Humana - Mente 13 Physics and Metaphysics

H
ISSUE 13 - APRIL 2010

Physics and Metaphysics
EDITED BY CLAUDIO CALOSI
UMANA
M
ENTE
EDITORIAL MANAGER: ALBERTO PERUZZI - UNIVERSITY OF FLORENCE
DIRECTOR: DUCCIO MANETTI - UNIVERSITY OF FLORENCE
VICEDIRECTOR: SILVANO ZIPOLI CAIANI - UNIVERSITY OF MILAN
INTERNATIONAL EDITORIAL BOARD
JOHN BELL - UNIVERSITY OF WESTERN ONTARIO
GIOVANNI BONIOLO - INSTITUTE OF MOLECULAR ONCOLOGY FOUNDATION
MARIA LUISA DALLA CHIARA - UNIVERSITY OF FLORENCE
DIMITRI D'ANDREA - UNIVERSITY OF FLORENCE
BERNARDINO FANTINI - UNIVERSIT DE GENVE
LUCIANO FLORIDI - UNIVERSITY OF OXFORD
MASSIMO INGUSCIO - EUROPEAN LABORATORY FOR NON-LINEAR SPECTROSCOPY
GEORGE LAKOFF - UNIVERSITY OF CALIFORNIA, BERKELEY
PAOLO PARRINI - UNIVERSITY OF FLORENCE
JEAN PETITOT - CREA, CENTRE DE RECHERCHE EN PISTMOLOGIE APPLIQUE
PAOLO ROSSI MONTI - ACCADEMIA NAZIONALE DEI LINCEI
CORRADO SINIGAGLIA - UNIVERSITY OF MILAN
CONSULTING EDITORS
CARLO GABBANI - UNIVERSITY OF FLORENCE
ROBERTA LANFREDINI - UNIVERSITY OF FLORENCE
MARCO SALUCCI - UNIVERSITY OF FLORENCE
Elena Acuti, Scilla Bellucci, Laura Beritelli, Alberto Binazzi, Matteo Borri,
Giovanni Casini, Roberto Ciuni, Chiara Erbosi, Marco Fenici, Riccardo Furi,
Matteo Leoni, Stefano Liccioli, Umberto Maionchi, Daniele Romano
HUMANA.MENTE - QUARTERLY JOURNAL OF PHILOSOPHY
Editorial
Board
Editorial
Staff
TABLE OF CONTENTS
INTRODUCTION

Physics and Metaphysics. Introduction p. III
REPORTS

The Third Interdisciplinary Ontology Conference
Tokyo, Japan, 27-28 February, 2010
reviewed by Michele Pasin
p. XIX
Winter School: "Open Problems in Philosophy of Sciences"
Cesena, Italy, 15-17 April, 2010
reviewed by Pierluigi Graziani p. XXV
PAPERS

Mauro Dorato
Physics and Metaphysics: Interaction or Autonomy? p. 1
Gabriele Veneziano
String Theory: Physics or Metaphysics? p. 13
John Norton
Time Really Passes p. 23
Sam Baron, Peter Evans, Kristie Miller
From Timeless Physical Theories to Timelessness p. 35
Vincent Lam
Metaphysics of Causation and Physics of General Relativity p. 61
Claudio Garola, Sandro Sozzo
Realistic Aspects in the Standard Interpretation of Quantum Mechanics p. 81
Giovanni Macchia
Expansion of the Universe and Spacetime Ontology p. 103
Adriano Angelucci, Vincenzo Fano
Ontology and Mathematics in Classical Field Theories and Quantum Mechanics p. 139
Tracy Lupher
Not Quite Particles, Not Quite Fields p. 155
Friedel Weinert
Relativistic Thermodynamics and the Passage of Time p. 175
BOOK REVIEWS

Bastiaan C. Van Fraassen - Scientific Representation: Paradoxes of
Perspective
reviewed by Bradley Monton p. 193
Yuri Balashov - Persistence and Spacetime
reviewed by Lorenzo Del Savio p. 201
Steven French and Decio Krause - Identity in Physics: a Historical,
Philosophical and Formal Analysis
reviewed by Giovanni Casini p. 209
Tim Maudlin - The Metaphysics Within Physics
reviewed by Emanuele Coppola p. 213
Carlo Rovelli - Anaximander
reviewed by Umberto Maionchi p. 219
COMMENTARIES

A Structural Interpretation of Pure Wave Mechanics
edited by Jeffrey Barrett p. 225
Metaphysical Language, Ordinary Language and Peter van Inwagen's
Material Beings
edited by Daniel Nolan p. 237
Michael Lockwood - The Labyrinth of Time: Introducing the Universe
commented by Giuliano Torrengo p. 247
Robert Geroch - General Relativity from A to B
commented by James Weatherall p. 259
INTERVIEWS

Adolf Grunbaum p. 267
Introduction

Claudio Calosi*

claudio.calosi@humanamente.eu

1. PHYSICS AND METAPHYSICS. AN OVERVIEW
It is notoriously difficult to define Metaphysics
1
, its content, its method, its
language, its scope. Thus I will not even try an attempt here. I will rest content
to point out some widely held characterizations. A long and highly influential
tradition maintains that Metaphysics is the study of being qua being. It is
concerned with what there is, what kind of things are the things that there are,
what properties do they have, how they are related. In this sense Metaphysics
deals with the more general features of reality, the most fundamental categories
of being. Call this tradition General Metaphysics.
It is well known that empiricism of any sort
2
has always been very skeptical
of the very possibility of such an enterprise, at least one it considered an
enterprise that should be carried out a-priori. Kants transcendentalism
somehow endorsed this skepticism about General Metaphysics yet it did not
dispense with Metaphysics in general. Metaphysics, according to this Kantian
standpoint, is nothing but the clarification of the most general structures at
work in our knowledge of the world. Call this Transcendental Metaphysics.
Strawson (1959) famously introduced a distinction between Descriptive and
revisionary, or Prescriptive Metaphysics. Descriptive Metaphysics aims to
describe the most general features of our conceptual scheme. Prescriptive
Metaphysics, on the other hand, attempts to revise our ordinary way of thinking
and our ordinary conceptual scheme in order to provide an intellectually and
morally preferred picture of the world. It could be argued, though this might
turn out to be a controversial claim, that in some sense Descriptive

* Department of Philosophy University of Florence
1
See for example discussion in Loux 1998 and van Inwagen 2007. There is whole industry now
in the analytical community that deals specifically with such questions. It is called Meta-metaphysics.
2
Both classical Empiricism of Locke and Hume and logical Empiricism of Carnap, Reichenbach,
Ayer.
IV Humana.Mente Issue 13 April 2010

Metaphysics is the continuation, or a contemporary variant, of Transcendental
Metaphysics and that Revisionary Metaphysics is a continuation, or a
contemporary variant of General Metaphysics. It remains the fact that both
traditions are alive in contemporary analytical philosophy.
3

Physics on the other hand might at first seems easier to define, at least since
the publication of Galileos Discourses and Mathematical Demonstrations
relating Two New Sciences in 1638, that can as well serve as the birthmark of
modern mathematical physics. A very rough characterization of it will probably
mention the study of matter and motions of matter through spacetime. In a
broader sense Physics deals with the general analysis of nature, the world and
their components.
4

It is immediately clear, even from these very sketchy presentations how wide
and deep is the area of overlap between these two disciplines. For clearly,
matter, space, time and so on do seem, at least at first sight, good candidates
for the alleged general categories of being or general categories of human
understanding. Another, I believe straightforward, empirical argument in favor
of the existence of such a deep an wide overlap, comes from a look at
contemporary introductions to Metaphysics, even if it is a brief and quick
look.
5
They almost invariably contain materials on space, time, causation,
constitution of material objects, identity, determinism and free will and so on.
But these are the very notions Physics is supposed to be about.
6

If my argument is sound then this overlap immediately raises deep issues
about the relationship between Physics and Metaphysics. There is almost an
infinite variant of positions one might hold. I cannot do justice to them here.
7

So I will rest content at rehearsing some of those. It seems to me that the two
extreme positions one might maintain can be labeled Metaphysical

3
I invite to read this claim not as a militant claim but rather as a pragmatical one. It just serves to
purpose of delimiting boundaries that, even if artificial, are sometimes useful if not implicitly
sustained.
4
I use these terms in a very loose way. Thus I do not want to suggest that there is a difference
between the world and nature or that components should be understood as parts or participants for
example. Mine is a claim of humility. An introduction is not the place to settle discussions that have
shaped the very course of Western Thought.
5
See for example again Loux 1998 and references therein.
6
Probably free will will not be explicitly mentioned in any physics textbook. However in
contemporary metaphysics the issue of free will is deeply connected to the issue of determinism, and
this is surely a topic physics deals with.
7
I refer the reader to the excellent Dorato (this volume).
Introduction Physics and Metaphysics V
Foundationalism (MF) and Physical Eliminativism (PE).
8
According to MF,
Metaphysics is the study of the most general feature of reality independently of
any particular science. Metaphysics provides the general framework in which
any empirical considerations, physics included, become meaningful.
9
Thus
Metaphysics is the foundation of every empirical science, Physics included. I
doubt that anyone would endorse such an extreme version of foundationalism
nowadays, although it could be argued that formal ontology is precisely a
contemporary variant of such attitude.
10

On the other hand PE maintains that there are no genuine metaphysical
problems. There are only the empirical questions asked by Physics. Suppose
for sake of argument that Physics could solve all of its problems. Then there will
be no problems left to solve.
Between these two extremes there is a wide range of options. One might
uphold a sort of Naturalistic Attitude (NA). Proponents of NA do not deny that
there are genuine metaphysical questions about, for example the nature of
existence, but they do deny that those questions can be solved or even
formulated or even arise
11
independently of any physical considerations.
12

According to such an attitude whether God exist is not a genuine problem after
all
13
. But it is a genuine metaphysical problem whether the electromagnetic
field exists and whether it supervenes or not on charged particles. But naturally
this question does somehow depend on physical considerations, mainly

8
Here and in what follows I will not make any attempt to decide whether any particular
philosopher would endorse any particular thesis.
9
I am perfectly aware that strictly speaking this claim does not follow from the previous one. But
I am not attempting to give a rigorous definition of Metaphysical Foundationalism here.
10
One of the authors in the volume, Vincenzo Fano, has suggested me a different taxonomy of
the possible relationships between Physics and Metaphysics. Here I briefly sum up his argument. i)
Genuine metaphysical problems are just foundational problems of physics, or foundational problems
of natural science in general. It would be interesting to assess whether this is my NA. ii) There are
genuine metaphysical problems but they have to be formulated keeping in mind the technical
resources used by physical sciences. Again, it is interesting whether this is an instance of my NA again.
iii) Metaphysical problems are independent from physics in their formulation but they might be solved
by physics. iv) Physics has no relevance whatsoever for metaphysics. This seems to be a stronger
variant of my IT. v) There are no metaphysical problems. It is probably a strong variant of my PE.
Thanks to Vincenzo Fano for helpful comments on a previous draft of this work.
11
Depending probably on the strength of such naturalistic attitude.
12
Or broadly speaking independently of any considerations drawn from natural sciences in
general.
13
Though this might be a strong controversial claim even for those who have naturalistic
inclinations. Some of them will probably argue that Physics do have something to say about that
question. And some would go probably as far as saying that Physics does settle that question.
VI Humana.Mente Issue 13 April 2010

classical electromagnetism and quantum electrodynamics.
Another possibility would be to hold an Independence Thesis (IT). IT
would probably claim that Physics and Metaphysics are two independent
disciplines with their own language, their own methodological and theoretical
components. Sometimes they do overlap. And when they do they are both best
understood as incomplete descriptions of the same portion of reality.
Someone willing to accept the distinction in Strawson 1959 that I have
mentioned above might want to argue that Physics and Metaphysics are both
independent and distinct and that the only viable Revisionary Metaphysics is
Physics and that the only viable Metaphysics is Descriptive Metaphysics. I leave
it to the reader to explore whether this option is just a variant of NA, IT or a
combination of both.
It is clear however, or, at least it should be clear, that whatever thesis one
might hold about the relationship between Physics and Metaphysics this calls
for substantive argument.
14

I am personally inclined to think that Metaphysics without Physics is blind
and physics without metaphysics is crippled. This claim should be understood
tentatively along the following lines. Our metaphysical theories should be
informed by our best, experimentally successful physical theories. I would
probably go as far as claiming that a contradiction with a well confirmed
physical theory should be a reason good enough to seriously consider the
possibility that a certain metaphysical theory is simply false But it is also the
case, I believe, that our best physical theories are not metaphysically
transparent. To read off a particular metaphysics from a physical theory
sometimes, if not always, requires substantive work that is not and cannot be
done by the physical theory itself. I would probably go as far as claiming that
there are genuine metaphysical questions for which physics by itself does not
have the answer. This conviction was what first motivated me to embark in the
present work. It follows from this conviction that an interaction between
Physics
15
and Metaphysics is necessary and should be welcome. In what follows
I will provide what I take to be an interesting case of fruitful interaction
between those two.

14
Again, see Dorato (this volume) and references therein.
15
And naturally philosophy of physics.
Introduction Physics and Metaphysics VII
2. PHYSICS AND METAPHYSICS. AN EXAMPLE
My arguments in section 1 notwithstanding, it is a widely recognized fact that
Metaphysics and Physics have been rather self isolated enterprises, even in the
analytical community. On one hand, metaphysical issues about identity,
location, persistence through time, material composition, causation and so on
have rarely been discussed within the framework of physical theories. On the
other hand, Physics and philosophers of physics have somehow endorsed a
form of skepticism along the lines I have sketched in section 1 about the
possibility for Metaphysics to provide a consistent and valuable view of how the
world is. In recent years however there has been a tendency to bridge the gap
between the two. Considerations drawn from physical theories have played a
major role in metaphysical disputes like the ontology of time, nature of
persistence, theory of identity and even mereology, to name just a few. This
section explores one particular case in which considerations drawn from
Physics and philosophy of Physics have been fruitfully used to deepen, clarify,
and arguably, solve classical metaphysical issues.
16
The case I have in mind is
Special Theory of Relativity
17
, and its consequences for Metaphysics of Time
and Metaphysics of Persistence. I am choosing this example for different
reasons. First of all providing a general and compelling argument about the
relationship between Physics and Metaphysics is far beyond my possibilities. I
leave this problem to better hands than mine.
18
Second this example is what I
am mostly concerned with. And finally it is briefly mentioned in various works
in the present volume
19
yet not directly addressed by any of those. So hopefully
this will not affect any reading of the papers, which are the main strength and
should be the main focus of the present issue. These seem to me good enough
reasons.

16
Again, I leave it to the reader to judge if my own take of the problems is an instance of NA or
not.
17
STR from now on.
18
And again here, see Dorato (this volume).
19
See in particular Norton, Barons, Evans and Miller, Weinert, Torrengo and Dorato (this
volume).
VIII Humana.Mente Issue 13 April 2010

2.1 SPECIAL THEORY OF RELATIVITY AND METAPHYSICS OF TIME
20

There are famously three different Metaphysical theories about Time
21
, namely
Presentism, Possibilism and Eternalism. They can be roughly defined along the
following lines:
(1) Presentism i) Only the Present exist, the Past and the Future do
not, and
22
ii) Only Present Objects exist.
23

(2) Possibilism i) Only the Present and the Past exist, the Future
does not, and ii) Only Present and Past Objects exist.
24

(3) Eternalism i) Past, Present, Future, they all exist, and ii) Past,
Present and Future Objects they all exist.
It is widely held that STR provides one of the most compelling arguments
against Presentism. If this is indeed the case
25
then this is a clear example in
which considerations drawn from a specific physical theory are used to solve
traditional metaphysical problems. It should be noted however that physical
considerations play a much subtler role than it is usually recognized. For they
do not enter only in the solution of the metaphysical problem. They enter, or
better, I think they should enter, even in the formulation of it. Claims (1)-(3)
are cast in temporal language. It is however controversial whether there is time
at all in a relativistic world.
26
Startling as it might be this claim has authoritative
defenders.
27
Even if someone is not willing to go as far as denying the existence
of time in a relativist world one might still worry about the fact that STR is best

20
This section is not supposed to be an exhaustive treatment of such issues. It should give the
reader a flavor, so to say, of how a fruitful interaction between Physics and Metaphysics might work.
For a careful analysis see Calosi (unpublished).
21
I am taking for granted that these debates are genuine metaphysical debates and not just
semantic debates in disguise. Those who are inclined to endorse such a semantical skepticism should
read my claims counterfactually.
22
The justification for this conjunction is based upon some implicit technical assumptions about
reducibility of objects to spacetime regions. It should be noted that the presentism eternalism debate
is not to be confused with another classical debate in philosophy of time, namely the debate between
A-theory of time and B-theory of time.
23
I am personally inclined to read this claim within the framework of a formal theory of location.
Such a theory is a formal theory in the logic sense. It is a set of definitions, axioms and theorems in the
language of the first order calculus. See again Calosi (unpublished).
24
In what follows I focus on Presentism. Most of the arguments will apply to Possibilism too so
there is no need to distinguish here.
25
As I maintain it is.
26
See the excellent Barons, Evans and Miller (this volume).
27
Most notably Barbour, Pooley and Stein.
Introduction Physics and Metaphysics IX
understood as a spacetime theory that attributes a particular geometric
structure on the world, namely that of Minkowski spacetime.
28
And Minkowski
spacetime is not just space and time.
29
Thus one might endorse some sort of
Supervenience thesis according to which temporal facts supervene on
spatiotemporal facts. If so, my formulation in (1)-(3) is, if not mistaken, at least
misleading. Fortunately enough it is not difficult to find a formulation of such a
debate that is more relativistic friendly. It can be reconstructed along the
following lines:
(4) Relativistic Presentism: i) There exists only one particular
subregion of Minkowski spacetime that is called the Present, and
ii) there exist only those objects that are located at that subregion
of Minkowski spacetime that is mentioned in i).
30

(5) Relativistic Eternalism: i) There is no ontological distinction
between different subregions of Minkowski Spacetime, and ii)
objects do not lose or acquire any particular ontological status
just by be merely located at them.
There are at least two main relativistic arguments against Presentism, both as
defined as in (1) or (4). I label them The Relativity of Simultaneity Argument
and the No Spatially Extended Present Argument. Here is a brief
reconstruction of both, starting with the Relativity of Simultaneity Argument:
(6) STR is true.
(7) If STR is true there is no absolute, i.e., frame independent,
relation of simultaneity.
31

(8) If Presentism is true than there is absolute simultaneity.
(9) Hence Presentism is false (by (6), (7), (8)).
The No Spatially Extended Present Argument instead runs roughly as follows:

28
That is a n-dimensional metric affine space with signature (1, 1-n) where n = 4.
29
The reader should grant that for sake of argument, namely the fact that time is not just the
timelike submanifold of Minkowski spacetime and space is not just the spacelike submanifold
orthogonal to it. Again, to see how an argument towards this conclusion can be constructed, see
Barons, Evans and Miller (this volume).
30
Even this formulation is not satisfactory in many ways. I cannot enter into these subtleties here.
31
This follows from the following facts about Minkowski Spacetime. Simultaneity is represented
geometrically by Minkowskian orthogonality, i.e., two events p and q are simultaneous relative to a
timelike line L iff <pq, u> = 0 where u is an arbitrary timelike vector that spans L. Vectors are written
in bold characters. But different timelike lines will single out different spacelilke submanifolds
orthogonal to them and so different events could count as simultaneous relative to different lines.
X Humana.Mente Issue 13 April 2010

(10) If STR is true there is no spatially extended present.
32

(11) If Presentism is true then there is a spatially extended present.
(12) Hence, Presentism is false (by (6), (10), (11)).
A careful and detailed analysis of such arguments is beyond the scope of this
introduction. But it is important to note something about them. They seem
33
to
assume implicitly the following premise, where event should be taken to be,
loosely speaking, as the content of a spacetime point.
(13) The Present of an event e
1
is the region of spacetime that
contains all and only those events that are absolutely
simultaneous with e
1
.
Then, given (13) it is possible to derive the geometry of such a region from
the geometry of Minkowski spacetime and go on to argue as in (7)-(9) and
(10)-(12). But (13) is not supported by STR itself. It is rather a metaphysical
claim that could be resisted on metaphysical grounds. Thus anyone who is
willing to question (13) should be able to resist both the arguments I have
presented. If I am right this is a very nice example of how much subtler and
deeper the interrelations between physical and metaphysical considerations
are. I should note here that I do believe that a new, more compelling and more
sophisticated relativistic argument can be given against Presentism. Here I can
only give a brief sketch of it.
34

(14) If Presentism is true, on pain of contradiction, every event should
belong to just one privileged subregion of Minkowski spacetime
that is suitable to represent geometrically the Present.
(15) If so the relation of belonging to such a subregion is an
equivalence relation (by (14)).
(16) There are no equivalence relations that are definable in terms of
the geometric structure of Minkowski spacetime
35
beside the
identity relation and the universal relation (by the geometric
structure of Minkowski spacetime).

32
This claim is allegedly based upon the causal structure of Minkowski Spacetime.
33
Though I know that this might be, again, a controversial claim.
34
For a detailed presentation see Calosi (unpublished).
35
This is due to facts about signature and facts about causal isomorphisms of Minkowski
spacetime, i.e., maps of the form : A A where A is the underlying affine space that preserves the
causal structure. Formally if pKq stands for p is causally connectible with q then invariance under
causal isomorphism can be written as pKq (p)K(q).
Introduction Physics and Metaphysics XI
(17) If the relation in question is taken to be the identity relation then
Presentism implies that there exists only one spacetime point.
(by (1) or (4) and (16)).
(18) It does not exist only one spacetime point.
(19) Thus the relation in question is the Universal Relation (by (16),
(17) and (18)).
(20) If the relation in question is the Universal relation it follows that
every spacetime point is in the present of every spacetime point.
And so every spacetime point is real. Thus Eternalism follows.
It is not possible to assess whether this argument is successful here.

2.2 SPECIAL THEORY OF RELATIVITY AND METAPHYSICS OF PERSISTENCE
While observations drawn from STR have played a considerable role in
Metaphysics of Time at least since the seminal Gdel (1949) and Putnam
(1967), they have been almost absent from another metaphysical debate,
namely the one concerning Metaphysics of Persistence, until fairly recently.
36

Things persist through time. This much seems uncontroversial. The
controversy is how they do so. Famously there are two main Metaphysics of
Persistence
37
, namely three and four-dimensionalism. Let me start by giving a
rough definition of a three and a four-dimensional object.
(21) x is a 3D object =
df
x is a persisting object that persist through
time by being wholly present at each time of its existence, thus
not having any temporal parts.
(22) X is a 4D object =
df
x is a persisting object that persists through
time by having a different temporal part at each time of its
existence.
Then three and four-dimensionalism can be stated as
(23) 3D = every material object is a 3D object.
(24) 4D = every material object is a 4D object.

36
Yuri Balashov is the one who first suggested a detailed relativistic argument in favor of a
particular metaphysics of persistence, namely four-dimensionalism. This argument is improved in his
Balashov (forthcoming). See the beautiful Del Savios review in this volume.
37
I am leaving the possibility of stage theory or exdurantism aside. From a strict ontological
point of view this theory can be seen as a variant of four-dimensionalism. From the metaphysical point
of view there is however, I believe, room for disagreement.
XII Humana.Mente Issue 13 April 2010

It is clear that this formulation of the debate over the Metaphysics of
Persistence is centered about the existence of temporal parts. Call this way of
formulating the problem Mereological Persistence. Even in this case STR has
played a considerable role even in the formulation of the problem. As it turns
out the task of giving a precise formulation of the central notions in the debate
over Mereological Persistence, namely the notion of temporal part and the
notion of being wholly present, is very far from trivial. This difficulty has raised
a considerable amount of skepticism over those notions and thus over the
entire debate. STR itself proves a substantive argument in favor of a
reformulation of the debate. The central notions are in fact again cast in purely
temporal terms and thus they fall short of the Supervenience argument of the
previous section. But STR itself, and more generally, spatiotemporal theories
have proved to be fruitful instruments to recast the debate in different terms.
An alternative formulation of the debate centers around the notion of location,
in particular the notion of exact location.
38
The driving intuition behind the
notion of the exact location is that an object and its exact location share all the
relevant geometrical properties. Thus the exact location of my hand will be a
hand shaped region, the exact location of a square with the side of one inch will
be a square spacetime region whose sides do measure one inch and so on. Let
me introduce some terminology. I will write
39

(25) ExL (x,R) for x is exactly located at spacetime region R =
df
x and
R do have the same geometrical properties.
(26) OvF (x,R) = for x overfills R=
df
ExL (x, R
1
) . (R <<
40
R
1
) = there
is no part of R which is free of x.
(27) Path (x) =
df
i
R
i
ExL (x, R
i
) = Path (x) is the union of all and only
those regions
41
x is exactly located at.
42

Then the distinction between a 3D and a 4D object can be make precise in
locational terms, along those lines. A three-dimensional object is an object that

38
I cannot enter into the details of different formal theories of location. I will therefore stick to
the theory that is almost invariably used, as it is found in Gilmore 2007 or Balashov (forthcoming). I
have personally some reservations about those theories of location.
39
Notation and definitions are taken from Calosi (unpublished).
40
<< stands for the mereological notion of proper parthood. The mereological system
presupposed in what follows is Minimal mereology. For those and other mereological details see Varzi
(2009).
41
If there is just one such a region than it follows that ExL(x, Path (x)).
42
Index i ranges over spacetime regions.
Introduction Physics and Metaphysics XIII
is multiply exactly located at different non overlapping temporally
unextended
43
spacetime regions while a four-object is an object that is singly
exactly located at a temporally extended spacetime regions. These claims
actually capture the powerful intuition according to which three and four-
dimensional objects do have different geometrical properties. The last piece of
notation is grounded in the geometric structure of Minkowski spacetime. I will
write
(28) Achr (R) for R is an achronal region of Minkowski spacetime
=
df
(p) (q) (p e R . q e R) pq is spacelike.
44

Then a persisting object x is defined via
(29) Pers (x) for x is a persisting object =
df
~ Achr (Path (x))
and three and four-dimensional objects can be given the following precise
definitions:
df
Pers (x) . (-R
1
) (-R
2
) (R
1
R
2
. ExL (x,R
1
)
. ExL (x,R
2
) . Achr (R
1
) . Achr (R
2
) . ~ Achrn (R
1
R
2
)).
45

df
Pers (x) . ExL(x, Path (x)) . (ExL (x,R) R
= Path (x)).
Informally definitions (31) and (32) say that a 3D object is an object that is
multiply exactly located at different achronal subregions, while a 4D object is
an object that is exactly located at a single non achronal subregion, namely its
path.
Then three-dimensionalism and four-dimensionalism can be formulated
again via (23) and (24). Call the present formulation of the debate Locational
Persistence.
46
This long formulation of Locational Persistence is again a fine
example of a case in which physical considerations have proved fruitful in
reformulating a typical metaphysical problem.
Within the background of locational persistence different relativistic
arguments against three-dimensionalism have been put forward. In what

43
The relativistic counterpart of this notion is achronality. See later on.
44
Where pq is spacelike iff <pq, pq> < 0.
45
Note that Pers (x) is redundant here. This definition of a 3D object can indeed be improved
upon, but I cannot refine it here.
46
I cannot enter here into the subtleties about the relationships between Mereological
Persistence and Locational Perisistence.
XIV Humana.Mente Issue 13 April 2010

follows I briefly rehearse two of them, the Explanatory Argument due to Yuri
Balashov
47
and the Location Argument due to Cody Gilmore
48
, before
addressing one of my own relativistic arguments the Relativistic Argument
from Change.
The Explanatory Argument is a typical inference to the best explanation
arguments. It follows from STR and thesis about Locational persistence that
the same object will have different 3D shapes relative to different frames.
Three-dimensionalism will not have any explanation of why different and loose
3D shapes form a remarkable unity and come together to form a smooth four-
dimensional volume. Four-dimensionalism on the other hand has a ready and
simple explanation. This is due to the fact that different 3D shapes are just
cross section of a four-dimensional object.
49

The Location Argument maintains that three-dimensionalism cannot
answer the so called Location question. Heres the location question. Let x be a
material object. What subregions of Path (x) does x exactly occupy? Four-
dimensionalism has a ready answer. Since x is a 4D object it will exactly occupy
the only proper subregion of Path (x), namely Path (x) itself. But if x is a 3D
object it exactly occupies just achronal slices of Path (x). But which ones?
Gilmore suggests different answers to this problem and discard them all using
arguments based on relativistic consideration. And this again suggests that
four-dimensionalism is somehow favored by STR.
To conclude the section I will briefly sketch one of my own relativistic
arguments. I can only give a rough presentation of it. I have labeled it
elsewhere the Relativistic Argument from Change. Think for a moment to the
classical case. Suppose x is a 3D object that changes from having the property
F at time t
1
to having the property ~ F at time t
2
. On pain of contradiction x
cannot have two incompatible properties. Traditional three-dimensionalist
solution to this problem maintains that properties should be somehow
relativized to times. Let me write F-at-t
1
(x) for x has the property F at t
1
. Then
F-at-t
1
and F-at-t
2
are not incompatible properties and the problem from
change vanishes. Given Locational Persistence this strategy will involve that 3D

47
See his Balashov (forthcoming) reviewed in this volume. This work contains at least another
influential relativistic argument in favor of four-dimensionalism, namely the so called coexistence
argument.
48
See his Gilmore 2007 for a detailed presentation.
49
This argument has been challenged many times. The interested reader should read Balashovs
own discussion in Balashov (forthcoming).
Introduction Physics and Metaphysics XV
objects have properties relativized to spacetime regions they occupy. I will
write
(32) F-at-R
1
(x) for x has property F at spacetime region R
1
.
Now suppose x is a 3D object that is exactly located at two different
overlapping achronal subregions of Minkowski spacetime R
1
and R
2
. The fact
that x can be exactly multi-located at R
1
and R
2
follows directly from the
definition of a 3D object. The fact that those regions can overlap comes from
the fact that the best answer to the Location Question presented above a three-
dimensionalist has is that x exactly occupies every achronal slice of Path (x).
50

Then possibility of change implies that x can have incompatible properties at
R
1
and R
2
. Consider now the following property, being uniformly F, defined via
(33) UnF-at-R (x) (R
1
) (OvF(x,R
1
) F-at-R
1
(x)).
Claim (33) just says that if x is Uniformly F at one of its exact locations it is F at
every subregion of that exact location. It follows from definition of Overfill that
every region x does overfill is a subregion of its exact location. Now everything
is in place for the new relativistic argument. I am presenting it in a shortened
version.
(34) UnF-at-R
1
(x) (assumption).
(35) Un~ F-at-R
2
(x) (from possibility of change).
(36) (-R) (R<<R
1
. R <<R
2
) (by definition of Overlap
51
).
(37) UnF-at-R
1
(x) (R) (OvF(x,R) F-at-R(x)) (definition of
UnF and (34)).
(38) Un~F-at-R
2
(x) (R) (OvF(x,R) ~F-at-R(x)) (definition of
Un~F and (35)).
(39) F-at-R(x) (by (34), (36) and modus ponens).
(40) ~F-at-R (x) (by (35), (36) and modus ponens).
Thus the conjunction of three-dimensionalism and possibility of change entails
a contradiction. Note that this argument crucially depends on the geometric

50
This claim calls for substantive argument. I think that Gilmore (2007) gives reasons enough to
believe that this is the only viable answer to the location Question for a three-dimensionalist. I have
also a different more general argument for this claim. Given Locational Persistence the every slice
answer to the location question is the only answer that can account for basic relativistic phenomena
such as length contraction.
51
Again, see Varzi 2009.
XVI Humana.Mente Issue 13 April 2010

structure of Minkowski spacetime. In a classical spacetime a 3D object will in
fact never be exactly multiply located at different but overlapping regions.
There are ways to resist the argument but I contend that they all fail. It is not
my intention to defend conclusively this argument of mine here. I just wanted
to give it as an example of interaction between physical and metaphysical
considerations.

3. PHYSICS AND METAPHYSICS. STRUCTURE
The present volume of Humana.Mente embodies perfectly the spirit of the first
two sections. It tries to bridge the gap between Physics and Metaphysics, both
in an historical
52
and in a more theoretical perspective. It offers examples of the
interaction between Physics and Metaphysics at their very best. The volume
contains discussion of various physical theories ranging from Relativity Theory
(Norton, Barons, Evans and Miller, Lam, Macchia), to Quantum Mechanics
(Garola and Sozzo, Angelucci and Fano), from Thermodynamics (Weinert), to
Quantum Field Theory (Lupher), from Electromagnetism (Angelucci and
Fano) to Quantum Gravity (Barons, Evans, Miller) to String Theory
(Veneziano). Considerations drawn from these physical theories are used to
clarify and solve traditional metaphysical issues. Among those who are
explicitly addresses in the volume we find the question of realism (Garola and
Sozzo), metaphysics of time (Norton, Barons, Evans and Miller, Weinert),
causation (Lam), questions about cosmology (Macchia) and fundamental
ontological questions regarding fields, particles and the spacetime structure
(Lupher and Macchia respectively). Commentaries provide new takes on
classical texts on Physics and Metaphysics. They provide new challenging
arguments on classical questions such as the relations between physical
theories and phenomenology of experience (Barrett), the relations between
composition, strict metaphysical language and loose common sense language
(Nolan), the relations between physics, mathematics and experience
(Weatherall) and paradoxes of time travel (Torrengo). The volume also
contains review of recent works that I am confident will provide a fundamental
contribution to the discussion in the field for many years to come. These works
include authors such as Maudlin, Rovelli, French and Balashov.

52
See mainly Angelucci and Fano (this volume).
Introduction Physics and Metaphysics XVII
Finally this issue has the privilege to have an interview with A. Grnbaum,
that many will probably regard as one of the highest vertex when it comes to the
philosophical enquiry on Physics and Metaphysics. And with this I leave the
reader to the volume.

Acknowledgments
I would like to thank all of the people that made this volume possible, all of the authors that
have written, all of those who have submitted papers, all of the referees that have done an
impressive work. I would also like to thank people that have supported me and this work in
many different ways, among them Paolo Parrini, Achille Varzi, Paolo Valore and Cody
Gilmore. Naturally this work could not have possibly seen the light without Duccio and
Silvano. To them not only a thanks but a glass of wine.

REFERENCES
Aristotle (1991). The Metaphysics. (tr. by J. H. McMahon). Amherst:
Prometeus Book.
Balashov, Y. (2010). Persistence and Spacetime. Oxford: Oxford University
Press.
Calosi, C. (unpublished). Time and Objects in Minkowski Spacetime. Ph.D.
Thesis. Florence: University of Florence.
Gilmore, C. (2007). Where in the Relativistic World are We?. Philosophical
Perspectives, 20(1), 199-236.
Lange. M. (2002). An Introduction to the Philosophy of Physics. Malden:
Blackwell Publishing.
Loux, M. (1998). Metaphysics: A Contemporary Introduction. New York:
Routledge.
Strawson, P. F. (1959). Individuals. London: Methuen and Company.
Van Inwagen, P. (2007). Metaphysics. Stanford Encyclopedia of Philosophy.
<http://plato.stanford.edu/entries/metaphysics/>
Varzi, A. (2009). Mereology. In Stanford Encyclopedia of Philosophy.
<http://plato.stanford.edu/entries/mereology/>
XVIII Humana.Mente Issue 13 April 2010

Report
Third Interdisciplinary Conference
InterOntology10
*

Keio University, Tokyo, 27-28 February 2010

Michele Pasin**
michele.pasin@gmail.com

The third Interdisciplinary Ontology Conference was held in Tokyo, Japan,
from February 27 to February 28 2010. Organized by the Japanese Center for
Ontological Research (JCOR) and cosponsored by the Japanese Governments
Ministry of Education and Science (MEXT), the stated goal of this forum is to
support the exchange of ideas and state-of-the-art technologies for those
working in the ontology domain from around the world. The event has a quite
unique f lavor, for it gathers researchers from disciplines as disparate as
computer science, logic and philosophy, as well as a variety of application
domains. The common thread is the discipline of ontology, which has
undoubtedly gone a long way since its early days in ancient Greece.
We all know that ontology began as a branch of philosophy, studying the
types of entities in reality and the relations between them. In the seventies, the
early researchers in artificial intelligence borrowed the word from philosophy
and applied it to their discipline. Consequently, if ontology used to be intended
as a systematic account of Existence, within this new context, what exists has
become that which can be represented using a computer. Disciplines such as
ontology engineering were soon to be born, which investigated (among various
other more technical aspects) how to best employ the rich body of theory from
philosophical ontology to the purpose of making conceptual distinctions in a
systematic and coherent manner. Nowadays ontology has become an
established branch of computer science, which offers solutions to problem in
areas as disparate as data integration, information retrieval, natural language
processing, industrial planning and many others.

* I would like to thank the Japan Society for the Promotion of Science (www.jsps.go.jp) for giving
me the means to attend this conference.
** Kings College London
XX Humana.Mente Issue 13 April 2010

As already mentioned, it is not uncommon for this conferences attendees
to be almost unable to follow a talk, for it uses the word ontology in a way
never heard before. This is, on the contrary, one of the most interesting aspects
of the strongly interdisciplinary meeting. In the review that follows we hope to
give to the reader a small taste of this feeling, and a better appreciation of the
many senses we can talk about ontology in 2010.
Achille Varzis paper titled On the Boundary Between Material and Formal
Ontology sets the scene for the whole conference. Material ontology, the one
made popular by Quine, is concerned with the question of what there is, while
formal ontology, often related to the work of Brentano and Husserl, focuses not
on what there is but on the formal structure of what there is. So, for example,
the former may include debates on whether abstract particulars exist or not, or
arguments in favor of the existence of holes. Instead under the heading of the
latter we might find issues such as whether an entity is self identical or not, or
attempts to prove that no entity can consist of a single proper part. In general,
the second type of ontology is concerned with the general features of what
exists, independently of what that is. Thus, formal relations such as identity,
parthood or dependence should be rightly investigated by formal ontology, and
kept separated by the more specific problems of material ontology. This is the
traditional view on the subject but, Varzi challenges us, is the boundary so
clear? With a well-developed argument, the author argues that one cannot
pursue one sort of theory without also engaging in the other. In other words,
the tasks of material ontology presuppose the backing of some formal-
ontological theory, and that not always formal ontology can be in the
material sense of the word, ontologically neutral.
Antony Galton with the paper How is a Collection Related to its Members?
gives us a first tasting of the depths of formal ontology by discussing the often
neglected relationship of membership between an object and a collection too
which it belongs. If a choir may be regarded as a collection of singers not just
an arbitrary collection of singers, but a group of singer that have formed an
agreement to come together for the purpose of performing certain types of
music how shall we characterize the relationship between the choir and the
singers? The obvious first answer can be that the singers are the choir; however
a more rigorous ontological analysis would impose us to define what we mean
by are in the previous statement. If are stands for an identity relation, then
we would immediately have to face the problem of a singular thing being
identical to a plural one. Galton takes us through two possible solutions to this
Report Third Interdisciplinary Conference. InterOntology 10 XXI

problem, which are considering a collection equal to the mathematical set of its
members, or considering it equal to the mereological sum of its members, but
in both cases he concludes that we have not gained any new ground. We must
then admit that the collection is sui generis, and cannot be identified with
anything that we can specify independently. Consequently, a relation other
than identity seems to be needed here. According to Galton, we can approach
this issue by looking at an analogous ontological problem about which the
relevant literature is much more abundant. In fact, the relationship between a
collection and its members is analogous to the relationship between a material
object and the matter it is made of, and the relation between an assembly and its
components. Therefore a thorough consideration of the principal
philosophical positions related to the latter problems (such as eliminativism,
constitution and four-dimensionalism) will help us elucidate matters with
respect to our initial issue. Furthermore, Galton points out a new possible
alternative, which is based on distinguishing synchronic and diachronic forms
of identity. This approach, he says, has often been discarded as impracticable
or bizarre, but it is worthwhile exploring especially in relation to the formal
characterization of collections.
The article of John Bateman, Ontological Modularity: Unity in Diversity,
gives us a brief tour of the type of problems computational ontologists have to
tackle. Bateman reports on the work of the Collaborative Research Center for
Spatial Cognition, an interdisciplinary team sponsored by Bremen and
Freiburg universities that investigates the acquisition, organization,
utilization and revision of knowledge about spatial environments, be it real or
abstract, human or machine. The range of computational systems that benefit
from this type of research is quite vast: for example, we can think of softwares
for human-robot interaction, ambient assisted living or architecture and
building specifications. One of the great issues in this area, says Bateman,
comes from the need to provide explicit models that do justice to a rather
diverse set of requirements. Such requirements, in general, can be grouped
under three major headings. First, we have formal ontology, which has long
been attempting to create a coherent formal characterization of the spatial
properties of objects. Second, there is linguistic, which accounts for all the
usage of spatial expressions in natural language. Third, we have qualitative
spatial representation and reasoning, a research area in artificial intelligence
that (broadly speaking) poses the accent on the creation of decidable spatial
models of reality, that is, models that can be used productively by computers. In
XXII Humana.Mente Issue 13 April 2010

his article, Bateman shows that too often all of these three approaches have
been working in isolation, and that if confronted with each other they tend to
choose reductionist solution, so that one might arrange the three approaches
hierarchically and make one set of distinctions more basic than other. In
contrast with this reductionist view Bateman advocates a radical
multiperspectivalism, a view that does not rule out the possible incompatible
nature of different representations of space, but attempts to find the synergies
among them. This can be achieved only if we use formal languages committed
to notions of strong structuring, modularity and heterogeneity. In
particular, he discusses the use of the Common Algebraic Specification
Language (CASL), which embeds the principles above and therefore supports
the creation of non-reductionist spatial ontologies.
Werner Ceusters and Barry Smith Malaria Diagnosis and the Plasmodium
Life Cycle: The BFO Perspective is a good example of applied ontology in the
biomedical domain. In their paper, the authors address the problem of
producing a formal representation of the concepts of diagnosis, disease,
symptom, disorder, pathological process and other biomedical notions. In
particular, they look at the specific case of diagnosing malaria: this disease can
be suspected on the basis of both symptoms reported by the patient and
physical findings detected at examination; however, for a definitive diagnosis to
be made, laboratory tests must demonstrate the presence within the patient of
malaria parasites. To make the situation more difficult, some people are
infected but not made ill by the parasites, thus requiring a further
differentiation between the concept of malarial illness and that one of malarial
infection. Ceusters and Smith show how their Basic Formal Ontology (BFO)
can be used to put logical order among these concepts, so to create a coherent
formal model which can support software applications in performing a number
of knowledge-intensive tasks. Another example of biomedical ontology is given
by Christopher Baker and colleagues, with a paper titled Lipid Ontologies. In
this case the ontology addresses the problem that lipid research lacks a
consistent nomenclature for lipids and that different lipid research groups
have developed customized classifications of lipids that are relevant only for a
restricted category of lipids. The authors therefore present their contribution
as a rigorous formal ontology aiming at covering the subject area in a
systematic and explicit way, to the aim of facilitating the process of data
integration between different users.
Report Third Interdisciplinary Conference. InterOntology 10 XXIII

A more computer science perspective on ontologies has been given by
Robert Meersman in his Hybrid Ontologies in a Tri-Sortal Internet of Humans,
Systems and Enterprises. According to the author computational ontologies
must be framed within an epochal change that includes the most recent
technological advances in our society, such as the pervasiveness of computing
devices, the participatory character of the web (the so-called Web2.0) and the
total transformation of traditional business processes by means of the internet.
Within such a scenario ontologies will achieve their intended purpose of
facilitating the integration of information only if they get closer to the world
and daily practices of human beings. In other words, they must become hybrid
ontologies, where concepts on the one hand are circumscribed linguistically
and (mostly) declaratively by agreement within (human) communities, and on
the other hand identified formally (and unambiguously) for use in computer-
based information systems.
While all the articles presented so far have been given by invited speakers, a
large portion of the conference consisted of contributed papers and research
reports. We are going to name just a few of them, as they are indicative of the
broad horizons the Interontology conference usually has. Existence and
Vagueness by Elisa Paganini attacks Siders claim that the word exist is non-
vague, even if intended as equivalent to an unrestricted existential quantifier.
Claudio Calosis Three-Dimensionalism and Formal Theories of Location
composes and argument in favor of a four dimensional spatiotemporal
ontology. In DNA Sequences from Below: A Nominalist Approach, Yu Lin
throws the basis for a formalization of molecular biology that does not require
the existence of abstract objects. In Building up a Large Ontology from
Wikipedia Japan with Infobox and Category Tree Takahira Yamaguchi and
Takeshi Morita discussed how they used computational methods to
automatically extract an ontology from the community-constructed Japanese
Wikipedia website. Finally, Makoto Sakai and Hiromichi Fukui have presented
Ontology Study for Analysis and Anatomy of English-language News Relating
to Human Security, discussing their use of ontologies to support
disambiguation in a system that integrates online news articles.

