). Merali described a recent incident: in the announcement of a
physics meeting on Higgs hunting at Orsay, only Brout, Englert and Higgs had been
quoted, leading some physicist to complain and even threaten to boycott the meeting.
She quoted one of the organizers of the conference as saying that People took this
very seriously, which we didnt expect (Merali 2010). She related an interview with
John Ellis from the CERN Theory Division, who took a clear position on the priority
issue: Ellis advised the committee to stick to their guns, arguing that it is undeniable
that Guralnik, Hagen and Kibble were the last to publish. He also notes that their
paper cited the earlier publications by Brout and Englert, and by Higgs, weakening
their authorship claim (Merali 2010). However, she added: That argument carries
little weight with particle physicist Tom Ferbel at the University of Rochester in New
York, who says that Guralnik, Hagen and Kibble should not be penalized for citing
other papers out of professional courtesy (Merali 2010). Ellis acknowledged that the
issue was delicate because there was a Nobel prize at stake, but suggested that the
disputes had less to do with scientic merit and more with national pride, and Merali
quoted him as saying: There is a lot of fuss being made about Guralnik, Hagen and
Kibble right now, and the American physics community seems to be listening. Im
just glad that Im not on the Nobel committee deciding who to throw out of the
lifeboat (Merali 2010). Ellis dramatic reference to a lifeboat is very telling indeed:
the Nobel prize, important as it might be, was not a matter of life and death for any of
its possible recipients, but in the long run it might be such for the high-energy-physics
community and in particular for the LHC and CERN. It is thus rather unsurprising
that a long-standing member of CERN Theory division like Ellis would have rather
strong views on the subject, although this fact of course in no way implies that he was
not sincere in believing that Brout, Englert and Higgs had the best claim to become
Nobel laureates, and that only fuss was being made around Guralnik, Hagen and

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

45

Kibble. For him, it was clearly possible to have both a Nobel prize for the mechanism
of mass generation and a fair apportioning of credit, and the community should focus
on obtaining the prize and not on questioning the rules of the Swedish Academy of
Science.
From 2010 on, the political aspects of the Higgs discovery took more and more
centre stage and the priority dispute was framed in terms of whether and how a
plausible case for Brout, Englert and Higgs against Guralnik, Hagen and Kibble could
be made, and not whether a priority claim made any sense at all. In 2011 Ellis and his
CERN colleague Luis Alvarez-Gaum`e took up once more the issue in a short article in
`
Nature bearing the title Eyes on a prize particle (Alvarez-Gaum`
e and Ellis 2011).
Altough Ellis, as we saw above, had already made his views public on who had a
priority claim for the discovery of the mechanism, here the authors avoided taking a
denite position, but stated that in 1997 the European Physical Society Prize had been
awarded only to Higgs, Brout and Englert, alleging that this was because Englert and
Brout had been the rst ones to formulate a non-Abelian version of the mechanism,
while Higgs the rst one to draw attention to the existence of the massive scalar (Ellis
and Alvarez-Gaume 2011). It must be noted that these arguments are not found in
the motivation of the prize, which simply states: For formulating for the rst time a
self-consistent theory of charged massive vector bosons which became the foundation
of the electroweak theory of elementary particles (European Physical Society 1997).
Moreover, as we saw in the previous sections, both spontaneously broken local nonAbelian gauge symmetries and massive scalar bosons breaking them were already
present in earlier works (e.g. Salam and Ward, Schwinger, Anderson) and it is rather
arbitrary to pinpoint these, and not other features as dening characteristics of the
mechanism. It is also interesting to note that, if one considered the issue of publishing
priority important, then Englert and Brout would have a stronger claim than Higgs,
so that the fact Higgs had allegedly been the rst one to explicitly mention a massive
scalar with nonvanishing vacuum expectation value became particularly signicant. In
September 2011 Guralnik published the text of a talk where he once again presented
a history of the Guralnik-Hagen-Kibble paper, giving detailed arguments to counter
the claim that the paper only spoke of a massless boson (Guralnik 2011). He explicitly
stated that There has been considerable recent discussion about the relative merit
and content of the work of the three groups that rst analyzed spontaneous symmetry
breaking as now used in the unied theory of weak and electromagnetic interactions.
Some of this discussion has been quite misleading and even wrong (Guralnik 2011: 8),
although he quoted no specic reference on this point. Yet Guralnik was not ghting
simply against technical interpretations of his paper, but against the possibility that
the enterprise of the LHC be crowned by a Nobel prize, and so he was destined to lose.
In January 2012 Ellis once more weighted in on the subject, this time together with
Mary Gaillard and Dimitri Nanopoulos who, as we saw above, in 1976 had written
with him A phenomenological prole of the Higgs boson. The three now published
A historical prole of the Higgs boson (Ellis, Gaillard and Nanopoulos 2012). It is
interesting to note that, other than earlier contributions by Ellis, this paper contained
absolutely no explicit mention of the Nobel prize or of priority and presented itself
as a disinterested attempt to trace the developments in Higgs physics in the last
fty years (Ellis, Gaillard and Nanopoulos 2012: 1). However, when considering the
context of publication, the previous statements by Ellis and the fact that none of
the three authors had so far taken an interest in history of science beyond occasional
personal recollections, it is dicult not to regard the paper as contributing to set
the stage for a Higgs-Englert Nobel prize, Brout having in the meantime deceased.
Without acknowledging Guralniks arguments, the authors repeated Elliss previous
statements that Englert and Brout had been the rst to discuss the non-Abelian
version of the mechanism and Higgs the rst one to introduce the massive boson,

