Rudolf Haags legacy of Local Quantum Physics

arXiv:1612.00003v1 [math-ph] 30 Nov 2016

and reminiscences about a cherished teacher and

friend
In memory of Rudolf Haag (1922-2016)

submitted to the Eur. Phys. J. H

Bert Schroer
permanent address: Institut fur Theoretische Physik
FU-Berlin, Arnimallee 14, 14195 Berlin, Germany
November 2016

Abstract
After some personal recollectioms about Rudolf Haag and his thoughts
which led him to Local Quantum Physics, the present work recalls his
ideas about scattering theory, the relation between local observables and
localized fields and his contributions to the physical aspects of modu-
lar operator theory which paved the way for an intrisic understanding
of quantum causal localization in which fields coordinatize the local
algebras.
The paper ends with the presentation of string-local fields whose con-
struction and use in a new renormalization theory for higher spin fields is
part of an ongoing reformulation of gauge theory in the conceptual setting
of Haags LQP.

1 First encounter with Rudolf Haag

On his return from the Niels Bohr Institute in Copenhagen to the University
of Munich Rudolf Haag passed through Hamburg to meet his colleague Harry
Lehmann, at that time the newly appointed successor of Wilhelm Lenz who
held the chair of theoretical physics since the 1920 foundation of the University
of Hamburg. It was the year 1958 shortly after the decision to construct the
DESY particle accelerator in Hamburg which created a lot of excitement. I had
nearly completed my diploma thesis under Lehmann and begun to worry about
my career.
Haag was about to accept an offer of a staff position at the University of
Illinois in Urbana, He asked me whether I would be interested to continue my

1
career in the US as his collaborator. The prospect of a scientific career and the
desire to change my somewhat precarious living conditions which I encountered
after my 1953 flight from East Germany to Hamburg made such a prospect
irresistible.
To better get to know each other Haag invited me to accompany him on a
visit to Daniel Kastler who at that time was a recently appointed faculty member
of the physics department of the University in Marseille. He had met Daniel a
year before and both participated in an international conference in Lille/France
where Rudolf for the first time presented his idea to base quantum field theory
on spacetime-localized operator algebras at an international conference. Daniel
was attracted by these new ideas and the purpose of Rudolfs visit was to obtain
Daniels help for the improvement of their at that time still shaky mathematical
formulation. In this way Daniel and Rudolf became soulmates in the exploration
of what was referred to as algebraic quantum field theory (AQFT) and later more
appropriately named local quantum physics which as a result of its frequent
use will be abbriviated as LQP. With Rudolfs acceptance of an offer from the
University of Illinois and his impending move to the US their collaboration was
delayed. Their first important joint publication appeared in 1962 [1].
The voyage to Marseille provided an opportunity to get to know each other
before my planned but not yet approved move to the US. The journey by car
through parts of Germany and across Switzerland and parts of southern France
to Marseille was an unforgettable experience. Having fled communist East Ger-
many and gone to Hamburg in 1953, it was my first travel outside the German
borders. In particular the journey along the Cote dAzur with its subtropical
vegetation and its new scents and cultural impressions remaines impressed in
my memory.
After my return to Hamburg Rudolfs offer to work with him in the position
of a research associate took a concrete form; I bought boat tickets on the Holland
America line for my family and the first birthday of our daughter was celebrated
in the middle of the Atlantic.
Arriving at the University of Illinois in Urbana I encountered a formal prob-
lem. Even taking into account the shock from the 1957 launching of the Sputnik
in 1957 which led to the creation of new positions for physicists and engineers,
the offer of a research associate position to somebody without a Ph.D. was un-
usual. As I learned later from Rudolf he cleared this problem in a conversation
with the department chairman Frederick Seitz.
Frederick Seitz, a renowned physicist with politival influence on US sci-
ence policies, was a former student of Wigner. This may have played a role in
Wigners recommendation of Haag for a full professorship at the University of
Illinois. Haags prior visiting position at the University of Princeton led to many
scientific contacts with Wigner and Wightman. In his reminiscenses [2] he gives
credit to Wightman for having directed his attention to Wigners 1939 path-
breaking work on the classification of all unitary representations of the Poincare
group. It is hard to understand why this important work of Wigners remained
unnoticed for more than a decade.
He also mentions contacts with other members of the Princeton universitys

2
physics faculty; in particular with Valja Bargmann, who extended Wigners work
on representation theory, as well as with Marvin Goldberger and Sam Treiman,
who at that time were working on the extension of the optical Kramers-Kronig
dispersion relations to particle physics. During this time in Princeton Haag
was the thesis adviser to Huzihiro Araki, a brilliant young student from Japan,
Araki visited Urbana several times and some discussions even led to a joint
publication [4].
Besides recalling personal events these notes present important ideas of
Haags local quantum physics (LQP) in their historical context. In order to
direct attention to its largely untapped innovative strength the last two sec-
tions include the beginnings of a LQP inspired positivity perserving string-local
renormalization theory for interactions involving higher spin s 1 fields whose
aim to replace the gostly BRST gauge theory by a LQP formulation which
only uses physical degrees of freedom. For Haag this was one of LQPs greatest
challenges [2].
Frequently occurring scientific expressions will be abbreviated: quantum
field theory (QFT). local quantum physics (LQP), point-like (pl), string-like
(sl), string-local quantum field theory (SLFT), power-counting bound (pcb),
spontaneous symmetry breaking (SSB), string theory (ST), the Becchi-Rouet-
Stora-Tyutin gauge formalism (BRST).

2 With Haag in Urbana

After the resounding success of renormalization theory in quantum electrody-
namics the main interest shifted to high energy nuclear interaction. It soon
became clear that these methods of perturbative renormalization theory do
not work for processes involving strong interactions (which in the field theo-
retic description at that time meant trilinear -meson-nucleon couplings and
-selfinteractions).
Moreover doubts were increasing as to whether the locality, as formally con-
tained in the relativistically covariant Lagrangian quantization, retaines its va-
lidity in the new high energy domain of nuclear interactions. From earlier work
on quantum optics it was well known that certain analytic relations, known
as the aforementioned Kramers-Kronig dispersion relations, had a rather direct
connection to relativistic causal propagation. The problem was to derive such
analytic relations for the scattering amplitudes of strongly interacting particles.
That positivity of energy together with Einstein causality leads to analytic
properties of spacetime correlation functions (vacuum expection values) of fields
was already well known. However fields in spacetime are not directly accessible
to measurements; in experiments; one rather measures scattering amplitudes of
particles in momentum space which have a large time asymptotic relation with
fields. For those on-shell amplitudes in momentum space the derivation of
analytic consequences of causality posed a harder problem than that of off-
shell correlation functions in spacetime (Wightman functions).
For the confidence in the validity of causal locality at the higher energies of

3
a new generation of accelerators it was important to obtain a rigorous deriva-
tion from the causal localization properties of field operators (micro-causality
of QFT) . The form of the expected dispersion relation was known from the
study of Feynman graphs; what was missing was a derivation from the spacelike
(anti)commutation relations of quantum fields.
The joint effort of Harry Lehmann together with Res Jost as well as con-
tributions from Freeman Dyson resulted in the derivation of dispersion relation
from first principles. The subsequent experimental verification at the at that
time highest energies at the Brookhaven accelerator brought the dispersion re-
lation project to a successful close. The confidence in the validity of the causal
locality principle in the new area of High Energy Physics was restored and the
interest in nonlocal modifications of QFT subsided.
For Haag quantum causality is not fully accounted for by Einstein causality
in the form of (anti)commutation of operators whose spacetime localization
regions are spacelike separated. He expected that in his LQP formulation in
terms of a net of causally related algebras the quantum counterpart of hyperbolic
propagation of Cauchy data, although closely related to Einstein causality, can
not be derived from it.
The relation of LQP to Wightmans axiomatic formulation in terms of fields
and their vacuum expectation values was from Haags LQP point of view analo-
gous to the relation between the coordinate-independent presentation of modern
geometry and its description in terms of coordinates. Decades later this analogy
was made more precise by H.-J. Borchers who showed that the quantum fields
form local equivalence classes and that the different members in one class (pro-
vided they their matrixelements between the vacuum and one-particle states do
not vanish) derscibe not only the same particles and their scattering matrix but
also generate the same localized algebras [10] [32].
The necessarily singular nature as operator-valued Schwartz distributions
(as a result of the omnipresence of vacuum polarization clouds) renders the
relation between fields and operator algebras very intricate. The presence of
these polarization clouds accounts for the fundamental difference of the intrinsic
causal localization and the Born localization in terms of an arbitrarily chosen
quantum mechanical position operator. Haag expected that causality properties
can be more natural described in his LQP setting.
In addition to Einstein causality there should also exist a time-like causality
property which is the quantum analog of the hyperbolic propagation of classical
waves. Classically the initial values within a sphere of radius r at time t = 0 cen-
tered at the origin have a region of influence which is the forward and backward
light cone emanating from the sphere. But they also lead to a compact double
cone region C = x | x x r2 inside which the radiation is completely
determined in terms of the t = 0 Cauchy data in the sphere. Any classical field
strength measured inside C which cannot be accounted for in terms of these
Cauchy data would be seen as a mysterious violation of causal propagation
since according to Einsteins causality requirement it could not have entered
from the causal complement.
Apart from free quantum fields whose propagation properties can be directly

4
related to those of classical fields, it is not clear how to formulate this hyperbolic
propagation property for interacting Wightman fields. Interestingly it turns out
that this is much easier in Haags algebraic formulation of LQP. It amounts to
an equality of two different localized (von Neumann) operator algebras

A(O) = A(O ) (1)

where O in the above illustration would correspond to a in time direction

slightly thickened spatial sphere1 . The causal complement O consists of all
points which are space-like with respect to O and the causal completion (or
causal shadow) is the complement taken twice which is generally larger than i.e.
O O with the equality defining the maximal extension which is consistent
with Einstein causality (the causal completion).
Our joint effort was to look at the (at that time simpler appearing) problem
of the time-slice property corresponding to the classical determination of the
future field in terms of the global t = 0 Cauchy data. A global time slice can be
patched together from infinitely many finite double cone O s since the additivity
property of LQP2 relates a violation of the causal completeness with that of the
time-slice property. Such a violation in a model which fulfills all other LQP
requirements implies that causal completeness is not a consequence of Einstein
causality.
In my recollection the start of my work on local quantum physics (LQP)
with Rudolf is inexorably related with a beautiful summer on Wisconsins lake
shores. Rudolf proposed to look for models which are Einstein causal but violate
causal completeness. At the beginning I was somewhat discouraged because I
considered my knowledge acquired under Lehmann as insufficient for the new
work on LQP. But it then turned out that at least some of it was useful.
Rudolfs heuristic idea was that too many degrees of freedom within a
bounded spacetime region and a certain bound on their invariant energy content
may (a kind of relativistic phase space) lead to violations of causal completeness.
In such a case an observer would see more degrees of freedom in the double
cone than his experimental friend had injected into the base region atound t = 0.
Such a poltergeist effect of increase of degrees of freedom apparently coming
from nowhere (since according to Einstein causality they cannot come from
space-like separated regions) is an unexceptable violation of causality and must
be excluded and the LQP setting is the best way of formulating this.
Our simple counterexample was provided by so called generalized free fields
with a suffiently increasing mass distibution; so my modest contribution to a
joint paper consisted in some calculation with generalized free fields with suitable
chosen continuous mass distributions (m). Rudolfs intuition was vindicated
1 Haag-Kastler nets of local operator algebras are localized in open spacetime. This is

similar to interacting quantum fields which as a result of their singular short distance behavior
have to be smeared with testfunctions supported in open regions.
2 The requirement that the operator algebra generated from the union of localized algebras

with overlapping localization regions is equal to the algebra localized on the union of the
localization regions.

5
by the result which showed that although all Wightman properties are satisfied,
the time slice property is violated [3].
With the hindsight of later work one may view this illustration as a first
indication of the importance of the notion of cardinality of degrees of freedom
which in the decades to come received various refinements; first the Haag-Swieca
compactness, then the Buchholz-Wichmann nuclearity [10] and the more
recent modular nuclearity which is used in existence proofs of certain d = 1+1
models of QFT (next section)
Together with the other postulates which already appeared in Haags contri-
bution to the 1957 Lille conference [2] our work was a rather complete account
of the axioms which define the framework of LQP. Two years later it was
superseded by a paper of Haag and Kastler [1] which contained a more detailed
account of their mathematical structure and physical consequences. The H-K
work is still considered to be the most authoritative reference for the algebraic
approach to QFT.
The time slice property played no role in most presentations of LQP but it
turned out to be important in recent formulations of QFT in curved spacetime.
For later reference it is helpful to collect the LQP causality requirements3

A(O ) A(O) Einstein causality, EC (2)

A(O)= A(O ) causal completion property, CC

A(O ) = A(O) , Haag duality, HD (3)

Haag duality is an important special case of Einstein causality. The CC property

is closely related to HD. Einstein causality can be directly expressed in terms
of covariant fields, whereas for HD and CC this is more subtle.
For a physical interpretation the EC and CC requirements are indispensable
whereas violations of HD occur in the presence of massless s 1 interaction for
multiply connected spacetime regions (tori). The most prominent illustration is

the operator of magnetic flux through a solid torus H(T ) which belongs A(T )
but not to A(T ) [6]. Interestingly the Aharonov-Bohm effect is related to this
breakdown of Haag duality [5].
The causal completeness property also severely limits relations between LQPs
in different spacetime dimensions (extra dimensions). This affects in partic-
ular the mathematical isomorphism between LQPs on n dimensional Anti de
Sitter space (AdS) and n 1 dimensional conformal spactime which share the
same spacetime symmetry group. On heuristic grounds one expects that the
AdS-CFT isomorphism leads to similar problems as the previous CC violation
for certain generalized free fields with too many degrees of freedom.
This is precisely what happens on the lower spacetime dimensional conformal
side [7]. In fact starting with a CC obeying AdS free field [8] one obtains
an overpopulated conformally invariant generalized free field [3], and this
problem does not disappear in the presence of interactions. Conversely the LQP
3 The dash on the operator algebra refers to its commutant i.e. the subalgbre of all bounded

operators A(O) B(H) which commute with A(O).

