0% found this document useful (0 votes)
24 views65 pages

Quantum Physics: Time and Truth Essays

The document presents three essays by Dick Hoekzema on Quantum Physics, Time, and Truth, exploring the conceptual foundations of quantum theory and its treatment of time. It introduces Quantum Event Theory (QET), which aims to establish a fully covariant quantum physics framework, contrasting it with traditional quantum field theory. The essays delve into speculative ideas about the dimensionality of spacetime, the nature of particles, and the implications of time symmetry in quantum mechanics.

Uploaded by

nidip42109
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Topics covered

  • Statistical Physics,
  • Information Theory,
  • EPR Paradox,
  • Philosophy of Science,
  • Quantum Event Theory,
  • Popper's Falsifiability,
  • Feynman Diagrams,
  • Quantum Mechanics,
  • Contextual Knowledge,
  • Causal Relationships
0% found this document useful (0 votes)
24 views65 pages

Quantum Physics: Time and Truth Essays

The document presents three essays by Dick Hoekzema on Quantum Physics, Time, and Truth, exploring the conceptual foundations of quantum theory and its treatment of time. It introduces Quantum Event Theory (QET), which aims to establish a fully covariant quantum physics framework, contrasting it with traditional quantum field theory. The essays delve into speculative ideas about the dimensionality of spacetime, the nature of particles, and the implications of time symmetry in quantum mechanics.

Uploaded by

nidip42109
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Topics covered

  • Statistical Physics,
  • Information Theory,
  • EPR Paradox,
  • Philosophy of Science,
  • Quantum Event Theory,
  • Popper's Falsifiability,
  • Feynman Diagrams,
  • Quantum Mechanics,
  • Contextual Knowledge,
  • Causal Relationships

See discussions, stats, and author profiles for this publication at: [Link]

net/publication/342521952

Three essays on Quantum Physics, Time, and Truth

Preprint · June 2020

CITATIONS READS

0 2,057

1 author:

Dick Hoekzema
Utrecht University
53 PUBLICATIONS 103 CITATIONS

SEE PROFILE

All content following this page was uploaded by Dick Hoekzema on 29 June 2020.

The user has requested enhancement of the downloaded file.


Three essays on
Quantum Physics, Time, and Truth
Dick Hoekzema
June 29, 2020

Introduction

With a bit of give and take, this year (2020) quantum physics is 100 years old.
The year 2000 was well chosen for a celebration, because of the introduction, in
1900, of Planck’s constant h and the expression E = hν for the energy of light
quanta. This was still decades away, however, from the formulation of more
lasting theoretical structures like Schrödingers wave equation (1926), matrix
mechanics, by Werner Heisenberg, Max Born, and Pascual Jordan (1925), or the
introduction of the Copenhagen interpretation by Bohr and Heisenberg ( 1927).
Assuming a sort of skewd bell shaped curve for the most important contributions
to its origins, the early 2020’s seem a suitable period for looking back at a
century of quantum physics.
For me personally, looking back roughly 1/3 of that time, about thirty years
ago, I was engaged in research about the way in which time is handled in
relativistic quantum physics. A couple of years earlier, the peculiar conceptual
structure of quantum physics had made me decide to finish my graduation at the
then newly founded institute for the foundations of physics at Utrecht university,
with a major program including physics and philosophy of science, to which I
added a minor in theoretical physics.
My first course on relativistic quantum theory and quantum field theory in-
cluded a major disappointment. I had fully anticipated that such a course would
start by redefining the structure of the Hilbert space, making sure that space
and time are treated in equivalent ways, resulting in a theory where important
expressions could (and would) all be written in explicitly covariant formulation.
Neither expectation was satisfied even in a remote sense. Hilbert spaces were
still spanned by vectors of type | x > (= | x1 , x2 , x3 >), lacking one important
dimension of spacetime. As a direct consequence it still remained impossible to
even define a time operator on such a Hilbert space.
Switching to quantum field theory (QFT), it is possible to state field equa-
tions in an entirely covariant manner, even without specifying a specific type
of Hilbert space on which the fields are supposed to act at all. Nevertheless,
the role of time still differs from that of space, because it is necessary to also
formulate equal time commutation relations. Moreover, the use of commutation
relations itself expresses that the fields are still defined as operators on, presum-

1
ably, a Hilbert space. Unless it can be shown that this is a relativistically well
defined Hilbert space, with equivalent roles for time and space, what progress
is achieved at all?
I decided to make it a research subject to design a fully covariant version
of Hilbert space quantum physics, and I started immediately with working in
this direction. After graduation, there was sufficient progress to start writing a
research proposal for a Ph.D. project. To my delight, as well as surprise among
colleagues, it was readily awarded. A suggestion, that, with the money available,
it was now time to start looking for a real subject, was joyfully ignored, and I
set out to continue my research. What good is a university position if you can’t
do why you wanted it for?
For a start, I wanted Hilbert spaces based on 4-vectors, i.e., like
Z
|ψ  = d4 x ψ(x0 , x1 , x2 , x3 )|x0 , x1 , x2 , x3 
Z
= d4 p ψ̂(p0 , p1 , p2 , p3 )|p0 , p1 p2 , p3 

and with inproducts like ≺ x | p = exp(ipx)/(2π)4/2 . At first sight, momen-


tum vectors of this type seem unacceptable, because they do not automatically
obey p2 − m2 = 0. This is not a real objection, however. It merely means that
there must be some mechanism such that some objects (particles) will obey laws
that enforce the validity of this equation (either fully or approximately).
Thus, after a lot of reading and similar amounts of thinking and writing,
results included a quantum theory based on a Hilbert space of the above type,
and written entirely in explicitely covariant expressions. It is called Quantum
Event Theory (QET), for the obvious reason that the coordinates (t, x, y, z)
refer to events in spacetime. Likewise, one might say that, in the momentum
representation, vectors |pt , px , py , pz  refer to events in the 4-momentum space.
In the position representation, of course, they represent wavelike solutions, and
describe entities we may identify as particles.
Although equivalent with respect to empirical content, at a conceptual level
there is a world of difference between QET and QFT, or other more conven-
tional types of quantum physics. For instance, in QET the uncertainty relation
between time and the timelike component of momentum has precisely the same
status as for the spacelike components of position and momentum, and can be
derived in the same manner, from a Hilbert space argument.
Further, a very significant difference is that probability is assigned in a very
different manner. There is no state reduction, no wave function collapse or
anything similar. Probabilities are assigned to histories rather than states. In
fact, this is not altogether very strange, because it is conceptually quite close to
how phenomena, like e.g. the Lamb shift, are calculated in QFT, i.e., by adding
a series of Feynman diagrams. Each diagram represents a possible history, so
the Lamb shift is actually the result of a superposition of histories.
In quantum physics, histories are typically represented by Green functions.
In QET, these are obtained, directly and in a fully covariant way, as matrix ele-
ments of a Green operator (instead of, e.g., taking a detour via, e.g., time ordered
products of fields). Moreover, probabilities are obtained, in an equally direct
manner, by assuming that they are proportional to Green functions squared.
From this, it can then be shown that the resulting probability measure pro-

2
duces a generalization of the time symmetric probability measure of Aharonov,
Bergmann and Lebowitz (ABL).1
Furthermore, with respect to timesymmetry, QET is very similar to timesym-
metric electromagnetism, as explored by Tetrode2 , Fokker3 , and Schwarzschild4 .
Either from this, or directly from the ABL timesymmetric quantum probability
measure, the further conclusion can now be drawn that the observed time asym-
metry must be related to a condition called complete absorption in the future.
Complete absorption is considered highly unlikely, but I beg to differ. I see it
as practically inescapable. Finding out how complete absorption is satisfied in
our universe is then one of the more interesting contemporary issues, closely
followed by the question whether we even have a definite past.
Yet another issue concerns the dimensionality of spacetime. Extra dimen-
sions, introduced for explaining the observed spectrum of elementary particles,
are often assumed to be spacelike, although no good reason for excluding time-
like dimensions seems available. Also, if there are extra dimensions, one should
expect that they occur in a generalized version of general relativity, so where
are they?
In view of its neat structure, and because of the conceptual differences it
generates as compared to more standard views in quantum physics, my opinion is
that QET deserves a place in the spectrum of methods investigated in relativistic
quantum physics. The more painful remark I need to make, however, is that
I have yet to find anyone to agree with this. Careerwise, the colleague who
suggested that, with the money available, it was time to start looking for a
serious subject, had been quite right. I am still glad, however, that I did what I
set out to do, and, bearing the consequences, after my university contract ended
I took up a career in High School teaching. I received my Ph.D. only a couple
of years later, after writing a book5 about quantum physics and philosophy of
science.
Teaching turned into a fulfilling career, mainly because of the contact with
colleagues and students, but certainly also by being given the opportunity of
chairing a curriculum renewal project on modern physics. The project ran suc-
cessfully at a growing number of test schools, until exterminated by a much
larger curriculum reform, making obtained results largely irrelevant. A surviv-
ing offspring, however, is a travel agency6 , taking, over the years, many teach-
ers and thousands of high school students to places like CERN, DESY, JET,
ILL/ESRF... So, over the years, I saw ATLAS and CMS steadily grow, disap-
pear underground, and start producing data, together with subsequent groups
of high school kids, but without ever having to participate in any of the hard
work.
In spite of all the to do’s of teaching, however, I never managed to get
entirely free of thinking about fundamental physics and philosophy of science.
I held on to the idea that, after retiring, there was still a job to finish. Hence
this document.
Now with respect to this document, clearly my ideas did not precisely result
1 Y. Aharonov, P.G. Bergmann, J.L. Lebowitz, [Link]., B134, 1410 (1964)
2 M. Tetrode, Z. Phys., 10, 317 (1922)
3 A.D. Fokker, .. Z. Phys., 58, 386 (1929)
4 K. Schwarzschild,Göttinger Nachrichten, 128, 132 (1903)
5 Dick Hoekzema: The Quantum Labyrinth. Kluwer, Dordrecht (1993)
6 Nowadays continued as an official activity of the Dutch physical society NNV.

3
in much of a research career, and, so far, preliminary versions of the ideas
behind it also failed to raise any positive comments at all. Considering the
circumstances, I am not really that eager to read myself entirely up to date, nor
even to start rereading a vast amount of older literature used in my previous
research. So what is the use of even trying to publish?
Frankly, I am a bit beyond caring. As the uselessness of this enterprise
seems to become more and more compelling, what is left of it is turning into,
first of all, an attempt to get rid of a compulsive addiction to thinking about
these matters. Simply stopping is not an option, but, whatever, a document on
my researchgate account, a copy in the archives, ... , enough is enough. Never
mind, I’ll be good, and pleasant about not publishing is that it yields unlimited
freedom in writing.
And after this? Well, I bought a new set of paints, spanned a few canvases,
and hope to paint away all traces of a physics addiction. Come to think of
it, painting has always been also my favourite way of looking at physics: as
creating a huge gallery of ideas about nature. Art, however, is not just a world
of finished paintings and other creations but also a playground for rough sketches
and wild ideas. This is certainly a thing to keep in mind, as a background for
the structure and content of this bundle. It is particularly evident in the second
essay, offering crude thought, in an almost embryonic phase of the process. Pure
speculation, perhaps, written entirely for pleasure, both cute and cutely naive
I would say, but it at least made delicious writing.

About the essays.

Clearly, the first essay must be, and is, an introduction to QET. One thing
that may still be worth mentioning in advance, is the role played by Wheeler’s
geometrodynamics, which exemplified, for me, a clear instance of how things
went wrong. Gravitation7 , and his conception of delayed choice experiments,
had made Wheeler one of my heroes.8 Geometrodynamics, however, bringing
the notion of superpositions of 3D-geometries of space as a function of time,
was a major set back. I cannot but disagree quite basically with his view. The
fact that someone so brilliantly aware of relativity theory, could come up with
a theory where space and time were treated so differently, added considerably
to my idea that something had to be very wrong with the general conception of
time in quantum physics. In my view, a theory like this can only be acceptable
by being a dimensionally reduced form9 of a more complete theory, that one
would desire to be formulated first.
The danger of retaining a formalism in which space and time are treated as
conceptually different, is that it reverberates with conceptual differences expe-
rienced in daily life. It seems true, of course, that there is something definitely
counter intuitive about treating space and time on the same footing. Anyway,
treating space and time on a really equivalent footing is certainly not standing
practice in current physics, and, of course, it really is difficult to withstand daily
intuitions about time. Intuitively, for instance, what happens now cannot de-
pend on future events, because the future doesn’t exist yet. In spite of relativity
7 Misner, Thorne, Wheeler;Gravitation; Freeman (1973)
8 OK, he still is, but...
9 See section (1.1.1) about dimensional reduction.

4
theory, such intuitive ideas may well continue to affect the development of ideas
in physics.10
Now there may be another side to this discussion. Suppose that the condition
of complete absorption is almost, but not entirely satisfied. Tiny variations
could still make it possible to establish some form of information exchange
between future and present. Now this is a subject with a Sting.11 Even in a
very limited form the strategic importance could be huge. So far, no clear effects
have been found, at least not in publicly accessible accounts of experiments.
Perhaps we should be quite happy for it to remain this way.

The second essay contains three sections, all revolving about time and space-
time. As warned above, the nature of the essay is highly speculative.
The first section starts from the quite common idea of a relation between
the dimensionality of spacetime and the spectrum of elementary particles. A
presently unsolved problem in the standard model of particle theory is the sta-
bility of protons, which should decay, very slowly, into leptons. Only, they fail
to obey. The suggestion made in the essay is that quark color is a spinlike
property in three dimensions, leptonness is a fourth color, and quarks and lep-
tons cannot transform into one another because they are separated by a Lorentz
transformation.
In particular, the speculation continues, they are part of a 4-vector time-
space representation, with a spacelike lepton and three timelike quarks. This,
then, forms the basis for a model of the universe: macroscopic spacetime (three
macroscopic space and one macroscopic time dimension), plus timespace (one
microscopic space and three microscopic time dimensions). We shall call this
the S3 T1 ⊗ s1 t3 model. Especially the idea that the particle spectrum provides
a quite direct imprint of the dimensionalities of space and time, in this form
or another, seems so natural, that it should require explicit argumentation for
assuming it to be wrong.
Alongside the spacetime × timespace assumption, the model uses yet an-
other, let us say, frivolous idea, more or less independent of it. It concerns the
role of the additional dimensions with respect to the dynamics of spacetime. If
microscopic dimensions exist, they must also be included in any theory of the
totality of space and time, i.e., in particular also as a generalization of General
Relativity. The essay presents some very rough suggestions in this direction.
The second section of the second essay addresses yet another important
theme, concerning subjects like EPR, and time symmetry versus asymmetry.
From the EPR thought experiment, and ensuing real experiments involving the
Bell equations, the conclusion, or at the very least a very strong suggestion, can
be drawn that causality must be basically time symmetric. Moreover, quantum
eraser experiments, delayed choice experiments and delayed choice quantum
eraser experiments lead to a similar conclusion. The time symmetric nature of
causality also follows directly from the time symmetric character of the proba-
bility measure in QET, with the direct consequence that phenomena like EPR
correlations cease to be enigmatic.
10 See, e.g.,The Singular Universe; Roberto Unger and Lee Smolin, Cambridge U.P.; 2015,

where they argue for the existence of a ‘real time’, including the idea of fundamental theories
of physics being time dependent, and for a science built on ‘natural philosophy’.
11 George Roy Hill, Universal Pictures (1973)

5
This, however, also leads quite directly to a design for a retrophone: a device
for communicating with someone in the future. It is the actually observed
impossibility of such a device, and more generally the observed time asymmetry
of causality, that is now in need of explanation. From the probability measure
of Aharonov, Bergmann and Lebowitz, or from the more general probability
measure in QET, it follows that complete absorption in the future universe is
a necessary and sufficient condition. This entails that the state of the future
universe must be a superposition and/or a mixture of many different futures
(which, by the way, also puts into question how unique our past really is).
Accordingly, the third section of the second essay is an investigation into
possible explanations for complete absorption.

The third essay is about the structure of science. A view of science as a quest
for the ultimate truth about the world is, I would say, a naive and in essence
also abhorrent idea. But then what?
Imagine a world with only a single species of plants. We may imagine cir-
cumstances where, indeed, only a single species survives, but is this the ultimate
aim of evolution? Maybe it can be realized locally, on an isolated mountain top,
perhaps. It would make mountain top botany a lot easier, but for the rest?
Does this lead to an ultimate, true, species? Maybe, on the next mountain top,
a different species survives. And, anyway, although at mountain peaks, in bar-
ren circumstances and freezing cold, life may reach to great heights, it spreads
through the valleys, where, in wild and beautiful landscapes with lush wildlife
and vegetation, there is also competition, strife, and chaos.
My time at university was roughly in the heydays of discussions about sci-
entific truth, the structure of theories, paradigms, research programmes, incom-
mensurabiliy, ... In the third essay of this bundle, I discuss a view on logical
structure, while using parallels with the structure and evolution of life.

Finally, both the overall lack of reaction, and the few reactions received so far,
have made me feel rather foolish, or even kind of ridiculous. Now I don’t mind
feeling foolish, as long as I am right. Therefore, perhaps unwise but still difficult
to avoid, a quick personal assessment of how right or wrong I might be.
With respect to the first essay, the fact that quantum mechanics, in at least
most versions and during most of the time, used a theoretical structure that
is conceptually at odds with the basic structure of relativity theory, is quite
clearly visible and hard to deny, I would say. So far, it may not have affected
the actual empirical content of the theory, but effects on the development of
quantum theory, either so far or in the future, may be quite difficult to trace.
The second essay is frivolous in design, but it does, nevertheless, contain a
few serious points. Firstly, the idea that the particle spectrum could (should)
contain a direct imprint of the dimensionalities of space and time, is an idea that
seems so obvious that it would require explicit explanation for being incorrect.
Secondly, from the basic time symmetry of causality, implicated by the EPR
correlations and similar effects, and from the basic time symmetry explicit in
QET, it follows that the observed time asymmetry must be due to the condition
of complete absorption in the future universe. Thinking about how this can be
is, therefore, an inescapable subject in quantum cosmology.

