A Quantum Mechanics Tourist

Quantum Mechanics Tourist
Bijou M. Smith
Geonworld
Wellington, New Zealand
January 26, 2022

i
Copyright © 2021 Bijou Murray Smith.
Permission is granted to copy, distribute and/or modify this

document under the terms of the GNU Free Documentation
License, Version 1.3 or any later version published by the Free
Software Foundation; with no Invariant Sections, no Front-
Cover Texts, and no Back-Cover Texts. A copy of the license
is included in the section entitled “GNU Free Documentation
License”.
ii
Contents
1 Introduction 1
1.1 The One and Only True Interpretation of Quantum Me-
chanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Particles and Symmetries 5

2.1 Introduction to particles and invariants . . . . . . . . . . . 6
2.1.1 Motivation — symmetries of the 𝑆-matrix . . . . . . 13
2.2 Group theory preliminaries . . . . . . . . . . . . . . . . . . 14
2.2.1 Quantum mechanics preliminaries . . . . . . . . . . 15
2.2.2 Symmetry representation theorem . . . . . . . . . . 21
2.2.3 First isomorphism theorem . . . . . . . . . . . . . . 23
2.2.4 Lie groups . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.5 Lie algebra . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.6 The groups SO(3) and SU(2) . . . . . . . . . . . . . 31
2.2.7 Topology of SO(3) . . . . . . . . . . . . . . . . . . . . 37
2.2.8 The isomorphism SU(2) Spin(3) . . . . . . . . . . 37
2.2.9 The isomorphism SL(2, ℂ) Spin+ (1, 3) . . . . . . . 37
2.2.10 Projective representation, part-II . . . . . . . . . . . 39
2.3 Poincaré Symmetries . . . . . . . . . . . . . . . . . . . . . . 43
iii
iv CONTENTS
2.4 Poincaré algebra . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.4.1 The vector representation . . . . . . . . . . . . . . . 48
2.4.2 Commutators . . . . . . . . . . . . . . . . . . . . . . 49
2.5 Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.6 What was accomplished? . . . . . . . . . . . . . . . . . . . . 55
2.6.1 The Standard Model . . . . . . . . . . . . . . . . . . 56
2.7 CPT Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.7.1 Why I like CP violation . . . . . . . . . . . . . . . . . 60
3 Scattering and Forces 61

3.1 Maximum parsimony . . . . . . . . . . . . . . . . . . . . . . 62
3.1.1 Who needs gravitons? . . . . . . . . . . . . . . . . . . 63
3.1.2 Two views of gravitons . . . . . . . . . . . . . . . . . 65
3.2 Geometric algebra preliminaries . . . . . . . . . . . . . . . 73
3.3 Spinor Helicity variables . . . . . . . . . . . . . . . . . . . . 81
4 Quantum Kinematics 83
4.1 The wave equations . . . . . . . . . . . . . . . . . . . . . . . 83
4.2 Probability and the Phase of the Wavefunction . . . . . . . 85
4.2.1 Complex Numbers are a Distraction . . . . . . . . . 86
4.2.2 Simple QM Preliminaries . . . . . . . . . . . . . . . 88
4.3 The Classical Action and Quantum Phase . . . . . . . . . . 93
4.4 Path Integrals and Phase . . . . . . . . . . . . . . . . . . . 94
4.4.1 Are Path Integrals Real? . . . . . . . . . . . . . . . . 96
4.5 Spin–Orbit Coupling . . . . . . . . . . . . . . . . . . . . . . 100
4.6 Probability Conservation . . . . . . . . . . . . . . . . . . . . 100
CONTENTS v
5 Kinematics and Amplituhedra 105

5.1 The S matrix origins . . . . . . . . . . . . . . . . . . . . . . 106
5.1.1 Inside the black box . . . . . . . . . . . . . . . . . . . 108
5.1.2 Fields or particles? . . . . . . . . . . . . . . . . . . . 109
5.1.3 What about the gauge symmetries? . . . . . . . . . . 110
5.2 Kinematics of scattering amplitudes . . . . . . . . . . . . . 113
5.3 Yang-Mills or GR and nothing else . . . . . . . . . . . . . . 113
6 Speculative Ideas 115

6.1 Miscellaneous Notes . . . . . . . . . . . . . . . . . . . . . . 116
6.2 Speculations on Vacuum . . . . . . . . . . . . . . . . . . . . 117
6.3 Speculations on Chromodynamics . . . . . . . . . . . . . . . 118
6.3.1 Continuous color rotation? . . . . . . . . . . . . . . . 118
Bibliography 123
Index 125
A GNU Free Documentation License 127

vi CONTENTS
Preface
I am a migrant worker in the landscape of physics academia. I like writ-

ing about topics in theoretical physics, and so that’s what I’ve done. No
one paid me, and I am a nomad in the world of academic physics. No one
should take my writing too seriously, it is mostly for amusement and fun
— the fun of imagining what the world might be like if we understood it
correctly.
This book began as a set of rambling teaching notes for unorthodox
students interested in quantum theory and gravity. Then I set aside one
chapter for “special topics” and it ballooned into a whole other book (see
The Quantum Nomad). This present book serves as auxiliary material
for general topics of interest.
The chapters are somewhat self-contained, like mini-essays, but
highly interwoven, which I hope makes some sense. It should be like
reading a bunch of essays in a journal of philosophy of physics. Pedagogy
is not tight, but I write in mostly plain English, and any jargon terms
can be searched almost entirely using Wikipedia as a resource, which for
physics is highly reliable. When researching a topic using Wikipedia, it
should suffice for understanding to take in just the plain English. when
physics is not explained in plain English (or some suitable international
lingua franca) it has no useful content, in my view, all the mathematical
derivations are merely demonstrations of what is expressed in plain
words.
In order to have fun writing these notes I have had to relax traditional
pedagogical standards. So,
• chapters do not build like a story from previous chapters, so there

is no systematic pedagogical order. I would have loved to tell a story
of theory building, but the ideas are too inchoate to fit a narrative
vii
viii CONTENTS
structure.
• each chapter could be considered is it’s own isolated vignette, al-
though the whole books is a One,
• there is no attempt to take readers on a course covering topics com-
pletely. My aim is not to teach or instruct, but to free-form lay out
some ideas.
These are notes very much intended as a fairly random collection of topics
the reader could dip into and out of on a whim, very much like a tourists
scrapbook of travel highlights.
If at any time in the future I can do some serious mathematical physics
and restate many of these ideas in a coherent form, from some sort of
starting postulates to a fully fledged theory, then I hope to be able to
publish it all. At present (c.2021) that seems unlikely, because late in life
I have found activism in political economy to be afar more important for
humanity, and so the problems of elimination of poverty and justice for
the working class have over-riding importance in my life, and physics is
just a hobby I indulge in to keep myself from going mad.
Note on typography. I am a big believer in the social goodness of shar-

ing information and not putting up unnecessary firewalls or gate-keeping
ideas. The cost is profligacy, which makes it hard for newbies to survey
a field. That is why peer review is still important. However, journal pay-
walls are evil, and peer review need not at all sit behind paywalls. If
donations from users cannot pay for the service then the state can pay,
and should pay if the public purpose is served — this is provided the state
runs a floating fiat currency (since then it is not revenue constrained, see
any good lecture son MMT). This is why I publish under the GNU Free
Document Licence.
I also think it is important for people to be able to play with and edit
free-libre documents, and that includes viewing them in the format you
please. If anyone wished the LATEX sources for my notes then they will be
welcome to them. You may then easily substitute the fonts you prefer to
read! The reason I chose the TEX Gyre Schola font is because it make
screen viewing much easier on my eyes. I do love Computer Modern Ro-
man, but my old eyes are not up to reading documents prepared in that
font beyond a few dozen pages. TEX Gyre Schola also retains the same
classic Victorian aesthetic typesetting style of CMR,
Chapter 1
Introduction
This book is a collection of loose notes, intended for “fun” reading. They
could be supplemental for an advanced undergraduate or beginning
graduate course on modern physics, with a focus on fundamental theory.
I am not going to pretend we will be preparing for practical applica-
tions. No! Most of this is all for fun, and we have no intention of ever
applying our studies in anger in the laboratory.
My philosophy, increasingly as I get older, is to do physics as an art.
This should not stop readers from joining laboratory teams to put their
learning into practical use, indeed such a motive would be highly laud-
able. But art too is a noble pursuit. The motive for art however has to
also be for the benefit of humanity in order to justify as noble in my opin-
ion, and that is the spirit of these chapters. I find this easy to justify: if I
can get good at this art of explaining physics, then it should be easier for
people to understand and apply to solve real problems, putting the art to
good practical use. There is no guarantee of this, but it is the motive.
One aspect of this fun are the diversions in algebraic geometry. For
a long time I wanted to understand in as simple terms as possible, but
no simpler, why 4-manifolds are so complicated. In Alexandru Scorpan’s
wonderfully titled The Wild World of 4-Manifolds the result is stated:
that dimension 𝑑 = 4 is the only dimension for which there exist infinitely
many distinct smooth structures on closed 𝑑-manifolds; and even more
wildly for open 4-manifolds, where for ℝ 𝑑 , 𝑑 ≠ 4 there is only one unique
smooth structure, whereas for ℝ4 there are uncountably many distinct
smooth structures! Closely related is the other unique result that only in
1
2 Introduction
𝑑 = 4 are the diffeomorphisms not the same as the topological homeomor-

phisms. Not only that, for physics it seems remarkable our space dimen-
sion is 3 𝑑, since 3-dimensions is the only dimension in which persistent
knots can exist. These are astounding facts. A naïve, but I think correct,
way to understand these results is that the low dimensional topology of
3 𝑑 and 4 𝑑 manifolds are what you would want to choose when (cough!)
designing a universe if you wanted your world to be as interesting as pos-
sible. Due to these algebraic topology results the dimension 4 is the only
smooth manifold dimensions where we, as observers within the space,
cannot be gods. Because the classification problem can be solved in all
other dimensions it is possible an observer in any different spacetime di-
mension could obtain sufficient information to solve hard problems to do
with their time evolution, these problems would be impossible for us to
solve because we live in (3 + 1)-dimensions. This is why I consider this
wild world to be on the itinerary of the adventurous quantum tourist’s
destinations.
1.1 The One and Only True Interpretation of

Quantum Mechanics
I wrote this section a long time after beginning this book, it is really just
to lay out my current opinion on the matter.
If you just stare at the principles of QM you should immediately see
the biggest difference with classical mechanics is that QM is a theory of
measurement, and CM is not. All CM takes place without measurement,
the idea of us making measurements like Archimedes, Galileo, Kepler,
Faraday or Michelson and Morley, is quite incidental to the physics. Not
so with QM which cannot be written down or worked on without reference
to measurement. This is the fundamental interpretation of QM.
All the other philosophy that goes on under headings like Copenhagen,
Many Worlds, Pilot Wave, Qbism, and so forth are more in the vein of
attempting some very bad models that seek to explain QM (except I guess
Copenhagen, which is a big nothing). There is no good model of QM yet.
But the interpretation of QM is unambiguous and clear: it is the physics
we get when we want to describe measurements in the context of time
evolution of a system. The time evolution part is critical (and definitely
The One and Only True Interpretation of Quantum Mechanics 3
not so critical in classical mechanics).

We need a time coordinate in both QM and CM, but in CM we can hy-
pothetically move our view forwards or back in time at will, so-to-speak
(just by doing some math on some big data), and we fundamentally can-
not do this in QM. QM has time reversal invariance (for most interactions,
with small violations only, where CPT symmetry is needed) but in QM
there is no sense in moving a system backwards in time: if you could do
that in QM then you would not need the probability axioms of QM (the
transition rule and the Born Rule).
So the wave function, or field variables, are instructions telling us
what we have to do to an object representing an observable (in the math-
ematical formalism, some operator) in order to get the chance of some
particular result. This always refers to some prepared system and mea-
surement apparatus. It is a practically complete theory of measurement
physics because it describes everything we can ever observe. But it is not
a model of physical reality. We have no such model from QM. We just have
a theory of measurement. QM is a model for a theory of measurement,
not a model for fundamental physics.
Standard Model particle phenomenology, on the other hand, or String
Theory, are models for physical reality. But they add nothing to our un-
derstanding of QM as a measurement theory.
I think I’ve made the main points I wanted to make. If it is not clear
then please write to me.
4 Introduction
Chapter 2
Particles and Symmetries
This is about the question — and this is a very ancient ques-

tion, at least I suspect 25 hundred years old — is nature dis-
crete [or continuous], ‘discrete’ meaning to say not that it is
polite. . . ” — Lenny Susskind, Fall Lecture, TTM, 2009.
As with most of my writing, you should find this chapter long but gen-
tle. I aim for anti-terseness but not rigor. Anti-terseness + rigor = hun-
dreds more pages, which I am not yet prepared to go through.
In this chapter I have attempted (probably ‘begun to attempt,’ since I
might not ever finish) to give what I think is the nicest introduction to
quantum theory of particles using the concepts of preservation of space-
time symmetries. This is what I regard to be a prelude to geon theory.
Geon theory seeks to construct geometrical and topological models for the
fundamental particles or ‘quanta’ that are observed in nature, at least
at low energy. This chapter is more primitive — meaning more abstract,
less exact — and only enquires into the abstract group structure of geons.
We will ignore the topological structure which must exist at around the
Planck scale. We really have to ignore this topological structure because
it is not yet well understood. However, this higher level of abstraction
has proven to be adequate for almost all quantum theory so far, other
than extreme conditions that are rarely observed in nature (black hole
interiors and big bangs).
Considering that we will not be seeking topological models, we only
want the bare algebra for the fundamental particles. The organizing
5
6 Particles and Symmetries
principle is that introduced by Wigner [2]. This is the idea that elemen-
tary quanta — whatever else they are — should at a minimum be de-
scribed by symmetries of spacetime. I will expand more philosophically
on Wigner’s ideas in the introduction, before then going into the theory
proper.
Background optional reading:
1. R. Hermann, Lie Groups for Physicists [3].
2. E. Wigner, On unitary representations of the inhomogeneous Lorentz

group [2].
3. V. Bargmann and E. Wigner, Group theoretical discussion of rela-

tivistic wave equations [4].
4. V. Bargmann, Irreducible unitary representations of the Lorentz

group [5].
5. N. Straumann, Unitary representations of the inhomogeneous

Lorentz group and their significance in quantum physics [6].
2.1 Introduction to particles and invariants
In all sciences we seek invariants in nature, that is the basic task — in-
variants are what we generally call laws of nature. And for fundamental
theoretical physics we seek the most primitive invariants. Physically we
can identify invariants as properties that do not change unless there is
significant interaction. In our mathematical language we have a highly
formal structure for such ideas known as group theory. To explore the
fundamental invariants of physics we thus attempt a group theoretic de-
scription as a first type of foundational theory.
Thus when Wigner asks for the elementary quanta to be described by
“irreducible unitary representations of the Poincaré group,” he was asking
for a minimal way to describe the elementary particles. It is not asking
for a complete description. Wigner’s scheme is interested only in the basic
symmetry properties of the particles and the consequent Lie algebra that
they obey. This is not asking for much! But it turns out to get us a long
way towards classifying most of the elementary particles.
Introduction to particles and invariants 7
A few questions for students (answers are given but I encourage you
to think before reading):
1. Why the Poincaré group? Answer: because the only known sym-
metries of flat spacetime are the Lorentz symmetries of special rel-
ativity.
2. Why irreducible representations? Answer: because we want
fundamental particles. In group theory a reducible representation
is in a plain sense more structure, and an irreducible representa-
tion is plainly minimal structure. We can build physics models of
composite or non-elementary particles (indeed nuclear physicists
have to) but for basic quantum theory we explore only the elemental
quanta, and call these the “elementary particles,” and define these
to be the observed quanta that can be described by the irreducible
representations of the spacetime symmetries.
3. Why unitary representations? Answer: for quantum mechanics
we require conservation laws to hold, the most basic of which is
conservation of probability. In both the Hilbert space formalism
and wave-function formalism this means unitary operators. If the
time evolution is non-unitary then we get probabilities that do not
add up to 1.0. That means we would have an incomplete theory.
If the probabilities added up to more than unity then we’d have a
non-sensible theory (and we would just impose some normalization
rule in that case, so it would end up being unitary anyway).
Follow-up questions. Does Q.1 mean quanta are not applicable in

non-flat spacetimes? To a degree yes, but the only known regimes in
physics where spacetime is not flat on the length scale of the quanta turn
out to be near black hole singularities and the very early universe. Note
that it is not “us” choosing Poincaré symmetry, it is Nature telling us
this is the fundamental symmetry group of the spacetime we happen to
live within. This also tells you spacetime is far more fundamental than
some people imagine — it is not easy to “emerge” spacetime out of some
more primitive structure, because the Poincaré group is already a simple
structure.
For all normal physics the description in terms of the symmetries of
flat Minowski spacetime will be perfectly adequate. The proper gener-
alization must include non-trivial topology, and that will be where geon
theory extends Wigner’s scheme. But note this also means on the scale
of the topological geons spacetime is again not flat, but that is regarded
as close to the Planck scale, which is again far too small for us to detect
presently and so mostly irrelevant for applications of fundamental the-
ory. The problem is this length scale is enormously important for theory
of fundamental theory!
Hence with regard to question 1, why Poincaré? The full answer is that
we need extra local symmetries for fully describing spacetime topology
on the scale of the fundamental strings or geons. But provided we are
doing physics asymptotically far enough from the Planck scale then the
symmetries of spacetime are the main ones to care about.
More on Q.2 — spacetime symmetry. We have not explained why

the Poincaré group is the symmetry of spacetime. We cannot! It is not a
theoretical result, it is an empirical result. We see light rays (or massless
particles) move at a fundamental invariant limit 𝑐 = 2.997 × 108 m/s, and
we always see massive particles always with invariant mass-energy 𝑝𝜇 𝑝𝜇 ,
if not then we do not class them as “elementary.” That’s not god-given,
that’s just our definition or classification of things. By “our” or “us” I
mean in terms of the sociology of physics. (There are some who beg to
disagree, and that’s fine, they do fringe physics, and they are welcome to
do their thing, but theirs is not what we call quantum theory.)
Also, the fact massless particles zip around at 𝑐 = 2.997 × 108 m/s is
no accident at all. The correct number in Natural units is 𝑐 = 1. This
is how we describe the light-cone or causal structure of spacetime. The
number 𝑐 = 2.997 × 108 is an artefact of our choice of a clock timer and
length ruler calibrated to the standards at the BIPM in Paris. This may
now sound like 𝑐 = 1 is a god-given number after-all, but that is still not
true. The proper value of 𝑐 = 1 is still an empirical determination, but
rather than seeing it determine a universal speed, it is better thought
of as telling us spacetime is homogeneous and isotropic, meaning that
in any direction we look the causal structure is the same: no physical
signals through flat space can get outside of the light-cone. It did not
have to be this way you see. There could be some exotic stuff that has a
rule for going outside a light cone then maybe back in. Or it could be the
universe has a preferred axis, say the one pointing along the longer arm
of the plane of the Milky Way galaxy — but relative to the background
stars, so after correcting for the bulk flow of the galaxy — and maybe only
along this direction the causal structure behaves like 𝑐 = 1/3 compared
to the other orthogonal directions where 𝑐 = 4/3. We do not observe any

such stuff, this is why 𝑐 = 1 is empirical.
It might have also turned out there were only two space dimensions
and two time dimension. Maybe in one time sub-dimensional set we’d get
a causal structure with 𝑐 = 2 relative to the other? I do not know what
could happen here. There might be some philosophy papers discussing
what that would imply, and it would probably be bizarre (I do not read
such essays).
The elementary particles we know about all appear to be invariants
absent interactions with external fields or other collisions, so they are
empirically what we take to be the “elementary quanta.” Of course this
is not 100% empirical. Every empirical observation here gets a mathe-
matical interpretation, the usual chosen one is a Minkowski spacetime,
scalar masses, and some choice of length, time, and mass measurement
units. If you want some pristine pure god-given mathematical theory
then I think you are not appreciating the beauty of physics. We take
what is observed, then try to find a suitable mathematical language to
describe it. There is no disconnect between theory and description then,
only either incorrect use of the language or an inadequate vocabulary in
the language. Mathematics is flexible enough to describe anything quan-
tifiable, and so we know the language is adequate for pretty much all of
physics. (This does not include subjective phenomena known to us, like
thought, qualia or consciousness, which are all not physics1 ).
On Q.2 — irreducibility. I think it is important to note that Nature

does not require irreducible representations of the spacetime symme-
tries. A theorist might imagine the classification of elementary quanta
in group theoretic language should conform to irreducible representa-
tions expressing the symmetry group of spacetime, but there is no divine
reason why this has to be true. The fact it seems to turn out to be true
makes sense to us aesthetically, but we do not really understand why our
sense of aesthestics works out to be how Nature behaves. There could be
some deeper philosophical or maybe even physical reason, which we do
not currently know.
The best way to put the situation for the elementary particles is the
1 All these subjective elements of known reality may be generated by material objects
and physical processes, but they are not described by any physics, and cannot ever be,
by definition. Physics is a science of objective reality. At least it is when I am lecturing.
old atomic theory: the atomic elements are not fundamental, but the lep-
tons, quarks and gauge bosons just might be fundamental, and so we’d
philosophically expect them to be like prime building blocks for matter.
As primes they cannot be factorized. That’s a metaphor with prime num-
bers in mathematics, but is nothing more than a metaphor. Perhaps the
deeper principle might be found in geon theory; there is a rough idea that
a geon is so fundamental that there is no way to rip up a geon without
tearing up or disconnecting spacetime. To my mind this would be a deep
enough principle that I would find thoroughly satisfying. You could ask,
“Why not allow spacetime to get ripped?” Sure, why not, but then I think
we are in the realm of scifi — where’s the evidence spacetime can be torn
apart? It is certainly not inside black holes which do the opposite, they
are extreme glue.
It is worth also noting that tearing a rip in spacetime that disconnects
a geon is precisely a loss of unitarity, because a quibit representing a geon
cannot just vanish without violating local unitary time evolution. This
was a puzzle in the early days of quantum gravity where baby universes
were thought to be allowed by topological fluctuation of the metric. In or-
thodox canonical Euclidean quantum gravity this is still an unresolved
problem. In geon theory there is no such problem because spacetime can-
not get disconnected using the mechanisms of metric fluctuations, be-
cause in geon theory there are no “quantum fluctuations” so-to-speak,
rather the metric merely appears to fluctuate due to the closed timelike
curve dynamical effects.
On Q.3 — why unitary representation? Why indeed? Well, we need

some mathematical object to represent the particles! Mathematics is
not physics. So the mathematical description is always a representation.
The question here really is why a group representation. I have already
explained why, it is not that groups are fundamental, it is that physical
particles we observe for real in nature seem to be invariant unless they
interact. That is why. Group theory is just the mathematical language
for describing anything that has invariants under symmetry transforms.
The physics part of this is that we observe rotation and translation invari-
ance, and with appropriate coordinate transforms we also see Galilean
or Lorentz invariance. So you see it is not the mathematics imposing any-
thing on the physics, it is always the physics demanding some particular
mathematics. It is never otherwise.
There are string theorists who might argue otherwise — for instance,
they argue that extra space dimensions are required, or predicted, by

string theory. Yes, but string theory is not physics. . . not yet, so there is
no prediction here, and there likely never will be. However this is a bit
subtle. Let’s consider the precedence of the “prediction” of anti-particles,
or the positron in particular. Dirac theory did not predict anti-particles,
because it did not require anti-particles. Dirac theory suggested anti-
particles were consistent with the theory of electrons. This is not an
existence prediction. Dirac also showed that magnetic monopoles were
consistent with the theory of electrons, and none have ever been found.
I guess when some magnetic monopoles are found people will start rav-
ing on about how Dirac accurately predicted them a thousand years or
whatever before one was discovered. That’d be baloney. Dirac made no
such prediction, he just showed magnetic monopoles might possibly exist.
(Most of you probably know there is a duality here: if magnetic monopoles
were common then isolated electric charge poles would not be.)
What is true is that string theory requires extra space dimensions, 6
extra to be precise. The story then is that if other results confirm string
theory is basically correct, then we will have a prediction of extra space
dimensions. The other way the story could go is that we first observe
extra space dimensions.
Technically what we will see is that;
1. If we desire positive-definite norms for particle probability density

in the orthodox QM formalism, then the representation for the par-
ticle has to be unitary.
2. We will find invariants, known as Casimir operators, of the group

have eigenvalues 𝑚2 and 𝑚2 𝑗 ( 𝑗 + 1). The closest (the only) way to
interpret these invariants without magic is that 𝑚 is mass, and 𝑗 is
spin. These will turn out to be the only labels we need to identify
particles in the Poincaré algebra.
3. Later we will add scattering theory, which is a bit sketchy, but intu-
itively appealing, and with scattering theory we will be able to see
why electromagnetism, the other Yang-Mills forces, and gravity are
the only possible forces in Nature. None others can exist. (Assum-
ing scattering theory is well-founded.) I find this to be one of the
most remarkable discoveries in physics of all time. When we get to
this stage we will have finished, for my present purposes.
The story of particle physics continues beyond our story however, because
further symmetries have been discovered in the atomic nuclei, namely
the color symmetry governing the strong nuclear force, and the flavour
symmetries governing the electroweak force. These are all Yang–Mills
forces. Essentially the same methods can be used as for the Poincaré
algebra, but using so-called internal symmetry groups.
I would ask all students what can possibly be the source of the sym-
metries of these internal symmetry groups? It has to be physical phe-
nomenology, unless you believe in magic. No one has ever offered a phys-
ical grounding for these internal symmetries other than fine structure of
spacetime geometry, such as the Calibi-Yau compactifications of String
theory, or fictional constructs like gauge field fibre bundles.
The problem with gauge field theory here is that if you stick a lit-
eral, not mathematical, fibre bundle at each point in spacetime then you
have added geometry. The mathematical properties of those fibre bundles
are describing something. What are they describing? No one has ever
proffered a convincing answer different to string theory or some Kaluza–
Klein theory.
We would use the color symmetry, SU(3), to find physical invariants
and with the index labels for the new irreducible representations for the
SU(3) group we can label further elementary particles. Remembering,
when we do this, that the particles are not the labels. The particles are
not even the Hilbert space objects that have these labels. The labels are
just abstract symbols identifying the particles by discrimination firstly,
and secondly by how the particle transforms upon Poincaré or now color
symmetry transform operations. We are not identifying particles by any
ontology. We have no idea what the quanta particles are, the closest
we can say (at the present date) is they might be fundamental strings
and branes. We will not carry out this quantum electroweak or quan-
tum chromodynamics procedure because although it adds complexity and
richness, yet without topological models that go beyond the group theo-
retic model we can find no further insight into the nature of fundamental
particles.
For sure, in geometric algebra and geon theory context the electroweak
and color symmetries might eventually help identify some underlying
topology, but so far no one has come up with a good scheme other than
some basic algebraic ideals [7–17]. But those explorations are pure alge-
bra, they have not yielded insight into the spacetime topology of putative
geons. It might be easy once someone gets the basic idea, but no one
seems to have found a good principle yet. String theory is still the only
game in town.
2.1.1 Motivation — symmetries of the 𝑆-matrix
Steven Weinberg gave the first account of the constraints on the types
of fundamental particles based on the orthochronous inhomogeneous
Lorentz group. Those are fancy words for ignoring the discontinuous
symmetries of space inversion ( 𝑃 -symmetry) and time reversal invari-
ance (𝑇 -symmetry) of the full Poincaré group.
It is a famous result that Nature does not actually respect those 𝑃
and 𝑇 symmetries anyway, so it was right to ignore them, however na-
ture does seem to respect 𝐶𝑃𝑇 -invariances, which means space inversion
combined with time reversal and charge conjugation. If all three are si-
multaneously applied then the physics is invariant. We are not going to
discuss 𝐶𝑃𝑇 in this chapter, but if you want a look ahead then chapter 3.3
of Weinberg Vol-I is a place to look.2
Also, as a teaser, a thread that I will leave hanging, doesn’t the fact
the combined 𝐶𝑃𝑇 symmetry get respected in nature tell you that electric
charge is probably not very mysterious, but might have some simple ge-
ometrical interpretation? It does to me, and I can guess what it is; since
charge is discrete, quantized, there is no doubt in my mind that electric
charge has something to do with topological indexes in gnarly wormhole
spacetime topology on the Planck scale. It cannot be too gnarly though,
since we only see three types of charge, ±1, ±1/3 and ±2/3. That’s where I
will leave this thread.
To dig in to Weinberg’s results we first need to wade through some
lengthy but fairly lightweight group theory. In the next few sections we
will go through the following pedagogic sequence:
• Use the above philosophy to justify searching for unitary irreducible

representations of the spacetime symmetries as abstract models
for our “fundamental particle.” We should not fool ourselves. The
2 Weinberg does not write in an easy-to-read manner, so you have to read much of the
previous chapters to get the flow of his text.
mathematical structures we find (the irreducible representations)

will not be the particles, they will describe algebraic transformation
laws for “fundamental particles” but nothing more than that. In
oither words, we are not going to be able to find geometric-topology
models for 4-geons and anything.
• Find that we require connected groups in order for Lie algebra the-
ory to be applicable.
• Once we have a Lie algebra we can find its generators.
• Once we have Lie algebra generators we can find the commutation
relations for the them.
• The commutation relations will allow us to compute unitary irre-
ducible representations of the parent symmetry group.
By the last bullet point we will be done with studying the Poincaré sym-
metries and will have representations for both classical and quantum
physics for a huge class of “fundamental particles,” not all of which are
seen in nature. We will then have to go further into scattering theory
basics to see how to further reduce the number of physical elementary
particles down to spins ≤ 2.
Unfortunately I am not aware of any way to obtain the scattering the-
ory constraints without some quantum field theory formalism. We will
need to know a very small basic bit of information about the poles for the
scattering amplitudes. It does not require much QFT, just a little. Once
we have that then we will be able to show why we can get spins up to a
graviton (spin= 2) but no higher. This will tell us why in Nature we see
only gravity and the Yang–Mills forces, no other fundamental forces. (All
other forces arising from particle exchange will be derivative of Yang–
Mills, for example Pauli overlap forces in fermions, van der Waals forces,
and all the rest).
2.2 Group theory preliminaries
It is kind of depressing to realize if you want to be a great theoretical

physicist instead of a merely good one, you have to limit your scope of
expertise. Normally this means either sucking at experiment or sucking
Group theory preliminaries 15
at mathematical rigour, or something in-between. It just takes too much

time in life to learn all the mathematics from scratch. So there are some
things we need to go and find a mathematician for and give them coffee
to produce, and then like a mercenary use their results as if we knew
all the proof, knowing all along we have no idea how to prove the results
(or at best some inkling). In short, if you bother to prove everything in
terms of the tools we want to use, from first principles, you will rapidly get
nowhere in advanced physics. Physics and mathematics truly are social
endeavours, do not let any libertarian or freak genius tell you otherwise.
In this section we collect a few such tools and if we can be bothered
will give a sketch of why they are true.
2.2.1 Quantum mechanics preliminaries
My overarching goal is to explore and promote Geon theory, which seeks

to derive QM from GR. For this chapter I wanted to cut straight to QM,
so I will skip over how Topological 4-Geon (T4G) theory gives us quan-
tum mechanics. I will assume we’ve already done that. This would then
justify use of a Hilbert space to represent knowable information about
states of a system.
The emphasis is on knowable information. Geon theory does not pre-
tend there are Hidden Variables, so T4G theory agrees that QM is as
complete an account of elementary systems as is possible — but from a
time oriented perspective. Geon theory asserts that QM is incomplete
but is as complete as we can manage if we demand a Hamiltonian time
evolution description. But T4G theory goes further to claim that general
relativity is already a quantum theory, and that the fundamental parti-
cles are real structures that involve non-trivial spacetime topology. While
existing T4G theory, unlike string theory, has no explicit models yet for
the elementary particles, it is expedient to use quantum mechanics as a
rubric. Any models T4G theory ends up providing will have to conform
with QM, and so it is no loss of paradigm to use QM here as our working
assumption.
We begin with a definition of physical states as elements of a Hilbert
space H under equivalence up to scalar multiples, which just means we
will want to normalize all states to unit norm, so probabilities are sensi-
ble — in accordance with the probability postulate 5 below. This means
the actual physical space is a ray space in the Hilbert space, where all
vectors with the same “direction”3 are equivalent physical states, so the
physical space is denoted, H /∼, which just means the Hilbert space mod-
ulo equivalence by normalization. This is also called Projective Hilbert
space, sometimes denoted 𝑃 (H ).
I should make a philosophical note here: these physical states are not
complete models for physical reality, they are rather minimal information
states. They suffice to carry all the information we need to make prob-
abilistic statements about measurement outcomes on the system. The
way we describe the information propagation in time is not equivalent
to the actual physical stuff out there. Need an analogy? How about; the
Fourier transform of a radio message is not the actual words of the mes-
sage. Something a bit like that is what a choice of Hilbert space means,
it means we are choosing convenience over exact reality. We lose some-
thing physical when we choose to do quantum mechanics with a choice
of Hilbert space representation of systems. We lose forever the ability to
precisely trace fundamental particles.
To quantum theorists it is a moot point whether or not there is some
way to retrieve determinism, the fact is Nature herself tells us determin-
ism is impossible unless you can time travel and find out the result of a
measurement before starting the measurement. If you could time travel
like that then yes, you would have access to information that would mean
you do not need a Hilbert space representation, and you could essentially
do classical mechanics, albeit in some new formalism that takes into ac-
count time travel. That would be an action principle of some sort aug-
mented by data from the future. Good luck with that.
However, the fact time travel can be considered a metaphysical possi-
bility means the Hilbert space representation is incomplete, it is not de-
scribing physical stuff as physics is, it is a formalism that acknowledges
in practice we (humans) do not have time travel. If some other sentient
species in some other galaxy had time travel they’d (a) be like gods, and
(b) have no need for quantum mechanics, unless they were lazy and time
travel was say super costly.
3 The
term “direction” does not have the same meaning for a Hilbert space over a
complex scalar structure as it does for Euclidean vectors over the reals. All vectors
that can be got from one to another by multiplication by a complex number (or a rotor,
in geometric algebra language) are taken to be “in the same direction,” but this more
technically just means “in the same ray.”
Before proceeding with the group theory stuff, one other notation I will
use is the Dirac ket symbol for vectors in the Hilbert space, and when
I want to indicate a ray (an equivalence class of vectors) I will use an
underbar, so,
|𝜓i is a vector in H
|𝜓i is a ray in H /∼ in the equiv. class [𝜓] .
|𝑟i will be a ray, so is the same as | 𝑟 i .
As per the usual postulates of QM;
1. Operators acting on the Hilbert space of states represent physical

observables if they have real eigenvalues, which means the physical
operators are Hermitian,
𝑂ˆ † = 𝑂.
ˆ
2. In highly abstract formalism, we pretend there are experiments we

can conduct on physical systems to record outputs on instruments
that are eigenvalues of some operator. Acting with the operator, say
𝐴ˆ , on the state of the system |𝜓i, yields the eigenvalue 𝑎 when,
𝐴ˆ |𝜓i = 𝑎 |𝜓i .
That mathematical statement of linear algebra is our model for