6
double cone algebras on the AdS side obtained from a healthy conformal LQP
are anemic in the sense that compact localized algebras do not contain any
degrees of freedom and one has to pass to noncompact localization regions to
encounter algebras which are not multiples of the identity. All those cases the
algebraic isomorphism preserves EC but violates the with degrees of freedom
connected CC.
The Kaluza-Klein proposal of extra or lowered spacetime dimensions works
for classical field theories as well as semi-classical approximations but it clashes
with QFT. Transplanting the matter content between worlds of different
spacetime dimensions preserves EC but fails on CC.
The issue of causal localization sustaining quantum degrees of freedom is a
very subtle one which is inexorably related with the role of vacuum polarization
clouds in causal localization and has no counterpart in quantum mechanics or
classical field theory. Using the standard formulation of QFT field coordinatiza-
tions one may easily overlook the breakdown of the causal completeness property
as a result of an overpopulation of degrees of freedom resulting from resettling
the degrees of freedom of a higher dimensional LQP into a lower dimensional
spacetime vessel. This is precisely what happened when the AdS-CFT isomor-
phism and the idea of extra dimensions became a focal point of interests in the
90s which led to thousands of publications.
During the almost 3 years of my time in Urbana there were many interesting
visitors. I remember that Gell-Mann on one of his visits asked us if we had a
more intrinsic understanding of the relation between the partially conserved
axial vector currents (PCAC) with the field of the -meson. At that time
gauge theoretic Lagrangian models with axial -mesons as proposed by Sakurai
enjoyed great popularity. At the end of the discussion Murray Gell-Mann joked:
you mean we can shoot Sakurai? before he enjoyed looking at our somewhat
helpless expressions.
Together with Haag I participated in a summer school in in Boulder, Col-
orado. My remembrances about the activities in physics are faint but I do recall
having been impressed by the beautiful nature of the Rocky Mountains and a
subsequent journey with my family through the Yellowstone National Park.
I also recollect an extremely peculiar occurrence. When I looked as usual
into the weekly Time Magazine I came across a story about two mathematicians
at the University of Illinois which were engaged in classified work for the NSA
before they defected via Cuba to the Soviet Union taking classified material with
them. The name of one of them was the same as that of somebody who lived in
an apartment in Urbana which I rented shortly before I went to Boulder. The
apartment in a university housing project became too small for my family after
the birth of my son. The former tenant whose name was Martin also sold his
piano and some furniture to me before he moved out. There was a picture of
the two mathematicians in Time Magazine, but the quality of printed photos in
those times was so poor that I could not identify him. I brushed the incidence
aside as a coincidence of names and enjoyed the rest of the stay.
Two agents of the CIA were already waiting for me. Apparently they found
the check of my payment for the piano in a Washington deposit. They really

7
knew a lot about my past, in particular that in 1953 I fled from East Germany.
Probably they obtained their knowledge from an archived protocolled hearing
in a transit refugee camp, a former concentration camp near Hamburg where
besides German officials also a US officer was present.
Rudolf assured me that this matter will be cleared up in a short time. Indeed
after several meetings in a restaurant I succeeded to convince them that my
involvement was coincidental and that I was not an East German spy.
Many years later when I mentioned the Martin-Mitchell spy story at an
international physics conference to Ludvig Faddeev, he told me that a week
before both of them applied for a position at the Steklov Institute in Leningrad.
By that time they had Russian wifes and and families. How was this possible;
did the communist ideology convert two homosexuals ?
Before my position at the University of Illinois came to an end, I met Jorge
Andre Swieca who, after having spent a year at the Werner Heisenberg Institute
in Munich (one of the largest Max Planck Institutes for physics in Germany),
passed through Urbana on his way to Brazil. The purpose of this visit was to
introduce himself to Rudolf as his new Research Associate. After he defended
his Ph.D thesis at the University of Sao Paulo (with Guettinger as his advisor)
he returned to Urbana to start his work with Haag.
During my stay in Urbana I had obtained some results which were appropri-
ate to be used for a Ph.D thesis. I returned 1963 to Hamburg where I submitted
my thesis. The terminology Infraparticles in its title [9] referred to the conjec-
ture that the infrared divergencies which appear in the scattering amplitudes of
electrically charged particles are related to a modification of the Wigner particle
structure. I was able to illustrate this in a two-dimensional model. The realistic
case was taken up two decades later by Buchholz. The issue of infraparticles
has remained a challenging topic of LQP [10]
After a one year at the IAS in Princeton, a short stay at the University of
Hamburg and a visit to the Middle East University in Ankara at the invitation
of Feza Gursey I returned to the US to take up my new position of associate
professor at the University of Pittsburgh.
Shortly before I left Urbana I shared an office with Derek Robinson who
became Haags second collaborator. During a visit by Kastler, Robinson Swieca
and Kastler investigated the properties of conserved currents within the new
algebraic Haag-Kastler setting of LQP. They found that the conservation law
only secures the existence of a partial charge which secures the existence of
a local symmetry within each finite spacetime localization regions but that the
global charge may diverge i.e. the inverse of the quantum Noether theorem may
be violated i.e. the current conservation may not secure the existence of a global
charge (the infinitesimal generator of a unitary symmetry.
It was known from perturbative investigations of self-interacting scalar fields
by Goldstone that the local current conservation may lead to a divergent global
charge resulting from the contribution of a massless scalar (Goldstone) boson
which impedes the large distance convergence and in this way causes a situation
which was appropriately referred to as spontaneous symmetry breaking (SSB).
Kastler, Swieca and Robinson showed that this cannot happen in the pres-

8
ence of a mass gap [12], and in a follow up paper (based on the use of the Jost-
Lehmann-Dyson representation) Swieca together with Ezawa [13] succeeded to
prove the Goldstone theorem in a model- and perturbation- independent way4 .
Goldstone constructed renormalizable SSB models of self-interacting scalar par-
ticles by applying the shift in field space prescription to formally symmetry-
preserving Mexican hat potentials.
This quasiclassical prescription leads to a model-defining first order inter-
action density which maintains the conservation of the symmetry currents in
all orders. There are symmetry-representing unitary R operators for each finite
spacetime region O but the global charges Q = j0 of same symmetry gener-
ating currents diverge. This is the definition of SSB whereas the shift in field
space procedure is a way to prepare such a situation whenever SSB is possible.
For the later presentation of the Higgs model it is important to be aware of a
fine point about SSB whose nonobservance led to a still lingering confusion. As
soon as scalar self-interacting fields are coupled to s = 1 potentials the physical
interpretation of the field shift manipulation on a Mexican hat potential as a
SSB is incorrect; one obtains the Higgs model for the wrong physical reasons
and misses the correct reasons why there can be no self-interacting massive
vectormesons without the presence of a H-field. Although this can be decribed
correctly in the gauge theoretic formulation, a better understanding is obtained
in the positivity preserving string-local setting of LQP (see section 6)
QFT is not a theory which creates masses of model-defining fields. The
masses of those free fields which define the first order interaction density) are,
together with the coupling strengths, free parameters5 . The only dynamic
masses are those of bound states created by acting with interacting composite
fields on the vacuum state but unfortunately there is no perturbative methods
which descibes bound states. space.
In a later paper Haag and Swieca investigatigated the cardinality of states
contained in a finite spacetime region with limited energy content [11]. In quan-
tum mechanics this corresponds to the number of degrees of freedom per cell in
phase space which is finite. They found that LQP leads to an infinite set whose
cardinality cannot exceed that of a compact set.

3 The Brazilian connection

Shortly before I left in 1962 I met Jorge Andre Swieca for the first time when,
on his return from the Max Planck Institute in Munich to the University of
Sao Paulo (USP) he passed through Urbana for an interview with Haag. His
thesis adviser was Werner Guttinger who, as several other German physicists,
was invited to the in 1952 newly founded Instituto de Fisica Theorica (IFT) in
4 The Goldstone theorem states that a Nother symmetry in QFT is spontaneously broken

precisely if aRmassless scalar Goldstone boson prevents the convergence of some of the global
charge Q = j0 = ..
5 Masses and mass ratios may appear in coupling strengths of induced higher order contri-

butions.

9
Sao Paulo. Guttinger recognized the potential of Andre Swieca and arranged a
visiting position for him at the MPI in Munich. When I met Andre in Urbana
he was on his way back to the USP in order to defend his thesis before taking
up the research associate position with Haag in Urbana.
Guttinger is one of the few theoretical physicists who, shortly after Laurant
Schwartzs presentation of the theory of distributions, saw the relevance of that
theory for the description of the singular nature of quantum fields. Before
he obtained a permament position at the University of Tubingen/Germany he
spend some years in the second half of the 50s at the ITF. It is interesting to
note that around 1952 Laurant Schwartz together with Alexander Grothendiek
spend some time at the USP. On my first visit in 1968 there still existed traces
of the legacy of Laurant Schwartz in the form of courses on distribution theory
at the USP physics department which were presented by a young Brazilian lady
who obtained her PhD with Laurant Schwartz.
My first chance to take a short leave of absence from the University of
Pitteburgh to follow Andre Swiecas invitation to the USP came in 1968. After
his return from the collaboration with Haag in Urbana to Brazil at the end of
1966 Andre held the position of a junior professor at the USP. When I arrived
he was surrounded by a group of enthusiastic young students of whom the most
advanced (Jose Fernando Perez) was assigned the task to take care of me and to
help me with the written version of my lectures on QFT. This was the beginning
of what Haag in his reminiscences called the Brazilian connection ([2] page
24).
During my visit Andre received the Moinho Santista prize for his quantum
field theoretic work on symmetries and their spontaneous breaking. After Jaime
Tiomno, one of the founders (together with Mario Schemberg and Jose Leite-
Lopes) of theoretical physics in Brazil, Andre was its second recipient. After
his collaboration wth Kastler and Robinson in Urbana on the LQP formulation
of symmetries and their conserved currents he had pursued this issue in more
depth with particular attention for spontaneous symmetry breaking (SSB) for
which the current remained conserved but the presence of a massless Goldstone
boson one looses the symmetry generator since the global charge diverges. In
a joint paper with H. Ezawa the Goldstone theorem, which previously only
existed as a perturbative property of a special class of models, was derived in a
model-independent way from the causal localization principles of QFT [13].
He lectured on his results in Erice [14] and up to date I know no clearer
model-independent presentation of Goldstones SSB theorem about SSB as a
consequence of the causality and spectral properties principle of QFT than that
in his notes. This is particularly important in times in which SSB became
somewhat misleadingly synonymous with a shift in field space on a Mexican hat
potential (see remarks in previous section).
At the time of my visits during the 60s and 70s Brazil was ruled by a mil-
itary junta whch took power in 1964 coupe. At the start of my visit in 1968 I
hardly noticed the presence of a military dictatorship, but the situation changed
abruptly in May 1968 when at the time of intensification of the Vietnam war
there were student demonstrations in Paris and Berlin and other places. I lis-

10
tened to the news on my short wave radio but soon became aware that there
was an increasing number of demonstrations against the military dictatorship
whose only connection with the Vietnam war was that those who started the
war were the same who supported the military regime in Brazil.
Many years after Swiecas premature death in 1980 somebody asked me
whether I knew something about a rumor that after having received the Santista
prize he was approached by the military government to explore the possibility to
offer the post of a scientific/cultural attache to Israel. I did not, but I was sure
that if this really happened Andre would have declined an offer of representing
a military dictatoship in a democratic country. Sometimes I saw military police
entering the USP campus and later I learned that one of my collegues Ernesto
Hamburger was taken into custody and his wive was tortured.
Andre told me a saddening story about an occurrance which happened
shortly before to one his professors from whom he took his first physics courses.
Plinio Susskind had a very strong personal contact with his students, after the
lectures he joint them to continue discussions about matters of physics and daily
events in nearby cafes and bars. He had a collection of books which included
the work of Marx and others which after the military coup were considered sub-
versive. When the military police searched his apartment they found a copy of
Sergei Eisensteins Couracado Potemkin (Battleship Potemkin). He was taken
into custody and after having been released he lost his university position.
He was not internationally known and had no chance to continue working
outside Brazil. He fell into poverty and Andre and some of his former fellow
students supported him for many years. The worst aspect of a military regime is
that it encouraged denunciations which in some people used to settle accounts.
Two of the founders of theoretical physics in Brazil as Jaime Tiomno and Leite-
Lopes who felt threatened by the regime accepted positions in the US or France.
For more than a decade, starting from the beginning of the 70s up to the return
of democracy in 1985, the catholic university of Rio de Janeiro (PUC) became a
refuge for many Brazilian scientists including Jorge Andre Swieca who worked
there for several years.
On this first visit to Brazil there was little time and peace of mind to talk
about how to use our shared knowledge acquired as collaborators of Haag for
establishing a joint project. We postponed the discussions of topics of joint
interests to future visits.
One week after my return to Pittsburgh I received a notice that military
tanks entered the USP at dawn and took positions around the CRUSP housing
and took everybody into custody. Apparently was released on the same day;
not because he was particularly cooperative but rather as the result of taking
notice that his fiance was the daughter of a high ranking military; an occurrence
which is easily understood for those who experienced the Brazilian jeitinho
which survived any system up to date.
When back in Pittsburgh I obtained informations about the worsening polit-
ical situation in Brazil I found myself in a unusual schizophrenic situation; here
I was living peacefully in a democratic country whose government supported
military dictatorships in other countries under which my colleagues suffered.

11
 Less than two years later Andre visited me at the University of Pittsburgh
where the QFT group was meanwhile strengthened by Ruedi Seiler, a mathe-
matical physicist who received his PhD shortly before from the ETH in Zurich.
Looking for a topic on which one could start a short time collaboration we found
it worthwhile to investigate to what extend Einstein causality and the causal
shadow property retain their validity for interactions of quantum fields with
external (classical) fields.
Using functional analytic methods it was possible to show with the help of
the energy norm that these causality properties hold for models of low spin quan-
tum fields coupled to time-dependent asymptotically vanishing classical fields
and for s > 1 interactions we extended previous observations about acausalities
[15]. In case of strong stationary external fields we were able to improve the
understanding about an inconsisteny of the Klein Gordon field in a strong po-
tential made thirty years before [16]. The result was that there are two ways of
quantizing bound states with negative E 2 namely either by using indefinite met-
ric or by abandoning the vacuum postulate and accepting repulsive (inverted)
oscillator degrees of freedom associated to such bound states..
This led me to take another look at tachyons described by fields with
m2 m2 . As the name suggests these fields were thought of as being associ-
ated to fields describing superluminal stuff. But how can this be in view of
the fact that a classical tachyon field has a perfectly causal propagation? The
answer was that in limiting oneself to real spacelike momenta one has left out
imaginary values m2 + p~2 < 0 whose momenta lead to inverted oscillators
which in quantum theory requires to substitute the vacuum state by a contin-
uum of negative energy jelly states. Such a situation without bottom becomes
chaotically unstable in the presence of interactions; this is reminiscent of Diracs
hole theory except that in the tachyon case there is no filling. with arbitray
large negative energies They correspond to those inverted oscillators which in
the problem of strong external potentials prevent the existence of a lowest en-
ergy vacuum state. In the tachyon problem they require the introduction of a
continuum of negative energy jelly states whose presence is indispensable for
maintaining causal propagation. Although the free theory exists, any perturba-
tion will cause a similar instability as the Dirac sea before filling it, except that
for tachyons such filling is not possible [17].
This instability argument was later used in the quasiclassical preparation
of spontaneous symmetry breaking (SSB), which is a mild form of symmetry
breaking in which there still exists a conserved current but its charge (the gen-
erator of the symmetry) diverges due to the presence of a massless scalar boson
(the Goldstone boson). Since it is somewhat tedious to prepare a first order
interaction density with this property one starts from a symmetric Mexican hat
kind of selfinteraction of a multiplett of particles and uses the quasiclassical
trick of a shift in field space which brings an apparently tachyonic situation of
a Mexican hat potential into a less symmetric one with a vacuum.
The test whether the quasiclassical shift in field space on a selfinteracting
multiplet with a tachyonic mass term which preserves the current conservation
of the multiplett leads really to a SSB is decided in terms of Q = . This and

12
not the manipulation is the definition of SSB. In the absence of couplings to
s 1 fields this is the case but it fails in models in which the scalar matter
couples to a vector potential. As will be demonstrated in the last two sections
the fattening of the photon does not require the presence of a Higgs field;
it is rather related to the appearance of an escort field which in turn is the
unavoidable consequence of maintaning positivity in the presence of massless
vector potential.
As far as one knows Nature provides no realization of exact internal sym-
metries or SSB in particle physics beyond the particle-antiparticle symmetry;
the application remains in the hands of phenomenologists. But there can be
no doubt that Nature supports the existence of a Higgs particle without which
there can be no self-interacting massive vector mesons.
Shortly after my 1968 visit of the USP John Lowenstein, Haags last PhD
student before he left Urbana and moved to the University of Hamburg, joined
Andre as a post-doc. Their joint work on QED in two dimensions [18] impressed
me as a thorough application of mathematical ideas and concepts from LQP.
For private reasons John wanted to return to the US and I was able to be helpful
in obtaining a position at the University of Pittsburgh. This was the start of
a fruitful collaboration on perturbative renormalization theory to all orders, in
particular to gauge theory and its axial anomalies, which continued after my
move to Berlin in 1971.
The collaboration with Andre and his research group continued during the
70ies; he came twice to Berlin and I met him 3 times, twice at the PUC in Rio
and a third time after he moved to the USP in Sao Carlos. We wrote several
papers on models with conformal symmetry in particular on global operator
expansions and one of my collaborators (A. Voelkel) had a short time visiting
position at the PUC.
In the middle 70ies a class of two-dimensional models with factorizing S-
matrices became the focus of attention. These integrable models were of partic-
ular interest since perturbative constructions did not permit to establish the ex-
istence of nontrivial models so that QFT was the only area of theoretical physics
for which the existence of interacting models within its conceptual framework
(causality, Hilbert space positivity) remained widely open.
The discovery of these d = 1 + 1 integrable models led to a close collabo-
ration between group of research associates at the FU Berlin (Berg, Karowski,
Thun, Truong and Weisz) with a group around Swieca (Koberle, Kurak, Marino,
Rothe) with myself representing the link between the two. Swieca death at the
end of 1980 also marks the end of what Haag in his reminiscences called the
Brazilian connection.
A collection of Swiecas publications appeared later as obras colligidas
[19] to which I wrote a long introduction with the title From the Principles
of Quantum Field Theory towards New Dynamical Intuition from Studies of
Models. This marked the end of a decade lasting collaboration to explore
and illustrate the content of Haags LQP in concrete models of QFT. Our last
joint project to detach operator product expansion from conformal QFT and in
this way obtain a nonperturbative construction remained an unfulfilled project.