6
Anyway, with respect to frivolity, how frivolous are fancy ideas like super-
symmetry or string theory? Both stem from beautifully wild ideas, but for-
give me my impression that they are both transforming into an advanced level
mathematics contest, with an intended but ever feebler relation to fundamental
physics. At least with respect to string theory, my impression is that slogans
like EPR= ER and GR = QM are missing a few points. I will elaborate a bit
on this in the course of the second essay.
By the way, don’t understand me wrong with respect to frivolity; it is a vital
component of thought, if properly restrained at later stages. In the second essay,
especially the final section is exceptionally frivolous, and left there mainly for fun
and inspiration. I would also say that workshops in creative and frivolous think-
ing, and making quick scetches of ideas without tearing them down immediately,
might make a fun and surprisingly useful addition to a physics curriculum.
The third essay is, to a large extent, an attempt to provide logical structure
for investigating the concept of incommensurability. As such, it is also a warning
against a conception of science as a quest for finding the (in principle single and
universal) truth. This latter conception I would consider as an unfortunate
leftover from monotheism, clouding our view on the true (???) structure of the
scientific enterprise.
Looking back at the essays, what are they: physics and philosophy, or phan-
tasy? Frankly, I am not in a position to decide or to care, and frankly, I don’t.
I tried to sort of let the stories tell themselves, and so they did. I thought,
I read, I wrote, and at several places I am nothing less than surprised about
the results myself. But now it is phynished, and I’m phfryyyyeyyyeeyyeeeyeeee.
Physics was quite an addiction, but I intend to go cold turkey. Please wish me
success and good luck.
And how do I, or should I, deal with respect to being foolish? Maybe it is OK
te be just a bit lenient on myself, maybe slightly silly is enough...? Whatever
may be, it is a delight not to have to judge about such things myself, and what
the heck, I’m getting out anyway:

7
1 THE CONCEPTUAL STRUCTURE
OF RELATIVISTIC QUANTUM PHYSICS

As is well known, Hilbert spaces of common versions of quantum theory do have


position operators for all directions of space, but not for time, as seems to follow
from an argument of Wolfgang Pauli12 against their existence, based on the spectral
properties of the Hamiltonian. This omission produces an obvious violation of the
relativistic law on equal treatment of time and space.
In this paper, it will be argued that the nonexistence proof can be evaded in a
very natural way, and that the most straightforward way of adding a time operator
works without problems. As a consequence, a manifestly covariant version of quantum
theory can be formulated, also at the Hilbert space level.
As an elegant side effect, one may also add as many dimensions, space and/or
time, as desired, including corresponding position operators. This may, for instance,
have implications for particle theory, where extra dimensions are usually assumed to
be spacelike.

1.1 Quantum spacetime


1.1.1 Position-operators for spacetimes
For a simple type Hilbert space in quantum physics, we may start from the space
L2 (R3 ) of squared integrable complex functions on R3 . This Hilbert space is
spanned by vectors | x >, and has a standard inproduct given by

< x | y >= δ3 (x − y), with x, y ∈ R3 (1)

Evidently, for an effectively two dimensional or one dimensional system we would


simply take L2 (R2 ) or L2 (R).
Assuming the existence of a fourth dimenson of space, the natural way to
proceed would be, quite straightforwardly, to replace L2 (R3 ) by L2 (R4 ), the
space of quadratically integrable functions on L2 (R4 ). The standard inproduct
thus changes into

≺ x | y = δ4 (x − y), with x, y ∈ R4 . (2)


Whether the added dimension is space- or timelike, however, should be irrele-
vant. There is no reason to assume that the metric properties of spacetime bear
any direct relation to the structure of the Hilbert space.
Nicely, this makes the procedure fully generic: dimensions of either type
can be added as needed (with particle physics as the most likely source of
need). A convenient notation for a spacetime with m dimensions of time and n
spacedimensions is x = (x−m , ..., x−1 , x1 , ..., xn ), which then produces a Hilbert
space of vectors:

|x = |x−m , ..., x−1 , x1 , ..., xn  such that: ≺ x | y = δm+n (x − y)


12 in particular Pauli, W. (1933), Die allgemeine Prinzipien der Wellenmechanik,

Handbuch der Physik, 2. Auflage, Band 24, 1. Teil (Springer-Verlag, Berlin), 83-272.

8
The new Hilbert space has position operators for both space and time, and
projection operators for whatever spacetime volumes one might wish to define
them for. So far, this looks fairly trivial. The interesting question is why
in actual quantum physics it doesn’t work this way, and the more interesting
question is whether it can be made to work this way.
It is fairly obvious that this should be possible, and, indeed, the answer is
affirmative, but it still took me quite some time and effort to get it done. The
result is a form of quantum physics I named Quantum Event Theory (QET),
for obvious reasons, and, as we shall see, it connects to more standard forms
of quantum physics in fairly straightforward, but certainly not entirely trivial
ways. However, contrary to other formulations of quantum physics, space and
time dimensions do play fully equivalent roles, and QET can be stated in a fully
relativistically covariant form.
With respect to Pauli’s argument against the existence of a time operator in
standard quantum physics, this rests on the assumption that the Hamiltonian
is the generator of time translations. The existence of a time operator requires
a generator with an unbounded continuous spectrum, however, which cannot
hold for the Hamiltonian.
In QET, on the other hand, the generator of time translations is not the
Hamiltonian, but the operator i∂0 (assuming natural units, with ~ = c = 1).
It is (sort of) replaced by the Hamiltonian in a process of dynamic reduction,
which kills the time dimension, and the covariant structure of QET, and which
essentially reduces QET to standard quantum physics.

1.1.2 Quantum Event Theory


It is quite evident that, by including time in the definition of the Hilbert space,
the resulting Hilbert space vectors are no longer adequate for describing the
dynamic behaviour of particles, or other quantum systems evolving in time,
because this requires time as a dynamic variable.
There are two ways to solve this problem. One way is to add an extra variable
τ that can take the role of dynamic variable. This produces Hilbert space kets
of type | x(τ ) >, and provides a way of quantizing worldlines13 . The other way
is to drop the dynamical approach, observe that a spacetime point x has the
dimensions of an event, and develop quantum physics as a theory for describing
correlations between events. Hence the name Quantum Event Theory.
In QET, quantum physics is construed as an (m + n)-dimensional static
image of the world, and the basic assumptions for obtaining this image are
that:
Green functions play the key role in correlating events,
Green functions are obtained as matrix elements of a Green operator.
The intention here is to show that this leads to an elegant, powerful and flexible
formalism, with clear relations to standard results. The latter hinges on the fact
that Green functions play a crucial role also in standard Quantum Field Theory
(QFT). In QFT, a Green function is essentially a way of deriving covariant
complex valued quantities from a, frame dependent expression, i.e., the unitary
evolution operator. In QET, the Green operator is a covariant operator valued
13 E.C.G Stueckelberg, Helvetia Physica Acta, 14:588-594

9
function, delivering the same quantities directly, as matrix elements. Increased
elegance is not the only benefit from this transformation. More importantly,
the roles of time and space are entirely equivalent, as one would expect in a
relativistic theory, and this involves a conceptual change in various ways. For
instance, the uncertainty relation between time and energy now follows from a
Hilbert space argument, in exactly the same manner as the uncertainty relations
for position and momentum. Another consequence is that the dimensionality
of time is no longer fixed. Dimensions of time can be introduced as easily as
space dimensions, with potential relevance in particle theory, for explaining the
observed particle spectrum.

1.1.3 Dimensional reduction 1:


from static (m + n) to dynamic (m + n − 1)
Any static (m + n)-dimensional picture can be transformed into a dynamic
picture by means of a technique that involves dimensional reduction. The idea
is to define a family of (m+n−1) - dimensional hypersurfaces, characterized by a
variable λ, in such manner that, by varying λ through its entire range of possible
values, every point in the (m + n)-dimensional picture is included in precisely
one hypersurface, with a unique value of λ. In this way, varying λ through its
entire range of values, the entire physical content of the (m + n)-dimensional
picture is reproduced as a dynamic process in an (m + n − 1)-dimensional world.
Aside from obvious requirements, it is, in principle, irrelevant how the hy-
persurfaces are chosen. Evidently, however, some choices may produce simpler
dynamics than others, and there is, of course, a long standing tradition of pre-
ferring time itself or at least timelike dynamic variables.
For the time being, however, we shall rest this point and return to standard
(1+3) spacetime. At this point it becomes enlightening to study how the method
of dimensional reduction works at a transition from a static (1 + 3)-dimensional
QET-picture to a standard dynamic picture of quantum mechanics, with time
as the dynamic variable over a 3-dimensional space. After just a few preliminary
steps, this is our next target.

1.1.4 The Fourier transformed (momentum) representation


Green functions play a crucial role in QET, just like they do in standard quan-
tum field theory (QFT). Anticipating more theory about how to handle Green
functions, we can use basic knowledge from QFT to assume that they consist
of vertices, connected by propagators. In QET, vertices are small clusters of
events, and propagators are about as close as we can get to the concept of a
particle.
Describing propagators requires introducing a Fourier transformed basis in
Hilbert space. In analogy with standard quantum physics, it will be called
the momentum representation, in spite of the fact that events, as we usually
understand them, do not carry momentum. This produces:

dN x
Z
|p  = exp (−ipµ xµ )|x  (3)
(2π)N/2
and ≺ x | p  = (2π)−N/2 exp (−ipµ xµ )

10
where N is the total number of dimensions of spacetime, i.e.: N = m + n. The
expression pµ xµ means, as usual:

pµ xµ = gµν pµ xν

so at this point the metric tensor of spacetime gets to play a role. Also, from
(3) it follows that the expression pµ can be defined as a differential operator,
such that:
≺ x|pµ |p  = i∂µ ≺ x | p = pµ ≺ x | p 
Reminder:
The word momentum is used because of the resemblance to the standard quan-
tum physics, and at this point it bears no relation (yet) to the mechanical
concept of momentum. Events do not carry momentum; they merely have a
spacetime position. It is also clear, that none of the components of pµ , in the
way in which it is used here, can have a confined spectrum, e.g. because this
would contradict the equality
Z
dN p ≺ x | p ≺ p | y =≺ x | y 

All this changes, however, once we start looking at propagators, which have
more particlelike properties.

1.1.5 Particles, propagators, and dimensional reduction


The simplest form of a propagator in QET is the operator :
1
K= (4)
p2 − m2 (+i)
Clearly, this K commutes with the operator p defined above.
The operator iK is also about the simplest practically useful Green operator
one may define, and it serves well for investigating a simple version of the relation
between a static picture and a dimensionally reduced dynamic picture.
If iK is sandwiched between two moments in time t0 and t, with t > t0 , the
effect is that the zero point of the denominator (in combination with choosing
 ↓ 0 for fixing the position of the singularity) produces a factor δ(p0 − E). This
effectively removes one dimension, the time dimension, as a degree of freedom
of the momentum vector.
The fact that the propagator commutes with the p0 component of the mo-
mentum operator, means that it has a lasting existence in time, which is com-
patible with the concept of a particle. What speaks against this, however, is
that it also commutes with the spacelike components of p, which means that it
is not confined in space. This suggests a more fieldlike character.
What comes closest to observing a particle, is a sequence of events as seen in
a particle tracker, with a bubble chamber as the most tangible example. This
defines the particle as an alteration of fieldlike expansions with a more or well
defined momentum alternated by more or less localized events.
With these reservations in mind, we shall look at the propagator as describing
the dynamics of a particle.
Starting from the matrix element ≺ φ,t|iK|ψ0 , t0  let us interpret time as a
dynamic variable and see how this leads to a picture in which the particle is seen

11
to live in a reduced Hilbertspace Hred (R3 ), spanned by vectors | p >, where the
quantities p are interpreted as momentum vectors in ordinary space. Taking t to
be a variable this defines a dynamics on the dimensionally reduced Hilbertspace.
Caught in formulas, the net effect can be summarized in the following manner:

Starting from the matrix element


i
≺ φ,t|iK|ψ0 , t0  = ≺ φ, t| |ψ0 , t0 
p2 − m2 + i
1 ie−ip0 (t−t0 )
= ≺ φ, p0 | 2 |ψ0 , p0 
2π p − m2 + i
1 ie−ip0 (t−t0 )
= ≺ φ, p0 | |ψ0 , p0  (5)
2π (p0 − E)(p0 + E) + i
assuming t > t0 and calculating the limit  ↓ 0 yields:

1 −2πi δ(p0 − E) × ie−ip0 (t−t0 )


... = ≺ φ, p0 | |ψ0 , p0 
2π p0 + E
e−iE(t−t0 )
= ≺ φ, E| |ψ0 , E  (6)
2E
For a transition to the reduced Hilbert space Hred , defining

1 e−iE(t−t0 )
< φ | =≺ φ, E| √ and | ψ(t) >= √ |ψ, E 
2E 2E
results in
| ψ(t) >= e−iE(t−t0 ) | ψ 0 > (7)
and
< φ | ψ(t) >=≺ φ, t|iK|ψ0 , t0  (8)
The conclusion is that a 4-dimensional static image, as produced by QET, is
converted into a 3-dimensional dynamic image, by means of a dimensional re-
duction, which removes the time dimension and converts time into a dynamic
parameter. A somewhat more extensive and more general discussion of dimen-
sional reduction14 will also provide greater clarity about the origin of the factors
(2E)−1/2 .

1.1.6 Existence and disappearance of the time operator


The dimensional reduction that removes the time dimension also removes the
timelike component of momentum. In its place (sort of), it produces an energy
operator, the Hamiltonian, with a bounded spectrum. In the above case, as
expressed
p by equations (5) and (6), this results in an energy spectrum with
E = p2 − m2 , bounded from below by the value E = m.
This shows how, in the reduced theory, a bounded spectrum of the energy
operator, and the ensuing impossibility of defining a time operator, can arise as
a direct consequence of the dynamic reduction that gives us quantum mechanics
instead of quantum event theory.
14 section 1.2.8

12
1.2 Between QFT and QET
As described in the previous section, the natural choice of a Hilbert space for a
(1+3)-dimensional spacetime is L2 (R4 ), the space of square integrable functions
on R4 , with an inner product given by

≺ x | y = δ4 (x − y) = δ(x0 − y 0 )δ(x1 − y 1 )δ(x2 − y 2 )δ(x3 − y 3 ) (9)

It was also discussed why this is different from more familiar expressions like

< φ | ψ(t) >=< φ |e−iE(t−t0 ) | ψ(t0 ) > (10)

Comparable differences appear when the QET approach is expanded to the level
of a field theory, which leads to a many events Fock space, with annihilation
and creation operators a(x) and a† (x) obeying commutation rules of type

[a(x), a† (y)] = δ(x − y) (11)


0 0 1 1 2 2 3 3
= δ(x − y )δ(x − y )δ(x − y )δ(x − y )

for bosonic events, and similar anticommutation rules for fermionic events. This
is, indeed, quite different from the usual equal time rules:

∀t : [ψ(x, t), ψ † (y, t)] = δ(x1 − y 1 )δ(x2 − y 2 )δ(x3 − y 3 ) (12)

In spite of such differences, QET and QFT are still supposed to describe the
same physical phenomena, so clear links must remain between the two theories,
but how? Let us tackle this in a few steps.

1.2.1 Event fields in spacetime


The closest thing to a state in QET is a vector in a Fock space F, containing
a collection of events. It will be convenient to write states in the bra-form, i.e.
≺ Ψ| ∈ F † , and, e.g.:
≺ Ψ| =≺ x1 , x2 , ..., xn |.
Further, like in QFT, F is generated by repeated application of the creation/annihilation
operators. For bosonic events we assume:

≺ ...; x[n]; ...| a(x) = n + 1 ≺ ..., x[n + 1]; ...| (13)

≺ ..., x[n]; ...| a† (x) = n ≺ ..., x[n − 1]; ...|

In QFT, the crucial link between fields and observational results is provided by
Green functions. This remains true in QET, and Green functions will provide
the crucial link between QFT and QET. They will now be obtained, however,
as matrix elements of a Green operator.

1.2.2 The Green operator


With event states written in bra-form, the Green operator is conveniently writ-
ten as a ket, i.e.:
|G  : F † → C
Whereas each ket on a Hilbert space can also be regarded as a linear operator
on the conjugated Hilbert space of bra’s, the reverse does not hold. |G  is a

13
linear operator, but not a vector, because it has no norm. It is a linear operator,
however, and the ket-notation remains convenient.
Our next desire is a procedure that enables us to calculate (approximations
of) the Green functions ≺ Ψ | G , and, evidently, these will be required
to match corresponding QFT Green functions. Because the starting point for
calculating Green functions in QFT is a Lagrangian, something similar is needed
in QET. Ideally, what we are looking for is:
• a way to translate QFT Lagrangians into QET Lagrangians;

• and a procedure for deriving Green functions from these Lagrangians.

1.2.3 Lagrangians: QFT → QET


QFT Lagrangians must be translated into QET equivalents. To this end, a
simple Lagrangian will provide a translation format. It describes a (for brevity
massless) bosonic matter field ψ, interacting with itself by means of a field φ
and with an interaction term of the form gψ(x)ψ † (x)φ(x), with g a coupling
constant:

LQF T = ∂µ ψ(x)∂ µ ψ † (x) + ∂µ φ(x)∂ µ φ(x) + gψ(x)ψ † (x)φ(x) (14)

The most straightforward way to proceed is to look at the resulting Feynman


diagrams, each consisting of some number of vertices connected by propagators.
• In QFT, a vertex is a point x where particles are created or annihilated,
by the action of the fields ψ(x), ψ † (x), and φ(x).
• In QET, a vertex is a point where corresponding events a(x), b(x) and
c(x) take place.