46

The European Physical Journal H

adding that the subsequent paper by Guralnik, Hagen and Kibble referred in its text
to the Englert/Brout and Higgs papers (Ellis, Gaillard and Nanopoulos 2012: 45). In
all of this, Peter Higgs was always very careful to avoid any controversial statement,
notoriously referring to the boson who was named after me and proposing that
it should be called ABEGHHtH boson from Anderson, Brouts, Englert, Guralnik,
Hagen, Higgs and t Hooft (Sample 2010b).
Eventually, a certain amount of political consensus was established in the physics
community that it would be acceptable to award the Nobel prize to Englert and Higgs
and, indeed, this is was happened in 2013. The relevant advanced information on
the ocial website of the Nobel Prize essentially took up the arguments by Ellis and
his colleagues, noting that Englert and Brout had proposed a non-Abelian model and
that Higgs had allegedly been the rst to note the presence of the massive boson, even
going so far as to claim that it was because of this that the boson was called after
him (Royal Swedish Academy 2013: 9). Guralnik, Hagen and Kibble only received a
mention as one of the further contributions (Royal Swedish Academy 2013: 1011).
In his Nobel lecture Englert referred to the BEH-mechanism, while Higgs carefully
avoided giving it any name at all, though acknowledging that Guralnik, Hagen and
Kibble published a little later (Englert 2014, Higgs 2014). Although no big public
disputes followed the awarding of the prize (hardly to be expected in a close-knit
community like the high-energy physics one) the fact that not everyone was happy
with this outcome can once again be seen in the comments on Woits blog entry
on the occasion of the 2013 Nobel prize show, where commentators regretted that
the award did not go to the experimental collaborations, noted again how many
theoretical contributions had been slighted and at times also criticized the advanced
information from the Royal Swedish Academy (Woit 2013). Woit also linked to a
Guardian article by experimentalist John Butterworth who criticized the lone genius
idea of science progress (Butterworth 2013). Although Butterworth was primarily
interested in speaking out in favour of awarding prizes to experimental collaborations,
he also noted: Even the theory breakthrough behind this prize [i.e. the 2013 Physics
Nobel prize] required a body of incrementally acquired knowledge to which many
contributed (Butterworth 2013).
In March 2013 a symposium was held at Imperial College, London, in honor of
Tom Kibbles 80th birthday and the lectures given in that occasion were published
in 2014 (Gauntlett 2014). Both Steven Weinberg and Frank Close contributed and,
although they did not mention the Nobel prize, they made their positions on the
priority issue quite clear. Close stated: It took six people, at least, to make the W
and Z boson massive (Close 2014: 134). Weinberg, describing how the problem of
massless Goldstone particles had been formulated and then solved, wrote: Other
[particles] show up as real physical particles of [. . . ] non-zero mass (because they
arent Goldstone particles). We needed a name for the latter massive particles, and
although they had already appeared in Goldstones 1960 work, we could hardly call
them Goldstone particles, so they came to be called Higgs particles, though they could
just as well have been called Brout-Englert-Guralnik-Hagen-Higgs-Kibble particles,
or just Kibble particles (Weinberg 2014: 130). Had the particle been called Kibble
boson, would the list of the 2013 Nobel prize winner have been dierent? Probably
yes, not because merit attribution is arbitrary, but because it is a primarily political
matter. Guralnik, Hagen, Kibble and their close colleagues and friends were of course
those who had the strongest perception of the 2013 Physics Nobel prize as unfair, but
in my opinion the tensions in the community cannot be simply explained as physicists
taking sides for one or the other camp, but rather as following from the perception
that in this case exclusive attribution of merit was not tting to the reality of research.