XXIV Humana.Mente Issue 13 April 2010

Report
A School on Open Problems in Philosophy of
Sciences
Cesena, 15-17 April 2010

Pierluigi Graziani*
pierluigi.graziani@uniurb.it

Philosophy, in its most serious and profound dimension, is about the
construction of a justified solution to a particular class of fundamental
problems; the philosopher is he who can argue for his answers to the problems,
who knows how to put them in their place historically, who can rationally justify
them and make them available for discussion and modification.
However, there are two points of view which concerns the relationship
between philosophy and sciences and which does not endorse this prospect
1
.
The first view rises from the fear of some philosophers that interaction
between philosophy and sciences could result in an expansion of the latter and
a reduction of the former; such a fear favors philosophical efforts for the
construction of rigid boundaries between sciences and philosophy.
The second view, although allows an opening of philosophy with respect to
sciences and takes scientific issues as starting points for philosophical
reflections, is however incapable of understanding the depth of those scientific
issues.
Therefore, the first view transforms philosophical analysis in a mere a priori
activity that is a sort of scholasticism; the second one, on the contrary,
considers sciences, but only in an extrinsic way.
In such a context, a research school which focuses on open problems in the
philosophy of sciences should be an attempt to support a more serious and
profound idea of philosophy, one which, first, addresses problems from a
historical point of view, by proposing solutions which are aware of the
development of the debates and argued in a style as cogent as possible and,

* Urbino University
1
For a wider reflection, see Boniolo, S. (2002), Quattro questioni per ridiscutere sulla filosofia.
CxC Calls for Comments Sito Web Italiano per la Filosofia.
XXVI Humana.Mente Issue 13 April 2010

second, picks the problems from the sciences and analyzes them with
competence, care, skill, by grasping their essence and putting them in a
general perspective.
With these purposes, following (mutatis mutandis) the experience of two
schools in Francavilla al Mare organized by the Italian Philosophical Society
2
,
the school Open Problems in Philosophy of Sciences
3
(Cesena, 15-17 April
2010) has promoted and stimulated a methodologically conscious and mature
philosophical analysis, capable of dealing with precision and deep cultural
awareness with the problems raised by sciences.
The school, opened by such a distinguished scholar of philosophy of
sciences as Evandro Agazzi, has allowed and stimulated a dialogue between
new generations of scholars (students close to completing their Doctors
degree, or Philosophy Doctors who have not yet achieved a permanent
academic position) and distinguished professors such as Giovanni Boniolo,
Mario Piazza, Alfredo Paternoster and Vincenzo Fano.
The school has fueled a dialogue open to scholars in the field, teachers of
every grade and the entire civil society. Rather than talking about philosophy of
science, scholars have shown how this discipline can and should be done by
reflecting on important issues such as the Philosophy of the Life Sciences
(Cecilia Nardini; Fridolin Gross; Fabio Lelli; Elena Casetta); the Philosophy of
Mathematics (Gabriele Pulcini; Gianluca Ustori; Andrea Sereni; Valerio
Giardino); the Philosophy of Mind (Maria-Erica Cosentino; Barbara Giolito;
Maria Grazia Rossi; Maria Francesca Palermo); the Philosophy of Physics
(Giacomo Mancin; Claudio Mazzola; Giulia Giannini; Giuliano Torrengo).
The school, organized by Gino Tarozzi, Vincenzo Fano, Mario Alai and
Pierluigi Graziani was made possible by a synergy between the Interuniversity
Centre for Research in Philosophy and Foundations of Physics, the
Department of Philosophy of University of Urbino, the Italian Society of Logic
and Philosophy of Sciences and the Municipality of Cesena, which has been
favoring for many years, mainly thanks to Franco Pollini, the Interuniversity
Center in organizing major international events in the philosophy of sciences.

2
See: Tatasciore, C., Graziani, P., & Grimaldi, G. (Eds.) (2007), Prospettive Filosofiche 2006:
Il Realismo. Napoli: Istituto Italiano per gli Studi Filosofici; Tatasciore, C., Graziani, P., & Grimaldi,
G. (Eds.) (2010). Prospettive Filosofiche 2009: Ontologia. Napoli: Istituto Italiano per gli Studi
Filosofici.
3
See <http://synergia.jimdo.com/>.
Physics and Metaphysics: Interaction or Autonomy?

Mauro Dorato*
dorato@uniroma3.it

ABSTRACT
In this paper I will argue in favor of the view that if physics is to become a
coherent metaphysics of nature, it needs an interpretation, namely (i) a clear
formulation of its ontological/metaphysical claims and (ii) and a precise
understanding of how such claims are related to the world of our experience,
which is the most important reservoir of traditional, merely aprioristic
metaphysical speculations. Such speculations especially if conducted in full
autonomy from physics, or imposed upon it from the outside risk to turn
analytic metaphysics into a rigorous but fully sterile intellectual game.

1. AN ATTEMPT AT CLASSIFYING POSSIBLE ATTITUDES
In order to defend this claim, I will begin by quoting some interesting remarks
of Robert DiSalles Understanding Spacetime, where we find an historically
grounded discussion about possible ways of characterizing the relationship
between physics and metaphysics (DiSalle 2006, pp. 57-60). Some of his
remarks will be very helpful not only as a starting point to survey possible ways
to characterize such a relationship in contemporary philosophy of physics, but
also to understand, to a certain extent, the historical development of this
relationship, and some future directions that it might take.
A first position that DiSalle does not discuss and that is, however, too
important and widespread to be neglected claims that physics and metaphysics
are to be regarded as completely independent of each other, so that they cannot
conflict even in principle. This position amounts to a reciprocal attitude that
when it does not amount to disrespect could be labelled Tolerance.
Tolerance means that the subject matter of the disciplines is to be regarded
by their practitioners as completely non-overlapping, so that neither of the two

* Department of Philosophy University of Rome 3
2 Humana.Mente Issue 13 April 2010
disciplines can claim to offer a more reliable description of the fundamental
structure of reality. This stance is widespread among a significant number of
contemporary philosophers or metaphysicians, who pursue their work without
paying any attention whatsoever to what happens in the field of science, or
physics in particular. Perhaps more significantly, this attitude is also advocated
by the vast majority of physicists, who unconscious disciples of early
neopositivism are convinced that physics has nothing to do with metaphysics.
Physics is, and ought to remain, immune from metaphysics: possibly, the only
metaphysical hypothesis that is needed by physics is, as Dirac used to say at the
beginning of his lectures, that there exists an external world.
A second option consists in the attempt to subordinate one discipline to the
other. Subordination, in its turn, here might mean two different things, as
argued by DiSalle (2006, p. 57). On the hypothesis that physics and
metaphysics both try to offer a description of reality and could give conflicting
claims about it, one could either claim that either physics or metaphysics is
closer to the truth than the other either on single questions or in general or
try to derive the principles of one discipline from those of the other. Let us
distinguish between these two attitudes, by naming the former Denial and the
latter Explanatory Imperialism. Denial and Explanatory Imperialism have been
exercised much more on the part of philosophers over physics, than in the
opposite direction. This asymmetry might depend on the sociological fact that
in the last two centuries science has acquired much more prestige with respect
to traditional metaphysics. Consequently, while physicists often implicitly
believe that they can afford to ignore what happens in the field of metaphysics
since they are tracking the truth, the converse does not hold. Metaphysicians
feel that they have to justify they own approach to things even if many of their
beloved concepts (space, time, matter, number, motion, etc.) have become the
subject matter of science.
The historical figures of Descartes and to a lesser extent that of Leibniz
represent the first clear example of Denial, namely the attempt to claim that
metaphysics is closer to truth than physics.
1
The reader will recall that Galileo
was accused by Descartes of lacking a systematic method in his approach to
natural philosophy, and of being too absorbed by scattered and isolated
empirical questions. On the other hand, Newtons physical hypotheses were
attacked by Leibniz on the basis of metaphysical and theological principles (the

1
See also DiSalle 2006, p. 57.
Mauro Dorato Physics and Metaphysics: Interaction or Autonomy? 3
Identity of the Indiscernibles, and the Principle of Sufficient Reason applied to
Gods choice). Notably, in contemporary metaphysics this attitude of Denial
seems lost: even those who tamper with Special Relativity for purely
metaphysical reasons
2
, do not deny the fact that the scientific theory they
consider has some claim to (approximate) truth; in the example at hand, they
simply supplement it with an unobservable reference frame which, as such,
does not contradict the theory, but only its spirit.
3

The third possibility, Explanatory Imperialism, also grants metaphysics
some sort of superiority over physics, and according to DiSalle is exemplified
by the great mathematician Euler and by the early Kant. By granting the
possibility that the principles of physics may be taken at face value, and need
not be rejected in the name of metaphysics, such a third view still assigns
metaphysics the task to understand why those physical principles hold (DiSalle
2006). In other words metaphysics has the task to explain the basic principle of
physics, by deriving them from deeper aprioristic, metaphysical truths. Also
this third attitude, in contemporary metaphysics seems completely lost. Both
Denial and Explanatory Imperialism seem two attempts at resisting a
fundamental change in ways of acquiring knowledge about reality: from
aprioristic analysis to empirical inquiry supplemented by mathematical models.
Interestingly, the mature Kant realised that after Newtons Philosophiae
Naturalis Principia Mathematica traditional metaphysics had no future.
Transcendentalism or Foundationalism is the attempt to using philosophy or
the theory of knowledge to inquire into the very possibility of physics (and
metaphysics), namely to dig out those apriori structures of the transcendental
subject that make physics (and mathematics) possible. In a nutshell, according
to Kants Critique of Pure Reason, Newtonian physics is a fact, and his theory
of knowledge explains why it holds necessarily and universally, against the
doubts raised by Humean scepticism with respect to the universal validity of
physical laws. Contemporary Neokantians give up the idea that our apriori
categories grant universality and necessity to physics, but still retain the
Reichenbachian view that some historically changing structure might be
constitutive of important physical theories (Friedman 2001).
According to DiSalle, this kind of Kantian Foundationalism is to be
interpreted at the same time as an inquiry into the right that physics has to

2
See Craig 2001.
3
See Dorato 2001.
address metaphysical questions (DiSalle 2006). This is why DiSalle, certainly
by making some violence to Kant, also attributes him what for me should be
regarded a distinct, fifth position, amounting to the claim that
physics is not a consequence of the metaphysics of nature. Quite simply it is the
metaphysics of nature. The metaphysical concepts that we find in physics
body, force, motion, space, time, become to us intelligible precisely, and only,
as they are constructed by physics itself; physics provide us with the only
intelligible notions we have on this matter. (DiSalle 2006, p. 60)
This fifth position, which we could call Physical Autonomy, is the claim that
physics, being itself a metaphysics theory of nature, does not need any
metaphysical analysis or intervention from outside. The reason why physics is a
metaphysical theory of nature is given by the fact that it aspires to describe
reality: as Einstein wrote to Schrdinger, the true difficulty lies in the fact that
physics is a kind of metaphysics. Physics describes reality. But we dont
know what reality is unless we describe it with physics (Allori et al. 2005, p.
13).
Let us now pass to comment those options that are nowadays still regarded
as alive, in particular Tolerance and Physical Autonomy.

2. TOLERANCE, DISTORTION AND THE COMPATIBILITY TEST WITH PHYSICS
2.1. Contemporary analytic metaphysic is replete with attempts at
gaining some precise explications of notions like substance, event,
persistence, object, sameness, identity, becoming, properties, disposition,
causation, etc. All these investigations proceed, in many cases at least, in ways
that are completely independent of what is taking place or has taken place
within physics. For this reason, such investigations seem to belong to the kind
of approach to metaphysics that I have named Tolerance.
Are these researches relevant for a deeper understanding of physics? The
answer seems to be in the negative, whenever these inquiries do not even feel
the need to confront their theories with physics. This autonomous
metaphysical theorizing is certainly interesting and worth-pursuing, but also
contains the danger of sterility and isolation. Of course, I would not be ready to
deny that, say, understanding the nature of causation, is important for its own
sake; what is less clear, however, and still open to philosophical debate, is
whether causation has any role in the object language of physics, rather than
being important in a purely pragmatic or explanatory sense.
4
Here is how
Norton put it in 2003: Mature sciences, I maintain, are adequate to account
for their realms without need of supplement by causal notions and principles
(Norton 2003, p. 2). The idea here is that in physics the notion of law replaces
the notion of cause
5
, even though the latter notion has still an important role in
helping us to recover old, approximate theories from newer ones (say,
Newtons cause of free fall, invoking a force, from the contemporary
geometrized theory of gravity).
In conclusion, I would like to put forward the following claim: if it is not
clear yet in what sense the notion of property or cause or object can
affect, and play a role in physics, a confrontation with physics looks dubious or
suspicious, and this is the main reason why many metaphysical inquiries within
the analytic tradition remain, and possibly ought to remain, safely within the
tradition of autonomy, with all the dangers that such an autonomy involves.

2.2. Another instance of tolerance, sometimes masked under the
invitation to coming to terms with real physics, is what I could call
distortion. This occurs when the confrontation with physics is only apparent,
since the kind of physics that is invoked is either distorted or highly simplified.
This sort of uses of a non-actual physics on the part of the analytic
metaphysicians is what Ladyman and Ross criticize in various parts of their
book, when they refer, for example, to the alleged physically-based
opposition between atomism and gunkism (the view that matter is infinitely
divisible):
it is preposterous that in spite of the developments in the scientific
understanding of matter that have occurred since [Descartes], contemporary
metaphysicians continue to suppose that the dichotomy between [partless]
atoms and gunk remains relevant, and that it can be addressed a priori.
(Ladyman and Ross 2007, p. 20)
This sort of appeal to an imagined physics really counts as an instance of
Tolerance, especially if the method of inquiry continues to be wholly a priori.

4
See the debate between Norton (2009) claiming that physics does not need causation and
Frisch (2009) trying to argue for the opposite claim.
5
For the relation between causation and law, see, among other texts, Dorato 2005 and Psillos
2002.
2.3. One way in which a confrontation between metaphysics and physics
really does occur, on the contrary, is when physical theories are invoked as a
sort of experimenta crucis to decide between two or more competing
metaphysical views: if one of the metaphysical theories is in conflict with
physics, it ought to be abandoned. An instance of this use of physics within the
community of analytic metaphysicians is given for instance by the dispute
between presentism and eternalism, namely between the view that only the
present exists (presentism), and the view that past, present and future events
exist on a par in a block universe (eternalism).
Interestingly, there are people nowadays who claim that this dispute is not
genuine (Dolev 2006, Dorato 2006a, Savitt 2006). But let us assume, for the
sake of the argument, that there is a genuine debate just in conceptual and
metaphysical terms, and let us assume that this holds also for another debate
with respect to which the special theory has been invoked as a decisive test,
namely the problem between endurantists and perdurantists. I quote from an
abstract of a recent article:
There are two main theories about the persistence of objects through time:
endurantism and perdurantism. Endurantists hold that objects are three-
dimensional, have only spatial parts, and wholly exist at each moment of their
existence. Perdurantists hold that objects are four-dimensional, have temporal
parts, and only partly exist at each moment of their existence. In this paper we
argue that endurantism is poorly suited to describe the persistence of objects in
a world governed by Special Relativity. (Hales and Johnson 2003)
Clearly, the meta-philosophical principle, in both of these cases, is that if a
metaphysical theory is not compatible with, or is not properly suited to adapt
itself to, a physical theory, we should abandon it. However, note that also in
these cases, the metaphysical debate is somewhat completely external to
physics: the solutions to the question whether the future is real or not, or to the
problem how entities persist, seem quite unrelated to what physicists nowadays
are really after. These metaphysical problems, in other words, are external to,
and independent of, the issues that are really debated within physics. This is
why also this type of relationship between physics and metaphysics, I take it,
fully belongs to the philosophical style that I referred to as Tolerance.

3. THE METAPHYSICS WITHIN PHYSICS AND THE NATURE OF INTERPRETATION
Taking stock from Einsteins quotation above (physics is a type of
metaphysics), or DiSalle (physics is the metaphysics of nature), I would like
to sketch a view of the relationship between metaphysics and physics that
regards the former as being strictly dependent on the latter. If the essential task
of the philosopher of physics is interpreting physical theories i.e., 1) coming
up with a precise and exact ontology to associate to the language and formulas
of physical theories and 2) relating such ontology to the world of our
experience it then seems that 1) necessarily involves a metaphysical task,
namely finding out how the world can be like if our physical theories are at least
approximately true. Note that this interpretation of the interpretation of
physics (which has variously defended by van Fraassen 1980, Giere 1988 and
Lange 2002) does not require truth from our physical theories, but can be
embarked upon also by instrumentalists, since the whole interpretative task
rests on a conditional statement (if the theories are at least approximately
true).
For instance, can the non-local correlations presupposed by entangled
states be interpreted as referring to some sort of causal correlations? And if the
answer is in the positive, which model of causation do they allow? When we ask
such questions, it is of the utmost importance to remark that we can never
exclude that one of the novelties of the whole worldview suggested by EPR-
Bohm correlations lies just in the fact that no causal explanation is really fit to
explain/interpret them, because these correlations are to be regarded as
fundamental, or natural in Aristotles sense, and as such they need no causal
explanation whatsoever (Fine 1989). In other words, without taking stance in
this complicated problem, here we simply want to suggest that sometimes the
old, metaphysical notions (causation, property, dispositions) may be unfit to
give an account of the new ontology suggested by a physical theory. In our
example, this sort of possibility makes all attempts of looking for a causal
interpretation of the correlations in question look outdated.
The same could be true for the dispute between substantivalism and
relationism, a metaphysical debate that might have been appropriate for the
times of Leibniz and Newton, but simply inappropriate after General Relativity,
which is a theory that overcomes the distinction between empty spacetime and
matter by identifying spacetime with a gravitational field (Rynasiewicz 1996,
Dorato 2008). In other words, the interpretative task must always be open to
the possibility that no current or traditional metaphysical category is really
appropriate for the new physical theory, so that, one more time, the
confrontation between metaphysics and physics is one between concepts and
categories that are imposed onto the latter from the outside.
As another example, consider the question: can the timelike-separation of
events in classical spacetime theories be interpreted as giving rise to a tenseless
form of local becoming? Philosophers who have recently defended this
minimalist interpretation (Savitt 2002, Dieks 2006, Dorato 2006b) are well
aware that the question remains whether such a metaphysical reading of
relativity is capable of connecting with, or explaining, the sense of passage of
time typical of our experience. If this second explanatory task is not fulfilled by
postulating a becoming of events defined as their timelike succession in the
partial order defined by special relativity, the first ontological interpretation
must be abandoned or at least corrected. This is why the ontological question
posed by 1) is relativity theory (special and general) interpretable as a theory
that metaphysically admits of becoming? can never be divorced from the
questions raised by 2), namely the connection with our experience.
In conclusion, I would like to stress that it is the connection of the
metaphysical interpretation of a physical theory with our experience that gives
us the final test for the plausibility of a proposed metaphysical interpretation.
This is particularly evident in some interpretations of quantum mechanics, in
particular in the no-collapse views linked to Everetts interpretation. In a
nutshell, Everetts interpretation solves the measurement problem by denying
the reality of the reduction process: the metaphysical interpretation here
consists in claiming that the fundamental equation governing the temporal
evolution of a quantum system is always linear and deterministic. Obviously, in
this case one must solve two correlated problems, both involving an agreement
with what we see, and therefore the relationship of the relevant metaphysical
posit with our experience. The first problem is why we never perceive
macroscopic superpositions. The second problem consists in trying to explain
the origin of the notion of probability, namely the impression that the
probabilities involved by the Born-rule play a fundamental role in quantum
theory. The first problem is attacked with the theory of decoherence, which
explains why we never perceive interferences (from within the same world)
of Schrdingers infamous dead cat with its alive counterpart, even though all
possible measurement outcomes do realize. This means that there is a world in
which the cat is dead and looks dead, and a world in which the cat is alive, and
looks alive, but in any of these two worlds macroscopic interferences are never
observed, due to decoherence processes. In other words, the final state is still a
macroscopic superposition of different worlds, even though this state cannot
be accessible to our experience. Whether this interpretation is satisfactory is of
course dependent also on how the second difficulty is tackled, a difficulty that
recently has involved attempts at explaining the notion of probability in a
physical theory with decision theoretic strategies of agents (Deutsch 1999,
Wallace 2007).
The appropriateness of this interpretation of quantum mechanics of course
cannot be judged in this context. Here it has been presented simply in order to
show the reader how complicated the interpretation of a physical theory really
is, and how far more promising a philosophical activity it is if compared with
external metaphysical theorizing. The question of interpretation is in fact
not external to physics at all, at least to the extent that in the past also physicists
have asked themselves whether atoms or the ether existed or not; in any case
the need to link the physical picture of reality with the world of our experience
remains one of the main tasks of philosophy of physics (Sellars 1963).
6

REFERENCES
Allori, V., Dorato, M., Laudisa, F., & Zangh, N. (2005). La Natura delle cose.
Unintroduzione ai fondamenti e alla filosofia della fisica. Roma:
Carocci.
Craig, W. L. (2001). Time and the Metaphysics of Relativity. Dordrecht:
Kluwer Academic Publishers.
Deutsch, D. (1999). Quantum Theory of Probability and Decisions.
Proceedings of the Royal Society of London, A455(1988), 3129-
3137. <http://arxiv.org/abs/quant-ph/9906015>
Dieks, D. (2006). Becoming, Relativity and Locality. In D. Dieks (Ed.), The
Ontology of Spacetime (pp. 157-176). Amsterdam: Elsevier.
Dieks, D. (Ed.) (2006). The Ontology of Spacetime. Amsterdam: Elsevier.

6
Of course, I admit that the criteria of what is internal and what is external to physics cannot
be left just to the judgments of physicists. Here this difficult question of the normativity of philosophy
can simply be mentioned.
DiSalle, R. (2006). Understanding Spacetime. The Philosophical
Development of Physics from Newton to Einstein. Cambridge:
Cambridge University Press.
Dolev, Y. (2006). How to Square a Non-Localized Present with Special
Relativity. In D. Dieks (Ed.), The Ontology of Spacetime (pp. 177-
190). Amsterdam: Elsevier.
Dorato, M. (2001). Review of William Lane Craig. Time and Metaphysics of
Relativity. Republished (2003) in Studies in History and Philosophy of
Modern Physics, 34(1), 154-158.
Dorato, M. (2005). The Software of the Universe. Aldershot, UK: Ashgate
Publishing.
Dorato, M. (2006a). The Irrelevance of the Presentism/Eternalism Debate for
the Ontology of Minkowski Spacetime. In D. Dieks (Ed.), The
Ontology of Spacetime (pp. 93-109). Amsterdam: Elsevier.
Dorato, M. (2006b). Absolute Becoming, Relational Becoming and the Arrow
of Time: Some Non-Conventional Remarks on the Relationship
Between Physics and Metaphysics. Studies in History and Philosophy of
Modern Physics, 37(3), 559-576. Reprinted in N. Oaklander (Ed.)
(2009), The Philosophy of Time, vol. IV (pp. 254-276). London:
Routledge.
Dorato, M. (2008). Is Structural Spacetime Realism Relationism in Disguise?
The Supererogatory Nature of the Substantivalism/Relationism
Debate. In D. Dieks (Ed.), The Ontology of Spacetime II (pp. 17-37).
Amsterdam: Elsevier.
Fine, A. (1989). Do Correlations Need to be Explained?. In J. Cushing & E.
McMullin (Eds.), Philosophical Consequences of Quantum Theory (pp.
175-194). Notre Dame, IN: Notre Dame University Press.
Friedman, M. (2001). Dynamics of Reason. Stanford, CA: CSLI Publications.
Frisch, M. (2009). Causality and Dispersion: A Reply to John Norton. The
British Journal for the Philosophy of Science, 60(3), 487-495.
Giere, R. (1988). Explaining Science. Chicago: University of Chicago Press.
Hales, S. D., & Johnson, T. (2003). Endurantism, Perdurantism, and Special
Relativity. The Philosophical Quarterly, 53(213), 524-539.
Ladyman, J., & Ross, D. (2007). Every Thing Must Go: Metaphysics
Naturalized. Oxford: Oxford University Press.
Lange, M. (2002). An Introduction to the Philosophy of Physics. Oxford:
Blackwell.
Norton, J. D. (2003). Causation as Folk Science. Philosophers' Imprint, 3(4).
<www.philosophersimprint.org/003004/>
Norton, J. D. (2009). Is There an Independent Principle of Causality in
Physics?. The British Journal for the Philosophy of Science, 60(3),
475-486.
Psillos, S. (2002). Causation and Explanation. Acumen & McGill-Queens
University Press.
Rynasiewicz, R. (1996). Absolute versus Relational Space-time: An Outmoded
Debate?. Journal of Philosophy, 43(1), 279306.
Savitt, S. (2002). On Absolute Becoming and the Myth of Passage. In C.
Callender (Ed.), Time, Reality & Experience (pp. 153-167).
Cambridge: Cambridge University Press.
Savitt, S. (2006). Presentism and Eternalism in Perspective. In D. Dieks (Ed.),
The Ontology of Spacetime (pp. 111-127). Amsterdam: Elsevier.
Sellars, W. (1963). Science, Perception and Reality. London: Routledge and
Kegan Paul. Republished (1991) by Ridgeview Publishing Company.
Van Fraassen, B. (1980). The Scientific Image. Oxford: Clarendon Press.
Wallace, D. (2007). Quantum Probability from Subjective Likelihood:
Improving on Deutsch's Proof of the Probability Rule. Studies in
History and Philosophy of Modern Physics, 38(2), 311-332.


String Theory: Physics or Metaphysics?

Gabriele Veneziano*
gabriele.veneziano@cern.ch

ABSTRACT
I will give arguments for why the enormous progress made during the last
century on understanding elementary particles and their fundamental
interactions suggests strings as the truly elementary constituents of Nature. I
will then address the issue of whether the string paradigm can in principle be
falsified or whether it should be considered as mere metaphysics.

1. THE CENTURY OF PHYSICS?
Very likely the 20th century will go down in history as the century of physics.
In my opinion no other field of human knowledge has undergone, in that
century, so much progress and so many revolutionary changes. Its very
beginning was marked by three developments that shook forever as many
scientific beliefs:
The belief in absolute determinism when, in 1900, Max Planck, in order
to eliminate an infinity in the energy emitted by a black body,
introduced a constant, h, that still carries his name. This marked the
beginning of the quantum revolution whose indeterminism was nicely
embodied later in Heisenbergs uncertainty principle,

, (1)
that bounds from below the product of position and momentum
uncertainties.
The belief in absolute time when, in 1905, Albert Einstein, starting
from the invariance of the speed of light in vacuum, c, introduced
Special Relativity and his much celebrated relation between mass and
energy:

* Collge de France, Paris; Theory Division CERN, Geneva
=
2
(2)
The belief in an absolute geometry when, only ten years later, again
Einstein formulated the theory of General Relativity according to which
matter curves spacetime and bodies simply move along geodesics
(i.e., minimal-length paths) in the ensuing non-trivial geometry. In
General Relativity Newtons constant, G, controls the amount by which
mass and energy affects the surrounding geometry of spacetime.
These three breakthroughs fed all subsequent developments of that branch
of 20th century physics that aims at uncovering the laws of physics at their
deepest level. In particular, the efforts made in the twenties and thirties to
combine Quantum Mechanics and Special Relativity led physicists to
formulate, in the forties, a very successful framework known as Quantum Field
Theory (QFT). The first successful application of QFT was Quantum
Electrodynamics (QED for short), a theory describing electromagnetic
phenomena at the quantum level with incredible accuracy (better than 10 parts
in a billion in the case of the anomalous magnetic moments of the electron and
the muon).
During the fifties and the sixties physicists tried to extend that framework
to the description of two of the other known forces, the strong (or nuclear)
force, responsible for binding protons and neutrons inside the atomic nuclei,
and the weak force, responsible for radioactivity. Such experimental and
theoretical effort was rewarded in the early seventies when physicists
formulated the so-called Standard Model of elementary particles, a milestone
that will certainly stay in the books of physics. Many tests of the Standard
Model, carried out in the seventies, eighties, and nineties, have so far
confirmed the validity of the Standard Model with no exception (the discovery
that neutrinos do have mass can be easily incorporated in the Standard Model
without any basic change on its structure).
Rather than describing what the Standard Model is, I will simply emphasize
its beauty in terms of the conceptual unification it brings about. The Standard
Model asserts that all non-gravitational interactions are described by one and
the same special class of QFTs, known as gauge theories. This was a known
fact for the QED description of electromagnetic interactions, but is a highly
non trivial and even surprising claim for the other two forces. The reason
why the same kind of theory, a gauge theory, can lead to such diverse
phenomena like Coulombs law, the short range force responsible for nuclear
Gabriele Veneziano String Theory: Physics or Metaphysics? 15
binding, and the slow process of radioactive decay, is due to the fact that the
underlying symmetry of a gauge theory (called gauge invariance) can be
realized in different ways, or phases, not unlike the way water can manifest
itself as a solid, a liquid, or a gas.
What is this single gauge principle underlying so many diverse
phenomena? The answer is quite simple: all these interactions are induced, at
the most fundamental level, by massless spin-1 particles.
1
Gauge theories are
the mathematical way to describe such kind of elementary particles. Thus, if we
wish to extract a single message from the incredible success of the Standard
Model, I would put it as follows:
Nature likes spin-1 massless particles
and therefore She likes Gauge Theories.
But then what about the fourth fundamental force known to us, gravity? For
several decades particle theorists were not very interested in gravity, since
gravitational forces are completely negligible for elementary particles in
normal situations (in the hydrogen atom, for instance, the ratio of the
gravitational and electromagnetic force between the electron and the proton is
a miserable 10
40
). Actually, physicists were quite happy to leave gravity to
Einsteins General Relativity. Also, since gravity is relevant only for
macroscopic bodies, they were happy to treat it classically, i.e., without any
appeal to quantum mechanics. However, when looked at more carefully, also
gravity reveals itself as a sort of gauge theory where the gauge symmetry is
replaced by Einsteins equivalence principle, an invariance under a generic
change of the coordinate system. Even more amazingly, one finds that the
symmetries of General Relativity are what they are precisely because the
quantum of gravity (called graviton in analogy with the electromagnetic
photon) is a massless particle of spin 2 (angular momentum 2 ). In other
words, we can add gravity to our previous reasoning by slightly enlarging the
message that Nature is sending to us:
Nature likes spin-1 and spin-2 massless particles
and therefore She likes Gauge Theories and General Relativity.

1
Spin-1 means an angular momentum =
2
. We recall that Quantum Mechanics implies
that angular momentum is an integer or half-integer multiple of .
What remains to be explained is why Nature likes precisely such kind of
massless particles.

2. CLASSICAL vs QUANTUM STRINGS
I will now argue that Relativistic Quantum String Theory (RQST) explains in a
very natural way the existence of those massless spinning particles that Nature
appears to like so much. What is RQST? It is simply what one obtains by
adding to the basic principles of Special Relativity and Quantum Mechanics
(whose combination, as I said, led to QFT) a third crucial ingredient: the
assumption that all truly elementary particles, rather than being pointlike, are
instead one-dimensional objects: strings.
By combining this assumption with special relativity and quantum
mechanics results in arguably the richest theory ever constructed by physicists,
RQST. The three ingredients: Relativity, Quantum Mechanics and Strings, of
RQST are all essential, but I will concentrate my discussion on the latter two,
Quantum Mechanics and Strings. I will argue that Quantum Mechanics is
essential to make string theory a candidate theory of elementary particles and
fundamental interactions by comparing the properties of classical,
deterministic strings with those of their quantum mechanical analogues.
A classical relativistic string is a well defined system containing a single
physical parameter, the so-called string tension T. The string tension plays, in
string theory, the same role that mass plays in point-particle theory. Mass can
be converted into energy via Einsteins eq. (2), whereas T denotes the energy
per unit length stored in the string. Its physical dimensions are thus
1
=

2
. A classical non-interacting pointlike particle moves along a
trajectory that minimizes its length (the already mentioned geodesic); similarly,
the classical string motion is such as to minimize the area of the two-
dimensional surface it sweeps during its motion.
Classically, neither point-particle nor string theory have a fundamental
length built in. At the quantum level, however, we can associate to the mass of a
particle a quantum length (so-called Compton wavelength) given by:

. Of course, such a wavelength is not a universal constant since it

varies from particle to particle according to its mass. Similarly, we can associate
with T a quantum length-scale, called the string length:

(3)
except that there is just a single T for all possible strings and therefore

unlike
, is a truly universal length scale. We can say that the three

fundamental constants c, h and
represent, respectively, the three basic

ingredients underlying RQST: Relativity, Quantum Mechanics and Strings.
Here comes the punch line: classically, string theory is scale-free. Given a
possible classical string motion we can always construct another one by
rescaling all lengths by a common factor. As a consequence, the mass of the
string gets rescaled by exactly the same factor. We can go to the limit in which
we rescale the size of the string to zero size and the string will become both
massless and spinless. The last property is physically very obvious: classically
we cannot have angular momentum without having both a finite mass and a
finite size (and indeed under a rescaling of the string size by a factor k its
angular momentum gets rescaled by a factor k
2
). Actually, one can prove a
strict inequality stating that
2
2
, which immediately implies that massless
spinning strings cannot exist classically!
Let us now turn on h, i.e., Quantum Mechanics. Its consequences are truly
amazing, even miraculous. They can be partly understood by the quantum
mechanics of known systems. Take for instance the hydrogen atom: the
classical theory had great problems in explaining its stability. The single
electron of the hydrogen atom would like to emit electromagnetic radiation
(like a charged particle circulating in an accelerator at CERN), would lose
energy and slow down, and eventually fall on the nucleus (a proton in this
case). But the uncertainty principle of Quantum Mechanics, Eq.(1), intervenes
and tells us that falling on the nucleus would make the relative position of the
electron and the proton very precisely determined at the cost of a lot of
momentum uncertainty, hence of a large kinetic energy. Quantum Mechanics
tells us which the best compromise is: the optimal (i.e., energy minimizing)
average distance between the electron and the proton is given by the so-called
Bohr radius:

0
=
~ 10
8
, (4)
where ~1/137 is the so-called fine-structure constant.
A similar mechanism is at work with a quantum string. In analogy with a
classical string it does not like to have a large size (since this costs a lot of
tension energy) but, unlike a classical string, it does not like to be very small
either since, in that case, its very small forces a very large i.e., a very
large kinetic energy (again the uncertainty principle at work!). Not
surprisingly, the best compromise happens to be for a quantum string to have a
size of order
, the only length scale available. Thus quantum strings acquire,

through Quantum Mechanics, a characteristic size, a minimal length. That
minimal length is essential for resolving the long-standing short-distance
problems with quantization of Einsteins General Relativity.
A second, no less important, miracle comes from the fact that the classical
inequality
2
2
suffers a quantum correction which is insignificant for
large J and M but essential when these are small. The inequality becomes:

2
2
+ (5)
where a can take integer or half integer values up to and including = 2.
Thus, as a quantum effect, massless strings of spin 1 and 2 become not only
possible but, actually, inevitable in RQST!
In conclusion, the combination of the two above-mentioned miracles
potentially makes RQST a realistic theory of all fundamental interactions
which, furthermore, is free from the UV infinities that plague ordinary QFTs
and that make quantization of General Relativity virtually impossible.