Therefore, even though annihilation operators like a(x) in QET and field oper-
ators like ψ(x) in QFT are different entities, they do, from within their different
theories, refer to the same physical things. In other words, the mappings

ψ(x) ↔ a(x), ψ † (x) ↔ b(x), and φ(x) ↔ c(x)

should provide a direct translation between QFT and QET Lagrangians. Very
convenient, because it will make it rather straightforward to compare results,
in spite of the major structural differences between QET and QFT, that are
apparent from the commutation rules (11) and (12), to which we may further
add:
[a(x), b(y)] = a(x), b† (y) = 0
 
(15)
Applying ψ(x) → a(x), ψ † (x) → b(x) and φ(x) → c(x) to the QFT Lagrangian
(16), has the result:

LQET = ∂µ a(x)∂ µ b(x) + ∂µ c(x)∂ µ c(x) + ga(x)b(x)c(x) (16)

It must be observed here, that in QET the demand of the Lagrangian being
self adjoint is replaced by the requirement that it is invariant under charge
conjugation. In QET this is not the same thing, as is apparent from the above
commutation rules.

14
1.2.4 The free field case 1
The next phase is that we want to calculate Green functions. As a first step,
we consider the free fields case, with g = 0, while, for the time being, ignoring
the c-field altogether. In this case it is clear that the following holds:
• Without interactions, the only lines appearing in a diagram are the direct
connections between a and b sources.

• A nonzero result can be obtained only if there is an equal number n of a-


and b-sources,
• On these conditions, the Green function consists of in times the sum of all
permutations of the product of all the propagators connecting the sources.
In the absence of interactions, there is a simple way to arrive at this result:
G now has a generator X , such that

|G = eiX |0 
Z
with X = K(p) a† (p)b† (−p) d4 p
p
1
and K(p) = K(−p) =
p2 + i
This leads, e.g., to the result that

G(a : x1 , x2 ; b : y1 , y2 ) (17)
= ≺ a : x1 , x2 ; b : y1 , y2 | G 
= iK(x1 , y1 )iK(x2 , y2 ) + iK(x1 , y2 )iK(x2 , y1 )

where
eip(x−y)
Z
K(x, y) = d4 p (18)
(2π)4 (p2 + i)

1.2.5 The free field case 2


The important next step is that the same result is obtained using the following
differential equation:
δS
|G = ia† (x)|G  (19)
δa(x)
where S is the action operator, defined by
Z
S = L(x)d4 x (20)

Using this equation |G  can be obtained by iteration. To see this, let us first
consider a case where the action is restricted to a spacetime volume τ :
Z
S= L(x)dτ (21)
x∈τ

15
In calculating δSτ /δa(x), terms containing no a(x) or ∂a(x) can be disregarded.
Then, kinetic terms that do contain ∂a(x), can be removed by means of partial
integration, and we arrive at:
δS
|G 
δa(x)
Z I 
δ  2 µ
= −a(x)∂ b(x) dτ + a(x)∂µ b(x)dV |G 
δa(x) τ ∂τ
= −∂ 2 b(x)|G  + O(x ∈ ∂τ ) (22)

For the time being, the hypersurface integral O(x ∈ ∂τ ) will be neglected,
although it is certainly not irrelevant. It expresses that the boundary ∂τ can,
and usually will, act as a source.
The next step is to evaluate an example matrix element ≺ x, ȳ | G . It
is convenient here to use that a propagator K(x, y) can be written as a two-
event operator, in the form ≺ x|K|y . The a-field propagator can therefore be
written as K = {p2 + i}−1 , and, while duly noticing that Kp2 = 1, we get:
Z
≺ x, ȳ | G  = ≺ x|Kp2 |z ≺ z, ȳ | G  d4 z (23)
Z
= K(x, z) ≺ z, ȳ|p2(z) |G  d4 z
Z
= K(x, z) ≺ ȳ| − ∂ 2 a(z)|G  d4 z
Z
= K(x, z) ≺ ȳ|ib† (z)|G  d4 z

= iK(x, y)

Generalizing to more extensive states ≺ Ψ| =≺ x1 ...xn , ȳ1 ...ȳm | it is readily


verified that, indeed, results are equal to what is obtained by using the generator
X , i.e.:

Conclusion:
Without interactions, a nonzero result ≺ Ψ | G  for a state:

≺ Ψ| =≺ x1 ...xn , ȳ1 ...ȳm |

is obtained only if n = m, and, if so, it consists of in times the sum


of all permutations of the products of all the propagators connecting
the sources. This result can be derived both by using the differential
equation (19) or by assuming |G = eiX |0 .

1.2.6 Including interactions


Clearly, finding a Green operator becomes more complicated when interactions
are included. It is not clear, for instance, that finding a generator X for |G  is
still possible at all. There is one argument, however, that saves at least part of
the day by stating that equation (19) remains valid, and provides a method for
finding approximations for |G  in ever higher orders, i.e.:

16
hypothesis:
Equation (19) still holds, because it is a local equation.
The idea behind this is that the identity of an event may depend on its
location in spacetime. For instance, a spin-up event at one position in spacetime
may have a different direction when moved to another position, due to spacetime
being warped. These changes caused by deformations of spacetime must be
accounted for, by means of interaction events that undo the warp difference
between the two ends of a propagator.
At the same time, however, these events themselves are also sources of the
warping process, leading to an infinite regress. As a consequence, success in
approximating Green functions in this way comes to depend on coupling con-
stants being small enough. For the present, this will be taken to be the case,
and for the rest it will be assumed that QET is sufficiently similar to QFT to
make more advanced techniques, such as renormalization, applicable as well.
Recalculating δSτ /δa(x), but now with the Lagrangian of equation(16), pro-
ducing an extra term, resulting in:
Z
≺ x, ȳ | G = iK(x, y) − g K(x, z) ≺ ȳ, z̃, z | G  d4 z (24)

The matrix element ≺ ȳ, z̃, z | G  can be further analyzed in a similar way,
and the result is a sequence of terms that can be represented graphically, as in
figure 1, where the open circle represents G and the lines touching it define the
matrix element.

Figure 1: Green functions by iteration

Continuing this way, the Green function G(x, ȳ) is expressed as an infinite
series of terms, and, provided that the coupling constant g is small enough, this
converges to an iterative approximation of the desired Green function. By the
way:

More details about calculating Green functions are given in an appendix at the
end of this bundle.

Assuming that the methods used here can be extended to more interesting
Lagrangians, this is clearly a nice result, providing a direct link between QET
and QFT results. Moreover, the link between QET and QFT can be nicely
tightened by observing, that the above iterative equation can be transferred to
QFT by letting −∂ 2 act on the QFT Green function, expressed as the vacuum
expectation of a time ordered product of fields.

17
Example:
For the Lagrangian in eq. (14), we obtain

δS
= −∂ 2 ψ(x) + gψ(x)φ(x) = 0 (25)
δψ † (x)
and, therefore: ∂ 2 ψ(x) = gψ(x)φ(x) (26)

where it is assumed understood that ∂ 2 = ∂ 2 /∂x2 . Applying −∂ 2 to the Green


function < 0 |T [ψ(x)ψ † (y)]| 0 > produces

−∂ 2 < 0 |T [ψ(x)ψ † (y)]| 0 >= R − g < 0 |T [ψ(x)φ(x)ψ † (y)]| 0 > (27)

where R is the sum of all terms containing a derivative of a θ, in this case:

R = −∂ 2 T [ψ(x)ψ † (y)]
= −∂ 2 {θ(x0 − y0 )ψ(x)ψ † (y) + θ(y0 − x0 )ψ † (y)ψ(x)}
= T [ψ † (y)∂0 ψ(x)] (28)

Inserting the result in equation (27), while using

∂0 θ(x0 − y 0 ) = δ(x0 − y 0 )
∂0 δ(x0 − y0 )f (y 0 ) = −δ(x0 − y 0 )∂0 f (y 0 ) (29)

and that ψ and ψ † are canonical conjugates, and therefore

δ[ψ(x), δ0 ψ † (y)] = iδ(x − y) (30)

then, solving by means of Fourier transformation yields

< 0 |T [ψ(x)ψ † (y)]| 0 > =


Z
iK(x, y) − g K(x, z) < 0 |T [ψ † (y)∂0 ψ(z)φ(z)]| 0 > d4 z (31)

in agreement with eq. (24).

1.2.7 Position and momentum representations


Switching between (spacetime) position and momentum representations of field
operators is performed by means of:
Z
a(x) := d4 p ≺ x | p  â(p) with ≺ x | p = (2π)−2 exp(ix · p) (32)
Z
â(p) := d4 x ≺ p | x  a(x) with ≺ p | x = (2π)−2 exp(−ix · p) (33)

Transforming Lagrangian and action operators into the momentum representa-


tion, thus, results in expressions like the following:

18
Z
S = d4 x ∂µ a(x)∂ µ b(x)
Z
= − d4 x d4 p d4 q i∂µ ≺ x | p  â(p) i∂ µ ≺ x | q  b̂(q)
Z
= − d4 x d4 p d4 q ≺ −q | x ≺ x | p  pµ q µ â(p)b̂(q)
Z
= − d4 p d4 q ≺ −q | p  pµ q µ â(p)b̂(q)
Z
= d4 p p2 â(p)b̂(−p)

The momentum representation is convenient for understanding theoretical


structure, and for keeping track of plus- and minus signs, but less for including
localized structure. This needs to be added in a Fourier transformed format.
Also, integration over all of spacetime either implicitely suppresses a surface
integral over a spacetime volume, or imposes hidden assumptions about the
structure of spacetime, e.g. that it is toroidly closed.
Applying a change into the momentum representation, the action operator
corresponding to Lagrangian (16):
Z
S = d4 x {∂µ a(x)∂ µ b(x) + ∂µ c(x)∂ µ c(x) + ga(x)b(x)c(x)}

changes into:

d4 q
Z  Z 
4 2 2
S= d p p â(p)b̂(−p) + p ĉ(p)ĉ(−p) + g â(p)b̂(q)ĉ(−p − q) (34)
(2π)2

where the interaction term is obtained by means of

Z
d4 x ga(x)b(x)c(x)
x
Z
= d4 x d4 p d4 q d4 k ≺ x | p ≺ x | q ≺ x | k  g â(p)b̂(q)ĉ(k)
Z
= d4 p d4 q d4 k (2π)−6 ei(p+q+k)x g â(p)b̂(q)ĉ(k)
Z
= d4 p d4 q d4 k (2π)−2 δ(p + q + k) g â(p)b̂(q)ĉ(k)

d4 q
Z
= d4 p g â(p)b̂(q)ĉ(−p − q) (35)
(2π)2

which, e.g., immediately shows conservation of momentum.

19
1.2.8 Dimensional reduction 2: states and fields
Consider the case of two events connected by a propagator, and let τ be a
spacetime volume such that x 6∈ τ and y ∈ τ . We then have the following
theorem:

I
iK(x, y) = dV µ (z) iK(x, z) i ∂µ (z) iK(z, y) (36)
∂τ
I 
= ≺ x|iK dV µ (z) (pµ |z ≺ z| + |z ≺ z|pµ ) iK|y 
∂τ
↔ → ←
where, as usual, ∂µ =∂µ − ∂µ .

Proof: From Gauss’ theorem we have:



I
dV µ (z) iK(x, z) i ∂µ(z) iK(z, y) = (37)
Z∂τ → ←
d4 z iK(x, z) i ∂ 2(z) iK(z, y) − iK(x, z) i ∂ 2(z) iK(z, y)

=
Zτ →
d4 z iK(x, z) ∂ 2(z) − m2 )K(z, y) −

=
τ

K(x, z) ∂ 2(z) − m2 iK(z, y)

Z
d4 z

= iK(x, z)δ4 (z − y) − δ4 (x − z)iK(z, y)
τ
= iK(x, y) because x 6∈ τ

As a next step, it may be observed that the part between curly brackets in
equation (36) is self adjoint, so we can take its square root and define:
I 1/2
µ
∆∂τ = dV (z) (pµ |z ≺ z| + |z ≺ z|pµ ) (38)
∂τ

which then gives us:


iK(x, y) =≺ x|iK∆2∂τ iK|y  (39)
Next, observe that the whole procedure may also be repeated. Taking two
nonintersecting hypersurfaces τ1 and τ2 , such that τ2 encloses τ1 , we can write

iK(x, y) = ≺ x| iK ∆2∂τ2 iK ∆2∂τ1 iK|y  (40)


= ≺ x| iK∆∂τ2 ∆∂τ2 iK∆∂τ1 ∆∂τ1 iK|y 
= < φ(x, ∂τ2 ) | V (∂τ2 , ∂τ1 ) | ψ(y, ∂τ1 ) >

and now we may call φ and ψ states, and V an evolution operator. The hy-
persurfaces ∂τ1 and ∂τ2 are arbitrary closed nonintersecting spacetime surfaces.
This generates the possibility of dividing spacetimes in differently described
regions of arbitrary form, that can be described in different ways. For in-
stance region τ2 may contain a measuring apparatus, described classically, and
| ψn (y, ∂τ2 ) > describes the state vector on ∂τ2 , corresponding to measurement
result n. Likewise, τ1 can contain a classically described preparation procedure,
and | ψ(y, ∂τ1 ) > describes the corresponding state vector on ∂τ1 .

20
Finally, we may also define a family of nonintersecting hypersurfaces ∂τ (λ)
around point y, and reformulate equation (36) into:

iK(x, y) =< φ(x, λ2 ) | V (λ2 , λ1 ) | ψ(y, λ1 ) > (41)

If instead of a family of hypersurfaces around point y we take a sequence of equal


time planes between the points y and x, and replace λ by t, the above reasoning
remains valid, and from (subsub)section 1.1.5: Propagators and particles, we
already know what it leads to. We get
1
< φ(t2 ) | = ≺ φ, E| √ |t2  (42)
2E
V (t2 , t1 ) = ≺ t2 |e−iE(t−t0 ) |t1 
1
| ψ(t1 ) > = ≺ t1 | √ |ψ, E 
2E
Clearly, however, the method of obtaining dynamics in a three dimensional
manifold by means of a dimensional reduction of a four dimensional manifold
offers more general options than merely dropping the time dimension.
The above method of dimensional reduction is readily extended for situations
with a larger number of propagators, but things become more complicated by
the presence of interactions. Letting λ run through its range of values, we
encounter vertices, where propagators begin or end. In order to take this into
account, annihilation and creation operators must be introduced.
Instead of annihilating or creating with respect to a sequence in time, how-
ever, they must now act with respect to the dynamic variable λ, i.e., anni-
hilations and creations must be defined in the direction perpendicular to the
hypersurfaces of constant λ.

1.3 Quantum probability


For the well functioning of quantum physics, it is important that there are phys-
ical systems that really behave in a quantum way. Equally important, though,
is that there are systems obeying, at least in very good approximation, the
laws of classical physics. This makes it possible, e.g., to manufacture machines
producing quantum systems for our experiment, make measuring instruments
producing data by interacting with such a quantum system, and to have mem-
ory systems, able to store and read data of quantum experiments, and have
living beings who write papers and books about quantum physics.
In specifically also an experimental setting, the interplay between quantum
and classical pictures is crucial for a meaningful application of quantum physics.
Accordingly, we shall assume that in any such application, the volume of space-
time in which the application takes place can be divided into quantum and
classical sectors (although not necessarily in simple and predefined ways). The
quantum sectors are where the actual interactions to be studied take place,
whereas the classical sectors contain systems that are sufficiently stable to keep
the experimental set up in place, and store data into a sufficiently permanent
memory. If desired, this sector may be taken to include humans, pen and paper,
libraries, desktops, and computer memories (even though the well functioning
of the latter may depend in direct and crucial ways on quantum physics).

21
Quantum physics is a probabilistic theory. Therefore, in experiments where
this aspect is present, the theory must provide a nontrivial probability measure,
and the experimental set up must provide a classical measurement procedure
supplying statistical data. This procedure must result in a data set with a
classical data structure, corresponding to the outcomes of the measurement,
and stored in a stable memory device obeying classical rules for storing and
retrieving the data.
It will not be assumed that measurement outcomes must correspond to pro-
jection operators on a Hilbert space. A somewhat vague and smeared out rela-
tion between quantum states and measurement results may be quite acceptable.
Also, it is not neccessary that quantum and classical sectors in an experiment
correspond to predefined volumes in spacetime. A particle trajectory in a cloud
chamber is a nice example. The positions where interactions between a quan-
tum particle and surrounding gas molecules take place are not predefined, but
they do result in a position measurement, which is also vague enough to produce
only a limited disturbance of particle momentum. Moreover, the measured tra-
jectory provides information about particle momentum, and this is information
needed for, usually, a quantum physical analysis of a particle collision. In other
words, the relations between classical and quantum sectors in spacetime may
be complex.
In this respect, astronomy is also a subject that provides an interesting case.
The universe is a place full of quantum matter, such as light and other kinds
of quantum matter flying around. In other words, with respect to astronomical
observation, it is a giantly large quantum sector. The classical sector in this type
of ‘experiment’ starts at the observation instruments, producing classical data.
As discussed in particular in section 1.2.8, the boundaries between a quantum
and a classical sector are the places which, in a traditional setting in quan-
tum physics, are associated with quantum states, see especially equation (42).
This enables the picture of a telescope as an instrument producing (classically
described) information, when hit by a (quantum) light wave from the universe.

1.3.1 Quantumclassical cocktails


The idea of clearly separating an experiment into preparation, quantum, and
measurement sectors is a bit on the simplistic side. From a macroscopic per-
spective, a typical experiment is a sequence of manipulations and observations,
but often in a quite flexible order. Stretching the meaning of the words a bit,
the macroscopic description of an experiment is a list of switch settings and
pointer positions.
Switch settings may usually precede the pointer positions in time, but not
necessarily. For instance, experimental data may very well be postselected on
certain criteria, and measurement values obtained during an experiment may
also be used to improve the quality of preparation quantities. About these and
other interventions, one can also say that they effectively amount to turning
pointer positions into switch settings or vice versa.
Intervening actions during an experiment can also be extremely useful by
allowing for runtime feedback loops in a process. The way in which the distance
between needle and substrate is controlled in a scanning tunneling microscope
is an example.