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

47

18 Concluding remarks
The story of the Higgs boson and its relationship to the question of the orign of mass
told in the previous pages started with a paper published by Schwinger in 1957, but
there is little doubt that also other, earlier or later points could have been chosen. My
aim was not to reconstruct how some specic mathematical construction of physical
notion came to be, but rather to show how a conviction central to todays particle
physics, namely that the Higgs boson has a special connection with the origin of mass,
did not arise as the result of some singular theoretical or experimental discovery, but
rather grew in time from the combination and positive feedback of a number of dierent, initially largely distinct research questions and methods. Some of these elements
are still today constitutive parts of the Standard Model, such as the scalar particle
with vacuum expectation value or the expressions spontaneous symmetry breaking
and mass generations. Other elements, such as the notion of a self-generation
of masses and particles, do not belong any more to mainstream high-energy physics,
although at the time of the emergence of the so-called Higgs mechanism they were
regarded as fully equivalent to the rst ones. As was the case in the 1960s, the special
status enjoyed today by the Higgs boson is primarily due to the relation of ancestry, to use Peter Higgs words, between the Lagrangian of the Weinberg-Salam model
and a Lagrangian invariant with respect to a local gauge symmetry and containing
no massive gauge vector bosons. Although today we understand in detail how the
renormalizability of the broken Lagrangian is connected to the renormalizability of
its ancestor, the unbroken one, it still remains open how and why the latter and not
the former one gives rise to observed particle phenomena, among them massive particles. However, future theoretical and experimental investigation of the Higgs boson
is regarded in todays high-energy-physics community as a promising path to learn
more about the origin of mass.

References

Alvarez-Gaum
e, L. and J. Ellis. 2011. Eyes on a Prize Particle. Nature Physics 7: 23
Anderson, C.D. 1932. The Apparent Existence of Easily Deectable Positives. Science 76:
238239
Anderson, C.D. 1933. The Positive Electron. Physical Review 43: 491494
Anderson, P.W. 1963. Plasmons, gauge invariance, and mass. Physical Review 130: 439442
Arnowitt, R. and P. Deser. 1965. Spontaneous symmetry breakdown and the -e- interaction. Physical Review B 138: 717723
Baker, M. and S. Glashow. 1962. Spontaneous breakdown of elementary particle symmetries.
Physical Review 128: 24622471
Bludman, S. 1958. On the universal Fermi interaction. Nuovo Cimento 9: 433444
Bogoliubov, N. 1958. A new method in the theory of superconductivity. London: Chapman &
Hall
Borrelli, A. 2009. The emergence of selection rules and their encounter with group theory:
19131927. Studies in the history and philosophy of modern physics 40: 327337
Borrelli, A. 2012. The case of the composite Higgs: the model as a Rosetta stone in contemporary high-energy physics. Studies in the history and philosophy of modern physics
43: 195214
Borrelli, A. 2014. The making of an intrinsic property: symmetry heuristics in early particle
physics, to appear in: Studies in the history and philosophy of science, special Issue on
Integrated History and Philosophy of Science. DOI:10.1016/j.shpsa.2014.09.009
Brout, R. 1967. Group theory of the possible spontaneous breakdown of SU(3). Nuovo
Cimento 47: 932934