3. IS THIS PHYSICS?
Physicists have a definite criterion for deciding whether a certain theory can be
considered to be a scientific one: the theory has to make testable predictions so
that it can be falsified, at least in principle, by experiments. (On the contrary, a
theory can never be proven to be correct, since it is impossible to exclude that
an alternative explanation of the same phenomena can be found.) When this
criterion is applied to string theory it is converted into something slightly more
demanding. Indeed, everybody would agree that string theory makes definite
predictions, like for instance the existence of very heavy (by particle physics
standards) string excitations, or modifications of gravity at very short
distances. What is under dispute is whether any conceivable experiment, now
or in the foreseable future, will ever be able to test those predictions.
Several respectable physicists have taken a negative attitude towards that
question. According to them RQST is a beautiful construction whose
predictions will never be accessible to experimental verification: hence RQST
is not even wrong to quote a sentence by famous physicist Wolfgang Pauli.
The reason underlying this statement is simply that the fundamental length of
RQST,
, is, most likely, of the same order of magnitude (or perhaps just a
factor 10 larger than) the so-called Planck length, a length scale that can be
constructed out of the three fundamental constants we introduced at the very
beginning, c, , and G :

3
~1.6 10
33
. (6)
But then, it is argued, such a length scale is so tiny that there will never be a
way to distinguish a string of that size from a zero-size point. Hence, it will be
impossible to compare the predictions of a RQST from those of some suitable
QFT if we only have experiments of limited energy and thus, by the uncertainty
principle, of limited spatial resolution. Indeed, the possibility of building an
accelerator capable of testing distances such as those in eq. (6) is definitely out
of question.
That reasoning appears to be awed on (at least) two grounds:
There is in Nature a very powerful accelerator: the Universe itself.
Because of its expansion, the Universe has been cooling down since the
big bang. On the contrary, if we go back in time, the Universe was hotter
and hotter as we proceed towards the big bang. Thus, the physics of the
very early Universe, and even the very existence of a Big Bang as the
beginning of time, should have been strongly affected by the
characteristic properties of quantum strings and would much differ from
what would come out of more conventional theories like General
Relativity.
It is generally accepted today that the quantum properties of the early
Universe left an imprint on the large scale structures of the Universe
that we observe today: stars, galaxies, clusters. Therefore, it is all but
excluded that RQST can be tested through its cosmological
implications. Any possible modification of physics at the string scale has
been stretched to macroscopic (or even astronomical) distances by the
expansion of the Universe. At present, the problem with such a way of
testing string theory is that it is very hard to solve it in extreme regimes
like the one that must have prevailed around the big bang epoch.
Techniques are being developed to study such regimes, but are not yet
at the level of providing robust predictions.
There is an even stronger argument against the claim that RQST cannot
be falsified. It is enough to recall how the first version of string theory
was abandoned at the beginning of the seventies. That original string
theory, born in the late sixties and thus predating the construction of
the Standard Model, was not invented for the purposes for which it is
studied now: it was instead an attempt to describe the physics of the
strong interactions outside the framework of conventional QFT (at that
time QFT looked inapplicable to strong interactions). At the beginning
string theory led to great hopes, but then it proved to be too tight and
constrained a framework and, in particular, it kept predicting the
existence of massless particles of spin up to 2. When the purpose of
string theory was to describe the world of protons, neutrons and pions,
there was no room for such massless particles. This was certainly one of
the main reason for abandoning the old string theory in the early
seventies in favour of the theory of quarks and gluons, now known as
Quantum-ChromoDynamics (or QCD in analogy with QED).When the
purpose of string theory was not to describe the carriers of gauge and
gravitational interactions but rather the world of protons, neutrons and
pions, there was no place for such massless particles. This was certainly
one of the main reasons for abandoning the old string theory in the early
seventies in favour of the theory of quarks and gluons, now known as
Quantum-ChromoDynamics (or QCD in analogy with QED).
Could history repeat itself? Well, hopefully not, but nothing is less clear at the
moment. At a first, crude level of approximation RQST provides not only the
nice massless spinning particles we like and need so much: it also gives us, in a
single package, a bunch of massless spinless particles generically called
moduli. Some of them are related to the sizes and shapes of the extra
dimensions of space in which quantum strings like to evolve. This is, by the
way, another gift of quantum mechanics, we cannot just take what we like
and refuse the rest: string theory comes as a package deal: take it all or leave it
all!
One can show that these undesired massless strings produce new
unobserved long-range forces whose strength is similar to that of gravity but
which, unlike gravity, do not obey the equivalence principle of General
Relativity and thus lead, for instance, to unacceptable violations of the
universality of free fall, a property now tested with exceedingly high precision.
Hopefully, that first approximation is indeed too crude and the moduli become
massive particles by the time the theorys full solution is worked out. This will
make the new forces short-ranged and thus avoid contradiction with
experiments.
So, not only string theory is falsifiable, the real question is: why is it not
already falsified? The answer, once more, is that the theory is not developed
enough to be able to answer such questions since they lie outside the regimes
in which string theory can be studied by presently available techniques. We
should not forget, in this respect, that RQST is an entirely new and relatively
young theoretical construction. It took many decades to develop QFT to such
an extent that it could be successfully applied to actual experiments, or even to
understand that the non-observation of free quarks was not in contradiction
with QCD. Indeed, the problem of proving quark confinement turned out to be
extremely hard to solve analytically and, even today, can only be addressed
numerically through powerful dedicated computers.
The conclusion stemming from both arguments given above is that RQST is
so constrained by its mathematical and physical consistency that, in principle,
its test should be easy. Only our present inability to draw firm predictions from
its complicated equations is preventing us from saying today whether it has any
chance to survive. It is not an improvement in experimental techniques, but
rather of the theory itself, that will tell us whether this beautiful theory has any
chance to survive as a physical theory, or whether it will remain forever a
beautiful construction in search of experimental confirmation.


Time Really Passes
*

John D. Norton**
jdnorton@pitt.edu

ABSTRACT
It is common to dismiss the passage of time as illusory since its passage has not
been captured within modern physical theories. I argue that this is a mistake.
Other than the awkward fact that it does not appear in our physics, there is no
indication that the passage of time is an illusion.

A common belief among philosophers of physics is that the passage of time of
ordinary experience is merely an illusion. The idea is seductive since it explains
away the awkward fact that our best physical theories of space and time have yet
to capture this passage. I urge that we should resist the idea. We know what
illusions are like and how to detect them. Passage exhibits no sign of being an
illusion.
We have no clear idea of how to describe this passage in ways comparable
to the precision of theories of time and space in physics. We usually end up
describing passage with metaphors that prove circular and then, in
desperation, gestures. That failure is not a good reason to doubt the objectivity
of temporal passage; it is a reason to doubt the reach of our understanding.
Explaining passage away as an illusion is an instance of a desperate
stratagem that has been used to ill effect elsewhere when we become too eager
to explain away an awkward fact. Immanuel Kant urged that the Euclidean
structure of space and the causal character of the world are illusory, merely
arising through the interaction of our sensory apparatus with the things in
themselves. He thereby grounded his defense of the impossible dogma that we
know these facts with certainty. Many worlds theorists in quantum mechanics

* My thanks for stimulation discussion and reactions to members of the seminar HPS 2675
Philosophy of Space and Time, Spring Term 2009, University of Pittsburgh; and to thoughtful
remarks from a helpful, anonymous referee.

** Department of History and Philosophy of Science University of Pittsburgh
protect the possibility of some superpositions of systems at the macroscopic
level by asserting that the most basic fact of laboratory experience that
experiments have unique outcomes is an illusion.

THE PASSAGE OF TIME
Time passes.
1
Nothing fancy is meant by that. It is just the mundane fact known
to us all that future events will become present and then drift off into the past.
2

Todays eagerly anticipated lunch comes to be, satiates our hunger and then
leaves a pleasant memory. The passage of time is the presentation to our
consciousness of the successive moments of the world.
Time really passes. It is not something we imagine. It really happens; or, as
I shall argue below, our best evidence is that it does. Our sense of passage is
our largely passive experience of a fact about the way time truly is, objectively.
The fact of passage obtains independently of us. Time would continue to pass
for the smoldering ruins were we and all sentient beings in the universe
suddenly to be snuffed out.
This passage of time is one of our most powerful experiences. What is not
in that experience is the idea of a present moment, the now that has any
significant extension in space. The now we experience is purely local in
space. It is limited to that tiny part of the world that is immediately sensed by
us. There is a common presumption of a present moment that extends from
here to the moon and on to the stars. That there is such a thing is a natural
supposition, but it is speculation. The more we learn of the physics of space
and time, the less credible it becomes. For present purposes, the essential
point is that the local passage of time is quite distinct from the notion of a
spatially extended now. The former figures prominently in our experience; the
latter figures prominently in groundless speculation.

1
Whether and in what sense time passes is a venerable topic in philosophy. For entries into this
enormous literature, see Savitt 2008 and Markosian 2008.
2
Do we really experience time passing? For my assessment, see Appendix.
John D. Norton Time Really Passes 25
AN ILLUSION?
The passage of time is a real, objective fact that obtains in the world
independently of us. How, you may wonder, could we think anything else? One
possibility is that we might think that the passage of time is some sort of
illusion, an artifact of the peculiar way that our brains interact with the world.
3

Indeed that is just what you might think if you have spent a lot of time reading
modern physics.
Following from the work of Einstein, Minkowski and many more, physics
has given a wonderfully powerful conception of space and time. Relativity
theory, in its most perspicacious form, melds space and time together to form a
four-dimensional spacetime. The study of motion in space and all other
processes that unfold in them merely reduce to the study of an odd sort of
geometry that prevails in spacetime. In many ways, time turns out to be just like
space.
In this spacetime geometry, there are differences between space and time.
But a difference that somehow captures the passage of time is not to be found.
There is no passage of time. There are temporal orderings. We can identify
earlier and later stages of temporal processes and everything in between. What
we cannot find is a passing of those stages that recapitulates the presentation of
the successive moments to our consciousness, all centered on the one
preferred moment of now.
At first, that seems like an extraordinary lacuna. It is, it would seem, a
failure of our best physical theories of time to capture one of times most
important properties. However the longer one works with the physics, the less
worrisome it becomes. We find through success after success that the reach of
modern physics has grown through the past century to embrace all conceivable
spatiotemporal processes. That embrace has long held the motions of tennis
balls and the return of comets. It now extends to the transitions of electrons
between their energy levels in atoms and on to the earliest moments after the
big bang. It even gives precise meaning to the extraordinary ideas that time
may have a beginning and an end.

3
For a classic, colorful statement of the view that the passage of time is a myth, see Williams
1951. For a recent, able defense of the reality of passage from its many detractors, see Maudlin 2007,
ch. 4 On the Passage of Time.
We start to get used to the idea that our theories of space and time are
telling us all that can be said about time objectively. When we start to believe
that, we begin to invert the reasoning. If these theories do not have an
objective, factual passage of time, then perhaps it is not something factual after
all. Perhaps the universe, in all its spatiotemporal glory, just is. It is extended in
spatial and temporal directions, with notions like closer and farther spaces and
earlier and later moments. But those moments of time intrinsically possess no
special unfolding that would comprise the passage of time.
In this spacetime, we exist as objects extended in space and time. We are
world tubes of matter winding through spacetime and interacting with all the
processes in it. Part of each of our bodys world tubes are eyes, ears and a
brain, all busily sensing and interacting and signaling. Then, in a process we
understand poorly but which assuredly happens, this physical system delivers
news of the moments of time in small, serially ordered doses to consciousness.
The illusion of the passage of time, the story continues, arises from our
brains strict regime of rationing this news to these small doses, all impeccably
arranged serially, so that earlier and later are never confused, or at least at most
by milliseconds. Nothing in the objective facts of the world requires that the
news of the moments must be delivered in this rigid, serial regimen. But
something in our neural systems leads to it being so. Whatever that something
is, it is universal across all humans. That part of our neural system that rations
the news of the moments is the same in all of us. So we all have the same
illusion of the passing of time.
4, 5

4
While I believe this attitude lies behind much of the popularity of the idea that passage is a
psychological illusion, it is rare to find the argument spelled out. It is implicit in remarks made by
Carnap when he reported his discussions with Einstein over Einsteins discomfort that physical theory
does not do justice to our experience of the Now. Carnap remarked: [] all that occurs objectively
can be described in science; on the one hand the temporal sequence of events is described in physics;
and, on the other hand, the peculiarities of mans experiences with respect to time, including his
different attitude towards past, present, and future, can be described and (in principle) explained in
psychology (Carnap 1963, pp. 37-38). Carnaps remark can only cohere if we assume he believes
that physics has failed to capture the peculiarities of our experience of the Now and that this failure is
permanent.
5
On March 21, 1955, shortly before his own death, Einstein wrote to console the family of his
dear friend of over half a century, Michele Besso: And now he has preceded me briefly in bidding
farewell to this strange world. This signifies nothing. For us believing physicists the distinction
between past, present and future is only an illusion, even if a stubborn one (Hoffmann and Dukas
1975, pp. 257-258). While the remark is much repeated, we must resist the temptation of reading it
as a serious claim in philosophy of time. That reading is precluded by the circumstances, a private
BACK TO REALITY
I was, I confess, a happy believer that passage is an illusion. It did bother me a
little that we seemed to have no idea of just how the news of the moments of
time gets to be rationed to consciousness in such rigid doses. Perhaps, I
wondered, could we turn that problem over to the neuroscientists? Then there
was the odd implausibility of the whole idea that became harder to suppress.
This is what we are supposed to believe. The world tubes of our brains are
embedded in a greater world of four-dimensional spacetime that is open for
interaction with it timelessly or all at once or however you want to describe a
world in which there is no fact of passage. Yet my world tube and yours and all
others like them convey those interactions to consciousness in serial doses.
Most significantly, the delivery of the doses is perfect. There are no
revealing dislocations of serial order of the moments. While there may be
minor dislocations, there are none of the types that would definitely establish
the illusory character of passage. We do not, for example, suddenly have an
experience of next year thrown in with our experience of today; and then one of
last year; and then another from the present.
There are some minor dislocations, but they not the sort that suggests
passage is illusory. They are the sort we would expect exactly if passage were
objective, but there were occasional malfunctions of our perception of it. Take,
for example, the odd experience we have under anesthesia of no time at all
passing between the administration of the drug and its wearing off. That is
easily explained in the passage view as a suspension of that part of our neural
system that is aware of the passing moments.
This world tube is like someone resting comfortably on a sofa. The sofa
presses uniformly over the body, whose mind could in principle sense the
pressure over the full length all at once. Yet the pressures are communicated to
consciousness in a slow series that starts at the feet and marches inexorably up
the length of the body to the pillow behind the head; and it is the same for every
reclining body, without failure or serious dislocation. The result is that the
reclining body and all others like it experience an illusory passage of pressure.

condolence to a grieving family, a lack of similar remarks in Einsteins published writing on time and
the ambiguous internal warning (for us believing physicists) that his remark concerned something
other than matters of fact.
If this sofa parable sounds fantastic, then you should find equally fantastic the
same idea when applied to world tubes of brains.

DETECTING ILLUSIONS
So where does that leave us? It would be convenient for physics if the passage
of time were an illusion. But there is something odd in the idea that an element
of our experience that is so universal, and so solid and immutable, is just an
illusion. So what do we have to do to show that passage is an illusion? Here we
can take cues from the many experiences we know to be illusory.
If I hold out my outstretched fingers nearly touching at their tips in front of
my face, I will see the illusion that has amused children since the beginning of
human time. There, floating in space right in front of my face, is a finger
sausage.
There is an
enormous literature
devoted to much more
elaborate illusions.
Here is a remarkable
one that is a variant of
the Pinna motion
illusion (Picture 1).
Gaze at the black dot
in the center and,
while continuing to
focus on the black dot,
move your head
towards and away from
the image.
6

Picture 1

6
Whether the illusion is visible depends essentially on the quality of the reproduction. When it is
visible, the effect is striking and unmistakable. For a good collection of these illusions and analysis of
them, see Pinna 2009.
The effect is that the rings of triangles are set into rotation in opposite
directions. The motion is an illusion; the rings of triangles are entirely static.
We convince ourselves that both the finger sausage and the motion of the
triangles are illusions by two means.
First, we note that the illusion can be controlled and eradicated. Merely
shutting one eye leads to the sausage disappearing. Or in the case of the
motion illusion, we can eradicate the motion in any triangle merely by focusing
on that triangle as we move our heads in and out; or, more simply, just not
moving our heads.
Second, we can often identify the mechanism through which the illusion
arises. In the case of the finger sausage, it results directly from an improper
fusion of the images received by each of our eyes. The motion illusion depends
upon the triangles being seen in peripheral vision so that the figures are
blurred and normal motion clues thrown out of balance.
7

If the passage of time is an illusion, it is quite unlike these familiar examples
of illusions. It carries none of the distinguishing marks that enable us to
identify other illusions.
First, it seems impossible to eradicate passage from experience in a way
that would reveal its illusory character. Indeed there is a healthy tradition in
experimental psychology that seeks to generate temporal dislocations in our
experience. Subjects hear sounds in each ear that are delivered slightly
dislocated in time. Yet they misperceive them as simultaneous. In other
experiments, subjects are led to misperceive the exact timing of an event they
see by hearing cleverly timed audible clicks.
These sorts of experiments are quite successful in leading to dislocations of
the order of milliseconds. That sort of dislocation is remote from what one
would expect if the entirety of passage is an illusion. With all the tricks at their
disposal, why cant an inventive researcher induce dislocations of the order of a
day or a year? But if the passage of time is an objective fact independent of our
neural circuitry, that failure is no surprise. The greatest dislocation possible
would only be the milliseconds of time involved in the neural processing of the
moments once they have been delivered to our senses and are routed to
consciousness.

7
See Pinna and Brelstaff 2000.
Second, what of identifying the mechanism that restricts the delivery of
moments to consciousness into the rigid series we experience? In particular,
what in the neural machinery blocks us from having perceptions of tomorrow
or next year? While neuroscientists have made enormous advances in recent
years, I do not think that circuitry blocking this avenue of perception has been
identified. But if passage is an illusion of our perception, there must be some
mechanism that blocks us perceiving the future.

THE DESPERATE STRATAGEM
There are occasions in which our best science requires us to dismiss some fact
of experience as an illusion. All our ordinary experience of water and air is that
they are perfectly continuous fluids. Yet our best science tells us that is an
illusion. On a sufficiently fine scale both have the granularity of molecules. The
appearance of continuity is an illusion. But it is one that is readily explicable by
the extremely small size of atoms.
Again, light appears to us to propagate instantaneously in ordinary
experience. Yet it is essential for relativity theory that it have a finite speed of
propagation. So we dismiss the appearance of instantaneous propagation as an
illusion. Once again, it is readily explicable by the extremely short propagation
times needed, which are well below those we can discern in ordinary processes.
Now consider the passage of time. Is there a comparable reason in the
known physics of space and time to dismiss it as an illusion? I know of none.
The only stimulus is a negative one. We dont find passage in our present
theories and we would like to preserve the vanity that our physical theories of
time have captured all the important facts of time. So we protect our vanity by
the stratagem of dismissing passage as an illusion.
We are not alone in this stratagem. There is a fine tradition of dismissing
uncomfortable facts as illusions. Immanuel Kant believed that we knew some of
the facts of science with certainty. They included the facts that space is
Euclidean and all processes causal. He was brought by Humes skeptical
arguments to see that ordinary inductive exploration of the world could not
give this assurance of certainty. Yet he was certain and that certainty needed to
be protected. So he applied the same stratagem to protect the certainty of his
belief from the fallibility of our epistemic devices.
That space is Euclidean and processes causal are not facts of the world
independently of us, that is, facts about things in themselves. Rather, he urged,
they arise through the operation of our sensory apparatus as forms of intuition.
This is fancy philosophers talk for a simple idea: the truths of Euclidean
geometry are, in the end, illusions. We have no idea of the mechanism through
which these illusions are produced. Any talk of eyes, brains and neurons
already presupposes the description of things in space, the very thing we are
trying to portray as illusory. Nonetheless we are assured these are illusions
over which certainty is possible.
The Everrett, or as it is popularly called, the many worlds interpretation
of quantum theory is based on the supposition that the linearity of quantum
theory of the atomic level also persists all the way up to the macroscopic level.
This linearity allows particles to evolve over time into linear superpositions. An
electron can evolve into an equal parts superposition of a spin up and a spin
down state. If this linearity persists unchecked at the macroscopic level, it is
possible for macroscopic objects to evolve into comparable superpositions.
This is what Schroedinger demonstrated with his celebrated cat thought
experiment. The cat evolves into a superposition of dead and alive. That
we do not see this macroscopic superposition the cat is just dead, say, when
we check would seem to put an end to the supposition of macroscopic
linearity. It is an experimental refutation that is replicated every time a Geiger
counter clicks, affirming that the counter has not evolved into a linear
superposition of detection and no detection, but has unambiguously detected a
radioactive decay.
The desperate stratagem intervenes. We are told that there is another live
cat we cannot see, so that the definiteness of its death is an illusion. If the
invisibility to us of the other cat is worrisome, we are further assured that there
is another one of us observing the live cat as well.

AS GOOD AS IT GETS
We should stop protecting our vanity and admit what is now becoming obvious
to me. We have no good grounds for dismissing the passage of time as an
illusion. It has none of the marks of an illusion. Rather, it has all the marks of an
objective process whose existence is independent of the existence of we
humans.
The real and troubling mystery lies in asking what more can be said. Beyond
the barest fact of objectivity, it is not at all clear how properly even to describe
the passage of time in precise terms. Our usual approach is to employ a
metaphor. The term passage itself literally means moving, which in turn
presupposes change in time. That makes the metaphor circular. The difficulty
is that temporal passage metaphors like these are so fundamental to our
descriptions that we have none more basic that can be used to describe the
passage of time itself.
It gets worse if we seek a description of passage consonant with familiar
physical theories. Yet should not an objective property of time admit
description by the hugely successful techniques of modern physical theories?
The natural candidate is to talk of time passing with respect to some other time.
But that is already too much. It introduces an extra time above and beyond the
one we use in our spacetime theories. While such a thing is possible, it is a
speculation every bit as dangerous as the idea that the now is extended
infinitely in space past Alpha Centauri and beyond.

APPENDIX
Do we really have an experience of time passing? Is it something we perceive
directly such as when we see the blueness of the sky or hear the noise of the
wind? I think we do experience it. To see why, let us break the experience up
into two parts.
First, I think it is uncontroversial that we experience time as moments. That
is, we experience the present moment directly. We have indirect access to the
past through memory and other traces and we have even more tenuous access
to the future through predictive methods. We do not experience temporal
relations directly. That is, we do not experience directly that last Wednesdays
snowstorm was earlier than todays rain shower. We experience todays rain
shower and recall the snowstorm, inferring from that their temporal relation.
Second, we do have a direct perception of the changing of the present
moment. That is clearest in our perception of motion. For example, consider
two bicycle wheels, one rotating and one not rotating. As long as the rotation is
not too fast or too slow, we perceive directly which is which and perceive the
motion directly. How that comes about is an issue for experimental
psychology. My guess is that the important fact will be that our sensed present
moment actually lasts some definite time less than a second. During that time,
the wheel turns appreciably and we grasp that as a perception of motion.
That we do perceive motion at some primitive level in our experience is
supported by the existence of motion illusions. We see the image in the main
text move, even though it does not.
This perception of motion is not the perception of the passage of time.
However it is part of it. It may well be impossible to give a non-circular
definition in ordinary words of such a primitive notion as the experience of the
passage of time. However I do think we come close with the idea of the totality
of our perceptions of changes underway in the present moment. Even if we are
in an environment that it totally static an empty, noiseless doctors waiting
room we still perceive our own bodily functions changing, such as our
breathing and heartbeat, and even the process of our thought.
Nothing philosophically fancy is intended by experience in this context.
All that is intended is the ordinary use of the word. There is no claim made that
we can divide what we know of the world into two parts: the pure sense data
that are somehow prior to cognition and then the inferences we make from
them. Indeed it seems incredible to me to that our knowledge or beliefs of the
world could be divided so cleanly in this way. At best we can identify things
whose experiential contribution is greater; and I urge that our momentary
sense of the passage of time has such a large experiential contribution.

REFERENCES
Carnap, R. (1963). Carnaps Intellectual Biography. In P. A. Schilpp (Ed.),
The Philosophy of Rudolf Carnap (pp. 1-84). La Salle, IL: Open Court.
Hoffmann, B., & Dukas, H. (1975). Albert Einstein. Frogmore, St Albans,
UK: Paladin.
Markosian, N. (2008). Time. In Stanford Encyclopedia of Philosophy.
<http://plato.stanford.edu/archives/win2008/entries/time/>
Maudlin, T. (2007). The Metaphysics within Physics. Oxford: Oxford
University Press.
Pinna, B. (2009). Pinna Illusion. Scholarpedia, 4(2), 6656.
<www.scholarpedia.org/article/Pinna_illusion>
Pinna, B., & Brelstaff, G. J. (2000). A New Visual Illusion of Relative Motion.
Vision Research, 49(16), 2091-2096.
Savitt, S. (2008). Being and Becoming in Modern Physics. In Stanford
Encyclopedia of Philosophy.
<http://plato.stanford.edu/archives/win2008/entries/spacetime-
bebecome/>
Williams, D. (1951). The Myth of Passage. The Journal of Philosophy, 48(15),
457-472.
From Timeless Physical Theory to Timelessness

Sam Baron*
samuel.baron@sydney.edu.au
Peter Evans*

peter.evans@sydney.edu.au
Kristie Miller*
kristie_miller@yahoo.com

ABSTRACT
This paper addresses the extent to which both Julian Barbours Machian
formulation of general relativity and his interpretation of canonical quantum
gravity can be called timeless. We differentiate two types of timelessness in
Barbours (1994a, 1994b and 1999c). We argue that Barbours metaphysical
contention that ours is a timeless world is crucially lacking an account of the
essential features of timean account of what features our world would need to
have if it were to count as being one in which there is time. We attempt to
provide such an account through considerations of both the representation of
time in physical theory and in orthodox metaphysical analyses. We
subsequently argue that Barbours claim of timelessness is dubious with
respect to his Machian formulation of general relativity but warranted with
respect to his interpretation of canonical quantum gravity. We conclude by
discussing the extent to which we should be concerned by the implications of
Barbours view.

1. INTRODUCTION
It is now ten years since Julian Barbour first introduced his rather revolutionary
views about the fundamental structure of our world to a general audience, and
it is fair to say that although his views have been well scrutinised by the
theoretical physics community, they have largely been ignored by the

* University of Sydney

philosophical community.
1
This is a pity, since the analysis that Barbour
provides of this particular attempt to unify general relativity and quantum
mechanics is rich in startling metaphysical consequences. He argues that ours
is a timeless world: the apparent experience of change, persistence and motion,
of memory and anticipation, are merely apparent. There is no unique way our
world was in the past, nor will be in the future. There is just a static
configuration space filled with three dimensional instants and no path
through that space that can rightly be thought as a history of a world. This is a
radical conclusion that would seem to overhaul almost all that we think we
know about the world.
Barbours project is multifaceted: beginning with his reformulation of
general relativity and using canonical quantisation techniques to yield a
particular representation of canonical quantum gravity, Barbour proposes a
picture of the fundamental structure of reality through his interpretation of
these physical theories. Taking the technical details of this project as given, it is
worth considering whether some of the radical metaphysical conclusions that
Barbour draws really are a consequence of this proposed picture of reality. We
begin, in section 2, by offering an exploration of Barbours views. This allows
us, in section 3, to consider the extent to which Barbours view entails that our
world has no time. In order for Barbour to extract a metaphysical conclusion
from his physical picture, a further premise about the nature of time is
required. In particular, we need a premise that states the necessary features of
time, and we need these necessary features to be ones that Barbours physical
picture tells us that our world lacks. We consider two possible ways to
understand the essential features of time: through an analysis of the
representation of time in physical theory, on the one hand, and in terms of
more orthodox metaphysical analysis, on the other. We argue that on either
way of construing the essential features of time only one arm of Barbours
project is genuinely timeless. Finally, in section 4, we consider some
outstanding worries with Barbours view if we take his conclusion seriously. In
particular we consider two issues: first we consider whether his account of the
experience of time, motion and change is a good one; and second we consider a
worrying sceptical scenario that may be entailed by his view. We argue that
although this scepticism arises, it is not as pernicious as one might have
supposed.

1
Some notable exceptions include Butterfield 2001, Healey 2002 and Ismael 2002.
Baron, Evans & Miller From Timeless Physical Theory to Timelessness 37

2. EXPLORING BARBOUR
Barbours interpretation of canonical quantum gravity is presented both in his
(1999c) and in a pair of companion papers which preceded it (1994a, 1994b)
and is best understood as comprised of two major parts. The first part consists
of an argument that classical general relativity can be formulated in a Machian,
and thus in some sense timeless, fashion. The key to this argument is the claim
that general relativity is an implementation of a dynamical theory (with dynamic
geometry) that can be formulated as a reparametrisation invariant geodesic
principle on the relative configuration space of all possible instants of time. In
the second part of his project, Barbour examines a theory of quantum gravity
constructed via the quantisation of this formulation of general relativity, his
interpretation of which is itself also timeless. The key to this second part is the
proposal that the Wheeler-DeWitt equation, the timeless dynamical law of
canonical quantum gravity, can be interpreted as a probability distribution,
defined in terms of the relative configurations, that concentrates the quantum
mechanical probability on time capsules. Importantly for our purposes in this
paper, the sense in which Barbours Machian formulation of general relativity is
timeless differs in some crucial respects to the sense in which his interpretation
of quantum gravity is timeless; more on this to follow. Firstly, let us begin
characterising Barbours position in more detail by considering the structure
of configuration space.
According to classical mechanics, the objects of the universe at any
particular instant are in some definite configuration relative to one another. It
is usual to refer to any one of these configurations as an instant of time. While
the notion of an instant loses some of its physical significance in relativity
theory (since an instant is a hyperplane or hypersurface in a four dimensional
spacetime, and specification of these in relativity theory is arbitrary), this
significance is regained when describing dynamical evolution: the essence of
dynamics is to describe the evolution of specific data on such hyperplanes or
hypersurfaces. The dynamical evolution of a particular physical system can be
formulated in terms of a configuration space Q which represents all the
successive configurations through which the system passes as it evolves in
time. The path that a system describes through its Q is a path of least action, or
geodesic, where the action is a function of the energy of the system and the
physical laws governing the dynamical evolution; this is the Hamiltonian
formulation of a physical theory.

In taking the system in question to be the universe, one can move from the
configuration space of the universe Q to a relative configuration space of the
universe Q
0
by factoring out the six frame variables that specify the centre-of-
mass coordinates and the orientation of the system; thereby one can remove
any absolute frame from the description of the universe. The formulation then
of a relational theory of dynamics requires that we define the action between
any two neighbouring points of Q
0
in terms of the best-matching of the
intrinsic differences between the two points in question, quantified by a
Pythagorean least-squares fit; this is the essence of a Machian formulation of a
physical theory. Barbour describes this process as like placing two relative
configurations on top of each other and then supposing them moved relative to
each other until the intrinsic difference between them is least. Via this Machian
definition of the action, geodesics through Q
0
can be obtained.
Barbour claims that general relativity is a special case of such a Machian
formulation of a physical theory where the instants of time in the relative
configuration space are no longer configurations of particles in Euclidean
space but three dimensional Riemannian spaces endowed with 3-geometries.
The sense in which such a Machian theory is timeless is the following:
time [ is] obtained from a timeless[] heap of relative configurations [] by
placing the configurations on top of each other in the best-matching positions
(horizontal stacking []) and spacing them apart (vertical stacking) in
accordance with their [] differences. (Barbour 1994a, p. 2863)
Thus, while time is not present in any individual three dimensional relative
configuration, the Machian principle on the relative configuration space
enables time to be reconstructed as an ordering of the instants along a
geodesic. Barbour contends that the fundamental property of general relativity
is that it is a timeless theory of the relationships of 3-geometries and the
relative configuration space is the arena in which we should fundamentally
describe reality.
2
With this Machian notion of timelessness in mind, let us turn
to the second part of Barbours project, his timeless interpretation of canonical
quantum gravity.
Canonical quantum gravity is a theory of quantum gravity that is obtained
from the Hamiltonian formulation of general relativity and canonical
quantisation techniques. Thus while in some sense Barbours interpretation of

2
Barbours Machian formulation of general relativity is not without its technical problems. See
Pooley 2001.

canonical quantum gravity is inescapably connected to the Hamiltonian
formulation of general relativity, the timelessness of the former is not related to
the Machian sense of timelessness of the latter. To see this, let us consider how
Barbours interpretation grows from the sort of union he envisages between
general relativity and quantum theory. If the arena for our fundamental classical
description of reality is the relative configuration space as above, then the
compatibility of quantum theory with this description is dependent upon the
viability of formulating quantum theory in terms of instantaneous relative
configurations of systems. Barbour notes that the time independent quantum
dynamical laws can be represented on a space of three dimensional relative
configurations; the Schrdinger wavefunction of any system is defined over all
its possible configurations, and thus instead of describing a unique classical
history in the configuration space of the system, the quantum wavefunction
explores all configurations. If we were then to extend the relative configuration
space, on which the time independent Schrdinger equation is usually applied,
to the relative configuration space of the universe, Barbour suggests that the
Wheeler-DeWitt equation could be used to describe a static wavefunction
that takes relative configurations as its argument.
On this view, the notion of a Hilbert space representing the state space of
some subsystem of the universe is simply redundant; Barbour proposes to treat
the universe as a single holistic quantum system. In any one configuration, no
distinction is possible between quantum system and measurement device: all
are simply part of a particular configuration of the universe. The sole role of the
wavefunction, as in Borns probability interpretation, is to say how likely the
actualising of a given configuration is. These probabilities are not, however,
time dependent nor are they conditioned on prior knowledge and tied to
measurement setups; they are given once and for all for the possible
configurations that the universe could be in. It is in this sense that Barbours
interpretation of canonical quantum gravity is timeless.
In trying to motivate an intuitive picture of his model, Barbour contrasts
two ways that we might imagine such a universe: externally and internally
(1994b, p. 2881). From an external viewpoint, we can imagine each relative
configuration of the space to exist as a heap of possibilities.
3
We can then
divide the space into infinitesimal hypercubes, take the value of in each

3
Barbour emphasises that this is called a heap because each point in the relative configuration
space has, unlike an ordinary manifold, an individual existence outside the space, i.e., a three
dimensional configuration.

hypercube, calculate
*
and place a number proportional to
*
of
identical copies of a representative configuration of that hypercube into a
second heap called the heap of actualities.
4
We may now suppose that drawing
one configuration at random from the heap of actualities actualises that
configuration. Thus, a probable configuration is more likely to be actualised
than one that is improbable.
From an internal viewpoint, we have it that our direct experience, including
that of motion, is correlated only with configurations in our brains.
Our seeing motion at some instant is correlated with a single configuration of
our brain that contains, so to speak, several stills of a movie that we are aware of
at once and interpret as motion. (Barbour 1994b, p. 2883)
The connection between the internal and external views is that while some
divine mathematician actualises (by random selection) one particular
configuration of the universe, it seems to us as though we are inside part of that
configuration and have direct awareness of that part as an experienced instant.
The problem in orthodox quantum theory concerning the reality of the
unactualised possibilities is compounded in Barbours quantum gravity
since one even has to ask whether events of which we have vivid memories are
actually experienced. This is because everything we experience in any instant,
including the memories themselves, must be coded in our instantaneous brain
configuration. Records of apparent past events are in fact details in the present
configuration. And all the timeless theory tells us is that each such
configuration has a certain probability. (Barbour 1994b, p. 2883)
Thus while we have direct evidence that the present configuration is actualised,
we are epistemically locked in this configuration and therefore have no warrant
for believing that any other instant is actually experienced. Barbours quantum
gravity does seem to come perilously close to solipsism of the instant
(Barbour 1994b p. 2883).
The most significant element of Barbours interpretation of quantum
gravity is the notion of a time capsule. A time capsule is a static configuration
of part or all of the universe containing structures which suggest they are
mutually consistent records of processes that took place in a past in accordance
with certain laws. It is the existence of such special configurations that Barbour

4
In what way we are to imagine this heap of actualities is unclear. We abstain from exploring this
issue.

claims allow us to recover the appearance of time from a timeless reality. Since
the set of all time capsules has negligible measure amongst the set of all
possible configurations, Barbours proposal is conditional upon his suggestion
that the solution to the Wheeler-DeWitt equation concentrates the quantum
probability distribution on time capsules, thereby making it probable that we
would find ourselves in a three dimensional configuration that contained
evidence of having been created by a dynamical process. With this, then, the
moral of the second part of Barbours project is that time is not a framework in
which the configurations of the world evolve [rather] time exists only so far as
concrete configurations express it in their structure (Barbour 1994b, p.
2885).

3. LOCATING BARBOUR
For those more familiar with work in the philosophy of time, Barbours claim of
timelessness may seem startling: it is not immediately obvious how or why such
a conclusion should follow from Barbours treatment of classical general
relativity or his interpretation of quantum gravity. Thus while we leave the task
of challenging the technical details of Machian general relativity and canonical
quantum gravity to Barbours fellow physicists, we pose here a challenge of a
different sort. If Barbours two theories are indeed fair descriptions of the
classical and quantum worlds respectively, what does this tell us about time in
our universe? In particular, should we conclude that ours is a timeless
universe?
One might begin by asking how we find ourselves in a position to decide
that some particular phenomenon, class of phenomena or kind of object does
not exist; after all, we do not discover absences. In the case of phenomena that
are posited by scientific theories, we usually decide that the posits are unreal if
we find that (i) the theory that posits them is false, (ii) there is no true theory
whose posits are sufficiently like those of the original theory and (iii) we are not
inclined to say that the posits of the former just are the posits of the latter
appropriately reconstrued. For instance, when the theory that posited
phlogiston is found to be false we become error theorists about phlogiston.
Had the posited features of phlogiston been sufficiently like those of oxygen in
crucial ways we might instead have discovered that there is phlogiston but it is
somewhat different than we first thought. Many philosophers are tempted to

say that this same story can be applied to terms that are not introduced via
scientific theories but are the folk terms we find in everyday discourse. Thus
they are inclined to say that we became error theorists about witches because
we discovered that the folk theory that posited witches was false and no close
successor of the theory was true (and hence we did not discover that witches
were rather different than we had supposed). put in broad terms, one might say
that one will become an error theorist about the xs, just in case the core or
essential claims about the nature of the xs (whether these are part of scientific
theory or folk theory) turn out to be false.
According to Barbour we ought to be error theorists about time. If he is
right, this must be because certain core or essential claims about the nature of
time turn out not to be true. It now becomes clear why his conclusion might
seem startling: for although Barbour offers an interesting reformulation of
general relativity and a solid interpretation of the formalism of canonical
quantum gravity, he says very little about what our universe would need to be
like for it to be a universe in which there is time; he does not characterise
times essential features. The crucial link in the argument for timelessness
therefore appears to be missing: namely the link that takes us from the claim
that particular features are essential to time to the claim that in virtue of
Barbours classical and quantum theories those features are absent, and thus to
the conclusion that there is no time.
Thus in this section we explore two routes for characterising the essential
features of time. The first route, section 3.1, proceeds via the features of time
as they are represented in physical theory while the second route, section 3.2,
is a consideration of the features of time as one might find them in more
orthodox metaphysical analyses. Through these considerations of the essential
features of time we hope to offer some suggestions for filling the
aforementioned lacuna in Barbours own argument for timelessness. We then
evaluate the extent to which Barbours argument thus construed is compelling.
Before we embark on this examination let us introduce some terminology
that will help us distinguish the different senses of timelessness in Barbours
work that we wish to highlight below. The core of Barbours claim of
timelessness is that the fundamental elements of our description of reality,
both in his classical theory and his quantum theory, are three dimensional
relative configurations, i.e., frozen instants of time which lack a temporal
dimension. However, as we noted above, there is a significant difference
between the timelessness of Barbours Machian formulation of general

relativity and the timelessness of his interpretation of canonical quantum
gravity. In the former sense, the timelessness of the relative configurations is
supplemented by a Machian reconstruction of temporal structure: using only
the data present within the set of relative configurations we can reconstruct
reparametrisation invariant geodesics through Q
0
and thus we can read off a
temporal metric from these geodesics. As Butterfield remarks, the Machian
formulation of general relativity will deserve to be called timeless, in that
there is no time metric in Q
0
; rather [] the time metric is definable from the
dynamics (Butterfield 2001, p. 15). Thus while the theory is timeless in the
sense that a time dimension is absent from the fundamental elements of the
theory, a temporal metric of sorts can be reconstructed from these timeless
elements. Let us call this Machian timelessness.
The same cannot be said of the latter sense of timelessness. The
timelessness of Barbours interpretation of canonical quantum gravity is again
manifest in the absence of a time dimension in the fundamental elements of the
theory but, in contrast to Machian timelessness, it is then compounded by the
further structure in the theory: there exists a time independent (static)
quantum probability distribution (QPD) across the relative configuration space
that is concentrated upon time capsules, i.e., special three dimensional
configurations that merely appear as though they have been created from a
dynamical process. Thus in quantising the Machian formulation of general
relativity to yield Barbours particular interpretation of quantum gravity we lose
an element of Machian temporal reconstruction and gain an account of
temporal appearances in the form of time capsules. Let us call this sense of
timelessness QPD timelessness. Barbour does not provide a clear statement
distinguishing these two senses of timelessness. Indeed, Butterfield again, the
book [The End of Time] gives the misleading impression that Barbours
various views are closely connected one with another (Butterfield 2001, p. 3).
The distinction between these two senses of timelessness will become more
clear through our consideration of the essential features of time.