22
The usual mode of measuring is that the tunneling current is kept constant,
by means of feedback to the actuator that controls the distance between needle
and substrate. Because the steering signal determines how much the needle
is going up or down, it also provides the measuring signal, which makes the
distance between preparation and measurement extraordinarily thin.
Coming back to an earlier example, mentioned above, the interplay between
classical and quantum modes of reasoning is clearly visible in the way in which
data are gathered in particle experiments. To begin with, quantum physics does
play an important role in planning experiments, because it is important for
particle theory, and, hence, for deciding what is interesting to investigate.
Then, classical electromagnetism is important, e.g. in the Nobel prize win-
ning beam focussing method developed by Simon van der Meer at Cern. It
provides a very nice example of a feedback mechanism in which measurements
of beam quality, as part of the preparation of the quantum state of the incoming
system, are fed back into the system in order to improve focussing.
Then, after a particle collision in the beam pipe, the resulting rain of sec-
ondary particles passes various arrays of detectors, and from the curvature of
the orbits in a magnetic field their momenta are deduced. All this requires little
or no quantum physics, but mostly or only electrodynamics. After acquiring
gigantic amounts of data, these are scanned on possibly interesting events, and
in the analysis of this final selection quantum physics is important again for
trying to understand what was going on.
Taking all this into account, it is sort of clear that there are no fundamental
distinctions between preparation and measurement, and that the interplay be-
tween classical and quantum descriptions of parts of an experiment can contain
all kinds of subtleties.

1.3.2 Uncertainties in preparation or measurement


Both preparations and measurements are not always as accurate as one might
like, but there are ways to take this into account, for instance by describing a
preparation by means of a density matrix instead of a wave function. Density
matrices provide a way to combine quantum probabilities with classical proba-
bilities, describing the limited accuracy of the preparation process. Replacing
the expression | < φ |U | ψ > |2 by T r[P U W U † ], where

W = | ψ >< ψ |, P = | φ >< φ |, and T r[...] is the trace operation,

the option is created to change the incoming state into a distribution


Z
W = ρ(ψ)| ψ >< ψ | (43)
ψ

where ρ(ψ) can be used for expressing uncertainties in the preparation process.
Recalling that there should not be a fundamental difference between prepa-
rations and measurements, it must be expected that the projection operator
P , describing the measurement outcome, can be treated in a similar manner.
Indeed, to some extent it can, but there is a bit of a problem with sequences of
consecutive measurements.

23
In this case the expression p = T r[P U W U † P ] for the probability must be
replaced by
p = T r[P2 U2 P1 U1 W U1† P1 U2† P2 ] (44)
where the P2 at the end is added for symmetry, which can be done safely because
it is a projection operator.
Now if we replace the P s by distributions, we have the problem that they
occur twice, so we must take their root in order to compensate. Moreover, if we
want to include the possibility that the measurement also changes the system,
a unitary part must be added, which must included in taking the root. Let us
say this changes the Pi into operators Qi , and if we now want to describe the
measurement uncertainty in Qi by means of a distribution ρi , we must have
that
Z
Qi Q†i = ρ(ψi )| ψi >< ψi | (45)
ψi
where the ψi are the eigenvectors of Qi .
Finally, in order to emphasize the equal status of W and the Qs, W can be
written as an action on something else, like the empty state | 0 >< 0 |, i.e.:
W = Q0 | 0 >< 0 |Q†0 (46)
It is nicely symmetric, then, to do the same for the final intervention QN ,
i.e.: QN = QN | 0 >. Using all this, we find for the probabilities that:
p ∝ T r[ < 0 | QN UN ...Q1 U1 Q0 | 0 >< 0 | Q†0 U1† Q†1 ...UN

Q†N | 0 > ] (47)
and, therefore:
p ∝ | < 0 | QN UN ...Q1 U1 Q0 | 0 > |2 (48)

1.3.3 Quantum probability


The final thing missing before we can complete an expression for calculating
probabilities, is a way to sum over, or conditionalize on, values or ranges of
values of observations or other interventions.
Therefore, let:
• Q be the sequence of all interventions: Q = (Q1 , ..., QN ),
• range(Qi ) the set of possible values of intervention Qi , and
• range(Q) the set of possible values of Q, i.e., each possible value is a
sequence of values of the Qi .
Because the order in which items appear in the list of Qi ’s is a fixed ascending
time ordering, we can also write it in the form < 0 | T [U, Q] | 0 >, as the time
ordered product of interventions and unitary evolution. Accordingly:
p ∝ | < 0 | T [U, Q] | 0 > |2 (49)
Now we select a set B ⊆ range(Q), of all sequences that are allowed into the
set of valid results of the experiment. This means that B defines the preparation
of the experiment (where the ‘pre’ before ‘paration’ is a bit of an anachronism,
in view of the fact that it was just concluded that such selections may take place
at every stage of the experiment).

24
Next, let A ⊆ range(Q) be a set of sequences defining a result for which we
want to know the probability, i.e., we are looking for a result p(A|B), that will
be given by15 :
| < 0 | T [U, Q] | 0 > |2
R
Q∈A∩B
p(A|B) = R (50)
Q∈B
| < 0 |T [U, Q]| 0 > |2
Finally, observe that the vacuum expectation of a time ordered product of evo-
lution operators and sources defines a Green function, allowing us to write
| G(Q) |2
R
Q∈A∩B
p(A|B) = R (51)
Q∈B
| G(Q) |2

1.3.4 Probability in QET


The next job is to translate equation (51) into a QET setting. Again we have
a list Q of N interventions (Q1 , ..., QN ), and, in analogy to the demand stated
in equation (45), each Qi in the list must be linked to a distribution of vectors
≺ Ψ| in the events Hilbert space, in such manner that :
Z
Qi Q†i = ρ(Ψi )|Ψi ≺ Ψi | (52)
Ψi

It makes sense here to formulate the result of each intervention Qi , and of


the entire list Q, directly in terms of field operators. This is possible, because,
corresponding to each Qi and Q, there must be compositions A(Qi ) and A(Q)
of event fields, such that:

≺ Qi | =≺ 0|A(Qi ) and ≺ Q| =≺ 0|A(Q). (53)


and:
| ≺ Q | G  |2
R
Q∈A∩B
p(A|B) = R (54)
Q∈B
| ≺ Q | G  |2
in agreement with eq. (51). Because the vectors ≺ Qi | and, hence also ≺ Q|,
obey equation (52), this equation can also be changed into:

ρ(Φ) | ≺ Φ | G  |2
R
p(A|B) = Φ∈A∩B
R (55)
Ψ∈B
ρ(Ψ) | ≺ Ψ | G  |2

which nicely distinguishes between the quantum probabilities, expressed by the


Green functions, and preparation and measurement uncertainties, as contained
in the distributions ρ.

1.4 To become or to be?


In standard quantum physics, time is treated as a dynamic variable, rather than
a dimension. As it turns out, and as described above, this is not a necessity.
A static spacetime picture is possible. Conceptually, and from a relativistic
perspective, it is quite significant that such a perspective is possible. It should
be.
15 In view of the time symmetric formulation of this expression, it can also be seen as a
generalization of the probability measure as described in Aharonov, Bergmann and Lebowitz,
Phys. Rev. B, 134: 1410-1416, (1964)

25
Further, starting from a static spacetime picture, it should also be possible,
by means of any sufficiently well behaved family of spacetime hypersurfaces, to
turn it into a dynamic one, by means of dimensional reduction. At first sight,
however, the relation between QET and QFT does not seem to be as simple as
this, so it needs some further investigation.
Let us first observe that dynamic reduction by means of a well behaved
family of hypersurfaces ∂τ (λ) is a quite general procedure. It works by slicing
all propagators of all terms of a Green function along the same ∂τ (λ), by means
of theorem (36) (and checking that it also works for propagators with form
factors, e.g. like (γ µ pµ − m)/2m) for spin-1/2 systems.
If done properly, this can be formulated to produce a quantum state Ψ(λ),
containing, for each propagator present, a particle described by a state that
develops in agreement with a set of field equations. Occasionaly, interaction
terms may produce or destroy these particles, as will be described by interaction
terms in a Lagrangian.
In addition to the Lagrange equations, there will also be an equal λ com-
mutation (or anti commutation) relation, describing the quantum character of
the fields. In principle, this whole scheme should work in a general manner,
for arbitrary well behaved families of spacetime hypersurfaces. As a special
case, then, one may choose equal time hyperplanes in spacetime for the surface
integrals ∂τ (λ), i.e., λ becomes t.
In actual history quantum physics developed in a quite different manner.
The crucial point was that the energy operator, the Hamiltonian, was assumed
to be the generator of time shifts, instead of being merely numerically equal
to it, under the proper conditions. This immediately made it impossible to
define a proper time operator, as became apparent by Pauli’s 1933 argument,
see footnote 12. The confusion is understandable, because quantum physics
is conceptually difficult to grasp, and takes a lot of time for getting used to
its strange features. On the other hand, by the time when a more or less
mature version of quantum physics became established, relativity theory was
already about twenty years old, and even general relativity had existed for
about ten years. At least in some corners of physics, physicists were getting
used to formulating theories in an explicitely covariant manner, and recognizing
the need to do this.
The failure of trying something similar with quantum physics probably
means it had not yet grown beyond a status comparable to statistical physics,
with Hilbert spaces replacing the phase space. It was apparently not yet recog-
nized as a basic theory of physics, in need also of a seriously relativistic version.
This much is still comprehensible, but it should have been corrected somewhere
in the 1930s, or certainly by the time that relativistic quantum theory and field
theory came about. Instead, it turned out, halfway through my first lecture on
the subject, that this item had been left there especially for me to deal with.
This was a challenge I took eagerly, with both hands, and, after finishing uni-
versity, I wrote a research proposal that was, by some miracle, actually granted,
resulting in a few papers, a book, a Ph.D., and, finally this bundle of essays.

26
2 SPACETIME, PARTICLES, and
COSMOLOGY
2.1 Particles and spacetime
The present section is based on the idea that proton decay, or more generally any
transition between quarks and leptons, is forbidden because they are separated
by a Lorentz transformation. With this in mind, let us take a fresh look at
the spectrum of fundamental fermions. Omitting isospin and the distinction
between particle generations, the spectrum of particles can be ordered in the
following array:
 
l(epton)    
 r(ed)
 ⊗ P(article) L(eft)


 g(reen)  ⊗ (56)
A(ntiparticle) R(ight)
b(lue)
In fact, we are overcounting a bit, because of an observed relation between
particleness and handedness, with particles being lefthanded and antiparticles
righthanded, changing the multiplet into
 
l  
 r 
  ⊗ P⊗L (57)
 g  A⊗R
b
Awaiting a better view on the background of this relation, however, I shall
ignore it and stick to multiplet (56).
Still to be accounted for is the rest of the particle spectrum, including the
three generations and the isospin variable, together producing the array
 
e(lectron)  
 µ(uon)  ⊗ U(p) (58)
D(own)
τ (auon)
With respect to the interpretation of these multiplets, the two columns on the
right in multiplet (56) are well known to provide a spin- 12 spinor representation
of (ordinary) spacetime. Following the suggestion that the particle spectrum as
a whole is actually a representation of an extended spacetime, let us look at the
first column of table (56) in this way. What, then, do we see?
As a representation of spacetime, given its clear (1,3) structure, this first
array can only be a spin-1 vector representation of a 4-dim spacetime subspace.
This idea comes with the considerable bonus that it nicely explains a significant
failure of the standard model of particle physics. Although the proton is pre-
dicted to decay, very slowly, into leptons, this decay has never been observed.
If the three quarks are timelike, however, and the lepton spacelike (or, as we
shall shortly discuss, vice versa), then the three quarks would be connected by
simple rotations, while being separated forever from the lepton by a Lorentz
transformation.

The result is a strict conservation law, prohibiting the decay of quarks into lep-
tons, and, consequently, also the decay of the proton.

27
The lepton-quarks multiplet, with its (1,3) dimensionality makes the impres-
sion of being a representation of ordinary spacetime. This might be possible, but
it would turn the quarks into spin-3/2 particles. It is questionable whether this
leads to the right kind of physics, but maybe we should keep in mind whether
this is nevertheless a possibility. Another, perhaps better option, is that it is a
vector representation of 4 different dimensions. Therefore we should now start
looking at the dimensionality of all of spacetime.

2.1.1 A spacetime ⊗ timespace hypothesis


In QET, introducing extra dimensions of time is as straightforward as adding
spacelike dimensions, and a proper look at the spectrum of fundamental fermions
seems to suggest one specific choice of dimensions in particular. As we shall see,
choosing
4 space and 4 time dimensions.
produces not only a particularly beautiful symmetry, but also a very nice fit to
the observed particle spectrum.
First, however, let us take a look at the eight presumed dimensions and con-
sider the validity of general relativity (GR). GR is a conceptually very beautiful
theory, and it is clear that the theory works beautifully at intermediate dis-
tances. On somewhat closer examination, however, it becomes quite unevitable
that things go wrong outside this domain. At large distances it is necessary to
introduce a new force called dark energy, and at small distances and high mass
density it leads to singularities at the centers of black holes and at the begining
of the universe.
Also from a conceptual perspective some things are missing. At very small
distances, one would expect that the microscopic dimensions, introduced for
explaining the particle spectrum, should become important in the dynamics of
spacetime, in a way that should arrest the formation of singularities. At very
large distances, on the other hand, dark energy should be somehow included in
the equations.
Taking these thing into account, it seems inevitable that something is missing
in general relativity, and holding on to it can hardly be anything else than
a combination of desperation and a lack of imagination. It can further be
observed, that, also from a purely theoretical perspective, something is missing,
in the sense GR does not contain the complete set of linear transformations
possible on a spacetime. The circular and hyperbolic rotations in spacetime,
i.e. proper rotations and Lorentz boosts, are generated by matrices of types:
   
0 i 0 1
R= and L =
−i 0 1 0

The odd ones out are the diagonal transformations:


 
1 0
T =
0 −1

A simple reason for including them is for completeness of the space of linear
spacetime transformations. For this reason they are standard ingredients of
transformation groups used in particle theory, such as SU3.

28
What are the resulting transformations of these three types of generators?
Clearly, rotations act between directions in space, but, allowing for multiple
time dimensions, presumably also between different directions of time. Boosts
act between spacelike and timelike dimensions, but what about T ? It generates
a combination of expansion and contraction:
 α 
e 0
exp(αT ) = (59)
0 e−α
This expands spacetime in one direction, at the cost of contracting another direc-
tion perpendicular to it. Being diagonal, it does not mix dimensions. Therefore,
at least for now, there is no reason to assume that it could not apply to arbitrary
combinations of space- and timelike directions. How should we further interpret
this type of transformations? They potentially explain why some dimensions
are of cosmological size, and the other ones ultramicroscopically small. There is
also the suggestion of a link with dark energy. Furthermore, properly included
in a general theory of spacetime transformations, they should be able to prevent
the occurrence of singularities.

2.1.2 Bangs, Crunches, Holes, Particles, Strings?


Viewing the history of the universe in reversed time, we are headed towards a
Big Crunch in about 14 billion years. What we, in forward time, observe as a
dark energy pressure speeding up the expansion of the universe, is now an ever
increasing braking force, delaying its collapse.
Equation (59), however, makes the collapse into a transition, by which
macroscopic dimensions become microscopic, and vice versa. In the above 8 di-
mensional ST⊗ts model, it implies that the universe will pass into an st⊗TS
state with a microscopic spacetime, formed by our present three dimensions of
space and one time, and a macroscopic timespace with three dimensions of time
and one of space. This provides a new picture of the inflationary period of
the universe, which is where this transition takes place, with, still in reversed
time, our familiar dimensions crumbling and timespace exploding. What about
physics in timespace, at the other side of the Big Bang? Actually, the turnover
might not make much of a difference. Events remain events, and propagators
keep the same form. Up to renaming time into space and vice versa, the universe
at the other side should be very similar to ours.
If this idea about the Big Bang singularity is correct, it is natural to assume
that it should apply to the inside of a Black Hole as well. The singularity at
the center of the hole should be just a similar type of bridge between spacetime
and timespace. Further, with respect to particle theory, the natural assumption
would be that black holes are quantized, and elementary particles are the small-
est possible black holes: tiny transitions between our side, with macroscopic
spacetime and microscopic timespace, and the other side, with macroscopic
timespace and microscopic spacetime.