48

The European Physical Journal H

Brown, L. and L. Hoddeson. 1983. The birth of elementary particle physics: 19301950. In:
The birth of particle physics, L. Brown and L. Hoddeson (eds.). Cambridge: Cambridge
University Press, pp. 336
Brown, L., M. Dresden and L. Hoddeson. 1989. Pions to quarks: particle physics in the
1950s. In: Pions to quarks: particle physics in the 1950s, L. Brown, M. Dresden and L.
Hoddeson (eds.). Cambridge: Cambridge University Press, pp. 339
Brown, L. and T.Y. Cao. 1991. Spontaneous breakdown of symmetry: Its rediscovery and
integration into quantum eld theory. Historical studies in the physical and biological
sciences 21: 211235
Brown, L.M., R. Brout, T.Y. Cao, P. Higgs and Y. Nambu. 1997. Panel session: spontaneous
breaking of symmetry. In: The rise of the Standard Model, L. Hoddeson et al. (eds.).
Cambridge: Cambridge University Press, pp. 478522
Butterworth, J. 2013. Nobel prize: well done Higgs theorists but what about the experimenters? The Guardian, October 8th 2013, http://www.theguardian.com/science/
life-and-physics/2013/oct/08/nobel-higgs-boson-experimenters, last accessed September
23rd 2014
Byrne, N., C. Iiddings and E. Schrauener. 1965. Meson states in a nonlinear spinor model
of elementary-particle theory with spontaneous symmetry breakdown. Physical Review
B 139: 933944
Carroll, S. 2012. The particle at the end of the universe. London: Oneworld Publications
Carroll, S. 2013. The Nobel prize is really annoying. Entry in Carrolls blog
The preposterous universe. http://www.preposterousuniverse.com/blog/2013/10/07/
the-nobel-prize-is-really-annoying/, last accessed September 21st 2014
CERN press release July 4th 2012. CERN experiments observe particle consistent with long-sought Higgs boson, http://press.web.cern.ch/press-releases/2012/
07/cern-experiments-observe-particle-consistent-long-sought-higgs-boson, last accessed
April 1st 2014
CERN press release Mar 14th 2013. New results indicate that particle discovered at CERN is a Higgs boson, http://press.web.cern.ch/press-releases/2013/03/
new-results-indicate-particle-discovered-cern-higgs-boson, last accessed on April 1st
2014
Chew, G. 1989. Particles as S-matrix poles: hadron democracy. In: Pions to quarks: particle physics in the 1950s, L. Brown, M. Dresden and L. Hoddeson (eds.). Cambridge:
Cambridge University Press, pp. 600607
Chew, G., M.L. Goldberger, F.E. Low and Y. Nambu. 1957. Relativistic Dispersion Relation
Approach to Photomeson Production, Physical Review 106: 134555
Chew, G., M.L. Goldberger, F.E. Low and Y. Nambu. 1957. Application of Dispersion
Relations to Low-Energy Meson-Nucleon Scattering. Physical Review 106: 13374
Close, F. 2010. How the Boson Got Higgss Name. Nature 465: 87374
Close, F. 2011. The infinity puzzle. Oxford: Oxford University Press
Close, F. 2014. Tom Kibble at 80: after dinner speech. In: Symmetry and fundamental physics:
Tom Kibble at 80, J. Gauntlett (ed.). Singapore: World Scientic Publishing, pp. 133135
Conversi, M., E. Pancini and O. Piccioni. 1947. On the Disintegration of Negative Mesons.
Physical Review 71: 209210
Domokos G. and P. Suranyi. 1965. [Spontaneous symmetry breaking in quantum eld theory
(in Russian)]. Journal of nuclear physics (U.S.S.R.) 2: 501511 (English translation:
Soviet journal of nuclear physics 2: 361367)
D
urr, H.-P., W. Heisenberg, H. Mitter, S. Schlieder and K. Yamazaki. 1959. Zur Theorie der
Elementarteilchen. Zeitschrift f
ur Naturforschung 14a: 441485
D
urr, H.-P. 1968. Goldstone theorem and possible applications to elementary particle physics.
Discussion. In: Fundamental problems in elementary physics. London: Interscience publishers, pp. 121
D
urr, H.-P. 1993. Unied eld theory of elementary particles I, II. In: W. Heisenberg,
Collected works. W. Blum et al. (eds.). Berlin: Springer, pp. 133140 and 325334
Ellis, J., D. Nanopoulus and M. Gaillard. 1976. A phenomenological prole of the Higgs
boson. Nuclear Physics B 106: 292340