“making a measurement.” If our written |𝜓i is a close approxima-
tion to some real physical system in our laboratory, and if in many
repeated trials we make some measurement and find a value= 𝑎
recorded on our instruments, then 𝐴ˆ is the operator which closely
models our act of making that measurement. If this all sounds very
imprecise and not “textbook,” to you, then . . . get over it.
3. (Born Rule.) The probability of a transition from a state |𝜓𝑖 i to a

state |𝜓 𝑓 i is given by,
2
𝑃 (𝜓𝑖 → 𝜓 𝑓 ) = h𝜓 𝑓 |𝜓𝑖 i
where h·|·i is the inner product on the Hilbert space, and we’ve
taken the squared norm.
4. A QM system can exist in a superposition of states, which just

means there is a basis set of states, say | 𝑎𝑖 i, for which the general
state |𝜓i can be written as a linear combination,
Õ
|𝜓i = 𝛼𝑖 | 𝑎𝑖 i
𝑖
5. (Measurement postulate.) The probability of obtaining the particu-

lar measurement result 𝑎𝑘 =the eigenvalue of the operator 𝐴ˆ when
we make an 𝐴ˆ type measurement on a system which is in a super-
position like the above, is given by,
𝑃 (𝛼𝑘 ) = |𝛼𝑘 | 2 , where 𝐴ˆ | 𝑎𝑘 i = 𝑎𝑘 | 𝑎𝑘 i .
This is the first clause of the postulate. Please note the notation
brevity: 𝑎𝑘 is a scalar number, | 𝑎𝑘 i is a Hilbert space vector (a pos-
sible state of the system, up to knowable information). The second
clause in the measurement postulate is that after actually perform-
ing such a measurement the system will be in the state | 𝑎𝑘 i, not in
the initial state |𝜓i.
Comment. There is no god-given reason why any of these postulates

are true. They could be considered pulled out of a hat, but in reality we
know a lot of sweat, tears and struggle went into collating and organizing
experimental data clues, hunches, random hypotheses, until folks like
von Neumann eventually synthesised quantum physics into this proper
quantum theory. Empirical reality has led us to these postulate like a
moth to a lightbulb.
Comment. I do not know what to call postulate 2 other than pretend-

ing, or rather estimation. Clearly it is a useful postulate, since it connects
QM theory with real experiments. However, that said, postulate 4 is even
worse pretending, or “pretending for real!” Physics awaits a decent fun-
damental justification for the superposition postulate. In my mind T4G
theory provides one, but it provides superposition in inseparable combi-
nation with the Born postulate 3 and measurement postulate 2, which
to my knowledge has never been achieved parsimoniously anywhere else
before, notwithstanding fantasy Many Worlds and Hidden Variables the-
ories. Those latter theories use exotic ontology, so they increase the pos-
tulates, whereas T4G theory reduces the number of fundamental postu-
lates.
Comment on postulate 4. Most textbooks and academics use complex

numbers for the scalars. This is just a convention. Little do most students
ever know. Any isomorphic structure will do, and in my work I often
choose the geometric algebra structure, either 𝐶ℓ (3) for non-relativistic
QM, or 𝐶ℓ (1, 3) for relativistic field theory. Why? Because geometric al-
gebra is more natural, has a proper geometric interpretation of the unit
imaginaries (and in fact distinguishes several roots of −1, which is highly
valuable). We only need a structure for the Hilbert space isomorphic to
the orthodox complex scalar choice. That is provided by the even sub-
algebra of the geometric algebra 𝐶ℓ (3). So they will be a basis for our
elementary particles.
But we do need some isomorphism to the abstract complex structure,
since “it can be shown” that reals, quaternions and octonions are not
the correct algebra for physics. The reals cannot give us interference,
so that’s just an empirical restriction — so the coefficients 𝑎𝑘 have to be
higher dimensional, see also [18]. The quaternions and any higher al-
gebra yield faster-than-light signaling (though Adler thinks one can get
around this, see [19, 20]). Having said that, the quaternion and octonion
algebra are implicit in the geometric algebra of spacetime, only with the
“correct” Lorentz signatures for the latter, so there is no point to using
quaternions or octonions once we have the spacetime algebra (STA) avail-
able.
By the way, 𝐶ℓ (1, 3), 𝐶ℓ 1,3 , 𝐶ℓ1,3 , G(3, 1), ‘STA’, ‘graded Dirac algebra’,
and R 1,3 are all synonymous. I believe David Hestenes uses R 𝑝,𝑞 strictly
for the ( 𝑝 + 𝑞)-dimensional Vector space with signature ( 𝑝, 𝑞), and re-
serves R 𝑝,𝑞 for the Mother algebra which is the full graded algebra, so he
writes R 𝑝,𝑞 = G(R 𝑝,𝑞 ).
Comment. It is possible to prove that a 2-norm as in the Born Rule

is the only viable rule for obtaining probabilities in QM. I will not go
through such proof, but see [18].
Comment on postulates 4 and 2. These, to my mind, are the only

funky postulates. All the others could be classical mechanics by just
an unorthodox description in terms of information that is gained from
measurement. There just is no superposition in CM. There is also no
measurement projection, or rather, since there are no superpositions the
state of a system just changes when we make a measurement in CM by
simple pure interaction due to the experimental apparatus, and in CM
we conventionally just assume we can make this disturbance as small as

desired.
Comment. Planck’s constant ℏ does not appear. We will see why later,
but the basic reason is that ℏ is just a unit conversion factor. All the
“quantumness” in QM comes from (i) superposition, and (ii) a further
postulate we have not yet written, which is the particle-postulate:
6. (Particle postulate.) There exist states of physical systems for which

energy is localized, these are called “quanta” or elementary parti-
cles.
It is worth noting general relativity already has this quantum postu-

late, since it contains black holes. What orthodox GR does not have is
a principle of superposition and entanglement, but we can give GR such
structure was well, with geon theory (not in this chapter).
I am not sure if this 6 should be called a postulate. It is very vague, it
tells us nothing about the structure of elementary particles, so it should
be regarded as more of a principle. You can violate principles, when you
need to, but not postulates.
So much for preliminaries.
* * *
We use ordinary old linear algebra (on the Hilbert space, since it is a
vector space) to get all the usual theorems such as existence of complete
basis states, possibility of orthonormalization and so forth. Do you re-
member eigenvectors of a Hermitian operator form a basis? No matter
if you forgot, you remember now. I will not go through all that since you
can find it in any standard textbook, like Sakurai, or Ballentine.
We will use Wigner’s theorem to help classify the possible fundamen-
tal particles in QM. For this we need a definition of a symmetry trans-
formation.
Definition 2.1 (Symmetry transform). An operator 𝑇 on the Hilbert
space represents a physical symmetry of the QM system if the inner
product is preserved as follows,
h𝑇 𝜙|𝑇 𝜓i = h𝜙|𝜓i
Notice that I am using the ray representation here. Since unit scalars
disappear when we take a norm, up to scalar multiples a ray represents
a unique state, since only the transition probabilities and measurement
outcomes are observable. By this definition a symmetry transform 𝑇 will
preserve transition probabilities,
𝑇 : 𝜓 → 𝜓 0,
⇒ 𝑃 (𝜓1 → 𝜓2 ) = 𝑃 (𝜓10 → 𝜓20 ) .
So everything of physical interest is preserved by such a transforma-

tion, which justifies identifying these with symmetries of a physical sys-
tem we are supposed to be describing in the QM formalism.
2.2.2 Symmetry representation theorem
Critical to finding unitary irreps. will be the result that unitary repre-
sentations can in fact be found. Without this theorem we might just get
lucky and find the Poincaré group does have a unitary representation,
but it is nice for the mathematicians to tell us one will exist!
Theorem 2.2 (Symmetry representation). Any symmetry transform on

a Hilbert space can be represented by an operator that is either,
(i) Unitary (hence linear), or
(ii) anti-unitary (hence anti-linear).
Hence from the definition above,
𝑇 𝜓 = 𝑈ˆ 𝜓
for some unitary operator 𝑈ˆ .
Since time reversal invariance is the only useful anti-unitary opera-

tion in fundamental theoretical physics, we can pretty much ignore the
second clause, and we are then guaranteed to be able to find a unitary
representation. This means we can satisfy the physical intuition that
fundamental particles go around the place preserving information.
Proofiness. It should not surprise anyone with physics intuition that

the Hilbert space formalism is closely associated with conservation of
probability, and that means unitary operators describe interactions and
measurements. The idea is that in QM a symmetry transform of the
physical system leaves the transition probabilities invariant, so if the
transform,
𝑇 : H /∼ ↦→ H /∼
then,
(a) 𝑇 is a bijection, so is an automorphism,
(b) for any ray | 𝑟 i, the transition probabilities transform invariantly,

so,
𝑃 (| 𝑟 i → | 𝑟 𝑛 i) = 𝑃 (| 𝑟0i → | 𝑟0𝑛 i)
with “a bit more work” a good mathematician can show with these prop-
erties a representation of the symmetry transform can be found which
satisfies,
(1) if a state |𝜓i is in the ray | 𝑟 i before the transformation, then after
the transformation the transformed state is in the transformed ray,
that is,
if |𝜓i ∈ | 𝑟 i then 𝑈ˆ |𝜓i ∈ | 𝑟0i
and,
(2) 𝑈ˆ † = 𝑈ˆ −1 .
The first requirements says 𝑈ˆ represents the symmetry, and the second
says it is (anti)unitary.
If you are feeling strong then have a go at proving this theorem. You
will need some more background in group theory and linear algebra than
I am prepared to provide, but you will not need too much. Also consider
just looking up the proof. This is the ethical hackers method for math-
ematical physics. You never write all your code from scratch, you read
clean code others have written, and only if you can follow it should you
copy it, but if you can follow it and it seems clean then you should copy
it and re-use.
* * *
What of the theorem itself? As part of the Theorem 2.2 we saw for
every physical symmetry transform of a QM system there is an associated
symmetry group. The theorem tells us for every element of the group
there is an (anti)unitary operator which represents that element (which
means that operator performs the transformation on the physical states,
just as the group element performs the precisely analogous transform on
the abstract group).
It does not tell us how to find a unitary representation! We did not
really need the mathematicians to tell us this theorem either, but it is
still nice to know, because it means we can use abstract group theory.
If unitarity was not guaranteed, then we would be stuck needing to find
a different set of algebraic tools to work out how to model fundamental
particles. What is nice about Theorem 2.2 is that it says group theory
will be useful to physics. It is a “lucky” break, because group theory is so
primitive — this means we can we can go a long way with it in theoretical
physics, for very little in we get a lot out. I hate to imagine what particle
physics would be like without applicability of group theory. It’d be like
zoology I suppose. We could do it as science, but it would not be pleasant.
2.2.3 First isomorphism theorem
The easiest way to find irreducible group representations is by examin-

ing the Lie algebra of the group, which is the algebra of commutators
of the group generators obtained from expanding the group composition
function about the identity (don’t worry if these are “just words” to you
now, they will become clear later).
Looking ahead, what we are going to find is that the group of rotations
SO(3) will be a disconnected group, which means we cannot easily find a
unitary irreducible representation. We need a connected Lie group to do
this. This will lead us to look for a universal cover of SO(3) which we will
see is the connected group SU(2). Thus the Lie algebra 𝔰𝔬(2), of SU(2),
will give us the irreducible representation.
We will then repeat all this for the proper Lorentz group SO+ (3, 1),
which is also disconnected, and we will find the connected cover is
SL(2, ℂ). So the Lie algebra 𝔰𝔩(2), of SL(2, ℂ), will give us the unitary
irreps. for the Lorentz group. Once we have done that we will have
completed our objective, which is just to show that one suitable repre-
sentation of elementary particles in QM is in terms of momentum and
spin eigenstates, because the mass squared 𝑚2 and the spin will be
invariants, and moreover the spin numbers, 𝑗 , will be shown to be whole
number multiples of half-integers. These spin indices are labels for the
particles, they are not quantized units of rotational motion, this is im-
portant to keep in mind, because there is nothing inherently “quantum
mechanical” about what we are going to do. It is unusual to select a
Hilbert space for our states, but we do not get quantum mechanics out
of this, we could formulate classical mechanics instead as also a set of
states in a Hilbert space. The quantum mechanical things occur when
we include the postulates of superposition and entanglement.
It will be useful just to know how the connected covers of the discon-
nected Lie groups arise. This is the following First Homomorphism The-
orem. First a definition:
Definition 2.3 (kernel of a map). The kernel of a map 𝑓 is the set of ele-
ments in the domain that get mapped onto the identity in the codomain,
so for,
𝑓 : 𝑋 → 𝑌,
kern( 𝑓 ) = { 𝑥 ∈ 𝑋 | 𝑓 ( 𝑥) = 𝟙𝑌 }
Theorem 2.4 (First homomorphism theorem). Given a surjective homo-

morphism between two groups, 𝐺1 and 𝐺2 ,
𝑓 : 𝐺1 → 𝐺2
then the kernel of 𝑓 is a normal subgroup of 𝐺1 , but more importantly, the

quotient of 𝐺1 by kern( 𝑓 ) is isomorphic to 𝐺2 , that is,
𝐺1 /kern( 𝑓 ) 𝐺2 .
Proofiness. It is not hard to show that kern( 𝑓 ) is a normal subgroup

of 𝐺1 . It is a subgroup because if 𝑔, ℎ ∈ kern( 𝑓 ) then so is 𝑥 = 𝑔ℎ, since,
𝑓 ( 𝑔ℎ) = 𝑓 ( 𝑔) 𝑓 ( ℎ) , because 𝑓 is a homomorphism

= 𝟙𝐺2 .
The fact kern( 𝑓 ) is a normal subgroup ensures the quotient space

𝐺1 /kern( 𝑓 ) is a group. It is normal because suppose ℎ ∈ kern( 𝑓 ), then
for any 𝑔 ∈ 𝐺1 ,
𝑓 ( 𝑔ℎ𝑔−1 ) = 𝑓 ( 𝑔)𝟙𝐺2 𝑓 ( 𝑔−1 )
= 𝑓 ( 𝑔) 𝑓 ( 𝑔−1 )
= 𝑓 ( 𝑔 𝑔−1 )
= 𝑓 (𝟙𝐺1 ) = 𝟙𝐺2
⇒ 𝑔ℎ𝑔−1 ∈ kern( 𝑓 )
most of which all follows just from using the fact 𝑓 is a homomorphism
(so also maps identity in 𝐺1 to the identity in 𝐺2 ). But you know any
homomorphism that is bijective will be an isomorphism, and we are given
𝑓 is surjective, so just show when you mod 𝑓 out by the kernel you get an
injection. To finish up then we need to show if [ 𝑎] ↦→ 𝑦, and [ 𝑏] ↦→ 𝑦, then
[ 𝑎] = [ 𝑏]. If 𝑎 ∈ kern( 𝑓 ) then 𝑦 = 𝟙 by definition, so 𝑏 ∈ kern( 𝑓 ) as well,
so [ 𝑎] = [ 𝑏]. Otherwise 𝑎 and 𝑏 will be either in the same or different
cosets of 𝐺1 by kern( 𝑓 ). (Remember, the equivalence classes here are the
cosets.) For ℎ ∈ kern( 𝑓 ), if 𝑎ℎ and 𝑏ℎ map to the same element of 𝐺2
they must be in the same coset, hence [ 𝑎] = [ 𝑏] again. So we’ve got an
isomorphism. □-ish.
2.2.4 Lie groups
Some background reading is necessary which you can get from almost
any textbook on Lie groups. The basic stuff we need is to know that a
Lie algebra is defined by the commutation relations for group generators
connected to the identity. That means if we are to use the simpler tool
of Lie algebra for our physics we need to know whether our Lie group
is connected or not, and if it is not connected then we need to find the
component that is connected to the identity. Once we know this, then
exponentiation of the Lie algebra generators will give us group elements
and group representations, but not only that — crucially — irreducible
representations.
Got it? Irreducible representations (irreps. for shorthand) can be
found from the Lie algebra generators, by exponentiation. We will start
with unitary groups, so unitarity will be taken care of, provided we study
physical symmetries as per Theorem 2.2.
If along the way we make sure the generators are anti-Hermitian

then the group elements they represent will be Hermitian, and so the
induced unitary irrep. will be physical. That means the invariants when
the group acts on physical states (of momentum and position) will give
us appropriate group theoretic labels for physical invariants. A physi-
cal invariant that cannot be reduced to something simpler is how we, or
Wigner, defined an “elementary particle.” To be more exact, the invari-
ants really should have something to do with position and momentum, to
be particles rather than magical things. So when we begin we will make
sure we start with symmetries of position and momentum, which are the
Poincaré (Lorentz) group.
After finding the physical invariants of the resulting unitary irreps.
of the Lorentz group then that will be the end of our journey (for this
chapter). These invariants will be the appropriate labels for the quantum
states of “elementary particles,” and their associated fields, should we
ever want to do quantum field theory.
Definition 2.5 (Connected Lie group). A connected Lie group can be de-
fined as follows:
(i) the group elements are parameterized by a set of real numbers, say
{𝜃 1 , . . . , 𝜃 𝑛 }, where 𝑛 is the dimension of the group. Abstractly then
the group is,
𝐺 = {𝑇 (𝜃 𝑎 )}, 𝑎 ∈ {1, . . . , 𝑛} .
(ii) the composition law is given by some analytic function 𝑓 :

𝑇 (𝜃 𝑖 ) ◦ 𝑇 (𝜃 𝑗 ) = 𝑇 𝑓 (𝜃 𝑖 , 𝜃 𝑗 )
so the key thing is the functions 𝑓 𝑎 (𝜃 𝑖 , 𝜃 𝑗 ) have a Taylor expansion.
(iii) The set of all parameters is path-connected in ℝ𝑛 , so {𝜃 𝑎 } ⊂ ℝ𝑛 .
• Also 𝟙𝐺 = 𝑇 (𝜃 𝑎 = 0 ) (by convention),

• by this convention, and since 𝟙𝐺 is unitary, and because all the
elements are path connected via 𝜃 𝑎 all being connected plus 𝑓
being smooth, this means all the elements are unitary. We can
write a general element as 𝑈ˆ (𝜃), using the physicists notation
𝑈ˆ for unitary operators.
The third property is critical for us: because we seek unitary representa-
tions (for quantum mechanics probability conservation) we must seek a
connected Lie group, not just any old Lie group for the Lorentz group.
The last few properties mean that the group will not have any anti-
unitary or anti-linear operators, since path-connectivity means all the
elements are smoothly connected to the unitary 𝟙𝐺 .
Note that we really should have a vector symbol on the 𝜃 since they
live in ℝ𝑛 , but in most physics courses the vector symbol is reserved for
Euclidean ℝ3 and so because we do not want to miss-prime the brain
we will just assume the Lie group parameters are understood to be 𝑛-
dimensional reals.
The unitary operators are linear so can be always written as matrices
if desired, we then get by definition a representation, because they satisfy,

𝑈ˆ (𝜃 1 ) ◦ 𝑈ˆ (𝜃 2 ) = 𝑈ˆ 𝑓 (𝜃 1 , 𝜃 2 ) .
(this is the definition of a (unitary) representation, provided det 𝑈ˆ ≠ 0,

which is true, since det 𝟙 = 1 and all elements are smoothly connected to
the identity.)
2.2.5 Lie algebra
It is tricky to introduce Lie algebra before providing some examples of

actual Lie groups. So the students should feel free to read ahead to § 2.2.6
and then come back to this section. It’s a disease to think you need to
read physics books in a sequential fashion, but if the writer is great at
pedagogy then sequential should be good, I am not so good at pedagogy.
As a mental primer, the prototypical Lie group most people can refer to
is the group of rotations. All Lie groups are smooth manifolds, so what
is the rotation group when conceived of as a manifold? Here it is even
more expedient to drop down to 2-dimension space, where the group of
rotations is pretty simple, it is split into two parts, clockwise and anti-
clockwise (choice of axial vector ‘up’ or ‘down’) plus an angle 𝜃 ∈ [0, 𝜋],
and that’s all. This rep. has a degeneracy though, because anti-clockwise
rotation by 𝜋 is the same as clockwise. Alternatively one can have only
clockwise rotation, in which case the manifold is 𝜃 ∈ [0, 2𝜋), then you
can see the manifold is the unit circle. (Not a filled-in disk, just the circle
perimeter, a 1D manifold, as befits 2D rotations which need only one

parameter.
If we embed the 2D rotations in 3-space then we need a further set of
two parameters to specify a rotation axis (two extra parameters suffice
since the axis vector can be normalized), but then this is a subgroup of
SO(3). This gives us a spherical shell manifold. One point on this sphere
is only for one plane embedded in 2D space (the plane normal to the
vector from the origin to the chosen point on the 2-sphere). To get all 3D
rotations, hence the whole of SO(3) we need to stick a 2-sphere at every
point on the 2-sphere. But this overdoes things, since as we noted before,
a rotation by 𝜃 in a plane with normal 𝑛ˆ is equivalent (the same group
element) as a rotation of 2𝜋 − 𝜃 in the plane with normal direction −𝑛ˆ .
(A bit of jargon: we say the ‘rejection’ of a vector 𝑣 in geometric algebra
is the set of all vectors orthogonal to 𝑣 .)
These are of course naiïvely the same plane, but that ignores the fact
planes have an orientation in geometric algebra, so the ‘rejection’ of +𝑛ˆ
is the same space as the rejection of −𝑛ˆ , but they define different planes.
In G 𝐴 it is more conventional to define a plane by a bivector that is the
outer product of any two non-parallel vectors in the plane, so,
𝑎 ∧𝑏
defines the plane with right-hand normal 𝑛ˆ provided 𝑎 · 𝑛 = 0 and 𝑏 · 𝑛 = 0,

and 𝑎 ≠ 𝑏 . Whereas
𝑏 ∧𝑎
defines a different plane with the opposite orientation, with right-handed

normal −𝑛ˆ .
So anyway, the manifold for SO(3) is a 3-ball with antipodal points
on the surface identified. The result is called the projective 3-sphere,
or 𝑃 3 (ℝ), which is the manifold of SO(3). It’s like the computer game
Asteroids, only circular instead of rectangular: if you smoothly go along
a path and hit the surface, you pop out on the antipodal point, so if you
go up through the North pole you reappear out of the South pole heading
back inwards to the origin.
This has importance in physics because it means (as we will see later
in § 2.2.7) that SO(3) is connected but not simply connected, so it pro-
vides projective unitary representations — representations that differ
by a phase factor. But enough of this preliminary talk, we now turn to

defining Lie algebras.
Every Lie group has a Lie algebra, which informally is defined as the
algebra of infinitesimal group generators near the identity 𝟙𝐺 . The Lie
algebra for a generic group 𝐺 is usually denoted in Gothic lowercase type-
setting, so 𝔤; so 𝔰𝔲( 𝑛) is the Lie algebra for a special unitary group SU( 𝑛).
This is powerful for Lie group theory because from near the Identity
one can get to absolutely any group element 𝑔 in the component of the
group connected to the identity. And the most important subgroup of
any Lie group is the one containing the identity, precisely because it will
be smoothly connected, thus a manifold, and often (though not always) a
compact manifold (a “nice” topological property).
We study Lie algebra because it is dead easy to get group representa-
tions from the Lie algebra. This is because the Lie algebra equips us with
an exponetial map from a Lie algebra generator 𝑋 onto the Lie group, for
every 𝑋 ∈ 𝔤,
exp : 𝔤 → 𝐺, via 𝑒 𝑋 = 𝑔 ∈ 𝐺
this map is defined by the limit,
𝑛
𝑋 𝑋
𝑒 = lim 𝟙 +
𝑛→∞ 𝑛
∴ 𝑒𝑥 = 𝟙 + 𝑋 , for 𝑋 infinitesimal.
The last result is of course a nice linearization. I’ve always liked this def-
inition more than the Taylor series, since it shows us the iterated map-
ping aspect of all exponentials. It’s a very physical sort of definition even
though it is pure mathematics (due to the ∞ symbol).
The entire set of possible 𝑔 mapped onto by 𝑒 𝑋 is of course only the
component of 𝐺 connected to the identity. But since we can compute ir-
reducible representations from this, they will be irreps. for the parent
Lie group too. Since in QM we are only interested in the unitary reps. —
because we want conservation of probability — it is only the component
connected to the identity that matters for us. So that’s perfect.
Now for the formal definitions. First, a Lie algebra is a vector space.
This means it has a multiplication operation and an addition operation,
which form a closed group, and these are linear. (Linearity is another
reason why the Lie algebra is easier to work with than the parent Lie
group. Think of it as working on the flat tangent plane at the identity of

the group, while the group itself can be a curved manifold.)
OK, but what vector space?
Definition 2.6 (Lie algebra). A Lie algebra 𝔤 is a vector space {𝑉, ·, +, [ , ]}
with a Lie bracket [·, ·],
[·, ·] : 𝑉 × 𝑉 −→ 𝑉
that is bilinear, meaning it is linear in both entries,
[ 𝑣1 + 𝑣2 , 𝑢 ] = [ 𝑣1 , 𝑢 ] + [ 𝑣2 , 𝑢 ]
So far this is a mere algebra. To be a Lie algebra the bracket must also
have the following further properties, for all 𝑥, 𝑦, 𝑧 ∈ 𝔤,
(i) (anti-symmetry): [ 𝑥, 𝑦] = −[ 𝑦, 𝑥].