13
Recently there has been significant progress on this old problem [20].
When, I revisited Brazil 20 years later, some of Andres closest colleagues
had retired or died (Jose Giambiagi) and his former younger collaborators were
working on different problems.
The old project received a new impulse when Karowski and Weiss extended
it to what nowadays is referred to as the formfactor program which consists
in the explicit nonperturbative construction of matrixelements of fields between
particle states. Besides presenting new insights into nonperturbative QFT its
aim is to construct a QFT in terms of vacuum expectation values of quantum
fields which can be formally represented as infinite sums over product of form-
factors. As in perturbation theory there is presently no control of such sums.
Meanwhile a different approach has led to the first existence proofs for inte-
grable models in the absence of bound states. It does not use individual fields
but rather directly Haags LQP setting in terms of net of local algebras. It
is a top-to-bottom approach which starts from the observation that the mod-
ular localization theory (see next section) connects the algebraic structure of
wedge-localized algebras with the S-matrix and uses the fact that for factor-
izing S-matrices without bound states there exist simple generating operators
for wedge algebras whose Fourier transforms fulfill the Zamolodchikov-Faddeev
algebra relations.
Knowing the structure of the wedge algebra the next step is to show the ex-
istence of nontrivial algebras associated to compact spacetime regions resulting
from intersection of wedges. This is the real hard part where estimates about
degrees of freedom in the form of nuclear modularity enter [46]. The termi-
nology top-to-bottom refers to obtain algebras of compact spacetime regions by
intersection of wedge algebras. Covariant fields which generate these algebras
would appear only in a later stage of this (top-to-bottom) construction. Since
the physical consequences can be directly extracted from the algebras they are
not needed. The protagonists of these ideas belive that future existence proofs of
interacting QFTs in d = 1+3 will be based on such top-to-bottom constructions.
The remainder contains some remarks which bear no relation to physics but
which form part of my personal Brazilian connection
During the collaboration with Andre the weight of the past was always
present. Andre was born in 1938 in Warsaw/Poland. His family had the good
luck to escape from the murderous anti-semitism of the Nazis to that part of
Poland which in 1939 according to the Hitler-Stalin pact was occupied by the
Soviet from Union. Before Hitlers assault of the Soviet Union and the Nazi
occupation of the rest of Poland, the Swiecas fled to the Soviet Union from
where they succeeded to reach Vladivostoc on the transsiberian railroad from
there they got by boat to Yokohama and finally to South America.
They had some relatives in Rio de Janeiro but Getulio Vargass anti-semitic
police chief Filinto Muller created problems which forced them to remain for
some months in Buenos Aires. In the 70s Muller was the senator of the states of
Mato Grosso and leader of the Arena party which was created by the military.
I was invited several times to the house of Andres parents and on one of these
visits I sensed a mood of commotion. It was the day on which Filinto Muller

14
died in a plain from Rio to Paris. In those days the seats in many airplanes
contained a material (polyvinyl chloride) which, if ignited by a cigarette, could
lead to a smoldering fire. This happened on Mullers flight; the captain made
an emergency landing but all the passengers and those of the crew who did not
succeed to enter the captains cabin perished in the toxic fumes.
Andre and his parents were not religious, yet there was a feeling of higher
form of justice since Filinto Muller was responsible for the deportation of Olga
Benario-Prestes on a Spanish ship via Francos Spain and her extradition to
Nazi-Germany [21]. Olga, a German communist, together with the Brazilian
tenent Luis Carlos Prestes were in opposition to the dictatorship of Getulio
Vargas. Their attempt to initiate a revolt within the Brazilian military failed
and both were jailed. Muller deported Olga om a Spanish ship and the Franco
extradited her to Nazi-Germany. Being of jewish descent this was like a death
penalty. Her deportation caused national and international protests in partic-
ular since such an extradition in a state of advanced pregnancy was against
the Brazilian law. Olga gave birth to her daughter Anita Leocadia Prestes in
a Berlin prison clinic. Using her connections to the Itamaraty (the Brazilian
Foreign Office) Prestes mother succeeded to take the baby to Brazil. Nowadays
she is a professor of history at the Federal University of Rio de Janeiro.
Olga was taken to the Ravensbruck concentration camp and killed by gas in
the Bernburg Euthanasia Centre which the Nazis created years before as part
of their euthanasia program for mentally ill people. When after protests from
part of the catholic church this clandestine murderous progran was halted, the
installation was used to kill prisoners from those near by concentration camps
which, as the womens camp at Ravensbruck, had no extermination facilities.
The fate of Filinto Muller, who died the same way as Olga Benario, is re-
markable even for those who do not believe in higher justice and destiny.
The fact that I spend my childhood in Bernburg, and that I remembered my
mother wispering with neighbors about busses with painted windows arriving
at the mental hospital, constitutes an encouter with my past in a manner which
I could never have imagined. In this way this became an inexorable part of my
Brazilian connection.

4 Local Quantum Physics and Modular Local-

ization
One of the meetings between mathematical physicists and mathematicians which
I attended in the 60s was a 1967 conference in Baton Rouge, Luisiana. In the
center of attention was the work of Yuji Tomita, an elderly Japanese math-
ematician who appeared with a thick still unpublished manuscript. Its title
contained the word modular, indicating that he wanted his new results on
operator algebras to be viewed as a kind of noncommutative generalization of
measure theory. I understood very little of these new mathematical results.
Richard Kadison, an authority on operator algebras who chaired the confer-

15
ence, had doubts about some of Tomitas arguments. He encouraged Tomitas
compatriot Takesaki to review the arguments and rework the presentation of
the results together with Tomita. This led to the still authoritative first book
on Tomitas theory which became known as the Tomita-Takesaki modular the-
ory [23].
Tomitas ideas led to a new formulation in which the unitary modular group
was associated with operator algebras satisfying certain conditions. This seemed
to be connected in some way to new formulation (statistical mechanics of open
systems) of equilibrium statistical mechanics directly in the infinite volume limit
(without using Gibbs trace formula) proposed by Haag, Hugenholtz and Win-
nink [22]. Their results were also presented at this conference. What these
authors referred to as the KMS property had its much more general operator-
algebraic counterpart in Tomitas work.
The KMS property appeared first in previous work by Kubo, Martin and
Schwinger (the historic origin of the terminology). In the work of these authors
it was merely a computational trick which converted the calculation of traces in
the Gibbs formula into more manageable analytic properties involving analytic
continuations. But in the new context it acquired a foundational meaning far
beyond a mere computational device.
This terminoly and also some of its physical content was afterwards adopted
by the operator algebraists; Alain Connes and Uffe Haagerup used it in his im-
pressive classification of the type III von Neumann factor algebras. Whereas the
mathematical concepts of quantum mechanics, such as the Hilbert space and
operators acting on it, existed before its discovery, the foundations of modular
theory are the result of a joint effort between mathematicians and mathemat-
ical physicists. More details about the path-breaking Baton Rouge conference
can be found in Haags reminiscences [2] and an authoritative account about
its impact on mathematical physics including an important interrelation with
causal localization is contained in a seminal article by H.-J. Borchers [24].
A decade later the work of Bisognano and Wichmann [25] revealed that
causal localization and the KMS thermal aspects are inexorably interconnected;
subsequently Geoffrey Sewell pointed out that this interrelation plays a fun-
damental role in the understanding of Hawkings black hole radiation and the
Unruh effect [26]. An account of the Hawking radiation from the viewpoint of
LQP can be found in [27].
The interest in the application of LQP to problems in curved spacetime is
reflected in an increasing number of publications starting in the 90s. A recent
review with references to earlier work can be found in [29].
In this context it is also interesting to note that modular localization sheds
some new light on a fascinating but for a long time incompletely undertood
controversy between Einstein and Jordan (the Einstein-Jordan conundrum)
which led Jordan to the first model of a field theory (the model of conserved
current in d = 1 + 1).
Haags view of localized quantum matter as a net of causally localized op-
erator algebras acting in a joint Hilbert space received important support from
the modular theory of operator algebras [28]. A particularly fruitful concep-

16
tual enrichment came from the application of modular localization to integrable
models6 and to the use of Wigners representation theory of the Poincare group
for the construction of noninteracting nets of operator algebras [30]. Since this
concept plays an important role in the later sections an at least rudimentary
understanding will be helpful.
There exists a weaker version of the T-T modular theory which does not refer
to operator algebras but uses the concept of a so-called standard subspace K of
a Hilbert space H. This is a closed real subspace K H whose complexification
is dense in H i.e. K+iK= H and KiK = {0} where the bar referes to the
closure.. The Tomita S operator is then defined as S( + i) = i; conversely
K can be represented in terms of a Tomita operator K = ker(S 1). As in the
algebraic Tomita-Takesaki setting the S has a polar decomposition S = J1/2
where it is an automorphism of K and the antiunitary J transforms K into its
symplectic complement (symplectic orthogonal) of K within H which is defined
in terms of the symplectic product i Im(f, g).
The simple physical illustration of the connection of causal localization with
the T-T modular theory is provided by the 2-point function of a scalar free field
Z
(f, g) = ((f ), (g)) , (f ) = (x)f (x)d4 x (4)

defines a scalar product between the forward mass-shell restriction of the Schwartz
test functions. The Hilbert space H is the space of Wigner wave functions H
of a scalar particle which is obtained from the closure of the forward mass-shell
restriction of the Fourier transformed test functions.
The closed subspace K(O) of H obtained by the closure of real test func-
tions localized in O turns out to be standard in the above sense; this follows
from the one particle projection of the cyclic and the separating property of
quantum fields known as the Reeh-Schlieder property [10] (or can be shown
directly). Since i Im(f, g) can be written in terms of the vacuum expectation of
the commutator
i Im(f, g) := ( [(f ) , (g)] )
the aforementioned symplectic orthogonality receives a physical interpretation
in terms of Einstein causality.
More interesting and important is the inversion of this relation i.e. the
construction of a net of causally localized subspaces K(O) HW ig of the
Wigners representation space using his representation theory of the Poincare
group. The key for this construction is the Bisognano-Wichmann property i.e.
the physical identification of the Tomita operator SW for a wedge region e.g.
W = {x3 > |x0 |} In this situation these authors showed (under mild technical
assumptions in a Wightman setting [25]) that the antilinear S-operator associ-
ated to the dense set of states obtained by applying a wedge-localized operator
algebra to the vacuum can be expressed in terms of phy,sical data. Whereas
6 For a recent account containing a rather complete list of references to previous work see

[46].

17
the unitary modular group it associated to the radial part of its polar de-
composition S = J1/2 is the W -preserving boost it = U (W (2t))), the
antiunitary angular part J is, apart from a -rotation in the plane of the edge,
the TCP operator which plays a funde,ental role in QFT.
With this physical identification one obtains the modular subspace K(W ),
and by covariance also all its Poincare transforms. Subspaces K(O) for general
localization regions O are obtained in terms of intersections, their modular
groups have no geometric interpretation (except in the presence of conformal
covariance); although they preserve the localization region their action inside is
fuzzy i.e. cannot be visualized in terms of geometric transformations inside O.
Using the functorial relation between real one-particle subspaces and operator
subalgebras, which is defined in terms of Weyl operators, one finally arrives
at an explicit construction of Haags net of local algebras in the absence of
interactions [30].
This functorial relation which maps localized real subspaces into local von
Neumann algebras in such a way that subregions correspond to subalgebras and
Einstein causality holds permits a generalization to all positive energy Wigner
representations. The fact that it is tied to the energy positivity shows s (per-
haps somewhat unexpected) close connection between geometric with spectral
properties of the translation operators (stability properties).
This relation breaks down in the presence of interactions. In this case one
may start from the Wigner-Fock space which in the presence of a gap is provided
by (LSZ or Haag Ruelle [10]) scattering theory. It has been known for a long
time that in this case the TCP operator differs from that of a free incoming
fields by the scattering matrix Sscat so that one obtains J = Sscat J0 where J0
refers to the free fields associated to the Wigner-Fock space.
An explicit construction of modular localized subspace in the presence of
interactions is possible for integrable models with known factorizing S-matrix.
An important supporting property is the existence of so called vacuum polar-
ization free generators (PFG) i.e. operators in an interacting theory whose
application to the vacuum creates a polarization-free one-particle vector. Their
existence is based on the relation between the tightness of causal localization
and the strength of interaction-caused vacuum polarization clouds. It is well-
known that the singular nature of states created by interacting quantum fields
is related to the strength of vacuum polarization clouds; the larger the space-
time localization region conceded to the clouds, the easier to find less singular
operators.
It had been known for a long time that covariant point-local fields which
create a polarization-free one particle state from the vacuum must be free fields
(the J-S theorem [32]). The more general concept of vacuum polarization free
generators (PFG) leads to theorem that for compact localized spacetime regions
such PFGs do not exist in interating theories. The tightest noncompact region
for which (under certain weak condition) PFGs exist are (arbitrarily narrow)
space-like cones [33]. The fact that they are always available in wedge regions
[34] makes the latter an ideal point of departure for existence proofs.
In the case of integrable models such PFGs are provided by the Fourier trans-

18
forms of the creation/annihilation operators which obey the Zamolodchikov-
Faddeev algebra [35] whose commutation structure is given in terms of the
known elastic part of the factorizing S-matrix. This observation is the starting
point of an LQP-based construction project starting from the PFG-generated
A(W ) operator algebra. A highly intricate part of the construction is the demon-
stration of nontriviality of double cone intersections of wedge algebras [46]. Here
concepts of cardinality of degrees of freedom in the form of modular nuclearity
play an important role.
The obtained results are complementary to those of the form factor program
for integrable models . Whereas the latter lead to concrete closed-form expres-
sions for formfactors of point-local fields (but without control of the convergence
of the resulting infinite series expressions for the correlation functions), the LQP
construction starts from the generators of the wedge algebra and establishes the
existence of a nontrivial double cone intersection of arbitrary small diameter
(but falls short of constructing the generating point-local fields).
Interacting models in dimensions d > 1 + 1 are not integrable and hence pos-
sess no closed form (analytically representable) solutions. Whether the exten-
sion of ideas based on modular localization to d = 1 + 3 dimensional interacting
models will lead to a nonperturbative control remains a dream of the future.
Different from all other areas of theoretical physics QFT remains an enigmatic
project.
QFT earned its standing as the most comprehensive description of Natures
physical properties from the observational success of its perturbative formu-
lation. The predictive success of the Standard Model is based on low order
perturbation theory complemented by phenomenologically supported propos-
als.
Contrary to a widespread misconception renormalized perturbation theory
does not depend on any quantization parallelism with classical field theory. As
shown in [31] covariant point-local (pl) free fields are constructed from Wigners
theory of unitary positive energy representations of the Poincare group; the
corresponding spaces of particle wave functions bear no relation to actions of
classical point-local particles (section5).
In terms of the creation/annihilation operators a# (p, s3 ) for massive parti-
cles and their anti-particles b# (p.s3 ) which act in a Wigner Fock space, the pl
covariant free fields are of the form
s
1 d3 p
Z X
A,B ipx A,B ipx A,B
(x) = 3/2
(e u (p, s 3 )a (p, s 3 )+e v (p, s 3 )b(p))
(2) s =s
2p0
3
(5)
A,B
where the intertwiner functions u (p, s3 ) and their charge-conjugate counter-
part are (2A + 1)(2B + 1) component which intertwine between the unitary
(2s + 1)-component Wigner representation and the covariant (2A + 1)(2B +
1) dimensional spinorial representation labeled by the semi-integer A, B which
characterize the finite dimensional representations of the covering of the Lorentz
group SL(2, C). The intertwiner functions are determined in terms of group
theoretic properties; the use of modular localization is not necessary.