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

49

Ellis, J., D. Nanopoulus and M. Gaillard. 2012. A historical prole of the Higgs boson.
arXiv:1201.6045 [hep-ph] last accessed on September 21st 2014
Englert, F. 2014. Nobel Lecture: The BEH Mechanism and Its Scalar Boson. Review of
Modern Physics 86: 84350
Englert, F. and R. Brout. 1964. Broken symmetry and the mass of gauge vector mesons.
Physical Review Letters 13: 321323
Englert, F., R. Brout and M.F. Thiry. 1966. Vector mesons in presence of broken symmetry.
Nuovo Cimento 43: 244257
European Physical Society. 1997. High energy and Particle Division of the European Physical
Society. Prize 1997, http://eps-hepp.web.cern.ch/eps-hepp/hepp-prize-awards.php. lst
accessed September 21st 2014
Feynman, R. and M. Gell-Mann. 1958. Theory of the Fermi Interaction. Physical Review
109: 19398
Fraser, G. 2008. Cosmic anger. Abdus salam The first Muslim Nobel scientist. Oxford:
Oxford University Press
Friedman, J.I. and V.L. Telegdi. 1957. Nuclear Emulsion Evidence for Parity
Nonconservation in the Decay Chain + -+ -e+ . Physical Review 105: 16811682
Garwin, R.L., L.M. Lederman and M. Weinrich. 1957. Observations of the Failure of
Conservation of Parity and Charge Conjugation in Meson Decays: The Magnetic Moment
of the Free Muon. Physical Review 105: 14151417
Gauntlett, J. (ed.) 2014. Symmetry and fundamental physics: Tom Kibble at 80. Singapore:
World Scientic Publishing
Gell-Mann, M. and M. Levy. 1960. The axial vector current in beta decay. Nuovo Cimento
16: 705726
Glashow, S. 1961. Partial symmetries of weak interactions. Nuclear Physics 22: 579588.
Glashow, S. 1963. Spontaneous breakdown of octet symmetry. Physical Review 130: 2132
2134.
Goldberger, M.L., and S.B. Treiman. 1958a. Decay of the Pi meson. Physical Review 110:
11781184
Goldberger, M.L., and S.B. Treiman. 1958b. Form factors in decay and capture. Physical
Review 111: 35461
Goldstone, J. 1957. Derivation of the Brueckner Many-Body Theory. 1957. In: Proceedings of
the Royal Society of London Series A. Mathematical and physical sciences 239: 267279
Goldstone, J. 1961. Field theories with superconductor solutions. Nuovo Cimento 19:
154164.
Goldstone, J., A. Salam and S. Weinberg. 1962. Broken symmetries. Physical Review 127:
965970.
Guralnik G.S. 2009. The history of the Guralnik, Hagen and Kibble development of the
theory of spontaneous symmetry breaking and gauge particles. International Journal of
Modern Physics A 24: 26012627
Guralnik, G.S. 2011. The Beginnings of Spontaneous Symmetry Breaking in Particle Physics
Derived From My on the Spot Intellectual Battleeld Impressions, arXiv:1110.2253
[physics.hist-ph], last accessed 15 August 2014
Guralnik, G.S., C.R. Hagen and T.W.B. Kibble. 1964. Global conservation laws and massless
particles. Physical Review Letters 13: 585587
Heisenberg, W. 1957. Quantum theory of elds and elementary particles. Review of Modern
Physics 29: 269278
Hendry, J. 1980. The development of attitudes to the wave-particle duality of light and
quantum theory, 19001920. Annals of Science 37: 5979
Higgs, P.W. 1964a. Broken symmetries, massless particles and gauge elds. Physics Letters
12: 132133
Higgs, P.W. 1964b. Broken symmetries and the masses of gauge bosons. Physical Review
Letters 16: 508509
Higgs, P.W. 1966. Spontaneous symmetry breakdown without massless bosons. Physical
Review 145: 11561163