3.1. REPRESENTING TIME IN PHYSICAL THEORY
According to Rovelli, when we use the word time there are many attributes of
time to which we might be referring. Indeed, in both his (1995) and his
(2004), Rovelli identifies up to nine distinct attributes of time that we find
littered throughout our contemporary physical theories and folk concepts,
including directionality, uniqueness and globality amongst others. Rovellis

project concerning the notion of time is a terminological one: in identifying
these different senses of time, he might begin to alleviate some of the
ambiguity that abounds in the philosophy of physics literature. We wish to
utilise this prescription to clarify Barbours error theoretic claim.
Rovelli proposes that our contemporary physical theories and folk concepts
that refer to time can be arranged in a hierarchical structure in which an
increase in universality corresponds to a decrease in the possible attributes of
time to which we might be referring. Thus when we refer to time in orthodox
general relativity, since it is one of our most universal physical theories, there
are only two possible attributes of time to which we can be referring: linearity,
time can be used to refer to a one dimensional substructure of ordered
temporal instants; and metricity, time can be used to refer to the meaningful
measure of distance between any two time instants. Given the focus on these
two features of time in general relativity, we can characterise the essential
features of time that Barbour might be denying as just these features: linearity
and metricity. This now gives us a straightforward manner in which to evaluate
Barbours claim that both his Machian formulation of general relativity and his
interpretation of quantum gravity entail that we live in a timeless world: in each
case we consider the extent to which linearity and metricity can be extrapolated
from Barbours interpretations of the relevant physical theories.
In Barbours Machian formulation of general relativity and his
interpretation of canonical quantum gravity the fundamental elements of the
theory are three dimensional relative configurations. If we consider a single
relative configuration in isolation, there is no one dimensional substructure
therein to identify as time and no way to meaningfully measure the temporal
distance from this configuration to any other. Thus in both his classical and
quantum theories, when we consider a single instant in the relative
configuration space, we notice that there is no fundamental linear or metric
structure to be found.
However, an integral part of Barbours Machian formulation of general
relativity is the specific and detailed Machian algorithm that enables one to
define a meaningful measure of distance between any two points in the relative
configuration space and thus describe a linear ordering of instants along a
geodesic, thereby recovering both linearity and metricity. Thus when it comes
to Barbours Machian formulation of general relativity, these two particular
features of time are not entirely absent from the theory. The relevant features
exist, it is just that they emerge out of the three dimensional points in the

relative configuration space via this specific best-matching algorithm. Thus it
becomes apparent that talk of time is not misplaced in Barbours Machian
general relativity; there is time qua linearity and metricity, it is just that time is
not a fundamental component of the theory, admitting of a straightforward
reduction to the relative configuration space.
Given this reduction, it would be very odd for Barbour to claim, with
respect to his Machian formulation of general relativity, that time ought to be
eliminated from our ontology. This is because this sort of reduction about
some phenomenon is rarely a reason to eliminate the reduced phenomenon
from our ontology, except in cases where there is an essential feature of the
phenomenon to be reduced that is not captured by its putative reductive base.
However this is clearly not the case here: the relative configuration space of
Barbours formulation of general relativity explicitly yields a temporal
parameter that corresponds directly with that of orthodox general relativity.
Thus it appears that Barbours claim that his Machian formulation of general
relativity is timeless involves a touch of hyperbole. For Barbour, his
formulation of general relativity is timeless simply because time is not
fundamental; it is not, however, timeless because time does not exist.
Is it possible to strengthen Barbours argument here turning his Machian
timelessness into a more full-blown error theory about time? Well, one option
might be for Barbour to claim that being one of the fundamental posits in our
best physical theory is essential to the nature of time. If this conceptual claim
were true then it would seem that a more robust error theoretic conclusion
would follow from the fact that no fundamental linear or metric structure is
present in the theory.
Even if this were correct, however, it seems to us that there is a ready
response available on behalf of the temporal realist: namely, that it is not at all
obvious that it is part of our conceptual grasp of the notion of time that,
whatever time is, it is fundamental in the sense that it is posited by the most
fundamental physical account of our universe. It might be that time is
fundamental in the sense that we find ourselves unable to imagine being able to
engage in ordinary talk without appealing to temporal relations, and unable to
imagine what it would be like to experience the world without experiencing it
in terms of temporal relations, but that is not at all to say that we suppose it
essential to time that it is physically fundamental.
Another option for turning Barbours Machian formulation of general
relativity into a robust error theory might be via some form of an

indispensability argument. These arguments are familiar, particularly in the
philosophy of science. They generally proceed via the claim that we ought to be
ontologically committed to all and only the entities indispensable to our best
scientific theories. Indispensability, in this context, is replacing the Quinean
idea that we regiment our best scientific theories in first-order logic and in so
doing reveal their ontological commitments. Rather, the idea is that we should
be committed to those entities that are indispensable to those theories, where
an entity is indispensable to a theory just in case, roughly, a nominalised
version of the theory (one that does not quantify over the entity in question) is
less theoretically virtuous than the non-nominalised version.
With this in mind, one might attempt to use the following argument to
bolster Barbours view:
(i) Time does not play a role in the Machian formulation of general
relativity.
(ii) If time does not play a role in the Machian formulation of general
relativity, then time is dispensable from one of our most basic physical
theories.
(iii) If time is dispensable from one of our most basic physical theories
then we ought not to be committed to the existence of time.
(iv) Therefore, we ought not to be committed to the existence of time.
Although this would give Barbour his error theoretic conclusion, it seems to us
that this argument is sound only if we ought to be committed not only to all of
the indispensable posits of our best scientific theories, but to only those posits.
However, this is a controversial claim: some philosophers are tempted to think
that our ontological commitments should be broader than science alone allows
(i.e., Lewis commitment to the existence of concrete possible worlds).
In addition, the argument is sound only if we are committed to all and only
the posits of fundamental physics. But again, this is controversial. Higher level
theories commonly quantify over temporal phenomena (like instants). For
example, theories of meteorology, economics, psychology, and likely some
higher level theories of physics itself. And it is not at all clear that talk of time
can be eliminated from such theories without a loss of theoretical virtue. If that
is right, then quantifying over time is not dispensable to our best theories
broadly construed. Indeed, instead we seem to find that we ought to be
committed to the existence of time. Only if we ignore these other theories and
focus on fundamental physics should we conclude that there is no time. But if

we do that then we should also conclude that most of the ordinary objects in
our common sense ontology do not exist, since they are dispensable to
fundamental physical theory. Since that seems too much to be plausible it may
not be the right set of theories to take into account when we are trying to figure
out what exists. But once we take into account the right set of theories, the
claim that time is dispensable may seem less obvious, and with it the conclusion
that we should be temporal error theorists may seem less compelling.
If what we have said so far is along the right lines, then it is hard to see how
Barbour might successfully argue for an error theoretic conclusion based only
on his Machian interpretation of General Relativity. When it comes to
Barbours interpretation of canonical quantum gravity, however, it seems that
Barbour makes a much stronger case for an error theory. This is because,
unlike Barbours Machian formulation of general relativity, there is no
algorithm specified for defining a meaningful measure of distance between the
relative configurations and thus there is no linear ordering of the three
dimensional instants; there is only the appearance of an illusory history from
within each time capsule. What this shows us is that QPD timelessness is
genuine timelessness in the error theoretic sense: not only is it the case that
there is no linear or metric structure in the theory, such structure is not
recoverable from the relative configuration space in any sense. At best, there is
the mere illusion of linear and metric structure via the time capsules.
5

Thus, it is only QPD timelessness and not Machian timelessness that gives
Barbour the sort of error theoretic conclusion that he seems to want. This is
important because, as we shall now show, this interpretation of the difference
between Machian timelessness and QPD timelessness in terms of the
representation of time in physical theory dovetails nicely with what we take to
be the most plausible metaphysical interpretation of the difference between
these two senses of timelessness. Indeed, both the physical and metaphysical
interpretations suggest that it is only QPD timelessness that is genuine
timelessness.

3.2. TIME IN METAPHYSICS
In the preceding discussion we considered one particular route towards
characterising the essential features of time, i.e., as they are represented in

5
For an interesting discussion of structure that might be associated with the set of time capsules,
see Healey 2002.

physical theory. In this section we pursue an alternate route to this end in
terms of more orthodox metaphysical analyses. Metaphysicians have long
worried about what sorts of features might be essential to time and, by
considering some widely held views in the metaphysics of time, we can begin to
assess the extent to which these views might be appropriate for filling the
lacuna in Barbours argument for timelessness. In fact, we will see below that
these two routes we consider are actually not so different; the metaphysical
characterisation of the essential features of time we present here corresponds
closely to the characterisation we presented in the previous section. One might
even go so far to say that the following metaphysical analysis is a mere
terminological variant of the argument in terms of the temporal structure of
physical theory (more on this below).
McTaggarts famous distinction between the A-series and the B-series will
be familiar to most (1908). The A-series and the B-series constitute two
distinct ways of ordering times, events and so on. The B-series orders times in
terms of the relations of earlier than, later than and simultaneous with. These
relations are taken to be unchanging. That is, for any two times (or any two
events) t
1
and t
2
in some world W, if t
1
and t
2
are related by some B-theoretic
relation, R, then they are R-related from the perspective of any time in W. Or,
as it is commonly described, t
1
and t
2
are tenselessly related in W. The A-
series, by contrast, orders times in terms of whether they are objectively past,
present or future.
The easiest way to get a handle on the A-series is in terms of the monadic
properties of pastness, presentness and futurity. The idea is that for any time in
the A-series, that time instantiates a particular A-theoretic property that
determines its place in that series. So, for example, suppose that there are two
times t
1
and t
2
that are located in the A-series. On this view, their location in the
A-series will be determined by the monadic properties they instantiate. Thus,
t
1
might instantiate the monadic property of pastness (say), and t
2
might
instantiate the monadic property of presentness. Unlike the B-series, however,
the monadic properties that t
1
and t
2
instantiate are dynamic. That is, it is
usually thought that if t
1
instantiates the property of presentness, then it will
eventually instantiate the property of pastness with the passage of time. Indeed,
some A-theorists are inclined to think that t
2
becomes more past as the
monadic property of the present shifts from one time to the next. Regardless,
this shift in A-theoretic properties is usually attributed to the objective flow of
time and marks the principle difference between the A-series and the B-series.

The B-theory of time is the view that once we have laid down the B-series,
we have thereby completed our description of time. The A-theory, by contrast,
is the view that the B-series constitutes an incomplete description of temporal
reality. In order to complete this description we must also accept the reality of
the A-series.
Given this distinction, there are really two routes to a temporal error
theory. Almost everyone supposes that if there is an A-series in a world, then
there is also a B-series in that world, but that the presence of a B-series does
not entail the presence of an A-series. Thus one might take the A-series to be
essential to time (as A-theorists do) and think that actually there is no A-series,
thus concluding that actually there is no time. Or one might think that the B-
series is essential to time (as B-theorists do) and think that actually there is no
B-series (and thus also no A-series), thus concluding that actually there is no
time. Or one might think that it is essential to time that there is either an A- or a
B-series, but that actually there is neither, and thus actually there is no time.
The last two of these options effectively collapse into one, since regardless of
whether one thinks that either the A- or the B-series is necessary for the
existence of time, or one thinks that the B-series alone is necessary for the
existence of time, since the absence of the B-series in a world entails the
absence of the A-series in that world
6
, in either case a world without the B-
series is a world without time. Thus we consider just the first and second of
these options since the second entails the third.
Is either of these routes to error theory one that Barbour intends? It should
be clear from section 2 that the physical picture he proposes is incompatible
with the existence of an A-series. He denies that the world is dynamical in the
way that A-theorists think it to be: there is merely the static state of the
configuration space, nothing comes into being or passes from being, and
nothing comes to be the present that was past and will be future.

6
The A-theorist may contest this conceptual claim. In particular, they may object to the
implication that the B-series is necessary for the existence of the A-series. Rather, the A-theorist might
contend that the A-series is, in some sense, metaphysically primitive and it is the A-series that is, in
fact, necessary for the existence of the B-series. If this is the case, then contra what we say below,
endorsing a view according to which there is no A-series would entail a robust error theory about time.
Nevertheless, given that Barbour does not discuss the A-theory in any great detail, it is unlikely that
this is what he means by the claim that there is no time. Still, there are some interesting issues here:
what is the right conceptualisation of time? And what implications does it have for standard
metaphysics? Unfortunately, we do not have the space to go into these issues in detail here.

So if Barbour thought that the A-series were essential to temporality, then
his error theoretic conclusion would follow. But if this were all there were to
his view, it would not be very startling. After all, most physicists hold that the
A-series is incompatible with our best physics and that, as a result, there is no
actual A-series. If this were Barbours view, then although he might offer a
substantially new physical theory, his claims about temporal error theory would
not issue from any particular features of that new theory, but merely from the
well canvassed fact that modern physics offers us a B-theoretic conception of
our universe that makes little room for the features posited by the A-theorist.
Moreover, if that were Barbours route to error theory, it would not, we
think, be a particularly compelling one. For suppose there were a world with a
B-series and no A-series. Then in that world there are events related by earlier
and later than relations, and indeed related by temporal durations (at least
relative to any given frame). It still makes sense to talk about when an event
happened, and what happened at the same time (though not in a frame invariant
manner) and it makes sense to make appointments at certain times in the
future, and to anticipate those future events and to wish they were happening
sooner. Thus it seems entirely open to someone to respond to such an error
theorist by arguing that she has misunderstood the folk concept of time, and by
doing so has invested that notion with essential features, namely the A-
theoretic features, that the folk notion simply never had. The presence of the
B-series, she will maintain, is sufficient for the existence of time since there are
perfectly good relations of earlier and later than and so forth and these are
sufficient to make it the case that our everyday term time refers. Thus if
Barbours view were simply that ours is not an A-theoretic world, and thus that
temporal error theory follows it would seem entirely plausible to respond that
the B-theory is all that is necessary for temporality and that ours is a B-theoretic
world.
Fortunately, Barbours claim that there is no time is not, we think, best
interpreted as the claim that the A-series does not exist. Rather, it seems more
plausible to read him as holding that it is essential to time that there is a B-
series, and that since actually there is no such series, there is actually no time.
Such a view is considerably more interesting since the case for temporal error
theory is much stronger if one supposes that actually there is no B-series. For
one might reasonably hold that if there are no relations of earlier and later than
or simultaneous with, then truly there are no temporal relations and hence no
time.

We can see just how this way of understanding Barbour provides some
insight into his claim that our world is a timeless one. Consider again the
distinction made above between Machian timelessness and QPD timelessness.
As noted above, Barbours Machian formulation of general relativity contains a
detailed algorithm for reconstructing four dimensional spacetime from the
three dimensional relative configurations. This reconstruction of spacetime
yields a reconstruction of a temporal ordering, and thus a B-series. If we read
Barbours claim of timelessness as error theoretic about the B-series, Machian
timelessness simply does not fit this bill.
Having said this, QPD timelessness provides a more conducive timeless
structure: there is no temporal ordering whatsoever in Barbours
interpretation of quantum gravity, reconstructed or not. The appearance of the
present configuration having evolved in time is merely an illusion brought
about by the mutually consistent records we find in each time capsule. Thus it
seems as though all that exists are the set of instants that neither bear B-type
relations to one another, nor instantiate A-type monadic properties. If that
were right, then in some sense Barbour would hold a similar view to
McTaggart. According to McTaggart, neither the A-series nor the B-series
exists. All that exists is what he calls the C-series, which is not a temporal series
per se, since the C-series is a series of times that are not related by the
temporal relations of earlier than, later than and simultaneous with, nor are
they related by the A-theoretic determinations of pastness, presentness and
futurity. Indeed, both McTaggart and Barbour would agree that if there is no B-
series, no objective temporal ordering whatsoever, then plausibly there is no
time. So if Barbour implicitly takes the B-series to be essential to time, and
since his interpretation of quantum gravity entails that there is no such series,
then we think his error theoretic conclusion is warranted with respect to QPD
timelessness.
As mentioned at the beginning of this section, the parallel between the
argument in section 3.1 concerning the temporal structure of physical theory
and the metaphysical argument presented here suggests a close
correspondence between the two. In both analyses we find that in Barbours
Machian formulation of general relativity certain essential features of time
remain as integral parts of the theory: a linear and metric temporal structure,
on the one hand, and a B-series, on the other. Put like this, however, it should
now be clear why there is a close correspondence between these two
arguments: the temporal structure that the B-series provides just is a linear and

metric temporal structure, and vice versa.
7
Thus, one might say, the above
metaphysical analysis is a terminological variant of the argument developed in
section 3.1. It should come as no surprise then that via both routes for
characterising the essential features of time we find Barbours error theoretic
conclusion dubious with respect to his Machian formulation of general
relativity but warranted with respect to his interpretation of canonical quantum
gravity.

4. TIME CAPSULES, TEMPORAL EXPERIENCE AND SOLIPSISM
So far we have considered the extent to which Barbours claim of timelessness
is compelling. Suppose, however, we accept that Barbour gives us some reason
to endorse an error theory of time. In this section we wish to examine the
position in which this leaves us. In particular, focusing on Barbours
interpretation of canonical quantum gravity, we consider two issues. First, we
will assess Barbours attempt to provide an account of why it is that we
experience motion and change given that nothing moves and nothing changes.
And, second, we will then consider a sceptical worry with Barbours view
concerning the ontological status of points in the relative configuration space.

4.1. ON THE EXPERIENCE OF MOTION AND CHANGE
There are really two components to Barbours metaphysical view of time as it
pertains to his interpretation of canonical quantum gravity. The first
component consists in the claim that time is unreal. It is this feature of the view
that has been the focus of the discussion thus far. The second component is the
claim that change and motion are unreal. Arguably, this second claim falls out
of the first: if there is no time, then nothing can change and nothing can move.
This is because, we suppose, it is something of a conceptual truth about change
and motion that they require time.
This second feature of Barbours view is startling. In experience it seems to
us that things change and that things move. However, Barbours view entails
that these experiences are systematic illusions: although we experience motion

7
Likewise, the temporal structure that the A-series provides might also be characterised in terms
of Rovellis attributes of time: linearity, metricity, globality, externality, uniqueness, directionality and
presentness.

and change, these experiences are non-veridical, on a par with hallucinations.
But as with any error theory this view incurs an explanatory cost: Barbour must
provide us with some explanation as to why it is that we nevertheless seem to
experience motion and change even though there is no motion or change in the
world. Barbour recognises this explanatory burden and attempts to provide an
account of such experiences via the central notion of a time capsule.
According to Barbour, time capsules are points in the configuration space
that possess a rich structure of what we would describe as records: fossils,
memories, geological phenomena and historical accounts of the past. Of
course, for Barbour these are not, strictly speaking, records of the past since
there is no sense in which there is an objective past from the perspective of any
point in the relative configuration space. As indicated, these records taken
together merely constitute a mutually consistent story. This story, however, is a
fiction and should not be taken to represent how things actually were since,
again, there is no way things actually were.
Now, for Barbour, our experience of change and motion is due to a specific
kind of structure found in a time capsule: memories (or perhaps apparent
memories, since they are only genuine memories if the events they are
memories of did in fact happen in the past). Consider a film. There is an
obvious sense in which the experience of motion and change when watching a
film is an illusion. Films consist of a series of static frames that are replaced one
after another at a certain speed such that it seems to us as though each frame is
moving smoothly into the next. Barbours explanation of why we experience
motion or change proceeds along similar lines. Our experiences are the result
of a huge series of experiential stills, namely apparent memories, which are
interpreted by our brains as the experience of motion and change.
It is unclear to us the extent to which one should think that Barbour offers a
reasonable account of the appearance of motion and change within a timeless
world. Here is a very quick argument that one might mount in response to
Barbours account of the experience of motion and change: everyone admits
that we experience the phenomena of change; but you might think that in order
to experience change, our experiences themselves must change (or,
alternatively, our brains must undergo change as part of the process of
interpretation); but either of those can be the case only if there is change in the
world; so if we experience change then there is actually change.
Whether or not this argument is sound is contentious. In particular, it
relies on claims about the nature of experience that we are not in a position to

defend. Nevertheless, it is worth noting that if the argument is sound then this
provides some evidence against the metaphysical claim that actually there is no
time. This can be shown by way of the following further argument: it is obvious
that we experience change in the actual world, the experience of change
presupposes the existence of change and motion, in order for there to be
change and motion in the world there must be time, therefore we do not live in
a timeless world.
Even so, without this metaphysical claim, do we have good reason to give up
on Barbours particular physical theories? Well, it depends upon which arm of
Barbours interpretive project we are interested in. If we are focusing on his
Machian formulation of general relativity then it is not clear that it does since,
as indicated, there is no reason to think that this view is error theoretic about
time. If, on the other hand, we are interested in his interpretation of quantum
gravity, then since the non-existence of time is entailed by the theory, the
above argument may give us some reason to doubt this interpretation.
It is worth noting that if Barbour does find the above argument worrying,
then he has at least two options available. First, he might defend his account of
motion and change by appealing more directly to the philosophy of temporal
consciousness where a view along these lines has long been defended by a
number of philosophers.
8
Second, he might prefer instead to reject the
conceptual entailment between the existence of motion and change in the
world and the non-existence of time. That is, he might argue that although time
does not exist, there nevertheless is motion and change in some sense. Note
that although strange, this manoeuvre resembles the sort of conceptual shift
that occurred with regard to the A- and B-theories of time. According to
McTaggart there could be no change or motion in a world in which there was
no A-series. Subsequent to this, however, philosophers have found
McTaggarts view to be implausible, maintaining instead that the B-series is
sufficient for the existence of change, so-called Cambridge change. Perhaps
then what Barbours view shows us is that a similar conceptual shift is in the
offing: contrary to what we might have thought, the B-series is not, in fact,
necessary for the existence of change.

8
This view is traditionally attributed to Husserl given in a series of lectures between 1893 and
1917, but has been more recently discussed by Barry Dainton 2000. See also Phillips 2008, Noe
2006 and Kelly 2005 for further discussion.

4.2. ON SOLIPSISM
Another reason to find Barbours view puzzling and perhaps worrisome is the
concern that it might lead to temporal solipsism: the view that only a single
moment exists, or will ever exist. Strictly speaking, Barbours view is not a form
of temporal solipsism. According to Barbour, the relative configuration space
that is constitutive of the metaphysical structure of the universe is such that
each and every point in the configuration space exists. A temporal solipsist
formulation of Barbours view would be the claim that only one point in the
relative configuration space exists (or will ever exist), a claim that Barbour
denies.
If Barbours view is not a form of temporal solipsism, then what threat does
temporal solipsism pose for his metaphysics? The trouble is that on Barbours
view we are epistemically locked within a single point in the configuration
space. This is because the only epistemic access we have to the world is via
empirical evidence, which is all encoded within a single time capsule.
If this is the case, then it seems that we have no warrant for believing that
the other points in the configuration space exist. Consider a relative
configuration space Q
0
in which only a single point t exists. Suppose that t is a
time capsule and thus has a rich structure of apparent memories, historical
accounts and so on. Finally, suppose that at t there exists at least one observer
O. Now consider a relative configuration space Q
0
* such that every point in
Q
0
*

exists. Suppose that in Q
0
* there is a point in the space t* such that t* is
just as rich as t with regard to mutually consistent records and t* is the home of
at least one observer O*. By Barbours lights, O experiences the world in
exactly the same way that O* experiences the world. Moreover, both O and O*
have precisely the same sort of empirical evidence available to them. As a
result, O* could have no a posteriori reason for thinking that she lives in a
Q
0
*-type configuration space rather than a Q
0
-type configuration space.
Could then O* have some a priori basis for thinking that she lives in a Q
0
*-
type configuration space? Well, there are two options here. First, Barbour
might provide some reason for thinking that the points in a relative
configuration space are somehow mutually dependent in an ontological sense,
such that if one point in the configuration space exists, then all points in the
configuration space exist and vice versa. The idea then would be that a Q
0
-type
configuration space is simply incoherent, given a correct specification of the
metaphysical constraints on configuration spaces in general. The upshot would
be that O* could, conceivably, have a reason for thinking that she lived in a

Q
0
*-type configuration space, since there are only Q
0
*-type configuration
spaces. Unfortunately, there is nothing in Barbours view (at least as set out
above) that would warrant this metaphysical claim. Moreover, more generally,
it is hard to see why the points in a relative configuration space like Barbours
need be ontologically dependent in the manner needed to avoid the worry:
ontologically sparse relative configuration spaces just do not seem to us to be a
priori incoherent.
This brings us to the second way in which Barbour might provide O* with
some a priori reason for thinking that she lives in a Q
0
*-type configuration
space. This option involves appealing to theoretical simplicity. The idea here is
that a theory that entails the existence of a Q
0
*-type configuration space is
theoretically simpler than a theory that entails the existence of a Q
0
-type
configuration space, since all of the points in the configuration space are on a
par, ontologically speaking. The trouble with this, however, is that the appeal
to parsimony cuts both ways. This is because a Q
0
-type configuration space is,
in fact, more parsimonious than a Q
0
*-type configuration space when it comes
to relative ontological economy. This is because a Q
0
-type configuration space
is committed to a single instant only, whilst a Q
0
*-type configuration space
boasts a lavish ontology of instants. As a consequence, it seems that the two
views have different theoretic virtues when it comes to parsimony. However, it
is very hard to see why one form of parsimony ought to be preferred over
another and, as a consequence, there seems to be no way that an appeal to
symmetry principles of this kind can provide any a priori reason for thinking
that O* lives in a Q
0
*-type configuration space rather than a Q
0
-type space.
What this shows us then is that Barbours view entails a sort of scepticism
about whether or not we live in a world in which temporal solipsism is true.
This is because there seems to be no a posteriori or a priori way for one to
determine whether one lives in a Q
0
*-type configuration space or a Q
0
-type
configuration space. But if this is correct, then it would seem that one can
never be warranted in believing in the existence of a point in the relative
configuration space other than the point at which one is located.
The crucial issue here then is just how serious a threat this sceptical
scenario poses for Barbour. That is, would it really matter if we were warranted
only in believing in the existence of a single point in the relative configuration
space? It seems to us that although temporal solipsism is a worrying doctrine, it
does not carry any negative implications that are not already contained within
Barbours view. In order to see why, it is useful to consider a more standard

formulation of temporal solipsism that makes no mention of configuration
spaces. The easiest way to get a handle on this form of temporal solipsism is to
consider a view commonly referred to in the metaphysics of time as presentism.
According to presentists, only a single time exists. However, presentists also
believe that the present moves via an ongoing process of the ex nihilo coming
into, and going out of existence of time slices. Thus, although only one time is
ever in existence it remains the case that there were other times that existed
and there will be other times that will exist in the future.
Temporal solipsism of the garden variety can be arrived at by taking
presentism and stripping away the claim that time flows. Thus, on this view,
only a single time, the present, exists and it is not the case that other times
existed, and it is not the case that there will exist other times in the future. All
there is, was and ever will be is a single, unchanging instant. Temporal
solipsism of this form is unattractive for at least four reasons. First, it goes
against the common sense view of time. This is because presentism is usually
taken to be the common sense view of time, and temporal solipsism entails that
presentism as traditionally conceived is false. Second, temporal solipsism
entails that all of our apparent memories are false memories, all of our
historical records are false records and all of our fossil evidence is not evidence
at all. This is because there never was a past. Third, according to temporal
solipsism, we do not persist through time. The most worrying consequence of
this feature of the view is that it turns out to be irrational to anticipate
experiencing what seem to us to be future events, or to regret those things that
we take ourselves to have done in the past. This is because there is no sense in
which we will be located at the future times in question to undergo the relevant
experiences, or that we were located at the past times in question, doing the
relevant deeds. Fourth, one might maintain that if temporal solipsism is true
then nothing moves or changes. Aside from the cost to common sense, the
temporal solipsist owes us an explanation of why it is that we experience
motion and change, and it seems that any such explanation must appeal to
certain instantaneous structures that exist within the one existing time (such as
memories).
But if this is why temporal solipsism is taken to be so worrying then
temporal solipsism does not in fact render Barbours view any more
unattractive than it already is. This is because Barbours view carries the same
commitments as garden variety temporal solipsism. That is, Barbours view
entails that presentism is false, that our memories are only apparent memories,

that we are ontologically (and, indeed, epistemically) locked to a single time
qua point in the relative configuration space and thus that we do not persist
through time and, finally, that the experience of motion and change is an
illusion that must be accounted for using instantaneous structures, i.e.,
apparent memories. Thus, it is not clear to us that it matters for Barbours view
all that much if only a single time in the relative configuration space exists. This
is because the move to temporal solipsism carries with it no added
disadvantages. Of course, the upshot is that Barbours view turns out to be only
as plausible as temporal solipsism of the garden variety, since the two views
have similar commitments. But we take it that Barbour would simply bite the
bullet on any such metaphysically peculiar consequences given his conviction
that his view is the best way to understand the physical structure of the world.

5. CONCLUSION
We do not claim to have given the definitive account of the metaphysical
consequences of Barbours timeless view. On the contrary, all we hope to have
achieved is the beginnings of an exploration into the intriguing picture of the
fundamental structure of our reality that Barbour presents. We warn, though,
that one should proceed with caution. The lesson of this examination is that,
due to the multifaceted nature of Barbours view, his claim of timelessness is
not one size fits all. While the timelessness of Barbours interpretation of
quantum gravity seems to hold up to scrutiny, the timelessness of Barbours
classical theory is somewhat questionable. Nevertheless, if there is any decent
chance that Barbours interpretation of quantum gravity is correct, it is
startling indeed to realise that this would entail that ours is a timeless world,
and that alone makes consideration of Barbours views important. For it is hard
to imagine a greater change in how we view ourselves and our world than the
discovery that all that we thought had gone before us, our memories, our
accomplishments and our regrets are all illusory, and that our dreams for the
future, our plans, our decisions and our choices are in some good sense
pointless. Our self conception as agents who are extended in time and whose
choices today in part create who we will be tomorrow would be radically
undermined by discovering that ours is a Barbourian world, and that would
require a radical rethink of ourselves as ethical and prudential agents, a task
that we leave for another time.

REFERENCES
Barbour, J. (1994a). The Timelessness of Quantum Gravity: I. The Evidence
from the Classical Theory. Classical Quantum Gravity, 11(12), 2853-
2873.
Barbour, J. (1994b). The Timelessness of Quantum Gravity: II. The
Appearance of Dynamics in Static Configurations. Classical Quantum
Gravity, 11(12), 2875-2897.
Barbour, J. (1999c). The End of Time. Oxford: Oxford University Press.
Butterfield, J. (2001). The End of Time?. ArXiv General Relativity and
Quantum Cosmology e-prints.
<http://arxiv.org/abs/gr-qc/0103055v1>
Dainton, B. (2000). Stream of Consciousness: Unity and Continuity in
Conscious Experience. London: Routledge.
Healey, R. (2002). Can Physics Coherently Deny the Reality of Time? In C.
Callender (Ed.), Time, Reality and Experience (pp. 293-316).
Husserl, E. (1980). On the Phenomenology of the Consciousness of Internal
Time. (tr. by J. Brough). Dordrecht: Kluwer Academic Publishers.
[1928]
Ismael, J. (2002). Rememberances, Mementos, and Time-Capsules. In C.
Kelly, S. (2005). The Puzzle of Temporal Experience. In A. Brook (Ed.),
Cognition and the Brain: The Philosophy and Neuroscience Movement.
Cambridge: Cambridge University Press
McTaggart, J. E. (1908). The Unreality of Time. Mind, 17(68), 457-474.
Noe, A. (2006). Experience of the World in Time. Analysis 66(298), 26-32.
Phillips, I. B. (2008). Perceiving Temporal Properties. European Journal of
Philosophy, 16(3). Published Online, DOI: 10.1111/j.1468-
0378.2008.00299.x.

Pooley, O. (2001). Relationism Rehabilitated? II: Relativity. Pittsburgh
Philosophy of Science Archive.
<http://philsci-archive.pitt.edu/archive/00000221/>
Rovelli, C. (1995). Analysis of the Distinct Meanings of the Notion of Time in
Different Physical Theories. Il Nuovo Cimento, 110 B(1), 81-93.
Rovelli, C. (2004). Quantum Gravity. Cambridge: Cambridge University
Press.
Metaphysics of Causation and Physics of General
Relativity
*

Vincent Lam**
v.lam@uq.edu.au

ABSTRACT
This paper aims to discuss two realist conceptions about causation in the light
of the general theory of relativity (GTR). I first consider the conserved quantity
of causation, which explicitly relies on the energy conservation principle. Such
principle is however problematic within GTR, mainly because of the dynamical
nature of the spacetime structure itself. I then turn to the causal theory of
properties, according to which (fundamental physical) properties are such that
insofar as they are certain qualities, they are powers to produce certain effects.
In order to be compatible with GTR, such theory has to assume non-trivial
global conditions on the spacetime structure; such assumptions seem to
deprive the singularist non-Humean feature of this theory of causation. The
question of the possible causal nature of spacetime (metrical) properties is
addressed in the conclusion.

1. INTRODUCTION
The question about the nature of causation is one of the most fundamental
questions in philosophy of nature. In an analytical approach, this paper aims to
discuss some aspects of this question in the light of one of our best
fundamental physical theories, the general theory of relativity (GTR). If the
possible link between causation, space and time is clearly not new (think about
the standard Humean conception for instance), GTR might put some novel

* This paper is a (very slightly modified) English version of Lam 2009. I would like to thank
Claudio Calosi for the invitation to contribute to this volume and the Swiss National Science
Foundation (113688) for partial financial support.
** University of Queensland
constraints on it (to the extent that it is relevant for such question). More
specifically, I will discuss the constraints exerted by GTR on two realist
conceptions of causation which have been widely discussed in the literature
recently.
I will first consider the conception of causation in terms of conserved
quantities, such as recently developed and discussed by Wesley Salmon (1998,
2002) and Phil Dowe (2000a) (section 2). This conception of causation
considers the fundamental causal relations as physical relations whose nature
relies on the physical principle of conservation of (mass-)energy. It seems clear
that such physicalist and mechanistic conception of causation must come to
terms with what our fundamental physical theories, such as GTR, say about our
actual world. So the fact that within GTR there is in the general case no
meaningful energy conservation (in a precise sense which is discussed below)
might have important consequences for the conserved quantity theory of
causation (section 3).
This conception of causation is a realist and non-Humean one in the sense
that it considers fundamental causal relations as objective (mind-independent)
and singularist (regularities-independent) relations in the physical world.
However, such empirical conception does not say anything about necessity or
about the irreducible causal feature of the relation between cause and effect
feature which is often hold as essential for any realist conception of causation.
1

Such aspect is taken into account within the causal conception of properties, or
conception in terms of dispositions or powers, according to which fundamental
physical properties are such that insofar as they are certain qualities, they are
powers to produce certain effects: these effects can so be understood as the
necessary consequence of the very nature of the considered properties (section
4). Within such realist conception of causation, developed and defended by
Rom Harr and Edward Madden (1975), Sydney Shoemaker (1980), Stephen
Mumford (1998) and Alexander Bird (2007) among others, the fundamental
physical properties then have an irreducible causal nature; causation is then
bound to the very nature of properties, reflecting its objective and fundamental
nature. It seems clear that the existence of such irreducible causal nature of
properties is not an empirical question that science and fundamental physics in
particular could settle: indeed, a world where the fundamental physical

1
See for instance Chakravartty 2005.
Vincent Lam Metaphysics of Causation and Physics of General Relativity 63
properties are categorical (non-causal) is not qualitatively distinct from a world
where the fundamental physical properties are irreducibly dispositional
(causal). This is a metaphysical distinction. However, insofar as we require that
a conception about causation has to be compatible with what fundamental
physics tells us, fundamental physics might provide relevant arguments in the
debate. In particular, I will discuss some arguments from GTR (section 5). It
should be made clear that this paper aims less to argue in favor or against these
two conceptions of causation than to provide an interesting case of fruitful
interaction between physics and metaphysics.

2. CONSERVED QUANTITY THEORY
In this section, I will briefly present the main elements of the theory of
conserved quantity, as presented by Dowe (the differences with Salmons
version are not relevant here). Within this theory, causation is understood in
terms of causal interactions and causal processes, which are themselves defined
in physical terms.
2
A causal interaction is an intersection of worldlines that
involves an exchange of a conserved quantity. A worldline is a (timelike or
null
3
) curve in spacetime that represents the history of an object. The notion of
exchange is understood here in a minimal way and only requires a change in the
value of the conserved quantity for at least one incoming and one outgoing
worldline involved in the interaction
4
and such that the corresponding
conservation law holds. A conserved quantity is any physical quantity that
satisfies a conservation law in any given spacetime region.
5
Our current best
physical theories give us indications of what these conserved quantities might
be; such conception of causation depends therefore strongly on what these
fundamental physical theories have to say. One generally considers (mass-)

2
I follow Dowe 2000a for most of the definitions of the conserved quantity theory.
3
A curve in spacetime (M, g
ab
) is said to be timelike or null if for all points p of the curve the
tangent vector t
a
(p) to the curve is timelike, that is, such that g
ab
(t
a
(p)t
b
(p))<0, or null, that is, such that
g
ab
(t
a
(p)t
b
(p))= 0, with (-, +, +, +) as the signature of g
ab
.
4
Incoming and outgoing are labels defined by the local light cone structure and are
interchangeable, see Dowe 2000a, p. 92.
5
As such, it does not need to be a law: what is required is only that the relevant conservation
statements hold in the actual physical world, independently of the question about their possible lawlike
nature (Dowe 2000a, p. 96; Salmon 1998, p. 259; Psillos 2002, p. 121). I will discuss below a
mathematical representation of such conservation statement.
energy as the relevant conserved quantity within the analysis of causation.
There are two main reasons for that. First, conservation of (mass-)energy is of
fundamental importance for contemporary physics and seems to be an
experimentally successful principle in our world. Second, there seems to be a
close connection between (mass-)energy and causation, which has already been
exploited by some philosophers in the past, such as Bertrand Russell and Hans
Reichenbach for instance, and which is at the basis of the conserved quantity
approach of causation
6
: the idea is roughly that two events are causally related
if and only if they are in a certain relation with respect to their (mass-)energy.
The second fundamental element of the conserved quantity theory is the
causal process, which is a worldline of an object that possesses a conserved
quantity. The possession of a conserved quantity by an object is to be
understood as the instantiation of this conserved quantity conceived as a
property. The notion of object is conceived here in a very broad sense. To the
extent that we are considering causation at the fundamental level, objects are
the fundamental elements that constitute the ontology, like particles or fields
for instance. The conserved quantity theory does not take position in the
traditional debates in the metaphysics of objects and properties.
7
The main
claim of the CQ theory is that there is a causal relation between two events if
and only if they are linked by a set of causal processes and causal interactions
such that any change of an object or any change of a conserved quantity occurs
at a causal interaction and the changes in the conserved quantities are governed
by the physical conservation laws.
8

3. CONSERVED QUANTITY THEORY AND PHYSICS OF GENERAL RELATIVITY
The notions of (total) energy of a physical system and of conservation of this
energy constitute a difficult and tricky topic in GTR. The difficulties come

6
See for instance Fair 1979 and Curiel 2000, .2.
7
However, this definition of a causal process requires that the object, whose world line is a causal
process, has identity through time, sometimes called genidentity, in particular in order to distinguish a
causal process from so-called pseudo-processes, such as a moving spot of light along a wall (Salmon
2002, p. 113-116).
8
An event is understood here as the instantiation of properties in a given space-time region
(ultimately a space-time point), the content of a space-time region (without any commitment to a
specific space-time ontology); in particular, a causal interaction is an event.
from the fundamental GTR property according to which the gravitational field
and the spacetime structure are described as one and the same physical
structure within GTR.
9
The spacetime structure is then fundamentally
dynamical and there is (in the general case) no non-dynamical background with
respect to which physical systems can be considered.
10
This property, often
called background independence, is taken by many physicists and
philosophers of physics to be the distinctive feature of GTR and one of the
main difficulties towards a coherent view between GTR and quantum field
theory (QFT), which relies on a non-dynamical spacetime background. As a
consequence of this dynamical nature, there is in general no natural family of
timelike curves representing observers all at rest with respect to each other;
therefore, within GTR, a given observer cannot in general define the energy of
a distant particle.
11
From a technical point of view, this comes in the general
case from the lack of the symmetries that are required for such family of curves
to exist (a general solution of the Einstein field equations does not possess
timelike Killing fields). Without such symmetries it is not possible to express
an energy conservation law in a given spacetime region: there is in general no
integral (mass-)energy conservation law within GTR.
12
It should be clear that

9
Spacetime is represented by an equivalence class of pairs (M, g
ab
), where the Lorentz metric
tensor field g
ab
encodes the inertio-gravitational effects as well as the fundamental spacetime relations.
Within GTR, spacetime and the gravitational field can be understood as a physical structure in the
precise sense of a network of physical relations (spacetime and gravitational relations) among relata
that do not possess any intrinsic identity (as made explicite by the invariance under active
diffeomorphisms); see Esfeld and Lam 2008, and Lam 2010a.
10
More precisely, the metric-gravitational field cannot be decomposed uniquely into an inertial
(non-dynamical) part plus a gravitational (dynamical) part.
11
See Wald 1984, p. 69.
12
Within GR, we have the differential energy conservation
a
T
ab
= 0, where T
ab
is the stress-
energy-momentum tensor representing the energy-momentum distribution and s the covariant
derivative associated with the metric. But such differential expression does not represent a meaningful
conservation statement valid on any (finite) spacetime region. Indeed, there is in general no unit
timelike vector field v
a
such that
a
v
b
= 0 or merely such that
a
v
b
+
b
v
a
= 0 (Killings equation): this
means that there is in general no family of observers all at rest with each other (with parallel 4-
velocities). Therefore, there is no meaningful (integral) energy-momentum conservation law in the
(integral) form of
S
J
a
n
a
dS = 0 where J
a
= - T
ab
v
b
is the (mass-)energy current density 4-vector
measured by the observers represented by v
a
, S is a 3-dimensional boundary of any 4-dimensional
space-time region and n
a
is the unit normal (like in the flat Minkowski space-time of special relativity:
see Wald 1984, pp. 69-70).
this fact is not an epistemological question but is a consequence of the very
nature of spacetime and of the gravitational field.
From the point of view of the universally interacting gravitational field, the
failure of integral (non-gravitational) energy conservation is an obvious
consequence of not taking into account gravitational energy: strictly speaking,
there cannot be non-gravitational energy-momentum conservation since any
physical system interacts with the gravitational field and its energy can
transform into gravitational energy and vice versa. So, it seems natural to think
that one would just need to take into account the energy of the gravitational
field in order to obtain an energy conservation law. Things are however a bit
more complicated. Indeed, gravitational energy cannot be represented by a
(unique) coordinate-free geometric object (that is, for instance, by a tensor
field). It is always possible at any spacetime point to find a coordinate system in
which (infinitesimally) there is no gravitational energy.
13
As a consequence,
gravitational energy can be understood as non-local in the precise sense that
the amount of gravitational energy in any given spacetime region cannot be
defined in a unambiguous way.
14
In the general case, there cannot be energy
conservation insofar as gravitational energy cannot be taken into account;
gravitational energy can be transformed into non-gravitational energy and can
therefore increase or decrease the amount of energy that is present in a given
spacetime region.
Although not much debated in philosophy of physics yet, this conclusion
and its consequences for the conserved quantity theory have been discussed by
Alexander Rueger (1998) and Erik Curiel (2000). Due to the lack of a
meaningful energy conservation law, the notions of exchange and of possession
of energy as well as the notion itself of energy as a conserved quantity become
problematic, so that the fundamental notions of causal interactions and of
causal processes are undermined. GR tells us that, in general, there can be no
genuine (mass-)energy conservation law and so no such conservation law can

13
This is often understood as a consequence of Einsteins equivalence principle, see for instance
Misner et al. 1973, p. 386; however, the meaning of the equivalence principle(s) is trickier, see
Norton 1993, .4.1.
14
Hoefer (2000) argues that this non-local feature of gravitational energy precisely shows that it
does not constitute a meaningful form of energy. Actually, there seems to be difficulties with the very
notions of energy and mass (gravitational or not) within GTR, see for instance Jaramillo and
Gourgoulhon 2010 as well as the discussion in Lam 2010b. These questions do not alter the point
that in general there is no meaningful energy conservation law.
rule the exchange of (mass-)energy as a conserved quantity in a causal
interaction. The characterization of causal for an interaction then loses its
fundamental meaning and one cannot say unambiguously whether two events
are causally related or not.
Insofar as the energy conservation constitutes a fundamental principle
within contemporary physics, it might still seem that the conserved quantity
theory, in grounding causation on this principle, does well capture a
fundamental aspect of the world. However, it seems that such a position would
amount to dismiss the moral of one of our two most fundamental physical
theories: the spacetime structure and the gravitational field are one and the
same physical entity, which possesses energy and momentum and which is
dynamically (and universally) related to the whole non-gravitational (mass-)
energy. From the GTR point of view, the relevance of the energy conservation
principle in most physical theories and in particular in quantum theory comes
from the fact that these theories require a non-dynamical spacetime structure
that possesses the high degree of symmetry needed to define an energy
conservation law. Indeed, within GTR, there exists a well-defined notion of
total energy for an isolated physical system, that is, a physical system in an
asymptotically flat spacetime (without entering into the details, spacetime
becomes Minkowskian very far from the considered system). Although very
useful and justified in many concrete cases, such idealizations cannot
constitute the empirical basis for any account of causation that pretends to be
fundamental.
Another possible position of the friend of the conserved quantity theory is
first to explicitly consider this conception of causation as contingent, for
instance on the existence of conservation laws in our actual world, and then to
maintain that our actual world does possess the required properties, namely
the high degree of spacetime symmetry necessary for conservation laws (in the
case of energy conservation, the spacetime structure has to be invariant under
temporal translations (such spacetime is called stationary). Such position is
endorsed by Dowe (2000a, 2000b), who considers the conserved quantity
theory as an empirical analysis of causation in our actual world in contrast to a
metaphysics of causation valid in all possible worlds; according to him, such
empirical analysis has to rely on the results of science this is precisely the
methodology here. The results from GTR applied to our actual world tend to
show that, contrary to what Dowe (2000a, p. 97; 2000b, p. 24) says, our
universe does not possess strictly speaking the required high degree of
symmetry (Friedmann-Lematre-Robertson-Walker solutions, which
constitute the standard model of contemporary cosmology, are highly
symmetric, in particular homogenous and isotropic, but it is clear that these
physically relevant solutions only constitute approximations of our actual
universe).
15

That the conserved quantity theory requires such specific conditions (from
a GTR point of view) seems to reflect the fact that this theory considers
causation among material objects against a fixed, non-dynamical
background.
16
But, as we have discussed above, GTR describes spacetime as a
fundamentally dynamical physical entity that possesses energy and momentum
and such that strictly speaking it is not possible to isolate a causal process
from causal interactions with the spacetime structure or gravitational field.
The conserved quantity theory provides a good example of the fruitful
relations between metaphysics of nature and physics: motivated by physical
considerations, this realist conception of causation is put into question by even
more fundamental physical considerations. In a similar way, any metaphysical
position about nature should be discussed in the light of what fundamental
physics tells us. It is clear that there is no direct implication between physics
and metaphysics of nature; however, physics can bring some relevant light on
certain metaphysical conceptions, such as in the case of the realist conceptions
of causation. In the same move, I now discuss the realist conception of
causation in terms of dispositions or powers.