19
 There is one annoying loophole in this construction in that the important
massless vector potential and more generally tensor potentials do not exist in
a point-local form since they violate positivity 7 . In that case the way out
has been to quantize classical gauge theory. The problem with this is that
Hilbert space positivity, which is classically irrelevant but indispensible for the
probabilistic properties of quantum theory, is violated in the quantized result. It
can be recovered only for the gauge invariant part of the theory which excludes
the important matter fields but includes the local observables in the form of
gauge invariant fields. There exist no perturbative approach based on point-
local fields which is able to avoid the use of unphysical fields. This requires
the introduction of additional indefinite metric degrees of freedom and ghost
fields which have no counterpart in the classical theory but are necessary to
implement the operator gauge transformations; the latter bear no relation with
physical symmetries but are nevertheless needed in order to extract the physical
quantities from an unphysical setting.
The physical reason for being forced to take recourse to quantum gauge
theory is that there is a clash between positivity and localization of which the
problem with massless pl free vector potentials is only the tip of an iceberg. It
also manifests itself in the nonexistence of massless conserved pl currents for
s 1 as well as that of massless energy-momentum tensors for s 2 [37].
In the presence of interactions its manifestation affects even massive QFTs in
that it is the cause of the nonexistence of positivity preserving renormalizable
interactions involving s 1 fields. Instead of combatting this phenomenon by
short distance improvement resulting from compensations of part of the positive
probability with negative metric contributions the positivity preserving way of
improving short distance scale dimensions of fields is a relaxation on tightness
of causal localization by passing from pl to sl free fields.

5 String-localized fields
Point-local free fields for spin s or helicity h are uniquely determined in terms of
their covariance. Massive tensor fields of spin s have short distance dimension
dsd = s + 1. Interaction densities L are defined in terms of Lorentz-invariant
Wick-ordered products of free fields and according to the power counting bound
of renormalizability dsd (L) 4 there are no renormalizable interactions with
s 1. Positivity obeying massless point-local tensor fields of helicity h 1 and
tensor degree h do not exist; this is a consequence of the absence of intertwiners
from massless helicity h unitary Wigner representations to (h/2, h/2) covariant
tensor fields.
Both problems are related to the positivity of pl fields which is in turn
a consequence of the unitarity of Wigners representation theory. For s = 1
this problem is formally solved by replacing the Hilbert space by a positivity-
violating indefinite metric Krein space which lowers the dsd of the Proca field
from 2 to 1. The indispensable positivity property is then recovered for the
7A similar problem exists for massless fermionic fields for s 3/2.

20
subtheory of local observables (which includes field strengths) whereas the im-
portant charge-carrying fields, which relate the causality principle with particles,
remain outside gauge theory. Hence gauge theory, although in its classical form
a complete theory with local gauge invariance, is an incomplete QFT in which
gauge symmetry plays the role of a formal device whose only purpose is to filter
the physical subtheory from an unphysical (negative probabilities containing)
description.
The new stringlocal field theory (SLFT) is a complete QFT whose construc-
tion is based on the observation that the culprit for the indefinite metric and
the resulting lack of positivity of probabilities is the use of covariant pl fields.
As soon as one uses their covariant sl siblings in a way which is consistent with
their weaker localization one is led to the beginnings of a new renormalization
theory in which s = 1 (and more generally s > 1) fields have a spin-ndependent
short distance dimension (dsd (bosons) = 1, dsd (f ermion) = 3/2) and thus per-
mit the formation of tri- or quadri-linear interaction densities L with the power
counting bound (pcb) dsd (L) 4.
The naturalness of sl fields follows from a theorem in LQP [10] which
states that in the presence of a mass gap there exists for each particle type an
interpolating sl field8 . From SLFT one knows that interactions containing s 1
fields lead (apart from pl observables) to sl interacting fields. The terminology
QFT and in particular LQP always refers to positivity maintaining descriptions;
indefinite metric decriptions will be referred to as gauge theory (GT).
The Wightman setting of QFT and Haags LQP cannot dispense with posi-
tivity; its absence does not only affect the probability interpretation but gauge
dependent fields also fail to describe the correct causal localization. In order
to solve the positivity problem it is important to understand the relation be-
tween tightness of localization and short distance dimensions in more detail.
Starting from a massive dsd = 2 Proca vector potential AP it is easy to see
that the covariant string-local solution of the operator-valued differential 2-form
F = AP v A
P

Z
A (x, e) = F (x + e)e d, e2 = 1 (6)
0

Here the linear form of the space-like string and the Lorentz transformation
of e in the covariant secures the covariance of sl fields and the integration to
infinity insures the lowering of the dimension from dsd = 2 to 1.
By starting from a massive general spin s tensor field and forming the field
strength which corresponds to a 2s-form the s-fold repetition of the line inte-
gration results in a dsd = 1 e-dependent string-local counterpart while their
iterative application to the pl degree s tensor potential define the s tensorial
escorts of maximal degree s 1 [36]. A similar idea applied to the point-local
spinor-tensor potential of half-integer spin s (one spinor- and s 1/2 tensor
indices) leads to a similar situation in which the resulting string-local spin s field
8 In the algebraic setting of LQP this corresponds to an interpolating operator which belongs

to an algebra of an arbitrary narrow space.like cone (whose core is a space-like string).

21
has the same dimension as a s = 1/2 Dirac field namely dsd = 3/2, independent
of s. By taking a more general integration measure d ()d one can vary
the dsd continuously down to zero.
Only the so constructed massive sl tensor potentials have smooth massless
limits9 . This does not only lead to the sl replacement of the missing pl massless
potential but it also defuses a No-Go theorem by Weinberg and Witten claiming
that massless energy-momentum tensors do not exist for s 2 [37]. The correct
statement is that pl conserved massless E-M tensors do not exist ; they have to
be replaced by sl E-M tensors which are different as densities but lead to the
same global charges (generators of the Poincare group).
One may think that the use of sl instead of pl fields which converts pcb
violating pl interaction densities into pcb obeying sl ones renders a model renor-
malizable. But as often in QFT, trying to patch up a problem creates another
problem at an unexpected place.. Without the fulfillment of an additional con-
dition, which prevents the total delocalization at higher orders, the validity of
the pcb is insufficient to guaranty consistency. This additional requirement will
be addressed in section 6.
Covariant sl fields can be constructed in a rather elementary way from their
pl counterparts without referring to the more foundational LQP. But overlooked
simple constructions arise sometimes in a roundabout way. The study of sl
fields did not start in the above form but rather developed in the aftermath
of solving the foundational problem, more than 7 decades old, of the causal
localization of Wigners infinite spin matter for which it was essential to use
modular localization theory.
In [30] it was shown that all positive energy representations are localizable
in arbitrary narrow (noncompact) space-like cones. Since it is well known that
the massive and finite helicity zero mass class is pl generated and that the
generating fields of the infinite spin representations cannot be pl Wightman
fields [38] it seemed likely that their generating fields are localized on the semi-
infinite string-like cores of a space-like cones. Using the modular localization
of the LQP setting it was possible to construct the intertwiner functions u(p, e)
which relate the momentum space Wigner creation/annihilation operators with
covariant sl fields [39]. Previous attempts in terms of Weinbergs group theoretic
method based on covariance had failed.
Meanwhile there appeared a rather sophisticated direct proof which excludes
the existence of nontrivial compact modular localized Wigner-Fock subspaces
[41]. It uses the spatial version of modular localization (the K-spaces) sketched
in the previous section . This raises the question about possible physical prop-
erties of those fields. The new setting of sl perturbative renormalization theory
strongly suggests that this infinite spin matter is inert with respect to interac-
tion with normal matter. Matter which only exists in the form of free fields and,
through the use of its energy momentum tensor in Einstein-Hilbert equations,
9 What is meant is that the 2-point massive correlation functions converge to those of the

massless helicity fields but the representations of the operators of course remain unitarily
inequivalent

22
may lead to backreactions on the gravitational field, is an interesting candidate
for dark matter since its coldness is natural [42]
The same method of modular localization applied to Wigners unitary repre-
sentations of ordinary matter led to the rather large class of massive and finite
helicity massless sl fields which also can be directly constructed in terms of
semi-infinite line integrals over pl fields.
The next section addresses the question of interest for many readers are
string-localized fields related to ST theory?.

6 21st century physics, or, the new phlogiston?

The naturalness of string-localization in LQP and its generic appearance in all
positivity preserving renormalizable interactions involving s 1 particles begs
the question of its relation to string theory (ST).
To understand how particle physicists arrived at ST it is helpful to recall
its historical roots which can be traced back to ideas about an autonomous
S-matrix theory. This refers to attempt to formulate a theory of scattering
amplitudes without the use of its large time asymptotic relation with QFT. The
problem of such a project is that causal localization principles of QFT are not
available in such a direct construction of global on-shell objects.
A possible way out was to look for analyticity properties which generalize
those of the dispersion relations. The most conspicuous model-independent
property of on-shell restrictions of Feynman diagrams is the analytic crossing
property. In order to separate this property from its (for strong interactions
useless) perturbative context Mandelstam proposed a representation for the
elastic scattering amplitude which incorporates such a property.
The historical origin of ST cannot be understood without Venezianos sub-
sequent Dual Model which replaces Mandelstams representation by a concrete
crossing symmetric meromorphic functions which substitutes the elastic scat-
tering continuum by a trajectory of particle poles.
ST started by viewing such on-shell particle mass trajectories as manifesta-
tions of strings in spacetime in analogy with the energy spectrum of a chain of
quantum mechanical oscillators. This new spacetime interpretation implied a
return to an off-shell description based on the quantization of actions of world
sheets traced out by strings in spaetime. The interaction between such strings
was assumed to be described in terms of splitting and recombining tubes repre-
senting world sheets.
The positivity requirement on this quantization selected one model in which
the spacetime was the target space of a certain d = 1 + 1 conformal field theory
associated to a 10-component supersymmetric current. Our 4-dimensional world
was to result from a Kaluza-Klein dimensional reduction.
Since Haags LQP comprises all models which fulfill the causal localization
principles in a (positivity obeying!) Hilbert space setting and ST falls according
its protagonists into this category, the obvious question is whether the objects
of ST describe, as string theorists claim, string-localized objects in the sense of

23
causal localization in Minkowski space. If localization in ST really means what
the terminology suggests, two string operators should commute if the strings
are spacelike separated (the quantum version of Einstein-causality); there is no
other physical meaning which one can attribute to quantum strings localized in
spacetime.
Freed from a quantization parallelism to classical physics, the LQP formu-
lation is synonymous with a realization of causal localization principles in the
context of quantum theory which means in particular that string-local operators
are defined as objects in spacetime which are causally localized i.e. two string
operators commute if they are relatively spacelike separated.
Causal localization is inexorably connected to vacuum polarization and the
strength of the vacuum polarization clouds depend on the tightness of localiza-
tion. This affect in particular the short distance scale dimensions of pl fields. If
the alleged stringy objects of ST bear any relation to spacetime strings they
must be related to the sl fields of LQP even if they had been constructed in a
different way from that of sl fields. The main point of contention is whether the
objects of ST are really string-local in any with relativistic causality compatible
sense.
In order to understand that string theorists use the terminology string for
something which bears no relation with localized quantum objects in spacetime
it is helpful to look at what they are doing and understand why they think
they are addressing propertie of quantum localization. Before addressing the
quantization of the Nambu-Goto action or constructing their 10 dimensional
superstring model from the action of a particular 10 component supersym-
metric d = 1 + 1 conformal current model string theorists it is helpful to take
a critical look at their view of the quantum theoretical counterpart of particle
world lines [43].
The model is defined in terms of the relativistic action ds2 but the re-
sulting covariant classical world line has no quantized counterpart since particle
operators ~ q (t) only exist in (nonrelativistic) quantum mechanics (the nonin-
trinsic Born localization) and the quantum theoretical description of a single
relativistic particle uses Wigner representation theory. From the latter one can
construct free fields which and the point-local free fields can be reformulated
in terms of a relativistic action. There is simply no access to wave functions
of relativistic particles in terms a quantization of actions describing relativistic
world lines and hence this construction turns out to be a squib load.
The theory which describes relativistic particles is Wigners construction
of unitary representations of the Poincare group which cannot be accessed by
quantization of classical actions; in fact his 1939 unitary representation theory
was the first successful intrinsic quantum construction of a relativistic particle
theory. As we know nowadays this theory already containes the germ of causal
localization10 in the form of modular localization of positive energy states which
10 Wigner tried ito find a representation theoretical signal of causal localization and became

disappointed when he realized that the Newton-Wigner localization did not solve the prob-
lem [2]. The conceptual prerequisites for the later modular localization did not exist at that
time.