29
By their bridgelike structure, one might interpret them them as strings.
There are fundamental differences, however, with ideas that are currently be-
coming popular in string theory, as characterized by the slogan EPR=ER. Here
the idea is that EPR correlated particles are connected by an Einstein Rosen
bridge, being a blackholish connection between two points in spacetime itself.
As a consequence, such an explanation does not hold in the ST⊗ts view, but
it is also superfluous, because in QET the EPR correlations result from the
probability measure being basically time symmetric.
A related slogan, enjoying increasing popularity, is GR = QM. It is not
my intention to discuss the fundamental ideas behind this to any length. Erik
Verlinde’s views on gravity (and other forces) as being emergent, i.e., basically
having a thermodynamic origin, is an important ingredient, and finding a way
of connecting this to QET is interesting, but beyond the scope of this essay.
Just a few superficial remarks should be in order, however.
• Verlinde’s analysis basically starts from the information content of physical
systems within volumes in space, and continues by enclosing the system
within a holographic screen, being a closed surface surrounding the system.
• He then observes that the surface can be seen as a storage device for
information, with a storage capacity proportional to its surface.
• The information content of the screen has an associated entropy S and
temperature T .
• Gravitational force (and other emergent forces) within the volume are
obtained from thermodynamics, by using F = T ∆S/∆x.
There is an undeniable touch of beauty in this approach, but it is beyond the
scope of this essay, and in the form in which I read it, it lacks a relativistic for-
mulation. From the point of view of QET, deriving an expression for forces is
not a very natural thing to do, because it includes choosing a coordinate frame
in spacetime, and obtaining expression relative to the chosen frame. I certainly
don’t have objections, however, against the idea of regarding a spacetime volume
as a container, for information that can also be expressed entirely on the hy-
persurface of the volume. In fact, this is precisely what happens in dimensional
reduction.16 Using this, I presume that, in principle, a fully frame independent
form of forces being emergent in some thermodynamic sense should be possible.
For the time being, in event theory itself, results, including probabilities,
can be obtained using only covariant expressions for the Lagrangian and the
sources. Further, increase of entropy comes as a direct consequence of complete
absorption, i.e., of having 1l as a( maximally high entropy) final state. These
may be ingredients for finding a QET generalization for emergence of physical
laws for emergence, but it is certainly outside the scope of this essay.
Finally, for me, as an outsider, looking from a distance and with a rather
superficial view, a thing that seems a bit worrying is the ease by which the idea
of emergent forces produces quite different phenomena. You can straightfor-
wardly produce Newtonian gravity, but with just a bit of modification you get
general relativity quite as easily, or so it seems from slogan GR = QM. I am not
sufficiently familiar with the details, and mainly quite advanced on my way out
of physics, but it doesn’t smell quite like something I would expect to be valid.
16 Section (1.1.1)

30
2.1.3 Multiple time dimensions and particle theory
In order to interpret the particle spectrum, we first assume that the spins on
both sides
 ofthe hole together determine the particle type. Then, in multiplet
e  
U
(58):  µ  ⊗ we see an array that looks like a vector and a spinor
D
τ
representation of a three dimensional space, or time.
Looking for a spacetime ⊗ timespace symmetry, this is interesting but not
very revealing. With a slight extension, however, the result is actually quite
spectacular, and also accommodates for the existence of dark matter. It involves
introducing another generation, that we shall call the δuisteron (δ), and with it
comes darkness, a new, isospinlike quantity, with the values F(ront) and B(ack).
The result is:
   
l     δ    
 r 
⊗ P L  e  U F

 g  ⊗ ⊗  µ ⊗ D
  ⊗ (60)
A R B
b τ

The resulting table has events ordered in two nicely similar multiplets, and with
a clear interpretation in terms of existing or expected particles. For instance, the
combination (r,A,R,µ,U,F) would be a red anticharmquark with righthanded he-
licity, and any combination containing a δuisteron or Backward darkness would
be dark matter.
The δuisteron, together with darkness, provides sufficient room for includ-
ing various different types of dark matter, such as the combination (l, δ, D),
(lepton, δuisteron, Down), which would accommodate the fourth (sterile) type
of neutrino that turned out to be compatible with WMAP results and other
observations.
The idea behind the above multiplet is that a particle is a tiny black hole with
two sides, one in our universe, where spacetime is macroscopic and timespace
microscopic, and one in the other universe, where spacetime is microscopic and
timespace macroscopic. In this interpretation, numbering the six arrays in table
60, we have, from left to right:
1 a vector representation of microscopic timespace (s1 t3 );
2,3 a spinor representation of macroscopic spacetime (S3 T1 );
4 a vector representation of microscopic spacetime (s3 t1 );
5,6 a spinor representation of macroscopic timespace (S1 T3 );
The multiplet also provides a clear prediction for the number of different dark
matter particle species. Counting all possible combinations, there are 160 types
of dark matter, which is 160/96 = 1,7 times as many as the types of known
matter, provided that all combinations are possible.
The latter is not the case, of course, because of the relation between par-
ticleness and helicity. Taking this into account, the number of dark matter
types could become as much as 3,3 times as many, provided that there is no
such relation between isospin and darkness. Evidently, this fraction does not
directly give us the mass ratio, but it may still provide some indication of why
the amount of dark mass is larger.

31
2.1.4 Evaluation
Surveying the first chapter of this second essay, we must now try to evaluate
the degrees of plausability versus phantasy involved in its various ideas. I would
say that the idea with the highest degree of plausibility is to accept the limited
validity of general relativity, and work towards a generalization:
- I see no reason for accepting the occurrence of singularities;
- A unification of GR with a theory of dark energy is a highly desirable
objective;
(and, evidently, the most desirable long range objective would still be a
unification of all interactions.)
Further, the idea I started from, i.e., to explain the impossibility of a transition
between quarks and leptons by their being seperated by a hyperbolic rotation,
is really quite fascinating, but very far removed from current theoretical de-
velopments. This is also true for multiplet (60). It is a stunningly beautiful
symmetry, but still far from being even the beginning of a theory.
On the other hand, my personal degree of belief in the existence of multiple
time dimensions is actually quite high, and I like the idea of ST → T S strings,
i.e., bridgelike structures between spacetime and timespace, a lot better than
standard ST → ST strings. Furthermore, the idea that this is also the structure
of black holes is really fascinating. Incidentally, by the way, it also generates
a view with respect to the black hole information paradox.17 Because (also
macroscopic) black holes are transitions between spacetime and timespace, they
are longer isolated systems within spacetime. Accordingly. there is no longer
a unitary connection between what falls into the hole, and what comes out in
the form of thermal radiation by black hole evaporation. On the other hand, it
is necessary to assume that the total amount of matter falling into the hole is
the same on both sides. This, however, should be arranged as a consequence of
symmetries of the probabilitiy measures at both sides of the hole.
Still far away, and hard to say, is whether all this would lead to anything like
the type of general field theory that could have satisfied Einstein, but it may
be as close as it gets. Needless to repeat, however, that, for the time being, all
this is pure speculation.

2.2 Time and space, the same, but different


The S3 T1 ⊗ s1 t3 model generates the following important problem:
If space and time are so alike, as seen by relativity theory, then how
can they be so different?
A first, obvious and macroscopic difference, is that there are three macroscopic
dimensions of space and only one of time. A second difference is that, with
respect to time, all our observations are about present and past moments in
time, and never about the future.
The first difference can be thought of as a coincidence. There might be
different types of universes, such as a universe with two macroscopic dimensions
of space and also two of time. We just do not happen to live in one of those, or
they might not be fit for life at all.
17 See, e.g., The Black Hole War; Susskind; Little, Brown, and co. (2008), p.209 ff

32
The second difference, however, is more difficult to understand, also because
we think about it in terms of cause and effect. Our basic intuition is that we have
no direct knowledge of the future because it hasn’t happened yet, and what is
going to happen depends, among other things, on our own actions, about which
we are, at least to some extent, still free to decide.
The upcoming sections are an attempt to organize some thoughts about the
direction of time, in a sort of cosmological perspective.

2.2.1 Cause and effect, and the EPR paradox


The basic intuition about cause and effect is that things happening now deter-
mine, or influence the probability for, what happens in the future.
Now the first problem, with respect to thinking about the temporal order of
things, is that simultaneity is not an observer independent concept. Never mind,
this is all nicely described by relativity theory. Except, however, that quantum
physics manages to reproblemize the situation, by means of the process called
state reduction. Even in a relativistic version of quantum physics this seems
to include nonlocal effects. Let me illustrate this by considering the following
version of the Einstein, Podolsky, Rosen paradox.
A pair of two spin 12 particles is prepared in a singlet state and then separated
in a spin conserving manner. One particle goes to physicist A, and the other
to a far away physicist B. They both know that, by measuring the spin of their
particle they can instantly change the state of the other particle too, because
an instantenous collapse of the wave function will take place, no matter how far
the particles are removed from each other. Competitive by nature, they both
want to be the one who changes the state of the other one’s particle, so, as soon
as the particles are bound to arrive, they hurry to their labs and measure the
spin.
It so happens that A measures Sx and finds a result + 12 , whereas B measures
Sz and finds − 12 . Therefore, A claims he changed B’s particle into an Sx = − 12
state, and B that he changed A’s particle into Sz = 12 . They then phone each
other and start quarreling about who was first, merely to find that they finished
their measurements at precisely the same moment. Therefore, the measurements
took place outside each other’s lightcones, and the time order comes to depend
on the state of motion. A definite answer cannot be given.
What remains is that, whatever measurement B performes, his result must
be consistent with what A finds, and vice versa, but how can this information
be exchanged when the particles are at a spacelike distance? Conclusion, in a
much (ab)used phrase: something weird is going on.

2.2.2 Delayed preparation

There is something definitely weird about the version of the EPR paradox dis-
cussed above. It uses a superposition of states, followed by a state reduction,
and, subsequently, a relativistic argument to conclude to the paradoxical charac-
ter of the situation. The concept of a quantum state, as used in the description
of the situation, is already by itself a fundamentally non relativistic concept,
however, by being based on state vectors | ψ > that contain no reference to a
time dimension.

33
It may be argued that a relativistic argument should be based on a relativis-
tic theory. And, anyway, it must be checked whether this makes a difference
and how these situations are actually described in a relativistic theory. In QET,
probabilities are calculated from squared Green functions, rather than squared
inproducts of states.18 It matches standing practice in relativistic quantum
physics, where all sorts of quantum effects, like Lamb shifts or magnetic mo-
ments, are calculated by adding Feynman diagrams, each of them a possible
history of a process. Clearly, these practices do include the time dimension.
Although this is quite a bit more satisfactory, for an analysis of EPR corre-
lations the result is not different than before. What remains is that:
Changing history in a Green function at a given point in spacetime
may require changes at other points not causally connected to it.
Still, if we only look at relativity theory, without imposing our intuitive ideas
about causality and the direction of time, there is still a clean way out. A and
B may both be right about determining the state of the particle the other one
receives. Causal influence back in time is not forbidden by relativity theory, and
it is explicitely in agreement with the QET probability measure. It may clash
with normal intuition, but it leads to correct results.
Accordingly, A may well be quite right in assuming that, because he mea-
sured Sx = + 12 , B must have obtained an Sx = − 12 particle. And B may be
quite right as well, in assuming that, because he measured Sz = − 12 , A must
have obtained an Sz = + 12 particle. Both for A and B, the preparation of their
systems may well have taken place after their measurement. We may call this
a delayed preparation, or delparation.

2.2.3 Design of a retrophone

Now suppose that A and B see no trouble in accepting the idea of delayed
preparation. They are annoyed, however, that they don’t know in advance in
what state the system is delpared for them.
This way, they may waste valuable observation time by measuring quantities
in which they are not interested. Therefore, they reason, if we can change things
backwards in time, why not build a system by which we can communicate back
in time? So A suggests the name retrophone, for an apparatus for communi-
cating with someone in the past, but B wants to call it an advancophone, for
communicating with someone in the future. A was first, however, so retrophone
it is. Moreover, retrophone sounds a lot better, and has a nicely oldfashioned
air about it.
I will first look at a design of how a retrophone could work, and this auto-
matically leads to a discussion about why it doesn’t. To begin with, B may
send messages to A, into the future, by means of pulses of photons transmitted
through glass fiber cables. This much is standard procedure.
In order to construct a well functioning retrophone, that also talks back
from future to past, we need a second beam of photons, that can be pointed to
a direction in the sky that doesn’t accept photons.
18 See eqs. (50) and (55).

34
Now what A can do to send an answer to B, is to direct the second beam
into this direction where the light is not accepted. B can then observe this by
observing the power output drop in this second cable. Accordingly, by shifting
the direction of the beam, A can signal to B in binary code.
Whereas it sounds awfully simple, it doesn’t work like this. Why not?
Clearly, it would work fine, if there were places in the sky that do not accept
light. We don’t know any such places, however. Their absence is a condition
known as complete absorption and it plays a role in discussions in cosmology,
originally initiated by discussions about time symmetry in the theory of elec-
tromagnetism.
Its significance is not restricted to electromagnetism, however. In as far as
known, it is a general condition, pertaining to all types of fields. Thinking about
it, however, did start with ideas about electromagnetism.

2.3 Direct action, time symmetry,


and complete absorption
I became aquainted with the electromagnetic theory of direct action by reading
the article of Hugo Tetrode19 for a seminar. The idea was independently de-
veloped, however, by Schwarzschild20 and by A.D. Fokker21 . The term ’direct
action’ refers to the absence of the electromagnetic field.
The idea is that there can be a direct exchange of energy and momentum,
between particles at positions separated by a lightlike distance, i.e., the idea is
that this is possible because the relativistic distance r2 − c2 t2 between the two
particles is zero. A quite brilliant idea, and the resulting theory is beautifully
timesymmetric, which, however, is also a problem. This idea appealed to me,
also because it suggests an interpretation linking it to QET. If the electromag-
netic field can be eliminated in this way, then why not do the same with all
other fields as well? It would leave us with nothing more than direct relations
between vertices, quite close in spirit to event theory.
A counter argument against this idea is that not all fields are massless,
which destroys the idea of direct interaction at lightlike distances. Pure fields
can be assumed massless, however. Mass then originates from adding terms in
a superposition of virtual processes, containing interactions: interactions with
the Higgs field as well as terms containing, e.g., self energy interactions with
the electromagnetic field. Even so, there is also a deeper and more worrying
point. The time symmetric character of direct action electromagnetic theory,
implies that electromagnetic interactions depend on the future distribution of
charges and currents as much as on past ones. This is, by the way, not due to
the assumption of direct action, but to its time symmetric character.
19 H. Tetrode. Zeitschrift für Pysik 10: 317, 1922
20 K. Schwarzschild, Göttinger Nachrichten, 128, 132 (1903).
21 A. D. Fokker, Zeits. f. Physik 58, 386 (1929); Physica 9, 33 (1929),

Physica 12, 145 (1932)

35
Wheeler and Feynman investigated a time symmetric version of electrody-
namics including fields. Their fields, however, were time symmetric, with re-
tarded and advanced components. They were assumed to depend as much on
future as on past charge and current distributions. They found that this the-
ory can be made compatible with standard electrodynamics, but only under the
condition of complete absorption, i.e., all radiation emitted in our era must be
absorbed somewhere in the future.
This condition of complete absorption also follows from the time symmet-
ric probability measure of quantum physics that was introduced by Aharonov,
Bergmann and Lebowitz22 . The probability measures (50) and (55), derived
in the previous essay, can be regarded as a generalization of it. Experiments
looking for deviations of complete absorption did not produce any result so
far.23
In a simple form, with only a single measurement, it can be written as

T r[Wadv U2 P (x)U1 Wret U1† P (x)U2† ]


p(x) = (61)
Σy T r[Wadv U2 P (y)U1 Wret U1† P (y)U2† ]

In this equation, Wret is a normal type density operator, describing the result
of a preparation procedure. In order to obtain a fully symmetric expression,
there is also an advanced density matrix Wadv , similarly describing the result of
a postparation procedure, executed at the end of the experiment. The peculiar
thing is that an adequate preparation usually requires all sorts of carefully man-
aged actions, such as collecting suitable equipment, building a set up, et cetera.
What we do after the final measurement, however, makes no difference at all,
unless we decide to cheat and postselect on the results of the final measurement.
If not, then surely we may be required to clean up and write a report, but this
has no effect on the measurement outcomes as such. For all that matters, Wadv
can always be omitted, which, in other words, amounts to

Wadv ∝ 1l or, dividing by a constant: Wadv = 1l (62)

This, again, expresses the condition of complete absorption. Clearly, this is


a condition in need of an explanation, and it can be seen as one of the basic
questions of cosmology. We shall proceed by discussing various solutions to this
problem.
22 Y. Aharanov, P.G. Bergmann, J.L. Lebowitz; Physical Review B, 134, 1964: 1410-1416,

reprinted in:
Wheeler and Zurek (eds.), Quantum Theory and Measurement, Pr.U.P., Princeton (1983).
23 B. Partridge; Nature 244, 1973, 263

36
2.3.1 From universal wave function to multiverse
Hugh Everett III was the first to assume the existence of a universal wave func-
tion24 . His theory is also called the relative state formulation, or the many
worlds theory, because the universal state does not undergo wave function col-
lapse. It is assumed to be a superposition of an ever increasing number of
branches, each containing a different version of the universe. Because different
branches contain different copies of ourselves, we always experience ourselves to
inhabit one particular branch and what we observe is relative to the branch in
which we actually find ourself. Hence the terms relative state formulation and
many worlds interpretation.
The branches are thought to originate from the interactions between separate
physical systems. In our universe, or at least in our part of the universe, it is very
common for physical systems to interact and become separated in a way that
creates correlations between them. In a standard quantum physical description,
this means that, in the state vector of the combined system, uncorrelated cross
terms tend to vanish quite quickly25 , e.g.:

| ψbefore > = | ψA , ψB > (63)


= ( α1 | ψA1 > +α2 | ψA2 >) ⊗ ( β1 | ψB1 > +β2 | ψB2 >)
becomes:
| ψafter > = α1 β1 | ψA1 , ψB1 > + α2 β2 | ψA2 , ψB2 > + vanishing cross terms

Looking at the state | ΨU > of the whole universe, the above process takes place
in a particular branch of the universe, and we can presumably write the initial
universal state, just before the interaction, in the form:

| ΨU >= c1 | Ψ1 > +c2 | Ψ >

where | Ψ >= | ΨR > ⊗ | ψA , ψB > is a branch of the universe containing the


particular correlating system we want to study.

For the state of this branch of the whole universe this means that it evolves
from :

| Ψ> = | ΨR > ⊗ | ψA , ψB >


into
| Ψ> = | ΨR > ⊗ ( α1 β1 | ψA1 , ψB1 > + α2 β2 | ψA2 , ψB2 > ) (64)

By learning about this process, we ourselves get correlated to this system in


a similar manner, and our actual experience is that, with probabilities |α1 β1 |2
and |α2 β2 |2 for the corresponding terms, we observe only one of them. Ac-
cordingly, we ourselves are present, in many different versions, in some subset
of the universal branches. Nowadays, this structure is also called a level III
Multiverse26 . I do not wish to enter into detail about different views on the
24 H. Everett, Rev. Mod. Phys. 29, 1957, 454
25 W.H. Zurek; Physical review D 24 (6), 1516; Physical review D 26 (8), 1862
26 e.g.: M. Tegmark Our Mathematical Universe, Penguin Books, Allan Lane, 2014 and:

[Link]
interpretation-of-quantum-mechanics/

37
existence and structure of multiverses, and what they mean or might mean for
the interpretation of quantum physics27 . I mainly have some questions and
suggestions.