4. CAUSAL THEORY OF PROPERTIES
Within the conserved quantity theory, causation is considered as an objective
feature of the world, but the more metaphysical question about the necessary
(or irreducibly causal) nature of the causal relation is not addressed. On the
one hand, despite their objective and singularist character, the causal
processes and interactions of Salmons and Dowes theory can be understood

15
Curiel (2000, pp. 47-48) underlines the fact that the slightest inhomogeneity breaks the
required symmetry and that nothing so particular would correspond to our actual world.
16
For instance, Curiel (2000, p. 47) underlines the fact that a timelike Killing field can be
understood as a kind of privileged fixed temporal background against which conserved quantities can
be considered.
as supervenient on a non-causal, Humean basis.
17
On the other hand, it seems
possible to consider properties within the conserved quantity theory as
irreducibly causal in the sense that their very nature is to produce certain
effects.
18
From an empiricist point of view, one may wonder what would be the
benefits of considering fundamental physical properties in terms of causal
powers whose irreducible causal nature seems out of reach of empirical
sciences (such consideration constitutes the crux of the Humean criticism).
Within the contemporary debate, a more or less explicit motivation among the
proponents of such conception comes from the way fundamental physics seems
to work: indeed, it seems that we only gain knowledge about a fundamental
physical property through the way it interacts, that is, through the causal
relations in which it stands. If this epistemic specificity suggests that
fundamental physical properties possess some dispositional nature, it does not
constitutes a powerful enough argument against some underlying categorical
basis for these properties. However, this epistemic specificity fits well with one
of the main (purely metaphysical) arguments in favor of the causal theory of
properties: insofar as we have access to the fundamental physical properties
only through the causal relations in which they stand, their possible categorical
nature is independent of their nomological and causal role. Such possible
categorical nature is therefore a primitive and inaccessible qualitative feature (a
quiddity) of the properties and forces us to some humility in the sense that
we cannot know what the properties are. From an empiricist point of view, it is
not very satisfying to accept such inaccessible qualitative nature of properties.
Two fundamental physical properties that are qualitatively distinct in virtue of
their distinct categorical nature can therefore stand in exactly the same causal
and nomological relations (corresponding to what we would consider to be the
charge for instance). Such metaphysical underdetermination vanishes if one
accepts a causal theory of properties according to which their nature is (the
power, the disposition) to produce certain effects. So, two distinct properties
in virtue of their very nature cannot produce exactly the same effects and
cannot stand in exactly the same causal relations. These latter are then a
consequence of the very nature of the relevant properties and have therefore an
objective and necessary character: insofar as the nature of a fundamental
physical property is to produce certain effects, these latter constitute the

17
See for instance Psillos 2002, .4.5.4.
18
Chakravartty (2005, .5) briefly discusses this possibility.
necessary consequence of the nature of the considered property. The
fundamental physical properties are then conceived as possessing some
irreducible dispositional essence or as irreducible causal powers, which do not
require any external triggering condition (it is the very nature of properties to
produce certain effects). The details and subtleties of the different versions of
the causal theory of properties in terms of causal powers or irreducible
dispositions are not relevant for the discussion here.
19
The main point is that
there is an important purely metaphysical argument in favor of the causal
theory of properties (although motivated by an empiricist stance, this
metaphysical argument is independent from the above mentioned epistemic
considerations), that is in favor of a strong causal realism, binding causation to
the very nature of fundamental physical properties.
20

As already mentioned, the main aim here is not to discuss this conception
as such. Rather, the idea is to consider the interactions between this
metaphysical conception about causation and the physics of GTR. More
specifically, I consider the consequences of the fundamental dynamical nature
of spacetime for the causal theory of properties. Some (non-trivial) global
constraints have to be imposed on the spacetime structure in order to secure
the coherence of the causal conception of properties. It then seems that an
important aspect of this conception is affected by these constraints: the
singularist character of the causal relation. If one considers two events A and B
(instantiations of fundamental physical properties for instance) as related by a
causal relation in the sense that the very nature of A is to produce the effect B,
then the causal nature of the relation as well as the very existence of the event B
rely entirely (in a local, singularist way) on the event A. On the contrary,
according to the non-singularist, Humean conception of causation, the causal
link between A and B depends on the regularities in the world between all
events of type A and all events of type B; actually, it may supervene on the
entire distribution of fundamental physical properties (as within David Lewis
thesis of Humean supervenience
21
).
22
Among the proponents of the causal

19
See for instance Shoemaker 1980, Mumford 1998, Bird 2007, Chakravartty 2007.
20
According to the proponents of the causal theory of properties, the main metaphysical
argument about quiddity is not the only one in favor of this conception; they also offer arguments
from quantum physics (entangled quantum systems possessing irreducible dispositions to dissolve the
entanglement for instance) and from GTR as well (spacetime properties as causal see section 6); see
recently Esfeld 2009 as well as references therein.
21
See for instance Lewis 1986, pp. ix-x.
theory of properties, this singularist aspect of causation is not controversial
and is considered as an advantage since only the nature of the relevant
properties have to be considered in order to account for a causal relation. But
the this singular character of the causal relation seems to be affected by the
global constraints on the spacetime structure that have to be assumed: it seems
that the causal nature of the relation does not only depend on the nature of the
property that is considered as the cause any more, but also on global properties
of the spacetime structure. Before discussing further this point, it is useful to
consider in some more details the constraints that the proponents of the causal
theory of properties have to assume.

5. CAUSAL THEORY OF PROPERTIES AND PHYSICS OF GENERAL RELATIVITY
One of the central points of GTR is that spacetime and the gravitational field
constitute one and the same dynamical physical structure, which possesses
(gravitational) energy and which universally interacts with gravitational and
non-gravitational (mass-)energy interaction that is encoded in the Einstein
field equations. As a consequence of the dynamical nature of the spacetime
structure, the spacetime topology can be non-trivial. Within special relativity,
the inert Minkowski metric defines for all spacetime points a future and past
light cone (future and past are here interchangeable labels, modulo a certain
consistency constraint called time orientability). The future light cone of any
spacetime point p determines the set of events that can be causally influenced
by an event at p, that is for instance the set of spacetime points that can be
reached by physical particle starting at p and travelling at a speed less or equal
the speed of light (and similarly for the past light cone). This light cone
structure exists only locally within GTR, that is, it is only valid for some

22
It is possible that the relata of the causal relations may not be local and pointlike entities; for
instance, within the framework of ontic structural realism, it has recently been suggested that the
physical structures (defined in the precise sense of a network of concrete physical relations among
concrete physical relata) are themselves causal (French 2006, Esfeld 2009); however, how structures
and relations can possess causal powers remains obscure (French 2006, .VI). In any case, the central
feature of the singularist aspect of this conception of causation is that the cause (for instance the state
of the world at a certain time) necessarily produces the effect (for instance the state of the world at a
certain later time this example is discussed in section 5) in virtue of its very nature and
independently of the rest of the fundamental physical properties.
neighborhood (called normal) of any spacetime point p; things can be
radically different at the global level. Indeed, in a dynamical (non-flat)
spacetime, the set I
+
(p) of events that can be reached by a physical particle
starting at p and following a (future oriented) timelike curve does not coincide
with the interior of the future light cone at p (which is actually only defined
locally for some neighborhood of p strictly speaking, the light cone is on the
tangent space at p). In particular, it is perfectly possible that pI
+
(p), which
means that there exists a future oriented closed timelike curve through p. In
such spacetime and within the framework of a realist conception of causation,
an event can therefore causally influence events in its past and can causally
interacts with itself; this latter aspect might be the most problematic one, in
particular within the framework of the causal theory of properties. It seems
indeed difficult to maintain that the nature of an instantiation of a property is to
produce itself. Two attitudes are now possible. First, to accept the causally
pathological behaviors as physically possible (in our actual world). Such
attitude is motivated by the fact that many solutions to the Einstein field
equations (among which some can apply to our actual world) possess closed
timelike curves (for instance, the Kerr-Newman solution, which describes
spacetime around a rotating charged mass, can contain such curves). In such
cases, it seems that the causal theory of properties as presented in the last
section is not compatible with GTR. Insofar as GTR provides an experimentally
successful description of the world, this consequence constitutes an important
drawback for the causal theory of properties. A second attitude is to dismiss
solutions containing closed timelike curves as physically non relevant. Such
attitude can be justified by the fact that many solutions of the Einstein field
equations that contain closed timelike curves are artificial and do not
correspond to physical situations in our actual world (for instance it is in
general considered that the famous Gdel solution does not correspond to any
region of our universe; in the above given example, the Kerr-Newman metric
more specifically describing spacetime outside a charged rotating black hole
does not contain closed timelike curves; moreover, global considerations on
certain physical laws, such as Maxwell equations, can be invoked to exclude the
existence of closed timelike curves). Such attitude is however controversial
among philosophers of science as well as among physicists.
23
I will however

23
See for example the discussion in Earman 1995, .6.4 for instance, the Kerr-Newman

adopt this attitude here in order to provide a physical framework in which the
causal theory of properties can be discussed. It is therefore possible to impose
a global constraint that excludes the solutions containing closed timelike
curves (chronology condition). It is straightforward that such condition is not
sufficient and that it is also necessary for our purpose here to exclude closed
null curves as well (simple causality condition). It seems that this latter global
condition provides the causal theory of properties with a safe (causally well-
behaved) environment. But this is far from certain. Indeed, the simple
causality condition does not exclude timelike curves that are almost closed in
the sense that for some (possibly infinitesimal) neighborhood U of a spacetime
point p, a timelke curve starting at p can cross U more than once. Such almost
closed curves can be excluded by the strong causality condition. Moreover, it
is likely (but not necessary) that the proponent of some strong causal realism
may want to derive some temporal order out of the more fundamental causal
order (this move allows her to ground the notion of change in the notion of
causal production). The existence of a global time function
24
seems mandatory
for such derivation to be possible (and maybe for the very notions of change
and temporal evolution too).
25
The existence of such function is guaranteed by
a stronger condition, called the stable causality condition. Indeed, there
exists a whole hierarchy of stronger and stronger causality conditions (a given
condition implying the weaker ones) that one can impose on the Einstein field
equations solutions. The strongest is the global hyperbolicity condition,
according to which spacetime has the topology of a product R x , where R

solution that contains closed timelike curves can be understood as representing the spacetime
generated by the gravitational collapse of a rotating star that does not lead to a black hole physical
possibility which cannot a priori be excluded)
24
A global time function is a function t from the manifold M to the reals such that t(p)<t(q) for all
p, qM that are linked by (future oriented) timelike curve from p to q. In general such function is not
unique. It should be clear that the existence of such functions does not imply any notion of objective
temporal distance between two events.
25
The notions of temporal evolution and change are notoriously problematic within GTR
(problem of time), in particular because of the (gauge-theoretic) diffeomorphism invariance of the
theory; accordingly, the definition of meaningful observables (in the sense of Dirac or Bergmann) and
the characterization of their evolution are very difficult tasks within GTR. It is however possible to
define gauge-invariant correlational observables (or coincidence observables as Earman 2002 dubs
them with respect to Einsteins intuitions), such as Komar events, together with a meaningful notion
of (not necessarily temporal) evolution; the Global Positioning System (GPS) can be considered as a
technological implementation of such correlational observables (see Rovelli 2004, ch.2; about the
problem of time, see Rickles 2006 and references therein).
represents the real line and a 3-dimensional hypersurface, called Cauchy
surface, such that no (inextendible) timelike or null curve crosses more than
once (in some sense, represents the universe at a certain time).
26
Such
condition provides the possibility very attractive for the proponent of the
causal theory of properties of an initial value formulation of the theory:
without entering into the details, the set of initial data on a Cauchy surface
completely determines (in a precise sense) the spatio-temporal (gravitational)
and material structures on the rest of the spacetime. The causal realist can then
be tempted to consider the relevant properties on the initial Cauchy surface as
causally generating (producing) the entire (physical state of the) universe.
27

The aim here is not to discuss to what extent the causal theory of properties
requires this whole hierarchy of global conditions imposed on the spacetime
structure. It is sufficient here to highlight the fact that such metaphysical
conception about causation requires to impose (at least) some of these
conditions; the simple causality condition has to be at least required for the
conception of causation as production to be make sense.
28
As already
discussed, it might be the case that further conditions from the hierarchy might
be needed for such conception about causation to be plausible. The important
point is that these conditions are global constraints on the spacetime structure.
Without entering into the technical details, it seems easy to see for instance
that the global hyperbolicity condition is a global condition in the sense that it
constraints the global topology of the (whole) spacetime structure to be
equivalent to R x .
As a consequence, it seems that within this framework the causal relations
depend on global spacetime properties (such as the global topology). As
mentioned at the end of the last section, this fact seems to affect the singularist
aspect of causation within the causal theory of properties. Considering again
the example of the last section, it seems that the causal relation between the
event A and the event B does not only depend on the nature of the cause A

26
The spacetime foliation into Cauchy surfaces is not unique and no foliation is privileged. This
fact constitutes one of the simplest formulations of the famous problem of time within GTR as well as
within the various attempts to quantize the theory.
27
For instance, Maudlin (2007, ch. 6) defends such position, where the laws of nature,
considered as primitive, rather than causation, underlie the notion of production.
28
Maudlin (2007, ch. 6) underlines very clearly the fact that according to him the existence of
closed timelike curves is incompatible with the notion of ontological production at the fundamental
level.
(which is to produce B within the causal, dispositionalist conception of
properties), but also on global properties of the whole spacetime (gravitational)
structure. The nature of the causal relation between A and B, as well as the
existence of B itself, cannot be ontologically grounded uniquely in the nature
of A. Strictly speaking, the fact that A causes B cannot be uniquely in virtue of
the nature of A, but also in virtue of the global spacetime (gravitational)
structure. For instance, let us assume that A and B are located within a region
U of the spacetime (M, g
ab
), which satisfies the necessary global conditions
such that it is meaningful to say that A causes B according to the causal theory
of properties. Let us further consider a distinct spacetime (M', g'
ab
), which
does not satisfy these necessary global conditions but which possesses a region
U' that is exactly similar (in a precise sense) to U (with the events A', B'
located in U' exactly similar to A, B in U). Now it seems that if A causes B in
(M, g
ab
) (by hypothesis), A' does not cause B' in (M', g'
ab
) because for instance
they might be related by closed timelike curve outside the region U'. So it
seems that the nature of the causal relation between A and B depends not only
on the nature of A, but also on the global properties of the spacetime structure
in which they are.
29

The reply from the proponent of the causal theory of properties is clear: she
cannot consider (M', g'
ab
) as metaphysically possible. It is the nature of
properties to produce effects that globally satisfy the above discussed
conditions; the global constraints are in some sense encoded in the very nature
of properties and the singularist aspect of the causal relation remains
unthreatened. If this reply may seem unsatisfying, there does not seem to be
any definitive objection against such attitude. Besides the question of the
relevance of this reply to the challenge of non-trivial topologies, the important
point here is that this metaphysical conception about causation strongly
constraints the solutions of the Einstein field equations. Indeed, it should be
clear that the above discussed hierarchy of causality conditions is not a
consequence of the theory; it is in general justified by some particular
conception about causation. As already mentioned, the lack of certain global
properties of the spacetime structure a purely empirical question would put
the causal theory of properties into a difficult situation.

29
A and B need not be local and pointlike for the argument; for instance, they can also represent
states of the universe at some given time.
6. CONCLUSION AND PERSPECTIVES
The main aim of this paper is not to provide a definitive argument against any
specific metaphysical conception about causation. However, within the here
adopted analytical approach, any metaphysical conception about nature can be
strengthened or weakened by arguments from contemporary physics. So we
have first considered the conserved quantity theory of causation; through its
use of the (energy) conservation principle, such conception clearly relies on
physics. However, we have seen that this it has to face certain difficulties from
fundamental features of the spacetime (and gravitational) structure as
described by GTR: the dynamical nature of the gravitational-spacetime
structure, the lack of non-dynamical, background physical structure with
respect to which physical entities as well as their (possibly causal) interactions
can be considered (background independence). Indeed, the symmetries that
are required by the conserved quantity theory constitute some fixed and
privileged temporal structure with respect to which the conserved quantities
can be considered. Strictly speaking, and contrary to what Dowe (2000b)
argues, our actual world most probably does not possess such symmetries,
even if these latter allow us to elaborate physically relevant and useful
approximations, which are justified in many practical cases (see section 3).
Another aspect of the problem is that, strictly speaking, one cannot isolate any
physical system from the interactions with the gravitational field (even if, again,
in many physically useful and justified approximations one does consider
gravitationally isolated systems). Moreover, the non-local nature of
gravitational energy (in the precise sense discussed in section 3) prevents the
proponent of the conserved quantity theory to consider any gravitational
interaction as a causal interaction (a gravitational wave destroying a rock would
not be considered as a causal interaction within this account).
In a certain way, the case of the causal theory of properties is more
complicated. We have seen that it requires that the world instantiates certain
non-trivial global properties and it considers certain (many indeed) solutions of
the Einstein field equations as metaphysically impossible. This conception can
therefore clearly be challenged by the lack of such properties in our world.
Insofar as it is a purely empirical question that remains debatable, this fact itself
does not constitute a definitive objection against the causal theory of
properties (section 5).
One can wonder how this conception accounts for the above mentioned
fundamental aspects of spacetime and gravitation that cause so much trouble to
the conserved quantity theory (background independence, non-local nature
of gravitational energy). This question belongs to the actual debate about this
conception of causation and properties. Another aspect of the same question is
about whether the causal theory of properties applies to spacetime
(gravitational) properties. At first sight, it seems that the dynamical nature of
the spacetime and gravitational structure within GTR provides an explicit
argument in favor of a causal interpretation of the spacetime properties. The
argument is that metrical properties within GTR (represented by the metric
tensor field and its functionals such as the Riemann tensor field) can be
considered as causal in the sense that they have a causal role that is manifested
by the action of tidal forces.
30, 31
So the nature of the metrical (gravitational)
properties is to produce tidal forces, which can be experienced by test particles
for instance, via spacetime curvature. The presence of non-gravitational (mass-)
energy is not necessary for such understanding, so that the causal
understanding of metrical properties remains valid in the purely gravitational
cases.
If such causal understanding of spacetime properties seems attractive at
first sight, a closer look shows that it is not free of difficulties. According to this
conception, the spacetime (gravitational) structure causally acts on non-
gravitational (mass-)energy (via curvature and tidal forces) as well as on
gravitational energy; this latter fact seems to imply that the spacetime
(gravitational) structure causally interacts with itself. Moreover, the Einstein
field equations do not privilege any direction in the interaction between the
spacetime (gravitational) structure and non-gravitational (mass-)energy; if one
understands one side of the Einstein field equations as causally acting on the
other (matter tells space how to curve), then it seems that the inverse is
equally true (space tells matter how to move). The conception of causation as
production of certain effects is far from being clear in this context. These
questions actually deserve to be discussed in details in a separate paper.

30
The argument is mainly defended by Bartels (1996, 2009) as well as Bird (2009); for a critical
point of view, see Livianos 2008.
31
Baker (2005) argues that the cosmological constant within GTR also provides an argument
in favor of a causal understanding of spacetime.
REFERENCES
Baker, D. (2004). Spacetime Substantivalism and Einsteins Cosmological
Constant. Philosophy of Science, 72(5), 1299-1311.
Bartels, A. (1996). Modern Essentialism and the Problem of Individuation of
Spacetime Points. Erkenntnis, 45(1), 25-43.
Bartels, A. (2009). Dispositionen in Raumzeit-Theorien. In C. F. Gethmann
(Ed.), Lebenswelt und Wissenschaft. XXI. Deutscher Kongress fr
Philosophie, Kolloquien. Hamburg: Meiner.
Bird, A. (2007). Natures Metaphysics. Laws and Properties. Oxford: Oxford
University Press.
Bird, A. (2009). Structural Properties Revisited. In T. Handfield (Ed.),
Dispositions and Causes. Oxford: Oxford University Press.
Chakravartty, A. (2005). Causal Realism: Events and Processes. Erkenntnis,
63(1), 7-31.
Chakravartty, A. (2007). A Metaphysics for Scientific Realism: Knowing the
Unobservable. Cambridge: Cambridge University Press.
Curiel, E. (2000). The Constraints General Relativity Places on Physicalist
Accounts of Causality. Theoria, 15(37), 33-58.
Dowe, P. (2000a). Physical Causation. Cambridge: Cambridge University
Press.
Dowe, P. (2000b). The Conserved Quantity Theory Defended. Theoria,
15(37), 11-31.
Earman, J. (1995). Bangs, Crunches, Whimpers, and Shrieks: Singularities
and Acausalities in Relativistic Spacetimes. Oxford: Oxford University
Press.
Earman, J. (2002). Thoroughly Modern McTaggart: Or, What McTaggart
Would Have Said if He Had Read the General Theory of Relativity.
Philosophers Imprint, 2(3), 1-28.
Esfeld, M. (2009). The Modal Nature of Structures in Ontic Structural
Realism. International Studies in the Philosophy of Science, 23(2),
179-194.
Esfeld, M, & Lam, V. (2008). Moderate Structural Realism about Space-time.
Synthese, 160(1), 2746.
Fair, D. (1979). Causation and the Flow of Energy. Erkenntnis, 14(3), 219-
250.
French, S. (2006). Structure as a Weapon of the Realist. Proceedings of the
Aristotelian Society, 106(1), 170-187.
Hoefer, C. (2000). Energy Conservation in GTR. Studies in History and
Philosophy of Modern Physics, 31(2), 187199.
Jaramillo, J. L., & Gourgoulhon, E. (2010). Mass and Angular Momentum in
General Relativity. In L. Blanchet, A. Spallicci & B. Whiting (Eds.),
Mass and Motion in General Relativity. Berlin: Springer.
Lam, V. (2009). Mtaphysique de la causalit et physique de la relativit
gnrale. Klesis, 13, 106-122.
Lam, V. (2010a). Aspects structuraux de lespace-temps dans la thorie de la
relativit gnrale. Forthcoming in S. Le Bihan (Ed.), Philosophie de la
Physique. Paris: Syllepse.
Lam, V. (2010b). Gravitational and Non-Gravitational Energy: The Need for
Background Structures, submitted.
Lewis, D. (1986). Philosophical Papers. Volume 2. Oxford: Oxford University
Press.
Livanios, V. (2008). Bird and the Dispositional Essentialist Account of
Spatiotemporal Relations. Journal for General Philosophy of Science,
39(2), 383-394.
Maudlin, T. (2007). The Metaphysics within Physics. Oxford: Oxford
University Press.
Misner, C. W., Thorne, K. S., & Wheeler, J. A. (1973). Gravitation. San
Francisco: W. H. Freeman.
Mumford, S. (1998). Dispositions. Oxford: Oxford University Press.
Norton, J. (1993). General Covariance and the Foundations of General
Relativity: Eight Decades of Dispute. Reports on Progress in Physics,
56(7), 791-858.
Psillos, S. (2002). Causation & Explanation. Chesham: Acumen.
Rickles, D. (2006). Time and Structure in Canonical Gravity. In D. Rickles, S.
French & J. Saatsi (Eds.), The Structural Foundations of Quantum
Gravity (pp. 152195). Oxford: Oxford University Press.
Rovelli, C. (2004). Quantum Gravity. Cambridge: Cambridge University
Press.
Rueger, A. (1998). Local Theories of Causation and the A Posteriori
Identification of the Causal Relation. Erenntnis, 48(1), 25-38.
Salmon, W. (1998). Causality and Explanation. Oxford: Oxford University
Press.
Salmon, W. (2002). A Realistic Account of Causation. In M. Marsonet (Ed.),
The Problem of Realism (pp. 106-134). Aldershot, UK: Ashgate
Publishing.
Shoemaker, S. (1980). Causality and Properties. In P. van Inwagen (Ed.),
Time and Cause (pp. 109-135). Dordrecht: Reidel.
Wald, R. (1984). General Relativity. Chicago: University of Chicago Press.

Realistic Aspects in the Standard Interpretation of
Quantum Mechanics
*

Claudio Garola**
garola@le.infn.it
Sandro Sozzo**

sozzo@le.infn.it

ABSTRACT
The belief that quantum mechanics (QM) does not admit a realistic
interpretation is widespread. According to some scholars concerned with the
foundations of QM all existing interpretations of this theory (except for the
statistical interpretation) presuppose instead a form of realism which consists
in assuming that QM deals with individual objects and their properties. We
uphold in the present paper that the arguments supporting the contextuality
and the nonlocality of QM are a significant clue to the implicit adoption of
stronger forms of realism (realism of theoretical entities and realism of
theories). If these kinds of realism are substituted by a simpler and more
intuitive semantic realism one can contrive a noncontextual and local
interpretation of the formalism of QM (SR interpretation). Moreover one can
provide a model for such an interpretation (ESR model) in which local realism
and QM do not conflict and some fundamental problems of the standard
interpretation of QM are avoided.

* The present paper was originally conceived as an English translation of a paper by C. Garola,
published in Italian in Giornale di Fisica, Vol. L, Suppl. 1, 2009. During the translation process the
paper has been entirely revised with the contribution of S. Sozzo and a final section has been added.
** Department of Physics and INFN (National Institute of Nuclear Physics) University of
Salento

I
In a recent paper T. Norsen (2007) listed four basic forms of realism, with the
aim of comparing them with the notion of realism implied by the term local
realism, widely used in the physical literature concerning Bells theorem and
the problems connected with it. Apart from the conclusions, from some aspects
disputable
1
, Norsens paper is interesting because it stimulates new reflections
on the different forms of realism introduced in physics and on their relations
with the notions of realism that one usually meets in the philosophical
literature. We briefly discuss some aspects of these problems in the present
paper by referring to one of the fundamental theories of modern physics,
namely quantum mechanics (QM). We intend to show that the prevailing
physicists attitude presupposes compelling philosophical choices,
notwithstanding the claimed antimetaphysical character of the orthodox
interpretation of the formal apparatus of QM
2
, and that these choices raise
some nontrivial problems.

II
The belief that QM does not admit a realistic interpretation is commonly
accepted among physicists, but the sense of this conviction is ambiguous if one
does not specify the form of realism to which one is referring. For example,
according to Busch et al. all existing interpretations of QM, but the statistical
interpretation, are realistic in the sense that they all agree that quantum
mechanics deals with individual objects and their properties (Busch et al.
1991, p. 5). Nevertheless such a form of realism is very weak, even though its
implications are not trivial. One can then wonder whether the standard
interpretation of QM presupposes more compelling forms of realism, and

1
According to Norsen the notion at issue does not fit in with any previous definition of realism
and should be avoided. But the term local realism has a precise meaning in the physical literature and
is widely used, hence we think that it should be preserved, even though it denotes a kind of realism that
does not fall within standard philosophical classifications.
2
Such an orthodox interpretation is often called standard, or Copenhagen, interpretation of QM.
This terminology is incorrect, both because the interpretation of a physical theory should be
considered as a part of the theory itself (see .II), and because there were different positions in the
Copenhagen school about the interpretation of the formalism of the new theory (Tassani 2004). For
the sake of brevity, however, we adopt the current terminology in the following.
Garola & Sozzo Realistic Aspects in the Standard Interpretation of QM 83

whether alternative interpretations exist that imply further different forms. An
exhaustive answer to this question would require a complex preliminary
analysis of the linguistic structure of scientific theories that cannot be
undertaken here. Therefore we limit ourselves to present an outline of the
epistemological perspective that we adopt in this paper (which is mainly based
on the received viewpoint.
3

According to our perspective, any scientific (e.g., physical) theory T is
stated by using a fragment of a natural language, enriched by conventional and
technical symbols (the metalanguage of T). This fragment contains in particular
a theoretical language L
T
and an observational (or pre-theoretical) language
L
O
. The former contains terms denoting theoretical entities and their relations
(in particular, L
T
contains the mathematical apparatus of T), and constitutes a
formal structure. The latter is interpreted on objects and events of the physical
(observable) world by means of assignment rules which provide an
interpretation of L
O
, usually in terms of operational definitions, which
constitutes the basis for any further interpretation, in particular of L
T
. L
O
is
thus provided with semantics and a notion of truth. Moreover a mapping exists
of L
O
onto a sublanguage L
TO
of L
T
, which implies that there are
correspondence rules (or bridge principles) that inversely relate L
TO
with L
O
,
supplying an empirical interpretation of L
T
(which is necessarily partial, for
L
TO
c L
T
, and indirect, for it concerns only derived and not primitive terms of
L
T
). The following remarks are then relevant.
(i) The mapping of L
O
on L
TO
generally is not bijective. Hence different
terms of L
O
may exist which are mapped into the same term of L
TO
.
(ii) The laws of T that are expressed by means of the mathematical
formalism of L
T
must be considered theoretical laws. Some of them are
expressed by means of the sublanguage L
TO
and allow one to state, via
correspondence rules, empirical laws that are expressed by means of
L
O
and can be experimentally checked, hence confirmed or falsified.
(iii) The mathematical apparatus of T is generally provided with a model
M, that is, it is mapped on another mathematical (usually geometrical
and intuitive) structure which provides a complete interpretation of
L
T
. It should be stressed, however, that such an interpretation does not
map L
T
on a physical domain unless the theoretical entities of the

3
See Braithwaite 1953 and Hempel 1965.

model are assumed to represent entities that actually exist in the
physical world.

III
Bearing in mind the scheme proposed in .II one can identify various kinds of
realism in the interpretations of physical theories. In particular, there exists a
widespread tendency to consider the theoretical entities in T (denoted by
terms in L
T
) as constituents of a reality existing beyond mere physical
phenomenology and sensory data (realism of theoretical entities, or, briefly,
RTE
4
). This tendency can be implemented in two different ways.
(i) By implicitly assigning to intuition the capability to grasp parts of
reality, hence by assuming the entities of the model M of T as real.
(ii) By directly considering the mathematical entities in L
T
as faithful
representations of some reality underlying the phenomena of the
physical world, hence identifying them with this reality.
Position (i) often occurs in classical physics. Newtons absolute space (which
opposes, because of deep ideological reasons, the relational notion of space
upheld by Leibniz), motions and trajectories of classical mechanics (whenever
mass points and forces are interpreted as parts of reality), waves of classical
electromagnetism (whenever also electric and magnetic fields are interpreted
as parts of reality) are examples of it.
5

Position (ii) is instead more frequent in modern physics whenever the
difficulties in working out consistent geometric and intuitive models of
nonclassical theories, as QM, lead one to directly assume the theoretical

4
RTE is also called scientific realism in the literature. We prefer the acronym RTE here to stress
the difference between the notion of scientific realism and the notion of semantic realism that is
introduced in the following, because these notions lead to antithetical conclusions (see .VII). We
note that, according to some authors (Alai 2009), scientific realism can be metaphysical (if
independence of the theoretical entities denoted by terms in L
T
from perception and thought is
assumed) or empirical (if reality only of the theoretical entities denoted by terms in L
T
is assumed).
This distinction, however, is problematical, and our arguments in this paper are independent of it.
5
It is interesting to note that also the old debate on the nature of the physical entities described
by QM, that is, whether they were waves or particles, was possible and meaningful only in a
perspective in which waves and particles were considered as alternative parts of reality.

entities in L
T
as real (we discuss in details some examples of this position in
.IV).
It is well known that RTE may have a positive role in the development of a
scientific theory, stabilizing its theoretical content and stimulating research
within the theory. But it may also constitute a serious hindrance to the research
progress because the empirical relations described in L
O
do not determine
univocally the theoretical apparatus, hence L
T
. Therefore, different theoretical
apparatuses generally exist which are equivalent with respect to the description
of a given empirical domain, hence the construction of L
T
requires the
adoption of nonempirical additional criteria (simplicity, intuitiveness, etc.).
These criteria, together with their interpretation, change with the cultural
context, and an enrichment of the overall knowledge may require their
modification, which in its turn implies a modification of L
T
. Moreover, since L
T

is not univocally determined, complex procedures of empirical control are
defined for each theory T, and a possible falsification may impose a
modification of L
T
. In both cases a conflictual situation occurs, since
significant variations of L
T
demand constructing a new picture of reality and
renouncing previously well-grounded parts of reality (which is psychologically
and conceptually difficult).
It should be noted, however, that nowadays physicists generally assert their
indifference (or hostility) towards any ontological commitment. Hence our
foregoing statement about the existence of forms of RTE in modern physics
should regard only a minority of scholars. But let us observe that recently there
has been a large increase in the theoretical apparatuses of some areas of physics
(e.g., string theory
6
), without a comparable rise in the empirical data, which
has strengthened the propensity to attribute an absolute role to the
mathematical formalism. Therefore many physicists implicitly assume that only
what can be described by the formalism of the theory T they are dealing with
has a physical meaning, thus considering the formalism as exhaustive of the
physical reality to which T applies. This attitude sets aside the distinction
between L
O
and L
T
and introduces a form of RTE of the kind described in (ii),
whatever the claimed epistemological attitude may be.

6
See Smolin 2005.

IV
We have seen in .II that the mapping of L
O
into L
TO
is generally not bijective.
Indeed, it may occur that terms in L
O
denoting different observational entities
correspond to the same theoretical term in L
TO
. From the point of view of the
mere mathematical formalism the observational entities at issue are then
equivalent, but they are not equivalent from the point of view of the predictions
of the theory following from the empirical interpretation (see .II).
Nevertheless the identity of the mathematical representations may induce, or
facilitate, the identification of such observational entities, which may lead to
physically disputable consequences.
Because of the relevance of the above argument we illustrate it by means of
examples, referring to QM, which is the theory that we intend to consider in
the present paper.
Example 1. A state S of a physical system E is operationally defined as a
class of physically equivalent preparing devices. Each preparing device, when
constructed and activated, produces an individual example of E or, briefly, a
physical object. A physical property E of E is operationally defined as a class of
physically equivalent registering devices. Each registering device, when
applied to a physical object x in a state S (that is, a physical object prepared by
means of a preparing device belonging to S) produces one of two possible
answers (for example, yes/no). States and properties have therefore quite
different operational definitions. Nevertheless, a subset of states and a subset
of properties exist which can be put in a one-to-one correspondence: namely,
the subset of pure states and the subset of atomic properties in the lattice of all
properties of E.
7
Moreover, a pure state S and the atomic property E
S

corresponding to it (usually called the support of S) are represented by the
same mathematical entity (a one-dimensional projection operator of the form
|)(|, with|) unit vector of the Hilbert space H associated with E) in QM.
This representation therefore suggests one to identify S and E
S
(that are
obviously terms of L
O
), especially if a role is attributed to the mathematical
apparatus which goes beyond the role of mere formal representation
(conversely, the identification itself may constitute a clue of a possible

7
For the sake of brevity, we do not enter into technical details defining explicitly these notions,
see, e.g., Beltrametti and Cassinelli 1981.

epistemological position of this kind, e.g., RTE). By analyzing the literature on
this argument one then finds that the identification between S and E
S
is in fact a
basic element in at least a conceptually relevant semi-axiomatic approach to
QM (Piron 1976). Furthermore, this identification is sometimes implicitly
accepted in the current literature (Bouwmeester et al. 1997).
On the other hand, missing the distinction between pure states and their
support can lead to misunderstandings. Indeed, a physical object in a pure
state S possesses the support E
S
of S with certainty, but physical objects may
exist which display the property E
S
in a measurement without being prepared
in the state S. From a semantic point of view one can say that the extensions of
S and E
S
do not necessarily coincide, while the identification of S and E
S
can
lead one to identify them, as it occurs in the example that follows. More
generally, representing S and E
S
by means of the same mathematical term in L
T

implies that the semantic differences between S and E
S
are not preserved in the
syntactic apparatus of L
T
, which constitutes a serious limit of the technical
language of QM.
Example 2. Many manuals and popular books on QM introduce the reader
to the surprising features of QM by discussing the well known two-slit
experiment, or a modern variant of it (Mach-Zehnder interference
experiment
8
). This experiment can be synthetically described as follows.
A monochromatic beam of light hits a screen with two close slits, say 1 and
2, and light is then collected on a faraway second screen. Whenever both slits
are open the light distribution on the second screen is not the superposition of
the two distributions that can be obtained by closing either slit 1 or slit 2.
Rather, an interference pattern appears which suggests that the beam should
be described as a wave. On the other hand the same experiment produces
isolated spots on the second screen if it is performed with a low intensity beam,
which suggests that the beam should be described as a bunch of particles.
According to current literature these contradictory results can be explained
only accepting a particle model for the beam but avoiding attributing to the
particles all the features that they should have according to classical physics.
The reasoning leading to this conclusion proceeds ab absurdo and can be
schematized as follows.
(i) A seemingly obvious premiss of objectivity is stated:

8
See, e.g., Albert 1992, etc.

(O) Each particle possesses the property E
1
of passing through slit 1
or the property E
2
of passing through slit 2.
(ii) Because of O one can consider all particles that possess the property
E
1
(E
2
): these should produce the same distribution on the final screen
that is produced by particles passing through slit 1 (2) when slit 2 (1)
is closed.
(iii) It follows that the overall distribution should be the superposition of
the two distributions obtained by closing either slit 1 or slit 2.
(iv) Experimental data do not confirm conclusion (iii).
(v) Because of (iv) premiss (O) is falsified.
(vi) Hence, for every particle, the property of passing through a given slit
cannot a priori be considered as possessed or not possessed by the
particle, as it occurs in classical physics. A property of this kind is a
nonobjective, or potential, property that may become actual only if a
measurement is performed determining which option occurs.
Therefore the quantum notion of particle differs in a fundamental way
from the classical notion.
The argument expounded above is widespread and commonly accepted. It is
therefore important to observe that statement (ii) does not follow from
statement (i). Rather, statement (ii) introduces an implicit additional
assumption, which can be made explicit as follows.
(A) A particle possessing the property E
1
(E
2
) whenever slits 1 and 2 are
both open is physically equivalent (at least with respect to the
distribution on the second screen) to a particle prepared by leaving slit
1 (2) open and slit 2 (1) closed.
The explicit statement of assumption (A) makes it evident that statement (ii)
postulates a physical equivalence between properties and preparations, hence
states (which implies, in particular, that E
1
and E
2
are mutually exclusive). As
long as the operational definitions of states and properties are not explicitly
given, assumption (A) seems intuitively obvious because one implicitly
assumes an elementary model according to which particles move along straight
trajectories. But if one refers to the operational definitions introduced in

Example 1 and avoids nave models, assumption (A) is questionable
9
, and the
experimental falsification of conclusion (iii) does not necessarily falsify (O),
because it could instead show that (A) does not hold. Nevertheless,
representing pure states and their supports by means of the same mathematical
entities provides a natural, though improper, backup to assumption (A),
especially if one accepts, more or less explicitly, RTE in the version (ii) of .III
(it is probably because of this backup that the critics to the two-slit argument,
which is implicit in the reasoning above, has never been propounded by other
authors
10
).
Example 3. Besides pure states, mixed states, or proper mixtures, are
usually introduced in QM. A proper mixture M can be operationally defined as
a set of pure states in which each pure state S in associated with a weight p
S
,
interpreted as the epistemic probability (which expresses a subjective lack of
knowledge about the physical situation) that the system be actually prepared in
the pure state S. Hence the term M belongs to L
O
. Moreover, the proper
mixture M is represented by a density operator on H in QM (the technical
features of a density operator are not relevant for our purposes). Hence the
term belongs to L
TO
. Then, several problems arise because of such
definition and representation.
(i) A given density operator corresponds to some different operational
definitions that are equivalent as far as probabilities of physical
properties are concerned, hence the corresponding mixtures are
identified in QM. But this identification overlooks the fact that
different operational definitions are not equivalent at an individual
level because the possible pure states of the physical object that is
considered are different when the operational definitions are different.
(ii) It follows from (i) that the knowledge of a density operator is not
sufficient to pick out a single operational definition, which implies that
the coefficients in the decomposition of into pure states cannot
generally be interpreted as epistemic probabilities (Beltrametti and
Cassinelli 1981).