(ii) (the Jacobi identity) [ 𝑥, [ 𝑦, 𝑧]] + cyclic perms. = 0.
End of definition.
We then can define a representation of 𝔤 for any Lie group 𝐺. It closely
follows how group representations were defined; by maps into the set of
invertible matrices. For quantum physics we will use a particular type
of representation, which is not the most general, which is matrices that
are operators on the Hilbert space of physical states. This is just for our
convenience so that our Lie algebra work is not too abstract.
Definition 2.7 (Lie algebra representation). A representation of 𝔤 is a
mapping,
𝑋ˆ : 𝔤 −→ GL(H )
which preserves the structure of the Lie algebra. So, for instance, for
every basis vector 𝑥1 in the algebra 𝔤, there is a mapping 𝑋ˆ 1 taking 𝑥1 to
some matrix in GL(H ).
Just as for groups, this mapping need not be bijective, if not then it is
not a faithful representation, but is still a representation.
In practice, to compute a representation we almost always use the defi-
nition that the Lie bracket operation is just the commutator product, and
then provided the maps 𝑋ˆ preserve the structure we will get a represen-
tation, thus for all 𝑣1 , 𝑣2 ∈ 𝔤;
(i) [ 𝑋1 , 𝑋2 ] ≡ 𝑋1 𝑋2 − 𝑋2 𝑋1 ,

(ii) [ 𝑋 ( 𝑣1 ) , 𝑋 ( 𝑣2 )] = 𝑋 [ 𝑣1 , 𝑣2 ] ,
ensures 𝑋ˆ form a representation of 𝔤.
The projective representations, part-I
Later, in § 2.2.10 we will return to the topic of projective representations

when we study the rotation group. Here I just give a wee preview. Recall
in QM the physical space is projective Hilbert space H /∼, which is the
space of rays |𝜓i, not vectors |𝜓i.
Just peeking ahead then, to get a desired projective representation of
a Lie group of interest, we need not only the representatives of the Lie
algebra, but also paths in the group parameter space. These are the 𝜃 𝑎
parameters mentioned above in the Lie group definitions.
2.2.6 The groups SO(3) and SU(2)
This section is an important example. But is more than an example, since

rotations are a proper subgroup of the Lorentz group. So irreps. of the
rotations will give part of the classification of the irreps. of the Lorentz
group SL(2, ℂ). Thus, the eigenvalues for some choice of basis in SU(2)
will also be labels for unitary irreps. of SL(2, ℂ). It’s just that for the
boost and translation symmetries of SL(2, ℂ) we will need other labels.
We will get to all that later. For now we will stick to the space rotations,
SO(3).
Hang a minute though, isn’t using SL(2, ℂ) for the Lorentz group a
sin in geometric algebra? Yes it is. Using SL(2, ℂ) (2 × 2 complex matri-
ces with +1 determinant) is really old white man tradition. The proper
(shall we say) group for Poincaré rotations is Spin+ (1, 3), the rotations in
Minkowski spacetime, with an algebra over the reals. This spin group
has precisely the algebra of rotors in spacetime. What could be more
elegant?
There is a subtlety with any isomorphism. Although SL(2, ℂ) is iso-
morphic to Spin+ (1, 3), with SL(2, ℂ) you introduce complex numbers,
whereas in Spin+ (1, 3) you have no need for that crutch, and in fact the
algebra over the reals is far more natural and intuitive, and you do not
lose any geometric insight by choosing the “correct” (or more natural)
group. The “correct” group is nearly always the one with an algebra over
the reals. Why is that? It is because whenever you employ the complex
numbers you are doing so for some geometric reason, so why not just use
the proper geometry instead? Sometimes the formula looks a bit shorter
on the line when you use the complex numbers, but that is often because
you are obscuring the full graded algebra, i.e., the full geometry. Which
do you prefer, compacter notation or full insight? Your answer to that
characterizes you as a type of physicist, the Julian Schwinger/Chen-Ning
Yang type (loves elegant formalism even when it is wrong) or the Gell-
mann/Feynman type (elegance is fine when you can get it, but being right
is overriding).
We will look at the isomorphism SL(2, ℂ) Spin+ (1, 3) later, if I have
the time and inclination. I might just skip it and go straight to using the
more natural Spin+ (1, 3) with the STA rotor representation.
SO(3) is the Lie group for rotation symmetries in ℝ3 , but abstractly in
its own right it has its own group theoretic properties that do not depend
upon physical space. Some properties;
• It is a smooth manifold.
• It is compact (closed and bounded).
• It has dimension 3, so there are always three independent genera-

tors, e.g., rotations about 𝑥ˆ , 𝑦ˆ , 𝑧ˆ cartesian axes, is one such set —
although that is a physical picture, while SO(3) itself is abstract, it
is not physical.
To illustrate what this manifold looks like we need to consider some ba-
sic topology. This will lead us to see SO(3) is not simply connected, which
means there can be a simpler group from which irreducible representa-
tions of the rotations will be found.
Then in searching for a simply connected group that covers all rota-
tions we will discover SU(2). Although in these lectures I basically did
the reverse, and go to Spin(3) + first as the more natural group for rota-
tions
Abstractly, SO(3) is the group of orthogonal 3 × 3 matrices with

determinant=+1. The determinant=+1 ensures SO(3) acts to also pre-
serve orientation, whereas O(3) which includes reflections does not
always preserve orientation.
For representations of SO(3) it is useful to switch to thinking physi-
cally about rotors (or, if you prefer, matrices) acting on vectors and bivec-
tor to produce physcial rotations of objects. In the matrix picture SO(3)
reps. are orthogonal,
𝑅𝑇 𝑅 = 𝟙3
where 𝟙3 is the unit 3 × 3 matrix., and the reps. have unit determinant,
det 𝑅 = +1.
In the geometric algebra 𝐶ℓ (3) or G(3) picture, the reps. are rotors,
− 𝐵𝜃/2
𝑅=𝑒
for any unit bivector 𝐵 defining the plane of rotation, which have the
property,
𝑅 · 𝑅 = 1, and 𝑅𝑅˜ = +1.
which are respectively the geometric analogues of orthogonality and unit

determinant. We will see these rotors are actually representations of
SU(2). This is a half-angle or “spinor” representation, which abstractly
belongs to SU(2).
For example, a rotation about the 𝑧-axis, has an invariant plane 𝑒1 𝑒2
in the G 𝐴, so the rotor is,
− 𝑒1 𝑒2 𝜃/2
R 𝑧 (𝜃) : 𝑅=𝑒
but by the G 𝐴 this can also be succinctly written 𝑅 = 𝑒−𝐼𝑒3/2 .

To get strictly SO(3) matrix-like representations we have to recognize
the rotors 𝑅 act double-sided, for any geometric multivector 𝐴,
𝐴0 = 𝑅𝐴 𝑅.
˜
Whereas, when acting only on vectors,
𝑅𝑣 = 𝑣 𝑅˜
which means we can move the right rotor over to the left, so,
𝑣0 = 𝑅2 𝑣
which gives the matrix-like representation (one-sided product). With

𝑅2 = 𝑒− 𝐵𝜃 . Maybe this should be marked as an aside, but you might
like to think about the relationship between the squared rotors and the
associated matrix. You can brute force get the one–one correspondence by
looking at the action on an arbitrary vector. Just bare in mind, the spinor
rep. is so much better because it works the same way on all objects in the
whole 8-dimensional geometric algebra 𝐶ℓ (3), not just 1-vectors. The G 𝐴
expression for the action on vectors is,
𝑣 → 𝑣0 = (cos 𝜃 − 𝐵 sin 𝜃) 𝑣
Test driving this out with 𝑧 axis rotations means putting 𝐵 = 𝑒1 𝑒2 , and
expanding 𝑣 = 𝑣𝑖 𝑒𝑖 . The rest follows. You can reconstruct any rotation
matrix this way by acting on column vectors 𝑒 𝑖 , for 𝑖 = 1, 2, 3.
The other advantage of the rotors/spinors is that it is dead easy to
get the commutation relations for the generators of the Lie algebra from
them, and without having to insert a stupid unit imaginary to concoct
anti-Hermitian generators.
Physically SO(3) is the set of transformations that preserve lengths,
angles, and orientations, because distance is preserved it also preserves
volume. Whereas O(3) is the orthogonal group in 3 𝐷, the group which
preserves only the norm ||𝑇 ( 𝑥)|| = || 𝑥 ||, it does not preserve orientations,
e.g., with a single reflection a right-handed coordinate system { 𝑥, 𝑦, 𝑧 }
flips into a left-handed system { 𝑥0 = 𝑥, 𝑦0 = 𝑦, 𝑧0 = − 𝑧 } for one example.
SU(2) is the group of 2 × 2 unitary (Hermitian) matrices with unit
determinant. We will see SU(2) and SO(3) have the same Lie algebra, but
in searching for irreducible representations SU(2) is favoured because it
has a simply connected topology.
Now the cool thing here, is that we’ve got the rotors, and they form
a representation of SU(2). Most textbooks start with SO(3), we can by-
pass all that. If you want to study the generators for SO(3) and get the
commutation relations, be my guest, but why would you bother? When
you understand the geometric algebra 𝐶ℓ (3) you see Spin(3) is really the
“correct” group for rotations, which has a nice isomorphism with SU(2)
via Pauli spinors. Hence the first thing I want to do is show you how
to get the generators and commutation relations for Spin+ (3) or equiva-
lently SU(2). We will then work backwards to most textbooks and just
for fun show SO(3) has the same Lie algebra.
The formality is the abstract definition of SU(2), which is the group
represented by 2 × 2 Hermitian matrices with unit determinant:
SU(2) = {𝑈 ∈ 𝐺𝐿 (2, ℂ) | det 𝑈 = 1 & 𝑈 † = 𝑈 −1 } .
A general element is given by a linear combination of Pauli matrices,

with 𝜎
ˆ 0 the unit matrix,
𝑈 = 𝑎𝜎
ˆ 0 + 𝑖𝑏𝜎
ˆ 1 + 𝑖𝑏𝜎
ˆ 2 + 𝑖𝑐𝜎
ˆ3
the unit imaginary 𝑖 pre-multiplies because the Pauli matrices are all
anti-Hermitian. The bijective correspondence with the geometric algebra
is that with 𝐶ℓ (3) + Spin+ (3), which is the even sub-algebra of 𝐶ℓ (3),
these are the rotors, which are the closed algebra of the set,
{1, 𝜎𝑖 𝑗 }
where 𝝈 𝑖 𝑗 = 𝝈 𝑖 𝝈 𝑗 , and each 𝝈 𝑖 is just the usual Cartesian unit basis

vector. To be explicit,

𝑎0 + 𝑖𝑎3 𝑎2 + 𝑖𝑎1
SU(2, ℂ) 3 ←→ 𝑎0 + 𝑎𝑖 𝐼 𝝈 𝑖 ∈ 𝐶ℓ (3) + .
− 𝑎2 + 𝑖𝑎1 𝑎0 − 𝑖𝑎3
Note that the pseudoscalar 𝐼 = 𝝈1 𝝈2 𝝈3 multiplying a vector 𝝈 𝑖 gives a

bivector 𝝈 𝑗𝑘 , so is even grade, as is the scalar 𝑎0 . If you really do not
believe this is a bijective correspondence and algebra isomorphism then
I encourage you for homework to check the defining relations for the Pauli
algebra, which are,
𝜎
ˆ 𝑖𝜎
ˆ 𝑗 = 𝛿𝑖 𝑗 𝟙 + 𝑖𝜖 𝑖 𝑗𝑘 𝜎
ˆ 𝑘.
In our bijection of course SU(2) 3 𝟙2×2 ↔ 1 ∈ 𝐶ℓ (3) + . The orthodox text-
book way of getting the rotation group generators and then commutation
relations is to go via SU(2) matrices. But we’ve just shown we can get
there very slickly using the rotors of 𝐶ℓ (3) + instead. In fact this is the
best pedagogical way to go, since the rotors are so obviously the proper
and correct rotation operators. The rotation matrices are really an ugly
abomination.
Theorem 2.8 (Rotors rotate). Every rotation is a composite of two reflec-

tions (basic fact). From this we can show every rotation is a double-sided
geometric product of a rotor and its reverse,
𝑣0 = 𝑅𝑣 𝑅˜
and the same operation will rotate any multivector, 𝐴0 = 𝑅𝐴 𝑅˜ .
Proof sketch. This can be seen from the fact every rotation is a compo-
sition of two reflections, and in any and all G 𝐴s the reflections are dead
easy to compute, you pick a unit normal vector to the plane of reflection,
say 𝑛 and to reflect a vector 𝑣 just do,
𝑣0 = −𝑛𝑣𝑛
This means every rotation is effected by a product of the form,
𝑣0 = 𝑅𝑣 𝑅˜
where 𝑅 = 𝑛𝑚 is a rotor, and 𝑅˜ = 𝑚𝑛 is it’s reverse. Since the geometric
product of two vectors is,
𝑛𝑚 = 𝑛 · 𝑚 + 𝑛 ∧ 𝑚
it will have scalar and bivector parts, hence is a rotor. The obstruction to
rotating bivectors and higher grade elements of the algebra is failure of
rotors to commute with them, the double-sided rotor transform takes care
of this. The rotor operation above is grade preserving since det 𝑅 = 1. It
thus takes 𝑟 -vectors to 𝑟 -vectors, is continuous under continuous change
in 𝜃, preserves the origin, preserves orientation and lengths and grade
preserves, for any multivector. We offer just an explicit demo for bivetors:
first note,
𝑎 ∧ 𝑏 = 12 ( 𝑎𝑏 − 𝑏𝑎)
with this,
( 𝑎 ∧ 𝑏) 0 = 𝑎0 ∧ 𝑏0 = 21 ( 𝑅𝑎 𝑅𝑅𝑏
˜ 𝑅˜ − 𝑅𝑏 𝑅𝑅𝑎 ˜ 𝑅˜ )
= 1 ( 𝑅𝑎𝑏 𝑅˜ − 𝑅𝑏𝑎 𝑅˜ )
2
˜
= 𝑅 ( 𝑎 ∧ 𝑏) 𝑅.
This concludes the proof. □
With this theorem under our belt, we can go straight on to finding the
infinitesimal generators of the Lie algebra 𝔰𝔲(2).
TODO: The generators

2.2.7 Topology of SO(3)
It is now time to take a little excursion in topology to see what the non-
simple connectivity of SO(3) means. What it ultimately means is that
there is a redundancy is the representations for rotations, meaning one
can get a projective representation rather than an ordinary representa-
tion. If we find a different simply connected Lie group with the same Lie
algebra, then it will have no projective representation, so will be better
for finding eigenvalues and hence labels for the quantum states.
First then, why is SO(3) non-simply connected? In naïve topology this
means the manifold has a “hole” in it, so some looped paths cannot be
shrunk (smoothly deformed) to the identity. But the manifold 𝑃 3 (ℝ (the
3-ball with antipodal points on the surface identified) does not appear to
have any obvious holes. Right. But in topology a “hole” can be a removal
of a single point. Still, it is not obvious 𝑃 3 (ℝ is a smooth structure with
a point removed.
TODO: Finish this section. Need some pictures.
2.2.8 The isomorphism SU(2) Spin(3)
The group of rotors, 𝐶ℓ + (3), restricted to the unit rotors makes the spin
group Spin(3). I think I have alreadt covered this isomorphism. You just
identify the Pauli spinors with the 𝐶ℓ + (3), but with a factor of 𝐼 = 𝝈1 𝝈2 𝝈3 ,
𝜎
ˆ 𝑖 ↔ 𝐼 𝝈𝑖
and the unit matrix goes over to the unit scalar. The old white due’s view
of rotations is SU(2), the vibrant young purist’s view is Spin(3).
2.2.9 The isomorphism SL(2, ℂ) Spin+ (1, 3)
Earlier I promised to show that the traditionally old white man Lorentz
group SL(2, ℂ) should really be thought of at first by innocent babies
straight out of the womb as the Spin group. Why? Because the Spin
group has a representation in the spacetime algebra (STA). Elements of
Spin+ (1, 3) are,
𝑅 = ± 𝑒 /2
−𝐵
for bivectors 𝐵. The bivectors 𝛾0𝛾𝑖 give the space-time rotations (Lorentz
boosts).
−𝛼𝛾0 𝛾3/2
𝑒
gives a boost in the 𝑧, i.e., the 𝛾3 direction, with rapidity 𝛼 given by,
tanh 𝛼 = 𝛽 .
To find an isomorphism, start with the spacetime algebra basis vec-

tors,
𝛾02 = −1, 𝛾𝑖2 = 1, 𝛾𝜇 𝛾𝜈 + 𝛾𝜈 𝛾𝜇 = 2 𝑔𝜇𝜈 (2.1)
𝐼 = 𝛾0 𝛾1 𝛾2 𝛾3 (2.2)
with metric 𝑔 = (− + ++). And as always the geometric product of any two
vectors splits into two grades,
𝑎𝑏 = 𝑎 · 𝑏 + 𝑎 ∧ 𝑏.
Now we can construct the isomorphism explicitly: just take coponents

( 𝑡, 𝑥, 𝑦, 𝑧) in the standard basis, then a general Hermitian matrix is,

𝑡 + 𝑧 𝑖𝑥 + 𝑖𝑦
SL(2, ℂ) 3 𝑀 =
+𝑖𝑦 −𝑡 + 𝑧
which has determinant,
det 𝑀 = −𝑡2 + 𝑥2 + 𝑦2 + 𝑧2
equal to the invariant spacetime interval.

𝑎 𝑏
SL(2, ℂ) 3 ↔∈ 𝐶ℓ (1, 3)
𝑐 𝑑
You remember from mathematics class how to establish a map is an iso-
morphism right? It has to be bijective and preserve the group composi-
tion law structure. Well, this does. Convince yourself it does for home-
work.
If you were a real mathematician, or feeling like a strong physicists,
you’d show the more general case,
𝐶ℓ 𝑝+1,𝑞+1 Mat2×2 (ℝ ⊗ 𝐶ℓ 𝑝,𝑞
see [?] or [?].

2.2.10 Projective representation, part-II
Recall that the Symmetry Representation Theorem 2.2 tells us for ev-
ery element of the symmetry transforms on a QM system there is an
(anti)unitary operator which represents that element (which means
that operator performs the transformation on the physical states, just
as that group element performs the precisely analogous transform on
the abstract group). For the continuous (Lie) symmetries of special
relativity (the Poincaré group) we then want to restrict to connected Lie
(sub)groups, because these will have the identity and so all elements
of these subgroups will be unitary. This will ensure we find unitary
representations, as per our guiding philosophy from §2.1.
Recall that the reason we are not going to be able to use the Lie algebra
to find unitary irreps. with a disconnected group like SO(1, 3) + is because
such disconnected groups have elements that cannot be obtained from
exponentiating the Lie algebra generators. Only a connected Lie group
can be completely obtained from the Lie algebra alone.
We thus find ourselves needing to find a universal covering group,
which turns out to be SU(2), which was guaranteed to exist by the Sym-
metry Representation Theorem.
The general 2 × 2 Hermitian matrix is

𝑎0 − 𝑎3 − 𝑎1 + 𝑖𝑎2
− 𝑎1 − 𝑖𝑎2 𝑎0 + 𝑎3
which acts natural on Pauli spinors. This same matrix has a represen-
tation in the Clifford algebra 𝐶ℓ (3),
®𝑖
𝑎0 + 𝑎 𝑖 𝜎
which acts on multivectors if we replace the matrix 𝜎 ® 𝑖 by the 3-space basis

vector 𝜎 𝑖 . The unit rotors (even elements) in 𝐶ℓ (3) are one-one with the
Pauli spinors, and we can identify them as one-sided rotation operators.
(All operators in a geometric algebra are also multivectors of the algebra,
a very nice unification.)
To check this correspondence is one–one all one needs to do is check
the Pauli algebra:
® 𝑖2 = 𝟙,
𝜎 ® 1 𝜎®2 = 𝐼 𝜎3
𝜎 and cyclic,
𝜎® 1 𝜎®2 𝜎3 ≡ 𝐼
{𝜎𝑖 , 𝜎 𝑗 } = 0, for 𝑖 ≠ 𝑗.
That is exactly the Clifford algebra 𝐶ℓ (3). So by using Pauli spinors we
have not in fact gone into “quantum mechanics” we are still in classical
Euclidean coordinate geometry.
The same applies for the Dirac algebra which is the classical geomet-
ric algebra of 𝐶ℓ (3, 1), that is, the classical spacetime algebra. Things
only start to get “quantum” when we introduce discrete energy packets,
together with their superposition and entanglement. It is not sufficient
to just have wave packets with energy 𝐸 = ℏ𝜔, to get actual QM we must
have superposition and entanglement as well.
However, there is one subtlety of importance to physicists in these
representations, which is that in QM we only need the representation
property to hold up to a phase factor 𝑒 𝐼 𝜙 . Recall that the group 𝐺, we are
concerned with, is associated with some physical symmetry or invariance
transform 𝑇 , we wrote this,
𝐺 = {𝑇 : H /∼ → H /∼ such that 𝑃 (| 𝑟 i → | 𝑟 𝑛 i) = 𝑃 (| 𝑟0i → | 𝑟0𝑛 i)} .
This slightly indirect definition of the group from the transform 𝑇 is nec-
essary because the physical states in H are represented by rays |𝜓i. For
instance, suppose we have some symmetry group 𝑇𝑖 , with composition
law
𝑇1 ◦ 𝑇2 = 𝑇12
then an ordinary representation will have,
𝑈ˆ (𝑇1 )𝑈ˆ (𝑇2 ) = 𝑈ˆ (𝑇12 )
But suppose we have a state |𝜓i in a ray | 𝑟 i. Then,

𝑈ˆ (𝑇1 )|𝜓i ∈ | 𝑟0i ⇒ 𝑈ˆ (𝑇2 )𝑈ˆ (𝑇1 ) ∈ | 𝑟00i
but then since the physical states are equivalent only up to their rays we
can have a looser requirement on our representations, namely,
𝑈ˆ (𝑇12 )|𝜓i ∈ | 𝑟00i
∴ 𝑈ˆ (𝑇2 )𝑈ˆ (𝑇1 ) = 𝑒 𝐼 𝜙(𝑇1 ,𝑇2 ) 𝑈ˆ (𝑇12 )
then if any of these 𝜙 are non-zero we have a projective representation of

the group 𝐺, rather than an ordinary representation. In QM these are
perfectly fine.
Lie algebra
First, lets cover soe motivatins. The reason we want to find the Lie al-
gebra for our physcial symmeteries is in order to be able to construct a
projective representaiton.
Lie algebra and structure constants
Due to analyticity of 𝑓 for our supposed connected Lie group, we can Tay-
lor expand about the identity, once for 𝑓 and also for 𝑈ˆ ,
𝑓 𝑎 (𝜃 1 , 𝜃 2 ) = 𝜃 1𝑎 + 𝜃 2𝑎 + 𝐶𝑏𝑐
𝑎 𝑏 𝑐
𝜃1𝜃2 + . . . (2.3a)
1 𝑎 𝑏ˆ
𝑈ˆ (𝜃) = 𝟙 + 𝑖𝜃 𝑎 𝑋ˆ 𝑎 + 𝜃 𝜃 𝑋𝑏𝑐 + . . . . (2.3b)
2 1 2
The requirement that 𝑈ˆ be unitary means the generators 𝑋ˆ 𝑎 must be

Hermition, so 𝑋ˆ † = 𝑋ˆ . If we dropped the factor of the unit imaginary 𝑖 in
the above expansion than we’d have 𝑋ˆ be anti-Hermition, so this is just
a convention for 𝑋ˆ . Convince yourself this is so.
Now recall the condition that the 𝑈ˆ are representations,

𝑈ˆ (𝜃 1 ) ◦ 𝑈ˆ (𝜃 2 ) = 𝑈ˆ 𝑓 (𝜃 1 , 𝜃 2 ) . (2.4)
If we plug Eq.(2.3a) and Eq.(2.3b) and into (2.4), then equate terms (Tay-
lor series are equal term-by-term) we find the following relations,
𝑎 ˆ
[ 𝑋ˆ 𝑏 , 𝑋ˆ 𝑐 ] = 𝑖 𝑓𝑏𝑐 𝑋𝑎 ,
𝑎 𝑎 𝑎
where 𝑓 𝑏𝑐 = 𝐶𝑐𝑏 − 𝐶𝑏𝑐 .
These are the Lie algebra defining commutation relations.

𝑎
The numbers 𝑓𝑏𝑐 are called the structure constants.
The imaginary 𝑖 — WTF?
When it comes time to do the Lorentz group (space rotations plus boosts)
we will see the boosts are just geometric rotations but with invariant
spacetime bivectors rather than invariant space bivectors. A beautiful
unifying bit of spacetime algebra. All the proper Lorentz group operators
are all just rotors!
This will mean we can drop the unit imaginary 𝑖 in the above expan-
sion of 𝑈ˆ , and in fact the operator 𝑈ˆ will be a simple G 𝐴 rotor, a geometric
object. This is also beautiful. This means we get Hermiticity automati-
cally, there is no need to plug in an 𝑖 like a kludge. With a rotor,
− 𝐵𝜃/2
𝑅=𝑒
the object 𝐵 is a unit bivector defining the plane of rotation, and this
works for space rotations as will as Lorentz boosts, beautifully. All unit
bivectors square to negative one,
𝐵2 = −1
and that will give us the replacement for the unit imaginary in the Lie
algebra generator expansions.
I am not sure if any of this is of much concern. The complex numbers
enter into QM in another way, which we will have to later on deal with in
another chapter. In this chapter the imaginary numbers disappear, since
the invariants are real numbers, which is only the correct physical thing
to get.
The complex numbers can re-appear if we choose a matrix represen-
tation for the operators 𝑈ˆ , however, little do most students know, we
can even get rid of these, since every complex valued matrix representa-
tion can be replaced by a higher dimensional real matrix representation.
That’s a pain if you have to write out a lot of matrices. So perhaps that
is why most textbooks go for the lower dimensional complex matrix rep-
resentation when they get down to computing wave functions. We will
indirectly show that is not needed, and the lowest dimension can be rep-
resented in the reals, using the bivector algebra of space/spacetime.
Although it is not the topic for this chapter, in case you were wonder-
ing, the complex numbers get squeezed back into quantum mechanics
because the probability amplitudes have to exhibit interference, and real-
valued probabilities cannot interfere, unless they are allowed to go neg-
ative. No one, except perhaps Feynman, figured out how to make sense
of negative probabilities, so everyone sticks with tradition and uses the
complex “square roots of probabilities” called amplitudes in QM.
Poincaré Symmetries 43
I hope to show, in a later chapter, how to reinterpret complex ampli-

tudes geometrically, so that we will have an entirely compact real-valued
geometric formalism for QM, which via the techniques already developed
will translate directly over to quantum field theory as well.
2.3 Poincaré Symmetries
The orthochronous (single time direction) and inhomogeneous (no space

translation) part of the the Poincaré group is what we wish to study. (The
full Poincaré group includes space translations.) This involves,
1. Space rotation.
2. Lorentz boost (general spacetime rotations).
The reason we are not interested in the space translations is because

the symmetry of the translations is related to momentum conservation,
which holds for all particles, so is not a discrimination for labelling par-
ticles or geons. If you like, think of the translational symmetry property
as distinguishing a particle from a non-particle.
To be more pedantic I will here, once, write out the Poincaré group in
full index notation, then the main three subgroups of interest. The full
group can be specified by Lorentz transforms plus spacetime translation,
and can be considered in the first instance to act on your chosen inertial
coordinates { 𝑥𝜇 }, as follows:
𝜇
𝑥0𝜇 = Λ𝜇 𝑥𝜈 − 𝑎𝜇 (2.1a)
with the transition matrix defined by partial derivatives,

𝜇
𝜕𝜈 𝑥0𝜇 ≡ Λ𝜈 (2.1b)
and for inertial frames these require the metric be unchanged, hence,
𝜇
𝑔𝜇𝜈 Λ𝜌 Λ𝜈𝜎 = 𝑔𝜌𝜎 . (2.1c)
The spacetime vector 𝑎𝜇 for translations is the position of the origin of

the { 𝑥0 } frame in the { 𝑥 } frame. The minus sign is conventional so that a
fully positive shift of origin has a full positive 𝑎𝜇 .
Comment. Given the Minkowski metric 𝑔 ( 𝑝) = 𝜂 = (−+++) you can eas-

ily see this means the inverse Lorentz transform is given by alternately
lowering and raising the Λ indices using the metric,
𝜇
(Λ−1 )𝜈 = 𝑔𝜇𝜌 𝑔𝜈𝜎 Λ𝜎𝜌
We are not quite done. To get the Poincaré group from the above defini-
tions we need to specify the group composition rule and show is satisfies
the group axioms. To compose two transforms,
𝑥 ↦→ 𝑥0 ↦→ 𝑥00
we only need to perform a sequence of two of the above Poincaré trans-

forms, Λ1 followed by Λ2 , as follows,
𝜇 𝜌 𝜇
𝑥00𝜇 = (Λ2 )𝜌 (Λ1 )𝜈 𝑥𝜈 − (Λ2 )𝜈 ( 𝑎1 ) 𝜈 − ( 𝑎2 ) 𝜇 (2.2)
In abstract notation, denoting P to be a general Poincaré transform, the

above expression says the group compoition rule is,
P (Λ2 , 𝑎2 ) ◦ P (Λ1 , 𝑎1 ) = P (Λ2 Λ1 , Λ1 𝑎1 + 𝑎2 ) . (2.3)
Note we should always apologise for using index notation, because we

have learned we can always choose to use geometric algebra multivectors
for all of this which avoids indices. Oh well, so. . . my apologies. Maybe
some day when I am retired I will redo all of this in pure G 𝐴.
Good. Now let’s consider the four main subgroups.
1. The full Lorentz group O(1, 3): preserves spacetime lengths. (The
full P (1, 3) does not, due to Lorentz contraction).
• det Λ = ±1,
• has all translations 𝑎𝜇 = 0,
• composition rule is: 𝐿 (Λ2 , 0) ◦ 𝐿 (Λ1 , 0) = 𝐿 (Λ2 Λ1 , 0),
• Λ00 ≥ 1 or Λ00 ≤ −1, so this group is still not simply connected.
2. the proper Lorentz group is SO(1, 3) which restricts to only allow

Λ with
• det Λ = +1, this removes the reflections in 𝑥 .

Poincaré Symmetries 45
• this is still disconnected, due to time reflection symmetry.
3. The proper orthochronous Lorentz group which restricts to a

single time direction, denoted SO+ (1, 3) or SO(1, 3) + .
• restrict to Λ00 ≥ 1, this removes reflections in 𝑡 = 𝑥0 .