50

The European Physical Journal H

Higgs, P. 2014. Nobel Lecture: Evading the Goldstone theorem. Reviews of Modern Physics
86: 85153
Higgs mechanism: Revision history. 20042014. Wikipedia. http://en.wikipedia.org/w/
index.php?title=Higgs mechanism\&action=history, last accessed September 20th 2014
Hoddeson, L., H. Schubert, S.J. Heims and G. Baym. 1992. Collective phenomena. In: Out of
the crystal maze, L. Hoddeson et al. (eds.). Oxford: Oxford University Press, pp. 489616.
Itzykson C. and J.B. Zuber. 1980. Quantum field theory. New York: McGraw-Hill.
Kaiser, D. 2005. Drawing theories apart. Chicago: University of Chicago Press
Karaca, K. 2013. The construction of the Higgs mechanism and the emergence of the electroweak theory. Studies in History and Philosophy of Modern Physics 44: 116
Kibble, T. 2009a. Englert-Brout-Higgs-Guralnik-Hagen-Kibble mechanism, Scholarpedia 4:
6441, DOI:10.4249/scholarpedia.6441 last accessed on September 23rd 2014
Kibble, T. 2009b. Englert-Brout-Higgs-Guralnik-Hagen-Kibble mechanism (history),
Scholarpedia 4: 8741, DOI:10.4249/scholarpedia.8741 last accessed on September 23rd
2014
Lee, T.D. and C.N. Yang. 1956. Question of Parity Conservation in Weak Interactions.
Physical Review 104: 254258
Lee, B. 1966. Renormalization of the -model. Nuclear Physics B 9: 649672
Marschak, R.E. 1989. Scientic impact of the Rochester conferences (1950-1960), In: Pions
to quarks: particle physics in the 1950s, L. Brown, M. Dresden and L. Hoddeson (eds.).
Cambridge: Cambridge University Press, pp. 645667
Marx, G. 1965. Decay KL 2 and spontaneous breakdown of CP symmetry. Physical
Review Letters 14: 334336
Mehra J. and K.A. Milton. 2000. Climbing the mountain: the scientific biography of Julian
Schwinger. Oxford: Oxford University Press
Merali, Z. 2010. Physicists get political over Higgs Nature News 4 August 2010,
DOI:10.1038/news.2010.390 last accessed September 18th 2014
Monaldi, D. 2005. Life of : The Observation of the Spontaneous decay of Mesotrons and
its Consequences, 19381947. Annals of Science 62: 41944255
Nambu, Y. 1960a. Quasi-particles and gauge invariance in the theory of superconductivity.
Physical Review 117: 648663.
Nambu, Y. 1960b. Axial vector current conservation in weak interactions. Physical Review
Letters 4: 380382
Nambu, Y. and G. Jona-Lasinio. 1961. Dynamical model of elementary particles based on
an analogy with superconductivity I. Physical Review 122: 345358
Nature Editorial. 2012. Mass Appeal. Nature 483: 374374
Neddermeyer, Seth H. and Carl D. Anderson. 1957. Note on the Nature of Cosmic-Ray
Particles. Physical Review 51: 884886
Nishina, Y., M. Takeuchi and T. Ichimiya. 1937. On the Nature of Cosmic-Ray Particles.
Physical Review 52: 119899
Nobel prize. 1979. http://www.nobelprize.org/nobel prizes/physics/laureates/1979/, last accessed August 20th 2014
Nobel prize. 1999. http://www.nobelprize.org/nobel prizes/physics/laureates/1999/, last accessed August 20th 2014
Nobel prize. 2013. http://www.nobelprize.org/nobel prizes/physics/laureates/2013/, last accessed on August 20th 2014
OED. 1989. Entry Bootstrap. In: Oxford English Dictionary. Second Edition. Online version
June 2012, http://www.oed.com/view/Entry/21553, last accessed August 27th 2012
Pais, A. 1986. Inward bound. Of matter and forces in the physical world. Oxford: Oxford
University Press
Peskin, M.E. and D.V. Schroeder. 1995. An introduction to quantum field theory. Boulder:
Westview
Pickering, A. 1984. Constructing quarks. A sociological history of particle physics. Chicago:
University of Chicago Press
Polkinghorne, J. 1958a. Renormalization of axial vector coupling. Nuovo Cimento 8: 170180
Polkinghorne, J. 1958b. Renormalization of axial vector coupling II. Nuovo Cimento 8: 781