24
is closely related to the causal localization of fields.
Only on this level of causal localization of fields can one make contact with
the quantization of pl fields (section 3). The more generic and important co-
variant sl fields cannot be accessed in this way (section4). They are objects
which are pure quantum in that the umbilical cord of an alleged quantization
parallelism has been cut. This is our main motivation for giving much space to
causal localization in an article dedicated to the memory of the protagonist of
LQP which places causally locaiized operator algebras into the center stage.
This leaves the question of what remains of ST if it is not a theory of quantum
strings in spacetime. An authoritative answer from somebody who has spend
a good part of his professional life to understand the physical content of the
Nambu-Goto action is that it describes an infinite set of conserved charges as
one finds in d = 1 + 1 integrable QFTs. But different from integrable d=1+1
QFT there is no trace of anyspacetime localization in N-G models [44].
The fusion and splitting of world sheets as a description of spacetime strings
in analogy to the interpretation of perturbative Feynman graphs as coalescing
and splitting of point-like particles represents an attempt of string theorists to
create localized interactions in terms of classical metaphors. On the other hand
the fact that this is based on misunderstandings of quantum causal localization
does not invalidate the mathematical use of such constructions as an inspiration
for interesting topological, algebraic and geometric constructions. ST also led to
some new computational techniques which are useful in other areas of particle
theory. If it did not prevent careers by occupying many research position and
distract many from problems of particle theory it would be easier for physicists
to appreciate its mathematical contributions.
One reliable result which was obtained by string theorists, although not
related to string localization, is a theorem by Brower [45]. It states that the ir-
reducible superstring algebra, defined in terms of the aforementioned supersym-
metric 10 component conformal field theory, carries a positive energy Wigner
representation which decomposes into a an infinite direct of sum of irreducible
(m > 0, s) and (m = 0, h) Wigner representations.
This has an interesting connection with an old project by Majorana. In anal-
ogy to the description of the discrete spectrum of the hydrogen atom in terms
of a O(4, 2) representation, Majoranas idea was to construct a group algebras
of a higher dimensional group which contain a tower of particle wave function
spaces. This idea underwent a revival in the 60s in the form of dynamical
groups leading to a discrete spectrum of particles. Apart from the fact that
the irreducible superstring algebra associated to the conformal field theory is
not a group algebra, Browers theorem is similar in that it refers to a particular
particle spectrum which originates from the action of the Poincare group on an
irreducible algebra.
In the eyes of string theorists the map of the two-dimensional conformal space
into the 10 dimensional target space describes what they call a string in form of a
world sheet defined in terms of a map from the two-dimensional conformal space
into the 10-dimensional target spacetime. Without these string glasses one
only sees a discrete direct sum of unitary Wigner representation (but no target

25
space localization) whose conversion into covariant free fields leaves the choice of
pl or sl. As for any unitary positive energy Wigner representation which carries
a modular localization structure it is the interaction which decides about the
localization: renormalizable interactions of s < 1 require the use of pl fields
whereas renormalizability and positivity in the presence of s 1 fields can only
be maintained in terms of string-localization.
The problem of localization of fields has nothing to do with that particles;
the latter remain what they always were: states described by in time dissipating
Wigner wave function which, as a result of their positive energy content, do not
admit a causal pl or sl localization. What may be idealized as a pl event is
the registering in a counter; the difference whether the fields whose application
to the vacuum create these states were sl or pl only manifests itself in a more
spacetime spread out region of clicks. It is important to have a clear view of
the relation between fields and particles in order to understand Haags stomach
ache with string theorists view of strings and particles (see below).
The heuristic picture of ST in terms of splitting and recombining world
sheets has led, particularly in the hands of Ed Witten, to highly interesting
new ideas and results in geometry and topology but this has not helped to its
physical use. Concepts as that of modular localization which are the raison
detre for local quantum physics have remained outside ST and its derivations
(extra-dimensions, AdS-CFT).
The promise to address the issue of string-localization (which is the origin of
the terminology ST) has remained unfulfilled and there is no way in which this
can change. It is simply not possible to create a new theory without a foun-
dational dispute with the in every aspect successful and comprehensive QFT.
Mathematical enrichments cannot hide the fact that the physical contributions
of ST to particle theory has remained smaller than any preassigned epsilon.
Historian of physics who would seriously attempt to take stock of viable the-
oretical physics concepts which originated within 50 years of ST will presumably
have a hard time to account for what had been achieved. Haags reaction to the
present situation in this respect is quite interesting.
On the occasion of presenting a seminar talk more than three decades after
having held a visiting position at the university of Princeton, he was hard pressed
by Ed Witten to join the ST community. Haags recollection of this situation
can be found in his published reminiscenses [2]. He writes:
I visited Princeton in the early 90ies. At that time Sam Treiman was head
of the physics department at the university. I had known him since 1958 and
highly appreciated his sober judgement. So I asked him about his assessment
of the future of string theory. He said that he had not occupied himself with
it but that he was supporting it without reservation because the people who
worked on it were very very good. He meant primarily Ed Witten who was now
the spearhead of this approach. I had been asked to give a physics colloquium
talk about my views on quantum gravity and hoped to have some discussion
with Ed Witten. Next morning he greeted me by saying: Your talk was very
interesting but I would really advise you to work on string theory. When
he saw the somewhat incredulous look on my face he added I really mean

26
it. I shall send you the manuscript of the first chapters of our book. This
ended our discussion. Back in Hamburg I received the manuscript but it did
not convert me to string theory. I remained a heathen to this day and regret
that meanwhile most physics departments believe that they must have a string
theory group and have filled their vacant positions with string theorists. To be
precise: It is good that people with vision like Ed Witten spend time trying to
develop a revolutionary theory. But it is not healthy if a whole generation of
young theorists is engaged in speculative work with only superficial grounding
in traditional knowledge.
Haags critical comments should be seen in the context of his conduct of
research which is distinguished by self-reliance and a self-critical scrutinizing.
This may have had its origin in the circumstances in which his interest in the-
oretical physics arose. As a teen ager he was on a private visit to his sister
who lived in the UK when in 1939 the war started and he could not return to
his mother in Stuttgart and finish high school (in 1939 his age was 17). As a
German citizen he was shipped to Canada where he spend the years of war in
a detention center. There he managed to get hold of a book on physics which
he used for self-studies.
Returning at the end of the war from the Canadian camp to Stuttgart he
found himself in a war-devastated city without a functioning academic teach-
ing program. In such a situation self-reliance and intellectual autonomy were
important.
To find his former strongly independent minded colleague from Princeton
Sam Treiman three decades later in a state of dependence on authorities con-
cerning a subject which he considered of prime importance was apparently some-
what unexpected for Rudolf Haag.
The reaction of Haag to both Witten and Treiman can be best commented
in form of a metaphor: it is not enough to believe to have discovered the Lapis
Philosophorum, one must also have the charisma to convince sufficiently many
prestigious persons to share this belief.
Haag was hardly impressed by mathematical work whose motivation did not
originate from fundamental physical problems. The situation of QFT after the
discovery of perturbative QED, in which different prescriptions of exorcising
infinities amazingly led to mutually compatible results, certainly motivated
him to look for a more coherent description which finally culminated in his
framework of what he later referred to as local quantum physics (LQP). He was
convinced that all properties of causally localized quantum matter was encoded
in the relation between observable algebras A(O) labeled by their spacetime
localization region O. His innovative strength resulted from his ability to find
the appropriate mathematical setting for his physical ideas. For more detailed
mathematical knowledge he relied often on mathematically more knowledgeable
collaborators.
The mere fact that ST did not arrive at any observationally testable proposal
throughout its 50 years of existence (which is its common critique) was of not
much concern to Haag. One can assume that both he and Witten shared the
belief that foundational ideas should have all the time they need to evolve.

27
 He would however have expected that the exploration of a foundational idea
should lead to a steady increase of knowledge about important theoretical prob-
lems of particle physics. His LQP led to a profound understanding of why the
local structure of QFT is much more powerful than its classical counterpart.
Together with his collaborators Sergio Doplicher and John Roberts he derived a
classification of internal symmetries and the absence of parastatistics as conse-
quences of properties of superselection rules which in turn were obtained from
localization structure of local observables. Theorems in Wightmans formula-
tion of QFT [32] have their counterpart in LQP and some properties (including
the causal completion property and Haag duality) permit no natural formula-
tion in Wightmans field theoretic setting. Most of the results were obtained by
a few researchers; the number of people working on foundational problems of
QFT in the 80s was rather small compared with that in ST.
On the other hand ST led to hot topics which produced thousands of
publications and provided university positions to their authors who in many
cases obtained their positions because they were working on such a fashionable
topic.
The formation of such transient fashions is reminiscent of bubbles in the
financial market but it is presently not clear to me whether this is the conse-
quence of the increasing dominance of financial capitalism in all areas of life
or whether this has its more specific explanation in the seducing charisma of
the protagonists of ST; probably it is the result of a Zrizgeist in which both
interplay.
An example of such a bubble in the wake of ST is the physical use of the
mathematically correct AdS-CFT isomorphism. Fronsdal observed already in
the 60s that the spacetime symmetry group of the so-called anti-de Sitter space-
time is isomorphic to the symmetry group of a one dimensional lower con-
formally invariant spacetime. As a result of a presumed connection between
five-dimensional gravity with gauge theories in four spacetime dimensions the
problem of a possible QFT isomorphism behind the AdS-CFT group theoretical
relation returned in the late 90s.
Since the mismatch between degrees of freedoms in comparing QFTs in dif-
ferent spacetime dimensions renders the use of fields unsuitable, the existence
of a AdS-CFT isomorphism was finally rigorously established within the alge-
braic LQP setting [7]. The proof showed that the mismatch of the cardinality of
degrees of freedom between isomorphically related QFTs in different spacetime
dimension only affects the causal completeness property, which is the quantum
counterpart of the hyperbolic propagation of classical Cauchy data.
As mentioned before (section 2) this problem is not limited to this particular
isomorphism but it affects all problems related to transplanting quantum
matter between spacetimes of different dimensions; metaphorically speaking the
resettling from higher to lower dimensions suffers from overpopulation whereas
in the opposite direction it causes anemia in that there are not enough degrees
of freedom to sustain AdS fields. The most appropriate way is to express the
isomorphism in terms of localized operator algebras [7].
Methods based on quantization of actions are not suitable for the study of

28
such isomorphisms because the notion of cardinality of degrees of freedom has
no counterpart in classical or semiclassical field theory and therefore tends to
be overlooked. This mismatch of cardinality of degrees of freedom removes the
rug from underneath the idea of extra dimensions.
This situation reveals a dilemma of present foundational theoretical research.
The increasing number of researchers in particle theory does not seem to lead
to a broadening of topics and an increase of critical knowledge: it rather tends
to favor monocultures and a loss of past knowledge and wisdom.
It is interesting to compare the present situation with that which Haag met
in the 50s during his stay at the Niels Bohr Institute in Copenhagen ([2] page
269) and which still dominated the scientific discourse during the 60s. This was
the high time of the European Streitkultur in which different views about
problems and the elimination of incorrect or misunderstood ideas as well as
what new directions to take was hammered out in often heated and sometimes
even polemic disputes between equals.
After the discovery of quantum theory in Europe different univerities of-
ten represented different schools of thought (the Copenhagen interpretation)
which led to rivalries and sometimes even to polemics between the protagonists
of these schools. When the political situation worsened many of the leading
scientist left for the US and this rivalry spread to the US. In the 50s and 60s
the discourse was dominated by individuals as Pauli, Jost, Kallen, Landau,
Lehmann, Feynman, Schwinger to name just a few.
A positive effect of this often somewhat rough way of communicating was
that futile or erroneous ideas (the S-matrix bootstrap, peratization, Heisenbergs
spinor theory, Reggeology, SO(6),..) could not survive for more than a decade.
In highly speculative research as particle theory the occurrence of wrong turns
is inevitable and therefore the existence of a lively Streitkultur is important.
In such a climate the survival of theory bases on misunderstood or evenfor more
with a Nearly all our important theoretical results and computational tools,
which later became household goods, originated in those times.
Compare this with the legacy of 5 decades of ST; apart from some new
calculational techniques and an enrichment of certain ares of mathematics it is
hard to find any remaining contribution to particle theory. ST and its legacy
appears increasingly as a gigantic bubble in particle theory which has led to extra
dimensions, branes, M-theory,...which contradict basic properties of QFT. The
more damaging legacy of this bubble is the incorrect view of the field-particle
relation which ignores previously gained wisdom.
It is interesting to quote Haag on this matter [2]. In many popularized
presentations the starting point of string theory is explained as the replacement
of the fundamental notion of particles with its classical picture of a point in
space or a world line in spacetime by a string in space respectively a sheet in
spacetime. This, I think, is a misunderstanding of existing wisdom. First of all,
paraphrasing Heisenberg , one may say Particles are the roof of the theory, not
its foundation. Secondly points in space cannot be defined as the position of
particles in a relativistic theory.
The understanding of the relation between fields and particles is one of the

29
most important and subtle achievements of the 50s and 60s. Quantum fields are
the carriers of the foundational causal localization principles and are generally
not objects of direct observations11 .
In particular the correct formulation of string-like localization of fields does
not imply that particles become stringy. Covariant free fields exist in pl as
well in a sl form; applied to the vacuum pl and sl fields create states in the
same Wigner representation. Their difference only shows up in interactions;
in particular renormalizability requires to use sl s 1 fields in the interaction
density (section 6) and higher order interactions transfers this sl localization to
the originally (in lowest order) pl s < 1 fields.
There naturalness of sl localization is supported by a theorem ([10] section
IV.3) which states that in models with a mass gap and local observables the
asymptotic particles and their scattering matrix can be described in terms of
the large time asymptotic behavior of operators which are localized in arbitrary
narrow spacelike cones whose cores are spacelike strings. Taking this theorem
from its algebraic LQP setting to that of QFT formulated in terms of covariant
fields it states that in such a theory the Wigner particles can be described in
terms of interpolating covariant sl fields.
What was not known at the time when Haag wrote his reminiscences was
that positivity violating local gauge theory can be reformulated in terms of a
positivity obeying sl theory and that the idea of positivity preservation requires
the use of sl s 1 fields in all interactions involving such fields. Viewing local
gauge theory as a prick in the flesh of QFT he certainly would have appreciated
this recent insight. It strongly suggests that sl is the standard situation and
interacting pl fields are limited to s < 1 interactions.
The state space generated by charge-carrying fields coupled to photons has
a more complicated particle structure (infraparticles) than a Wigner-Fock
particle space and its description remains essentially unknown, although there
exist efficient momentum space descriptions for photon-inclusive cross sections
(Bloch-Nordsiek, Yenny-Frautschi-Suura [61]). The loss of foundational knowl-
edge on the road to a theory of everything bodes ill for the future of particle
theory.
Research at the frontiers of particle theory is an intrinsically highly specula-
tive intellectual activity. According to one of Feynmans allegorical comments
it is sometimes necessary to dive into the blue yonder but, as he continues
to point out, such jumps should be only undertaken from a platform of solid
knowledge of QFT, so that one can return and try other directions instead of
getting lost for the rest of ones life in a hopeless project. In his last years of his
life he saw the problems originating from the popularization of ST but he was
unable to influence its course.
For the first three decades of post-renormalization QFT it was possible to
make important discoveries without deep conceptual investments. With some
basic knowledge about computational techniques of QFT and a heuristic under-
11 However their quasiclassical approximations (usually expectation values in coherent

states) are measurable in QED (the massless photons are important).

30
standing of the field-particle relation one could make important discoveries by
pulling up ones sleeves and starting a calculation and, if necessary, correcting
it or trying other directions.
It was not important whether a consistent and interesting-looking result was
derived from a fully correct theory since there was always the possibility to con-
sider incomplete or faulty theoretical ideas which led to important discoveries
as a temporary placeholder and hope for a future more appropriate understand-
ing. In this way Dirac discovered antiparticles within the less than correct hole
theory.
This way of conducting research was exhausted at the end of the 70s. ST
and its derivates are the result of attempts which tried to extend this success
without making new conceptual investments. According to Phil Anderson the
overwhelming success of particle theory in its first decades of existence had cre-
ated a kind of intellectual arrogance about Nature. It was easier to speculate
about how to go beyond QFT and claim to arrive at a theory of everything than
to do the hard work necessary for the understanding of the deeper conceptual
layers of our most successful and comprehensive theory of the material nature
of the world. The superficial image of QFT which the leading influential repre-
sentatives of ST painted and transmitted was that of old QFT being replaced
by ST.
The title of this section contains the term phlogiston which in pre-oxygen
times represented a substance which allegedly escapes in the process of burning.
The phlogiston theory only disappeared when Lavoisier at the time of the French
revolution discovered oxygen and its role in combustion. ST cannot disappear
in this way because unlike phlogiston it has no observables consequences. As
long as there are renown scientists (including bearers of Nobel prizes) among its
protagonists it will persist. The times in which it was possible to clarify issues
in disputes between equals as in the old European Streitkultur are long gone.