9
For instance, one cannot a priori exclude that the particle possesses both the property of
passing through slit 1 and the property of passing through slit 2 whenever both slits are open. Indeed,
these properties could be possessed by the same particle at different times (a straight trajectory of the
particle would of course be excluded in this case).
10
See Garola 2000.

Example 4. Also improper mixtures are introduced in QM (dEspagnat
1976). An improper mixture N is operationally defined in a complex way,
which is equivalent to assigning a set of pure states in which each pure state S is
associated with a weight p
S
, as in the case of a proper mixture, hence the term
N belongs to L
O
. In this case, however, p
S
can never be interpreted as an
epistemic probability. This notwithstanding also the improper mixture N is
represented by a density operator on H in QM. It follows that the same term
in L
TO
can represent both a proper and an improper mixture. This identity of
representations may lead one to disregard the distinction between the two
kinds of mixtures, reaching the doubtful conclusion that the problems of the
quantum theory of measurement can be solved within the standard
interpretation of QM (Garola and Sozzo 2007, Genovese 2005, Schlosshauer
2004).
Examples 1-4 illustrate the theses expounded at the beginning of this
section. Summarizing, they show that the technical language of QM has serious
limits because the mathematical representations do not distinguish some
different physical entities, hence the language L
T
of QM does not adequately
express the semantic differences existing in the language L
O
on which L
T
is
interpreted (L
T
should therefore be suitably extended: attempts in this
direction have been forwarded in particular by the Brussels school
11
, but they
are generally ignored by physicists involved in the foundations of QM). These
expressive limits of L
T
may suggest or support the improper identification of
different observational terms having the same mathematical representations. It
is now important to observe that such an identification naturally follows if one
accepts a form of RTE of the kind discussed in .III, (ii). Therefore,
overlooking the important distinctions illustrated in the previous examples
constitutes in our opinion a significant clue to an implicit adoption of RTE.

V
The conclusions expounded at the end of .IV would probably be considered
inessential by most physicists, who usually rely on the mathematical formalism
and are scarcely sensitive to epistemological analysis of the kind carried out in

11
See Aerts 1999.

.IV. Moreover many scholars would claim that our invalidation of the two-slit
argument on the basis of the distinction between states and properties is
unimportant because, for every physical object in a given state, the existence of
nonobjective properties can be asserted on the basis of more rigorous
arguments, i.e., the Bell theorem (Bell 1964) and the Bell-Kochen-Specker, or
Bell-KS, theorem (Bell 1966, Kochen and Specker 1967) that are usually
maintained to prove the nonlocality and contextuality, respectively, of QM (we
recall that these features of QM play a relevant role in quantum information and
quantum computation). It is therefore relevant to our aims to observe that the
proofs of these theorems depend on some epistemological assumptions that
are unanimously accepted without being explicitly recognized, and that these
assumptions can be questioned when looked into more deeply. Since this
statement is compelling, we resume its proof in .VI, limiting ourselves to the
Bell-KS theorem for the sake of brevity, and introduce here some preliminary
notions that are needed to this end.
First of all, let us observe that, bearing in mind the framework introduced in
.II, we can single out two disjoint subclasses in the class of all theoretical laws
of classical and quantum physics.
(i) The class of all laws stated by means of L
T
\L
TO
, consisting of
mathematical expressions that have not a direct physical interpretation
on the domain of physical facts.
(ii) The class of all laws stated by means of L
TO
, consisting of mathematical
expressions, generally deduced from the laws of the former class,
which allow one to state in L
O
, via correspondence rules, empirical
laws that can be confirmed or falsified (by abuse of language, we briefly
call empirical laws also the laws belonging to this class in the
following).
This remark allows one to point out a fundamental distinction between
classical and quantum theories. In classical theories there is no theoretical limit
to the possibility of confirming or falsifying an empirical law. In QM instead,
this possibility is restricted because incompatible observables occur in QM
whose values can be neither measured nor predicted simultaneously. Indeed,
incompatibility entails that physical situations may exist in which an empirical
law can be, in principle, neither confirmed nor falsified. A situation of this kind
occurs, e.g., whenever a physical object (individual example of a physical
system, see .IV) is known to possess a property E, and the empirical law that

one is considering establishes a relation between two further compatible
properties that are not compatible with E. Then, two different positions can be
adopted about the truth value of a sentence o of L
TO
expressing an empirical
physical law (Garola and Pykacz 2004, Garola and Solombrino 1996a, Garola
and Solombrino 1996b).
Metatheoretical classical principle (MCP). o is true in every physical
situation that can be devised, even if this situation is such that QM does not
allow one to check the empirical law expressed by o.
Metatheoretical generalized principle (MGP). o is true in every physical
situation in which the law can be checked (epistemically accessible physical
situation), while it may be true as well as false in physical situations in which
QM does not allow one to check the physical law expressed by o.
Accepting MCP implies maintaining that empirical physical laws express
relations on the set of theoretical entities that hold at a deeper level of reality,
beyond the level in which empirical confirmation is possible. Accepting MGP
implies instead a weaker truth mode of empirical physical laws, is consistent
with the antimetaphysical attitude underlying QM and avoids any form of
RTE.

VI
Let us come now to the Bell-KS theorem. As we have anticipated in .V, this
theorem is maintained to provide a rigorous support to the conclusions
traditionally attained by means of the two-slit experiment (.IV, Example 2)
stating that, if one accepts the assumption that QM deals with physical objects
and their properties, then for every physical object in a given state there are
properties which depend on the set of measurements that are performed on the
object (contextual, or nonobjective, or potential, properties). Therefore these
properties cannot be considered as possessed or not possessed by the object
independently of the experimenters choices.
All proofs of the Bell-KS theorem assume (sometimes, implicitly) the
following condition (Bell 1966, Kochen and Specker 1967; see also Mermin
1993).
KS. Let A, B, be compatible observables and let
f(A, B, )=0 (1)

express an empirical quantum law. Then, whenever measurements of A, B,
are performed obtaining the outcomes a, b, , respectively, the following
equation
f(a, b, )=0 (2)
holds.
Condition KS is needed if one wants to get predictions from empirical laws,
hence it must be accepted for physical reasons. All proofs then proceed ab
absurdo. They consider several empirical quantum laws,

.. ..........
0 = ) , B' , g(A'
0 = ) B, f(A,
(3)
assume that the values a, b,; a', b',; of the observables A, B,; A', B',;
respectively, are defined independently of any measurement procedure for
any physical object x, apply the KS condition repeatedly, which implies that f(a,
b,)=0, g(a', b',)=0, must hold simultaneously, and finally show that a
contradiction occurs. A seemingly unavoidable conclusion is that the values a,
b,; a', b',; are not defined independently of the set of measurements that
are performed (contextuality), hence cannot be considered as preexisting to
the measurements.
The proofs schematized above are mathematically correct. However, a
careful analysis shows that their premises follow from the adoption, implicit but
essential, of the epistemological position MCP introduced in .V. Indeed, we
have seen that each proof requires a repeated application of the KS condition.
But direct inspection shows that in every proof there are observables in a law
(say, f(A, B,)=0) that are not compatible with some observables in another
law (say, g(A', B',)=0). Whenever the values a, b,; a', b', are
simultaneously attributed to the physical object x, a nonaccessible physical
situation is devised in which only one (at choice) of the empirical laws can be
checked. If one adopts the position expressed by the weaker principle MGP in
.V, the proofs of the Bell-KS theorem cannot be completed because one
cannot assert that f(a, b,)=0 and g(a', b',)=0 hold simultaneously. Hence
the repeated application of the KS condition implies postulating the
unrestricted simultaneous validity of all empirical quantum laws, that is, MCP.

To conclude this section, let us note that the adoption of MCP can be
defended by observing that physicists can choose the empirical law that they
want to check among the laws listed in Eqs. (3). Since all experiments show
that, for every choice, quantum predictions are fulfilled, it is difficult to
understand how a breakdown of a law in Eqs. (3) may occur just when another
law is experimentally proven to hold (conspiracy of nature
12
). We show in
.VIII, however, that this argument can be overcome by suitably reinterpreting
quantum probabilities.

VII
Our analysis in .VI can be repeated by considering further theorems of the
same kind, as the Bell theorem (Garola and Pykacz 2004, Garola and
Solombrino 1996a, Garola and Solombrino 1996b). Our conclusion can be
resumed and integrated as follows.
(i) The deduction of fundamental and universally accepted theorems
which state the contextuality and nonlocality of QM requires the
adoption of an epistemological position (MCP) that assumes the
validity of empirical laws also in physical situations in which the laws
cannot, in principle, be checked. On the other hand, MCP necessarily
follows whenever RTE is assumed and extended to relations on the set
of theoretical entities formalized by means of L
T
(realism of theories;
Boniolo and Vidali 1999). Hence the adoption of MCP constitutes a
clue, if not a proof, that RTE has been more or less implicitly accepted.
(ii) The contextuality of QM entails that, for every state of a physical
system, nonobjective physical properties exist (.VI). It follows that
RTE (which implies MCP, hence the contextuality of QM) prevents
one from adopting in QM any form of realism assuming objectivity of
properties of physical objects. In particular, RTE does not allow one to
adopt semantic realism in QM, i.e., a purely semantic form of realism
which avoids ontological commitments and only assumes that every
sentence attributing a property to a physical object is semantically
objective, in the sense that it can be provided with a truth value

12
See, e.g., Lalo 2001.

(true/false). This result is relevant because the opposition between
RTE and semantic realism is a priori unexpected, for it does not occur
in classical physics and characterizes QM.
(iii) A change in the epistemological perspective (in particular, the
replacement of MCP with MGP) may invalidate some deeply-rooted
beliefs, opening the way to new interpretations of the formalism of QM
(in particular, an interpretation in which semantic realism is assumed,
hence contextuality and nonlocality are avoided).

VIII
An interpretation of the kind conjectured in .VII, (iii), has been propounded
several years ago by one of the authors, together with other authors (semantic
realism, or SR, interpretation
13
). Successively various models have been
provided to show the consistency of this interpretation. The last of these,
named extended semantic realism (ESR) model, is a new kind of noncontextual
hidden variables theory for QM which introduces, besides hidden variables, a
reinterpretation of standard quantum probabilities.
14
The ESR model modifies
(and in some sense extends) the original SR interpretation, but preserves its
basic features, that is, semantic realism and the substitution of MCP with the
weaker principle MGP. Within this model many problems (e.g., the
objectification problem of the quantum theory of measurement
15
), paradoxes
(e.g., the EPR paradox and the Schrdingers cat paradox) and the ambiguities
illustrated in .IV disappear. These results are interesting not only from a
physical point of view (interpretation of entanglement, quantum information,
etc.) but also from a philosophical perspective. Therefore we devote this
section to illustrate the main features of the ESR model.
The basic set of theoretical entities introduced by ESR model is a set E of
hidden variables, whose elements are interpreted as microscopic properties
(some additional parameters may occur that we do not discuss here). Given a
physical object x and a microscopic property f, x either possesses or does not

13
See Garola and Solombrino 1996a, 1996b.
14
See Garola 2002, 2003, 2007, 2009a, Garola and Pykacz 2004, Garola and Sozzo 2009a,
2009b, 2010a, 2010b, Sozzo 2007, Sozzo and Garola 2010.
15
See, e.g., Busch et al. 1991.

possess f. The set of all microscopic properties possessed by x then defines the
microscopic state of x.
Each feE corresponds to a macroscopic property F. If F is measured on x
and x displays F, then x possesses f. But the converse implication does not
hold, because the set of microscopic properties possessed by x (that is, the
microscopic state of x) might be such that x cannot be detected when F is
measured, independently of the specific features of the apparatus measuring F.
Hence a detection probability is associated with the measurement of F which
depends on the microscopic state of x, not only on F, and must not be mistook
for the detection probability that occurs because of the reduced efficiencies of
real measuring apparatuses.
The introduction of detection probabilities depending on the microscopic
state characterizes the ESR model. The model does not say anything about the
deep causes of them: rather, introduces them as overall results of these causes.
Intuitively, one can think that something is happening at a microscopic level
which underlies the standard quantum picture of the physical world and does
not reduce to it, so that a broader theory is needed. The ESR model aims to be
a first step in this direction.
The macroscopic part of the ESR model (briefly, the macroscopic ESR
model) rests on the features of the microscopic part specified above, but it can
be presented without mentioning hidden variables, as a self-consistent
theoretical proposal. In particular, detection probabilities occur in it that can
be considered as unknown parameters whose value is not predicted by any
existing theory. More generally, the main features of the macroscopic ESR
model can be summarized as follows.
16

(i) It brings into every measurement a no-registration outcome which is
interpreted as providing information about the microscopic world that
is inquired, as well as any other possible outcome, hence it substitutes
the observables of QM with generalized observables with enlarged sets
of possible values.
(ii) It embodies the mathematical formalism of standard (Hilbert space)
QM into a broader noncontextual framework, which explains how the
objectification problem and the quantum paradoxes quoted above can
be avoided.

16
See in particular Garola and Sozzo 2009b.

(iii) It reinterprets the quantum rules for calculating the probability that a
physical object x in a state S display a property F when a measurement
of F is performed as referring to a selection of x in the subset of all
physical objects that have been prepared in S and can be detected
(conditional probability) rather than in the set of all physical objects
prepared in S (absolute probability).
17

(iv) It provides some predictions that are formally identical to those of QM
but have a different physical interpretation and further predictions that
differ also formally from those of QM. Therefore it can be empirically
checked, hence it is falsifiable.
(v) It implies, by introducing some additional assumptions, that the Bell-
Clauser-Horne-Shimony-Holt (BCHSH) inequality, a modified
BCHSH inequality and (reinterpreted) quantum predictions hold
together because they refer to different parts of the picture of the
physical world supplied by the model, hence it overcomes the
opposition between the BCHSH inequality and the formal apparatus of
QM in a framework in which physical properties are (semantically)
objective, hence local realism (Bell 1964; Einstein, Podolsky and
Rosen 1935) holds.
(vi) It introduces a kind of unfair sampling that explains the breakdown of
the BCHSH inequality at a macroscopic level.
By formulating the foregoing theoretical proposal in mathematical terms, we
have recently obtained some further relevant results in the ESR model.
18

(a) Each generalized observable is represented by a (commutative) family
of positive operator valued (POV) measures parametrized by the set of
all pure states of the physical system that is considered. It follows that
every physical property is represented by a family of bounded positive

17
This reinterpretation of quantum probabilities allows one to explain, without resorting to any
conspiracy of nature, how it may occur that an empirical quantum law may fail to be true whenever
another empirical quantum law is checked and proven to hold. Indeed, a physical object x that is
detected when measuring, say, the observables A, B, in Eqs. (3), which implies that the equation f(a,
b,)=0 must hold, could possess such microscopic properties that it would not be detected if instead
the observables A', B', were measured, hence one cannot assert that the equation g(a', b',)=0
must also hold. But, of course, if A', B', were measured and x were detected, then g(a', b',)=0
would hold, because quantum laws hold for every detected object according to the ESR model.
18
See in particular Garola and Sozzo 2010a.

operators (effects) rather than by a single projection operator, hence
pure states can never be identified with atomic properties and the
problems pointed out in .IV, Examples 1 and 2, are avoided.
(b) A generalized projection postulate (GPP) rules the transformations of
pure states induced by nondestructive idealized measurements.
(c) The conciliatory conclusion mentioned in (v) can be recovered without
introducing additional assumptions.
(d) Each mixture is represented by a family of density operators
parametrized by the set of all properties characterizing the physical
system that is considered.
(e) The new representation of mixtures avoids the problems pointed out
in .IV, Examples 3 and 4. Indeed, proper mixtures having different
operational definitions are represented by different families of density
operators, even if they are probabilistically equivalent, which avoids
problems (i) and (ii) in Example 3. Moreover, the distinction between
proper and improper mixtures does not occur because all probabilities
are epistemic, which avoids the problem sketched in Example 4.
(f) A generalized Lders postulate (GLP) that generalizes GPP rules the
general transformations of states induced by nondestructive idealized
measurements.
(g) GPP can be (partially) justified by describing a measurement as a
dynamical process in which a nonlinear evolution occurs of the
compound system made up of the (microscopic) measured object plus
the (macroscopic) measuring apparatus.

REFERENCES
Aerts, D. (1999). Quantum Mechanics: Structures, Axioms and Paradoxes. In
D. Aerts & J. Pykacz (Eds.), Quantum Structures and the Nature of
Reality (pp. 141-197). Dordrecht: Kluwer Academic Publishers.
Alai, M. (2009). Realismo scientifico e realismo metafisico. Giornale di Fisica,
L(1), S19-S27.
Albert, D. Z. (1992). Quantum Mechanics and Experience. Cambridge, MA:
Harvard University Press.

Bell, J. S. (1964). On the Einstein-Podolsky-Rosen Paradox. Physics, 1(3),
195-200.
Bell, J. S. (1966). On the Problem of Hidden Variables in Quantum
Mechanics. Reviews of Modern Physics, 38(3), 447-452.
Beltrametti, E. G., & Cassinelli, G. (1981). The Logic of Quantum Mechanics.
Reading, MA: Addison-Wesley.
Boniolo, G., & Vidali, P. (1999). Filosofia della scienza. Milano: Bruno
Mondadori.
Bouwmeester, D., Pan, J.-W., Mattle, K., Eibl, M., Weinfurter, H., &
Zeilinger, A. (1997). Experimental Quantum Teleportation. Nature,
390, 575-579.
Braithwaite, R. B. (1953). Scientific Explanation. Cambridge: Cambridge
University Press.
Busch, P., Lahti, P. J., & Mittelstaedt, P. (1991). The Quantum Theory of
Measurement. Berlin: Springer.
dEspagnat, B. (1976). Conceptual Foundations of Quantum Mechanics.
Reading, MA: Benjamin.
Einstein, A., Podolsky, B., & Rosen, N. (1935). Can Quantum-Mechanical
Description of Physical Reality be Considered Complete?. Physical
Review, 47, 777-780.
Garola, C. (2000). Objectivity versus Nonobjectivity in Quantum Mechanics.
Foundations of Physics, 30(9), 1539-1565.
Garola, C. (2002). A Simple Model for an Objective Interpretation of
Quantum Mechanics. Foundations of Physics, 32(10), 1597-1615.
Garola, C. (2003). Embedding Quantum Mechanics into an Objective
Framework. Foundations of Physics Letters, 16(6), 605-612.
Garola, C. (2007). The ESR Model: Reinterpreting Quantum Probabilities
Within a Realistic and Local Framework. In G. Adenier et al. (Eds.),
Quantum Theory: Reconsideration of Foundations - 4 (pp. 247-252).
Melville, NY: American Institute of Physics.

Garola, C. (2009a). A Proposal for Embodying Quantum Mechanics in a
Noncontextual Framework by Reinterpreting Quantum Probabilities. In
L. Accardi et al. (Eds.), Foundations of Probability and Physics - 5 (pp.
42-50). Melville, NY: American Institute of Physics.
Garola, C. (2009b). An Epistemological Criticism to the Bell-Kochen-Specker
Theorem. In L. Accardi et al. (Eds.), Foundations of Probability and
Physics - 5 (pp. 51-52). Melville, NY: American Institute of Physics.
Garola, C., & Pykacz, J. (2004). Locality and Measurements Within the SR
Model for an Objective Interpretation of Quantum Mechanics.
Garola, C., & Solombrino, L. (1996a). The Theoretical Apparatus of Semantic
Realism: A New Language for Classical and Quantum Physics.
Garola, C., & Solombrino, L. (1996b). Semantic Realism Versus EPR-like
Paradoxes: The Furry, Bohm-Aharonov, and Bell Paradoxes.
Garola, C., & Sozzo, S. (2007). The Physical Interpretation of Partial Traces:
Two Nonstandard Views. Theoretical and Mathematical Physics,
152(2), 1087-1098.
Garola, C., & Sozzo, S. (2009a). The ESR Model: A Proposal for a
Noncontextual and Local Hilbert Space Extension of QM. Europhysics
Letters, 86(2), 20009-20015.
Garola, C., & Sozzo, S. (2009b). Embedding Quantum Mechanics into an
Objective Framework: A Conciliatory Result. International Journal of
Theoretical Physics. Published online, DOI 10.1007/s10773-009-
0222-8.
Garola, C., & Sozzo, S. (2010a). Generalized Observables, Bells Inequalities
and Mixtures in the ESR Model for QM. Foundations of Physics.
Published online, DOI 10.1007/s10701-010-9435-1.
Garola, C., & Sozzo, S. (2010b) The Representation of Mixtures in the ESR
Model for QM. In A. Y. Khrennikov et al. (Eds.), Quantum Theory:
Reconsideration of Foundations - 5. Melville, NY: American Institute of
Physics. In print.

Genovese, M. (2005). Research on Hidden Variables Theories: A Review of
Recent Progresses. Physics Reports, 413(6), 319-396.
Hempel, C. G. (1965). Aspects of Scientific Explanation. New York: Free
Press.
Kochen, S., & Specker, E. P. (1967). The Problem of Hidden Variables in
Quantum Mechanics. Journal of Mathematics and Mechanics, 17(1),
59-87.
Lalo, F. (2001). Do We Really Understand Quantum Mechanics? Strange
Correlations, Paradoxes, and Theorems. American Journal of Physics,
69(6), 655-701.
Mermin, N. D. (1993). Hidden Variables and the Two Theorems of John Bell.
Reviews of Modern Physics, 65(3), 803-815.
Norsen, T. (2007). Against Realism. Foundations of Physics, 37(3), 311-
340.
Piron, C. (1976). Foundations of Quantum Physics. Reading, MA: Benjamin.
Schlosshauer, M. (2004). Decoherence, the Measurement Problem, and
Interpretations of Quantum Mechanics. Reviews of Modern Physics,
76(4), 1267-1305.
Smolin, L. (2005). Why No New Einstein?. Physics Today, 58(6), 56-57.
Sozzo, S. (2007). Modified BCHSH Inequalities Within the ESR Model. In G.
Adenier et al. (Eds.), Quantum Theory: Reconsideration of
Foundations - 4 (pp. 334-338). Melville, NY: American Institute of
Physics.
Sozzo, S., & Garola, C. (2010). A Hilbert Space Representation of
Generalized Observables and Measurement Processes in the ESR
Model. International Journal of Theoretical Physics. Published online,
DOI 10.1007/s10773-010-0264-y.
Tassani, I. (Ed.) (2004). Quanti Copenhagen? Bohr, Heisenberg e le
interpretazioni della meccanica quantistica. Modena: Societ Editrice
Il Ponte Vecchio.


Expansion of the Universe and Spacetime Ontology
*

Giovanni Macchia**
lucbian@hotmail.com

The theory of the expanding universe is
in some respects so preposterous that
we naturally hesitate to commit ourselves to it.
It contains elements apparently so incredible that
I feel almost an indignation
that anyone should believe in it
except myself.
Arthur Eddington
1

ABSTRACT
The debate on the ontological status of spacetime in General Relativity has
historically seen two principal philosophical contenders: substantivalism,
roughly the view that holds that spacetime exists apart from the material
contents of the universe, and relationism, the doctrine that spacetime does not
exist, i.e., it is a mere abstract web of spatiotemporal relations among bodies.
This dispute, however, has rarely been fought on a cosmological battlefield. In
this paper an attempt in this direction is made. The question at issue is the
following: is there any feature of our universe that requires or is best explained
in terms of a substantival space? I claim that there is indeed: the expansion of
the universe, perhaps the most important phenomenon in cosmology, can play
such a role.

* I am deeply grateful to Silvio Bergia, Claudio Calosi, Vincenzo Fano, Sona Ghosh, Chris
Smeenk for very useful observations and suggestions to a previous version of this paper. Needless to
say, I am the only responsible for the conclusions I advance.
** Urbino University
1
Eddington 1933, p. 124.

The expansion of the universe appears as a simple idea: it means that the
proper physical distance between a [typical] pair of well-separated galaxies is
increasing with time, that is, the galaxies are receding from each other
(Peebles 1993, p. 71) with velocities proportional to their distances. The
systematic redshifts of radiation emitted by distant galaxies are the primary
piece of evidence regarding expansion we can gather from direct
observations.
2
General-relativistic interpretation of these cosmological
redshifts is widely accepted, though often challenged and sometimes
misinterpreted; in any case, it is still not an exaggeration to say that
cosmological redshift is the most important fact of modern cosmology
(Merleau-Ponty 1965, p. 28). According to this physical interpretation, the
wavelength of the emitted radiation from a distant source is stretched during its
travel from this source to the observer so that the cosmological redshift is an
effect of the large-scale expansion of the universe when regarded as an
expansion of space.
3
Hence the global recession of galaxies originates in the
dynamical evolution of the universal spacetime metric, and not from the
effective motion of galaxies through a static space, in which case redshift would
be interpreted as a classic Doppler shift. However, some authors argue that the
expansion of space is a mere coordinate-dependent effect and redshifts are
Doppler shifts. Accordingly, the interpretation of the universal expansion is
still being debated among cosmologists, especially in these last years in light of
new cosmological data.
In this paper I analyse some aspects (confined to classical General Relativity
(GR) and to those cosmological models describing homogeneous and isotropic
universes) of such a rich debate, favouring those related to the nature of the
redshift. I will try to show how its different interpretations, correlated with
astronomical observations, might become a crucial discriminating link not only
between possible choices of cosmological models but also between ontological
commitments on spacetime. I argue that the cosmological interpretation of
redshift privileges a substantivalist ontology of spacetime insofar as universal
metric field at large scales has an active role in wavelength stretching. On the

2
However, the expansion of the universe, and the Big Bang theory in which it is mostly codified,
is also supported by other phenomena, in primis by the interpretation of the cosmic microwave
background radiation as the relic radiation from a hot big bang, and by the cosmic abundances of
helium and deuterium produced in a primordial hot phase of the universe.
3
Distant objects are those with redshifts of order one or more, for which, as will be clear later,
the light-travel distances involved are comparable to the curvature length scale of the universe; thus, at
these distances, spacetime cannot be approximated as flat.
Giovanni Macchia Expansion of the Universe and Spacetime Ontology 105

contrary, a relationist point of view regarding expansion would be sustainable
only in arguing in favour of effective galactic motions through a non-expanding
space; in particular, relationists would have to resort to Doppler
interpretations and accordingly fall back on cosmologically unfruitful
descriptions based on Special Relativity (SR).
The plan of the paper is as follows: in section 1, I review the Doppler and
cosmological redshifts with their differences in order to introduce, in section
2, the most common approach to the expansion of the universe, the so-called
expanding space paradigm. In section 3, I explain in which sense the expansion
of space is regarded as a coordinate-dependent concept. After having
expounded, in section 4, the theoretical and observational reasons favouring a
cosmological interpretation of distant redshifts, I take a closer glance at the
expansion: its simpler physical aspects (section 5) and its effects on local
systems (section 6). Starting from section 7, some philosophical analyses of the
expansion are pursued, showing how it sustains a substantivalist metaphysics
on spacetime (section 8), whereas a relational one is not satisfactory (section
9).

1. A LITTLE BIT OF TECHNICAL BACKGROUND
It is well known that it is possible to recognize the atomic origin of incoming
light from the pattern of the received spectral lines, each line being a precise
measurable wavelength. Hence it is possible to compare different light spectra
of chemical elements whose emitting sources have different status (at rest,
moving, large or small distance from the detector) and eventually detect
systematic shifts (towards the red or the blue of the spectrum) of their
characteristic wavelengths.
4
Two reasons for these systematic displacements
are here analysed.

4
Obviously, these comparisons make sense only if it is assumed that the frequency with which
light is emitted by a distant object (then in a remote past) is the same with which light, of the same
element at rest in our laboratories, radiates now.

1.1 DOPPLER SHIFT
Doppler shift is the fractional shift of the spectral lines of an object due to a
decreasing or increasing radial distance between the emitting object and the
observer. For a received wavelength
rec
and an emitted wavelength
em
,
Doppler shift z is defined as:

.
If the object is receding, the spectral lines are shifted toward longer
wavelengths (redshifts), whereas if it is approaching, spectral lines are shifted
toward shorter wavelengths (blueshifts). Doppler shift is the result of the
relative motion of objects moving through space: it depends upon the relative
velocity between the source and the receiver and it is caused by peculiar
velocities v (i.e., physical motions in space) of objects. If (where c is the
speed of light in a vacuum), z is expressed by the Fizeau-Doppler formula:
= .
5

1.2 COSMOLOGICAL SHIFT
Cosmological redshift is usually drawn from the most common cosmological
model of our universe whose metric, describing the large-scale structure of the
universe, is the so-called Friedmann-Lematre-Robertson-Walker (FLRW)
spacetime
6
, which can be written as
7
:
2
=
2
2
+
2
2
+
2
2
+ sin
2
d
2
,
where t is the synchronous cosmic time coordinate; , and are the
comoving space coordinates; the function
= sin, sinh

for
closed, flat, or open spatial sections, respectively; and the scale factor R(t),
with dimensions of distance, characterizes the relative size of the space

5
This formula is an approximation of the relativistic Doppler effect (with ) given by (for a
source moving at an angle to the observer) 1 + = (1 + cos )/1
2
/
2
(for small ,
indeed, it reduces to ).
6
The currently adopted basis for the interpretations of astrophysical observations is a model
< ,
> in which M is the spacetime manifold,
is the following FLRW metric

(representing the gravitational field and the geometric structure of spacetime), and
is the stress-
energy tensor (representing the material contents of the universe) taking a perfect fluid form with zero
pressure (dust). In this idealization, valid only at large scales, clusters of galaxies may be regarded as
the particles the so-called fundamental particles out of which this fluid is made.
7
See Cook and Burns 2008, p. 3; Davis and Lineweaver 2003, p. 18; Peacock 1999, p. 69.

sections (each one having constant curvature) as a function of cosmic time and
hence describes is the essence of the expansion of the universe when
regarded as the expansion of space. Indeed, speaking of an expanding space
makes sense in these particular highly symmetric cosmological models
8
in
which it is possible to decompose in an unique way the spacetime into space
and time, namely, into spatial hypersurfaces of simultaneity expanding over
time.
9
The idealized fluid of fundamental particles is always at rest relative to
the space coordinates (and for this reason are called comoving). Thus the
constant spatial coordinates mark the position of the fundamental particles
whose worldlines are geodesics of the metric. Fundamental particles, when
regarded as mere geometric points, constitute the kinematic substratum. At
every point of such a substratum, a locally inertial reference frame, along with
its own fundamental observer, can be defined. Along their worldlines,
fundamental observers measure the same cosmic time and see the universe
spatially isotropic and homogeneous.
From the radial null-geodesic ( = = = 0) of the FLRW line
element, we may obtain the cosmological redshift z from the Lematres
equation:
1 + =
=
(
)
(
)
,
where
rec
(
em
) is the wavelength received (emitted) at t
rec
(t
em
) when the factor
scale of the universe was R(t
rec
) (R(t
em
)). In an expanding universe such as ours,
R(t
rec
)>R(t
em
), so wavelengths increase in time and z is always positive, while in
a contracting one wavelengths decrease and z is negative. This equation
contains no velocity terms, and hence z cannot be measuring any sort of
kinematic redshift: z here depends on the increase of distance between emitter
and receiver during the time of propagation, not from their actual movements
in space. How then to interpret the cosmological redshift? Rindler is clear:

8
The FLRW spacetime is the most general metric compatible with an isotropic and spatially
homogeneous distribution of dust (such a requirement is named Cosmological Principle). The form of
this metric is derived from these symmetries alone, and not from the assumptions of GR (Rindler
2006, p. 368). The Einstein field equations intervene only afterwards to determine (by the so-called
Friedmann equations) how R(t) changes with time for a given universe with specified matter and
radiation contents.
9
In other words, in FLRW models there are invariant quantities that we can think of as indicating
in some clear sense that space is expanding, whereas, in more general (less symmetric) models, where
that decomposition is not even possible, it is not evident what the term space means.

What is remarkable about this formula is that the frequency shift depends only
on the values of R(t) at emission and reception. What R(t) does in between is
irrelevant. Regarded in this way, the cosmological redshift is really an
expansion effect rather than a velocity effect. (Rindler 2006, p. 375)
In other words, cosmological redshift cannot be a Doppler velocity effect
because the increase of wavelength does not depend on the rate of change of
R(t) at emission or reception, but on the increase of [R(t)] in the whole period
from emission to absorption (Weinberg 2008, p. 11, my emphasis).
10

2. THE EXPANDING SPACE PARADIGM
Although Doppler and cosmological effects cannot be distinguished from one
another by observing their spectra, the interpretative difference, as deduced
from their mathematical expressions (quite similar at small distances, but
significantly different at large ones), is very important: the latter effect does not
contain any relative velocity, but only a relative distance, namely the ratio
R(t
rec
)/R(t
em
). This relative distance tells us how much the universe has
expanded during the time the light has been travelling. Due to the fact that the
difference between these two redshifts is evident only at large scales,
historically the cosmological nature of high redshifts was recognized only when
observations of the distant universe became more accurate. For this reason,
some authors
11
speak of a paradigm change between an expanding universe in
space (in which galaxies are hurled through a static space) and a universe
consisting of expanding space (in which a continual expansion of space is
pulling along galaxies fixed to it).
12
Thus in the expanding space paradigm,

10
This cosmological interpretation is adopted by most cosmology textbooks. In particular, some
authors, for example MTW (1973, p. 776) and Peebles (1993, p. 96) describe cosmological redshift
as an effect of standing waves (generated, for instance, between comoving points in the universe)
independent of: 1) why expansion came about; 2) the rate uniform or nonuniform at which it came
about; 3) the source-receptor distance at emission, at reception, or at any time in between.
11
In primis Harrison 1993; Harrison 2000. See also Baryshev 2005.
12
This paradigm change is evident in the laws stating the expansion of the universe. Indeed,
whereas the so-called velocity-distance law = ( is the Hubble constant) is a linear theoretic
law (derived from the FLRW metric) valid for all distances, the redshift-distance law = is an
empirical approximate relation that can be used only in the limit of small (when compared with the
Hubble distance
= () ) distances. The former is often called Hubble law, but it is only the
latter that was really discovered by Hubble; they are theoretically equivalent only in small redshift
domains (Harrison 1993, Harrison 2000).

the universal expansion, incorporated within a spacetime metric through the
use of a scale factor, manifests itself as a geometric effect.
The sense in which the universe is experiencing a metric expansion can be
seen in mathematical terms by analysing the comoving coordinate separation
of the FLRW metric. The proper radial ( = = 0) distance D, at time t,
between two fundamental reference frames increases with time as does the
scale factor R(t). Consider, indeed, the proper distance (defined to be along a
surface of constant time, = 0 between an observer at = 0 and a distant
cluster whose local frame is at = . The FLRW metric is thus reduced
to = , which, upon integration, becomes = ().
Differentiating with respect to time, one obtains: (
= (
+ ), that is to
say, the total velocity
has two components: the recession velocity
and the peculiar velocity
= , so that
.
The peculiar velocity is the normal physical velocity of a cluster in space: it
expresses the local movement through comoving coordinates. The recession
velocity, instead, is the rate of increase of the metric distance D as a function of
time: it expresses the global movement, as it were, of comoving coordinates,
so that the increasing separation of the inertial frames of the fundamental
observers is referred to as the expansion of space.
13
It is the peculiar velocity
not the total one that must always be less than (or, for light, equal to) c (as SR
claims) because that is the velocity of a material object with respect to its local
reference frame. The total velocity, on the other hand, governed by the rules of
GR (we are generally in curved spacetimes), is given also by the additional
stretching of distances between observers as codified in the recession
velocity, so it can increase without bounds.
14
It goes without saying that

13
The different nature of recession and peculiar velocities is evident in their mathematical
expressions given respectively, in GR and SR, by:
, = (
0
)
1
() [
]
1
, and
= [(1 = )
2
1][(1 + )
2
+ 1]
1
(see Davis and Lineweaver 2003, p. 5). These
formulae give the same results only in the first order of v/c.
14
When = 0, i.e., for fundamental particles (defined as those for which = const.), the total
velocity is given only by the term referred to the expansion of space. In this case, the exact relativistic
theoretical form of the Hubble law (recall footnote 12) is yielded. Indeed,

implies
(), and from = () one has
= (). Thus, whereas in the

Minkowski space of SR, or in other non-expanding solutions, for any spatial separation of objects their
relative peculiar velocity is necessarily less than c, in expanding universes, at distances greater than the
Hubble distance, recession velocities exceed c: if, for instance,
= , then
= () + > .

Doppler and cosmological shifts are evidence, respectively, of peculiar and
recession velocities.