A. Borrelli: The story of the Higgs boson: the origin of mass in early particle ...

51

Polkinghorne, J.C. 1989. Rochester roundabout: the story of high energy physics. Harlow:
Longman
Pontecorvo, B.M. 1989. Recollections on the establishment of the weak interaction notion.
In: Pions to quarks: particle physics in the 1950s, L. Brown, M. Dresden and L. Hoddeson
(eds.). Cambridge: Cambridge University Press, pp. 367372
Rechenberg, H. 1989. The early S-matrix theory and its propagation (19421952). In: Pions
to quarks: particle physics in the 1950s, Brown L., M. Dresden and L. Hoddeson (eds.).
Cambridge: Cambridge University Press, pp. 551578
Royal Swedish Academy, Class of Physics 2013. The BEH-mechanism, interactions with
short-range forces and scalar particles, Advanced information about the Physics Nobel
prize 2013, http://www.nobelprize.org/nobel prizes/physics/laureates/2013/advanced.
html, last accessed on September 23rd 2014
Salam, A. 1961. Some speculations on the new resonances. Review of modern physics 33:
426-430
Salam, A. 1968. Weak and electromagnetic interactions. In: Elementary particle theory, N.
Svartholm (ed.). Stockholm: Almqvist and Wiskell. pp. 367377
Salam, A. 1979. Gauge unication of fundamental forces. In: Nobel Lectures. Physics 1971
1980, S. Lundqvist (ed.). Singapore: World Scientic Publishing Co. 1992, http://
www.nobelprize.org/nobel prizes/physics/laureates/1979/salam-lecture.html, last accessed August 28th 2014
Salam, A. and J.C. Ward. 1959. Weak and electromagnetic interactions. Nuovo Cimento 11:
568577
Salam, A. and J.C. Ward. 1960. I = 1/2 rule. Physical Review Letters 5: 390
Salam, A. and J.C. Ward. 1961. On a gauge theory of elementary interactions. Nuovo
Cimento 19: 165170
Sakurai, J. 1958. Mass Reversal and Weak Interactions, Nuovo Cimento (19551965) 7:
649660
Sakurai Prize. 2010. http://www.aps.org/programs/honors/prizes/sakurai.cfm, last accessed
on April 10th 2014
Sample, I. 2010a. Massive. The missing particle that sparked the greatest hunt in science.
New York: Basic Books
Sample, I. 2010b. The Long Story of How the Boson Got Only Higgss Name. Nature 466:
689689
Sanyuk, V.I. 1992. Genesis and evolution of the Skyrme model from 1954 to the present.
International Journal of Modern Physics A7: 140, reprinted in: T. Skyrme, Selected
papers, with commentary, of Tony Hilton Royle Skyrme. G.E. Brown (ed.). Singapore:
World Scientic, 1994, pp. 126165
Schweber, S.S. 1994. QED and the men who made it: Dyson, Feynman, Schwinger, and
Tomonaga. Princeton: Princeton University Press
Schwinger, J. 1957. A theory of the fundametal interaction. Annals of Physics 2: 407434
Schwinger, J. 1962a. Gauge invariance and mass. Physical Review 125: 397398
Schwinger, J. 1962b. Gauge invariance and mass II. Physical Review 128: 24252429
Shaw, R. 1955. The problem of particle types and other contributions to the theory of
elementary particles. PhD thesis, Cambridge University, unpublished
Skyrme, T.H.R. 1958. A non-linear theory of strong interactions. In: Proceedings of the Royal
Society of London. Series A. Mathematical and Physical Sciences 247: 260278
Skyrme, T.H.R. 1959. A unied model of K- and -mesons. In: Proceedings of the Royal
Society of London. Series A. Mathematical and Physical Sciences 252: 236245
Skyrme, T.H.R. 1988. The origin of skyrmions. International Journal of Modern Physics A3:
27452751, reprinted in: T. Skyrme, Selected papers, with commentary, of Tony Hilton
Royle Skyrme. G.E. Brown (ed.). Singapore: World Scientic 1994, pp. 116125
Strocchi, F. 2008. Symmetry breaking. Berlin: Springer
Streater, R.F. 1965. Spontaneous breakdown of symmetry in axiomatic theory. In:
Proceedings of the Royals Society of London. Series A. Mathematical and Physical
Sciences 287: 510518