• this is now a connected group, and so by elementary theorems
in group theory will have a simply connected cover.
There is a fifth subgroup of the Poincaré group, so a fourth subgroup

of 𝑂 (1, 3) which takes the proper orthochronous Lorentz group (so
SO(1, 3) + ) and puts back in the translations 𝑎𝜇 and this is also simply
connected. This group has no special name.
We can now diagrammatically summarize the full Lorentz group as
four disconnected components shown in Fig. 2.1, each component is con-
nected by one of the discrete transforms 𝑃 =parity inversion, or 𝑇 =time
reflection. Notice Time inversion flips both det and Λ00 , whereas Parity
only flips the det.
O(1, 3)
SO(1, 3) +
𝑃 det Λ = −1
det Λ = 1
Λ00 ≥ 1
Λ00 ≥ 1
𝑇 𝑇
det Λ = −1 𝑃 det Λ = −1
Λ00 ≤ −1 Λ00 ≤ −1
Figure 2.1: The four components of the full Lorentz group.
The plan now is to find the generators of the Lie algebra for the
Poincaré group. Once we find the generators we can compute their
commutation relations, and use those to build a basis of unitary states
and find the eigenvalues, and those will be our unitary irreducible
representations, hence our “elementary particles.”
Notice also that the Lie algebra for the full Poincaré group will be the
same as the Lie algebra for the one connected sub-component containing
the identity. That will be SO(1, 3) + . So the problem of finding irreducible
representations reduces to studying the Lie algebra of SO(1, 3) + .
2.4 Poincaré algebra
We’ve motivated the need to look at the symmetries of spacetime (the

Poincaré group). Any further symmetry found in nature can be accom-
modated by either adding extra compactified space dimensions or by
admitting non-trivial topology. In this chapter we are only interested
in the symmetries of bulk spacetime, these being the only symmetries
needed for studying gravity (spin=2), electromagnetism (spin=1), and
Yang–Mills forces (also spin=1).
We will not look at space or time translations because we assume the
student is already familiar with Noether’s theorem in the context of clas-
sical mechanics, and thus knows why and roughly how momentum con-
servation is the associated invariant related to space translation symme-
try; and time translation invariance is associated with energy conserva-
tion (or invariance of the Hamiltonian).
We will skip then to the remaining elements of the Poincaré group,
which are the orthochronous homogeneous Lorentz transforms, rotations
and boosts, or SO(1, 3) + . For short-hand I will just call this the Lorentz
group.
The other inhomogeneous symmetries are space & time inversion,
which have consequences later when we come to spin-statistics, but we
will leave that topic for later. I will also skip some other mathematical
preliminaries and use physical arguments wherever possible under the
philosophy that physics determines the mathematics we use, mathe-
matics is only a consistency constraint to check our physics is not too
wild.
We begin then with the physical idea that the Lorentz transformations
in Einstein’s theory together with rotations are the only symmetries re-
Poincaré algebra 47
maining once we put aside space translations and the discrete symme-
tries of time and space inversion. One might worry this will miss some
elementary particles, but we can look ahead and say,
• time inversion will only add anti-particles,

• space inversion distinguishes bosons from fermions,
and so we should get all the quanta for the classical fields of gravity and
electromagnetism just by examining the homogeneous Lorentz group.
In section 2.3 we already went through the whole Poincaré group. Re-
call that we want to now look at the components connected to the identity
to study the Lie algebra. Restricting now to the connected component
SO(1, 3) + we introduce some symbols for the generators:
1. translations (displacements) in time and space (𝑃 ), forming the

abelian Lie group of translations on space-time;
2. rotations in space, forming the non-Abelian Lie group of three-
dimensional rotations (𝐽 );
3. boosts, transformations connecting two uniformly moving bodies
(𝐾 ).
The last two symmetries, 𝐽 and 𝐾 , together comprise the Lorentz group.
We may also say the Poincaré group comprises the full set of isome-
tries of spacetime or ISO(1, 3) — meaning action by the group transforms
leaves the metric invariant (preserves spacetime vector lengths and an-
gles). Space lengths and time intervals are not preserved by the general
transform, as is familiar for students of special relativity (we get 3-vector
length contradiction and time dilation). Only the pure space rotation
and translations preserve 3-vector lengths. This is so far all physics, the
mathematical language of group theory is only that, a language. Without
Michelson–Morley we would not know to be using the Poincaré group.
In Minkowski space, 4 𝐷 flat spacetime with metric signature (− + ++),
there are ten degrees of freedom for the isometries, which may be thought
of as translation through time or space (four degrees, one per dimen-
sion); reflection through a plane (three degrees, the freedom in orienta-
tion of this plane); or a boost in any of the three spatial directions (three
degrees). Proper rotations are also produced by composition of an even

number of reflections. This will be important when we come to geomet-
ric algebra, because it will show us that rotations are best thought of
as double-sided operations of rotors; each rotor being a product of two
vectors representing the pair of reflections. A unit vector represents a
reflection in 3 𝐷 because it is a normal to a 2-plane, or bivector, the plane
of the bivector is the plane of reflection.
These 10 generators (in four spacetime dimensions) associated with
the Poincaré symmetry, by Noether’s theorem, imply 10 conservation
laws: 1 for the energy, 3 for the momentum, 3 for the angular momentum
and 3 for the velocity of the center of mass. Note that it is a very strong
hint that in superstring theory the 6 extra space dimensions have to be
either compactified or cosmological (the two extremes) so that we do not
see them, roughly because we do not find more than these 10 isometries
in the world of physics.
2.4.1 The vector representation
We begin our study of the Lie algebra by considering the way spacetime
vectors transform under SO(1, 3) + . This is because our only tenuous con-
nection to physics, for now, is energy-momentum conservation. To study
the Lie algebra we need to consider infinitesimal transformation about
𝜇
the identity. To do so we introduce an infinitesimal tensor 𝜖𝜈 to param-
terize these infinitesimal Lorentz transforms,
𝜇 𝜇
Λ𝜈 ≈ 𝟙 + 𝜖𝜈
Then we can write the transform for a spacetime vector 𝑣, as
𝜇
𝑣𝜇 ↦→ 𝑣0𝜇 = Λ(𝜖)𝜈 𝑣𝜈 .
𝜇
NB: if you want tensor notion throughout then 𝟙 = 𝛿𝜈 . This definition for
𝜇 𝜇
the infinitesimals 𝜖𝜈 of course requires the 𝜖𝜈 satisfy certain conditions
𝜇
for yielding Lorentz transforms, we cannot just pick anything for 𝜖𝜈 . We
can see what the main condition is by remembering the Λ are required
to be metric preserving, hence,
𝜇
𝑔𝜇𝜈 Λ𝜌 Λ𝜈𝜎 = 𝑔𝜌𝜎
𝜇
plug the previous definition for 𝜖𝜈 into that condition and we find,
𝜇
𝜖𝜈 = −𝜖 𝜇𝜈
Invariants 49
so the condition on the generator, 𝜖, is that it be anti-symmetric. An

anti-symmetric 4 × 4 tensor has only 6 independent components.
2.4.2 Commutators
2.5 Invariants
TODO: Need to do the Casimir mass invariant . . .
Looking now at the rotations, 𝐽 , there is a bit of jujitsu involved in

getting the Casimir invariants, and I have not figured out a slick way
to do it using the G 𝐴 rotors, so I am going to show you the traditional
method.
First we form some derived generators, called 𝐽± form the 𝐽𝑥 and 𝐽 𝑦,
as follows,
𝐽± = 𝐽𝑥 ± 𝑖𝐽𝑦
using the previous commutation relations we find,
[ 𝐽𝑧 , 𝐽± ] = ±𝐽 ±, (2.1)
[ 𝐽+ , 𝐽− ] = 2𝐽𝑧 . (2.2)
Since 𝐽± is a mix of Hemrtian and anti-Hermitian, we have to scrutinize

the Hermiticity for the 𝐽𝑖 , to get a relation for say 𝐽+ †, and (homework
one-liner) we find,
𝐽+† = 𝐽−
Also, from (2.2), we can see 𝐽− 𝐽+ and 𝐽+ 𝐽− must both be Hermitian be-
cause 𝐽𝑧 is. This means,
𝐽− 𝐽+ = 𝑂ˆ † 𝑂,
ˆ for some operator 𝑂ˆ
We’ll need this in a minute.

Now, here is a first bit of mild jujitsu. I’ve never properly figured out
why the 𝐽𝑧 component is preferentially chosen, it has something to do
with a traditional convention for sticking magnetic fields up or down in
the 𝑧 direction when measuring spin/magnetic moments. Blame the ex-
perimentalists! So the “thing to do” is chose some standard basis for
states, and the choice is 𝐽𝑧 , so there is some label 𝑠 say, for the eigenvec-
tors of 𝐽𝑧 (this is just a definition and convention, ok):
𝐽 𝑧 | 𝑠i = 𝑠 | 𝑠i .
Jujitsu done! Next, consider the commutator from eq.(2.1)a acting on a

state | 𝑠i,
( 𝐽𝑧 𝐽+ − 𝐽+ 𝐽𝑧 )| 𝑠i = 𝐽+ | 𝑠i
∴ 𝐽𝑧 𝐽+ | 𝑠i − 𝑠𝐽+ | 𝑠i = 𝐽+ | 𝑠i
∴ 𝐽𝑧 𝐽+ | 𝑠i = ( 𝑠 + 1) 𝐽+ | 𝑠i
⇒ 𝐽+ | 𝑠i ∝ | 𝑠 + 1i
and similarly,
𝐽− | 𝑠i ∝ | 𝑠 − 1i or |0i
This implies the only possible 𝐽𝑧 eigenvalues from a discrete sequence,
. . . , 𝑠 − 1, 𝑠, 𝑠 + 1, . . .
Next little bit of jujitsu, although I guess you can come back to this point
and check it does not make any difference, but the “thing to do” is define
eigenvalues for 𝐽− and 𝐽+ as follows,
𝐽− | 𝑠i = | 𝑠 − 1i , 𝐽+ | 𝑠i = 𝜆 𝑠 | 𝑠i
we can do this because eigenvalue/eigenvector equalities for 𝐽𝑥 and 𝐽𝑦 are

not changed by an overall multiple, so we’ve taken this freedom away bey
setting the 𝐽− part to unity. Recall, 𝐽𝑥 = 𝐽+ + 𝐽− , and 𝐽𝑦 = 𝑖 ( 𝐽+ − 𝐽− ) is
where we get this one factor setting freedom from. Anyway, this directly
gives, provided 𝑠 ≠ 0,
𝐽+ 𝐽− | 𝑠i = 𝜆 𝑠−1 | 𝑠i , 𝐽− 𝐽+ = 𝜆 𝑠 | 𝑠i
Acting with the commutaor on | 𝑠i then gives,
[ 𝐽+ , 𝐽− ] = (𝜆 𝑠−1 − 𝜆 𝑠 )| 𝑠i
= 2 𝐽 𝑧 | 𝑠i = 2 𝑠 | 𝑠i
⇒ 𝜆 𝑠−1 − 𝜆 𝑠 = 2 𝑠. (2.3)
Invariants 51
Next a third little bit of jujitsu, we do one of these cheat look-ahead (“to
make things nice” the teacher always says. . . sure) and define an eigen-
constant 𝑗 as follows,
𝜆 𝑠 = 𝑗 ( 𝑗 + 1) − 𝑠 ( 𝑠 + 1) . (2.4)
Why? Well, just because this solves e.q(2.3). Check that it does! 𝑗 ( 𝑗 + 1)
is just a constant, so it disappears upon the subtraction, so yeah, it is a
solution. A difference equation like that always has an arbitrary constant
in the solution.
Now consider, if 𝑗 is a constant, and we suppose 𝑠 can get arbitrarily
large (positive or negative) then clearly at some worse case 𝑠 we’ll get
𝜆 𝑠 < 0. OK, but recall both 𝐽− 𝐽+ and 𝐽+ 𝐽− are both of the form 𝑂ˆ † 𝑂ˆ ,
which has a positive (or zero) eigenvalue. The eignenvalue is zero only if
𝐽− or 𝐽+ annihilate the state. This is inconsistent with 𝜆 𝑠 < 0, so it must
be true that there is a minimum and maximum eigenvalue for 𝑠 labelling
| 𝑠i, call these 𝑠min and 𝑠max .
Since, as we’ve defined them, 𝐽− lowers the 𝑠 value and 𝐽+ raises the 𝑠
value, we must have,
𝐽+ | 𝑠max i = 0 ⇒ 𝜆 max = ( 𝑗 − 𝑠max )( 𝑗 + 𝑠max + 1) = 0

𝐽− | 𝑠min i = 0 ⇒ 𝜆 min−1 = ( 𝑗 + 𝑠min )( 𝑗 − 𝑠min + 1) = 0
note that we get these from factorizing eq(2.4) in two different ways:
𝑥 ( 𝑥 + 1) − 𝑦 ( 𝑦 + 1) = ( 𝑥 − 𝑦)( 𝑥 + 𝑦 + 1) , or ( 𝑥 + 𝑦)( 𝑥 − 𝑦 + 1)
The difference between these 𝑠 extrema is also some whole number,
𝑠max − 𝑠min ∈ {0, 2, 3, . . . }
We have freedom to choose a sign for 𝑗 , so choosing 𝑗 > 0, then because

𝑠max cannot be negative, and 𝑠min cannot be positive, the left brackets in
the above equalities have to be zero. This gives us the roots,
𝑠max = 𝑗, and 𝑠min = − 𝑗.
For these last two conditions to both be true we require,
𝑗 ∈ {0, 21 , 1, 32 , 2, . . . }
and from before,

𝑠 ∈ {− 𝑗, − 𝑗 + 1, . . . , 𝑗 − 1, 𝑗 }
In other words, the basis of states | 𝑠i form a (2 𝑗 + 1) dimensional

eigenspace for the rotational symmetry operators exp[− 𝐼 𝜎𝑘 𝜃/2]. We say
these are the “spin- 𝑗 representation.”
OK, so when 𝑗 = 1/2 we get a “spin-1/2 representation” of SU(2), and
hence also of the connected component of SO(3) + , which has just (2 12 +1) =
2 possible states | 𝑠i. Usually labelled | ↑i and | ↓i. Or sometimes |+i and
|−i. Or sometimes | 𝑢i and | 𝑑i, but then not to be confused with quark
flavour. Fermions!
* * *
Mathematicians must have known this since way back, but they were
never much doing quantum mechanics. Joe Numberson says: “If you
wish to rotate use a damn matrix on a space vector! Why go to a Hilbert
space?” Jane McMath says, “Hey, do not even use matrices, use | 𝐺 𝐴
rotors.”
Since the physicists go to a Hilbert space, and want these eigenstates
and eigenvalues, for the damn Measurement postulate P.2, we have these
half-integer labels possible. But they are not about quantization of any-
thing. They are just a choice of representation for measurements. What
are they measuring? The answer is anything to do with rotational sym-
metry invariance. What’s that? Angular momentum dumb ass! The
group is a smooth group of rotations, the representions of a basis in
Hilbert space have these discrete labels, and the two are not the same
thing, they are just related. They help classify elementary particles.
Got it? So the “discrete spin quantum numbers” are classifications
of possible robust (invariant) elementary particles, not quantization of
angular momentum or anything like that.
The same comments apply if we go and look at solutions to the
Schrödinger Equation for the Hydrogen atom or whathaveyou. We there
see orbital angular momentum, which is a rotational symmetry of the
electron orbits in the H atom. We should find a set of basis states in the
Hilbert space for the atom to represent this symmetry, and it will be the
same Lie algebra as these more fundamental particles. All rotationally
symmetry system everywhere will have these integer or half-integer
Invariants 53
Hilbert space representations in their theory of measurement, quantum

or classical.
So a basketball (up to forgetting about the pump nipple and grip pat-
tern) has these numbers too, we just cannot tell if it is a boson or fermion
because it’s 𝑗 number is huge. If you get the number of atoms just right
a basketball could be a Bose-Einstein system like the Helium-4. But it
ain’t going to Bose condensate any time soon because it is too massive,
its electrons get entangled with all sorts of sh∗t all the time.
Ohmigod then! When you measure the orientation of a charged par-

ticle you need a magnetic field, since it can spin around it has a mag-
netic moment. Whichever direction you orient the magnetic fields will
dictate the eigenvector the measurement puts the particle into. Postu-
late 2. With a 𝐵 𝑧 field you are not going to be able to detect any 𝑥 or 𝑦 spin
eigenstates, so fugghedaboutit. This does not mean electron rotational
motion is quantized. It means, at best, the electron circulation (which
we know not what that is) is constant, hence the well-defined magnetic
moment, and what can be altered by a measurement is the orientation
of the spin plane. When you impose a strong enough 𝐵 field to deflect
the electron then the electron current will experience a torque, as usual
for classical magnets, and it will align with the field, and that’s why you
observe only one of two values, you are measuring a set orientation, not
an electron’s current loop orientation.
Still, it is weird, when you arrange a sequence of filters with 𝐵 fields
in crossed directions the non-classical non-distributive logic shows up
for the electrons. This could be considered the highly non-classical part
of the spin/rotational symmetry business. It’s not about the half-integer
representation of the rotational symmetry — all that does is conveniently
give you just two eigenstates to observe, which is lovely for simple filter
experiments, and then the non-distributive logic shows up clearly.
To me, the slightly weirder thing about rotational symmetry in quan-
tum mechanics is scattering. Since there is no magnetic field in funda-
mental scattering process, what is there to “align” the spins of the parti-
cles? Also, some of them will not be charged, so they will not align nicely
anyway. Yet spin number conservation applies at Feynman vertices. How
does that work? Clearly (maybe not clearly to anyone else?) the spin rules
for scattering vertices are a different phenomenon. I think a way to see
what might be going on is the total symmetry of a system plays a role, and
so in-going must match out-going. So this scattering phenomenon is not

really about particular spatial orientation, but is about local symmetry
conservation at a vertex.
A couple of other things, and to relate both the above discussions: I do
not think electron rotational motion is “quantized,” what I think is the
electron is bloody small and elementary, so if it has a charge that rotates,
that will be a pretty good bet to be a constant of nature, because electrons
have close to nil structure (they are fundamental). The “quantum nature”
of spin thus, for me, boils down just to the fact electrons, like quarks,
are the tiniest packets of charge. When I try to measure the magnetic
moment of an electron I have to hit it with photons. But where do those
photons come from? They come from the imposed (fairly strong) magnetic
field. Those photons have an imposed helicity, this is why the 𝐵 field
(a fictional concept) is oriented. So those photons are not going to be
disturbing any particles that have an incommensurate spin orientation,
let’s say. So the only electrons that are going to get deflected are the ones
with commensurate spin, they’ll gt knocked one way or the other by the
scattering rules. Once again we see we are setting up an experiment to
detect electrons of one spin or the other, whatever our laboratory frame
is designed to detect is what we detect.
What about electrons in the beam that do not get hit by a photon, don’t
they “go straight through”? Yes, but think about the number of photons
hitting each electron. How many scattering events are there? I’ll tell you
how many . . . , on second thoughts I won’t, it’s a huge number, let’s say
quadrillions. It’s something big like that. Is it any wonder almost all the
electrons do not “go straight through”? No wonder at all. They are going
to align with the majority of photons that the 𝐵 coils are generating, one
way or another. If you dial down the magnetic field you will lose such
crispness, and gradually dialing down more and more electrons will go
straight through without deflection. I am not sure of all these numbers,
but I think to see a gaussian blurred electron beam you have to dial the
magnetic field down quite a bit, at which point you cannot reasonably talk
about a majority of fixed helicity photons hitting the electrons with any
regularity. At this stage most electrons reaching the detector in the beam
have random orientation. I cannot say this for sure because we cannot
measure spin orientation if we have no field! Sort of a Catch-22. But
this is really at the heart of quantum mechanics and the measurement
postulate. If you are not disturbing something you are not measuring
What was accomplished? 55
something.
Yeah, I know about the Elitzer–Vaidman Bomb Test and all that, but
it’s the same thing. You set up a spacetime manifold so the photon must
go via the “upper” path to ever, ever, get to detector 𝐷. (See en.wikipedia-
.org/wiki/Elitzur-Vaidman_bomb_tester.) Here you are still disturbing
something, the detector at 𝐷 and the optic fibres and mirrors in the upper
path. The weirdness of QM is not because of anything mystical here, it
is the plain measurement postulate in action, along with superposition.
Remember 50% of the tested bombs that are not duds do explode. If an
iterated qubit system is used to reduce the live bomb explode rate to near
zero you will then have problems with stray photons setting the bomb off
just randomly. It’s a nice system, but it does not scale well. Same for
most awesome things about QM, except Josephson junctions, transistors,
superfuids and black holes, they have nice quantum effects at large scale
(I guess you could say the cosmos as a whole too). Or, you know what I
mean: all things show QM at large scale, or anything with an atom in
it at least. But the effect of superposition and entanglement at scale is
what I am talking about.
To conclude: there is always interaction and the interactions cannot
do what the experiment is not set-up to do, and quantum mechanics is
only a little bit weird.
2.6 What was accomplished?
After all this effort we have not really done a lot of physics. Group the-
ory is a very blunt instrument. All we’ve discovered is that in quantum
mechanics fundamental particles can be usefully labelled by momentum
and spin. All that really says is that elementary particles move around
in space by translations and rotations in Lorentz invariant manner. Big
deal!
Well, it is a kind of big deal. We had no right to think group theory
would tell us two labels suffice to characterize the elementary particles.
It turns out we were right to be suspicious, we do have no right, because
there are color and flavour symmetries we have not taken into account.
The Poincaré symmetries are only the most obvious.
The fact we now have labels for doing relativity in quantum mechanics
(move particles around unitarily (QM) and preserving Lorentz invariance

(SR)) is considered a great “unification.”
2.6.1 The Standard Model
The current hints from the LHC are that there is physics “beyond the
Standard Model.” That said, in this chapter we have not even got up to
the Standard Model. But it is not hard to qualitatively describe how to get
to the Standard Model: we need to find other symmetries of fundamen-
tal particles, then repeat the same analysis — find the group, find the
unitary irreducible representations, and postulate “new particles” from
them. We’d find this is how we get the quarks and gluons, from flavour
and color4 symmetry. But what is the physical basis for those symme-
tries? There must be a physical basis otherwise we are doing magic, and
that is no good for me since I failed Wizardry 101 school.
The remarkable thing you “never get taught in school” is that the
quarks and gluons are not unexpected in Geon Theory. Of course, we had
to discover chromodynamics empirically, because the colors and flavours
are hidden in the composites, the protons and neutrons. We had to smash
them up to see.
Personally I do not think there was ever going to be any other way. The
experimentalists are the heroes, the theoreticians just get to tell stories,
and occasionally, if they are lucky, tell scifi stories that turn out to be
sci-fact, but that’s due to experimentalists finding the facts. However,
it is legitimate to tell fantasy stories of imagined alternative histories
where the theoretician predicted all the new fundamental particles in
advance, and the experimentalists were just confirming; firstly because,
as I just said, that’s rarely but sometimes how it happens for one or two
new particles, and secondly because theoreticians need encouragement.
On the encouragement side, how exactly were quarks and gluons not
unexpected in Geon theory? This has to do with the Gell-Mann matrices,
which comprise the algebra for chromodynamics. The entire Gell-Mann
algebra lives in the Clifford spacetime algebra 𝐶ℓ (1, 3), and so could have
been obtained from a geon theory, had anyone been using geometric al-
gebra instead of matrix algebra. Thus, so far, all known physics has
4I
use English spelling, but refuse to call chromodynamics “colour” symmetry, it is
nicer to think of color as a mere analogy to colour.
What was accomplished? 57
been found to live in spacetime, no extra dimensions or fibre bundles

tacked on to space have been necessary, though such decorative models
can be useful in aiding some computations. The point is, we do not need
those decorations. When you have localized topology in a quantum the-
ory, meaning that 4-geon cna zip around and be anywhere in space at
any time (up to what the amplitudes for superpositiosn allow) then the
most interest hack to describe this sort of gnarly spacetime is with fibre
unless attached to each point in space. These fibres are mathematical,
not physical. What is physical is the topological 4-geon.
Of course you do not have to buy T4G theory, and you are free to con-
sider fibre bundles as more fundamental or “real,” i.e., physical. I think
that gets you into severe trouble figuring out how the Measurement pos-
tulate 2 can sensibly work, and most gauge fibre bundle enthusiasts gloss
over this entirely. Gauge fibre theory starts to lack parsimony when com-
bined with basic QM. The average theoretician cannot juggle all physics
on their plate at the same time. Often Lorentz invariance is the focus,
sometimes the thing benignly ignored, other times the fact elementary
particles cannot be points is benignly ignored, no theoretician, to my
knowledge, juggles all considerations perfectly at once.
Suppose we do take T4G theory seriously? Then it is not hard to ex-
haust all the elementary particles one can get from geon theory, since
𝐶ℓ (1, 3), the STA, is an algebra with a finite number of unitary irre-
ducible representations, and tacking on more symmetry groups just to
get more representations is fundamentally an abomination for geon the-
ory! That said, geon theory only weakly predicts the dimension of space-
time, which comes from matching known particle phenomenology with
the synthesis of geon theory geometric algebra. So it does not destroy
geon theory if extra space dimensions are discovered, which will yield
higher symmetry groups. In that event T4G would still be a distinctly
different theory to strings, since strings conceive of strings and branes
as objects living in spacetime, whereas geon theory regards them all as
topological properties of spacetime. The spirit of these two approaches
is the same, but the theory comes out very differently, and future exper-
iments should be able to tell which theory is closer to mimicking nature.
If experimental physics ever exhausts and then overloads STA this
will not necessarily the death of geon theory. Composite particles and
higher modes (dynamics of geons) offer a similar freedom to string theory
strings and branes, and so the possibility of interesting ultra high energy
physics, beyond the Standard Model.
2.7 CPT Symmetry
Experiment rules the roost in physics, a fact everyone knows except the
more zealous theoreticians. Zeal and ardour in searching are however
not a bad thing. The key is balance in diversity, not in homogeneity. It
is good to have all sorts of types of people doing physics and cooperating
and sharing results, even the crackpots. How many times did a genius
physicist “steal” an idea from a crackpot and claim greatness, perhaps
even without thinking they were borrowing the idea? I do not know,
perhaps never, but the chance greatness can spring from insanity is not
something to always shy away from, it is something to temper with some
appropriate moderation, and let a hundred flowers bloom (there, you see,
even Mao Zedong was not all despot).
But experiment rules the roost.
So when parity violation was observed in kaon decay by Cronin and
Fitch, few thought they were crackpots with a mad theory because this
was experiment. Although experiments can go astray (cold fusion, eu-
genics and other madness) when you gt the experiments right no one
will accuse of of being a crackpot, at least not for the reason of your ex-
perimental results. If you test a novel unproven drug on yourself then its
you who is the crackpot, not your experiment.
Subsequently 𝑇 (time symmetry) was found to be violated, for example
in neutral meson systems. Since CPT symmetry is thought to hold true,
this means CP must also be violated in some systems.
This section discusses some basic aspects of CPT symmetries for stu-
dents who are unfamiliar, it is not meant to be a detailed review. My
main motive is to give younger students some good intuitive feel for how
physics works, so to speak. To motivate the student I present this little
puzzle.
How can physics depend upon whether we mirror invert? Surely this
would mean given a ‘blind’ trial look through a mirror into our
universe we could tell which side of the mirror we were on, and would
CPT Symmetry 59
this not violate a kind of Copernican Principle?

Also, what about the idea physics changes if we change coordinates
to a left-handed system instead of right-handed? Does that not
violate observer-independence? What are your thoughts on this?
Murray Gell-mann explains the resolution of both these types of ques-

tion in his interviews recorded on the Web of Stories— Life Stories of Re-
markable People series available on YouTube. He points out none of the
inversion symmetries (C, P or T) need to hold for a given system. Their
violation does not violate the general Copernican Principle (observer-
independence of laws of physics). A lot of physicists who had a good grasp
of the messy difference between idealizations and te real world under-
stood this, Gell-mann himself, also Feynman and Dirac.
As a teenager I confess I found CP and T violations exciting and
strange, but also hard to fathom. Why should Nature care about
whether we run time ‘forwards’ or ‘backwards’? Why would looking
through a mirror obviously appear to be the “wrong universe”?
Obviously plenty of things violate parity, our own bodies, for instance.
This is sort-of what Gell-mann was explaining: any particular system
can violate parity, but it is of no big consequence for philosophy. You
find a fundamental system with chirality and you say, “oh nice, it’s like
our bodies, cool.” At least that’s the attitude of someone like Gell-mann,
Dirac or Feynman. Gell-mann explained it by pointing out the parity
violation was a property of a particular Lagrangian or Hamiltonian, not
a violation of observer-independence. It’s the system that violates P or
T, not the whole universe.
Here is another way to appreciate the situation: if we mirror invert
our coordinate system then the left chiral particle become right chiral
particles, but they are still chiral, and they do not car about our coordi-
nate system. This answers the second naïve part of the above puzzles.
The coordinate invariance of physics holds absolutely. When a applying
a coordinate inversion you change how you describe chirality, not how the
particles behave.
2.7.1 Why I like CP violation
The 𝐶𝑃 symmetry violation (or equivalently 𝑇 violation) is the fascinating

one for me. I think it strongly hints that electric charge is a geometrical–
topological property of particles envisaged as 4-geons, not a mysterious
number tacked onto the ends of strings or branes. This could also apply
to color charges in strong interactions, though I believe we do not have
evidence for this yet.
I never really know if the simple arguments are valid (I keep reading
physics blogs and forums like physicsforum and stackexchange, where
reasonable sounding ideas get knocked down like dodos) but to my mind
𝑃 and 𝑇 are geometric symmetries, so when 𝐶 inversion is the only way to
balance two geometric symmetry violations I have to tell myself charge is
a geometric–topological quantity. Why then is no one working on looking
for topological indices as models for electric charge? Might it help to note
that the charge on the electron should really be +3 not −1? That would
just be a convention of course, but the idea behind it is that charge should
(naïvely derive from a discrete topological index, like a winding number.
Then the quarks have integer charges as well, everything does.
That’s all I have to say on CP and T, I do not actively work on this area,
so whatever else I add would probably be worse than misinformation.
* * *
The next chapter, I think, is as big an advance, if not greater. We

will take what we’ve done so far and then figure out that if elementary
interactions are localized events (so, generically, they are “scattering”
events) then we will find the only possible forces of nature are Yang–Mills
(spin=1) and gravity (spin=2).
Chapter 3
Scattering and Forces
In quantum theory all force arises from particle interaction, even gravity.
There are no fields. This means a nice way to analyse what forces are
described by our physics is to consider elementary interactions. If we find
there are interactions that are not allowed, then we will get constraints
on the possible types of forces.
Once we have the known forces, then later on we can invent fictional
fields to calculate dynamics without needing to think about the elemen-
tary interactions. The elementary particle–particle interactions are
however always primary ontology, at least in all quantum mechanics we
know, including string theory.
A quick aside: (and since I write for students) when you read the fu-
ture press, do not get sucked in too much by the hype when the LHC or
the next generation of colliders has “discovered a new force.” All the pos-
sible elementary (single boson exchange) forces of nature are described by
Yang–Mills theory.1 It is like MMT in macroeconomics. One framework
can describe all. What a “new force” means is journalistic laziness for
discovery of a new type of boson. That is going to be a huge momentous
event when it happens, but it is not really a new type of force. It will
still be a Yang–Mills force. There is going to be a lot to hype when this
1Iwrite this with false confidence. We may eventually discover entirely new physics
that cannot ever be explained by a Yang–Mills theory. All I am saying, really, is that
I personally think the chances of that are none to zero. But that’s just a subjective
personal opinion. You can take that and Bayesianize it and deduce anything you like,
depending upon how much weight you give my opinion.
61
62 Scattering and Forces
happens, but it is not necessarily a “revolution” in physics. It could be a

revolution in how we understand quantum particles, or their associated
fields. For instance, the discovery could rule out point particle models, or
maybe rule out string theory, or maybe at least rule out supersymmetry,
or confirm one variety of supersymmetry, it could confirm or dis-confirm
extra space dimensions, and a discovery of that magnitude in significance
would be amazing.
3.1 Maximum parsimony
A big conceptual hurdle here is gravity, since we already know of a very

good theory for gravity, general relativity, which while it has a quantum
postulate (existence of masses) this is rather shoddily tacked on to GR,
since there is no reason why mass has to be found in little packets. The
idea there are little quanta or particles of mass comes from combining
special relativity and quantum mechanics, as we showed in the previous
chapter. But this did not require general relativity.
Maybe this should not be thought of as a puzzle. But it is. The thing
is, GR does not need gravitons. Whereas SR+QM very naturally should
have gravitons (spin=2 quanta). GR thus seems a redundancy, like a high
level effective field theory approximating some microscopic dynamics.
This would be perfectly fine. . . except that,
(a) the graviton theory is mathematically unworkable, unless we admit

exotica like extra compactified space dimensions of string theory
plus unobserved supersymmetries,
and,
(b) GR works really brilliantly provided it is fed the massive particles

to work upon.
This is only a matter of philosophical taste for now, but I am going to

say String/M-theory is unpalatable, for the reasons given — we do not
see supersymmetry, and we have no evidence for extra dimensions. This
could change tomorrow, but I am working conservatively with what we
know today.
Maximum parsimony 63
I do not want extra space dimensions, ok. If you do, for some reason,
then fine, go off and study string theory. Please. If you like, think of this
as just a research effort to see how much we can get from 4 𝐷-spacetime.
Maximum parsimony.
Since the quantum theory of gravitons is unworkable, does this mean
gravity is not a quantum theory?
My answer is no! — and in fact GR is already a quantum theory be-
cause it admits wormhole topology on the Planck scale (at least). I mean,
how do you tell the cosmos to not create extremal wormholes? You can-
not. If one exists and it’s ends are minimal black holes, it cannot Hawking
evaporate, so it will be stable. Macroscopic wormholes are a no-go, but
Planck scale wormholes are fine. This provides ordinary old GR with a
definite quantum postulate: minimal wormholes exist. They are elemen-
tary particles of a sort.
3.1.1 Who needs gravitons?
With massive particles already around, thanks to SR+QM, we do not