7 String-local perturbation theory

In the absence of interactions massive pl fields and their sl counterparts are two
physically equivalent ways of coordinatizing a free LQP model; in particular the
sl fields maintain the asymptotic relation between fields and particles which is
the basis of time dependent (LSZ, Haag-Ruelle) scattering theory. The Wigner-
Fock particle structure breaks down in case the interaction involves massless
s 1 fields.
Haags LQP and in particular the concept of modular localization played
an important role in raising awareness about the important role of covariant sl
fields and led to their first constructions [39]. It turned out that, apart from
fields associated to Wigners infinite spin representation class, all covariant sl
free fields can be directly constructed in terms of pl free.weighted line integrals
over pl fields.
Perturbative constructions involving sl fields are very much in their infancy,
and if the following simple illustrations encourage other particle theoreticians to

31
engage in the exploration of this extremely rich area of research this section will
have accomplished its purpose. The only perturbative interactions which have
been considered up to date are couplings of massive s = 1 vector potentials to
lower spin (s = 1, 1/2) matter fields.
In order to pass from a nonrenormalizable pl interaction density to its less
singular sl counterpart one needs a linear relation between the s 1 pl fields and
their less singular sl counterparts. For massive vector potentials this relation
reads
Z
A (x, e) = A (x) + (x, e), (x, e) = de AP
P
(x + e) (7)
Z
A (x, e) = de F (x + e), F (x) = AP P
A

It involves an additional scalar sl field.

Scalar covariant sl fields for any integer spin have been first constructed in
[39]: their semi-integer fermionic counterparts are sl Dirac fields for any halfin-
teger spin. They violate the connection between physical spin and the covariant
transformation property of their pl counterparts. It turns out that only those
sl fields which appear in a linear relation between covariant pl fields and their
sl counterparts as in (7) play an important role in the new SLFT renormaliza-
tion theory. On a purely formal level their appearance in the form of a gauge
transformation is reminiscent of scalar negative metric Stueckelberg fields which
appear in the operator gauge transformation between the Feynman gauge and
its unitary counterpart; they convert the renormalizable but unphysical matter
fields into their formally physical but very singular12 counterparts [50] [48] [49].
The conceptual and mathematical situation in (7) is very different. The three
linearly related fields live on the same Wigner Fock space of s = 1 particles and
belong to the same sl localization class (i.e. they are relatively Einstein-causal
in the sense of string-localization). The sl setting avoids the introduction of
additional (unphysical) degrees of freedom and in this respect may be viewed
as the result of the application of Ockhams razor to gauge theory [36]. It is
not a gauge theory because the local operator gauge transformations (different
from the global U(1) transformations) cannot be defined without the presence
of additional (indefinite metric) degrees of freedom.
The computational tests and the conceptual coherence of the new SLFT
setting leave no doubt that after more than 70 years one finally arrived at a
new setting which reunites the s < 1 renormalizable interactions with those
of s 1 under the same conceptual roof of causally localized and positivity
obeying (quantum probability preserving) genuine quantum theories.
The sl scalar -fields will be referred to as an escort of the pl Proca poten-
tial. Only the correlation functions of the sl potential permits a massless limit
whose reconstructed Wightman field [32] is the vector potential of Wigners
h = 1 helicity representation13 . The purpose of this care about massless limits,
12 Theyare not polynomially bounded and hence they cannot be Wightman fields.
13 Sincemassless Wigner representations are unitarily inequivalent to massive ones, the
mooth behavior refers to the expectation values and not to the operators.

32
which in the present context appears pedantic, is to raise awareness about the
reconstruction problem of a physical Hilbert space which corresponds to the
Wigner Fock space provided by scattering theory in the presence of a mass gap.
One knows very little about a particle like description of this limit (the problem
of infraparticles and confinement).
The linear relation (7) between AP , A and the escort is really a linear
relation between their intertwiners. Computing the intertwiners u,s3 (p, e) of
A (x, e) and us3 (p, e) of (x, e) in terms of of the intertwiner u,s3 (p) of AP

1
Z X
AP (x) = eipx u,s3 (p)a (p, s3 ) + h.c.
(2)3/2 s 3

using their definition in (7), one verifies the linear relation14 ; for general spin
the corresponding formula contains s escorts [36] [5]. A more geometric inter-
pretation views the escort field in the context of the Poincare lemma applied to
the differential 2-form F .
It is interesting to note that the massless limit preserves the number of
degrees of freedom: 2 are accounted for by h = 1 and one is carried by the
massless limit of the pl scalar field P (x) = limm0 m(x, e).This prevails for
spin s tensor fields for which the linear relation (7) contains s tensorial escort
fields.
Starting from the 2-point function of the unique positivity-obeying long-
range massless vector potential and switching on the mass one cannot return
to the short range 2-point function of the Proca potential without the escort
plying its (7). The difference from the Higgss mechanism is that escort
fields do not introduce new degrees of freedom; so whenever the presence of an
additional Higgs field is necessary it must be for other reasons.
The important property of the sl vector potential A (x, e) can be seen in
its 2-point function15
3
1 d p
Z
i(xx )p
hA (x, e) A (x , e )i = e M (p; e, e ) (8)
(2)3/2 2p0
p p p e p e
M = g +
pe pe+ pe pe+

where the signs refer to the distributional boundary values lim0 1/(pe i)
from the Fourier transforms of the Heaviside function of the semi-infinite linear
string.
The gauge theoretic Feynman 2-point function (without the additional ra-
tional p-dependent contributions) looks much simpler but contains longitudinal
positivity-violating unphysical degrees of freedom which in the presence of in-
teractions infect the matter degrees of freedom and account for the physical
14 The Fourier transforms of the Heaviside funtions (x) account for denoinators 1/pe.
15 All pl fields have polynomial two-point functions in p.

33
limitations of local gauge theory which, as a result of the positivity-localization
interrelation also affects causal localization16 .
In contrast to the Proca 2-point functions
P p p
M (p) = g + (9)
m2
which has a quadratic mass divergence the sl 2-point function (8) admits a
well-defined massless limit in that it passes smoothly to its sl helicity h = 1
counterpart (the mass only enters the p0 ).
The price for having used Ockhams razor is the appearance of nondiagonal
2-point functions of free fields, e.g. mixed 2-point function hAi . This is not
surprising since both A and are linear combinations of the same Wigner cre-
ation/annihilation operators. This leads to a slightly more involved perturbation
theory. But it is very worthwhile to pay this increase computational expendi-
ture since the new formalism does not only maintain the quantum probability
but secures also the physical localization.
Already for the free sl potentials this implies a slightly more refined inter-
pretation of the Aharonov-Bohm effect. It can be shown that Wilson loops keep
a topological memory of the string dependence [5] which leads to a violation of
the Haag duality. The latter is slightly stronger than Einstein causality (= in-
stead of ) which is intuitively often identified with the latter17 . The violation
of Haag duality is a feature of all massless physical s 1 fields which only exist
in the form of positivity preserving sl fields.
The conversion of dsd = s + 1 potentials into their dsd = 1 sl counterparts is
a general phenomenon [13] [5] and has an extension to Fermions. For instance
for the massive s = 3/2 Rarita-Schwinger potential the corresponding escort
shares the dsd = 1 with the above scalar escort but reveals its Fermi statistics
through the presence of gamma matrices in the propagator. The claim that
there exist gamma independent dsd = 1, s = 1/2 Elko fields is based on
a misunderstanding of the relation of free fields with Wigners representation
theory [51].
New physical properties arising from the reorganization of already existing
degrees of freedom into new fields represent a quite common phenomenon in
quantum mechanical many body problems. For example the Cooper pairs in
superconductivity are the result of such regrouping of electrons into bosonic
bound pairs at low temperature18 Among other things they account for the
change of long-range classical Maxwell vector potentials into their short-range
counterparts inside a superconductor (F. Londons screening).
In fact this analogy between escorts and Cooper pairs goes much further.
It clears the head from the tale about fattening of photons by swallow-
ing massless Goldstones and facilitates the correct understanding why massive
16 The interacting positivity-obeying electric charge carrying covariant field of is necessary

sl, even so the particle counter events can be well localized.

17 This is the origin of the quirky feeling about causality which makes the A-B effect a

subject of public interest.

18 Haags presentation of Cooper pairs is particularly close to the spirit of LQP [10].

34
neutral Hermitian H fields are really needed to save the second order renor-
malizability of self-interacting massive vector mesons through short distance
compensations. This is similar to what what was expected to be a fringe benefit
of supersymmetry but in the present case it is the raison detre for the H field
(more details below).
By analogy the long range vector potentials of photons cannot be converted
into their massive Proca counterparts by just switching on a mass; one also
needs the intervention of the escort field. QFT. Such escort fields do not
appear in renormalizable lower spin s < 1 interactions or in the s = 1 indefinite
metric local gauge theory19 , but their presence turns out to be an indispensable
aspect of renormalizable positivity preserving LQP interactions involving s 1
particles. The conversion of dsd = s + 1 spin s pl fields into their better behaved
dsd = 1 sl counterparts requires the introduction of s (for half-integer spin
s 1/2) escort fields .
Before addressing the dynamic use of sl vector potentials it is interesting to
note a relation of the massless sl potential with the dsd = 1 radiation potential.
The angular integration of the sl potential over the directions emanating from
the point x in a equal time hypersurface leads to the radiation potential. Both
the non-covariant radiation potential and the covariant sl potential live in the
same Hilbert space and remain infrared convergent in the massless limit.
It has been known for a long time that the radiation (Coulomb) potential
(in contrast to vector potentials in gauge theory) lives in the Hilbert space of
the h = 1 Wigner representation. This is the reason why investigations of
long distance (infrared) properties of charged particles have been preferably
discussed in the Coulomb gauge [52]. But the radiation potential is not a
gauge in the sense in which we have used the word gauge theory and gauge
transformation in the present work since any gauge theory needs additional
(generally indefinite metric) degrees of freedom to implement operator gauge
transformation for passing from one gauge to another.
The main reason for using the gauge theoretic setting in QED is that the
lack of covariance makes the Coulomb potential unsuitable for renormalization.
The new renormalization theory based on covariant sl fields permits to compute
the renormalized Coulomb equivalent of any operator by angular averaging over
all string directions. This shows in particular the equality of e-independent
operators (local observables, S-matrix) in both descriptions.
The guiding idea of the new sl renormalization theory is the conversion of the
power-counting bound (pcb) violating first order pl interaction density LP with
dint int
sd > 4 into a dsd 4 renormalizable sl density L. In this way one maintains
the heuristic physical content while improving the short distance properties.
This passing from pl to sl does not affect the Hilbert space positivity (unlike
gauge theory which achieves this by a brute force compensations of part of the
positive with negative metric contributions in a Krein space setting.
Let us take a brief look at how this is done. Using the linear relation (7) one
19 This is the reason why they played no role in the more than 80 years history of QFT.

35
finds

LP (x) = AP
j = L V (10)

L := A (x, e)j , V := (x, e)j
Z Z
S (1) = LP = L (11)

In this way the dint P P int

sd = 5 (3 from the j , 2 from A ) of L is lowered to dsd = 4
int
of L at the expense of dsd (V ) = 5 of the divergence term. But in models
with a mass gap this term does not contribute to the first order S-matrix in the
adiabatic limit (11)
The problem is whether the use of the renormalizable L in the adiabatic
limit extends to the higher order S-matrix and whether this idea of adiabatic
equivalence also works for the construction of correlation functions of sl fields.
Formally speaking this corresponds to the independence from internal e s in
suitable sums of Feynman graphs after having integrated over inner x s. For
the S-matrix (i.e. in the absence of external lines) this is reminiscent of gauge
theory except that covariant gauge fixing parameters are spacetime independent
unphysical object which bear no relation with the independently fluctuating e-
directions in inner propagators.
The LQP localization theorem [59] does not specify for which models one
needs sl instead of pl fields; here one has to appeal to the pcb of perturbative
renormalizability criterion which reveals that there are no positivity obeying
interaction densities within the pcb limitation dpcb
sd = 4 which involve dsd = s+1
pl s 1 fields. PCB violating pl interactions exist only for s < 1.
It is convenient to formulate the adiabatic equivalence in terms of the differ-
ential form calculus of the 2 + 1 de Sitter space of spacelike directions e2 = 1

de (L V ) = 0 (12)

here shortly referred to as the L, V pair condition; it states that the zero
form L V is in fact exact. Its second (and correspondingly also higher) order
extension [53] [36]

(de + de )(T LL T V L T LV + T V V ) = 0 (13)

P P
TL L (de + de )(T LL T V L T LV + T V V )

would be a trivial consequence of (13) if the time-ordered products would not be

distribution valued. It is in fact a normalization condition on the time-ordering
which extends the e-independence to a singular set of intersecting strings, in
particular (which include coinciding end points x which carry the strongest
singularities).
The higher order L, V pair condition can be seen as a round about way to de-
fine time ordered products of T -products of pl interaction densities whose direct
calculation in the pl renormalization setting would lead to a with the number of
LP growing number of undetermined parameters (second line). A higher order

36
implementation of this formalism requires an extension of the Epstein-Glaser
renormalization theory [56] to sl crossings.
The pair condition (12) is a requirement on interaction densities. Whereas
there is no problem to satisfy the pcb condition for tri- and quadri-linear in-
teraction densities L involving s 1 sl fields, to construct an L, V pair for a
specified field content (including the escorts) imposes restrictions on L and is
not always possible. But without it the higher order perturbation would cause
a total delocalization and such a first order L would not define a perturbative
model of LQP.
There exists a simpler version of the pair condition which replaces the V by
Q = de V

de L = Q , Q = de V = j u, u := de (14)
(de + de )T LL = T Q L + T LQ (15)

Assuming asymptotic completeness (no problem in perturbation theory) the

Hilbert space in the presence of massive vector mesons is the Wigner-Fock tensor
product space of massive vector mesons with that of the s < 1 matter particles.
The loss of the Wigner-Fock structure in the limit of massless vector potentials
presages itself in the form of infrared divergences of the perturbative scattering
amplitudes.
The application of the pair formalism up to second order leads to interesting
new phenomena as well as new views of old phenomena. In the following we
will report on the results for four different models

Massive spinor QED

The simplest model is massive spinor QED with j = ; In that case

the use of the standard kinematic time-ordered propagator20 yields the tree
contribution to the 2-particle scattering contribution to which we restrict our
presentation. It is important to note that the e, e dependence is only lost in
the on-shell S-matrix Using e = e in off-shell relations leads to infinite fluctu-
ations. Since the e s can be pictures as points on d=1+2 de Sitter space these
fluctuations are similar to those of coincident points in x-space. Although these
e-fluctuations have no counterpart in the spacetime independent gauge fixing
parameters of gauge theory the SLF formalism shares with gauge theory the e-
respectively gauge-independence of scattering amplitudes [5]. Off shell correla-
tion functions are independent of inner e s but depend on the (fluctuating) e s of
the fields in the vacuum expectation values. Second order off-shell calculations
have been started in [47].
One expects that their leading short distance behavior is shared with that of
gauge theory but anticipates significant differences in the long distance regime
where the incorrect localization of gauge theory has its incorrect localization
properties of gauge dependent fields have their strongest ramifications. In the
20 obtained d3 p d4 p
by the replacement pp
p2 m2
.