3. THE EXPANSION OF SPACE AS A COORDINATE-DEPENDENT CONCEPT
Some authors
15
counter the expanding space paradigm by arguing that the
observed redshifts are not consequences of some sort of stretching of space
but are classic kinematic Doppler (or, at most, gravitational) shifts due to the
outward effective radial motion of clusters in static space. They maintain that
the interpretation of the expansion of the universe as an expansion of space
itself results from and is merely an artefact of choosing certain particular
coordinatizations of spacetime. Indeed, one can reasonably describe the
redshift as due to the expansion of space only when standard comoving
coordinates are used, namely only when one has decided to establish a
correspondence between fixed spatial coordinates and fundamental observers
(so that the latter can be thought to be at rest). But since using noncomoving
reference frames, observers do not necessarily correspond to constant
coordinates, they argue that we thus have at least two different coordinate
systems, and two equally valid descriptions along with them. One of them (the
former) is surely more convenient than the other (and thus mainly used) but
this does not make the other one (the latter) less true. The profound reason for
this equality resides in that fundamental principle on which GR is based,
general covariance, which roughly states that the form of physical laws under
arbitrary coordinate transformations is invariant, that is to say, all spacetime
coordinate systems are physically equivalent for the description of nature. In
the light of general covariance, many arguments are nothing more than
interpretations of the same physical fact in different mathematical forms.
Accordingly, if the characteristics (for example, the spatial curvature)
pertaining to a given spacetime are non-invariant, i.e., they disappear when
that same spacetime is described by a different coordinate system, then those

15
For instance: Chodorowski 2007, Bunn and Hogg 2008, Cook and Burns 2008.

characteristics do not exist are not real
16
in the physical world
correspondent to that spacetime.
17

One example of this way to proceed is given by Cook and Burns (2008).
They consider a particular instance of an expanding FLRW metric (an empty
universe,
= 0, with negative spatial curvature and = for which a

coordinate transformation leads to a Minkowskian metric which is no longer
expanding (its spatial part does not depend on time), and thus shows that many
of the conclusions usually derived from the initial FLRW expanding metric
the Hubble law, the expansion of space itself, the recession of distant objects
faster than light, the observed redshifts considered as cosmological, the
impossibility to apply SR at large scales, the negative spatial curvature are not
invariant properties but coordinate-dependent effects. Therefore, according to
Cook and Burns, the expansion could also be interpreted as given by peculiar
motions in a static flat space (as in the Milne kinematic model
18
).
The same line of reasoning, but mainly focused on the redshift
interpretation, is pursued by Bunn and Hogg (2008). In a few words, their idea
(based also on Peacock 1999, p. 87; Peacock 2008, p. 4) is that the most
natural interpretation of the redshifts of any cosmic object is as a Doppler
shift. But, since redshifts of distant objects cannot be thought of as global
Doppler shifts
19
, their effects are supposed to be given by the integration of
many infinitesimal Doppler shifts caused by photons passing, along their

16
This view obviously endorses the claim that invariance is at least a necessary condition for
existence.
17
Think about black holes: the Schwarzschild singularity, the radius that makes singular the
Schwarzschild metric (describing the spacetime outside a spherical, non-rotating body), is not a real
singularity of spacetime, but it is a coordinate singularity as it occurs only in some coordinate systems.
18
This universe was proposed by Edward Milne in 1932 to be a Minkowski spacetime described
from an expanding reference frame. He did not agree with the GR dictate that matter and space
(geometry) are linked together. His model of the expansion, based on the so-called kinematic
relativity, is built without the presence of gravity as an initial assumption and is not a dynamical model:
it lives in Minkowski spacetime and may be treated by SR. For Milne, indeed, the big bang is like a real
physical explosion of matter (that does not affect the universes geometry): an infinite number of test
(no mass, no volume) particles shot out radially with all possible speeds (less than c) in all directions, at
a particular unique creation event, into previously empty infinite Minkowski space. Thus he rejected
the expansion of space, insisting instead on expansion through space, hence redshifts are Doppler.
His model, however, suffers many fatal flaws (Rindler 2006, p. 360) and today is used only as a
pedagogical tool.
19
Indeed SR would be violated. Recall that such a redshift is due to the relative peculiar velocity
of emitter and receiver, so that for large distances, the redshift would be given by superluminal
velocities.

paths, between fundamental observers separated by small distances (each
observer, placed on the photons path, measures her infinitesimal Doppler
shift in her local inertial frame in which SR hold). Bunn and Hogg also address
the question of whether it makes any physical sense to interpret the redshift as
one big Doppler shift, rather than the sum of many small ones. Technical
details apart
20
, the major difficulty in this kind of argument as admitted by
Bunn and Hogg themselves regards the necessity, when one wants to read the
redshift of a distant object as a Doppler shift, of taking into account the velocity
of that object (at the instant in which light is emitted) relative to the present of
the observer. This is because relative velocity in curved spacetime is undefined
for widely separated objects.
21
Accordingly, it is not correct to call their
relative motion a motion of one of the objects relative to the other (at most
their relative velocity can be considered a mere coordinate velocity and not
an actual velocity). The only meaningful thing to say is that they are
increasing their separation. Consequently, it seems to me that such a
procedure based on partitioning large redshifts into infinitesimal Doppler ones
is only a mathematical way to relate the special-relativistic Doppler redshift,
holding locally, to the global cosmological redshift; however, this does not
imply that the nature of the cosmological redshift is a Doppler one.
22

About Cook and Burns conclusions regarding, in particular, the nature of
the expansion of space as a reference frame-dependent concept, I have some
criticisms. It is in general true that observations carried out in different
reference frames can lead to very different, but equally correct and physically
equivalent, descriptions of phenomena. However, the SR and the GR
descriptions of universal expansion can be equivalent descriptions at most in
the particular case of a flat spacetime, whereas in curved spacetimes they
coincide only locally. Thus, the only FLRW metric as Cook and Burns
themselves admit for which measurements in an expanding universe can be

20
And noting previously that this interpretation holds mathematically even for 1 (Peacock
2008, p. 4).
21
Indeed velocities can only be compared at the same spacetime point, and the only way to
compare velocities of separated objects is by parallel transporting one velocity four-vector to the
spacetime location of the other, but the result of such parallel transport is path-dependent.
22
In fact, if we imagine to decompose the cosmological redshift phenomenon in its three
fundamental physical processes we find that only two (emission and reception of light in the emitter
and receivers proper reference frames) are special-relativistic phenomena, whereas one (the
propagation of light from emitter to receiver) is a general-relativistic process governed by the law of
geodesic motion in curved spacetime (MTW 1973, p. 776).

directly compared with measurements in a static universe is the SR spacetime.
Even if in such a case the nature of the redshift, depending on the reference
frame, can have an ambiguous status, in the general case (universe not empty)
the cosmological interpretation globally cannot be avoided insofar as the
Doppler one is valid only locally.
23

Furthermore, Cook and Burns consider only a change of coordinates, but a
mere change of coordinates is not enough to obtain a different cosmological
model
24
: To have a cosmological model one has to specify, besides a
spacetime (M, g), a congruence of timelike curves to represent the mean
motion of matter (Ehlers 1990, pp. 29-30). In other words, there is nothing
intrinsic to a geometrically defined spacetime which is able to determine which
is the family of preferred worldlines representing the average motion of matter
at each spacetime point.
25
Coordinate systems and reference frames are not at
all equivalent concepts, and general covariance states that all coordinate
systems, not all reference frames, are equivalent. In general, if a reference
frame can be naturally associated with the actual movement of a system of
bodies (as happens in FLRW models where the comoving frame is naturally
associated to the divergent motions of clusters), the ability to perform a change
of coordinates does not necessarily imply that such an association is still
possible under the new coordinatization. This difficulty, however, does not
imply that we must abandon general covariance; rather, it necessitates a move
from general covariance to a restricted or physical covariance (Ellis and
Matravers 1995, p. 787) in which the coordinates used for physical
applications are well-adapted to the system at hand, and despite the restricted
coordinate choices, such physical studies do indeed make sense (Ellis and
Matravers 1995, p. 778).
26
In the FLRW models, it turns out that the physical

23
Even worse, Davis (2004, p. 64) has showed that the SR Doppler shift formula can be used for
an object in an empty universe only if the velocity in this formula is the velocity in Minkowski space,
namely a velocity which does not obey Hubbles law. This means that the SR Doppler shift equation
does only relate redshift to velocity in the Milne description of the empty universe, but does not relate
redshift to the recession velocity appearing in Hubbles law. And indeed, in the Cook and Burns
derivations, Hubbles law of recession does not hold.
24
At most, the change suggests a different foliation of spacetime into spacelike hypersurfaces,
and thus (possibly) different spatial geometries.
25
The motion of a fundamental particle cannot be deduced directly from the Riemannian metric
as the particle moves along geodesic. Indeed, the worldlines of fundamental particles are geodesics,
but the contrary is not necessarily true (see Infeld and Schild 1945).
26
Ellis (2007, p. 1215) highlights another misconception regarding the fact that a preferred
FLRW frame would contradict relativity theory according to which all reference frames should be

motion of the substratum corresponds to fixed spatial coordinates, so the
interpretation of the redshift as evidence for the expansion of the universe, is
then unique (Ellis et al. 1978, p. 440).
Moreover, this choice of preferred worldlines the character of which is
described by Weyls Principle
27
has an ambiguous conceptual status:
It is neither a law, since it is not contained in the equations of GR for which it
even presupposes the explicit solutions, nor a general principle, since
empirically it can at most be realized on the average. (Pauri 1991, p. 324; his
italics)
In a similar way, Cosmological Principle (CP) which is strictly related to,
though is logically independent of, Weyls Principle (Bergia 1997, p. 187)
has a status that is independent from GR. As already said, indeed, when the CP
is adopted (not only on observational bases
28
) the FLRW metric, and the
characteristics it contains, is uniquely determined. In this way, FLRW models,
relying on these principles, acquire, at least in their foundational aspects, a sort
of independence from GR, so that appealing to coordinate changes and to
general covariance seems to us not the correct way to remove the supposed
spatial nature of the expansion. The expansion of the space, indeed, occurs at
those large cosmic scales which are exactly the areas under the jurisdiction
of both Weyls Principle and CP. Thus, from this point of view,
the expansion would be an effect of the large-scale homogeneity and isotropy
[]. [It] should be viewed as a de facto behaviour of the substratum []
which is merely compatible with the equations of GR. (Pauri 1991, p. 323)
29

In other words, it is only at the largest scales (where special-relativistic
descriptions are inadequate) that the expansion of the universe is, as it were,

equally valid: But this equivalence of frames is true for the equations rather than their solutions.
Almost all particular solutions will have preferred worldlines and surfaces; this is just a particular
example of a broken symmetry the occurrence of solutions of equations with less symmetries than
the equations display.
27
In modern terms it states that the worldlines of clusters form a 3-bundle of non-intersecting
diverging geodesics orthogonal to a series of spacelike hypersurfaces. This claim was stressed for the
first time by Weyl, in 1923, in the original de Sitters universe.
28
See Fano and Macchia 2008, sect. 6.
29
Actually, Pauri opts for a subtle relational view of the expansion. Indeed, if one lays at the
foundations of cosmology Weyls Principle (even before of the CP), it results adopting a particular
mathematical construction principally due to Ehlers, Pirani and Schild that the spacetime itself is
relationally constituted by fundamental particles. However, I will analyse this view on another
occasion.

the expansion of space as resulting from the special way in which space and
time are combined in FLRW universes, so that the expanding space paradigm
is a legitimate global concept (Peacock 2008, p. 1) as Peacock sums up, and
this is most clear-cut in the case of closed universes, where the total volume is
a well-defined quantity that increases with time, so undoubtedly space is
expanding in that case (Peacock 2008, p.1).

4. HIGH REDSHIFTS ARE NECESSARILY COSMOLOGICAL
Most cosmologists interpret distant redshifts as cosmological redshifts,
sustaining, more or less tacitly, the idea that this kind of redshift due to the
universal expansion is a physical phenomenon really different from the
Doppler effect (classical and special-relativistic)
30
, or even that the physics of
space expansion is different from motion in static space.
31
In this paragraph, I
want to analyse briefly the principal arguments promoting the GR
interpretation of cosmological redshift.
32

Firstly, Prokhovnik (1985, p. 25) reminds us that this interpretation is
strongly supported by the accepted astronomical and physical evidence
regarding the behaviour of light which is affected by gravitational fields.
Moreover, as I have already said, general-relativistic expansion of space can
explain superluminal recession velocities of distant objects and distances to the
particle horizon greater than ct (where t is the age of the universe).
33
It is
important to remark that these faster than light motions imply no violation of
SR in that their relative velocities are not local velocity differences, i.e., these

30
For instance: Harrison 2000, MTW 1976, Rindler 2006.
31
For instance: Abramowicz 2008, Abramowicz et al. 2007, Baryshev 2005.
32
It should be noted that other explanations, not operating on Doppler or cosmological
premises, have been attempted. Historically, the first one is the so-called tired-light theory: it
hypothesizes that photons might lose bits and pieces of their energy while traveling across vast regions
of extragalactic space. Many physical reasons for this suffering fatigue by light have been proposed,
but no one has proved to be satisfactory. Another explanation of distant redshifts, concerning some
particular quasars, attributes the cause of their so-called intrinsic redshifts to unknown internal
mechanisms of quasars themselves. Neither this last hypothesis has never gained significant support in
the astronomy community.
33
The particle horizon marks the size of the observable universe. Since the expansion of space
provides an additional stretching of the distances, the size of the entire universe could be bigger than
the size of the observable one. This is not possible in SR where the largest distances, since nothing
travels faster than light, cannot exceed the observable universe given by the age of the universe
multiplied by c.

motions occur outside the observers inertial frames and no information is
transferred between them.
34
On the other hand, a kinematical Doppler view,
necessarily based on the extension of a Minkowski frame into the expanding
universe, is not able to explain the absurdity of light traveling faster than
light because it does not see any difference between peculiar and recession
velocities.
From a theoretical viewpoint, a closer look at Lematres equation reveals
that the cosmological redshift cannot result from a Doppler shift due to a
source receding with the Hubble velocity = . Indeed, if
= 0 and
= 0, that is, if the universe were not expanding at the times of

emission and reception, the Doppler redshift would be null, whereas
cosmological redshift would be positive (obviously provided that
> (
), namely, that some expansion had occurred during the

intervening time). More specifically, if objects in Minkowski space are at rest
with respect to an observer then they have zero redshift, but this does not
happen in GR for
= 0 does not imply
= 0.
35
Surprisingly, this fact
holds also in empty expanding FLRW universes (without cosmological
constant), whereas we might have expected (incorrectly) that these universes
could have been well-described by SR in flat Minkowski spacetime. Davis et al.
(2003), indeed, derive that in such an empty FLRW universe, a galaxy with
v
tot
=0 will be blueshifted, concluding that, in general:
The fact that approaching galaxies can be redshifted and receding galaxies can
be blueshifted is an interesting illustration of the fact that cosmological
redshifts are not Doppler shifts. (Davis et al. 2003, p. 362)
36

This shows that even an empty FLRW universe does not trivially reduce to
Minkowski spacetime. All the more, one must use GR in universes with matter
and energy because spacetime is curved and there is a cosmic gravitational
field.

From an observational viewpoint, these two wavelength shifts reveal
noteworthy differences. Cosmological redshift is a global phenomenon: it

34
See Davis and Lineweaver 2003, pp. 5-6; Ellis 2007, p. 1215.
35
Zero velocity approximately corresponds to zero redshift only for
0.3 (Davis et al.

2003, p. 362). Note that the total redshift is given by: 1 + =
= 1 +
(1 +
), with
the cosmological redshift,
the Doppler shift caused by the local peculiar motion of the object
observed (Ellis 2007, p. 1197).
36
See also Grn-Elgary 2006, pp. 12-13, showing similarly that an object at large distance and
at rest relative to the observer has 0.

affects all spectral lines alike of a given cluster, and we measure redshifts for all
clusters whose movements are therefore always in a radial direction. This
cannot be evidence of simple peculiar motions because these motions tend to
be in all directions (including towards us) so that their effects should become
on average null for a large number of clusters at a certain distance.
37
In
addition, peculiar velocities should be smaller than a certain limit (of the order
of a few hundred kilometers per second), while those recorded dreadfully
increase with distance. Thus, the importance of peculiar motions, whose
redshifts or blueshifts contribute to the total redshift of a certain object,
decreases with distance: closer objects (like the members of the Local Group)
are dominated by peculiar motions
38
, but, as distance increases, dilatation of
space enormously exceeds peculiar motions. According to Hawley and
Holcomb:
The systematic increase of redshift with distance is the strongest argument that
the cosmological redshift is truly cosmological []. [It] is due to the properties
of space itself. (Hawley and Holcomb 2005, p. 294)
Here are three particular pieces of evidence supporting cosmological
redshifts and thus the expanding space hypothesis. The first one is exposed by
Davis (2004, p. 28), who analyses recent data concerning the Type Ia
supernovae magnitude-redshift relation concluding: SR fails this
observational test dramatically (Davis 2004, p. 33).
39
The second one is
about the temperature of the cosmic microwave background measured at
redshift z: it is observed (Srianand, Petitjean and Ledoux 2000) that it is given
by (1+z)T
0
(where T
0
is its present value), just like all GR Friedmann models
predict, whereas it would be T
0
if only matter were expanding. The last one,
provided by Misner et al. (1973, p. 767), concerns quasar redshifts which are
so high that neither a gravitational nor a Doppler origin are acceptable (in the
former case, quasars with high gravitational redshifts would be unstable against
collapse; in the latter, quasars with high accelerations would be disrupted).

37
Obviously, we are not referring to expanding peculiar motions la Milne.
38
On small scales, indeed, we measure blueshifts Doppler too; for example, the Andromeda
Galaxys one, due to its peculiar approaching motion towards Milky Way.
39
This relation can be used to test the validity of different models concerning the predictions of
some cosmological parameters. Roughly, the relation indicates that the model using a special-
relativistic redshift does not fit the observational data as well as the model using cosmological redshift.

5. A PHYSICAL GLANCE AT THE EXPANSION
The universe is expanding because it is in a dynamical state determined by its
initial conditions. The expansion of space manifests itself at large scales and
does not necessarily imply expansion into something; thus it does not occur
into a previously existing larger entity (an edge in the universe would be a
special location and this is not permitted by CP), but is rather an expansion of
the universe as a whole and one that takes place without any centre: the
universe is just getting bigger, while always remaining all that is (Ellis 2007,
p. 1214). Hence its volume is increasing
40
, and more and more space seems to
appear. But what does it mean? Is it an actual incessant creation of space, that
is a kind of production of a larger spatiotemporal container added with
vacuum?
41
Is it a sort of stretching of an infinitely elastic substance that
extends its points? These are very fascinating but risky questions: To
speak of the creation of space is a bad way of speaking []. The right way of
speaking is to speak of a dynamic geometry (Misner et al. 1973, p. 740). This
is because we only know that all distances in the 3-dimensional spatial
hypersurfaces {t=const.} of FLRW universes scale as (), all areas as
2
(),
and all volumes as
3
() thus each comoving finite cosmic box
unceasingly increases its volume and that the spacetime itself evolves, for its
curvature is given by two varying contributions: the curvature of the 3-spaces
(which varies as
2
()), and their expansion through time which affects the
density of matter and consequently part of the curvature of spacetime itself.
To figure out where this expansion might take place, if anywhere, is
obviously not easy, and may not even be possible: we cannot extrapolate our
familiar ideas about the growth of things placed in certain regions of space,
because they always expand in a bigger spatial container. Even the usual
simplifying example of an inflating sphere in a 3-dimensional space is not
completely analogous, because the 4-dimensional curved metric of the
universe does not necessarily expand in a 5-dimensional flat physically real
container.
42
At first sight, it seems plausible think that, if something grows,
it must grow in something bigger, so, in order to give a meaning to the

40
Even if the universe is spatially infinite, the concept of an increasing volume locally still makes
sense.
41
For example, Baryshev (2005, p. 5) looks at the creation of space as a new cosmological
phenomenon.
42
Even worse, perturbations need spacetimes of still more dimensions (Ellis 2007, p. 1215).

operation of, for instance, doubling the size of space, some higher dimensional
space that contains it must be supposed. Such a thesis is sustained by Nerlich
(1991), who says: to have a decent syntax, a well-formed description of such an
operation (or better, to give it even something stronger: a semantic meaning),
we need this further wider spatial container. However, the latter plays no
role in GR, so we ought to deem the doubling-of-space sentence meaningless,
though open to acquiring a meaning (Nerlich 1991, p. 187). Agazzi couches
a similar consideration:
When one says that the expansion of the universe is not to be conceived like a
process that occurs in space, but it is rather space itself that expands [this
expression] is evidently not so much intelligible if one does not refer to a space
relative to which it is possible to speak of expansion of space when regarded as
a physical magnitude in a given theory. (Agazzi 2006, p. 2362; my translation)
I do not agree with Nerlich and Agazzi on the necessity, at least from a
mathematical-physical point of view, of this further wider container. The
reason is simple. For we can distinguish between intrinsic and extrinsic
curvature namely, a curvature intrinsic to space and not dependent on some
higher dimensional containing space, and a curvature embedded in a higher
dimensional space
43
, respectively the expansion of space can be regarded as
intrinsic phenomena too: doubling the radius of curvature, i.e., halving the
curvature, or expanding a certain volume of space are possible operations
having nothing to do with any higher dimension space.
44
On the other hand,
Nerlich himself highlights this objection, judging it meaningful. Therefore, the
dynamical spacetime of our cosmos could expand (and shrink and curve)
without being embedded in a higher-dimensional one: the universe could be
self-contained. From this point of view a question like What does the universe
expand into? does not necessary make sense simply because a bigger spatial
container is not an essential requirement to the expansion.
45

A way to overcome this conceptual impasse is suggested by Schutz (2003).
Even if cosmology is the science of the universe as a whole, he invites us to

43
For instance: a curved line necessitates a plane, a curved plane a volume, and so on.
44
If some characteristics, as curvature, can only exist in conjunction with extrinsic
characteristics, or, contrariwise, curved spaces can exist also intrinsically without being embedded in
higher dimensions, is an issue obviously not solvable by appealing to human imagination. See Ross
1999 on the ontological axioms covering the possible spaces that Euclidean and non-Euclidean
geometries can describe.
45
However, about where, or better how, space could grow see Ohanian 2000, p. 690.

regard its expansion as a local property (a phenomenon pertaining each of its
single parts), not a global one: The Hubble expansion [] does not define the
global structure of our rubber-band universe []. It only tells us how it
stretches, locally (Schutz 2003, p. 349); in this way,
if the distance between two typical galaxies doubles over some period of time,
then the size of the cosmology has effectively doubled. These relative size
changes are the important aspects of cosmological expansion, not the overall
size of the universe. (Schutz 2003, p. 363)

6. EXPANSION OF THE UNIVERSE AND DYNAMICS OF LOCAL SYSTEMS
Rejecting the paradigm of expanding space from the observation that space
locally, at small scales (galaxy, solar system, house, atom), shows no sign of
expansion, is not pertinent. Cosmology considers only the largest scales, and
solutions to the Einstein equations, obtained by assuming some averaged
matter-energy distribution, correspond to the behaviour of the overall
gravitational field of the universe. At scales smaller than 100 Mpcs, the CP
does not hold (even approximately), i.e., the symmetries of the FLRW metric
do not match the matter distribution in the universe. The FLRW metric is only
a large scale approximation, thus the expansion of space is global but not
universal (Francis et al. 2007, p. 7).
In fact, the geometry of spacetime inside a galaxy, or near a planet, is
dominated by the curvature produced by local masses, so that there is no global
expansion for these objects to oppose, since locally the dynamics of spacetime
has already been modified (given that the local gravitational field of these
masses is stronger than that of the universe). In terms of classical physics,
inside a galaxy, the forces holding atoms and molecules together have
decoupled their constituents from the general expansion; the gravity that holds
the stars in a galaxy together has decoupled them from the expansion (Rindler
2006, p. 353), so that the local situation in the universe is quite analogous to
the case of the Schwarzschild metric rather than to a FLRW one, and the
planetary orbits are unaffected, as Birkhoffs theorem states, by the existence
of expanding surrounding mass shells.
46

46
Remind that Birkhoffs theorem roughly states: in order to calculate the motion of a certain
galaxy A relative to a given galaxy B, it is only necessary to take into account the mass contained within
the sphere around B passing through A (the effects of all matter outside that sphere are negligible).

Accordingly, the space in a house, for instance, does not expand insofar as
its walls, held together by electromagnetic forces, do not follow geodesics, and
the distribution of matter is not at all uniform but has collapsed so that, as
already claimed, the geometry of spacetime is completely different from the
FLRW metric. Davis and Lineweaver speak of coherent objects: galaxies,
cities, atoms are objects whose size has been set by a compromise among
forces (Davis and Lineweaver 2005, p. 44). On the contrary, photons are not
coherent objects, and thus their wavelengths expand with the space. Expansion
by itself they go on to state regarded as a coasting expansion neither
accelerating nor decelerating, produces no forces: it is only a changing rate of
the expansion that adds a new force to the cosmic objects, but even this new
force does not make them expand or contract. In our universe, the accelerating
expansion exerts a tiny outward force on bound bodies, thus the equilibrium
among forces is reached at a slightly larger size, and consequently they are
slightly larger than they would be in a non-accelerating universe. But what
happens if the acceleration itself is not constant but increases? Some
cosmologists think that an acceleration growing strong enough to tear apart all
objects structures could eventually lead to a kind of big rip, not caused,
however, by expansion or acceleration per se, but by an accelerating
acceleration.
However, since the Einstein and Straus pioneering paper on the Solar
System (Einstein and Straus 1945), showing that the Solar System is
completely immune from the cosmological expansion, the problem regarding
the effects of cosmic expansion on local non-comoving systems has been
analysed many times without unanimous agreement among scholars: some
think that the tendency to expand is merely negligible in practice, others that it
is completely non-existent.
47
In any case, the large scale phenomenon of the
expanding space is not invalidated by such a local uncertainty.

However, Cooperstock et al. 1998, p. 3, point out that this analysis holds only for a spherical cavity
embedded in a FLRW universe, but if spherical symmetry is absent satisfactory quantitative
calculations are missing in the literature.
47
For instance, recent supporters of the former view are Cooperstock et al. (1998), Davis et al.
(2003), of the latter Peacock (2006). For a rich bibliography see Carrera and Giulini 2006. A classic
analysis is the so-called tethered galaxy problem, in which a galaxy is imagined to be tethered to the
Milky Way so that their distance is constant; the point is to understand what happens when the tether
is cut: does the galaxy join up with the Hubble flow, starting to recede with the expansion of the
universe? In general, however, it is important to note that many characteristics of this issue do not
depend just on the simple expansion of the universe but on its acceleration (or deceleration).

As regards the redshifts, the different interpretations of these at small
versus large scales is physically motivated by the different physical states of the
emitting objects: in the first case, objects within bound systems are not
participating in the universal expansion, so SR applies and their redshifts are
Doppler, in the second one, such objects are participating, so FLRW and GR
apply, and redshifts are cosmological.

7. A PHILOSOPHICAL GLANCE AT THE EXPANSION
Is it possible to relate, by means of the redshift interpretations, the expansion
of the universe to the nature of spacetime? It seems rather metaphysical to
argue whether (on the one hand) two points are actually moving apart, or (on
the other) the space between them itself is growing (Whiting 2004, p. 4). It is
true: it seems rather metaphysical, but, hopefully, not completely.
48

7.1 SUBSTANTIVALISM AND RELATIONISM
The two principal philosophical views concerning the nature of spacetime are
spacetime substantivalism and spacetime relationism. The former roughly
maintains that spacetime is a sort of thing, a container which, though
different and existing independently of its physical contents (material things
and energy-forms), is in some sense just as substantial and real.
49
In particular,
for substantivalists, unobservable spatial and temporal properties of matter

48
Among other things, it seems that some experiments could directly reveal which situation is
real. In F. Melchiorri and B. Melchiorris opinion, there could be, in principle, an experiment
concerning the motion of the Earth around the Sun and measurements on the dipole anisotropy of the
cosmic microwave background radiation, that could discriminate between peculiar motions of clusters
and the expansion of space (see Melchiorri and Melchiorri 1994). Morgan (1988) is of the same
opinion. However, a philosophical analysis of the universal expansion has, strangely, never been a
tasty subject. The only author, at least to my knowledge, to have dealt with it is Whitrow (1980, p.
288-294). His conclusions are different from mine, but unfortunately I do not have here enough
space to unfold and criticize them.
49
Two main kinds of substantivalism are recognized: manifold substantivalism, which considers
spacetime represented by the manifold of events (the bare set of points with a topological and a
differential structure); metric field substantivalism, which identifies spacetime with the gravitational-
metric field. One of the matter in dispute regards just the nature of this field: is it a physical field
containing a form of energy (for instance, gravitational waves) and hence must be considered as part of
the contents, or is it the container insofar as its spatiotemporal properties (it determines the spacelike-
timelike distinction, the affine connection of spacetime, the distances between points) cannot be
expunged from a meaningful notion of spacetime? See Earman and Nortons 1987 classic paper.

(e.g., is at position x) are not reducible to observable relational properties of
matter (e.g., coincidence, betweenness) (Earman and Norton 1987, p. 515).
The latter is regarded as a denial of the main substantivalist thesis: the world is
constituted by its actual material objects and physical events, and spacetime is
viewed as a mere abstraction instantiated by their spatial and temporal relations
(only the contents exist, not the container). No ontological commitment
to spacetime points (or spatial points and temporal instants) is claimed.
50

Now, before looking at the cosmos through the dictates of substantivalism
and relationism, in order to better understand what cosmological expansion,
interpreted by the expanding space paradigm, is, it will be useful to briefly
analyse the expansion in the light of two philosophical approaches Nerlichs
detachment thesis and the nocturnal doubling thought experiment showing
essentially what cosmological expansion is not.

7.2 NERLICHS DETACHMENT THESIS
According to Nerlich, the distinction between flat and curved spaces may have
a significance for the reality of space, so this distinction has a great bearing on
the debate between substantivalists and relationists. His pro-substantivalist
argument is essentially based on the tangible manifestations of spatial
curvature.
Substantivalists sustain that objects inherit their spatial properties from the
regions of space that they occupy; thus space is a sort of intermediary between
objects, in the sense that their spatial relations are instantiated by a space
conceived of as a material entity in its own right. Relationists, on the contrary,
claim that objects possess distinctive relational properties, i.e., objects are
directly related to one another by spatial relations regarded as distances
separating them, and space is a mere invisible redundant entity. In brief,
substantivalism appeals to space-thing relations (position in space; the quantity
of space filled or occupied by a material object), whereas relationism appeals to
thing-thing relations (spatial distance between things). Since relationism needs
a sort of innocuousness of space, it must seek to cut the substantivalist tie
between the spatial relations among things and the mediation of space. In other
words, relationism has to appeal to the so-called detachment thesis: Thing-

50
In the rest of the paper I take into account only this traditional view of relationism. It is not
possible here to extend my analysis to other modern forms.

thing spatial relations are logically independent of thing-space relations
(Nerlich 1991, p. 172).
In conditional form it then states: if one changes thing-space relations,
there are no consequential changes for thing-thing relations. Leibnizs famous
argument, for example, relies on this thesis: if the physical universe as a whole
were shifted some arbitrary distance in a certain direction with respect to
space, all thing-space relations would change whereas all thing-thing relations
(i.e., spatial relations empirically detectable) would remain the same. Since all
thing-space spatial relations are idle, space, i.e., the relatum of these relations,
does not manifest consequences, and thus it is idle too and its existence is not
necessary. Thus relationism has no point without the detachment thesis
(Nerlich 1991, p. 189) insofar as such a thesis permits, or even requires, in a
way, the reduction of space to thing-thing spatial relations. It is important to
note that at the core of this argument lies the fact that thing-thing spatial
relations are considered privileged properties because they are observable,
whereas thing-space relations are considered as inconsequential properties
because they can change without any accompanying change in specifiably
privileged properties (Nerlich 1991, p. 171); in this way inconsequential
properties cannot be considered real properties.

7.3 NOCTURNAL DOUBLING
Another thought experiment (a variant of Leibnizs shift argument) based, at
least in a relationist perspective, on the detachment thesis, is the so-called
nocturnal doubling (also proposed by Henri Poincar). It proposes that if
everything (objects, distances, etc.) were to double in size overnight, there
would be no real difference because everything would still be related just as it
was to everything else. (You would wake up in your bed doubled in size, but
also you yourself would be twice your previous size, and so would be your
room, your house, and your town; the distances between all these things, and
even the laws of nature, would have been altered to conceal the doubling). But
this is true thanks to a tacit assumption: as Nerlich (1991) shows, only a
doubling and, in the same manner, a Leibniz shift happening in a Euclidean
space has no discernible consequences. In a non-Euclidean space, on the other
hand, the doubling would be different: displacements of things could yield
significant differences in their shapes if the curvature of space varies from point
to point. Imagine, for example, a 2-dimensional undulate surface containing
valleys, mountains and plains (i.e., places of negative, positive and zero

curvature, respectively). Now, the doubling experiment depends on the place
we start it: we obtain different results in the size and shape of objects if they get
pushed from a valley on to a plain, or from a mountain top into a saddle-back.
This fact would cause many problems to a relationist, because in such a
context the variations in thing-space relations produce manifest variations in
observable thing-thing relations. Therefore, thing-space relations are directly
consequential and thing-thing relations are not logically independent of thing-
space relations as detachment thesis claims. If thing-thing spatial relations
changes are real, so are real the properties (the thing-space spatial relations)
that induce those changes (the thesis stating that consequential properties are
real properties is called by Nerlich Discernibility Principle, DP), accordingly
must be real the entity possessing the consequential properties: the space.
From a relationist point of view, the only way to apparently weaken
Nerlichs criticism is to consider a complete nocturnal doubling, that is, a
thought experiment that doubles strictly everything, space included, thereby a
doubling that does not happen in space, but with space. In this case, as in a
Euclidean space, there is no change at all in thing-thing spatial relations
because they vary in correspondence with variations of thing-space relations
(objects do not get pushed, for example, from a valley on to a plain, but it is the
valley itself that is doubled in size correspondently with the objects contained
in it). Obviously, such a weakening is only apparent because the pro-
substantivalism move is actually still stronger because the entire argument
relies on an ontological commitment to space itself.
I have made this long detour in order to point out how cosmological
expansion, at least in the expanding space paradigm, is different from both
nocturnal and complete doubling. In fact, in the expanding universe the
doubling (which can be considered as a sort of temporary stage of a continuous
expansion) refers not to the size of every ordinary object (as happens in the
nocturnal doubling both in Euclidean and non-Euclidean spaces), but only to
the size of configurations of large scale cosmological objects (their respective
distances like, for example, a cluster triangle) when is compared to the size of
local reference standards. This is the crucial difference that makes the
expansion of the universe an observable phenomenon and gives it a physical
meaning. We verify empirically that these detectable reference standards
namely, the local gravitationally-bound systems, such as galaxies, double stars,
planetary systems, or even atomic standards do not expand, and thereby
deduce that the expansion is not universal as the complete nocturnal doubling

requires. In brief, universal metric doubling happens neither in space (as
nocturnal doubling) nor with space (as complete doubling), but, as often
repeated, is a phenomenon of space.
More generally, the detachment thesis argument itself is not strictly
applicable to the cosmological expansion case because there is no change in
thing-space relations: clusters remain embedded in the same (approximately)
point of space, and so thing-space relations are neither inconsequential nor
consequential but simply silent. On the other hand, the variation itself of
thing-thing spatial relations
51
depending on the expansion of space cannot be
detached from dependence on the space itself. This is evident, if one accepts
the cosmological interpretation of distant redshifts, by the changes in thing-
thing relations (distances among clusters) appearing to be mediated by paths
(intervals or stretches) of physical space insofar as wavelengths stretchings
instantiate a sort of thing-space relations. Indeed, if the stretching can be
considered as a kind of consequential property owned by a path because of its
action on electromagnetic waves, then it is a real property as Nerlichs DP
claims, so that space itself in a way the totality of all paths becomes a real
entity necessary both to explain and predict the detectable behaviour of cosmic
objects and light. Accordingly, compared with the following Nerlichs thought,
it seems to us that something more, i.e., a kind of action of space, is at stake
in our cosmological context:
The DP together with the nature of non-Euclidean geometry suggests that the
hypothesis of space can do genuine explanatory work (even though it makes no
appeal to the action of space). (Nerlich 1991, p. 188)

8. HOW EXPANSION IMPLIES A METRIC FIELD SUBSTANTIVALISM
In the last paragraph I argued for the necessary existence of stretching paths in
order to explain the cosmological redshift, and this has plain consequences
52

for our aforementioned ontologies. Substantivalists, indeed, naturally require,
in order to have spatial relatedness among things, the existence of paths
connecting these things, whereas relationists, speaking of spatial relations,

51
Note that here the situation is almost the opposite of Leibnizs shift argument: all thing-space
relations remain the same while thing-thing relations at large scales change.
52
Remind: As ontologists, we should be no less worried by the nature of the parts of space
(volumes, paths, points) than by the nature of the whole (Nerlich 1991, p. 179).

refer to distances separating objects, namely to relations operating, as it were,
across space but not through it (Dainton 2001, p. 145): relations are held to
connect spatially separated objects directly, without passing, or extending,
through the medium of intervening empty space.
53
In the case of cosmological
redshift, how wavelength is stretched claims a relation between emitter and
observer operating just through space, not across it: light is redshifted just
because it clears a path through an expanding metric that, point by point,
influences its wavelength. And this way to operate is unacceptable for
relationists.
54

In this way, the cosmological interpretation of redshift is like an epistemic
access to the ontology of spacetime, which must exist as a substantial entity
since it is provided with a property causally efficacious with respect to some
events involving matter (Hinckfuss 1975, p. 141). However, this causal
efficacy, despite being the necessary condition by which we may discover, as
Nerlich (1994a, p. 178) affirms, that space has a certain property, is truly a
slippery concept, above all in questions pertaining spacetime. Due to this fact,
it is better to specify that this concept means only that the explanation of the
changing shape of photons wavelengths can be regarded as causal, not that
space acts as an expanding force. Indeed, in the FLRW metric, clusters are free
fall particles, i.e., they are following geodesics in curved spacetime and have
zero acceleration: clusters worldlines merely extend, diverge and endure.
From this point of view no causal efficacy of space is at stake. In other words,
cosmological spacetime satisfies, in the behaviour of freely falling particles,
that twofold explanatory role recognized by Nerlich:
It explains (familiarly enough) how the apparent gravitational dynamics of free-
fall particles in general frames of reference vanishes into the mere kinematics of

53
Nerlich again is exhaustive: One reason for taking space as a real thing is the strongly intuitive
belief that there can be no basic, simple, binary spatial relations. Just such relations are the foundation
of relationism, so long as its basic spatial facts lie in spatial relations among objects or occupied points
of space or spacetime. Consider the familiar (though not quite basic) relation x is at a distance from y.
There is a strong and familiar intuition that this can be satisfied by a pair of objects only if they are
connected by a path. Equivalently, if one thing is at a distance from another then there is somewhere
half way between them. Distances are infinitely divisible, whether the intervening distance is physically
occupied or whether the space is empty. [] We dont understand how spatial relations can hold
unmediated (Nerlich 1994b, p. 19).
54
As previously underlined, I am referring only to the most traditional relationist view according
to which space is a sort of abstract web of relations between actual objects. However, as Chris Smeenk
has pointed out (private conversation), other more slippery forms of relationism, accepting also
relations between possible objects, should be analysed.

geodesics in flat or curved spacetimes. It explains also by citing identities in
various ways. E.g., the deviation of geodesics is not caused by spacetime
curvature: it is spacetime curvature. (Nerlich 2008, p. 2)
55

Therefore, we have clusters whose gravitational dynamics is subsumed into the
FLRW geodesics, and we have the identity between their diverging worldlines
and spacetime curvature. However, in our cosmological case we also have the
stronger necessity to explain where the energy lost by redshifted photons is.
In fact, the worldline of a photon is null geodesic, but its energy, in the
cosmological redshift case, is not constant along its worldline
56
, and this
physical fact is motivated by is necessarily ascribed to an entity that causes
the increase of its associated wavelength. And such an entity is not only a
theoretical term. It is an unobservable thing that acts directly producing an
observable effect.
57
And even if we want to be more cautious and remain
silent about causation, we can make sense of certain counterfactuals like this: if
this path had not been stretching, there would not have been that loss of
energy.
58
As in the case of gravitational waves, in the cosmological redshift case
it is of relevance that something physical exists between the times of physical

55
On the subtle meaning of the concept of geodesic see also DiSalle 1995, p. 327 and Nerlich
1991, p. 177.
56
The question regarding the loss of energy of photons in an expanding space, and a possible
non-conservation of energy on cosmic scale, is debated. Harrison (1995; 2000) maintains that energy
in expanding, spatially unbounded, homogeneous and isotropic universe (conforming to the FLRW
metric) is not conserved. He notes that, whereas in the Doppler and gravitational redshifts in spatially
bounded systems the lost energy is manifest in identifiable alternative forms (thus it is conserved),
in the case of cosmological redshifts this does not occur, neither if we take into account the possibility
that the lost energy of photons transforms into metric disturbances, namely in deformations of the
FLRW metric. Also in this case, indeed, the propagating deformations would lose energy because of
the cosmological redshift and the question would result the same: what happens to the lost energy of
the gravitational waves? The violation of conservation laws in expanding space is sustained also by
Baryshev (2005). Carlip and Scranton (1999, p. 8) remark that the electromagnetic energy of the
cosmic background radiation is not conserved during expansion, but there is nothing particularly
cosmological about this loss a photon rising in a static gravitational potential experiences a similar
energy loss. In the energy accounting, they conclude, one has to include gravitational potential
energy, but in GR is difficult to define a local gravitational energy density. The point is indeed that
energy conservation is only a good local concept: there is no general global energy conservation law in
GR (Peebles 1993, p. 139).
57
Electromagnetic waves are unobservable but the physical effect of their stretching is revealed
on our spectrometers.
58
Also in this weaker formulation, space assumes an important role in explaining why
wavelengths are stretched. Mellor (1980, p. 287) maintains that the cause-effect relationship,
mediated by spacetime, should be expressed by counterfactuals.

phenomena of emission and reception: in the former case, energy released
from a certain star needs an entity that possesses it to transfer to a detector on
Earth (this entity is a region of metric curvature that propagates at c), in the
latter, the same medium, this time revealed by its stretching properties, is
needed to explain where this lost energy goes: it is transferred to the
gravitational field (Pitts 2004).
59
Such a change in the energy-momentum of
individual photons dictated by the evolving FLRW metric is an example of how
spacetime geometry redefines matter (Sumner and Sumner 2007, p. 2).
Therefore, both in the cosmological redshift and in the gravitational wave
cases, a substantival metric field, i.e., a structure that can carry and store
energy, is needed for the explanation of these cosmological phenomena. In
such a way a substantivalist ontology for the spacetime emerges.
60

Furthermore, thinking about the expansion of the universe, the fact of the
matter naturally highlighted by the expanding space paradigm is that, whereas
the peculiar velocity must make reference to a material object (including
photons), the recession velocity
should not be regarded as the property of a source; rather, it should be
considered as the property of the point of space in question, whether that point
happens to be occupied by a source, a passing photon, or nothing at all. (Kiang
2003, p. 12)
61

This is a further typical substantivalist commitment to the ontology of
spacetime (and of its points), completely different from the relationist claim
committed to reduce all spatiotemporal properties to properties of objects, that
is to say, in our case, to motions, through space, in which velocities are to be
regarded as belonging to cosmic objects.
62

59
The status of gravitational waves is debated among philosophers. For instance, Hoefer 2000
maintains that we do not have sufficient reasons to accept that the metric field can possess genuine
energy because gravitational energy is not clearly localizable; thus, it behaves differently from
material energy. See also Baker 2005 for a reply to Hoefer and a clear explanation of the difficulties
that relationists have to face in order to explain gravitational waves emitted from a binary star system.
60
Note that I am not considering the metric field as a field in spacetime but as the spacetime: if
such a field and its properties (see footnote 49) were removed, spacetime could not be imagined to
exist. The properties possessed by the mere manifold of points, indeed, are only dimension and
topological structure, and it does not even distinguish between time and spatial dimensions. See
Hoefer 1996 for an elucidation of these points.
61
For a similar analysis see also Prokhovnik 1985, p. 73. He assigns at every point of space a
vectorial velocity so that the expanding universe can be represented as a velocity space.
62
Speaking of temporal becoming, also Dorato recognizes a form of spacetime substantivalism
in contemporary cosmologies: The expansion of the universe is in some sense the expansion of

9. IS A RELATIONAL READING OF EXPANDING UNIVERSE POSSIBLE?
For relationists, even though objects can change their distance relations with
one another, at any given time only objects and spatial relations exist, and we
can imagine drawing up representations, or maps, that reflect all these different
possible dispositions of objects. Therefore, all propositions about the
distances between objects, or about their sizes, are not false, or meaningless
(relationists, indeed, only claim that spatial facts do not require or involve a
spatial substance). By seeing things from this perspective, could a relationist
say that the increase in cosmological distances is only described by the
framework of expanding space, that is, this latter is only a useful
representation, and one that does not fit a real thing?
In my opinion this is not possible. Let us look at the differing relationist
and substantivalist conceptions of movement. The substantivalist view roughly
sustains that an object moves if and only if it occupies different spatial locations
at different times, whereas for the relationist view a body moves if and only if its
distance relations with other material objects change (obviously, also a
substantivalist recognizes that objects move relative to one another, but the
change of their distance relations is not a necessary condition for the
movement). Now, in the global comoving coordinate system, clusters occupy
the same locations at different times, but their distance relations increase so
that relationists must say that clusters are moving, whereas their spatial
coordinates do not change; the only way to overcome this contradiction is by
regarding the comoving coordinate system as a mere mathematical
representative tool. However, in doing so, relationists have to adopt a global
Minkowskian frame that is not able to account for the high redshifts and related
problems seen so far. On the contrary, substantivalists naturally can both say
that clusters are really at rest because they do not occupy different spatial

spacetime itself, since galaxies clearly do not expand in a pre-existing spacetime (Dorato 2006, p.
565; his italics). A substantivalist conclusion about the universal expansion is reached by Baker 2005
too, who charges the substantiality of spacetime to the cosmological constant insofar as it provides
an amount of curvature not entirely created by matter. I agree with his reasoning, however I think that
usually related, though its physical status still remains unclear, to an accelerating expansion is
not necessary for a substantivalist commitment. Indeed, such a commitment is already satisfied, as
hopefully shown so far, at the more fundamental level of the expansion itself as deduced from the
symmetries of the FLRW metric, without involving the dynamics of the Friedmann models (remind:
the cosmological redshift is independent of the way the universe expands quickly, slowly, with jerks
i.e., of the evolution of R(t): its rate of expansion R and acceleration R ).

locations (in fact the change of distance relations is not a binding assumption
for the movement) and explain why distance relations are changing.
Furthermore, it is worth noting that even if a picture of clusters moving in a
static space were possible, it would not necessarily contradict a substantivalist
ontology. Consequently, whereas cosmological shift implies solely
substantivalism, a Doppler shift interpretation would not uniquely sustain
relationism. In any case, trying to obtain a relationist reduction of the
cosmological redshift phenomenon induces us to look for physical
explanations chiefly based on clusters displacement in space, that is, to look
inevitably for special-relativistic (for instance Milnian) analogues to the
general-relativistic expansion and to its related phenomena. Nonetheless,
special-relativistic views undeniably break down outside of a local domain.
A last possible route for relationists might be to accept the standard
interpretation of the cosmological redshift and in the meantime deny that
metric field constitutes the spacetime itself: the metric tensor, incorporating
the gravitational field and carrying energy and momentum, should be
considered as a matter field, namely as part of the contents of spacetime.
63

Thus the stretching action would not be a property of spacetime. However, I do
not think that this is a promising route because it is undeniable also by the
relationists that the metric field has peculiarities of ontological priority respect
to the other matter fields (it can exist without any material content, but the
opposite is not true). Moreover, as briefly mentioned above, it is really hard to
show how a spacetime deprived of the typical spatiotemporal characteristics of
the metric field, might have any sense.