need gravitons. Spacetime is already a given with SR (and with QM, in
case you were wondering) so we can account for all gravitational phenom-
ena with two possible exceptional circumstances, black holes and primor-
dial big bang physics, where gravity dominates over the other Yang–Mills
forces.
These “black physics” regimes are not well understood, but some
things are known. There is a known logical inconsistency with QM
and gravity involving Hawking evaporation: is black hole evaporation a
unitary process or non-unitary? This innocent question (which prima
facie seems to be “it must be unitary, or QM is dead,” but that naïve
answer gets into trouble with general covariance) has led to the whole
industry of holography and AdS/CFT concepts. This has really expanded
our understanding of gravity, and brought even plain GR very close into
alignment with QM. But with remaining puzzles.
It is not a mainstream physics resolution to ditch orthodox quantum
gravity. But it is my preferred resolution. Gravity is still a quantum
theory in my view, but not by virtue of being a theory of gravitons.
However, graviton modes can still exist, we just do not need them.
There is precedent for such weird “possible but unobserved” physics. The
Dirac theory of electrons permits magnetic monopoles, and yet we have
never observed them. String theory permits all manner of exotic parti-
cles, but none of them are observed. String theory requires supersym-
metry, which is not observed. A lot of possible physical states are totally
allowed to exist but do not apparently exist.
Now I am not saying gravitons are fictional. My view is that pack-
ets of gravity waves are gravitons, and gravitons are nothing more. So
in the extreme conditions of black physics I would expect particle-like
gravitons: highly localized energetic gravity waves, soliton solutions in
general relativity. I would not see how nature can avoid them.
This means in black physics we will have to grapple with the mathe-
matical difficulties of having gravitons in the path integrals or scatter-
ing calculations. But that is probably going to be fine, because cutting
edge theory, like the Amplituhedron program, suggest perturbation the-
ory with Feynman diagrams is not the way to do honest quantum field
theory. If that turns out to be the case, then renormalizable calculations
can probably be accomplished in the future with gravitons included.
I’d beg some empathy here. These ideas are radically conservative
and highly unifying. Consider that gravitons are massless (gravity is
long range) so are gravity waves.
Consider (indulge the imagination) that in an alternative reality space
was filled with some kind of ‘luminiferous aether’ or phlogiston, respon-
sible for light waves. Although that turned out not to be the case, if it
had then electromagnetism would be a lot like gravity, only with pho-
tons being actual physical waves rather than topological particles. We
would not then be talking about the photoelectric effect and photons.
But, . . . small soliton-like waves of phlogiston might arise, and they’d be
photons, spin=1 massless “particles.” If we pushed physics to extremes
we would probably say these ‘photons’ can be generated. They’d radiate
like crazy, like gravitons would, but they’d exist, and we’d describe them
quantum mechanically by spin=1, if they were small enough to be treated
in superposition and entangled states.
Because photons are localized packets of energy, there is no aether
model for them. But for gravitons there is an aether, and it is just space-
time. Since spacetime is smooth on the macroscopic scale we see grav-
ity waves, not gravitons, and we have no need for gravitons at all. But
we have not gone beyond the paradigm of quantum mechanics here, the
spin=2 “particle” is still a possible object that can be generated, precisely
by sufficiently localizing a gravity wave packet. How does one do that?
With great difficulty is the answer. Just as solitons in water are incred-
ible rarities, so would gravitational solitons. They are not stable, and
require carefully manufactured boundary conditions to get propagating.
Who is to say nature does not fashion such conditions all the time though?
I do not know. But I am pretty confident people who study gravity wave
solitons will tell you it is not easy to motivate them.
Because all the other fundamental particles in nature are stable and
appear soliton-like, they are probably not wave modes of any aether. They
are far more likely to be topological in character, not wave-like. But then
we need a decent theory to explain why the elementary particles behave
like waves in QM. To my knowledge, only geon theory has a reasonable
answer to this, which I have, and will, write about elsewhere. It is a long
topic. The short answer is that to get information-theoretic (epistemic)
probability amplitude waves, as in Schrödinger–de Broglie–Dirac theory,
you do not need a wave medium, you need a mechanism for a fundamental
geon to appear to be in more than one place at one time, and yet appear
in only one of those places whenever detected.
Although it might be sore on your eyes and ears to hear or read this,
the most conservative way I can think of getting such phenomenology is
with closed timelike curves. This is where geon theory begins. A beau-
tiful unifying thing here is that gravity furnishes these for us, they are
wormholes. They do not need to be large, they only need permit elemen-
tary particles to traverse them, and that’s entirely plausible given all
known theoretical physics. Remember monogamy of entanglement? Why
would nature restrict entanglement to pairs, for godssake? Can it be that
entanglement is formed by Einstein–Rosen bridges? That explains why
there are only two ends to an entangled state, and why exotic multipar-
tite entanglement can occur, as with GHZ states, but upon measurement
only pairs are found to be fully entangled.
3.1.2 Two views of gravitons
(For philosophy majors minoring in physics.)

This is where I try to mess you up with one analogy then another so
you cannot tell what is real.

To recap: I am not saying gravitons do not exist. What I am saying
is that we do not gain much by doing calculations with gravitons instead
of gravity waves and ordinary GR. Previously I gave a phlogiston anal-
ogy, but I think we can do better. Consider instead how light cusps form
from reflections, say in your tea cup. Those are pretty intense concentra-
tions of photons, remember your eyes only see a tiny fraction that scatter.
These cusps are created without photon lensing. Likewise gravity wave
cusps can form and lead to intense gravitational fields in highly localized
regions. Because masses gravitate, whenever two or more massive parti-
cles get close enough there is a good chance gravity wave cusps can form.
If they are intense enough they will liberate what looks like gravitons.
Now consider what I think is the proper analogy: Huygens wavelets.
We know we can approximate electromagnetic waves with wavelets. And
who is to say the wavelets are not the reality and the macroscopic waves
the fiction?
OK, but then we realize light is not a wave, it’s collective behaviour of
photons, small packets of energy, and that starts to sound an awful lot
like Huygens wavelets are much closer to reality. But it did not have to
be that way. Light could have turned out to be a classical Maxwell field.
It just didn’t in our universe.
I think of gravitons as Huygen wavelets for gravity waves. But un-
like the mathematical approximation of classical wavelets, if we can suf-
ficiently localize a graviton then it should become subject to quantum
effects. Then it will appear to behave much like a particle, detected in
clumps, and you can conceive of a graviton equivalent to the photoelec-
tric effect if you supposed there were weird metals with work functions
sensitive to graviton energies (there aren’t so this is pure gedankenex-
periment).
Now consider that the classical GR bending of light around a star can
in fact be calculated by graviton scattering approximation, at the tree
level in Feynman diagrams [21]. That’s the approximation where there
are no radiative corrections. And without a preponderance of gravity
wave cusps (which of course there are none in these cases), this is a totally
valid approximation. The calculation of the bending angle is exactly the
same as calculated with classical GR.
So who the hell knows what is more real, the warping of spacetime or
the graviton?
All I know is that most gravity effects are enormous in scale, and so it
does little good to analyse most of these phenomena in terms of graviton
interactions. But does expedience dictate ontology? No. However, when
it comes down to aesthetics, I have a definite preference. I see elemen-
tary particles as topology obstructions in otherwise smooth spacetime. I
cannot model a graviton in the same way, whenever I model a graviton in
T4G theory, it is just a gravity wave, of one kind or another, and if these
waves get sufficiently localized they can propagate along closed timelike
curves through wormholes, and then they’ll have to be treated quantum
mechanically.
Does that mean the spin=2 graviton is a particle now, not a gravity
wave? I think not. It is still a gravity wave, but it can get highly local-
ized in special circumstances, and in black physics. It is still in some
sense a classical warp in spacetime, and never is otherwise. When I do
graviton Feynman or Amplithedron computations, I am most definitely
thinking in terms of the Huygens wavelet approximation in these cases.
I do not for one second believe there are tiny packets like gravitons in-
volved. The usual (non-black physics) gravitational scattering physics is
highly macro, and the graviton is a fiction used to do calculations.
I do not want to theoretically yoyo too much here, but I think it is
important to draw a sharp distinction now with photons. You can say
soft photon scattering is similar to graviton scattering, in that you can
treat the photon as a particle and do Feynman diagrams, but you have
no hope of localizing these photons, they are going to be practically un-
detectable. However, I can imagine getting lucky and detecting a soft
photon, and when I do it will be highly localized, and will jiggle some
electron — that’s how it gets detected, ultra-sensitively somehow. We do
not have such sensitive detectors, but this gedankenexperiment assumes
we’ve got some such detector from some advanced alien civilization. In
T4G theory the same is not going to be possible with gravitons. This is
a sharp prediction of geon theory in my view. The graviton can get local-
ized, but then will be high energy. Most gravitons are not, and so will
never be detected as localized quanta, because they are not topological,
they are geometrically smooth.
There is one “out” for geon theory here: if we can find a model where
a spin=2 topological geon exists, then we have topological gravitons and

they will get a quantum theory like photons. My bet is that this will not
be the case.
Now I realize most physicists take the exact opposite view: they think
gravitons are real and warping of spacetime is the fiction. They believe
gravity waves do not exist, and that all we are detecting at LIGO is col-
lective graviton action. The problem here is the cognitive bias of fluid
mechanics.
Every young physicist grows up learning that water and gases are
really particle phenomena, but the most efficient way to model them is
with a continuum medium approximation, called the hydrodynamic ap-
proximation. This is how the Navier–Stokes and magnetohydrodynamics
equations are derived. But we can derive them also from the Boltzmann
equation for particles, which although unwieldy, is the more realistic
model.
Unfortunately for geon theory, this learning is well-ingrained, as it
should be! There are no hydrodynamic effects in nature, and to explain
aerofoil lift generation you should use particle theory, not Bernoulli or
any other hydrodynamics approximation. But the hydrodynamic theories
still work, if you take enough care with boundary conditions.
What I am saying is that for gravity the reality is the reverse to fluid
dynamics: I think spacetime is real, and the graviton quantum theory is
fictional (except for black physics regimes), which is the exact reverse to
the case of fluid mechanics.
It is only a hunch, but my hunch is that gravitons are not topolog-
ical, they are smooth geometry, like gravity waves, and so they do not
easily localize like quanta. It is worth pointing out that this has some
pretty exact justification. In graviton scattering theory, if you use the
Einstein–Hilbert action and expand around flat space, then to maintain
diffeomorphism invariance (the symmetry of GR) you find you need to
include off-shell terms in the Lagrangian. There turn out to be infinitely
many of these. However, when computing the tree-level diagrams all
these terms cancel! So for on-shell 𝑆-matrix scattering the terms that
maintain diffeomorphism invariance are irrelevant. But for all practi-
cal purposes, in near flat space graviton scattering is not a terrifically
complicated process. You might get just one single graviton interaction
at closest approach if you are a massless particle getting bent around
a star. But think about this for just a second. It is absurd. Any parti-
cle (even the massless) going near a star gets continually and smoothly
bent around the geodesic. There is no single graviton doing this. Yet the
correct angle deviation is obtained from the single graviton scattering
formula. What is going on here?
To my mind this is a strong hint that the graviton theory is a gross
quantum idealization, it is a pure 𝑆-matrix theory object. The 𝑆-matrix
recall, only cares about asymptotic states, so the incoming and outgoing
geodesics are those of flat space, no star, and you get the angle devia-
tion from just a single graviton. The path integral pretends to sum over
infinitely many interactions, but all that is not really physical in some
sense — in the sense you do not need them.
It is like gross overkill.
The imagery that springs to mind is that of Ayrton Senna driving the
particle. Senna was famous for pumping the accelerator wildly when
turning corners, in the old F1 cars this somehow gave him more corner-
ing control and made him fast. Fernando Alonso does something similar
but with steering. My imagination thinks of all the off-shell interactions
as wobbling the particle around as it goes by the star, with all the in-
finitely many wobbles doing nothing at the end of the day except pray to
the mathematical symmetry gods and give alms to the lord of symmetry.
How the hell do all these off-shell interactions know to conspire to cancel
each other out? Clearly it is a physical fiction that has mathematics that
just works out, because of the demands of diffeomorphism invariance you
write into your Lagrangian.
I think it is pretty obvious that a lot more happens as the particle
goes by the star than a single graviton exchange. I think that graviton
is a fiction. It is being exploited to tell a story that avoids all mention of
Einstein curvature. You can comb a cat in more than one way. You can
have gravity in more than one way.
I do not think it is wrong to do the scattering calculation with the on-
tree graviton. It has to work out, because it has to be possible to replace
the curved spacetime near the star with a fictional standing gravity wave
or such-like, as long as what the particle sees is similar geodesics the
angle of deviation will be the same. And it has to be possible to replace
a macroscopic gravity wave disturbance with a spin=2 gravity-wave type
of soliton that hits the particle just once. Same 𝑆-matrix asymptotics.
Totally different business going on in-between. But,. . . you know . . . which

picture seems more realistic to you?
And as I said before, highly localized gravity waves that inherit quan-
tum properties from Planck scale dynamics could exist, but they are of no
consequence for these processes, we have to go to the extreme situations
of black physics for the graviton to be of interest.
I know quantum mechanics is “weird,” and I know you can always tell
a story of how we only measure the particle’s angular deviation once, we
never see a smooth curved path. So why the hell not suck-it-up and believe
the particle only gets hit by one very important graviton, and all the rest
are there just for the show? Ya know, . . . I can suck up a lot of things, but
this is not one of them. I can tell a better more beautiful story using geon
theory, and I do not care that Sabine Hossenfelder says beauty is not a
reliable guide. My idea of beauty in science is parsimony, not Helen of
Troy. I place my bets on geon theory, and I do not care if I am wrong.
The Einstein picture of gravity has to (as in: I need to win this bet
that I do not care about) reign supreme here.
* * *
End of story? No, not end of story, but the end of my tale of gravitons.
Scale invariance
I just lied, I do have three more things to say about gravity and gravitons.
One slightly fanciful, let’s be generous and say whimsical, way of
thinking about gravitons, which I do not think is totally wrong, is that
there is a graviton, there is exactly one graviton, and it has a radius of
about 15 billion light years. All sorts of interesting stuff happen internal
to this one graviton, which is called “life.”
A second thing concerns scale invariance. It does not surprise me (only
in retrospect of course) that the AdS/CFT correspondence works. The
CFT is a theory invariant to scale. When you do not care about scale then
the hydrodynamics approximation works well: one graviton scattering
becomes indistinguishable from warping of spacetime. The conformal
invariance says you cannot possibly tell the difference. The fact our world
does not have perfect conformal invariance is why we can say meaningful
things about the difference between gravity waves and gravitons.
A third thing harks back to my earlier comments on the absurdity
of thinking gravity lensing by stars is a single graviton scattering (for
massless particles). For a massive particle it does take more than one
graviton interaction to get the angular deviation, but the basic idea is
the same. The problem is that all the infinite other interactions have to
cancel out precisely to get the correct agreement with GR. QFT is a way
to get precise cancellation because it says you have to add up the ampli-
tudes in a precise way, and you must impose all the correct symmetries,
if you order the sums differently in the perturbative approximation all
hell breaks loose and you can get an unfathomable mess. It is only when
either,
(a) you do the exact path integrals non-perturbatively, or
(b) use the Feynman rules to do the perturbative approximation cor-

rectly ordered,
that the miracle cancellations occur. If you are a religious believer in QFT
then this is fine. For the believer, the miracles do occur because nature
does take all possible paths (or acts like it knows them all).
For my taste, this is a massive loss of parsimony. The GR spacetime
warping model suffices, and I think is more realistic in some sense. This
is the sense in which there really are no graviton quanta, and what we
have is the warping of spacetime, which can be thought of in the abstract
as one gigantic graviton the size of a solar system. The effect of space-
time curvature is thus a lot like a Yang–Mills interaction, but in a to-
tally different regime, a full hydrodynamics regime. Think of it this way:
what happens to your fluid water viscosity and compressibility when the
molecule of water is a metre in diameter and your boat or aeroplane is
about the same length? You do not have hydrodynamics, you’ve got bil-
liards. Clearly, at least to my mind, gravity lensing and planetary orbits,
and ourselves sitting on the Earth, are not the billiards regime. For us,
there are no gravitons. There is only Ricci curvature and gravity waves
(Weyl curvature).
* * *
That’s all I have to say for now on motivations. I thought it necessary

because it can confuse students who think spin=2 theory means gravi-
tons must exist and Einstein was wrong. I am saying not so fast! Einstein
was right, and gravitons can exist. We begin on the bulk of this chapter
now with basic scattering theory.
Geometric algebra preliminaries 73
3.2 Geometric algebra preliminaries
I will try to introduce the basics of Amplituhedron theory purely in ge-

ometric algebra (in G 𝐴), all multivectors, no matrices. Should be a fun
challenge. If you gt stuck and prefer to retreat to the comfort of ugly ma-
trices then Elvang & Huang give a decent review [22]. However, I am not
going to be covering Amplituhedron theory. I only need the spinor he-
licity formalism and some basic scattering theory for our purposes. Our
purpose is to understand the Weinberg-Witten proposition: there are no
fundamental forces other than Yang–Mills and gravity. We do not need
to compute actual scattering amplitudes for this.
I encourage you to take the effort to compare my formalism (which
probably does not exist anywhere else) with that of say Elvang & Huang.
It is good to know how to do things in at least two different languages.
I use a (− + ++) metric. Because in my language 4-vectors are grade-
4 spacetime multivectors I use the words ‘spacetime vector’ for position
and momenta, they are not 4-vectors, they are spacetime 1-vectors, so
they have four components:
𝑝𝜇 = ( 𝑝0 , 𝑝 ) = ( 𝑝0 , 𝑝𝑖 ) , 𝑝𝜇 𝑝𝜇 ≡ 𝑝2 = −𝑚2 .
We do not need matrices. The Pauli and Dirac matrices are better
thought of as 3-space and 4-spacetime basis vectors respectively. For the
Pauli algebra,
𝜎𝑖2 = 1, 𝜎𝑖 𝜎 𝑗 = 𝜎𝑖 · 𝜎 𝑗 + 𝜎𝑖 ∧𝜎 𝑗 ,
𝝈 𝑖 · 𝝈 𝑗 = 0, 𝜎𝑖 ∧𝜎 𝑗 = −𝜎 𝑗 ∧𝜎𝑖 , 𝐼 ≡ 𝝈1 𝝈2 𝝈3 .
For the Dirac algebra 𝑖, 𝑗, 𝑘 ∈ {1, 2, 3}, while 𝜇 , 𝜈 ∈ {0, 1, 2, 3}.
𝛾𝑖2 = 1, 𝛾02 = −1, 𝛾𝜇 𝛾𝜈 = 𝛾𝜇 · 𝛾𝜈 + 𝛾𝜇 ∧𝛾𝜈 ,

𝛾𝜇 · 𝛾𝜈 = 0, 𝛾𝜇 ∧𝛾𝜇 = −𝛾𝜈 ∧𝛾𝜈 , 𝑖 ≡ 𝛾0 𝛾1 𝛾2 𝛾3 .
The symbols 𝐼 and 𝑖 here are respectively the space and the spacetime
pseudoscalar, for Minkowski geometry. Both have properties of a unit
imaginary except that 𝐼 commutes with both vectors and with bivectors,
while the spacetime pseudoscalar 𝑖 anticommutes with vectors but com-
mutes with bivectors. The reason being that when there is a term 𝑒𝑘 𝑒𝑘
in basis vectors upon a permutation, then we do not need an extra per-
mutation to swap these two, so to move the psuedoscalar across a single
base vector only takes ( 𝑛 −1) permutatons. We illustrate here for a vector
𝑎 = 𝑎2 𝝈 2 :
𝑎 𝐼 = ( 𝑎2 𝝈2 )(𝝈1 𝝈2 𝝈3 )
= 𝝈2 (𝝈1 𝝈2 𝝈3 ) 𝑎2 , (here, the scalar 𝑎2
commutes with everything)
= −𝝈1 𝝈2 𝝈2 𝝈3 𝑎2
= +𝝈1 𝝈2 𝝈3 𝝈2 𝑎2
= (𝝈1 𝝈2 𝝈3 )𝝈2 𝑎2
𝑎
= 𝐼𝑎
Now try this again in spacetime algebra (STA), again for 𝑎 = 𝑎2𝛾2 :
𝑎𝑖 = ( 𝑎2𝛾2 )(𝛾0𝛾1𝛾2𝛾3 )
= −𝛾0𝛾2𝛾1𝛾2𝛾3 𝑎2
= +𝛾0𝛾1𝛾2𝛾2𝛾3 𝑎2
= −𝛾0𝛾1𝛾2𝛾3𝛾2 𝑎2
= −(𝛾0𝛾1𝛾2𝛾3 ) (𝛾2 𝑎2 )
= −𝑖𝑎
In any dimension the pseudoscalar will commute with even grade vec-
tors such as bivectors. A complex algebra structure, which requires a
commuting root of −1, is the even subalgebra of 𝐶ℓ (3).
By admitting vectors of all grades up to the vector space 𝑉 dimension
𝑛, for 𝑉 = ℝ3 the graded algebra is 𝐶ℓ (3), which has one scalar, 1, three
vectors, three bivectors and a pseudoscalar, so is a 𝑑 = 8 dimensional al-
gebra, and the Pauli algebra is just the even subalgebra of 𝐶ℓ (3), denoted
𝐶ℓ + (3). The Dirac algebra is the algebra of the unit scalar combined with
the vectors {1, 𝛾𝜇 }.
Formally, the Pauli algebra is the closed algebra generated by,
{1, 𝝈 𝑖 𝐼 }
Because 𝐼 is a trivector in 𝐶ℓ (3), the geometric product 𝝈 𝑖 𝐼 is a bivector.

The bijective correspondence with the Pauli spinors is,
0
𝑎 + 𝑖𝑎3
−→ 𝑎0 + 𝑎𝑘 𝐼 𝝈 𝑘
− 𝑎2 + 𝑖𝑎1
where on√the left we have the 2-component Pauli spinors and the imagi-
nary 𝑖 = −1, while on the right we just have plain basis vectors and the
pseudoscalar 𝐼 = 𝝈1 𝝈2 𝝈3 and no need for imaginary numbers. You can
check for yourself as a homework exercise that with the unit scalar re-
placing the identity matrix, the above map is indeed a bijection with each
Pauli matrix 𝜎ˆ 𝑖 replaced by the corresponding Euclidean base vector 𝝈 𝑖 .
In case you forgot, the Pauli operators can be defined by the relations,
𝜎
ˆ 𝑖𝜎
ˆ 𝑗 = 𝛿𝑖 𝑗 + 𝑖𝜖 𝑖 𝑗𝑘 𝜎
ˆ 𝑘.
which makes it easy to check the bijective correspondence.

No matter what the grade of a multivector, 𝐴 or 𝐵, there are two other
useful products we can define from the geometric product. One is the
scalar product ∗, which is just the scalar part of the geometric product,
𝐴 ∗ 𝐵 ≡ h 𝐴𝐵i0
inside the brackets is the full geometric product of 𝐴 and 𝐵, which could
have terms of all 𝑛-grades in the full algebra, the bracket with subscript
𝑘, so h· · · i𝑘 , by definition projects out the 𝑘-grade part only, and the scalar
part is the 𝑘 = 0 grade.
The other useful product operator is the commutator product ×, de-
fined by
1
𝐴 × 𝐵 = ( 𝐴𝐵 − 𝐵𝐴) .
2
again, inside the parentheses are two full geometric products. We will
use the commutator product in our Lie algebra technology.
Bivectors in particular are the only geometric object we need to define
representations for all Lie groups. A bivecotr, say 𝐵 = 𝑎 ∧ 𝑏, where 𝑎 and
𝑏 are 1-vectors, is an oriented area element. It is shapeless, so you can
think of it as a parallelogram if you like, the parallelogram swept out
by moving 𝑏 along the vector 𝑎. But there is no need to think 𝑎 ∧ 𝑏 has
any particular shape, this is powerful, since it means infinitesimal bivec-
tors can represent arbitrary shape area elements in a geometric calculus.
Determinants will handle the area transforms.
Likewise, a trivector in a 𝑛 ≥ 3 dimensional vector space, say 𝑎∧𝑏∧𝑐, is
an oriented volume element, so a parallelepiped, but does not have to be
thought of as having a parallelipiped shape, though again one can sweep
the parallelogram defined by first two along the third vector to get an
oriented parallelipiped. Again, determinants will handle volume forms

in geometric calculus. So we have no need for exterior algebra.
All of Maxwell’s Equations can be written succinctly in geometric al-
gebra multivectors, as,
2𝐹 = 𝐽
where 𝐹 is a Faraday bivector and 𝐽 the 4-current spacetime vector (a
four vector in dumb textbooks). The geometric operator 2 = ( 𝜕0 , ∇) is a
full geometric product, and can be decomposed into divergence (vector
part) and curl (pseudovector part) equations,
2·𝐹 = 𝐽
2∧ 𝐹 = 0
and to contact orthodox textbooks,
𝐹 = 𝐸 + 𝑖𝐵
𝐵,
𝑖 ≡ 𝛾0 𝛾1 𝛾2 𝛾3 , is the spacetime pseudoscalar.
If you are interested in Maxwell theory see Hestenes’ Spacetime Algebra

book [23] (a rather beautiful work). I am not interested, so will leave the
motivation for classical mechanics with G 𝐴 at that.
There are a few further useful products. Note that for 1-vectors, 𝑎,
and 𝑏 say, (so ordinary vectors, which in spacetime are ‘four-vectors’ in
other textbooks, confusingly, but remember that I am calling these 1-
vectors), the inner product (·) and outer product (∧) operators can always
be derived from the full geometric product,
1
𝑎·𝑏= ( 𝑎𝑏 − 𝑏𝑎)
2
1
𝑎 ∧ 𝑏 = ( 𝑎𝑏 + 𝑏𝑎)
2
But we can also define the inner product of a vector and any multivector
as the grade lowered part of the geometric product,
𝑎 · 𝐴𝑟 = h 𝑎𝐴𝑟 i | 𝑟−1|
where 𝐴𝑟 is a pure 𝑟 -grade multivector (a vector with terms of only grade

𝑟 ). by linearity we then can compute the inner product of a vector with
any multivector even one of mixed grades. Similarly, the wedge or outer
product is the grade raised term,
𝑎 ∧ 𝐴𝑟 = h 𝑎𝐴𝑟 i | 𝑟+1|
Without too much mental effort this leads naturally to the inner and
outer products of any two pure grade multivectors,
𝐴𝑟 · 𝐵 𝑠 ≡ h 𝐴𝐵i | 𝑟−𝑠 |
𝐴𝑟 ∧ 𝐵 𝑠 ≡ h 𝐴𝐵i | 𝑟+𝑠 |
We have now pretty much covered all the different types of product opera-
tors in a full graded geometric algebra. This will give us enormous power
of dopes who still use tensors. Although in my books I never use all this
power, so you can wonder why I bother. I just think geometric algebra is
super cool. I’m that that guy who’d drive a Bugatti Chiron around just
to go grocery shopping and drop the kids at playschool (like, I wish).
If you ever need a dictionary translation back to matrix elements, the
matrix elements 𝑎𝑖 𝑗 , of a matrix acting on vectors 𝑎® always refer to a
chosen basis, {®𝑒𝑖 },
𝑎𝑖𝑘 = 𝑒®𝑖 · 𝑎
®𝑘 .
here each 𝑎® 𝑘 is an entire vector, the index is a label not a scalar compo-
nent. The traditional matrix of this matrix is the array where all the 𝑎® 𝑘
are the columns. Determinants are then,
det( 𝑎𝑖𝑘 ) = ( 𝑒 𝑛 ∧ . . . ∧ 𝑒 1 ) · ( 𝑎 1 ∧ . . . ∧ 𝑎 𝑛 )
= 𝐼˜ · ( 𝑎 1 ∧ . . . ∧ 𝑎 𝑛 ) .
where 𝐼˜ is the reverse of the pseudoscalar 𝐼 . I will try to not use matri-
ces, we do not need them. Occasionally to show the correspondence with
orthodox textbooks I may write one or two of them.
All multivectors in a geometric algebra 𝐶ℓ ( 𝑛) are our linear operators.
Consider any linear map, an endomorphism 𝑓 , on the vector space,
𝑓 : 𝑎 ∈ ℝ𝑛 −→ 𝑎 0 = 𝑓 ( 𝑎 ) ∈ ℝ𝑛 .
This induces an outermorphism denoted 𝑓 on the entire graded geometric

algebra via,
𝑓 (𝑎1 ∧ . . . ∧ 𝑎 𝑘 ) = 𝑓 (𝑎1) ∧ . . . ∧ 𝑓 (𝑎 𝑘 ) , 𝑘 ≤ 𝑛.
In matrix language we use traces and determinants a lot, when our linear
operators are G 𝐴 multivectors like 𝑓 , we still have traces and determi-
nants, and they are very easy to compute,
det 𝑓 = 𝑓 ( 𝐼 ) 𝐼˜
or, 𝑓 ( 𝐼 ) = det( 𝑓 ) 𝐼 . Here 𝐼 is the pseudoscalar, and 𝐼˜ is its reverse, the

reverse of a multivector is defined by reversing the order of all vectors in
products,
𝐼˜ = ( 𝑒0 𝑒1 . . . 𝑒𝑛 ) ˜ ≡ 𝑒𝑛 𝑒𝑛−1 . . . 𝑒0
it is easy to check that 𝐼˜ = ± 𝐼 −1 depending on the metric signature. 𝐼˜ =
− 𝐼 −1 for vector spaces with an odd number of negatives in the metric
signature, so that is the case for spacetime. The familiar product rule for
determinants holds, and so forth.
The generalized trace is,
tr 𝑓 = 𝜕𝑎 · 𝑓 ( 𝑎) .
To understand this we need the definition of the vector derivative, the