37
massless QED limit these expected changes even affect the particle structure
(infraparticles) of the Hilbert space in a way which has not been fully under-
stood. From a physical viewpoint the terminology local gauge symmetry is
misleading since behind this more local looking symmetry is the result of the
presence of unphysical degrees of freedom. The physical localization properties
are only correctly described in a positivity-respecting setting.

Massive scalar QED

For scalar massive QED with j = the appearance of a derivative
leads to the well known second order quadratic in A contribution

(c x )A (x, e) (x)A (x , e )(x ) + h.c (16)

This comes about because the undetermined parameter c of the Epstein-Glaser

renormalization formalism which modifies the kinematic time-ordering by adding
a delta function term

hT i = hT0 i + cg (x x ) (17)

is fixed by the requirement (13) to c = 1. This insures that the in A quadratic

contribution does not introduce a new counterterm parameter. In the classical
gauge theory this results from the substitution D = igA required
by the differential geometry of fibre bundles whereas in SLFT it is follows from
the causal localization property which leads to he independence of particles and
the S-matrix on string directions. We will refer to renormalization terms whose
parameters are fixed by the locality principle as induced interaction terms. As
in the previous case the fluctuations in the e s become e-independent on-shell
after adding all contributions to a particular second order scattering process.

The abelian Higgs model

In this case the L in the L, V pair conditions depends explicitly on the scalar
sl escort

LP = mAP AP H = L V (18)
m2H
L = m(A AH + A H 2 H)
2
1
V = m(A H + 2 H), Q = m(A uH + u H)
2
Here the vector meson mass m factor in front accounts for the correct mass
dimension dsd =4 of the interaction density21 . The requirement of second and
third order preservation of the pair property in the tree approximation comes
with a surprise: in addition to the expected delta contributions (x x )A
A2 and (x x )A AH 2 which can be encoded into a T0 T change of
21 Note that according to its definition the escort has mass dimension dm = 0.

38
time-ordering, there is a second order induced potential of the form of a linear
combination of H 3 , H 4 , 2 H 2 , 4 [53] There is a formal similarity with the gauge
theoretic calculations in [48].
This similarity should however not be permitted to obscure the significant
conceptual difference: whereas the induction in the SLFT setting is a direct
consequence of the perturbative implementation of the causal localization prin-
ciples, the BRST gauge theory results from the imposition of a formal symme-
try which rescues a perturbative subtheory (local observables, the pserturbative
S-matrix) from a positivity- and causal localization- violating point-like descrip-
tion22 . That SLFT contains only physical degrees of freedom would certainly
have pleased Rudolf Haag (section 5).
SSB applies to internal symmetries of interacting s < 1 matter for which
the same field content can perfectly exist in the form of SSB or less symmetry
(independent coupling parameters and masses and a reduced number of con-
served currents). For s 1 the causal localization principle leads to the new
phenomenon of a fibre bundle like structure.
Some of these differences were suspected by Rudolf Haag and his LQP school
when they realized that their classification of superselection sectors led to inner
symmetries and the exclusion of parastatistics but confronts serious conceptual
problems in attempts to construct field algebra which extends the local observ-
ables of gauge theory [54] [55]. The perturbative SLFT adds the new viewpoint
of a causal localization caused instrinsic quantum fibre-bundle structure in s 1
interacting theories. The suggestion of perturbative SLFT is that interpolating
fields for particles in interacting models involving s 1 fields exist only in the
form of sl Wightman fields.
The arguments in this subsection are a good preparation for understanding
the true physical reasons why Higgs fields are needed in the presence of self-
interacting massive vector mesons which will be addressed below.

Self-interacting massive vector mesons

An even bigger surprise arises in the presence of selfinteracting massive vector

mesons. In this case the pair condition and its second order iteration leads to
two quite remarkable observations. On the one hand the second order restriction
requires the first order fabc coupling strengths in the general ansatz for a self-
interacting vector potential
X
L= fabc Fa A,b Av,c + ..A, contr. (19)
abc

to fulfill the Lie algebra relations of a reductive.Lie group; to find this Lie al-
gebra structure one has to implement the L, Q pair condition23 up to second
order. In the BRST gauge setting this has been known for a long time [48]
22 As pointed out before the e-independence of the S-matrix is not a postulate but rather

the consequence of the LSZ scattering formalism in the presence of a mass gap [59]
23 The Q formalism is somewhat easier to handle and maintains a formal similarity with

the CGI gauge formulation [48].

39
but this is hardly surprising since this BRST formalism resulted from achieving
formal compatibility between Lagrangian quantization of classical gauge the-
ory (where this relation follows from the fibre bundle requirements) with the
algebraic structure of QFT.
But in the new sl setting the Lie algebra structure follows from the s
1 causal localization principles in the form of the L, Q pair requirement; the
calculation is in this regard formally similar to that based on the BRST gauge
formalism [48].
Another important observation is that the second order leads to dsd = 5 delta
contribution which, if left uncompensated, would destroy the renormalizability
and hence the perturbative existence of the model. It is saved as a renor-
malizable QFT by extending the field content and adding a nonabelian A AH
interaction of the massive vector mesonsn with a H-field whose second order con-
tribution contains (after adjusting its coupling strength) compensating dsd = 5
second order terms generates such a second order compensating24 . This is rem-
iniscent of short distance compensations between different spin components in
supermultiplets except that in the present case it is not an epiphenomenon of
an extended symmerry but rather the raison detre for the H-particle.

Added comments

The new string local quantum field theory (SLFT) shows many formal sim-
ilarities with the prior causal gauge invariance (CGI) reformulation of BRST
in the Epstein-Glaser operator setting [57] . It has the advantage of clearly
distinguishing between properties which holds only on-shell such as the BRST
invariance sS = 0 of the S-matrix and off-shell properties as SSB
SLFT does not disqualify gauge theory, it rather shows its physical limita-
tions. Before commenting on this it is interesting to recall what Haag said about
the BRST formulation. In his reminiscences [10] one finds the following remarks
this elegant scheme is generally accepted today as the adequate formulation of
the local gauge principle in perturbation theory. But it bears no resemblance
to the conceptually simple picture in classical theory with its continuous group
acting on the fibres of a bundle.
He goes on to expresses his problem with the ghost degrees of freedom which
at the end of the day have to be removed with the help of BRST gauge invari-
ance. This is necessary in order to recover the most important property of any
quantum theory namely the positivity which secures the quantum theoretical
probability interpretation. Indeed the problem with pl s 1 interactions reveal
a deep clash between pl localization and positivity. Either one permits nega-
tive contributions in sums over intermediate states as in the Krein space setting
of local gauge theory, or one saves positivity and uses the more natural SLF
formulation in terms of L, V pairs25 .
24 Contrary to its abelian counterpart for which a corresponding second order d
sd = 5 term
vanishes on the e = e diagonal, the nonabelian contribution provides precisely the necessary
compensating contribution.
25 The physically preferred choice is supported by the B-F theorem [59] which states that

40
 In philosophical terms one may say that SLFT is the result of applying
Ockhams razor to the ghostly BRST Krein space setting. This leads to the
concept of sl escort fields which depend on the same degrees of freedom as those
already contained in the fields which they are escorting..Such free fields
Phave

necessarily

P mixed two-point functions i.e. all nondiagonal contributions A ,
hAi , A u , ..of the linear sl (Borchers) equivalence class are nonvanishing.
This makes perturbative sl calculations somewhat more involved than those in
the pl Krein space renormalization theory.
Interactions which involve s = 1 fields are subject to additional requirements;
In the CGI gauge setting this is the BRST invariance of the S-matrix sS = 0,
whereas in the SLFT formulation the causal localization requirement demands
the independence of the S-matrix from the fluctuating string directions which
in nth order reads n
X
d(n)
e S
(n)
= 0, d(n)e := dei
i=1

In the SLFT setting this requirement on the S-matrix follows directly from
the causal localization principle whereas in the CGI setting it is part of the
BRST formalism (whose spacetime interpretation is restricted to gauge invariant
observables).
For the implementation one uses the L, Q pair property and its higher
order extension. The main difference between couplings of vector potentials to
complex and Hermitian matter is that in the latter case one obtains a richer
set of second order induced terms including selfinteractions of the H and the
escort fields. The fact that these induced contributions have the appearance
of a field-shifted Mexican hat potential does not mean that the result bears a
relation to the physics of SSB.
A renormalized interaction density is uniquely fixed in terms of its field con-
tent (including their masses and internal symmetries). The interpretation of a
renormalized model of QFT cannot be described by the calculating theoretician,
it is uniquely determined by intrisic properties; Qsc = 0, Qsym < , QSSB =
represent the 3 different mutually exclusive realizations of the causal local-
ization principles, they correspond to screening, inner symmetry and SSB.
As mentioned in the previous subsection selfinteracting vector mesons lead
to two new phenomena. Such models are subject to the SLFT renormalization
theory based on the L, V pair condition. This requires the fabc selfcouplings
to obey a fibre-bundle like structure (19) which, in contrast to gauge theory, is
not imposed but by quantum adjustments to classical fibre bundles but rather
a consequence of the causal localization principles. For massive self-interacting
vector-mesons there is the additional phenomenon of second order violation
renormalizability violation which requires the compensatory presence of H fields
in order to save the Standard Model.
It is interesting to note that apart from the particle-antiparticle symmetry
Nature has no use of the concept of inner symmetries and their SSB (apart from
in the presence of a mass gap one needs no weaker localized interpolating fields than sl fields
(i.e. no need for branes in LQP)

41
phenomenological applications by theorists). As the success of the Standard
Model shows Nature prefers the fibre-bundle like structure of s 1 selfinter-
actions and renormalizability-saving compensations (the raison detre for the
H).
The next and final section contains remarks about possible extensions to
higher spins s 2 and the challenge their perturbative verification would pose
to LQP.

8 New challenges to Local Quantum Physics

The new perturbative SLFT originated from modular localization theory within
Haags LQP; in particular the observation that Wigners infinite spin represen-
tations does not permit compact localization and requires the construction of sl
fields [39] played an important catalyzing role. This begs the question whether
the extension of perturbation theory could also lead to an enrichment of LQP.
One challenging question is if the sl nature of interpolating fields in the
presence of s 1 massive particles is a general (perturbation-independent)
structural property of LQP. The naturalness of sl localization26 suggest that
this is the case. Part of the problem of proving such a conjecture is that one
has no intrinsic nonperturbative spacetime local (off-shell) understanding of
interaction; the reference to the nontriviality of the global (on-shell) S-matrix
is too far removed from properties of causal localization. The reformulation
of axiomatic QFT in the sense of Wightman [32] in terms of sl Wightman
fields is expected to be straightforward apart from the sl replacement of the pl
extended tube analyticity (since the representation of sl fields in terms of line
integrals over pl fields is limited to free fields).
An important part of such a reformulation is a better understanding of
problems which are outside the physical range of gauge theory, as e.g. the
construction of electrical charge-carrying fields in terms of properties of local
observables [10]. The use of the Gauss law in the LQP setting shows that such
fields are necessarily string-local in a very strong sense [60]27 . In fact a physical
description of the Hilbert space of QED (which is not a Wigner-Fock space !)
and the operators acting in it is still outstanding.
The particle structure of the Hilbert space is synonymous with the existence
of the S-matrix i.e. with the large time behavior of the charge-carrying fields.
Observationally important momentum space prescriptions for photon-inclusive
cross sections are no during replacement for a spacetime understanding of in-
frared aspects. Since the gauge theoretic indefinite metric destroys the physical
localization, the correct spacetime properties require the use of the positivity
preserving sl localization; in fact the correct analogy of quantum mechanical
long range (Coulomb) interactions are rigid (i.e. consistent with the Gauss law)
26 Weaker localization on e.g. spacelike branes is not needed ([10] section IV,3).
27 As a result of photonic vacuum polarization clouds along the space-like string direction the
strings are rigid and, different from massive vector mesons, cause a spontaneous symmetry
breaking of the Lorentz symmetry.

42
string-local quantum fields.
The proposed physical spacetime explanation for the appearance of the loga-
rithmic divergencies is that the coupling to massless photons changes the mass-
shell delta functions of the charge-carrying massive particles into a milder cou-
pling strength dependent singularity which leads to vanishing spacetime scat-
tering amplitudes. The logarithmic divergencies are the result of an illegitimate
expansion of the softened mass shell singularity into a power-series in the
coupling strength accounts. In [61] one finds rather convincing arguments that
the introduction of an infrared cutoff parameter28 and taking the limit of its
vanishing after summing over the leading logarithms to all orders indeed leads
to a vanishing amplitudes of photonless collisions of charge-carrying particles..
The correct spacetime scattering theory is expected to be a description in
terms of a large time behavior of expectation values (probabilities). This is
outside the range of gauge theory and can only be achieved within a positivity
preserving sl setting.
Another ambitious project outside the range of gauge theory is a LQP un-
derstanding of confinement. Different from the on-shell infrared phenomenon
whose cause is a change of the mass-shell properties of charged particles (which
leads to the vanishing of the large time limits of fields but has no direct effect
on fields and their vacuum expectation values), confinement is a more radical
phenomenon in which correlation functions containing self-interacting massive
vector mesons (massive Yang-Mills fields) coupled to spinor or scalar quarks and
the fields disappear in the massless gluon limit and only leave their composite
hadron-, gluonium- and quark-antiquark string-bridged fields behind.
In analogy with the vanishing scattering amplitudes for photonless charged
particle collisions one expects that all correlation functions which contain in
addition to hadron and gluonium fields as well as string-bridged q-q fields also
gluon or quark fields vanish, so that only those which contain no gluon and quark
operators are nontrivial. The only known way to describe theories in which the
basic model-defining fields leave only their composite shadows behind in our
present perturbative setting is in the form of zero mass limits of the conceptually
much clearer situation of selfinteracting between massive vector mesons.
Do the perturbative correlation function show such a behavior? A system-
atic construction of massless correlation functions of nonabelian gauge theories
can be found in [62] and the for the present purpose relevant result is that there
are no infrared divergent correlation functions in covariant gauges apart from
the expected on-shell logarithmic divergencies which are already present in the
abelian case. This had to be expected in view of the fact that gauge dependent
fields, although possibly revealing the correct short distance behavior in the
sense of having the physically correct beta-function29 will be maximally incor-
rect for long distances where the string-localization plays an important role.
In the presence of SLFT perturbative self-interacting gluons one however
28 In the SLFT formulation one would preserve covariance by viewing QED as a massless

limit of a vector mesons.

29 The Callen-Symanzik equation and in particular the beta function may turn out to be be

independent of e.