2.3.2 Complete absorption in the multiverse

Questions concerning branching are also important because of a direct relation


between branching and the condition of complete absorption. To see this, let us
investigate the evolution of the universe, or a collection of its branches, in time.
First, a complicating factor is that the movement of time is not a globally ho-
mogeneous process. It depends on gravitational fields and velocities, quantities
that depend on the particular branche followed through the multiverse.
Nevertheless, it may not be too much to ask that there be timelike hyper-
surfaces (i.e. with a timelike normal vector everywhere in each branche) that
can be parameterized by a variable t, and that we may describe a collection
of worlds in terms of a quantum state | Ψ(t) > evolving as a function of this
parameter t.
With | Ψ(t0 ) > the state of the universe at the timelike hyperplane corre-
sponding to a time t0 , the evolution to an arbitrary later hyperplane t is given
by a unitary operator U (t, t0 ):

| Ψ(t) >= U (t, t0 )| Ψ(t0 ) > (65)

Assuming that evolution continues this way into an infinite future, what does
this mean with respect to the question of complete absorption? Lacking a final
time, or some other significant future moment, we can choose an arbitrary future
value of t and consider the density operators

W0 = | Ψ(t0 ) >< Ψ(t0 ) |


and
W (t) = | Ψ(t) >< Ψ(t) | = U (t, t0 )| Ψ(t0 ) >< Ψ(t0 ) |U † (t, t0 )

Calculating the resulting probability, by means of eq.(50), see eq. (66),

T r[W (t)U (t, t0 )W0 U † (t, t0 )]


p(Ψ(t)|Ψ(t0 )) =
T r[W (t)U (t, t0 )W0 U † (t, t0 )]
T r[W0 W0 ]
= =1
T r[W0 W0 ]
T r[1l U (t, t0 )W0 U † (t, t0 )]
= (66)
T r[1l U (t, t0 )W0 U † (t, t0 )]

then shows that the condition of complete absorption is trivially satisfied, as


is clear from the conclusion Wadv = 1l in the last line. Further, the argument
still holds in separate branches that evolve from the original state, i.e.: with
| Ψb (t0 ) > the initial state of such a branch, and | Ψb (t) > developing from it by
means of unitary evolution, the above derivation remains exactly the same.
27 See e.g. [Link]

38
In the reverse direction of time it all seems very different. The state that resulted
from the evolution between t0 and t, consists of a variety of branches. As we
find ourselves in one particular branch, without knowledge of the other ones
that branched away between t0 and t, applying the reversed evolution operator
U −1 (t, t0 ) does not bring us back to the original state | Ψ(t0 ) >. For this we
would need the interference with all branches that were lost on the way. As a
consequence, this type of evolution is irreversible, and the increase of entropy,
or loss of information, is one of its results.

2.3.3 Reversed branching and history dissolving

The above derivation of complete absorption hinges on the assumption that


branches are stable. Once branched remains branched. Is this a likely, or even
tenable assumption?
Why does history branch anyway, and why in such a time asymmetric man-
ner? It cannot be in the structure of the evolution operator. There is no
structural difference between U −1 and U . They differ only by a minus sign in
the exponent. Therefore, applying U −1 to a state Ψ in the present will produce
an evolution branching into the past, instead of merely producing the state it
evolved from. Accordingly, the reason for branching must be in the boundary
conditions.
Does reverse branching appear at all? ...
At a small scale it certainly does. For instance, a charged particle may send
out a photon and catch it back. Such self energy diagrams contribute to the
renormalization of masses and charges.
Could reverse branching, however, also be relevant at a larger scale?
As U −1 Ψ produces an avalanche of pasts leading to the same present, it may
also produce alternative timelines that are compatible with the present, but
different from each other. These alternatives must, of course, be compatible
with everything known about our history.
We know, for instance, that William of Orange was killed in the city of Delft
on the 10th of July in 1584, but there are also many things unknown about that
day. Did William, perhaps, also have a mosquito bite on his cheek? If so, there
is no evidence left about it.
Could it be, however, that this is not only unknown, but also in a deeper
sense no longer a part of our history? A fact, then, that has entirely disappeared
from reality, and to which U −1 Ψ may attach and make it into a superposition
of possibilities, rather than merely an unknown past state of affairs?
A counter argument to this possibility is that, even if the fact as such has
gone unnoticed, evidence about it should still be present somewhere, for in-
stance in the form of light escaped into the universe. Some group of aliens,
situated about 450 lightyears away, might currently be building a very powerful
telescope, and, pointing it at the Earth at the right moment, they might notice
the mosquito bite’s reddish swelling on poor William’s cheek.

39
This is not going to happen, however, at least in almost no branches of the
alledged multiverse. No alien is really going to build a telescope with galaxy
sized lenses in order to study minute details about our planet. In this case, how-
ever, all light escaped from the event is now simply a random bunch of photons
spreading out through the universe, instead of a collection of information.
This creates a situation somewhat comparable to EPR and delayed choice
experiments, where facts come to depend on what quantities are actually mea-
sured in the future. In short, if relevant information from escaped photons or
other historical evidence is not actually retrieved in some form or another, then:
• not only do we have no facts about historical details or a clear picture of
what happened;
• but there may also be no reason why U −1 could not attach us to more
than one possible state, i.e., to a number of different histories.
If this reasoning really holds true, but this may require a more extensive
examination, it means that backward branching on a macroscopic scale can
occur, and details about our entire history may be evaporating all the time.
Forgetting details about the past may have a deeper significance than merely
losing knowledge about what happened. It may be that the past really becomes
a superposition of different pasts. In this case, our history in the multiverse may
have a treelike head of branches, but it also leaves a tail of histories behind,
evaporating into an ever larger superposition of possible pasts. Instead of a
system of branches the result is more like a network.
If we disregard the possibility of backward branching, Everett‘s theory intro-
duces a very fundamental form of time asymmetry. If backbranching is included,
then clearly, in as far as ideas about the multiverse are correct at all, branching
seems to be extremely dominant over backbranching, at least in the present
period of the history of the universe.
Evidently, if backbranching would occur on a significant scale this changes
things rather fundamentally. This would also depend, however, on the details
of how it takes place. Merely a chaotic type of attaching the present to different
possible pasts, on a not too enormous scale, leads to a loss of history and maybe
not much more. Much like forgetting things, but then in a more fundamental
sense.
More systematic forms of reversed branching could lead to stranger things.
In principle it could interfere with the above derivation of complete absorption
and produce advanced effects. It could in principle also lead to strange effects,
such as precognition. (By the way, during my own lifetime I remember hearing
about two quite strong cases, one from first and the other from second hand,
and the first one even harder to explain as extreme coincidence than the other.
This said, and taken into account, I have no strong opinions on the subject. In
as far as I know, scientific research with paragnosts did never lead to significant
results.)

2.3.4 Reality and virtuality

In the previous section I have been nibbling a bit on the margins of our under-
standing of the many worlds view. Let us see if we can nibble just a little bit
more.

40
First, in what sense can we say that the universe splits into branches? Does
the entire universe split into different copies, or is it a local phenomenon, spread-
ing out into the universe if and when information about the split travels outward
to other parts? This at least seems to be clear from equation (64). In the ex-
pression on th r.h.s.:

| Ψ >= | ΨR > ⊗ ( α1 β1 | ψA1 , ψB1 > + α2 β2 | ψA2 , ψB2 > )

referring to the moment in time immediately after the split, the factor between
brackets refers to information restricted to the object and the measuring instru-
ment. It can get out into the universe only by means of further interactions,
such as reading the outcome and spreading the news in papers and broadcasts.
Before this phase, the information can be destroyed, and if this is done well,
then, if the arguments in the preceding section hold true, the rest of the universe
doesn’t split at all.
There are some more things to say, however. The interaction in eq. (64) is
of a rather special type. Many interactions may produce only weak correlations
between systems. What does this mean with respect to the development of
universal branches? Should we say that there can also be weakly separated
branches, or how should we describe this?
Another problem about the multiverse is that, if it is real, reality is a lot
bigger than we used to think. Moreover, (100−)% of it is forever invisible to us,
resulting in a theory that is almost entirely unfalsifiable. What we really have is
only the small part relating to observable physics. All the rest is extrapolation,
staying afloat only because it is a more or less straightforward extrapolation of
the current state of theoretical achievement.
There are certainly reasons for being serious about these views, and spend
work on their further development, but there are also reasons to stay critical
and be careful about replacing science by faith. We already have some existing
and rather perfect terminology for creating a useful distinction here. What
is actually observable, are real phenomena, and probabilistic relations between
them.
For explaining what we observe, we calculate Green functions containing
superpositions of virtual processes. I would suggest to maintain, and further
develop this into, a clear distinction between reality and virtuality, with reality
reserved for phenomena with a more direct relation to observation. The various
types of multiverses and speculations and theories about their structure can
then be regarded as referring to virtuality.
The suggestion is not that this makes the concept of a multiverse uninter-
esting or less serious, but it may provide some buffer against developing it into
a religion. The distinction between reality and virtuality can also serve to delay
having to choose between a many worlds and a single world view, perhaps in a
manner that stays closer to the Copenhagen interpretation.
Taking event theory as a starting point, the idea would then be that the
world consists of real events, that actually happen and can be observed and
measured. Relations between these events, in particular probabilistic relations,
are described by connecting these events by Green functions, containing prop-
agators and virtual events. Is this a somewhat likely position at all? I don’t
know, but it is nice to at least have some different alternatives in a situation as
strange as the one we appear to be in.

41
A consequence of this Copenhagenlike point of view would be that the mul-
tiverse explanation of complete absorption doesn’t work anymore. This view,
therefore, needs an alternative explanation for complete absorption, and I will
try to formulate some ideas in this direction.
In as far as known, the condition of complete absorption is satisfied, at least
for electromagnetic radiation. Barring the multiverse explanation, in what kind
of future could all this radiation be absorbed in such a homogeneous way? I
will briefly discuss two quite different ideas.

2.3.5 Complete absorption and Higgs decay


Suppose that the Higgs boson is or becomes unstable. Then, in a very far future,
all Higgs will decay, and all matter becomes massless. The result would be a
plasma of massless particles, many of them charged. All electromagnetic radia-
tion will obtain a finite penetration length and the universe looses transparency,
resulting in complete absorption.
Will this happen? Maybe not, maybe rather probably not. It is a possible
scenario, however, and it also provides a possibility of discussing some thermo-
dynamic considerations of a more general nature.
Assuming a very simple thermodynamic model take:
1 −E/kTret 1 −E/kTadv
Wret = e and Wadv = e (67)
Zret Zadv
from which we derive
1 −E/kTeff 1 1 1
p(E) = e with = + (68)
Zeff Teff Tret Tadv
From this, it follows that advanced effects vanish if Tadv >> Tret . This con-
dition, however, raises the question how, in an expanding universe, Tadv could
ever be that much higher than Tret . Average temperatures should decrease, not
increase, or some enormous input of energy would be required. Something like
Higgs decay could actually deliver this. All the energy presently contained as
mass would be gradually liberated by the decay of massive matter, resulting in
a plasma of particles moving at light speed, and in a much stronger radiation
field than otherwise expected.
Compressing this radiation back in time to our present era, the resulting
Tadv will indeed be a lot higher than the retarded temperature Tret . As a
consequence, the exchange of radiation is dominated entirely by the present
state of matter, and for the advanced state we have approximately Wadv ∼ 1l.

2.3.6 The universe inside a black hole


The Big Bang is the aftermath of what is usually considered to be a spacetime
singularity. This provides reason to assume that the entire universe as we see
it is actually the inside of a black hole, being the natural home for a spacetime
singularity. Thinking in terms of the produces quite a few things to think about.
In terms of the S3 T1 ⊗ s1 t3 model, the blacke hole becomes a timespace to
spacetime transition, i.e., formulated more precisely:
s3 t1 ⊗ S1 T3 → S3 T1 ⊗ s1 t3
This transition represents the internal structure of hte black hole.

42
To begin with, some thoughts about the external structure of a black hole.
Counting from the outside, in external time, the founding matter of the hole, and
anything falling in at a later time, disappears into the hole, whereas Hawking
radiation comes out at the surface, shrinking the hole until, in principle, it
disappears entirely.
The inside view of the hole shows a complementary, but rather different
picture.

Figure 2: A structure diagram of a black hole

Figure 2 shows an adapted version of a conformal diagram. As a two-


dimensional representation of an eight-dimensional situation, it may easily lead
to confusion. Anyway, as introduced before, the idea is that the left part shows
us TS⊗st (macro timespace ⊗ micro spacetime), and the right part vice versa,
ts⊗ST (micro timespace times the macro spacetime we regard as home). In the
spacetime part, on the right, parts of a few worldlines of matter and radiation
are shown. The arrows, in this case, are not meant to distinguish matter from
antimatter, but show the local direction of time, i.e., the direction of entropy
increase.
On a large scale, as seen in the outside superverse, there is a difference
between inflowing and outflowing matter, but from within the universe, this
does not seem to have a very obvious meaning. There is just space and time,
and radiation and matter. Is there, nevertheless, a way to identify infalling
against outflowing matter? With a few moments of thought, the outflowing
matter is on its way to becoming Hawking radiation, so it must be thermal.
Therefore, the 3K background radiation is a good candidate for being part of
it.
The 3K temperature, by the way, is nowhere near the Hawking radiation
temperature of a universe sized black hole. This means there is still a very long
time to go before reaching the surface of the hole, even with the accelerated
expansion rate caused by dark energy. Further, there must also be a different
component of background radiation, because conservation laws must be satified
between matter streaming in and out of the hole. Most ordinary matter in the
universe, however, seems too clumpy to ever become part of the Hawking radi-
ation. Neutrinos are better candidates, but then the question remains whether
they are sufficient for meeting all conservation requirements.

43
Anyway, at least at first sight, macroscopic objects, let alone cosmic struc-
tures like galaxies, stars, and planets, are not a likely ingredient of future Hawk-
ing radiation. Therefore, any kind of macroscopic stuff, including ourselves, is
almost certain to be made of matter falling into our universal black hole (in
some distant future, as reckoned in our present sense of time).
As an inspiration for science fiction writing, this raises the question whether,
in the distant future, our descendents will pass the Schwarzschild radius and,
moving out of our universe, become superverse inhabitants. Or, depending on
the direction of time they experience, might they be superverse inhabitants
falling into the large hole that is our universe? An intriguing question: what
kind of experience would it be to live in a situation where the direction of time
is not well determined? Presumably, at least in the present case, it is no use to
consider this matter, because the surroundings of a large black hole are likely
to be a hostile and violent environment. A site of mass destruction, no matter
from whatever direction of time it is entered.
Still, an interesting question is whether the universal black hole is rotat-
ing, and whether matter surrounding it has a nonuniform distribution, and a
preferred direction of movement, like in an accretion disc. Assuming time sym-
metric causality, such phenomena could lead to observable effects, by means of
a nonuniform distribution of mass and radiation also in the present universe,
and/or by an uneven distribution of advanced radiation. This could be, and/or
remain, a worthwhile subject for monitoring, as observational technique keeps
developing.

2.3.7 Strings
Going back to the S3 T1 ⊗ s1 t3 model for elementary particles (section 2.1.3) as
tiny black holes (section 2.3.6), it is clear that they can effectively be regarded
as strings. The fundamental difference with standard string theory is that they
are bridges between spacetime and timespace, instead of spacetime and itself.
There are a couple of reasons why this is, for me, a strongly preferred view.
A strong reason is that I am very much devoted to the concept of spacetime,
with all its dimensions intact. The current idea in string theory, of a universe
essentially three dimensional but emergent in time, repulses me deeply. Erik
Verlinde’s idea of emergent forces, as underlying present day versions of string
theory, is certainly interesting. For me, however, it would need to be trans-
lated into fully consistent spacetime formulation, and extrapolated to allow for
whatever number of space and time dimensions is needed in future theories.
For the rest, I certainly don’t have any objection against the idea of re-
garding a spacetime volume as a container, for information that can also be
expressed entirely on the hypersurface of the volume. In fact, this is precisely
what happens in dimensional reduction.28 At least in principle, this seems nicely
compatible with the idea of forces being emergent in some sense. For me, as
an outsider, looking from a distance with a rather superficial view, a thing that
seems worrying, however, is the ease by which the idea of emergent forces pro-
duces quite different phenomena. You can straightforwardly produce Newtonian
gravity, but with just a bit of modification you get general relativity quite as
easily, or so it seems, and even up to the slogan GR = QM.
28 Section (1.1.1)

44
Further, it may also be clear that, with respect to the model described in
section 2.3.6, the slogan EPR = ER is simply false. In this model Einstein Rosen
bridges connect spacetime to timespace (or, to be more precise, see section 2.1.3,
table 60, they are connections between ST⊗ts and st⊗TS). Moreover, EPR is a
problem that is not a problem, but merely a sign of the time symmetric nature
of causality.

2.4 Concluding remarks


The basic idea of the present essay was:
1 to present ideas about the structure of spacetime while:
(a) taking into account the structure of the particle spectrum, and
(b) adopting the idea that General Relativity is a theory for intermediate
distances, while failing at very small and very large distances.
2 to argue that the physics of this universe is, with respect to causality, ba-
sically time symmetric, with the observed time asymmetry a consequence
of boundary conditions, and more specifically a condition of complete ab-
sorption in the future.
3 to investigate possible explanations for complete absorption.
With respect to several ideas discussed about complete absorption, the final one,
about the universe as a black hole, is the one I find the most intruiging. Various
combinations of ideas are possible, however. For instance, the idea of Higgs
decay29 would automatically make final boundary conditions at the surface of
the universal black hole almost or totally invisible, independent of further ideas
about the future universe.
Having no career to loose, it was easy to feel quite unrestricted in what
I wanted to write in this essay, and it certainly added to the fun of writing.
Whether it also added to the truth content is not for me to decide, and, more-
over, with respect to ideas about truth I may refer to the third essay in this
bundle. With respect to sources of inspiration, I can refer to a few footnotes in
this essay, but more background is provided in my book (see footnote 5).