“derivative w.r.t. a vector”; we can define this using the D’Alembertian
2,
2 = 𝛾 𝜇 𝜕𝑎𝜇 , where 𝑎𝜇 = 𝛾 𝜇 · 𝑎
so in 3 𝐷 we get 𝜕𝑥 ≡ ∇, which has the usual derived properties of div,
grad and curl. So 𝜕𝑎 just means, “differentiate w.r.t. each component of
the vector separately, then collect them all in a vector (in the appropriate
basis ordering).”
Adjoints are also easy to define, though without referring to bra and
ket vectors, so it takes a little getting used to adjoints in G 𝐴. I am still
not used to it. The adjoint is denoted 𝑓 , and is defined by,
𝑏 · 𝑓 ( 𝑎) = 𝑓 ( 𝑎) · 𝑏.
Pretty simple huh, so don’t you forget it! If you recall that the Dirac
bra-ket was just an inner product all along this should be easy for you
to digest. In QM terminology the adjoint, remember, is the Hermitian
transpose of the operator. It only takes one line, but it is tricky to show
the familiar adjoint product still holds for us,
𝐹𝐺 ( 𝑎) = 𝐺 𝐹 ( 𝑎)
Using these adjoints there is a nice computationally efficient expression

for the inverses,
𝐹 −1 ( 𝐴) = 𝐼𝐹 ( 𝐼 −1 𝐴)(det 𝐹 ) −1
−1
𝐹 ( 𝐴) = 𝐼𝐹 ( 𝐼 −1 𝐴) (det 𝐹 ) −1 .
This is so slick, you should try recalling how long it took to find a for-
mula for a matrix inverse. These multivectors, the ones with non-zero
determinants, are isomorphic to matrices, so we’ve just done a whole lot
of traditional matrix algebra technology in a single page.
Our linear maps also have eigenvectors of course,
𝐹 ( 𝑒) = 𝜆 𝑒.
We cannot however have complex values eigenvalues 𝜆, since our G 𝐴 is

defined only over the reals. We thus have to extend this technology to
eigenblades;
𝐹 ( 𝐴𝑟 ) = 𝜆 𝐴𝑟 .
The geometric structure obscured by complex roots to the polynomial
equations,
det( 𝐹 − 𝜆 𝐼 ) = 0
are now taken care of in the graded structure of our G 𝐴. No more complex
eigenvalues. Each eigenblade determines an invariant subspace of the
linear transformation 𝐹 . Ahhh, so beautiful. The geometry of the eigen-
stuff is now transparently geometric. It is almost self-explanatory. But
not quite for the uninitiated.
You will probably be more comfortable computing eigenvalues the ma-
trix way, solve that polynomial root equation, and in general get some
complex roots. Suppose you do. For example,
˜ − 𝑒1 𝑒2 𝜃/2
𝑓 ( 𝑎) = 𝑅𝑎𝑅, 𝑅=𝑒
then
𝑓 ( 𝑒1 ) = cos 𝜃 𝑒1 + sin 𝜃 𝑒2 , 𝑓 ( 𝑒2 ) = − sin 𝜃 𝑒1 + cos 𝜃 𝑒2
as you’d expect, since this is a rotation by the rotor 𝑅. Using the endomor-
phism extension of 𝑓 to act on the bivector 𝑒1 𝑒2 , we find an eigenblade,
specifically and eigen-2-blade or just “eigenbivector”;
𝑓 ( 𝑒1 ∧ 𝑒2 ) = (cos 𝜃 𝑒1 + sin 𝜃 𝑒2 ) ∧ (− sin 𝜃 𝑒1 + cos 𝜃 𝑒2 )

= 𝑒1 ∧ 𝑒2 .
Or, for shorthand, 𝑓 ( 𝑒1 ∧ 𝑒2 ) = 𝑒1 ∧ 𝑒2 . So the rotation has an eigenbivector
with eigenvalue =1. We see this, and say: rotation in a plane leaves the
plane unchanged ( 𝑒1 ∧ 𝑒2 ), but objects living in the plane ( 𝑒1 itself say, or
𝑒2 ) get rotated. How come 𝑒1 and 𝑒2 individually change but 𝑒1 ∧ 𝑒2 does
not? It is because in G 𝐴 the bivectors are primitive oriented elements of
area, they have no residual shape or vector aspect. In G 𝐴 vector parts are
described by vector parts, surprise, surprise. Although bivectors can be
described by parallelograms, you could equally describe a bivector just
as well by an oriented disk. Rotate the disk in its own plane leaves it
unaltered.
Now what if we have a different 𝐹 , with a “complex” eigenvalue from
the characteristic polynomial roots, say,
𝑓 ( 𝑢 + 𝑖𝑣) = (𝛼 + 𝑖𝛽)( 𝑢 + 𝑖𝑣)
This is the general case because complex eigenvalues are generally as-
sociated with complex eigenvectors, the eigenvector might have all real
components (so 𝑣 = 0)® in a particular 𝑓 instance, but not in general. Now
to see how to shift to a geometric interpretation, first decompose into real
and imaginary parts,
𝑓 ( 𝑢) = 𝛼 𝑢 − 𝛽 𝑣, 𝑓 ( 𝑣) = 𝛽 𝑢 + 𝛼 𝑣
then find your eigen-bivector using the outermorphism,
𝑓 ( 𝑢 ∧ 𝑣) = (𝛼 𝑢 − 𝛽 𝑣) ∧ (𝛽 𝑢 + 𝛼 𝑣)
= (𝛼2 + 𝛽 2 ) ( 𝑢 ∧ 𝑣)
We thus see the eigen-bivector picks up the magnitude of the otherwise
complex eigenvalue. The phase of the complex eigenvalue specifies the
rotation of object in the 𝑢 ∧ 𝑣 plane. Overall we’ve decomposed the action
of 𝑓 into a dilation combined with a rotation. Well, that’s what your tra-
ditional linear algebra course taught you, only in 3 or so pages instead of
1 page.
Thus we’ve solved the problem of geometrically interpreting complex
eignvectors and complex eigenvalues. In our continuing theme, we again
find we do not need the complex numbers, except perhaps as an interme-
diate crutch.
TODO: Some more prep was needed here I think?
Spinor Helicity variables 81
3.3 Spinor Helicity variables
For this section I am going to use my Huygens wavelet-like approxima-

tion for gravitons and treat them as spin=2 quanta, even though I suspect
this is a fiction. It does not invalidate any of the scattering physics, since
scattering off warped spacetime curvature can be approximated by suf-
ficient ‘fictional’ graviton interactions. Nature may know the difference,
but the mathematics does not.
There are now several decent reviews available for Amplituhedron
methods. If you know a little QFT and want to get a thorough review
of scattering then the book by Elvang and Huang is good, and there is a
preprint available for free [22]. I also found the shorter introduction for
astrophysicts useful [21]. My treatment will be more verbose and gentle,
and will only be interested in the Weinberg–Witten theorems on the pos-
sible forces of nature. I have no interest whatsoever in computing any
scattering amplitudes (I used to, for nuclear medical physics teaching for
postgrads, but that was decades ago).
Recall we had a (−, +++) metric. We use the word ‘spacetime vector’ for
position and momenta (not ‘4-vector’ which do not exist in ℝ3 and in ℝ1,3
are multiples of the psuedoscalar 𝑖 = 𝛾0𝛾1𝛾2𝛾3 ). The physical momenta
are,
𝑝𝜇 = ( 𝑝0 , 𝑝 ) = ( 𝑝0 , 𝑝𝑖 ) = ( 𝐸, 𝑝𝑖 ) , 𝑝𝜇 𝑝𝜇 ≡ 𝑝2 = −𝑚2 .
Chapter 4
Quantum Kinematics
The style for this chapter is to assume the reader has a background in
quantum mechanics and just wants to practise some problem solving or
grounded theory exploration. So I present minimal theory refreshers
when needed as Hints to selected problems. For GNU Free Document
contributors, each problem should be introduced with some motivation,
not just as a mathematical curiosity.
However, there is a running theme through all these notes in the orig-
inal version, which is the quest for a geometric understanding of quan-
tum mechanics through; (1) geometric algebra representations, and (b)
ER=EPR duality (wormholes are entanglement) which in the author’s
view provides us with a unification of QM with GR which is the reverse
of the orthodoxy, that is, through the geometric lens we will find general
relativity implies quantum mechanics and gravity is “already a quantum
theory.”
4.1 The wave equations

𝜕 𝜕 𝑖ℏ
𝑖ℏ 𝜓 = 𝐻𝜓 , or 𝑖ℏ 𝜓=− ∇𝜓 ,
𝜕𝑡 𝜕𝑡 2𝑚
or
( + 𝑚2 )𝜙 = 0
What do they mean? Where do they come from? They are just expres-
sions of energy conservation. The wave solutions for free particles mean
83
84 Quantum Kinematics
if you get a wave at any time, after a little time you will still have a wave.
Energy conservation. The wave in a free-field does not have anything
to interact with, so it goes nowhere except where it was first heading
for. This is why the famous “wave equations” can be derived from simply
writing down the law for energy conservation. You write down a Hamil-
tonian, then set it to the differential time evolution operator as Noether
told us to do.
What’s that you say? Who is this Noether? Noether’s results fol-
low from Lagrangian mechanics. The reason Noether’s principles hold
in quantum theory is because they are deep. You only need a least action
principle and some symmetries. Well, quantum theory has that.
The only real mystery here is twofold: (1) why waves? (2) Why the 𝑖?
Brief answers are:
(1) Particles behave like waves, so a wave equation is a first approxi-

mation. It is a probability amplitude wave though, so it is carrying
information, not a particle. Quantum theory asserts all this: we
need the waves because in quantum theory there is (a) superposi-
tion, and (b) entanglement. Geon theory (potentially) explains why
we have quantum theory.
(2) The mystery 𝑖 is concealing the fact that wave equation is for parti-
cles that have an orientation, so either a spin, or a preferred direc-
tion —like they could be extended objects, but extended objects can
rotate, so this all amounts to about the same thing; we need a way
to know how to rotate our measurement frame onto the orientation
frame of the elementary particle.
These two answers are naïvely in severe conflict. A particle, even if ex-
tended, like a string or brane or wormhole, can certainly rotate, so needs
some geometric factor like a pseudoscalar 𝑖 or a bivector 𝑖𝑒𝑘 , in it’s wave
equation. But fields do not rotate on the spot. So how can we use a wave?
The answer is given in the answer: the wave is not a real wave, it is de-
scribing propagation of information. Real information, but nonetheless
it is a theory of information. Why is QM a theory of information? Well,
this has nothing to do with weird It from Bit or Copenhagen nonsense,
in my opinion, quantum theory is about information because all physics
is about information, it’s just that quantum mechanics is a theory that
Probability and the Phase of the Wavefunction 85
describes a certain approach to physical information, and that can be

summarized succinctly by the statement:
Quantum mechanics is a theory of measurement.1
That’s the explanation in a nutshell.

Why the information in QM travels like a wave, not like a sharp ray
as in classical physics, is the guts of quantum kinematics. Why non-local
effects exist in an otherwise purely local theory is the other part of the
guts, maybe the heart. In my lectures I use geon theory to try to explain
all of this.
It is just one perspective though. There are many other interpreta-
tions of quantum mechanics. They tend to make the same predictions,
whereas I like geon theory because it makes different predictions in the
extreme scenarios (gravitational singularities and low energy macro-
scopic physics). Geon theory is not just an interpretation, because it is a
way to derive quantum mechanics using general relativity.
4.2 Probability and the Phase of the Wave-

function
I have always wanted to get deep into what the heck complex numbers
are doing in quantum mechanics as probability amplitudes? The thing
is, quantum mechanics looks superficially a lot like ordinary classical
statistical mechanics, but the role of superposition of the complex valued
probability amplitudes make all the difference in the world. People mar-
vel at the strange consequences and invent outlandish ideas like Many
Worlds Theory to try to explain these things, but I prefer to not worry
too much about the metaphysical consequences and worry instead about
why the amplitudes are so fundamental in the first place. So far I have
not been able to convince myself that Many Worlds Theory is the reason.
1 Classical
mechanics is not a theory of measurement. You can perform measure-
ments in CM, but CM is about real trajectories in phase space independent of any mea-
surement.
4.2.1 Complex Numbers are a Distraction
With all due respect to Roger Penrose and others, the role of complex
numbers in QM is vastly over-stated. The complex number field ℂ ap-
pears naturally in the real spacetime algebra of Clifford algebra. Since
fundamental particles are assumed to move around in spacetime there
should then be no mystery why complex numbers might appear. The field
ℂ is a subalgebra (the even subalgebra) of the Clifford algebra 𝐶ℓ 3 , and as
the algebra of rotors (generators of rotations) in the relativistic spacetime
algebra 𝐶ℓ 1, 3.
The rotors, of course, include the Lorentz rotations (boosts).
So why are quantum particles best represented by complex valued
wave-functions?
One answer has to be that we have historically just chosen a bad rep-
resentation. The wave-function is really a Clifford algebra rotor. The
algebraic dimension of the rotors is 2D, and the algebra is isomorphic
to the complex number algebra. In the Clifford (geometric) algebra 𝐶ℓ 3
the rotors are combinations of the scalars and bivectors, which square to
unity,
𝑅 𝑅˜ = 1.
From now on in my lectures I will refer to all Clifford algebras by the

more accurate and useful term Geometric Algebra or G 𝐴 for short.
The Schrödinger wave-function is really a field with locally the plane
of a rotor. The complex plane at each point in spacetime that the wave-
function “lives on” is really just a plane of rotation of a rotor or spinor. In
the Dirac theory it’s a spinor of course.
What all textbooks on quantum mechanics gloss over, or are totally
ignorant of, is the fact the Schrödinger wave equation also describes a
spinning particle. The spin is hidden because of the mistaken use of
the complex plane. The wave-function projects operators onto this plane,
which is something mysterious in orthodox QM treatments. But in a
proper G 𝐴 treatment we can see it for what it is, the projection is hiding
a physical process implicit when we make measurements, which is a pro-
jection of the laboratory observer’s frame onto the frame of the rotating
particle.
This should have been understood from the beginning of QM, but it
was missed. That’s an accident of history and excusable. What is not so

excusable is the continued neglect of this profound geometric origins for
quantum mechanics.
Perhaps ironically, one of the best ways to see all this is to start with
the Dirac theory for a spinor, then work backwards to the Schrödinger
equation as the non-relativistic limit. By going to the NR limit we of
course are not eliminating 3-space rotation, only the 4-space Lorentz ro-
tations. The result is the Dirac algebra morphs into the Pauli algebra.
And that’s how we will see Schrödinger was all along describing particles
with spin.
As a bonus, in the Dirac theory we will see the mysterious “gamma
matrices” are nothing but the natural geometric spacetime algebra basis
vectors of the fermion particle’s rest frame! But all this is for a later
section of this chapter. I want to warm-up first with some orthodox QM
before returning to geometric quantum theory.
A note of warning: in subsequent reading I have found the scattering

amplitude approach to multi-particle quantum mechanics, exemplified
in the Amplituhedron program [24–30], to be superior to using field vari-
ables and spinors. The fields and spinors are effective descriptions, and
not fundamental it appears. The Amplituhedron at least strongly sug-
gests this is so, by showing that treating scattering processes (the most
elementary interactions) on-shell results in amazing algebraic simplifi-
cations. The old Lagrangian field theory approach thus seems almost
defunct, a relic, of course still useful in computations where Amplituhe-
dra forms have not yet been discovered.
The trouble is I have yet to translate all the Amplituhedron formalism
into geometric algebra. So for now I’ve used the old Schrödinger-Pauli-
Dirac spinor formalism. I worry that this will introduce needless arte-
facts that we may be tempted to try interpreting geometrically, when in
fact they are just fictions that result form using the hugely (gauge) re-
dundant field theory descriptions.
* * *
I will now revert back to some standard QM to ease into some of the
mathematics. But I want the studnet in the back of their mind to be
always thinking of the wave function in the ℂ Hilbert space as really just
a dopey way of representing a normal spacetime spinor or rotor.
4.2.2 Simple QM Preliminaries
This next problem explores the relation between the real valued probabil-
ity density and the phase of the wavefunction, and we find the analogue
to the phase in classical mechanics. We will not gain a great mystical
insight into why our description of nature must use complex amplitudes
instead of straight-forward probabilities this way, but it is a good exercise
for developing some insight into what quantum mechanics is all about.
Problem 4.1. Starting with Schrödinger’s equation,

𝜕Ψ
𝑖ℏ = 𝐻ˆ Ψ
𝜕𝑡
for a non-relativistic particle, with KE= 𝑝2 /2𝑚 and potential 𝑉 (𝑟 ).
(a) Using the standard interpretation of 𝜌 = Ψ∗ Ψ = |Ψ| 2 as the parti-

cle’s probability density, derive a continuity equation for the par-
ticle’s probability. To do so, note that a “continuity equation” for
a conserved quantity is a relationship between the rate of change
of that quantity in an infinitesimal volume of space and the flow
of that quantity into/out of the volume.
(b) (i) Use your continuity equation to identify the probability cur-
rent 𝑗 , (flow of probability) in terms of the wavefunction Ψ.
(ii) Then show that your result can be reduced to,
ℏ
𝜌= =m{Ψ∗ ∇Ψ} .
𝑚
(c) By writing the complex wavefunction Ψ in polar coordinate form,
Ψ = 𝜌 1/2 𝑒−𝑖𝑆/ℏ
find an expression for the phase, 𝑆, of the wavefunction in terms

of the probability current 𝑗 .
(d) Derive an equation for the time derivative of the phase factor 𝑆,
and thus start exploring the physical meaning of the wavefunc-
tion phase 𝑆 by exploring the classical limit where ℏ → 0. A first
hint is to note that the Schrödinger equation for 𝜕𝑡 Ψ has an ℏ

factor in front of ∇2 Ψ, while ∇2 Ψ itself will get a ∇2 𝑆/ℏ2 factor.
And on the other side taking a time derivative of Ψ will give an
𝑆/ℏ factor, so you can begin to see several terms in the 𝐻 Ψ part
of the Schrödinger equation will vanish when ℏ → 0 but other
terms will “blow up”, and so when we divide through by a suit-
able power of ℏ and then take the limit ℏ → 0 we will end up
with some semi-classical residual relationships. These will tell
us what 𝑆 corresponds to classically.
(An alternative approach would be to find the expectation value
for 𝑆, because in the standard lore of quantum mechanics expec-
tation values correspond to classical observables. But we will be
able to get a semi-classical interpretation for 𝑆 without compu-
tating expectation values.)
Hint: The Hamilton–Jacobi equation in classical mechanics re-
lates a particle’s actions 𝑆 to the potential energy function 𝑉 , ac-
cording to:
1 2
𝑆¤ + ∇ 𝑆+𝑉 = 0
2𝑚
Hints for Problem 4.1.(a) The suggested method is to start by ex-

pressing 𝜕𝑡 𝜌 and working backwards to see it as a divergence of some
vector field.
Hint 2. You might need the vector calculus identity,

∇ • 𝑓 ∇ 𝑔 = ∇ 𝑓 • ∇ 𝑔 + 𝑓 ∇2 𝑔.
Hint 3. So beginning with,

𝜕𝜌 ∗ 𝜕Ψ 𝜕 Ψ∗
=Ψ +Ψ
𝜕𝑡 𝜕𝑡 𝜕𝑡
then note that,

𝜕Ψ 𝑖 𝑖 𝑝2
= − 𝐻Ψ = − Ψ + 𝑉Ψ
𝜕𝑡 ℏ ℏ 2𝑚
𝑖ℏ 2 𝑖
= ∇ Ψ − 𝑉Ψ
2𝑚 ℏ
∗

𝜕Ψ 𝑖 ∗ 𝑖 𝑝2 ∗ ∗
and = 𝐻Ψ = Ψ + 𝑉Ψ
𝜕𝑡 ℏ ℏ 2𝑚
𝑖ℏ 2 ∗ 𝑖
=− ∇ Ψ + 𝑉 Ψ∗
2𝑚 ℏ
¤
Putting these expressions into 𝜌,
𝜕𝜌 𝑖ℏ ∗ 2 𝑖 𝑖ℏ 𝑖
∴ = Ψ ∇ Ψ − Ψ∗𝑉 Ψ − Ψ∇2 Ψ∗ + Ψ𝑉 Ψ∗
𝜕𝑡 2𝑚 ℏ 2𝑚 ℏ
and use the fact wavefunctions are just complex numbers, so they com-
mute with 𝑉 , so the terms in 𝑉 cancel, and we get,
𝜕𝜌 𝑖ℏ ∗ 2
= Ψ ∇ Ψ − Ψ∇2 Ψ∗
𝜕𝑡 2𝑚
From here you can probably finish up.
Solution for Problem 4.1.(a) Using the vector calculus result, we can
write,
Ψ∗ ∇2 Ψ = ∇ • Ψ∗ ∇Ψ − ∇Ψ∗ • ∇Ψ

and likewise for Ψ∇2 Ψ∗ , and noting ∇Ψ∗ • ∇Ψ = ∇Ψ • ∇Ψ∗ , we get,
𝜕𝜌 𝑖ℏ ∗ ∗

= ∇ Ψ ∇Ψ − ∇ Ψ∇Ψ
• •
𝜕𝑡 2𝑚
= −∇ • 𝑗
𝑖ℏ ∗
where 𝑗 ≡− Ψ ∇Ψ − Ψ∇Ψ∗ .
2𝑚
The continuity equation for particle probability density is thus,
𝜕𝜌
+ ∇ • 𝑗 = 0.
𝜕𝑡
That’s it. ,
Hints for Problem 4.1.(b) The solution to part (a) has done most of
the work, we have already identified the current 𝑗 in terms of Ψ. Since
Ψ is a complex valued function it has real and imaginary parts, and Ψ∗
is it’s complex conjugate. Then noting,
∗
Ψ∗ ∇Ψ = Ψ∇Ψ∗
and,
𝑧 − 𝑧∗ = 2𝑖 =m( 𝑧) , for any complex number 𝑧,
Solution for Problem 4.1.(b) we find,

𝑖ℏ
𝑗 =− 𝑍 − 𝑍∗ , where 𝑍 = Ψ∗ ∇Ψ
2𝑚
𝑖ℏ
∴ 𝑗 =− 2𝑖 =m Ψ∗ ∇Ψ
2𝑚
ℏ
= =m Ψ∗ ∇Ψ .
𝑚
That’s the desired result. ,
Hints for Problem 4.1.(c) The polar form for Ψ can be taken as a def-
inition of the phase factor 𝑆,
Ψ = 𝜌 1/2 𝑒−𝑖𝑆/ℏ
we note firstly that 𝑆 has the dimensions of action (energy×time). So we
might expect a classical analogue for 𝑆 to be the classical action, which
is a functional integral of the particle Lagrangian. We want to see if this
is the case, basically. So this question is only a Lemma towards this goal.
Hint 2. It seems obvious to start by plugging Ψ(𝜌 , 𝑆) into the continuity

equation. Why obvious? Because differentiating (or taking the gradient)
of Ψ will pull down a factor of ∇ 𝑆 by the chain rule on the exponential.
We can then (naturally) expect to be able to divide out the exponential
and be left over with some relation between 𝜌 and ∇ 𝑆.
We should note that 𝜌 and 𝑆 are real numbers (scalar fields actually),
so
Ψ∗ = 𝜌 1/2 𝑒−𝑖𝑆/ℏ .
Hint 3. But for this lemma we do not need the full continuity equation,
you should be able to squint your eyes and see that the imaginary part
of Ψ∗ ∇Ψ is all we need to obtain a relationship between 𝑗 and 𝑆.

Solution for Problem 4.1.(c) Start with Ψ = 𝜌 1/2 exp 𝑖𝑆/ℏ , and take
the gradient, using the product and chain rules for differentiation,

1 √ 𝑖
∇Ψ = √ ∇𝜌 + 𝜌 ∇ 𝑆 𝑒𝑖𝑆/ℏ
2 𝜌 ℏ
√
and since Ψ∗ = 𝜌 𝑒𝑖𝑆𝑡 ,
∇𝜌 𝑖
∴ Ψ∗ ∇Ψ = + 𝜌 ∇𝑆
2 ℏ
Now ∇𝜌 is real, and so is ∇ 𝑆, so the imaginary part of the above is the

second term,
𝜌
∴ =m Ψ∗ ∇Ψ = ∇ 𝑆
ℏ
𝜌
∴ 𝑗= ∇ 𝑆.
𝑚
This is the result we sought. ,
Hints for Problem 4.1.(d) The question hinted plugging Ψ(𝜌 , 𝑆) into
Schrödinger’s equation and then taking the semi-classical limit ℏ → 0.
So that’s what we will do.
Hint 2. So start with Schrödingr’s equation for a potential 𝑉 (𝑟 ),

𝜕Ψ ℏ2 2
𝑖ℏ = − ∇ Ψ + 𝑉Ψ
𝜕𝑡 2𝑚
and then using the Ψ(𝜌 , 𝑆) form, and the product and chain rules for
differentiation w.r.t. time 𝑡,
√
𝜕Ψ 1 𝑖 𝜌
= √ 𝜌¤ + 𝑆¤ 𝑒𝑖𝑆/ℏ
𝜕𝑡 2 𝜌 ℏ
√
√ 𝑖 𝜌
and ∇Ψ = ∇ 𝜌 + ∇ 𝑆 𝑒𝑖𝑆/ℏ
ℏ

2 2√ 𝑖 𝑖√ 2 1√ 2 𝑖𝑆/ℏ
∴ ∇ Ψ = ∇ 𝜌 + ∇𝜌∇ 𝑆 + 𝜌 ∇ 𝑆 − 2 𝜌(∇ 𝑆) 𝑒
ℏ ℏ ℏ
Hint 3. Plugging these into Schrödinger’s equation and dividing by

𝑒𝑖𝑆/ℏ , gives us,

𝑖ℏ √ ¤ ℏ2 2√ 𝑖 𝑖√ 1√ 2 √
√ 𝜌¤ − 𝜌 𝑆 = − ∇ 𝜌 + ∇𝜌∇ 𝑆 + 2
𝜌 ∇ 𝑆 − 2 𝜌(∇ 𝑆) + 𝜌 𝑉
2 𝜌 2𝑚 ℏ ℏ ℏ
Hints for Problem 4.1.(d) Now just let ℏ → 0 in the above relation,
and you will see all we have left over classically is,
√
√ ¤ 𝜌 √
− 𝜌𝑆= (∇ 𝑆) 2 + 𝜌 𝑉
2𝑚
The Classical Action and Quantum Phase 93
1
∴ 𝑆¤ + (∇ 𝑆) 2 + 𝑉 = 0
2𝑚
which is precisely the Hamilton–Jacobi equation for a classical particle.
,
* * *
4.3 The Classical Action and Quantum Phase
We have just seen that as ℏ → 0 the quantum mechanical phase 𝑆, ap-

pearing in 𝑒𝑖𝑆/ℏ is the classical action.
We now want to develop the above result further. The Hamilton–
Jacobi equation relates the classical action to the potential energy func-
tion. Recall (from your classical mechanics courses) that the classical
action is a functional of the position of the particle, 𝑆 = 𝑆 [𝑥 ( 𝑡)]. This idea
of a ‘functional’ means to compute 𝑆 we need to feed it an entire trajectory,
not just a single set of parameter variables.
Here 𝑥 ( 𝑡) does not have to be the actual path of the particle, since 𝑆 is
defined for all times 𝑡 no matter what path 𝑥 ( 𝑡) we choose. The classical
mechanics dynamical principle is that the actual path of the particle is
the path which uniquely minimizes (or extremizes) the integral of the
action over the path.
This action minimization principle is equivalent to Newton’s Second
law of motion. But it is a global or non-local principle: it requires con-
sidering all possible paths and selecting the one which minimizes (or ex-
tremizes) the action from 𝑡 = 𝑡1 to 𝑡 = 𝑡2 . The interesting point here is
that 𝑡1 and 𝑡2 can be any times you like, they do not have to be infinites-
imally close. That’s why I said the action minimization principle is a
global principle, it uses integral calculus and does not require infinitesi-
mal calculus.
It should perhaps not surprise us that Newton’s equations (which are
local and infinitesimal calculus forms) have an equivalent formulation
in global integral calculus. After all, once you specify the potential en-
ergy function and some initial conditions then there is a unique solution
to Newton’s equations for all times, provided the potential energy func-
tion is smooth. So there has to be something special about the global
solutions. That ‘special thing’ is that the solutions minimize the action
between any two chosen points of the motion.
But note the Hamilton–Jacobi equation is a differential equation. So
it is a local dynamical differential equation, not global. How does this
arise form the global action minimization principle?
The way to turn the global action minimization principle into an in-
finitesimal calculus form is to first note that 𝑆 [𝑥 ( 𝑡)] can be viewed as a
function of the end points of the unique path which minimizes 𝑆,
𝑆 (𝑥 𝑖 , 𝑥 𝑓 , 𝑡 )
Normally we do not know the future, so we know 𝑥 𝑖 but not 𝑥 𝑓 , so we

would think of 𝑆 by taking a known fixed position of the particle at 𝑡 = 0,
and consider 𝑆 to be a function of the final position and time,
𝑆 = 𝑆 (𝑥 𝑓 , 𝑡 )
This is now an ordinary function and we can partially differentiate it

w.r.t. time 𝑡 or positions ( 𝑥, 𝑦, 𝑧) and get a differential equation. The D.E.
we get is precisely the Hamilton–Jacobi equation. We leave the detailed
proof to a course on classical mechanics.
4.4 Path Integrals and Phase
Since quantum mechanics is the more fundamental theory, one way of

interpreting the previous results would be to argue that the Lagrangian
formulation of classical mechanics is the more direct inheritor from foun-
dations of physics than the Newtonian formulation.
In practice we often find neither the Newtonian nor the Hamiltonian,
nor the Lagrangian formulations are inherently more practical than the
other, since it depends upon the system under question which is the eas-
ier to handle. However, quantum mechanics seems to teach us that the
Lagrangian and Hamiltonian formulations are truly more fundamental.
In quantum mechanics we never use the concept of forces, instead abso-
lutely everything is cast in terms of energy potentials. More generally,
we know that fundamental physics is driven by interactions. The funda-
mental interactions are, moreover, geometrical in origin, that is to say
Path Integrals and Phase 95
they involve either topology or vertex algebra, not action-at-a-distance

force fields. Force fields are fictional.
We know this from clues like the Feynman diagram pictures of funda-
mental interactions. In the Feynman diagram formalism fundamental
particles have to interact at vertices, they do not influence each other over
separate distances. And string theory gives us more clues, suggesting to
us that the vertices are not truly fundamental and that the Feynman ver-
tices are more like topological interactions, not point-like interactions.
So when quantum physics seeks to, “unify the fundamental forces”
this is a huge misnomer. The quest should be to, “unify the fundamental
interactions”.
If it were all topology we would have a hope of retrieving a classical
theory of physics. But quantum mechanics also tells us that the topolog-
ical interactions need to be averaged over somehow, because they are not
deterministic. This leads to formulations like the path integral formula-
tion of quantum mechanics.
A Feynman path integral is a summation (over all possible histories)
and normalization procedure. So it is a type of averaging over all his-
tories. What is fascinating about the path integral formulation is that
this averaging does not compute for us an expectation value, instead it
computes a probability, the probability of going from some initial state to
some final state. That is because, unlike an averaging summation in con-
ventional statistics and probability theory, the Feynman path integral is
a summation and normalization over probability amplitudes.
Probability amplitudes are complex square roots of probabilities. So
it makes some intuitive sense that if we take averages over probability
amplitudes we will get something like probabilities as a result, not ex-
pectation values.
To compute classical expectation values we need to do one further level
of averaging over the probabilities derived from the path integrals.
Now think about what this means for the classical action 𝑆 which be-
comes ℏ times the phase of the wavefunction in QM.
The phase factor tells us how fast the complex probability amplitude
is oscillating as we move around in space and time. As a complex val-
ued function Ψ varies from point to point in space and time. From point
to point 𝑖𝑆/ℏ will rotate the wavefunction (viewed as a complex number

in the Argand plane). This is not a physical rotation in space, it is a
probability-related rotation in an abstract probability amplitude space.
It has no consequences unless the path of a particle has more than one
alternative path. If there is more than one path a particle could take from
( 𝑥𝑖 , 𝑡𝑖 ) to ( 𝑥 𝑓 , 𝑡 𝑓 ) then the probability amplitudes need to be summed, and
since they are complex numbers they can sum non-additively, and can
sometimes cancel one another. This is the phenomenon of interference,
usually associated with waves, but now also associated with fundamental
particles.
4.4.1 Are Path Integrals Real?
The formal mathematical development of the path integral formulation

of QM procedes by something like the following sequence:
1. Write down the Green’s function expression for the Schrödinger

equation — this is called the propagator, and it is just the expres-
sion,
h𝑥 𝑓 , 𝑡 𝑓 |𝑈 ( 𝑡 𝑓 , 𝑡𝑖 )| 𝑥𝑖 , 𝑡𝑖 i
which is the solution to Schrödinger’s equation for the special ini-
tial condition of a highly localized particle at known position and
time ( 𝑥, 𝑡), i.e., a Dirac delta function initial state. (Green’s func-
tion solutions to P.D.E.’s are an important case to know, because
all other well-behaved solutions can be built up from a (possibly in-
finite) combination of point sources.)
2. By integrating over these propagators between infinitesimal time

intervals, we will arrive at a path integral expression for the macro-
scopic propagator from any given ( 𝑥𝑖 , 𝑡𝑖 ) to any given ( 𝑥 𝑓 , 𝑡 𝑓 ). For
which we will need some appropriate integrand. We will not know
exactly what this integrand should be.
3. When you obtain the above expression for the propagator, and you
know the potential 𝑉 , you will still not know how to compute it as a
function of 𝑥 and 𝑡. The usual process is to guess a solution and show
that it solves Schrödinger’s equation. The best guess can be worked
out fairly simply by figuring out what a solution would be in the
classical limit as ℏ → 0. We find an action extremization principle,
which are the Euler-Lagrange equations for classical mechanics.