52

The European Physical Journal H

Street, J.C. and E.C. Stevenson. 1937. New Evidence for the Existence of a Particle of Mass
Intermediate Between the Proton and Electron. Physical Review 52: 10031004
Sudarshan, E.C.G. and R. Marshak. 1958. The Nature of the Four-Fermion Interaction.
In: Proceedings of the Conference on Mesons and newly-discovered particles. Bologna:
Zanichelli, reprinted in A Gift of Prophecy, E.C.G. Sudarshan (ed.). Singapore: World
Scientic, pp. 508515
Suzuki, M. 1963. Foundation of spontaneous breakdown of symmetry. Progress in Theoretical
Physics 30: 627635
t Hooft, G. 1997. Renormalization of gauge theories. In: The rise of the Standard Model, L.
Hoddeson et al. (eds.). Cambridge: Cambridge University Press, pp. 179198
t Hooft, G. and M. Veltman. 1972. Regularization and renormalization of gauge elds.
Nuclear Physics B 44: 189213
Talk: 1964. PRL symmetry breaking papers. 2009. Wikipedia. http://en.wikipedia.org/wiki/
Talk%3A1964 PRL symmetry breaking papers, last accessed on September 21st 2014
Treiman, S. 1989. A connection between the strong and weak interactions. In: L. Brown,
M. Dresden and L. Hoddeson (eds.), Pions to quarks: particle physics in the 1950s.
Cambridge: Cambridge University Press, pp. 384389
Weinberg, S. 1967. A model of leptons. Physical Review Letters 19: 12641266
Weinberg, S. 1979. Conceptual Foundations of the Unied Theory of Weak and
Electromagnetic Interactions. In: Nobel Lectures. Physics 19711980, S. Lundqvist (ed.).
Singapore: World Scientic Publishing Co. 1992, http://www.nobelprize.org/nobel
prizes/physics/laureates/1979/weinberg-lecture.html, last accessed August 28th 2014
Weinberg, S. 2014. Tom Kibble: Breaking ground and breaking symmetries. In: Symmetry
and fundamental physics: Tom Kibble at 80, J. Gauntlett (ed.). Singapore: World
Scientic Publishing, pp. 125131
Wentzel, G. 1956. [untitled]. In: High energy nuclear physics. Proceedings of the sixth annual
Rochester conference, J. Ballam et al. (eds.). New York: Interscience, pp. VIII-15VIII-17
Woit, P. 2010. The Anderson-Higgs mechanism. Entry and comments in Woits blog Not
even wrong. http://www.math.columbia.edu/woit/wordpress/?p=3282 last accessed
September 21st 2014
Woit, P. 2013. Nobel for Englert and Higgs. Entry and comments in Woits blog Not
even wrong. http://www.math.columbia.edu/woit/wordpress/?p=3282 last accessed
September 21st 2014
Wolf
Prize
in
Physics.
2004.
http://www.wolund.org.il/index.php?
dir=site\&page=winners\&name=\&prize=3016\&year=2004\&eld=3008,
last
accessed on April 10th 2014
Wu C.S., E. Ambler, R.W. Hayward, D.D. Hoppes and R.P. Hudson. 1957. Experimental
Test of Parity Conservation in Beta Decay. Physical Review 105: 141315
Yang, C.L. and R.L. Mills. 1954. Conservation of isotopic spin and isotopic gauge invariance.
Physical Review 96: 191195
Yukawa, H. 1935. Interaction of Elementary Particles. I. In: Proceedings of the PhysicoMathematical Society of Japan 17: 4857

Das könnte Ihnen auch gefallen