2.4.1 Icing on the cake


Now, finally, it so happened that I stumbled into an idea so strange that I’d
better forget about it immediately. On the other hand, as said above, I don’t
have a career to lose, and it reminds me of days long gone, when lazy summer
evenings could be spent on discussing the wildest ideas about whatever. So, for
the sake of nostalgia, a serious attempt to demolish whatever chance remains of
being taken seriously at all.
First, let us go back to the model of the universe inside a black hole in
section 2.3.6, and observe that it leads to a regression. The black hole which is
our universe must be inside the universe addressed in figure 2 as the superverse.
This superverse presumably originates from a Big Bang of its own, creating an
automatic progression towards ever larger universes. Are we, either by cosmic
coincidence or necessity, at the bottom of this structure?
29 Section 2.3.5.

45
The universe contains many black holes, large numbers of star mass sized
and center of galaxy sized black holes. At the inside, could these in any way
resemble a universe? In either case, the amount of time and space inside doesn’t
seem nearly enough to suggest any resemblence. The progression seems bound
to stop, and, in terms of black holes containing universes, our universe indeed
appears to be at the very bottom of the regression series.
Looking still further down, however, reaching the very bottom, in section
2.1.3, table (60), leptons and quarks were already introduced as a representation
of Spacetime ⊗ Timespace, i.e., very very tiny, but non evaporating black holes.
Combine this with the circumstance that particle theory can only be made
to work by means of extensive renormalization techniques, involving clouds of
unbound numbers of virtual particles, and gigantic or even infinite rescaling of
energies and other parameters involved.
Perhaps, here, at the very bottom, a regress is going on. Particles, each
containing a universe, containing particles...... A true multiverse fractal of uni-
verses. And isn’t there a beautifully charming strangeness in the idea that our
universe might be, e.g., the inside of, a quark?
In the meantime, we are now suddenly talking about countless numbers of
universes, but the total number of different universes is limited by table (60).
This table, however, may be read as refering to different orientations in Space-
time ⊗ Timespace, so maybe there is only one type of universe, that can be
turned in different directions. Or maybe, then, there is only one universe, being
a string, connected with countless numbers of strings to itself.
From this perspective, let us look again at our very far descendants, that
make it to the surface of the universal black hole and travel into the superverse.
What happens to them? They, presumably, become part of the cloud of virtual
particles surrounding the quark, or electron or whatever particle, that is our
universe. There, they will presumably be disintegrated in the gigantic scale
transformations associated with renormalization.
Summarizing this model into a slogan produces a nice finale for this essay:

Every table (60) particle in the universe, is the universe.

46
3 THE ORGANIC NATURE OF KNOWLEDGE
3.1 Truth and relevance:
the contextual nature of knowledge

Almost a century ago, by now, logical positivism stated that empirical science
rests on predictions that can be confirmed by empirical facts. It took a few
decades before Karl Popper came by and realized that general predictions cannot
be confirmed by any number of observations, and, instead, science must aim at
producing statements that are capable of being falsified30 . Still later, it became
clear that this is really only a very limited part of what scientists actually
do. Thomas Kuhn stated that scientists work on both sides, inside a logical
framework he called a paradigm31 . This is a large program, which, like a ship,
may encounter successes and drawbacks, but which may take quite a beating
before it sinks. Imre Lakatos32 takes a sort of middle position, and then there
was Paul Feyerabend, colourfully defending a form of scientific anarchism33 .

3.1.1 Truth
By their nature, different paradigms, or research programs, are difficult to com-
pare. Each one may have its own goals and organisational structure. Like ships
they may be firing at each other, but it may take a lot of hits to finally sink
them. The difficulties in comparing them, because of things like a theory de-
pendence of meaning, and other reasons for a vagueness in what they say or
might say, if correctly applied or further developed, are captured in the term
incommensurability.
While studying various aspects of theories of meaning and truth, it occurred
to me that relevance judgements could play an important role in explaining
the origins of incommensurable views. In particular, it occurred to me that a
relevance criterion could provide in an interesting way of generalizing the theory
of truth of Saul Kripke34 .
Kripke demands that a statement, in order to be true, must be grounded,
i.e., roughly, that its truth does not depend on any feedback loops. In this way,
the liar paradox, defined by the sentence: “This statement is false”, or :

φ = ¬T φ
is false because the sentence is circular. The idea behind this definition is that
truth should ultimately be decided by a dependence on nonlinguistic grounds,
in particular on empirical facts.
30 Popper, Karl, The Logic of Scientific Discovery, Harper and Row, New York, 1959.
31 Thomas Kuhn, The Structure of Scientific Revolutions, The University of Chicago Press,
Chicago, 1962
32 Imre Lakatos and Alan Musgrave (eds), Criticism and the growth of knowledge. Cam-

bridge University Press, 1974


33 Paul Feyerabend, Against Method: Outline of an Anarchistic Theory of Knowledge, New

Left Books, Londen, 1975


34 Saul Kripke: “Outline of a Theory of Truth”, The Journal of Philosophy, Vol 72, No. 19

(Nov. 6 1975), pp. 690-716

47
As a more general criterion, however, relevance can be required as a con-
dition for truth, i.e., a statement φ, in order to be true (T φ), must first be
required to be relevant (Rφ), i.e.:

T φ ≡ (Rφ ∧ φ) (φ is true if and only if φ is relevant and φ)


For the liar sentence, defined by φ = ¬T φ, we then obtain that:
- from φ follows ¬T φ, hence ¬Rφ or ¬φ,
- hence only ¬Rφ remains, because ¬φ would contradict φ.
Therefore, instead of a paradox, φ simply becomes a sentence from which its
own irrelevance can be deduced. Instead of linking relevance directly to ground-
edness, however, it can now be used in a wider sense, and in different and
methodologically stronger ways.
As a general principle, I would say that truth cannot be seen as separate
from its method of validation. Truth, at least in science, is not an eternal
God given entity. Because truth depends in an essential way on its method of
confirmation, relevance, in the form of satisfying the requirements of method,
is already an implicit but essential condition for truth. Further, as scientific
method develops, so does (do) the associated concept(s) of relevance, which can
also be generalized to include various methodological requirements and degrees
of trustworthiness and precision. Maybe, seen solely as a solution to the liar
paradox, the introduction of relevance as a condition for truth may seem rather
trivial, as the easy way out. Viewed from the other side, however, what is
the meaning of a totally abstract concept of truth, stripped from any method
of valuation? Frankly, I would say that relevance is at the core of our logical
machinery and truth its colouring at the surface.
Furthermore, now, I would also want to use the concept of relevance as
a general type of filters that govern the flow of information between contexts.
This, in turn, means that truth becomes contextual, and the flow of information
between contexts comes to depend on relevance judgements. In this way, I
would expect, it can also become an instrument for studying the phenomenon
of incommensurability, and provide tools for a more detailed understanding of
relations between different views.

3.1.2 Contexts
In order to understand how relevance is relevant to understanding more about
the postpositivism discussions, we need to see a connection between contextual
knowledge, relevance judgements, and the phenomenon of incommensurability.
Therefore, it must be clear how to define contexts, describe relations between
them, and deal with these things in a more formal analysis.

48
The most basic form of a context, in the sense in which I have used it, is,
what Bas van Fraassen calls, a semi interpreted language.35 This is a triple
< L, Γ, γ >, containing

• a language L;
• a state space, or logical space, Γ;
• a satisfaction function γ, assigning to each sentence ψ ∈ L the set of states
γ(ψ) ⊆ Γ in which it is satisfied.

Evidently, there are all kinds of syntactic and semantical items that can be
added to obtain a richer structure, but the above is sufficient for the present
purpose: introducing multiple contexts, with connective structure tying them
together.
The first thing, needed then, is a set of different contexts. This means that,
instead of a single triple < L, Γ, γ >, we need some, or even a lot of such triples
< Lu , Ψu , γu >, where u is a context defining variable.
The next thing needed is a structure that ties them together. For this latter
function, I introduced, in earlier work36 , the notion of a connective context,
i.e., a context with sufficient structure to relate to other contexts, and handle
relations between them.
A further item on the wishlist for a connective context was that there should
be sufficient freedom in connecting to other connective contexts, including itself.
In a practical setting, from daily life to scientific discourse, moving between
contexts, and introducing new ones whenever needed, can be seen as a very
generally and fluently applied mental ability.
In order to satisfy these conditions, I introduced the idea of a connective
context c, defined as a quadruple < Lc , Ψc , Γc , γc >, with:
• Ψc is a set of contexts u that can be accessed from c;
• Lc is a language capable of referring to at least part of the statements in
the contexts it can access.

• Γc is a state space, containing sufficient structure for interpreting all Lu


statements translatable into Lc . A trivial choice is to simply start with
the direct product of all the Γu for u ∈ Ψc . Imposing relations between
the contexts then boils down to carving out a substructure.
Allowing relevance to be context dependent, i.e., each context u can have its
own relevance concept Ru , makes the contextual structure a lot more interest-
ing, as well as more complicated. It makes it possible to study hierarchies of
relevance, and the way in which partially ordered relevance systems may lead
to incommensurable views. Particularly in a scientific environment, however, it
will often be considered important to increase conformity of relevance between
contexts, and/or to keep track of differences.
35 Bas van Fraassen: A Formal Approach to the Philosophy of Science, University of Pitts-

burgh Press, 1972


36 Dick Hoekzema: The Quantum Labyrinth, Kluwer Academic Publishers, Dordrecht, 1993

49
For handling relations between contexts, two types of contextual operators
were introduced:

• operators [u: ... ]: Lu → Lc such that a statement ψ ∈ Lu translates into


a statement [u : ψ] ∈ Lc
i.e. with ψ remaining in the syntax of u;
• operators [ ... ]u: Lc → Lc such that a statement ψ ∈ Lu translates into
a statement [φ]u ∈ Lc , where φ := Tcu ψ provides a translation into Lc of
the sentence ψ ∈ Lu .
(The difference roughly reflects the difference between “Jan says: ’het
regent pijpestelen in Amsterdam’ ” and “Jan says it is raining cats and
dogs in Amsterdam”).

3.1.3 Logical labyrinths


In earlier work, I referred to a network of connective contexts as a logical
labyrinth. One might opt for a more optimistic term, like a body of knowledge,
but, with respect to questions concerning incommensurability, and considering
the subtle sliding of concepts between subsequent theories and versions of the-
ories, the term labyrinth still strikes me as quite appropriate. In fact when I
first visualized this system, somewhere in the early 1980’s it actually dizzied
me. Later, I came to see the contextual referencing system as defining a sort of
www-like structure.
Probably the most important part of a logical labyrinth is a sufficiently
sophisticated relevance system, but this is really complicated matter. In actual
life, people deal quite fluently with sliding concepts and relevance assignments,
although often not in a very conscious way, not always successful, and not always
without disagreement and quarrel.
Already at a young age, children pick this up effortlessly. They become
interested in fairy tales, for example, but, unless explicitely fooled, no child
will take Cinderella in a literal way. What does attract them is the intensity of
the narrator, and hidden notions about suffering, perseverence, courage, dealing
with the unexpected, et cetera. These, apparently, convey notions that have a
clear relevance to them.
Also later in life, e.g. at university, teachers convey notions of what is, or
isn’t, relevant, not only by what they say, but also by enthousiasm, body lan-
guage, selection of tasks et cetera. Clearly, relevance comes, and is picked up,
in many, and sometimes hardly noticed ways. Precisely this ability, the only
partly conscious fluency, makes the subject hard to study.
A further complication is the ease at which we manage to assign and use
degrees of relevance, which certainly adds applicability, but also complexity.
Obviously, however, it is a very important skill, and we are quite accustomed
to using even this quite successfully in many situations. Also animals, by the
way, will usually deal quite effectively with , for example, the presence of food
and the simultaneous presence of a predator.

50
My motivation for defining a contextual logic of the above type was that it
is powerful, fits a wide range of human thought and communication, and still
has a fairly transparent logical structure. I also considered it as a promising
approach to artificial intelligence, but a follow up on this never came to be. It
can be argued, however, that the internet, and in particular the www, is moving
into the direction of a system of artificial intelligence and that the hardest and
least successful part of this is very strongly related to the lack of an effective
relevance system.
Considerations about these matters gradually moved me in the direction of
the views expressed in the next section.

3.2 Knowledge as a life form


The point I want to make in this section is that knowledge is similar to, and
develops very much like, life itself. In this picture, furthermore, relevance judge-
ments are steering its development. They are the genes of knowledge.
The first thing to observe about the similarity between life and knowledge
is that they both have a structure with fractallike properties, covering many
interconnected levels. Starting at the microscopic level of a single cell, we ob-
serve that even here we find a very high degree of structure, with a nucleus,
mitochondra or chloroplasts, a cell wall, and microscopic molecular machines,
transporting proteines and other stuff along molecular cables..., and all of this
finetuned in subtle ways in order to maintain some form of dynamic stability.
One level higher, cells are combined into organs, cooperating, again, in highly
structured and interdependent ways inside a body. Still higher, the organism is
part of some community of similar organisms, and of a biotope it shares with
different species of organisms.
Each of these levels itself, however, has a structure that can itself be said
to have the characteristics of a kind of organic structure, balancing itself in a
sort of slow moving equilibrium, as long as its environment does not change
too abruptly. Whenever change occurs, it can be due to any mixture of causes,
somewhere between completely random or completely predictable: a genetic
mutation, the arrival of an illness, a new predator, a competitor, a natural
disaster, ...
Comparing this to the history of knowledge, the picture is remarkably sim-
ilar. Specializing to science, we can observe it has many different fields, with
usually a considerable number of competing theories and research programs,
and with large numbers of models, pertaining to its various applications.
There is organisational structure, with university faculties, buildings, ma-
chines, and all kinds of further equipment. There are large volumes of text
and other educational materials: books, journals, lecture handouts and other
instruction texts, conference reports, exercises, instruction videos .... Looking
at humanity as an organism, the academia should be comparable to something
like a part of its brain.

51
3.2.1 Truth and reality
Among the various disciplines, natural science has the special function of inter-
preting nature, with the aim of constructing a truthful picture of the world we
live in, or more particularly of specific substructures of it. The idea of such a
picture is that it describes reality.
The concept of reality differs from truth in that it is thought to be indepen-
dent of theory or language, which is why it makes sense to say that a theory
can be wrong. However, because we can think about reality only in terms of
theories, or more elementary expressions of knowledge, reality itself remains an
abstraction. It is accessible only by means of theories, or descriptions of a more
basic type, acquired by observation and learning.
An interesting point, then, is the meaning of the words ”truthful picture”.
This can only relate to layers of information, gradually integrated into a more or
less consistent system, by which we can interpret observations in a meaningful
way. In other words, it relies on a long process, in which a large collection of
intentionally truthful pictures is improved, again and again, in order to obtain
a more reliable picture, which, tentatively can be considered to be true.

3.2.2 Research levels

In particular in the natural sciences it is often, or at least sometimes, or rather


implicitely, assumed that the scientific endeavour converges towards a true pic-
ture of the world. According to this view, adequate research should eventually
lead to THE truth about a subject. What would something like THE truth be
like, and how can we get there?
The idea of adequate research contains an implicit reference to scientific
method and organization, and to relevance decisions made during research.
About the organization of knowledge, I already mentioned terms like paradigm
and research program. Related concepts include theory, model, fact, and obser-
vation.
What is the precise meaning and usage of these terms and the relation be-
tween them? The reason for me to I ask this, is because I have been rather
confused about it myself. Actually, the most sense I could make of them is by
understanding them not in an absolute sense, but in relation to each other, as
a rough division in relative levels, and in sort of scale invariant manner.
Actually, thus, if I interpret it in this way, the relation between research
program and theory is much like the relation between theory and model. Putting
concepts in, what I would think is, the right order, I get the following sequence:

paradigm, research program, theory, model, fact, observation

From left to right, each step provides structure to investigation. From right to
left, information from reality that streams into the system can be used as a basis
for theory development, for instance by trying to hammer theoretical ideas into
a more fitting structure.

52
In agreement with the above suggestion to interpret the scheme in a scale
invariant manner, there is some fluidity in the way in which the above six levels
are used. What is called theory one occasion, may be called a research pro-
gram some other time, or by different people. What some call a fact, others
may describe as a theory laden model. Further, investigating specific scientific
endeavours, it may be that more levels can be distinguished than the six ones
mentioned above, or less of course. Getting used to a flexible use of terminol-
ogy may then be necessary to streamline discussions without introducing ever
more elaborate schemes. By and large, this should provide sufficient reason for
supporting a fluid use of these terms, as mainly expressing a relative status,
between them.

3.2.3 The dynamics paradigm

Long ago, one of my university teachers, introducing us to Lagrangian and


Hamiltonian systems, began his first lecture by stating that:

Physics is the science that studies how physical systems develop in time.

It immediately pushed me into contemplation for a while, because it struck


me as a rather awkard statement, for various reasons. It is partly circular, it
seems to discard the field of statics, and it singles out time as a special variable,
different from space.
In a nonrelativistic theory, the latter item is not a problem, of course. Spatial
variables are part of the state of a system, and in a dynamical theory one studies
how the state of a system varies in time. Let us call this structure of a theory
the dynamics paradigm.
Quite a while later, at the beginning of a course on relativistic quantum
physics and field theory (QFT), I had a similar, but much stronger experience.
An important feature of relativity is that space and time are dimensions of
spacetime, and should be treated as conceptually similar. During my very first
lecture on relativistic quantum physics it took me about two seconds to see that
this requirement was violated. I could not understand how, in a relativistic
theory, Hilbert spaces could still be spanned by vectors of type | x >, with x
a 3-vector, and how a quantum field ψ(x), with x a 4-vector, could be defined
as an operator on a Hilbert space of this type, without every serious physicist
bursting out in tears.
In a truely relativistic theory, one should always try to formulate everything
in a manifestly covariant manner, or so I was taught. Evidently, then, this should
especially be true for the Hilbert spaces that serve as the basic structure of a
relativistic quantum theory. I never doubted that this should be corrected, and,
in due time, as described in the first essay in this bundle, I became convinced
this can actually be done, with quantum event theory (QET) as a result. Indeed,
I hold that QET provides a really elegant way of dealing with this.