This leads us to guess a path integral expression for the propagator
of the form,
ˆ
h𝑥 𝑓 , 𝑡 𝑓 |𝑈 ( 𝑡 𝑓 , 𝑡𝑖 )| 𝑥𝑖 , 𝑡𝑖 i ∼ D 𝑥 ( 𝑡) exp 𝑖𝑆 𝑥 ( 𝑡) /ℏ
where the operator D denotes a path integral form rather than a

normal space integral 𝑑𝑥. This is a crucial part of the axioms of
quantum mechanics! Take note! In classical physics this formal-
ism would be considered a complete fictional artifice — because why
would the non-physical non-classical paths have anything to do with
the propagation of a particle from ( 𝑥𝑖 , 𝑡𝑖 ) to ( 𝑥 𝑓 , 𝑡 𝑓 )? Why on Earth
would we have to integrate over paths we know that particles can-
not possibly traverse?
Later in this chapter we will complete the above mathematical develop-

ment. But, to my mind, the mathematics is relatively boring. But don’t
get me wrong. The mathematics is cool, it’s just not the most fascinat-
ing thing. The idea we can perform an integral of an expression over an
abstract “space of all possible spacetime paths” is completely audacious
and exciting. Especially exciting is that with such a procedure can obtain
probabilities for outcomes of realistic physical experiments. Although it
should be tempered with the knowledge that once we stick in an actual
energy potential 𝑉 we will not be able to analytically compute the path
integral except for very special idealized types of 𝑉 . However, we should
be used to that sort of disappointment. Most integrals arising in the real
world are not soluble analytically and must be treated numerically or
asymptotically. The awesome thing is not being able to analytically solve
problems, it is being able to analytically state the problem. That’s what is
truly awesome. To paraphrase Einstein; asking the right question is the
hardest thing.
Solving analytically stated problems exactly is vastly over-rated from
a theoreticians point of view. The beauty of physics is that we can ex-
press the laws of nature analytically, not necessarily that we can solve
them. Forget the mess of obtaining solutions. If you are excited about
aesthetics then stating a problem analytically is all the beauty you need.
It means nature is not an unfathomable mystery. In fact, it is more beau-
tiful that we cannot then solve the equations analytically, because that
means nature is both fathomable and hugely complicated beyond simple
computational tractability. I, for one, would be truly disappointed if all

the governing equations of physics could be exactly solved.
Another more important reason why I do not want to immediately pro-
ceed to solving the path integral problem is the philosophical project I
have in mind to better understand the foundations of quantum mechan-
ics. Using QM in applications has never much interested me, though
people who are QM engineers have my greatest respect and the best of
them are geniuses of the highest order.
So I want to ask a seemingly naïve question: is the path integral real?
What I mean by this is something like, “do particles like electrons and
photons (assuming they are localized particles, not waves) actually follow
all possible histories, in some sense? If if so, in what sense?”
The problem is we do not experimentally observe all possible paths,
our evidence of the reality of Feynman sums over histories is entirely ex
post facto. In which case someone can always blithely assert that the
sums over histories are fictions which just happen to allow us to compute
accurate probabilities. This is not a satisfying answer to my mind, it
reeks of the weak non-committal that is characteristic of the Copenhagen
Interpretation of quantum mechanics (to which the Sum Over Histories
interpretation is a rival).
A more satisfying explanation of the success of a Sum Over Histories
formulation is to take the histories seriously as real events that actually
occur, at least statistically in some sense as a sampled ensemble. We only
need the path histories to be sampled in order to explain the success of
the theoretical formalism developed by Feynman, because the outcome
of Feynman’s formalism are probabilities.
The question is then how do real particles sample the set of all possible
paths from ( 𝑥𝑖 , 𝑡𝑖 ) to ( 𝑥 𝑓 , 𝑡 𝑓 )?
I will not claim these are the only mechanics but a serious attempt
to discern a non-magical way particles can explore all possible histories
must include at least the following:
1. Waves Only. Abandon the idea particles are localized, and embrace
fully the wave function as reality. I reject this proposal because
every single experiment designed to detect systems described by a
wave function has found localized particles, not distributed energy
over all of space. We know waves can be localized, in the form of
soliton solutions, but we cannot possibly be expected to believe wave

functions are always solitons. There is no known constraint in na-
ture which would predict wavefunctions always end up as solitons
when observed. We would have to postulate this as an unproven
axiom, which, to me, is inelegant and likely incorrect.
2. Fields Only. Particles are not real, non-localized fields are, but
whereas fields obey the laws of Schrödinger time evolution, when-
ever field energy is detected is coalesces into a localized quanta. I re-
ject this proposal because similarly to the waves-only picture it is far
too inelegant and post hoc. Why would fields magically coalesce into
lumps only when measured, but not at other unobserved times? It
is clear fields cannot be too lumpy in the unobserved times because
then the Feynman histories would not get sampled. Remember, we
are trying to take the path integral formalism seriously.
3. Particles in a Field. The approach to Sum Over Histories sam-

pling I find most believable is to envisage the particles as topo-
logical features of a field, the field being spacetime. The particles
are localized, but if the field of spacetime admits non-trivial topol-
ogy then exotic non-macroscopic structure like wormholes can ex-
ist. The wormhole is a very natural way to get a particle dancing
wildly all over spacetime, thus sampling all possible histories, in-
cluding looping backwards in time through closed timelike curves.
The main problem I see with this approach is that particles are not
intrinsically ‘interfering objects’, only entities described by super-
posable waves (vibrations in some physical medium) can interfere.
But when particles are considered to be topological structures of
spacetime then there is probably a way to get them self-interfering,
but only if they can loop back in time to do so. A particle on a pure
timelike path, no matter how wild, cannot interfere with itself.
Some other notes: the Sum Over Histories formalism does not explic-
itly assume interference effects. Interference arises as a secondary result
of the sums of the complex phase factors. This is hard to identify as a lit-
eral wave interference. The phase factors are not particle probability
densities, they are just a mathematical component thereof. We can use
complex amplitudes or sinusoids to describe waves, but this is not how
the phase factors are motivated.
As we have seen, the phase factors are closely related to the classical
action. Geometrically the action is a phase space volume, energy×time,

which is closely related to a particle’s “worldline” or more correctly
“world-volume”. The world-volume is a physical spacetime 4-volume,
whereas the action is this same volume transformed into energy×time
units.
4.5 Spin–Orbit Coupling
The Schrödinger hydrogen atom model predicted an 𝑛 = 2 fine structure

splitting of 𝛼4 𝑚𝑐2 /12, this was about 3/8 too large. Schrödinger did re-
alize this was due to the electron spin. So the multiplicity of states is
larger than in the naïve model (each electron orbit has a degeneracy= 2).
Goudsmit and Uhlenbeck had suggested the electron has an intrinsic
spin, analogous to an internal angular momentum ℏ/2. The electron in
the hydrogen atom is nearly relativistic, so it sees an effective magnetic
field from it’s frame of motion around the nucleus. The spin interacts
with the magnetic field, splitting the electron energy levels. This is an
“intrinsic Zeeman effect” (the magnetic field is from the relativistic elec-
tron motion, not from an external field). The Zeeman effect is usually
refers to the splitting due to an external magnetic field. So this intrin-
sic splitting is instead just called the fine structure of the hydrogen atom
spectrum.
4.6 Probability Conservation
The main goal of early quantum theory was to unite QM with special
relativity (SR). The first obstacle was trying to find a Lorentz invariant
wave equation. Schrödinger actually found the Klein–Gordon equation
before the Schrödinger equation, but did not publish it, probably because
he knew or suspected it was not physical. We now know the Klein–Gordon
equation is kind of physical, but cannot describe the motion of a single
particle, rather, it can describe a massive spin-0 field. So it is a field equa-
tion, not a particle–wave equation. In the early years of QM physicists
had not realized quantum theory could be best expressed as a field the-
ory. So Schrödinger (unsuccessfully) and then later Dirac (successfully)
went in search of a relativistic wave equation using other tricks.
Probability Conservation 101
Recall, the Klein–Gordon equation was considered a fundamental

blend of QM and SR by naïvely using the relativistic energy–momentum
relation in Schrödinger’s equation, the squaring both sides p
to treat space
and time equally. Setting units so that ℏ = 𝑐 = 1, and 𝐻 = 𝑝2 𝑐2 + 𝑚2 𝑐4 ,
we get,
𝑖𝜕𝑡 Ψ = 𝐻 Ψ
becomes, −𝜕𝑡2 Ψ = ( 𝑝2 + 𝑚2 )Ψ. (Klein–Gordon)
To see why the Klein–Gordon cannot describe a single particle it is

best to go through the arguments pedagogically.
Problem 4.2. We first want to show probability is conserved for the

Schrödinger wave-functions. Then we’ll show probability cannot be
conserved for a hypothetical Klein–Gordon particle, which means there
are no such particles (unless you wish to tolerate a theory of physics
where particles can arbitrarily vanish or appear).
(a) Starting with the non-relativistic Schrödinger’s equation, per-
form a time derivative of the probability density to show the total
probability of measuring the particle somewhere is unity.
(b) Repeat for the Klein–Gordon equation to show the probability

density integrated over all space is not constant for all time.
Hints for Problem 4.2.(a) The probability density is 𝜌( 𝑥, 𝑡) = Ψ∗ Ψ.

Go to 1-dimension w.ℓ .o.g.
Hint 2. Insert 𝑖𝜕𝑡 Ψ = −𝜕𝑥2 Ψ + 𝑉 ( 𝑥)Ψ. Integrate this by parts.
Solution for Problem 4.2.(a) So we must have,

ˆ +∞
𝑑𝑥 𝜌( 𝑥, 𝑡) = 1, for all 𝑡.
−∞
Then the time derivative of the left-hand side must be identically zero.
Let’s see what it is?
ˆ ∞ ˆ +∞
𝜕𝑡 𝑑𝑥 𝜌( 𝑥, 𝑡) = 𝑑𝑥 𝜕𝑡 Ψ∗ Ψ + Ψ∗ 𝜕𝑡 Ψ
−∞ −∞
ˆ ∞
= 𝑑𝑥 (−𝑖𝜕𝑥2 Ψ∗ + Ψ∗ )Ψ + Ψ∗ ( 𝑖𝜕𝑥2 Ψ −
𝑖𝑉
𝑖𝑉
Ψ)
ˆ−∞
∞
= 𝑑𝑥 −𝑖𝜕𝑥2 Ψ∗ Ψ + 𝑖Ψ∗ 𝜕𝑥2 Ψ
h−∞ i∞
∗ ∗
= 𝑖 −𝜕𝑥 Ψ Ψ + Ψ 𝜕𝑥 Ψ
−∞
ˆ ∞
∗ ∗
−𝑖 𝑑𝑥 − 𝜕𝑥Ψ 𝜕𝑥 Ψ + Ψ
𝜕𝑥 𝜕𝑥 Ψ
−∞
h i∞
= 𝑖 −𝜕𝑥 Ψ∗ Ψ + Ψ∗ 𝜕𝑥 Ψ
−∞
= 0.
The boundary terms must of course vanish for physical reasons. You
could integrate by parts a second time and the boundary term would then
cancel identically, but it is already clear the whole integral vanishes. This
does not prove the probability density is unitary, it just shows that it is
constant. But that’s all we need for the present purposes. □
Hints for Problem 4.2.(b) TODO.
Solution for Problem 4.2.(b) TODO.
* * *
Note the caveat in the problem statement, “(unless you wish to toler-
ate a theory of physics where particles can arbitrarily vanish or appear).”
Well, maybe you already know that this is exactly what physicists have
indeed learned to tolerate! Quantum theory (more precisely, quantum
field theory, QFT) does not work relativistically unless we consider vir-
tual particles. The existence of virtual particles is basically the Feynman
diagram manifestation of quantum field theory. Feynman diagrams are a
particle picture, while the amplitudes of the field theory which the Feyn-
man diagrams are used for computing are what we get when the particles
are reinterpreted as “modes” of the fields.
The difference with the QFT picture and the Klein–Gordon wave-
equation problem is that in QFT the virtual particles or ghost field
modes must disappear when we integrate over all possible histories, and
we can tolerate this if we can make them field modes rather than real
particles. Whereas a Klein–Gordon particle that could appear or vanish
Probability Conservation 103
simply violates causality in a most fundamental way. The Feynman dia-

gram picture perhaps muddies this a little since it allows us to think of
the virtual articles as real particles. But the wave-equations do not exist
in the Feynman diagram formalism, so there really is no inconsistency
here. So as long as you stick to one or the other model you will not get a
problem with virtual particles.
In short, wave-equations have no place in relativistic QFT. The above
exercises showed that probability is not conserved if we allow relativistic
wave-equations. So that was the end of the road for relativistic wave-
equations. Or was it? Dirac did manage to derive a relativistic wave-
equation with a conserved probability current. However, we later learn
that the Dirac theory was implicitly a multiparticle theory, not a single
particle wave-equation.
So this is the real lesson: relativistic quantum theory forces us to con-
sider either fields or multiparticle states.
Chapter 5
Kinematics and Amplituhedra
The purpose of this chapter is to investigate a major new alternative to

quantum field theory. The finer purpose is to show — in I think the sim-
plest way — why any fundamental physics involving abstract “quanta”
(so any and all particle-like theories, including the Standard Model (SM),
strings, branes, geons, the whole lot) must by necessity after imposing the
most basic principles, be a theory of spin 0, 1, 2 particles including the
fermions in-between, and no other particles possible.
In other words, we will see from the seemingly most mild assumptions
about physics we are highly constrained in the possible basic theories
of fundamental interactions (hence forces) that we can describe. What
that means is those “mild assumptions” (such as unitarity, causality, lo-
cality, conservation of momentum–energy) are not really so mild after
all, they are highly constraining. Such understanding should be in the
background education of any theoretical physicist, and it is a shame past
education in QM could not teach such lore, but that was for good reason,
because the only why to understand it in the past was by wading through
introductions to quantum field theory. Thanks to the work of the ampli-
tuhedronistas we no longer have to do that, and there is this far more
natural way to understand this essence of the natural physical world.
The approach, which I will call scattering theory and amplituhedra
is in the tradition of 𝑆-matrix theory. First let’s briefly explain these
origins, and then later get on to the theory of Amplituhedra proper.
105
106 Kinematics and Amplituhedra
5.1 The S matrix origins
𝑆-matrix theory which was a failed attempt to do away with QFT (quan-
tum field theory) and just apply principles like the conservation laws plus
locality and unitarity to derive scattering amplitudes. The idea is that
we isolate a compact region of spacetime, send some stuff in, and observe
some stuff come out. Every time we repeat the experiment sending the
exact same stuff in, different stuff might come out, there is no consis-
tency in the scattering, we see only consistency in some abstract rules,
like momentum-energy conservation, angular momentum conservation
and unitarity. The thought occurs that maybe those abstract rules are
all there is to physical reality! This could be considered the essence of
quantum mechanics, though as we learn it is not the whole story, which
is why 𝑆-matrix theory fails. But for plenty of practical purposes this
idea of scattering processes being like a black box suffices to understand
the world form an elementary point of view.
What we know is that 𝑆-matrix theory is nevertheless very close to
being all there is no know about the world. The attempt to modernize
𝑆-matrix theory and tell the whole story about fundamental physics is
what the Amplituhedron program is all about.
The thing is, only tools one needs to do physics are the scattering am-
plitudes (or equivalently correlation functions). In this sense it is a darn
shame 𝑆-matrix theory does not work. Nature is telling us we need some-
thing like QFT or string theory to say a little bit more about the inner
workings of that abstract black box that is the scattering process, in order
to correctly calculate the scattering amplitudes.
Before we move on to describing the Amplituhedron program, it is
worth relating the 𝑆-matrix approach to the most recent advances in
theoretical physics and cosmology, namely the gauge/gravity dualities,
or AdS/CFT correspondence.
AdS/CFT, or gauge/gravity duality, tells us that in the fictional world of
anti-de Sitter spacetimes a consistent quantum field theory on the bound-
ary of the spacetime is dual to a pure gravity theory in the bulk with one
higher space dimension. Thus, if the bulk spacetime has a (3 + 1) signa-
ture then we would expect a 2 𝑑-quantum field theory to exist describing
the physics on the boundary. But we can add hidden space dimension by
either compactification (string theory) or decorating the space manifold
The S matrix origins 107
with topology (geon theory) to produce an effective higher space dimen-

sion, so for instance an effective (4 + 1)-dimensional spacetime will have
a 3 𝑑-quantum theory on it’s boundary. The gravity theory can have even
higher dimensions, say 9-space dimensions if 𝑑 − 4 dimensions are com-
pactified. Then we retain all the unique advantages of 4-dimensional
manifolds. (These are advantages for getting interesting physics, they
are severe disadvantages for calculations. Physics takes precedence to
ease of computation.)
These gauge/gravity correspondences are nice, but they cannot de-
scribe the real world. In these AdS spacetimes light can propagate all
the way to the boundary and then reflect back into the bulk (so it is what
Nima Arkani-Hamed calls “physics in a tin can”), which means scattering
amplitudes are well-defined and so the QFT is well-defined. The space
dimension in AdS has negative curvature, this is why it is a tin can. All
computations can be done exactly (in principle).
Those universes are thus, in a meaningful sense, boring. We can be
like gods in them. Now there is no reason why not having the capacity
of gods is a necessary metaphysical requirement for life, but it turns our
actual universe is one such place where the spacetime geometry is such
that we cannot be gods. This is de Sitter spacetime, spacetimes where the
space curvature globally is positive. In between Ads and dS we have Flat
or Minkowski spacetime, and physics is difficult enough in flat space.
The problem here is that the boundary is at infinity, and yet our ob-
servable horizon is distinctly closer to us than infinity, it is defined by the
light horizon or Celestial sphere. In our actual de Sitter space the cosmo-
logical constant is positive, so the Celestial sphere is shrinking relative
to all of spacetime — more space escapes us forever each moment.
It is thus doubtful QFT is well-defined for Minkowski or de Sitter
spacetimes. Some people view this as a great problem, but it is a math-
ematical problem, not a physical problem. Do you think nature cares
whether a QFT is defined? If we do not demand an exact QFT, and
are instead comfortable working with merely effective QFTs, then there
is no great problem. However, it does mean scattering amplitudes can-
not be precisely calculated. The question you have to ask yourself then
is why would you want to have precisely calculable scattering ampli-
tudes? There is no god-given reason to expect it to be so.
For the practical physicist the 𝑆-matrix idea is still fruitful. We do not
have to live on the boundary of our spacetime nor has access to the infor-
mation there (as we could in AdS) to do good physics. As far as fundamen-
tal particles are concerned, our position a few dozen metres distant from
a collider experiment may as well be at spatial infinity. So in everyday
terms, we are living on an effective boundary. We just have to deal with
the noise from nearby stuff that messes up the ideal approximation that
nothing can get in from our surrounding laboratory box here on Earth
to infect our collider experiments. We know how to do this, it is called
statistics.
Nevertheless, tell a pure theorist they have no exact theory and they
will be quite upset. You cannot stop them searching for an exact theory, it
is in their nature. The Amplituhedron idea is one way to conservatively
extend 𝑆-matrix theory so that we capture as much essential physics as
possible, in the spirit of Feynman, but without the painfulness of seem-
ingly endless Feynman diagram calculations.
5.1.1 Inside the black box
When physicists look inside the black box of a scattering process they
pull out Feynman diagram methods. The Feynman method also has more
modern operator calculus formalism, which do the same thing only less
diagrammatically. It is of no use at all to argue about which method is
“more fundamental” because they are describing the exact same physics.
Nature does not care how we compute the amplitudes.
However, it is not so simple. To use the standard Feynman diagram
approach one assumes all manner of virtual particle process can occur,
one has to add terms in the path integral for all of them, even if they are
not real. The end result is a mess, and if you make no mistake you get
some approximation to some loop order for your scattering amplitude.
But amazingly it all tends to simplify with “miraculous” cancellations.
The reason why the Feynman diagram methods starts off as a horrible
mess is due to the Lagrangian treatment given. I have often considered
the Lagrangian formulation of QM to be the most natural and obvious
thing to do. It says nature explores all possible paths. However, because it
is used as a treatment of Hamiltonian time evolution it is not manifestly
Lorentz invariant. This means off-shell processes have to be included in
the perturbative expansion. When a particle is off-shell it is not propa-
gating in the light cone, so is effectively a tachyon. Such particles cannot

be real. They are virtual.
Thus the deceptive appearance of the Feynman path integral method
as a nice literal description of what happens in nature is just that, it is
deceptive.
The thing is, an off-shell particle can always be Lorentz boosted to a
frame where it is on-shell. Only now previously on-shell processes will
appear off-shell. The problem here seems to the the QFT framework is
asking for Hamiltonian time evolution when there really is no such global
thing. The field variables can be thought of as fictions that we need to
introduce to get all the accounting right — unitarity, causality, locality,
CPT and all that.
What if there was another way? Without the fields?
One is led to wonder about something else behind the scenes that
forces us to imagine the virtual particles. If we understood what this
“something else” was then we could dispense with the fiction of virtual
particles. The amplituhedron method has some promise of doing just
this for us, which is one of the main motivations I had for learning am-
plituhedron theory. This is how I began to approach amplituhedron the-
ory, I think o it as Feynman diagrams done right. The amplituhedron
method is roughly what you get if there are no fields, and instead all you
have are particles that scatter, but according to the rules of QM not CM.
This is more in the spirit of Feynman than Feynman diagrams based on
Lagrangian field theory.
5.1.2 Fields or particles?
Before you start to worry all your previous quantum field theory learning
gets thrown out the window, it is worth bearing in mind for the practic-
ing physicists employed say at CERN or Fermilab, QFT or Yang-Mills
fields theory is still the way to compute scattering amplitudes. It’s just
that over time the Amplituhedron on-shell methods are proving to be far
more efficient and gradually replacing a lot of the traditional computa-
tion methods.
It is also not hard to interpret every quantum particle theory (but not
any classical particle theory) as a field theory. In QM there is a non-
zero amplitude for the particle to be somewhere, pick a place in time

and the particle could be there. In free-field (non-interacting) case this
amplitudes is the scattering amplitude, only it is trivial, since there is
no interaction there is no scattering, and we just describe such “trivial
scattering” by the free particle propagator. If we make the field theoretic
identification of the amplitude 𝜓 ( 𝑥) with a field 𝜙( 𝑥) we get a quantum
field theory. This allows us to describe negative energy modes — which
initially make no physical sense in the particle picture — which we only
latter identify as representing anti-particles. Remember, the only ele-
mentary particle not to first be discovered as a particle was the photon.
Also, bear in mind one of the masters of QFT, Steven Weinberg. He did
warn you over 30 years ago [31] that particles are the way to go. Wein-
berg considered introducing particles as the irreducible representations
of the Poincaré group as the best introduction. Field theory follows from
making idealizations about the particles. I also like Weinberg’s quip on
his page one:
“If it turned out that some physical system could not be described by a
quantum field theory, it would be a sensation; if it turned out that the sys-
tem did not obey the rules of quantum mechanics and relativity, it would
be a cataclysm”
Weinberg is thus careful to note that he is not saying “the particles are
more fundamental than the fields,” — but while you can protest you can
protest too much. Weinberg did not know about the development of Am-
plituhedron methods at the time he wrote his QFT volumes, had he un-
derstood the amplituhedron he might well have capitulated to the “par-
ticles are more fundamental” point of view. Who knows. The Weinberg
that lives in my head certainly did.
5.1.3 What about the gauge symmetries?
We cannot, or morally should not, use both particle and field descriptions
at the same time. The great success of QFT is mainly in the gauge theory
of the fundamental forces. The basic way gauge theory works is that in
the field theory one (might) discover there is gauge redundancy, meaning
a gauge transformation can be applied to the field without changing the
physics, it should therefore not change the Lagrangian.
I should mention here that most QFTs are described by specifying a

Lagrangian, which expresses the interactions between the fields. For
example, charged fermions coupling to photons is the particle view. In
the field view the electron field multiplies the electromagnetic field in the
Lagrangian.
Anyway, suppose we find such a gauge redundancy, and apply the
gauge transform, say,
𝜙( 𝑥) ↦−→ 𝑒𝑖𝑞𝜃( 𝑥) 𝜙( 𝑥)
for any arbitrary gauge parameter 𝜃( 𝑥) which can take on any scalar
value at any point 𝑥 in space. This is indeed the case for the electromag-
netism: the Dirac and Klein-Gordon fields are equally well physically
described up to a unit complex phase factor, hence there is a 𝑈 (1) sym-
metry in electrodynamics. We had better make sure this does not change
the Lagrangian!
The standard Lagrangian has a term in the field derivatives, 𝜕𝜇 𝜙 (if
not then you get no free particle propagation, just check the Klein-Gordon
or Schrödinger equations!). Of course we know this is really the momen-
tum operator 𝑝® but now in relativistic form, so it is ( 𝜕𝑡 , −∇𝑥 ). However,
the derivatives 𝜕𝜇 𝜙 do not transform like the field 𝜙 here. To fix this we
find we are forced to introduce another field in the Lagrangian with an
index 𝐴𝜇 , which multiplies 𝜙, so it is some kind of new interaction term
− 𝑞𝐴𝜙, then we find if we demand under the gauge transformation that
this new field transforms like,
𝐴𝜇 ↦−→ 𝐴𝜇 + 𝑖𝜕𝜇 𝜃
then this retrieves an invariant Lagrangian. You can check for yourself.
In this case a symmetry in the fields (charged fermions) forced us to as-
sume nature has some other field (photons).
Note we do not really need the 𝑖 in the phase factor, that just expresses
the charged particle is a fermion (the 𝑖 is hiding a bivector in the proper
spacetime algebra). So a similar argument could be made for a charged
boson.
* * *
The whole previous section can be totally reversed: we could begin

with our knowledge of classical electrodynamics and note the gauge re-
dundancy in defining the 4-potential 𝐴𝜇. We can then figure out how
the Schrödinger or Dirac quantum mechanical equations change when

applying a classical re-gauging. We will find 𝜓 ↦−→ 𝑒𝑖𝑞𝜃 𝜓, if the gauging
was,
𝐴𝜇 ↦−→ 𝐴 + 𝜕𝜇 𝜃 .
This just confirms our quantum mechanics is consistent with out classi-
cal electrodynamics.
However, something is gained, because the gauge symmetry for 𝐴𝜇 ( 𝑥)
is local, we can change it anyway we please at any point in space. So in the
quantum theory this implies there has to be a local field representing 𝐴.
This has to be some other kind of particle. Particles being the localization
of fields.
Strictly speaking we do not need to regard the necessity of the field 𝐴𝜇
as meaning there are new particles called photons. We can still insist all
we have are fields. It is really experimental observation that convinces
us the local nature of the 𝐴𝜇 field looks like a particle.
This now brings us back to the point of this section, which is to point
out there is no fundamental reason why the fields have to be regarded as
fundamental. They might be. But if we observed particle-like behaviour
of the type described by a gauge field like 𝐴𝜇 , then another valid interpre-
tation is that the charged fermion is interacting with a photon. This is
indeed the case in nature. We can use a field to describe interactions with
photons, and in doing so it looks like a quantum theory of electrodynam-
ics, since the gauge transform mapping is direct onto the electromagnetic
4-vector. But experimentally light behaves also like particles, so we can
also regard the field description as a useful fiction.
We can see this all in reverse, again. Suppose we had started with the
idea things interact locally, like particles. Forget for now that historically
light was probably always going to be viewed first as a field, just suppose
Einstein and the photoelectric effect had come along miraculously early.
Then we would have just the charges, spins and masses and couplings,
no fields, no gauge symmetry. The same picture we have in the ampli-
tuhedron program. If we later decided to now try a field theoretic model
we would find ourselves forced to use gauge symmetries, there would be
no other option if we wanted the known particle physics to be consistent
with relativity and quantum mechanics.
Now going back to the particles: once we realize we do not need fields,
we might in retrospect consider the gauge symmetry groups to have been
Kinematics of scattering amplitudes 113
just as redundant as the Amplituhedron program found them to be. We

would realize we can use locality to just look at the symmetries near
the identity, in other words, we would only need the Lie algebra. Particle
physics is basically to Lie algebra what field theory is to Lie groups. That
is the motivation for geon theory and is also, we now see, in the same
spirit as the amplituhedron program.
5.2 Kinematics of scattering amplitudes
We begin with assuming quantum mechanics postulates without stating

them. For our purposes it suffices to know that in all elementary interac-
tions momentum is conserved. What we do not know is what will happen
given we throw two or more particles into a collision. The interactions
are only probabilistically controlled by amplitudes for every possible pro-
cess. What is critical is to know the possible processes. Classical scatter-
ing theory is fine for employing conservation laws, but not for predicting
probable outcomes.
It turns out we will only need to know about three and four particle
scattering in order to deduce all the possible forces in nature. In this sec-
tion we introduce the notation for describing such scattering processes.
One-particle scattering (one particle in, zero particles out, or the con-
verse) does not occur in nature, which we describe by conservation laws,
which means particle “creation” or annihilation” always occurs in pairs.
Two-particle scattering is trivial, it is just propagation of particles. If
the particle changes type then some more-than-two particle process oc-
curred.
5.3 Yang-Mills or GR and nothing else
A Yang-Mills theory involves scattering intermediated by spin-0 (Higgs)

or spin-1 particles (gauge bosons). General relativity can be derived as
a continuum limit of interactions of spin-2 particles. In this section I
will try to show why there are no other fundamental forces in nature.
Yang-Mills and GR is it.
Chapter 6
Speculative Ideas
This chapter collects some random thoughts that I know are probably all
wrong, they are the sorts of ideas you have when day-dreaming, maybe
one out of 100 turns out to be a good insight. I have by no means kept
all such ideas that spring to my mind, most are obviously flawed. The
remaining I think are crazy, absolutely crazy, but as we know the future
physics often looks absolutely crazy to the past, so I hope there are some
gems in here.
* * *
115
116 Speculative Ideas
6.1 Miscellaneous Notes
This section is for main chapter contents that I either have not bothered
writing up or may never get to.
Yang–Mills massive vector bosons
In Yang–Mills theories the force is said to be mediated by boson exchange.