43
expects such logarithmic divergences. SLFT corresponds to the noncovariant
axial gauge which has been abandoned since it generates an entangled mix of
incurable ultraviolet and infrared divergencies. But the role of e in SLFT is
very different from that of a gauge fixing parameter. In contrast to a global
gauge parameter the e in SLFT is as x a spacetime variable in which each field
fluctuates independently in such a way that on-shell objects as particles and
the S-matrix as well as pl local observables remain e-independent, but fields
and their composites depend on e and transform covariant as linear spacelike
strings S = x + R+ e, e2 = 1. A low order calculation of two-point correlations
correlation functions for self-interacting massless vector mesons which could
reveal whether SLFT contains a signal of confinement is more elaborate but
feasible.
The SLFT renormalization theory enlarges the number of renormalizable
positivity maintaining interactions. There are two requirements which a pre-
scribed field content containing s 1 fields must fulfill in order to define a
positivity maintaining renormalizable SLFT. There must exist a L, V pair with
dsd (L) 4 which fulfills the pair requirement de L V = 0 and it must be
possible to compensate induced higher order anomaly terms with dsd 5 by
extending the field content of L.
The first requirement is a lowest order consistency condition which pre-
vents the short-distance improving string-localization of fields to destroy the
large time field-particle relation and maintains on-shell objects as the S-matrix
e-independent. Its preservation in higher orders is a normalization condition
which leads to induced higher order contributions. In contrast to renormal-
ization counterterms which enlarge the number of coupling parameters, higher
order induced terms preserve them. They are similar to the second order induced
A A ||2 term in scalar QED except that in SLFT they do not originate from
quantization of classical fibre bundle structures but are an autonomous conse-
quence of the positivity maintaining causal localization principle of QFT. This
applies also to the Lie-algebra structure of self-interacting vector potentials. It
shows that QFT does not need quantization crutches but can perfectly stand
on its own feet.
The second requirement maintains renormalizability to all orders. It has
no counterpart in s < 1 pl interactions for which first order renormalizability
dsd (L) 4 guaranties renormalizability to all orders (no induction). It is a
new phenomenon for interactions involving s 1 fields (unless one wants to view
it as an analog of the alleged renormalizability-improving role of compensation
between different spin components within a supermultiplet).
Interactions of abelian vector mesons with spinor-, complex scalar- or real
(Higgs)- matter do not require the compensatory extension of the field content;
the implementation of the pair condition suffices in those models. The need
for a compensatory enlargement in order to preserve second order renormaliz-
ability leads to the Higgs field in the presence of massive self-interacting vector
potentials [5] [53]. Both requirements have their counterpart in the CGI oper-
ator setting of BRST gauge theory [48] where the causality implementing pair
requirement corresponds to the BRST invariance of the S-matrix.

44
 The SLFT renormalization theory is still in its infancy. For s > 1 there are
as yet no SLFT results apart from the qualitative observation that higher spin
fields will enhance the short distance dimension of Q which in turn may lead to
renormalizability violating induced higher order delta terms whose compensa-
tion requires the enlargement of the field content. The most plausible scenario
in analogy to the compensatory role of the Higgs field is that the highest spin s
requires the presence of all lower spin fields (e.g. for s = 2 the presence of s = 1
and s = 0)
Fields belonging to the zero mass infinite spin Wigner class fail on the
L, Q pair requirement; they exist only in the form of sl free fields [42]. We
will refer to such matter as non-reactive or inert in the sense of SLFT pertur-
bation theory. Hence the problem posed by the two requirements is the question:
up to what spin does matter remain reactive ?
Since a further lessening of the tightness of localization beyond sl as a lo-
calization on spacelike hypersurfaces (branes) brings no gain for renormaliz-
ability, it is not unreasonable to expect that a field content which permits no
renormalizable perturbative interaction in the SLFT formulation has also no
counterpart outside perturbation theory. This belief is based on the natural-
ness of string-localization i.e. the fact that particles in LQP always admit sl
interpolating fields with pl being a special case of sl [59].
This does not require the convergence of the perturbative series; the singular
nature of fields due to the omnipresence of vacuum polarization clouds limits
their use in mathematical existence proofs. But a field of spin s > 1 which allows
no renormalizable interactions with itself and lower spin fields is also believed
to be inert par excellence.
All positive energy matter can be shown to admit a conserved energy-
momentum tensor; the No-Go theorem in [37] for massless higher spin mat-
ter refers to pl fields, whereas conserved weaker localized sl E-M tensors whose
global charges are identical to those of their pl siblings exist and have a well-
defined massless limit. Hence also inert matter which only exists in the form
of free fields couples to gravity and leads to gravitational backreaction, which
makes it interesting as candidates for dark matter. Intrinsically sl infinite spin
matter is inert [42] but as a result of its fleeting nature resulting from its mass-
lessness it does not seem to be compatible with the halo like accumulation of
dark matter around galaxies.
This is the content of a structural theorem of LQP which states that in order
to describe particles one does not need weaker than string localized interpolating
fields. Hence one expects that interactions which fail on both previous properties
do not exist as the result of lack of reactivity of the highest spin component.
The string-localization of matter fields in interactions involving sl s 1
potentials in SLFT renormalized perturbation theory begs the question to what
extend its occurrence can be understood in the nonperturbative LQP setting.
The problem is that there is no nonperturbative localization-based intrinsic
definition of interaction; the existence of a nontrivial S-matrix is too remote
from the spacetime properties of interacting fields. A proof would amount to a
theorem stating that a particle spectrum with mass gaps which includes s 1

45
particles is either a free field theory or an model whose interacting fields are
sl Wightman fields. The adjustment of Wightmans axiomatic framework to sl
fields would then be the lesser problem.
Perturbative SLFT also directs attention to a new problem of formal sym-
metries which are not inner symmetries in the sense of the DHR superselection
theory. As mentioned before such a problem is posed by the perturbative Lie
algebra structure of self-interacting vector mesons. Such a situation can not
be subsumed under inner symmetry since the latter always permit interactions
with the same field content but less or no symmetry. There is as yet no natural
conceptual place in LQP.
In the BRST gauge formulation this is less surprising since that formalism
is the result of a repair job which is necessary to control the indefinite metric
aspects of a formulation obtained from adjusting a classical fibre bundle setting
to the exigencies of a quantum theory which is only possible at the price of
indefinite metric and ghosts. In a somewhat metaphoric sense result from the
quantization (of a classical theory which has no use for positivity). Why
does Nature not present particle multiplets associated to internal symmetries
(or Goldstone particles of an exact SSB)? Why does she prefers the Lie algebra
structure of self-interacting vector mesons (the Standard Model) ?
LQP is still far from its ultimate goal of establishing the mathematical exis-
tence of nontrivial models and finding mathematically controlled approximation
procedures. However there are good reasons to expect that the pursuit of this
goal will lead to important more new insights. Rudolf Haags general LQP view
of QFT as causally localized quantum matter [2] remains a valuable compass
which helps to avoid a cul-de-sac as that mentioned in section 5.

Acknowledgement 1 The last two sections are part of an ongoing joint project
with Jens Mund. Its ultimate aim is to replace local gauge theory by a formula-
tion which is compatible with Haags LQP. For a critical reading I am indebted
to Joe Varilly.

References
[1] R. Haag and D. Kastler, An algebraic approach to quantum field theory, J.
Math. Phys 5, 848 (1964)
[2] R. Haag, Some people and some problems met in half a century of commit-
ment to mathematical physics, Eur. Phys. J. H 35, 263307 (2010)
[3] R. Haag and B. Schroer, Postulates of quantum field theory, J. Math. Phys.
3, 248 (1962)
[4] H. Araki, R. Haag and B. Schroer, The Determination of a Local or Almost
Local Field from a Given Current, Nuovo Cimento 19, 90 (1961)
[5] B. Schroer, Beyond gauge theory; positivity and causal localization in the
presence of vector mesons, Eu. Phys. J. C 76 (2016)

46
 [6] P. Leyland, J. Roberts and D. Testard. Duality for Quantum Free Fields
(Centre de Physique Theorique, CNRS Marseille, 1978)
[7] K-H. Rehren, Local Quantum observables in the Anti-deSitter-Conformal
QFT Correspondence, Phys. Lett. B493, 883 (2000)
[8] M. Duetsch and K-H. Rehren, A comment on the dual field in the AdS-CFT
correspondence, Lett. Math. Phys. 62, 171 (2002)
[9] B. Schroer, Infrateilchen in der Quantenfeldtheorie, Fortschr. Phys. 173,
1527 (1963)
[10] R. Haag, Local Quantum Physics, Fields, Particles, Algebras, Springer-
Verlag 1992
[11] R. Haag and J.A. Swieca, When does a quantum field theory describe par-
ticles?, Commun. Math. Phys. 1, 308 (1965)
[12] D. Kastler, D.W. Robinson and J.A. Swieca, Conserved currents and as-
sociate symmetries; Goldstones theorem, Commun. Math. Phys. 2, 108
(1966).
[13] H. Ezawa and J.A. Swieca, Spontaneous breakdown of symmetries and zero-
mass states, Comm. Mat. Phys. 5, 330 (1967).
[14] J. A. Swieca, Goldstones Theorem and Related Topics, Cargese Lecture
Notes in Physics, Vol. 4, 215, (1970)
[15] B. Schroer, R. Seiler and J.A. Swieca, Problems of Stability for Quantum
Fields in External Time-Dependent Potentials, Phys. Rev. D, 2 2927 (1970)
[16] B. Schroer and J.A. Swieca, Indefinite Metric and Stationary External In-
teractions of Quantized Fields, Phys. Rev. D 2, 2938 (1970)
[17] B. Schroer, Quantization of m0 Field Equations, Phys.Rev. D3, 1764
(1971).
[18] J. H. Lowenstein and J.A. Swieca, Quantum Electrodynamics in Two Di-
mensions, Annals of Physics 68, 172 (1971)
[19] Jorge Andre Swieca, Obras Collogidas, Projeto Galileo Galilei, CNPq,
Brasilia 1981
[20] J. Holland and S. Hollands, Recursive construction of operator product ex-
pansion coefficients, Commun.Math. Phys. 336, 1555 (2015)
[21] Olga, by Fernando Morais, published by Companhia das Letras 1985
[22] R. Haag, N. Hugenholz and M. Winnink, On the equilibrium states in quan-
tum statistical mechanics, Commun. Mat. Phys. 5, 215 (1967)

47
[23] M. Takesaki, Tomitas Theory of Modular Hilbert Algebras and its Appli-
cations, Springer, Berlin 1970
[24] H.-J. Borchers, On revolutionizing quantum field theory with Tomitas mod-
ular theory, J. Math. Phys. 3604 (2000)
[25] J.J. Bisognano and E. H. Wichmann, On the duality-condition for quantum
fields, J. Math. Phys. 17, 303 (1976)
[26] G. Sewell, Relativity of temperature and the Hawking effect, Phys. Rev.
Lett. 79A, 23 (1980)
[27] K. Fredenhagen and R. Haag, On the derivation of Hawking radiation as-
sociated with the formation of a black hole, Commun. Math. Phys. 127,
273 (1990)
[28] K. Fredenhagen, On the modular structure of local algebras of observables,
Commun. Math. Phys. 97, 79 (1985)
[29] K. Fredenhagen and K. Rejzner, QFT on curved spacetime; axiomatic
framework and examples, J. Math. Phys 57, 031101 (2016)
[30] R. Brunetti, D. Guido and R. Longo, Modular localization and Wigner
particles, Rev. Math. Phys. 14, (2002) 759
[31] S. Weinberg, The Quantum Theory of Fields I, Cambridge University Press
1991
[32] R. S. Streater and A. S. Wightman, PCT, Spin and Statistics and all that,
New York: Benjamin 1964
[33] J. Mund, An Algebraic Jost-Schroer Theorem for Massive Theories, Com-
mun. Math. Phys. 315, 445 (2012)
[34] H-J Borchers, D. Buchholz and B. Schroer, Polarisation-Free Generators
and the S-Matrix, Commun. Math. Phys.219, 125 (2001)
[35] B. Schroer, Modular Wedge Localization and the bootstrap Formfactor Pro-
gram, Ann. Phys. 275, 190 (1999)
[36] J. Mund and Erichardson T. de Oliveira, String-localized free vector
and tensor potentials for massive particles with any spin; 1. Bosons,
arXiv:1609.01667
[37] S. Weinberg and E. Witten, Limits on massless particles, Phys. Lett. B 96
(1-2) (1980) 59
[38] J. Yngvason, Zero-mass infinite spin representations of the Poincare group
and quantum field theory, Commun. Math. Phys. 18 (1970), 195203.
[39] J. Mund, B. Schroer and J. Yngvason, String-localized Quantum Fields
and Modular Localization, CMP 268 (2006) 621, math-ph/0511042

48
[40] J. Mund, String-localized quantum fields and modular localization, Com-
mun. Math. Phys.268, (2006) 621
[41] R. Longo, V. Morinelli and K.-H. Rehren, Where Infinite Spin Particles
Are Localizable, Commun. Math. Phys. 345 (2016) 587, arXiv:1505.01759
[42] B. Schroer, Can the inert matter corresponding to Wigners infinite spin
representations be dark matter ?, submitted to EPJC
[43] J. Polchinski, String theory, Vol I, Cambridge University Press 1998
[44] C. Meusburger and K.-H. Rehren, Algebraic quantization of the closed
bosonic string, Commum.Math.Phys.237 (2003) 69
[45] R. Brower, Spectrum-generating algebra and no-ghost theorem in the dual
model, Phys. Rev. D6, 1655 ()972)
[46] S. Alazzawi and G. Lechner, Inverse Scattering and Locality in Integrable
Quantum Field Theories. arXiv:1608.02359
[47] F. Pedrosa and J. Mund, String-local Dirac fields in massive QED, in prepa-
ration
[48] G. Scharf, Quantum Gauge Theory, A True Ghost Story, John Wiley &
Sons, Inc. New York 2001
[49] H. Ruegg and M. Ruiz-Altaba, The Stueckelberg field, Int.J.Mod.Phys A19,
3265 (2003)
[50] J. Lowenstein and B. Schroer, Gauge Invariance and Ward Identities in a
Massive Vector-Meson Model, Phys Rev D6, 1553 (1972)
[51] D. V. Ahluwalia, A story of phases, duals, and adjoints for a local Lorentz
covariant theory of mass dimension one fermions, arXiv;160103188
[52] G. Morchio and F. Strocchi, The infrared problem in QED: A lesson
from a model with Coulomb interaction and realistic photon emission,
arXiv:1410.7289
[53] B. Schroer, Peculiarities of massive vector mesons and their zero mass
limits, E.P,J.C. 75 (2015)
[54] D. Buchholz and J.E. Roberts, New Light on Infrared Problems: Sectors,
Statistics, Symmetries and Spectrum, arXiv:1304.2794
[55] D. Buchholz, F. Ciolli, G.Ruzzi and E. Vasselli, The universal C* -algebra
of the electromagnetic field, arXiv:1506.06603
[56] H. Epstein and V. Glaser, The role of locality in perturbation theory, Ann.
Henri Poincare Phys. Theor. A 19, 311 (1973)

49
[57] M. Duetsch, J. M. Gracia-Bondia, F. Scheck, J. C. Varilly, Quantum gauge
models without classical Higgs mechanism, Eur. Phys. J. C89, (2012) 599,
arXiv:1001.0932
[58] R. Jost, The General Theory of Quantized Fields, American Mathematical
Society 1965
[59] D. Buchholz and K. Fredenhagen, Locality and the structure of particle
states, Commun. Math. Phys. 84, (1982) 1
[60] D. Buchholz, The physical state space of quantum electrodynamics, Com-
mun. Math. Phys. 85, (1982) 40
[61] D. Yenni, S. Frautschi and H. Suura, The infrared phenomena and high
energy processes, Ann. of Phys. 13, (1961) 370
[62] S. Hollands, Renormalized Quantum Yang-Mills Fields in Curved Space-
time, Rev. Math. Phys. 20, (2008) 1033, arXiv:0705.3340

50