53
As I learned to understand a while later, the noncovariance of QFT can
also be polished away in a different manner, by defining it directly by means of
only the field equations, consisting of a set of differential equations describing
their behaviour in spacetime, and of equal time commutation relations express-
ing the quantum physical character of the fields. As a consequence, the Hilbert
space can now be regarded as just some structure suitable for providing a rep-
resentation of the field theory. It stays around merely as a convenience, not
as a fundamental entity. This way its nonrelativistic structure becomes less
problematic.
As discussed in the previous essay, however, it still remains a problem that
the equal time commutation relations force time into the role of dynamic vari-
able. This has a major effect on the heuristics of the theory, i.e., on expectations
about the direction in which the theory is expected to develop. After all, a liv-
ing theory is a project under construction, and, as part of a research program,
scientists are working on solving specific problems they see as a challenge.
QFT and QET are certainly not the same when it comes to their heuristics.
With field equations describing the flow of matter in spacetime, quite in style
with, e.g., a continuity equation for the flow of a fluid, and with time as the basic
dynamic parameter, QFT can be placed firmly within the dynamics paradigm.
This paradigm has ruled for a long time in large parts of physics, and physicist
feel more than familiar with it.
QET is clearly different, by basically describing the world in terms of a
static picture in spacetime. As explained in the first essay, the two descriptions
can be transformed into one another, if the necessary theoretical equipment is
present, which includes the insight that time is not the only possible choice for
the dynamic parameter. This view makes a basic difference with respect to the
heuristics of the theory.
Because QET and QFT are basically equivalent, however, the interesting
result is that we get a combined theory with expanded heuristics. There are
more ways to go, problems can be viewed from different sides, and perhaps even
combined in certain ways.
For example, with some phantasy: one might describe events ocurring in
a solid state system in an event theoretical manner, but with the background
system described by field theory.

3.2.4 Truth, with or without The ?

Among the various disciplines, natural science has the special function of inter-
preting nature, with the aim of constructing a truthful picture of the world we
live in.
An interesting point, then, is the meaning of the word ”truthful”. This can
only relate to many layers of information, gradually integrated into a more or less
consistent system, by which we can interpret new observations in a meaningful
way. In other words, it relies on a long process, in which a large collection of
intentionally truthful pictures is improved, again and again, in order to obtain
a more reliable picture.

54
Three important connotations of such a collection of intentionally truthful
pictures are consistence, completeness, and convergence. In this context:

• consistence means that two true accounts of the same situation cannot be
contradictory,
• complete is that, to a complete account, no relevant additions can be made.
• convergence is the idea that ongoing inquiry does, or can, or should, result
in getting ever closer to a consistent and complete account of the object
of investigation.
Now an interesting question is whether, and if so to what extent, convergence
leads to a result that is, in some relevant way, unique. If a number of separate
research groups, without mutual consultation, investigate the same subject and
each group independently converges toward a final theory, do they arrive at the
same result?
In particular in the natural sciences, it is often, or at least sometimes, or
rather implicitely, assumed that the scientific endeavour converges towards a
true picture of the world. According to this view, adequate research should
lead to THE truth about a subject. With truth being dependent on relevance,
however, and without mutual consultation as a way to reach conformity about
relevance judgements, how can one expect the two different converging views to
converge to the same point?
Part of this problem is, of course, that it is difficult to decide on what it
means to arrive at the same point. An analogy from evolution may serve to
clarify this question. Throughout history, and at different locations, evolution
leads to stable, or semistable local ecosystems, with more or less fixed ratios of
plants, plant eaters, and predators.
Contact between two ecosystems usually leads to dramatic changes, with
some species dying out, and other ones booming like pests spreading through
the system. It will usually take considerable relaxation times for the system to
settle down into a new equilibrium. In the process, major ecological structures
may survive, but be occupied by entirely different species, i.e., changes at the
level of genetics may be dramatically different from those at the ecological level.
Similar processes characterize the contacts beween cultures. Change now
takes place primarily at the cultural level, although this kind of contact may
also involve war, conquest, and resulting long term genetic changes.
Next question: what happens if two scientific cultures meet? Difficult to
test, because science is too internationally organized to provide us with two
really different sciences. Therefore, let us introduce an alien, space travelling
species, visiting us somewhere in the near future, and contemplate on whether
they might have, say, a truly different kind of physics. Is this possible, or is
our present theory of physics so close on the skin of the real world that this is
unimaginable?
Because truth depends on relevance, this would mean that relevance judge-
ments can be made in only one correct manner, and that there is a guaranteed
way to arrive at these correct decisions. Lacking a foreign species, drawing con-
clusions is difficult, but what we can try to do is to identify possibly interesting
relevance choices and examine them carefully.

55
Taking personal experience as a suitable case to study, the development
of QET has totally depended on a relevance judgement made during my first
lecture on relativistic quantum physics, and on a stream of subsequent coinci-
dences. I do not think that these personal experiences were unique in being
dependent on contingencies, however. It happens at much larger scales. Diffi-
cult to say, for instance, what things would have been like today, if Bohr and
Heisenberg would have pursued careers in pure mathematics, and De Broglie
had run the quantum show.
From this I do not want to conclude that science is a random walk. I certainly
do belief in some form of scientific progress, just like I do see that evolution does
seem to have some sort of direction. In both cases, however, it is difficult to
grasp precisely what it means, where it is going to, and how it comes about.

56
QET
Quantum Event Theory
appendix:
from Lagrangian to diagrams

57
Relativistic Quantum Physics:
• Standard versions of relativistic quantum physics still use Hilbert space
vectors of Rtype:
| ψ(t) >= ψ(x, t)| x > d3 x = ψ(x1 , x2 , x3 , t)| x1 , x2 , x3 > d3 x
R

• With t only as a dynamic parameter,


not represented as a separate dimension inside Hilbert space vectors.
• This breaks the relativistic concept of 4-dimensional spacetime.

Relativistic Quantum Field Theory:

Relativistic field equations, based on a Lagrangian, e.g.:

1
∂µ ψ∂ µ ψ † + ∂µ ψ † ∂ µ ψ − m2 ψ † ψ + ∂µ φ∂ µ φ − M 2 φ2 + gφψ † ψ

LQFT = 2

in which space and time do occur in the same way.


but:
• The fields are operators on a Fock space, i.e., a multi-object Hilbert space.
What is the structure of this Fock space?
• 1-particle states are still of type: | x >= | x1 , x2 , x3 >.
• and the fields have equal time commutation relations, like

∀t : [ψ(x, t), ψ † (y, t)] = δ(x1 − y 1 )δ(x2 − y 2 )δ(x3 − y 3 )

This still treats time in a special manner, forcing it into the role of a dynamical
variable, and it blocks the introduction of additional timelike dimensions.

Quantum Event Theory


Hilbert spaces H, built from vectors of type: |x = |x0 , x1 , x2 , x3 
Fock spaces F, for various types of events, like:
• multi-event vectors such as: ≺ a: x1 , x2 , ..., xn ; b: y1 , y2 , ..., yn ; c: z1 , z2 , ..., zn |

generated by a field a(x), its anti-field b(x), and an interaction field c(x),
• with commutation rules:
[a(x), a† (y)] = [b(x), b† (y)] = [c(x), c† (y)] = δ(x − y)

and all other commutators 0:

[a(x), b(y)] = [b(x), c(y)] = [c(x), a(y)] = [a(x), b† (y)] = [b(x), c† (y)] =
[c(x), a† (y)] = 0

58
Lagrangians correspond directly to QFT lagrangians; e.g., from:

1
∂µ ψ∂ µ ψ † + ∂µ ψ † ∂ µ ψ − m2 ψ † ψ + ∂µ φ∂ µ φ − M 2 φ2 + gφψ † ψ

LQFT = 2

follows:

LQET = ∂µ a∂ µ b − m2 ab + ∂µ c∂ µ c − M 2 c2 + gabc

With a corresponding action operator :

S = {∂µ a(x)∂ µ b(x)−m2 a(x)b(x)+∂µ c(x)∂ µ c(x)−M 2 c2 (x)+ga(x)b(x)c(x)} d4 x


R

Green functions:

Given a process state of the form

≺ Ψ| =≺ a: x1 , x2 , ..., xn ; b: y1 , y2 , ..., yn ; c: z1 , z2 , ..., zn |

Green functions are obtained as matrix elements of a Green operator |G , by


means of:

δS
Postulate 1: |G  = ia† (x)|G 
δa(x)
(69)
δS
i.e., for arbitrary Ψ: ≺ Ψ| |G  = ≺ Ψ|ia† (x)|G 
δa(x)
(70)
Postulate 2: ≺0|G = 1,

This leads to an iterative equation, from which ≺ Ψ | G  can be approximated


to arbitrary order.
Example: Starting from the free action:
Z
S = {∂µ a(x)∂ µ b(x) − m2 a(x)b(x)}d4 x

using Fourier transformation produces:

≺x|p = ≺ p | x ∗ = ≺ x | − p 
= eipx /(2π)2
0
x0 +p·x)
= ei(−p /(2π)2

which can be used to change the fields momentum representation.


Z
â(p) = ≺ p | x  a(x) d4 x
Z Z
4
b̂(p) = ≺ p | x  b(x) d x = ≺ x | − p  b(x) d4 x

59
This also gives us:
Z Z Z
â(p)b̂(−p)d4 p = a(x) ≺ p | x ≺−p | y  b(y)d4 xd4 yd4 p = a(x)b(x)d4 x
Z Z
because ≺ p | x ≺−p | y  d p = ≺ y | p ≺ p | x  d4 p = δ(y − x)
4

which further leads to:


Z Z
a(x)b(x)d4 x = â(p)b̂(−p)d4 p
Z
S = (p2 − m2 )â(p)b̂(−p) d4 p
δS
|G  = (p2 − m2 )â(p)|G  = ib̂† (−p)|G 
δ b̂(−p)

From (p2 − m2 )â(p)|G  = ib̂† (−p)|G  we obtain,

ib̂† (−p)
â(p)|G  = lim |G  = iK̂a (p)b̂† (−p) |G 
→0 p2 − m2 + i

where K̂(p) := {p2 − m2 + i}−1 .

Changing back to position representation:


Z
a(x)|G  = ≺ x | p  â(p) d4 p |G 
Z
= ≺ x | p  iK̂a (p)b̂† (−p) d4 p |G 

(and, because: ≺ z | − p =≺ p | z  )


Z
= ≺ x | p  iK̂a (p) ≺ p | z  b† (z) d4 z d4 p |G 
Z
= iKa (x, z) b† (z) d4 z |G 

with: Ka (x, z) = lim ≺ x|{p2 − m2 + i}−1 |z 


→0

The result is that:

≺ a : x; b : y | G  = ≺ b : y|a(x)|G 
Z
= ≺ b : y| iKa (x, z) b† (z) d4 z |G 

= iK(x, y) ≺ 0 | G 
= iKa (x, y) i.e., a Feynman propagator, as expected

60
Proceeding in the direction of Feynman diagrams, the outcome
≺ a : x; b : y | G = iKa (x, y) can be graphically represented by:

Figure 1: The Green function ≺ a : x; b : y | G = iKa (x, y)

Next challenge: bring back the interaction term(s), in our case consisting of:
Z
g
â(p)b̂(q)ĉ(−p − q) d4 q in the momentum representation.
(2π)2

Going back to the expression:


!
δS
|G  = (p2 − m2 )â(p)|G  = ib̂† (−p)|G 
δ b̂(−p)

the interaction term now changes it into:


 Z 
g
(p − m )â(p)|G  = ib̂† (−p) −
2 2 4
â(q)ĉ(p − q)d q |G 
4π 2

and, therefore:
 Z 
ig
â(p)|G  = iK̂a (p) b̂† (−p) + â(q)ĉ(p − q)d4 q |G 
4π 2

Transforming back to the position representation, this results in:


Z  
† ig
a(x)|G  = iKa (x, z) b (z) + a(z)c(z) d4 z

Looking back at the matrix element ≺ a : x; b : y | G , the result is now that

≺ a : x; b : y | G  = ≺ b : y|a(x)|G 
Z  
ig
= ≺ b : y| iKa (x, z) b† (z) + a(y)c(z) d4 z|G 

Z
ig
= iKa (x, y) + iKa (x, z) ≺ a : z; c : z | G  d4 z

61
As before, one item is removed from the state, giving a Feynman propagator in
return, but there is also an extra term, returning two new events into the state,
at the vertex z. In a diagram it looks like this:

Figure 2: Interactions develop new vertices

The second term contains a live |G -part. Changing the notation a bit, because
of a lack of characters beyond xyz, it can be further developed into, for instance:

Figure 3: into ever higher orders of g

Graphically it looks as if we are pulling interactions out of the G-circle. We


can do this in different ways, however, by making different choices. Any of the
lines touching the G-circle, can be taken as the starting point for adding further
interactions to the process, because any of the equations

δS/δa|G = ia† |G  or δS/δb|G = ib† |G  or δS/δc|G = ic† |G 

can be used to continue the calculation, e.g.:

In the wiggly line represents an interaction

boson. According to the Lagrangian, it may decay into a particle-antiparticle


pair, which would change the diagram into:

Figure 4: and in various ways

62
Evidently, whatever choice we make should lead to the same outcome. So sup-
pose we stop at the second order of g, and let us compare the outcomes. Let us
first go back to the diagram of figure 3. In second order, we can stop introduc-
ing further vertices, and finish calculations by connecting the remaining events
touching the G-term.

For then, the final second order term is:

, giving:

= iKa (x1 , x3 )iKa (x3 , x4 )iKc (x3 , x4 )iKa (x4 , x2 )

Alternatively, for the diagram

the two final second order terms are:

and

Any term of the latter type vanishes, however, so, in both cases, the final second
order result is that:

i.e.:
≺ a : x; b : y | G =
 2
ig
iKa (x1 , x2 )+ (2π) 2 iKa (x1 , x3 )iKc (x3 , x4 )iKa (x3 , x4 )iKa (x4 , x2 )+O(g 3 )

In the end, the procedure is similar to using parturbation theory: All diagrams
up to given order are collected, and all contributions calculated and added.

63
Question: How does this compare to standard quantum field theory?
Answer: For QFT, an iterative procedure similar to the one obtained above can be
obtained by letting the ∂ 2 operator work on the vacuum expectation of a
time ordered product of fields, i.e. like G =< 0 |T [ψ(x1 )ψ † (x2 )]| 0 >.
(derivation given in Hoekzema[1993]. A shorter example is contained in
this document)
Conclusion: It is possible to develop a quantum theory that is fully covariant in every
respect, including the Hilbert space level. It enables a full symmetry
between time and space, straightforwardly including, e.g., the possibility
of adding multiple space and time dimensions.

64

View publication stats

Common questions

Powered by AI

The concept of spacetime-timespace duality suggests that particles are tiny black holes with sides in both a macroscopic spacetime and a microscopic timespace, and vice versa for their other half. This view implies that black holes are transitions between these dual spaces and are no longer isolated systems, thus affecting the unitary connection typically considered in black hole evaporation .

Integrating Quantum Event Theory (QET) with general relativity could provide a model that unifies the understanding of interactions across different scales and resolves singularities. It could potentially address inconsistencies at very large and small distance scales, complementing general relativity's framework by incorporating mechanisms for event correlation within a covariant structure .

Dynamic reduction in QET influences the formulation of a time operator by eliminating the need to align it with the Hamiltonian's bounded spectrum. This approach allows a covariant structure where the time operator functions independently of the Hamiltonian by using the operator i∂0, thereby maintaining the integrity of covariant expressions .

Quantum Event Theory (QET) is a quantum theory based on a Hilbert space constructed with 4-vectors, involving coordinates that refer to events in spacetime. It maintains the same status for the uncertainty relation between time and the time-like component of momentum as for spatial components, and assigns probabilities to histories rather than states, avoiding state reduction or wave function collapse. QET differs conceptually from Quantum Field Theory (QFT) as it connects to standard forms of quantum physics in non-trivial, yet straightforward ways and promotes a covariant, event-based approach to quantum physics .

In Quantum Event Theory (QET), the generator of time translations is not the Hamiltonian, as in standard quantum physics, but the operator i∂0. This allows a covariant structure where the time dimension is not reduced dynamically, differing from traditional quantum physics where the Hamiltonian's bounded spectrum prevents the existence of a time operator .

Pauli argued against the existence of a time operator in conventional quantum physics, given the Hamiltonian's role in generating time translations. QET, however, replaces the Hamiltonian with the operator i∂0 as the generator of time translations, enabling a covariant formulation that maintains a time operator .

Green functions serve as the crucial link between Quantum Event Theory (QET) and Quantum Field Theory (QFT) by correlating events. In QET, Green functions are obtained as matrix elements of a Green operator, which also connect to QFT where Green functions link fields and observational results .

Considering the universe as a black hole suggests that observed time asymmetry is due to boundary conditions, specifically complete absorption in the future. It indicates that the universe's physics is fundamentally time symmetric concerning causality, where complete absorption explains the observed time directionality .

Boundary conditions, particularly the concept of complete absorption in the future, are significant in explaining the observed time asymmetry in the universe. They suggest that while the universe's fundamental physics is time symmetric, these boundary conditions give rise to the apparent directional flow of time, providing a coherent account of time asymmetry in a cosmological context .

The S3T1 ⊗s1t3 model addresses the differences between space and time by considering three macroscopic spatial dimensions versus one temporal dimension as coincidental. It suggests an alternate universe configuration with different macroscopic dimensions. This model also explores the directional flow of time from past to present, reflecting fundamental causal relationships perceived by observers .

You might also like