The character of the force (particle types and coupling) is given by irre-
ducible representations of a symmetry group, as per Wigner’s classifica-
tion, but extended beyond just Poincaré symmetry. In T4G theory I al-
ways think of the internal symmetry group as a result of 4-geon topology,
not as a fibre bundle, however the gauge fibre bundle is a decent math-
ematical approximation, given quantum mechanics effectively spreads
particles out over spacetime — where you want a gauge boson you can
have one, with some probability.
To get a short range force the gauge boson has to be massive. Massless
bosons lead to long range 1/𝑟 2 potentials generically. But if it is massive
then it can change character. Massless particles cannot change character
because they experience no proper time. This explains how the weak
bosons can “change flavour” and turn leptons into quarks and vice versa,
totally critical to explaining nuclear beta decay (the atomic nucleus has
no leptons) and hence critical for life (stars will not shine without beta
decay processes).
However, Yang–Mills gauge bosons begin life (theoretically) as mass-
less, this is because the unitary symmetry group requires them to be
massless. So a mass has to be induced by a symmetry breaking mecha-
nism. This is now recognized as the Higgs mechanism, which has been
confirmed thanks to the LHC.
It is called a “symmetry breaking” mechanism because at high ener-
gies the Higgs coupling goes to zero relative to other coupling constants,
then the Yang–Mills forces unify somewhat since they all become long
range (though only in some asymptotic limit of very high energy, perhaps
the Planck energy).
Speculations on Vacuum 117
6.2 Speculations on Vacuum
Haven’t had too many crazy ideas about the vacuum recently. With geon
theory I am biased towards thinking non-mystically, so I think vacuum
energy is just remnants of closed time-like curve traversals. This makes
it hard to compute a vacuum energy, but it makes it well-defined, since
there is an absolute flat Minkowski space to gauge here, whereas in or-
thodox QM there is no such thing since the Minkowski metric is not stable
to fluctuations in orthodox QM.
* * *
6.3 Speculations on Chromodynamics
6.3.1 Continuous color rotation?
We know from G 𝐴 that spin is not quantized really, fundamental parti-

cles, if they are literally spinning, will spin freely. What is quantized are
the eigenvalues, but that’s natural, since we are treating the fundamen-
tal geons as irreducible representations of the Poincarë group, so they are
labelled by integer or half-integer spin numbers less than or equal to 2.
The spin number tells us how to operate on the particle’s wave-function
(or field variables) to rotate our laboratory frame onto the rest frame of
the geon. That rest frame if spinning will be the bivector plane of the
spin motion. This transformation is indexed by integer or half integer
spin number, but the geon’s bivector plane is continuous in space. The
bivector remains fixed if the particle is not disturbed, but if it is disturbed
the bivector might re-orient, for example to a change in orientation of an
external magnetic field. But those are continuous operations. It is the
act of measurement which is discrete.
So it occurred to me that maybe the color charge is similar? Perhaps
quark color is like spin number and can change continuously? Or maybe
it is topological, but we still must project onto one of the three colors
when we “measure” color. The reason we observe three colors and their
anti-colors is then topological, and is indexed by a topological index. This
would, I think, help interpreting the Chisholm-Dixon-Baylis-Schmeikal-
Furey (CDBSF) type of algebraic models for chromodynamics.
Now I realize 𝑞𝑔-color is more like charge, rather than spin. So why
this crazy note on continuous change? What I mean by this is not that
color is continuous, it’s clearly not, clearly like charge it is a topological
index. What I mean is that the quarks might constantly rotate in col-
orspace, so change 𝑔 → 𝑏 → 𝑟 → . . ..
That was the idea I had. I think it is wrong! Now I think about it. The
impetus was in worrying about Schmeikal’s scheme, where the algebraic
Ideals of the Clifford basis multivectors determine the colorspace. My
thinking was how the hell can that be topological?
Now my thinking is that the stability within a colorspace has to be
based on some deeper topology that Schmeikal’s schemes does not sur-
face.
Speculations on Chromodynamics 119
Recall Schmeikal’s scheme uses the positive definite subspace of the

spacetime algebra 𝐶ℓ 1,3 (signature (− + ++)) as the colorspace basis. (STA
from hereon will be my abbreviation for spacetime algebra.) The six re-
sultant colorspaces are the six Cartan subalgebras of the positive set,
{1, 𝑒1 , 𝑒2 , 𝑒3 , 𝑒14 , 𝑒24 , 𝑒34 , 𝑒124 , 𝑒134 , 𝑒234 }
as you can verify for yourself, all these square to +1, hence the positive
definitieness.
The other thing is that Schmeikal has colorspace defined only up to
scalar multiples. So the ideals generated say in the first colorspace,
𝜒1 = {1, 𝑒1 , 𝑒24 , 𝑒124 }
can be multiplied by arbitrary scalars and will remain Ideals mod

𝐺𝐿 (1). It is difficult to conceive tat goen topology knows about identity
up to scalar multiple, so in my mind the colorspace is just an algebraic
representaiton of topologicla feautures of spacetime. The topology is
invariant to rescaling by multiples of th identity, and to my mind that’s
why the Clifford STA suffices.
Bibliography
[1] A. Scorpan, The Wild World of 4-Manifolds. Providence, Rhode Is-

land: American Mathematical Society, 2005.
[2] E. Wigner, “On unitary representations of the inhomogeneous
Lorentz group,” Annals of mathematics, vol. 40, pp. 149–204, 1939.
[3] R. Hermann, Lie Groups for Physicists. WA Benjamin, 1966.
[4] V. Bargmann and E. P. Wigner, “Group theoretical discussion of rel-
ativistic wave equations,” in Part I: Particles and Fields. Part II:
Foundations of Quantum Mechanics, pp. 82–94, Springer, 1997.
[5] V. Bargmann, “Irreducible unitary representations of the Lorentz
group,” Annals of Mathematics, vol. 8, no. 3, pp. 568–640, 1947.
[6] N. Straumann, “Unitary representations of the inhomogeneous
Lorentz group and their significance in quantum physics,” in
Springer Handbook of Spacetime, pp. 265–278, Springer, 2014.
[7] B. Schmeikal, “Minimal spin gauge theory — clifford algebra and
quantumchromodynamics,” Advances in Applied Clifford Algebras,
vol. 11, no. 1, pp. 63–80, 2001.
[8] G. Trayling and W. Baylis, “A geometric basis for the standard-model
gauge group,” J. Phys. A, vol. 34, no. 15, pp. 3309–3324, 2001.
[9] B. Schmeikal, “Transposition in Clifford algebra: SU(3) from reori-
entation invariance,” in Clifford Algebras, pp. 351–372, Springer,
2004.
[10] W. Baylis, R. Cabrera, and J. Keselica, “Quantum/classical inter-
face: classical geometric origin of Fermion spin,” Advances in ap-
plied Clifford algebras, vol. 20, no. 3, pp. 517–545, 2010.
121
122 BIBLIOGRAPHY
[11] J. Baez and J. Huerta, “The algebra of grand unified theories,” Bul-
letin of the American Mathematical Society, vol. 47, no. 3, pp. 483–
552, 2010.
[12] G. M. Dixon, “Division Algebras; Spinors; Idempotents; The Alge-

braic Structure of Reality,” 2010. arXiv preprint hep-th/1012.1304.
[13] G. M. Dixon, Division Algebras:: Octonions Quaternions Complex

Numbers and the Algebraic Design of Physics, vol. 290. Springer
Science & Business Media, 2013.
[14] C. Furey, “Unified theory of ideals,” Physical Review D, vol. 86, no. 2,
p. 025024, 2012.
[15] C. Furey, “Generations: three prints, in colour,” Journal of High

Energy Physics, vol. 2014, no. 10, p. 46, 2014.
[16] C. Daviau and J. Bertrand, The standard model of quantum physics

in Clifford algebra. World Scientific, 2015.
[17] C. Furey, “A demonstration that electroweak theory can violate par-

ity automatically (leptonic case),” International Journal of Modern
Physics A, vol. 33, no. 04, p. 1830005, 2018.
[18] S. Aaronson, “Is quantum mechanics an island in theoryspace?,”

in Proc. of the Växjö Conference on Quantum Theory: Reconsider-
ation of Foundations (A. Khrennikov, ed.), 2004. Available as arXiv
preprint quant-ph/0401062.
[19] S. L. Adler, “Quaternionic quantum field theory,” in Quantum Me-

chanics of Fundamental Systems 1, pp. 1–15, Springer, 1988.
[20] S. L. Adler, “Quaternionic quantum mechanics and Noncommuta-

tive dynamics,” arXiv preprint hep-th/9607008, 1996.
[21] D. J. Burger, R. Carballo-Rubio, N. Moynihan, J. Murugan, and

A. Weltman, “Amplitudes for astrophysicists: known knowns,” Gen-
eral Relativity and Gravitation, vol. 50, no. 12, pp. 1–59, 2018.
[22] H. Elvang and Y. tin Huang, “Scattering Amplitudes,” arXiv hep-

th:1308.1697, 2014.
[23] D. O. Hestenes, Space-Time Algebra. Birkh 2 ed., 2015.

Bibliography 123
[24] J. Bourjaily and H. Thomas, “What is the Amplituhedron?,” Not.

Amer. Math. Soc, vol. 65, p. 167, 2018.
[25] N. Arkani-Hamed and J. Trnka, “The amplituhedron,” Journal of

High Energy Physics, vol. 2014, no. 10, p. 30, 2014.
[26] N. Arkani-Hamed and J. Trnka, “Into the amplituhedron,” Journal

of High Energy Physics, vol. 2014, no. 12, pp. 1–32, 2014.
[27] N. Arkani-Hamed, A. Hodges, and J. Trnka, “Positive amplitudes

in the amplituhedron,” Journal of High Energy Physics, vol. 2015,
no. 8, pp. 1–25, 2015.
[28] N. Arkani-Hamed, H. Thomas, and J. Trnka, “Unwinding the am-

plituhedron in binary,” Journal of High Energy Physics, vol. 2018,
no. 1, pp. 1–41, 2018.
[29] N. Arkani-Hamed, C. Langer, A. Y. Srikant, and J. Trnka, “Deep

into the amplituhedron: amplitude singularities at all loops and
legs,” Physical review letters, vol. 122, no. 5, p. 051601, 2019.
[30] N. Arkani-Hamed, J. Henn, and J. Trnka, “Nonperturbative Neg-

ative Geometries: Amplitudes at Strong Coupling and the Ampli-
tuhedron,” arXiv preprint arXiv:2112.06956, 2021.
[31] S. Weinberg, The Quantum Theory of Fields. Vol. 1: Foundations.

Cambridge University Press, 1995.
124 Bibliography
Index
Ads/CFT, 106 spin-2, 113

amplituhedra, 105–113
S matrix, 106–109
gravity Yang-Mills, 113
125
126 INDEX
Appendix A
GNU Free Documentation

License
Version 1.3, 3 November 2008

Copyright © 2000, 2001, 2002, 2007, 2008 Free Software Foundation,
Inc.
https://fsf.org/
Everyone is permitted to copy and distribute verbatim copies of this

license document, but changing it is not allowed.
Preamble
The purpose of this License is to make a manual, textbook, or other

functional and useful document “free” in the sense of freedom: to assure
everyone the effective freedom to copy and redistribute it, with or without
modifying it, either commercially or noncommercially. Secondarily, this
License preserves for the author and publisher a way to get credit for
their work, while not being considered responsible for modifications made
by others.
This License is a kind of “copyleft”, which means that derivative works
of the document must themselves be free in the same sense. It comple-
ments the GNU General Public License, which is a copyleft license de-
signed for free software.
127
128 GNU Free Documentation License
We have designed this License in order to use it for manuals for free
software, because free software needs free documentation: a free pro-
gram should come with manuals providing the same freedoms that the
software does. But this License is not limited to software manuals; it can
be used for any textual work, regardless of subject matter or whether it
is published as a printed book. We recommend this License principally
for works whose purpose is instruction or reference.
1. APPLICABILITY AND DEFINITIONS
This License applies to any manual or other work, in any medium,

that contains a notice placed by the copyright holder saying it can be
distributed under the terms of this License. Such a notice grants a world-
wide, royalty-free license, unlimited in duration, to use that work under
the conditions stated herein. The “Document”, below, refers to any such
manual or work. Any member of the public is a licensee, and is addressed
as “you”. You accept the license if you copy, modify or distribute the work
in a way requiring permission under copyright law.
A “Modified Version” of the Document means any work containing
the Document or a portion of it, either copied verbatim, or with modifi-
cations and/or translated into another language.
A “Secondary Section” is a named appendix or a front-matter sec-
tion of the Document that deals exclusively with the relationship of the
publishers or authors of the Document to the Document’s overall sub-
ject (or to related matters) and contains nothing that could fall directly
within that overall subject. (Thus, if the Document is in part a textbook
of mathematics, a Secondary Section may not explain any mathemat-
ics.) The relationship could be a matter of historical connection with the
subject or with related matters, or of legal, commercial, philosophical,
ethical or political position regarding them.
The “Invariant Sections” are certain Secondary Sections whose ti-
tles are designated, as being those of Invariant Sections, in the notice
that says that the Document is released under this License. If a section
does not fit the above definition of Secondary then it is not allowed to
be designated as Invariant. The Document may contain zero Invariant
Sections. If the Document does not identify any Invariant Sections then
there are none.
GNU Free Documentation License 129
The “Cover Texts” are certain short passages of text that are listed,
as Front-Cover Texts or Back-Cover Texts, in the notice that says that
the Document is released under this License. A Front-Cover Text may
be at most 5 words, and a Back-Cover Text may be at most 25 words.
A “Transparent” copy of the Document means a machine-readable
copy, represented in a format whose specification is available to the gen-
eral public, that is suitable for revising the document straightforwardly
with generic text editors or (for images composed of pixels) generic paint
programs or (for drawings) some widely available drawing editor, and
that is suitable for input to text formatters or for automatic translation
to a variety of formats suitable for input to text formatters. A copy made
in an otherwise Transparent file format whose markup, or absence of
markup, has been arranged to thwart or discourage subsequent modifica-
tion by readers is not Transparent. An image format is not Transparent
if used for any substantial amount of text. A copy that is not “Transpar-
ent” is called “Opaque”.
Examples of suitable formats for Transparent copies include plain
ASCII without markup, Texinfo input format, LaTeX input format,
SGML or XML using a publicly available DTD, and standard-conforming
simple HTML, PostScript or PDF designed for human modification. Ex-
amples of transparent image formats include PNG, XCF and JPG.
Opaque formats include proprietary formats that can be read and edited
only by proprietary word processors, SGML or XML for which the DTD
and/or processing tools are not generally available, and the machine-
generated HTML, PostScript or PDF produced by some word processors
for output purposes only.
The “Title Page” means, for a printed book, the title page itself, plus
such following pages as are needed to hold, legibly, the material this Li-
cense requires to appear in the title page. For works in formats which
do not have any title page as such, “Title Page” means the text near the
most prominent appearance of the work’s title, preceding the beginning
of the body of the text.
The “publisher” means any person or entity that distributes copies
of the Document to the public.
A section “Entitled XYZ” means a named subunit of the Document
whose title either is precisely XYZ or contains XYZ in parentheses follow-
ing text that translates XYZ in another language. (Here XYZ stands for a
specific section name mentioned below, such as “Acknowledgements”,

“Dedications”, “Endorsements”, or “History”.) To “Preserve the Ti-
tle” of such a section when you modify the Document means that it re-
mains a section “Entitled XYZ” according to this definition.
The Document may include Warranty Disclaimers next to the notice
which states that this License applies to the Document. These Warranty
Disclaimers are considered to be included by reference in this License,
but only as regards disclaiming warranties: any other implication that
these Warranty Disclaimers may have is void and has no effect on the
meaning of this License.
2. VERBATIM COPYING
You may copy and distribute the Document in any medium, either com-
mercially or noncommercially, provided that this License, the copyright
notices, and the license notice saying this License applies to the Docu-
ment are reproduced in all copies, and that you add no other conditions
whatsoever to those of this License. You may not use technical measures
to obstruct or control the reading or further copying of the copies you
make or distribute. However, you may accept compensation in exchange
for copies. If you distribute a large enough number of copies you must
also follow the conditions in section 3.
You may also lend copies, under the same conditions stated above, and
you may publicly display copies.
3. COPYING IN QUANTITY
If you publish printed copies (or copies in media that commonly have
printed covers) of the Document, numbering more than 100, and the Doc-
ument’s license notice requires Cover Texts, you must enclose the copies
in covers that carry, clearly and legibly, all these Cover Texts: Front-
Cover Texts on the front cover, and Back-Cover Texts on the back cover.
Both covers must also clearly and legibly identify you as the publisher of
these copies. The front cover must present the full title with all words
of the title equally prominent and visible. You may add other material
on the covers in addition. Copying with changes limited to the covers, as
long as they preserve the title of the Document and satisfy these condi-
tions, can be treated as verbatim copying in other respects.
If the required texts for either cover are too voluminous to fit legibly,
you should put the first ones listed (as many as fit reasonably) on the
actual cover, and continue the rest onto adjacent pages.
If you publish or distribute Opaque copies of the Document numbering
more than 100, you must either include a machine-readable Transpar-
ent copy along with each Opaque copy, or state in or with each Opaque
copy a computer-network location from which the general network-using
public has access to download using public-standard network protocols a
complete Transparent copy of the Document, free of added material. If
you use the latter option, you must take reasonably prudent steps, when
you begin distribution of Opaque copies in quantity, to ensure that this
Transparent copy will remain thus accessible at the stated location un-
til at least one year after the last time you distribute an Opaque copy
(directly or through your agents or retailers) of that edition to the public.
It is requested, but not required, that you contact the authors of the
Document well before redistributing any large number of copies, to give
them a chance to provide you with an updated version of the Document.
4. MODIFICATIONS
You may copy and distribute a Modified Version of the Document un-
der the conditions of sections 2 and 3 above, provided that you release the
Modified Version under precisely this License, with the Modified Version
filling the role of the Document, thus licensing distribution and modifica-
tion of the Modified Version to whoever possesses a copy of it. In addition,
you must do these things in the Modified Version:
A. Use in the Title Page (and on the covers, if any) a title distinct from
that of the Document, and from those of previous versions (which
should, if there were any, be listed in the History section of the
Document). You may use the same title as a previous version if the
original publisher of that version gives permission.
B. List on the Title Page, as authors, one or more persons or entities

responsible for authorship of the modifications in the Modified Ver-
sion, together with at least five of the principal authors of the Doc-
ument (all of its principal authors, if it has fewer than five), unless
they release you from this requirement.
C. State on the Title page the name of the publisher of the Modified
Version, as the publisher.
D. Preserve all the copyright notices of the Document.
E. Add an appropriate copyright notice for your modifications adjacent

to the other copyright notices.
F. Include, immediately after the copyright notices, a license notice

giving the public permission to use the Modified Version under the
terms of this License, in the form shown in the Addendum below.
G. Preserve in that license notice the full lists of Invariant Sections

and required Cover Texts given in the Document’s license notice.
H. Include an unaltered copy of this License.
I. Preserve the section Entitled “History”, Preserve its Title, and add
to it an item stating at least the title, year, new authors, and pub-
lisher of the Modified Version as given on the Title Page. If there is
no section Entitled “History” in the Document, create one stating
the title, year, authors, and publisher of the Document as given on
its Title Page, then add an item describing the Modified Version as
stated in the previous sentence.
J. Preserve the network location, if any, given in the Document for

public access to a Transparent copy of the Document, and likewise
the network locations given in the Document for previous versions
it was based on. These may be placed in the “History” section. You
may omit a network location for a work that was published at least
four years before the Document itself, or if the original publisher of
the version it refers to gives permission.
K. For any section Entitled “Acknowledgements” or “Dedications”, Pre-

serve the Title of the section, and preserve in the section all the
substance and tone of each of the contributor acknowledgements
and/or dedications given therein.
L. Preserve all the Invariant Sections of the Document, unaltered in

their text and in their titles. Section numbers or the equivalent are
not considered part of the section titles.
M. Delete any section Entitled “Endorsements”. Such a section may

not be included in the Modified Version.
N. Do not retitle any existing section to be Entitled “Endorsements” or

to conflict in title with any Invariant Section.
O. Preserve any Warranty Disclaimers.
If the Modified Version includes new front-matter sections or appen-

dices that qualify as Secondary Sections and contain no material copied
from the Document, you may at your option designate some or all of these
sections as invariant. To do this, add their titles to the list of Invariant
Sections in the Modified Version’s license notice. These titles must be
distinct from any other section titles.
You may add a section Entitled “Endorsements”, provided it contains
nothing but endorsements of your Modified Version by various parties—
for example, statements of peer review or that the text has been approved
by an organization as the authoritative definition of a standard.
You may add a passage of up to five words as a Front-Cover Text, and
a passage of up to 25 words as a Back-Cover Text, to the end of the list
of Cover Texts in the Modified Version. Only one passage of Front-Cover
Text and one of Back-Cover Text may be added by (or through arrange-
ments made by) any one entity. If the Document already includes a cover
text for the same cover, previously added by you or by arrangement made
by the same entity you are acting on behalf of, you may not add another;
but you may replace the old one, on explicit permission from the previous
publisher that added the old one.
The author(s) and publisher(s) of the Document do not by this License
give permission to use their names for publicity for or to assert or imply
endorsement of any Modified Version.
5. COMBINING DOCUMENTS
You may combine the Document with other documents released under
this License, under the terms defined in section 4 above for modified ver-
sions, provided that you include in the combination all of the Invariant
Sections of all of the original documents, unmodified, and list them all as
Invariant Sections of your combined work in its license notice, and that
you preserve all their Warranty Disclaimers.
The combined work need only contain one copy of this License, and
multiple identical Invariant Sections may be replaced with a single copy.
If there are multiple Invariant Sections with the same name but different
contents, make the title of each such section unique by adding at the end
of it, in parentheses, the name of the original author or publisher of that
section if known, or else a unique number. Make the same adjustment to
the section titles in the list of Invariant Sections in the license notice of
the combined work.
In the combination, you must combine any sections Entitled “History”
in the various original documents, forming one section Entitled “His-
tory”; likewise combine any sections Entitled “Acknowledgements”, and
any sections Entitled “Dedications”. You must delete all sections Entitled
“Endorsements”.
6. COLLECTIONS OF DOCUMENTS
You may make a collection consisting of the Document and other doc-
uments released under this License, and replace the individual copies of
this License in the various documents with a single copy that is included
in the collection, provided that you follow the rules of this License for
verbatim copying of each of the documents in all other respects.
You may extract a single document from such a collection, and dis-
tribute it individually under this License, provided you insert a copy of
this License into the extracted document, and follow this License in all
other respects regarding verbatim copying of that document.
7. AGGREGATION WITH INDEPENDENT

WORKS
A compilation of the Document or its derivatives with other separate
and independent documents or works, in or on a volume of a storage or
distribution medium, is called an “aggregate” if the copyright resulting

from the compilation is not used to limit the legal rights of the compila-
tion’s users beyond what the individual works permit. When the Docu-
ment is included in an aggregate, this License does not apply to the other
works in the aggregate which are not themselves derivative works of the
Document.
If the Cover Text requirement of section 3 is applicable to these copies
of the Document, then if the Document is less than one half of the en-
tire aggregate, the Document’s Cover Texts may be placed on covers that
bracket the Document within the aggregate, or the electronic equivalent
of covers if the Document is in electronic form. Otherwise they must ap-
pear on printed covers that bracket the whole aggregate.
8. TRANSLATION
Translation is considered a kind of modification, so you may distribute

translations of the Document under the terms of section 4. Replacing In-
variant Sections with translations requires special permission from their
copyright holders, but you may include translations of some or all In-
variant Sections in addition to the original versions of these Invariant
Sections. You may include a translation of this License, and all the li-
cense notices in the Document, and any Warranty Disclaimers, provided
that you also include the original English version of this License and the
original versions of those notices and disclaimers. In case of a disagree-
ment between the translation and the original version of this License or
a notice or disclaimer, the original version will prevail.
If a section in the Document is Entitled “Acknowledgements”, “Ded-
ications”, or “History”, the requirement (section 4) to Preserve its Title
(section 1) will typically require changing the actual title.
9. TERMINATION
You may not copy, modify, sublicense, or distribute the Document ex-
cept as expressly provided under this License. Any attempt otherwise to
copy, modify, sublicense, or distribute it is void, and will automatically
terminate your rights under this License.
However, if you cease all violation of this License, then your license
from a particular copyright holder is reinstated (a) provisionally, unless
and until the copyright holder explicitly and finally terminates your li-
cense, and (b) permanently, if the copyright holder fails to notify you of
the violation by some reasonable means prior to 60 days after the cessa-
tion.
Moreover, your license from a particular copyright holder is reinstated
permanently if the copyright holder notifies you of the violation by some
reasonable means, this is the first time you have received notice of vio-
lation of this License (for any work) from that copyright holder, and you
cure the violation prior to 30 days after your receipt of the notice.
Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License. If your rights have been terminated and not permanently
reinstated, receipt of a copy of some or all of the same material does not
give you any rights to use it.
10. FUTURE REVISIONS OF THIS LICENSE

The Free Software Foundation may publish new, revised ver-
sions of the GNU Free Documentation License from time to time.
Such new versions will be similar in spirit to the present version,
but may differ in detail to address new problems or concerns. See
https://www.gnu.org/licenses/.
Each version of the License is given a distinguishing version number.
If the Document specifies that a particular numbered version of this Li-
cense “or any later version” applies to it, you have the option of following
the terms and conditions either of that specified version or of any later
version that has been published (not as a draft) by the Free Software
Foundation. If the Document does not specify a version number of this
License, you may choose any version ever published (not as a draft) by
the Free Software Foundation. If the Document specifies that a proxy
can decide which future versions of this License can be used, that proxy’s
public statement of acceptance of a version permanently authorizes you
to choose that version for the Document.
11. RELICENSING
“Massive Multiauthor Collaboration Site” (or “MMC Site”) means any

World Wide Web server that publishes copyrightable works and also pro-
vides prominent facilities for anybody to edit those works. A public wiki
that anybody can edit is an example of such a server. A “Massive Multi-
author Collaboration” (or “MMC”) contained in the site means any set of
copyrightable works thus published on the MMC site.
“CC-BY-SA” means the Creative Commons Attribution-Share Alike
3.0 license published by Creative Commons Corporation, a not-for-profit
corporation with a principal place of business in San Francisco, Califor-
nia, as well as future copyleft versions of that license published by that
same organization.
“Incorporate” means to publish or republish a Document, in whole or
in part, as part of another Document.
An MMC is “eligible for relicensing” if it is licensed under this License,
and if all works that were first published under this License somewhere
other than this MMC, and subsequently incorporated in whole or in part
into the MMC, (1) had no cover texts or invariant sections, and (2) were
thus incorporated prior to November 1, 2008.
The operator of an MMC Site may republish an MMC contained in the
site under CC-BY-SA on the same site at any time before August 1, 2009,
provided the MMC is eligible for relicensing.
ADDENDUM: How to use this License for

your documents
To use this License in a document you have written, include a copy of
the License in the document and put the following copyright and license
notices just after the title page:
Copyright © YEAR YOUR NAME. Permission is granted to

copy, distribute and/or modify this document under the terms
of the GNU Free Documentation License, Version 1.3 or any
later version published by the Free Software Foundation; with
no Invariant Sections, no Front-Cover Texts, and no Back-
Cover Texts. A copy of the license is included in the section
entitled “GNU Free Documentation License”.
If you have Invariant Sections, Front-Cover Texts and Back-Cover

Texts, replace the “with . . . Texts.” line with this:
with the Invariant Sections being LIST THEIR TITLES, with

the Front-Cover Texts being LIST, and with the Back-Cover
Texts being LIST.
If you have Invariant Sections without Cover Texts, or some other com-
bination of the three, merge those two alternatives to suit the situation.
If your document contains non-trivial examples of program code, we
recommend releasing these examples in parallel under your choice of free
software license, such as the GNU General Public License, to permit
their use in free software.

A Quantum Mechanics Tourist

Uploaded by

Copyright:

Available Formats

You might also like

A Quantum Mechanics Tourist

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Quantum Mechanics Tourist

Uploaded by

Copyright:

Available Formats

Quantum Mechanics Tourist

January 26, 2022

Copyright © 2021 Bijou Murray Smith.

Permission is granted to copy, distribute and/or modify this

2 Particles and Symmetries 5

2.4 Poincaré algebra . . . . . . . . . . . . . . . . . . . . . . . . . 46

3 Scattering and Forces 61

5 Kinematics and Amplituhedra 105

6 Speculative Ideas 115

A GNU Free Documentation License 127

I am a migrant worker in the landscape of physics academia. I like writ-

• chapters do not build like a story from previous chapters, so there

Note on typography. I am a big believer in the social goodness of shar-

𝑑 = 4 are the diffeomorphisms not the same as the topological homeomor-

1.1 The One and Only True Interpretation of

not so critical in classical mechanics).

Particles and Symmetries

This is about the question — and this is a very ancient ques-

1. R. Hermann, Lie Groups for Physicists [3].

2. E. Wigner, On unitary representations of the inhomogeneous Lorentz

3. V. Bargmann and E. Wigner, Group theoretical discussion of rela-

4. V. Bargmann, Irreducible unitary representations of the Lorentz

5. N. Straumann, Unitary representations of the inhomogeneous

2.1 Introduction to particles and invariants

Follow-up questions. Does Q.1 mean quanta are not applicable in

More on Q.2 — spacetime symmetry. We have not explained why

to the other orthogonal directions where 𝑐 = 4/3. We do not observe any

On Q.2 — irreducibility. I think it is important to note that Nature

On Q.3 — why unitary representation? Why indeed? Well, we need

they argue that extra space dimensions are required, or predicted, by

1. If we desire positive-definite norms for particle probability density

2. We will find invariants, known as Casimir operators, of the group

2.1.1 Motivation — symmetries of the 𝑆-matrix

• Use the above philosophy to justify searching for unitary irreducible

mathematical structures we find (the irreducible representations)

2.2 Group theory preliminaries

It is kind of depressing to realize if you want to be a great theoretical

at mathematical rigour, or something in-between. It just takes too much

2.2.1 Quantum mechanics preliminaries

My overarching goal is to explore and promote Geon theory, which seeks

As per the usual postulates of QM;

1. Operators acting on the Hilbert space of states represent physical

2. In highly abstract formalism, we pretend there are experiments we

That mathematical statement of linear algebra is our model for

3. (Born Rule.) The probability of a transition from a state |𝜓𝑖 i to a

4. A QM system can exist in a superposition of states, which just

5. (Measurement postulate.) The probability of obtaining the particu-

Comment. There is no god-given reason why any of these postulates

Comment. I do not know what to call postulate 2 other than pretend-

Comment on postulate 4. Most textbooks and academics use complex

Comment. It is possible to prove that a 2-norm as in the Born Rule

Comment on postulates 4 and 2. These, to my mind, are the only

we conventionally just assume we can make this disturbance as small as

6. (Particle postulate.) There exist states of physical systems for which

It is worth noting general relativity already has this quantum postu-

So everything of physical interest is preserved by such a transforma-

2.2.2 Symmetry representation theorem

Theorem 2.2 (Symmetry representation). Any symmetry transform on

2.2.8 The isomorphism SU(2) Spin(3)

2.2.9 The isomorphism SL(2, ℂ) Spin+ (1, 3)