10. CONCLUSIONS
I have discussed the meaning of the expansion of the universe and shown how
the large-scale expansion, when regarded as expanding space, is a natural
feature of FLRW models that allows us to unambiguously explain observational
phenomena and data, in particular the high redshifts of distant cosmic objects.
The direct interpretation of these redshifts is indeed cosmological: photons
wavelengths are stretched by the underlying dynamical geometry of the
universe. Thus, without the concept of stretching space we are not able to

63
Some authors, for instance Earman and Norton (1987) and Rovelli (1997), defend this view,
but the former attain manifold substantivalism, whereas the latter relationism.

understand the expansion from a global viewpoint insofar as special-relativistic
descriptions constrained to consider the expansion as given by clusters
peculiar motions in a static space and the wavelength shifts as Doppler effects
necessarily result as approximations valid only in local domains of general-
relativistic curved spacetimes. Therefore, if the cosmological interpretation is
correct, the expansion of space is a large-scale phenomenon occurring among
clusters, and even if no force is involved, its effects explaining both wavelength
stretching and where the energy lost by photons is conserved, are evidence
of the substantial nature of the metric field. These general-relativistic facts
support a substantivalist position on spacetime since a traditional relationist
one is to be necessarily committed to fallacious special-relativistic
descriptions. Hence the expansion of the universe reveals the inadequacies of
such a metaphysics.

REFERENCES
Abramowicz, M. A. (2008). Spacetime is not Just Space and Time. New
Astronomy Reviews, 51(10), 799-802.
Abramowicz, M. A., Bajtlik, S., Lasota, J.-P., & Moudens, A. (2007). Eppur si
Espande. <http://arxiv.org/abs/astro-ph/0612155v3>
Agazzi, E. (2006). Temi filosofici della cosmologia. In Enciclopedia
Filosofica, Vol. 3 (pp. 2355-65). Fondazione Centro Studi Filosofici,
Milano: Bompiani.
Baker, D. (2005). Spacetime Substantivalism and the Cosmological Constant.
Barnes, L. A., Francis, M. J., James, J. B., & Lewis, G. F. (2006). Joining the
Hubble Flow: Implications for Expanding Space. Monthly Notices of
the Royal Astronomical Society, 373(1), 382-390.
Baryshev, Y. (2005). Conceptual Problems of the Standard Cosmological
Model. <http://arxiv.org/abs/astro-ph/0509800v1>
Bergia, S. (1997). Problemi fondazionali e metodologici in cosmologia. In G.
Boniolo (Ed.), Filosofia della fisica (pp. 169-244). Milano: Bruno
Mondadori.

Bunn, E. F., & Hogg, D. W. (2009). The Kinematic Origin of the
Cosmological Redshift. <http://arxiv.org/abs/0808.1081v2>
Carlip, S., & Scranton, R. (1999). Remarks on the New Redshift
Interpretation. <http://arxiv.org/abs/astro-ph/9808021v2>
Carrera, M., & Giulini, D. (2006). On the Influence of Global Cosmological
Expansion on the Local Dynamics in the Solar System.
<http://arxiv.org/abs/gr-qc/0602098v2>
Chodorowski, M. J. (2007). A Direct Consequence of the Expansion of
Space?. Monthly Notices of the Royal Astronomical Society, 378(1),
239-244. <http://arxiv.org/abs/astro-ph/0610590v3>
Cook, R., & Burns, M. (2008). Interpretation of the Cosmological Metric.
<http://arxiv.org/abs/0803.2701v2>
Cooperstock, F., Faraoni, V., & Vollick, D. (1998). The Influence of the
Cosmological Expansion on Local Systems. The Astrophysical Journal,
503(1), 61-66. <http://arxiv.org/abs/astro-ph/9803097v1>
Dainton, B. (2001). Time and Space. Chesham: Acumen.
Davis, T. M. (2004). Fundamental Aspects of the Expansion of the Universe
and Cosmic Horizons. Ph.D. Thesis. Sydney: University of New South
Wales. <http://arxiv.org/abs/astro-ph/0402278v1>
Davis, T. M., & Lineweaver, C. H. (2003). Expanding Confusion: Common
Misconceptions of Cosmological Horizons and the Superluminal
Expansion of the Universe. Publications of the Astronomical Society of
Australia (2004), 21(1), 97-109.
<http://arxiv.org/abs/astro-ph/0310808v2>
Davis, T. M., & Lineweaver, C. H. (2005). Misconceptions About the Big
Bang. Scientific American, 292(3), 36-45.
Davis, T. M., Lineweaver, C. H., & Webb, J. K. (2003). Solutions to the
Tethered Galaxy Problem in an Expanding Universe and the
Observation of Receding Blueshifted Objects. American Journal of
Physics, 71(4), 358-364.
DiSalle, R. (1995). Spacetime Theory as Physical Geometry. Erkenntnis,
42(3), 317-337.

Dorato, M. (2006). Absolute Becoming, Relational Becoming and the Arrow
of Time: Some Non-Conventional Remarks on the Relationship
Between Physics and Metaphysics. Studies in History and Philosophy of
Modern Physics, 37(3), 559-576.
Earman, J., & Norton, J. (1987). What Price Space-Time Substantivalism?
The Hole Story. The British Journal for the Philosophy of Science, 38,
515-25.
Eddington, A. (1933). The Expanding Universe. New York: The MacMillan
Company.
Ehlers, J. (1990). [Discussion]. In B. Bertotti et al. (Eds.), Modern Cosmology
in Retrospect (pp. 29-30). Cambridge: Cambridge University Press.
Einstein, A., & Straus, G. (1945). The Influence of the Expansion of Space on
the Gravitation Fields Surrounding the Individual Stars. Review of
Modern Physics, 17(2-3), 120-124.
Ellis, G. (1978). Is the Universe Expanding?. General Relativity and
Gravitation, 9(2), 87-94.
Ellis, G. (2007). Issues in the Philosophy of Cosmology. In J. Butterfield & J.
Earman (Eds.), Philosophy of Physics, part B (pp. 1183-1285).
Ellis, G., & Matravers, D. (1995). General Covariance in General Relativity?.
General Relativity and Gravitation, 27(7), 777-788.
Ellis, G., Maartens, R., & Nel, S. (1978). The Expansion of the Universe.
Monthly Notices of the Royal Astronomical Society, 184, 439-465.
Fano, V., & Macchia, G. (2008). How Contemporary Cosmology Bypasses
Kantian Prohibition Against a Science of the Universe.
Francis, M. J., Barnes, L. A., James, J. B., & Lewis, G. F. (2007). Expanding
Space: The Root of all Evil?. Publications of the Astronomical Society of
Australia, 24(2), 95-102. <http://arxiv.org/abs/0707.0380v1>
Grn, ., & Elgary, . (2006). Is Space Expanding in the Friedmann
Universe Models?. <http://arxiv.org/abs/astro-ph/0603162v2>

Harrison, E. (1993). The Redshift-Distance and Velocity-Distance Laws.
Astrophysical Journal, 403(1), 28-31.
Harrison, E. (1995). Mining Energy in an Expanding Universe. Astrophysical
Journal, 446(1), 63-66.
Harrison, E. (2000). Cosmology. The Science of the Universe. Cambridge:
Hawley, J., & Holcomb, K. (2005). Foundations of Modern Cosmology.
Oxford: Oxford University Press.
Hinckfuss, I. (1975). The Existence of Space and Time. Oxford: Clarendon
Press.
Hoefer, C. (1996). The Metaphysics of Space-Time Substantivalism. Journal
of Philosophy, 93(1), 5-27.
Hoefer, C. (2000). Energy Conservation in GTR. Studies in History and
Philosophy of Modern Physics, 31(2), 187-99.
Infeld, L., & Schild, A. (1945). A New Approach to Kinematic Cosmology.
Physical Review, 68(11), 250-72.
Kiang, T. (2003). Time, Distance, Velocity, Redshift: A Personal Guided
Tour. <http://arxiv.org/abs/astro-ph/0308010v1>
Melchiorri, B., & Melchiorri, F. (1994). Cosmologia del Big Bang. In G.
Russo & E. Verondini (Eds.), Dentro la fisica (pp. 52-72). Bologna:
Clueb.
Mellor, H. (1980). On Things and Causes in Spacetime. British Journal for the
Philosophy of Science, 31(3), 282-288.
Merleau-Ponty, J. (1965). Cosmologie du XX
e
sicle. Paris: Gallimard.
Misner, C., Thorne, K., & Wheeler, J. (1973). Gravitation. San Francisco:
Freeman.
Morgan, J. (1988). Are Galaxies Receding or Is Space Expanding?. American
Journal of Physics, 56(9), 777-8.
Nerlich, G. (1991). How Euclidean Geometry Has Misled Metaphysics. The
Journal of Philosophy, 88(4), 169-189.

Nerlich, G. (1994a). What Spacetime Explains. Cambridge: Cambridge
University Press.
Nerlich, G. (1994b). The Shape of Space. Cambridge: Cambridge University
Press.
Nerlich, G. (2008). Why Spacetime Is Not a Hidden Cause: A Realist Story.
<www.spacetimecenter.org/conferences/2008/Nerlich.pdf>
Ohanian, H. (2000). What Space Scales Participate in Cosmic Expansion?.
American Journal of Physics, 68(8), 689-690.
Pauri, M. (1991). The Universe as a Scientific Object. In E. Agazzi & A.
Cordero (Eds.), Philosophy and the Origin and Evolution of the
Universe (pp. 291-339). Dordrecht: Kluwer Academic Publishers.
Peacock, J. A. (1999). Cosmological Physics. Cambridge: Cambridge
University Press.
Peacock, J. A. (2006). Cosmological Physics: Additional Topics.
<www.roe.ac.uk/~jap/book/expandspace.pdf >
Peacock, J. A. (2008). A Diatribe on Expanding Space.
Peebles, P. J. E. (1993). Principles of Physical Cosmology. Princeton, NJ:
Princeton University Press.
Pitts, J. B. (2004). Has Robert Gentry Refuted Big Bang Cosmology? On
Energy Conservation and Cosmic Expansion. Perspectives on Science
& Christian Faith, 56(4), 260-265.
Prokhovnik, S. J. (1985). Light in Einsteins Universe. Dordrecht: D. Reidel
Publishing Company.
Rindler, W. (2006). Relativity. Special, General, and Cosmological. Oxford:
Oxford University Press.
Ross, K. L. (1999). The Ontology and Cosmology of Non-Euclidean
Geometry. <www.friesian.com/curved-1.htm>
Rovelli, C. (1997). Halfway Through the Woods: Contemporary Research on
Space and Time. In J. Earman & J. Norton (Eds.), The Cosmos of
Science (pp. 180-223). Pittsburgh: University of Pittsburgh Press.

Schutz, B. (2003). Gravity From the Ground Up. Cambridge: Cambridge
University Press.
Srianand, R., Petitjean, P., & Ledoux, C. (2000). The Cosmic Microwave
Background Temperature at a Redshift of 2.33771. Nature,
408(6815), 931-935.
Sumner, W. Q., & Sumner, D. Y. (2007). Coevolution of Quantum Wave
Functions and the Friedmann Universe.
Weinberg, S. (2008). Cosmology. Oxford: Oxford University Press.
Whiting, A. B. (2004). The Expansion of Space: Free Particle Motion and the
Cosmological Redshift. The Observatory, 124(1180), 174-189.
<http://arxiv.org/abs/astro-ph/0404095v1>
Whitrow, G. J. (1980). The Natural Philosophy of Time. Oxford: Clarendon
Press.


Ontology and Mathematics in Classical Field Theories
and in Quantum Mechanics

Adriano Angelucci*
adriano.angelucci@alice.it
Vincenzo Fano**
vincenzo.fano@uniurb.it

ABSTRACT
A draft of a possible comparison between the use made of mathematics in
classical field theories and in quantum mechanics is presented. Hilberts space
formalism, although not only elegant and powerful but intuitive as well, does
not give us a spatio-temporal representation of physical events. The picture of
the electromagnetic field as an entity which is real in itself i.e., as a wave
without support fostered by the emergence of special relativity can be seen as
the first step, favored by many physicists and philosophers, of a gradual
escape from intuition into a purely mathematical representation of the
external world. After the introduction, in recent theoretical physics, of fiber
bundle formalism the classical notion of field acquires a new spatio-temporal
intuitiveness. This intuitiveness is clearly foreshadowed in the Kantian and
Meinongian analysis of the notion of magnitude. At the end of the paper we
show that, contrary to what happens in quantum mechanics, mathematics plays
a truly explicative role in general relativity, without any loss of spatio-temporal
intuitiveness.

INTRODUCTION
In the chapter on Field and ether of their popular book, Einstein and Infeld
(1938) write that to the modern physicist the electromagnetic field is as real as
the chair on which he sits. It is easy to find statements of this kind in handbooks

* Department of Philosophy Verona University
**

Department of Philosophy Urbino University

of physics, but we mentioned Einstein because he was the first to show the
truth of this statement, by introducing the special relativity theory. As a matter
of fact, at least until 1905 but even after that, many physicists considered
electromagnetic waves to be similar to sea waves and sound waves, that is they
thought of them as oscillations of some kind of matter. Nevertheless it was not
so easy to identify the carrier of electromagnetic waves, whereas for sea and
sound waves they are, respectively, water and air; due to the enigmatic
character of the stuff which transports light, it was dubbed ether.
1

The vanishing of luminipherous ether from electromagnetic theory brings
with it the idea that waves are realities in themselves without support. But it is
easy to understand what an ether wave is, on the basis of the analogy with a sea
wave, whereas it is quite difficult to visualize a wave that isnt a wave of
anything. Therefore one can state that with the coming of special relativity
theory and the refutation of the hypothesis that electromagnetic waves are the
oscillation of some sort of matter, a withdrawal
2
was begun in the direction
of a mere mathematical representation of what happens in the physical world.
This withdrawal was favored by many physicists and philosophers and reached
its apex in the renunciation of the possibility of a spatio-temporal picture of
what occurs in the microphysical world after the advent of quantum mechanics.
It is clear that mathematics is essential to physics, but no type of formalism
is suitable for every domain of objects. We will show that only with the
introduction of fiber bundles does Einsteins above mentioned statement
become fully understandable. Nevertheless not every implementation of new
mathematics favors the comprehension of physical reality. For instance we will
see that in quantum mechanics the formalism, though elegant and powerful, in
a certain sense, overshadows, instead of shedding light on, physical reality.
In the first section below the sense in which Hilberts space formalism
intuitiveness is not really physical is presented. In the following, on the
contrary, it appears that fiber bundles clarify the notion of the classical wave. In
the third section the importance of the notion of intensive magnitudes for this
clarification emerges. Finally we present the intuitive character of general
relativity moving from the concept of tensor in elasticity theory.

1
On these topics see the still wonderful Whittaker 1951.
2
See for instance Heisenberg 1958, ch. 10. The history of the withdrawal is told very well by
Hanson 1963, albeit in a little too hagiographic way for our taste.
Angelucci & Fano Ontology and Mathematics 141

QUANTUM MECHANICS
If on entering a caf we ask for the rest room and we are told to go first left and
then right, but we go first right and then left instead, we usually end up in a
different place. This example shows that though sum and multiplication are
commutative operations, sometimes by inverting the order of an operation we
change the result. Here the operation is the composition of two translations on
the plane. This is also true in the case of rotations in space. Indeed in algebra
the rotation of a vector (x,y,z) around the z axis is represented by a matrix of
the type:
cos sen 0
sen cos 0
0 0 1
R

,
where is the rotation angle. If R
is the matrix representing the rotation of

angle around the x axis, then:
1 0 0
0 cos sen
0 sen cos
R

Generally it holds:
R R R R

,
which means that the composition of two rotations in space one around the z
axis and the other around the x axis is a non commutative operation. Let us
propose an example. Let us consider a vector lying on the x axis:

Picture 1
x
y
z

If we rotate it clockwise by a right angle we obtain:

Picture 2

That is, the vector now lies on the y axis. Then we rotate the vector by a right
angle clockwise around the x axis and we obtain:

Picture 3

Returning now to the situation of picture 1, we invert the order of
rotations; that is, first we rotate the vector by a right angle clockwise around
the x axis. Obviously, since the vector lies on the x axis itself, such a rotation
does not produce any effect. Therefore we stay still in the situation of picture
1. Then we rotate the vector by a right angle clockwise around the z axis and we
obtain the situation presented in picture 2, that is the vector lies on the y axis.
It follows that, since in the first case the final configuration is that of picture 3,
by inverting the order of rotations the result changes.
3

When Heisenberg realized that the p and q variables, i.e., moment and
position, do not commute in the new quantum mechanics, Max Born proposed
representing the new kind of entity no longer through a number, as in classical

3
Remember that, on the contrary, rotations on the plane commute.
x
y
z
x
y
z

physics, but through a matrix.
4
Thus in new physics a new kind of mathematics
was introduced, which had been developed many decades before, in order to
deal with rotations in space. At first the new formalism seems very abstract.
Indeed Heisenberg repeated many times that in the mathematical
representation of nature we have to abandon spatio-temporal intuition.
Nevertheless the intuitiveness of the new formalism appears again when von
Neumann interpreted Schrdingers equation that is, the law which governs
the evolution of quantum state as a rotation in Hilberts space (von Neumann
1932, .III.1). Many handbooks of quantum mechanics emphasize, in fact, the
possibility of representing geometrically what occurs through Hilberts space
picture (e.g., Fano 1971). Yet we have to remember that the space in which
such processes happen is not a physical one, but a mathematical manifold
without representative capacity.
In a certain sense one can now formulate the great unsolved problem of
microphysics i.e., the measurement muddle as the impossibility to come
back from the abstract representation in Hilberts space to what happens in
physical space. For instance in a 2-dimensional Hilberts space, a dichotomic
variable, such as the spin of one particle, can have at the same moment two
values, even though each one is not completely ascribed to the particle, but
only partly, determined by its probability amplitude. Nevertheless, when one
measures spin in physical space, the particle shows only one of the two values;
but to identify which presumptive physical process would allow us to abandon
the representation given by the superposition of the two states, returning to
the more prosaic determination of the physical reality, is not easy.
5
Indeed one
forgets that, although mathematical language is absolutely essential to the
physical understanding of the world, it is nevertheless dangerous to state that
there is only a mathematical explanation of what happens, i.e., an explanation
totally deprived of intuitive character. Such a perspective favors Platonist and
instrumentalist interpretations of contemporary physics: i.e., either theoretical
terms of physics represent entities completely different from those of our
experience, or they have no representative capacity. Thus one opens the
Pandoras box of the wildest speculations of contemporary physics, lamented,
for instance, by Smolin (2006).

4
See Cassidy 1992, ch. 10. See also Born (1978, ch. 19) where the German physicist tells how,
probably on July 10th 1925, he noticed that the kind of multiplication, which Heisenberg needs, was
that between matrices.
5
See, for example, Ghirardi 1997.

THE ELECTROMAGNETIC FIELD
As mentioned above, this escape into mathematical representation begins
when one affirms that the electromagnetic field is not a wave of any stuff, but is
real in itself. On the other hand it is usual to represent an electromagnetic field
as a wave in physical space through drawings of this kind
6
:

Picture 4

The apparent crests and valleys lie in physical space, as happens in the
case of sea waves, but nothing oscillates in space! Therefore one is compelled
to ask: waves made of what? The standard answer is that it is not possible to
find an intuitive picture of what happens, because these drawings have only a
heuristic and didactic value, but they are not able to express what actually
occurs. Indeed, according to physicists, only the mathematics of Maxwells
differential equations provides a suitable representation of reality, that is a
non-intuitive representation of it.
On the other hand, in our opinion, at least for classical field theories, the
situation is quite different. In picture 4 something goes wrong, i.e., there are
waves, but they are not extended into physical space. A different kind of
mathematics can help provide a more intuitive representation. To understand
the point let us consider a 1-dimensional physical space. Then let us ascribe to
each point of this straight line a 1-dimensional vector space. Roughly speaking
one can image the situation as a bundle of infinite parallel straight lines all
perpendicular
7
to the unique spatial dimension we are considering.

6
For the sake of simplicity we have represented the wave in 1-dimension.
7
This perpendicularity is not essential to all fiber bundles, but only when fibers are tangent
spaces.

Picture 5

In other words, each point of the basis must correspond to a straight line. Now
let us associate each point of the basis to a vector, which could represent, for
instance, an electric field.
Each point of the 1-dimensional space has a determined value of the
electric field, to which a given vector corresponds.
8
One can put the latter
inside each fiber (as the infinite perpendicular straight lines are dubbed).
Probably the electric field varies like a wave along the straight line. But we now
understand where the wave lies:

Picture 6

That is, as is evident in picture 6, the electric field oscillates in the fiber bundle
not in physical space.
It is easy to generalize such a reasoning to the case of a 3-dimensional
physical space and therefore to 3-dimensional vectors, which lie in 3-

8
In the 1-dimensional physical space the electric field is represented by a 1-dimensional vector.

dimensional fibers. In a fiber bundle it is evident how the electromagnetic field
is real in itself, that is it doesnt need a physical support such as ether.
9

One could say that this mathematical representation is even more artificial
than that of Maxwell. Indeed one might ask where are these three additional
dimensions, which one has to ascribe to each point of physical space? Yet all
this is quite intuitive. In the Critique of Pure Reason (B202ff.) Kant already
distinguished between intuitions (Anschauungen) and anticipations
(Antizipazionen) of perception: the former are extensive quantities extended
in space and time the latter are intensive quantities, which are ascribed to a
single spatio-temporal point. To understand this, let us consider a yellow
colored stripe, which passes slowly from left to right from a milder yellow to a
darker one. In this situation color has an intensity peculiar to each point of the
stripe, as in the case of electric field. In general it is quite reasonable that the
complexity of reality
10
, at each spatio-temporal point, is so articulated as to
require a rich fiber space to represent it. The reality that, as it were, fills our
spatio-temporal intuitions possesses different qualities, such as yellow and red,
smooth and rough. Moreover these qualities present different degrees: more or
less red, more or less smooth. Something similar occurs in the representation
of an electromagnetic field through fiber bundle geometry.

INTENSIVE MAGNITUDES
As we were saying at the beginning of the present paper, in order to play a truly
explanatory role and to ensure the cognitive power of physics, theories, in their
attempts to describe reality, ought to make use of what we might call the
appropriate mathematics. By appropriate mathematics we mean those
mathematical theories which seem to be closer to our empirical intuitions. We
have also seen that the topological notion of a fiber bundle provides a good
example of appropriate mathematics. As a matter of fact, contrary to
Euclidean space or to Minkowsky space-time, fiber bundles constitute an
intuitive representation of electromagnetic waves and help us to grasp the
sense in which it is said that electromagnetic waves are real.

9
For an intuitive and elegant presentation of the role of fiber bundles in contemporary
theoretical physics see Penrose 2004, ch. 15.
10
Realitt; it is Kants terminology.

In this section we would like to suggest that the definition of intensive
magnitude given by Kant might be held to provide the intuitive ground for the
mathematical description of classical field theories in terms of fiber bundles. In
order to do this, a few historical considerations might be helpful.
By the second half of the nineteenth century experimental psychology had
opened a whole new chapter in the study of the mind by showing the possibility
to introduce an experimental approach to the matter. This attempt soon raised
a heated epistemological debate concerning the limits of measurement
procedures and the definition of measurement itself. As a consequence of the
reductionism inherent in any experimental approach, the qualitative side of
sensations dropped out of the picture and the new psychology started
considering sensations as a particular kind of magnitude. Positivist
epistemology had also adopted Kants distinction between extensive and
intensive magnitudes together with the corresponding definitions provided
by the first Critique. According to these definitions, sensations had to be
regarded as intensive magnitudes: A magnitude, Kant wrote that is
apprehended only as a unity, and in which multiplicity can be presented only by
approaching [from the given magnitude] toward negation, = 0, I call an
intensive magnitude (Kant B 210). The possibility to measure something
directly or indirectly allows a distinction between fundamental and
derivative magnitudes. In fact, in the case of indirect measurement, what we
are really measuring is not the magnitude we are interested in, but some other
magnitudes, which have the empirical feature of being functionally related to it.
The prevailing stance within the epistemological debate mentioned above
can be traced back to the work of Hermann von Helmholtz (1821-1894).
According to Helmholtz (1887) extensive magnitudes only enjoy the status of
fundamental magnitudes, whereas intensive magnitudes ought to be
considered and treated as merely derivative ones. A good example of this
attitude towards magnitudes can be found in the ideas concerning
measurement endorsed by Helmholtzs pupil Johannes von Kries (1882). The
intensive magnitudes that physics deals with, according to Kries, are merely
combined units (combinierte Einheiten), i.e., they are made up of (and are
nothing more than) the three fundamental physical dimensions of space, time
and mass, and they are therefore considered to be the result of a general
agreement regarding their practical applicability. From his point of view then,
the notion of intensive magnitude constitutes a sort of derivative concept, a

concept one cannot grasp without the help of the more fundamental and
intuitively evident notion of extensive magnitude.
11

Towards the end of the nineteenth century, an unexpected turn was given
to the debate on magnitudes by the Austrian psychologist and philosopher
Alexius Meinong (1853-1920). Meinong had studied in Vienna as a pupil of
Franz Brentano (1938-1917) and had founded in 1894 in Graz the first
Austrian laboratory for experimental psychology. In 1896 he published a long
essay with the title On the Meaning of Webers Law. Contributions to the
Psychology of Comparing and Measuring. In this essay, rejecting the view
endorsed by Kries and generally accepted by physicists, Meinong suggested a
radical revision of the debate concerning measurement. Measurement, he
believed, must be considered primarily as a mental process and its logical
analysis should therefore come only after a purely psychological account of its
main phenomenological features. Moreover this logical analysis, in order to be
grounded in our intuitions, has to account for those psychological features.
Meinongs treatment of measurement commits him therefore to the view
according to which intensive magnitudes must be regarded as fundamental. As
a consequence, extensive magnitudes should be thought of and treated as
derivative magnitudes.
The logical analysis of measurement, in other words, has to mirror specific
features made visible by its psychological description and this includes
accounting for the new relationship that now occurs between the two kinds of
magnitudes. In practice, this amounts to acknowledging the fundamental status
of intensive magnitudes and to giving a definition of the general notion of
magnitude which encompasses in itself the case of extensive magnitudes as
well. The definition given by Meinong is the following: If one can think of a y
such that, so to speak, if seen from x it lies in the same direction as non-x, then
x has or is a magnitude and non-x is zero ( Meinong 1896, .1).
12

It might be helpful at this point to try to render explicit the tacit, but deeply
intuitive, assumptions that underlie this apparently mysterious definition.
According to Meinong, a thorough psychological examination of our

11
Kries assumes here that space, time and mass are extensive magnitudes. We would like to
point out, though, that while the inertial mass (quantitas materiae) is an extensive magnitude, the
gravitational one may also not be so.
12
It should nevertheless be kept in mind that Meinong didnt think of the above formulation as of
a fully fledged definition. He would regard it rather as an empirical criterion useful to distinguish
between things that can be considered magnitudes from things that cannot.

perceptual field reveals the presence of certain directions within this field.
The idea is that if we start by considering the absence of a particular
phenomenon these directions allow us to order a series of phenomena similar
to it. As everyday language shows, in the case of magnitudes the notion of
dissimilarity (Verschiedenheit) itself entails the idea of a direction. If, by
comparing two magnitudes A and B, we found that they are not identical, we
probably wouldnt say that A and B are dissimilar, but rather that A is bigger
(or smaller) than B. The reason for this is that the second expression allows us
to assign to A and B a specific position on the imaginary line which connects
them to the zero point.
According to the general definition of magnitude suggested by Meinong
then, there are, within our perceptual field, several imaginary non-overlapping
lines along which different kinds of magnitudes (spatial, temporal, intensive
magnitudes etc.) draw near to each other toward their zero point. To use
Meinongs own words a magnitude is the disposition (Eignung) of a quality
to belong to one of the many lines which converge toward a zero point. It is
precisely in virtue of this capacity to move along this imaginary line which
connects it to its zero point that we can attribute a magnitude to a given quality.
The general condition that has to be satisfied in order for two magnitudes to be
comparable is the fact of lying on the same line. This means for instance that
two points a' and a'' which lie on different lines are certainly sufficient to
determine a specific direction, but this direction doesnt approach any zero
point and the two points cannot therefore be compared.
13

The above considerations seem to us sufficient to support the view that the
notion of magnitude does not necessarily have to be bound to any extensive
representation of space or time. Magnitude can also be seen as a sort of degree
which takes place in space and time and which is capable of varying from one
point to the other. We consider fiber bundles as an appropriate
mathematical instrument, useful to connect classical field theories to our own
intuitions concerning the external world.

13
See Meinong 1896, .7. It is worth pointing out that Meinongs approach was partially
adopted by Russell, in 1903, but since then has never been seriously considered in subsequent
epistemological debate on this same topic and would therefore merit more careful analysis. But see
Fano 1999.

GENERAL RELATIVITY
Quantum mechanics and general relativity theory are the most important
theories of contemporary physics. As is well known, till now it has not been
possible to conciliate them. Both make a massive use of mathematics, but in the
former, as mentioned above, formalization partly hides an as yet unresolved
difficulty, whereas in the latter, as we shall see, mathematics is a powerful
instrument useful to specify our intuition.
An essential peculiarity of general relativity is the dependence of physical
space metric on the distribution of masses. In classical physics the distance
between two points of coordinates (x
1
,y
1
,z
1
) and (x
2
,y
2
,z
2
) is given by:

(
2
1
)
2
+(
2

1
)
2
+(
2

1
)
2
(1)
On the contrary, in general relativity, for each couple of points, distance is
evaluated by a much more general formula:

2
+
2
+
2
+
(2)
Where, as usual, the s indicate differences between coordinates. In a certain
sense (2) is a generalization of Pythagorass theorem, which we apply in (1). In
relativistic space
14
coefficients
could always change point by point. For this

reason space is no longer mathematically represented by an Euclidean space,
but by a more general geometrical entity dubbed differentiable manifold, for
which the metric is not constant, so that it is possible to ascribe to each point
the correct series of coefficients
, that is the metric peculiar to that point.

Such metric depends on Einsteins celebrated equation, which allows us to
calculate
15
coefficients
given the distribution of masses and energy

16
.
Contrary to the first impression, representing space by a geometry, in
which the metric varies point by point, is not against our intuition. For
instance, in our perception, the same couple of points, if embedded in different
pictures, could appear nearer or farther, as occurs in the famous Mller-Lyer
illusion:

14
Here for the sake of simplicity we do not consider the 4th dimension, that is the temporal one.
15
Such a calculation is actually quite difficult and often does not produce univocal results.
16
For an intuitive and more exhaustive presentation see Penrose 2004, ch. 19.

Picture 7

Therefore perceptual space does not have a constant metric, so that
representing physical space with a varying metric is not so bizarre, because the
same, in a certain sense, occurs for visual space. Of course the former variation
of metric is much more objective.
We have seen that in classical theory of the electromagnetic field it is
possible to ascribe to each point of physical space a 3-dimensional vector
space, in which, as it were, the vector electromagnetic field lives. Nevertheless
we already know from elasticity theory that each point of physical space could
possess complex properties, which could not be represented by a simple 3-
dimensional vector. For instance let us consider a rock salt crystal; if one
applies a shear force parallel to the crystal plane, the stone quite easily cleaves,
whereas, if at the same point one applies a force perpendicular to the crystal
plane, the stone opposes more resistance and in the end it breaks. Therefore
the resistance to stress of the crystal at that point is not isotropic, so that to
represent in physical theory a much more complex situation, a vector is no
longer sufficient. For this reason already at the end of nineteenth century
physicists
17
introduced the concept of tensor, which is able to ascribe to a
point of physical space as many variables as one wills.
18

All these values could, as it were, live in multi-dimensional fibers, which
one applies to each point of physical space, as we have done in the case of the
electromagnetic field, but in a more general form. Neither is this against
intuition, if one considers that we are, for instance, able to distinguish
hundreds of thousands of different hues of color, which could be psycho-
physically ordered on the basis of many variables. And this holds for color, but
our perceptual space is much more complex. Therefore the qualitative

17
See Voigt 1898.
18
On this topic Brillouin 1938 is still valuable.

character of perceptual space has a very articulated structure, so that it is not so
strange that in physical theory one is compelled to introduce such complex
fibers to represent what occurs.
All this is of utmost importance for general relativity, because Einsteins
equations connect the metric tensor, i.e., the set of g
ii
coefficients which
changes point by point with the energy-moment tensor, which describes the
distribution of masses and energies. It is thus evident that the mathematization
implicit in the theory, though not simple, stems from our intuition; therefore
formalization here facilitates our understanding, without overshadowing
unsolved problems. As we have seen, the same does not hold for quantum
mechanics. Our opinion with respect to the latter theory is similar to that
expressed at the beginning of the paper for Einsteins statement that the
electromagnetic field is real in itself, before we found the more intuitive
representation of fiber bundles. It follows that, as far as the microphysical
world is concerned, either we have not yet identified a suitable formalism, or
there are some physical phenomena which still escape our understanding.

REFERENCES
Brillouin, L. (1938). Les tenseurs en Mcanique et en Elasticit. Paris:
Masson.
Born, M. (1978). My Life: Recollections of a Nobel Laureate. London: Taylor
& Francis.
Cassidy, D. (1992). Uncertainty. The Life and Science of Werner Heisenberg.
New York: Freeman.
Einstein, A. & Infeld, L. (1938). The Evolution of Physics: From Early
Concept to Relativity and Quanta. Cambridge: Cambridge University
Press.
Fano, G. (1971). Mathematical Methods of Quantum Mechanics. New York:
McGraw Hill.
Fano, V. (1999). Meinong e l'interpretazione della legge di Weber. Teorie e
modelli, 4, 99-109.
Ghirardi, G. C. (2004). Sneaking a Look at Gods Cards. (tr. by G. Malsbary).
Princeton, NJ: Princeton University Press. [1997]

Hanson, N. (1963). The Concept of the Positron. A Philosophical Analysis.
Heisenberg, W. (1958). Physics and Philosophy. London: Allen and Unwin.
Helmholtz, H. (1887). Zhlen und Messen, erkenntnisstheoretisch betrachtet.
In Philosophische Aufstz. Eduard Zeller zu seinem fnzigjhrigen
Doctor-Jubilum gewidmet. Leipzig: Fues.
Kant, I. (1996). Critique of Pure Reason. (tr. by W. S. Pluhar).
Indianapolis/Cambridge: Hacket Publishing Company. [1787]
Meinong, A. (1896). ber die Bedeutung des Weberschen Gesetzes.
Beitrge zur Psychologie des Vergleichens und des Messens. Zeitschrift
fr Psychologie und Physiologie der Sinnesorgane, 11, 81-133, 230-
285, 353-404. Also in R. Haller & R. Kindinger (Eds.) (1969-1978),
Gesamtausgabe, vol. II, (pp. 215-376). Graz: Akademische Druk-und
Verlaganstalt.
Von Neumann, J. (1932). Mathematische Grundlagen der Quantenmechanik.
Berlin: Springer.
Von Kries, J. (1882). ber die Messung intensiver Grssen und ber das
sogenannte psychophysische Gesetz. Vierteljahrsschrift fr
wissenschaftliche Philosophie, 6, 257-294.
Penrose, R. (2004). The Road to Reality. London: Vintage.
Russell, B. (1903). The Principles of Mathematics. Cambridge: Cambridge the
University Press.
Smolin, L. (2006). The Trouble with Physics. The Rise of String Theory, the
Fall of a Science and What Comes Next. Orlando, FL: Houghton
Mifflin.
Voigt, W. (1898). Die fundamentale physikalische Eigenschaften der Kristalle
in elementarer Darstellung. Leipzig: Von Veit.
Whittaker, E. (1951). A History of the Theories of Aether and Electricity.
London: Nelson.

Not Particles, Not Quite Fields:
An Ontology for Quantum Field Theory
*

Tracy Lupher**
lupherta@jmu.edu

ABSTRACT
There are significant problems involved in determining the ontology of
quantum field theory (QFT). An ontology involving particles seems to be ruled
out due to the problem of defining localized position operators, issues
involving interactions in QFT, and, perhaps, the appearance of unitarily
inequivalent representations. While this might imply that fields are the most
natural ontology for QFT, the wavefunctional interpretation of QFT has
significant drawbacks. A modified field ontology is examined where
determinables are assigned to open bounded regions of spacetime instead of
spacetime points.

1. INTRODUCTION
Looking to current physical theories for insights into metaphysics and
ontology has a long tradition in philosophy. Applying this principle today, it is
natural to look at quantum field theory (QFT) for insights into the fundamental
types entities in the physical world, or at least for constraints on possible
ontologies. QFT is the successor to quantum mechanics and provides the
mathematical framework for the standard model of particle physics. The
predictions of QFT for electromagnetic interactions, also known as quantum
electrodynamics, are extremely accurate when compared to experimental
results. However, using QFT for insights into ontology is far from
straightforward. The typical choice given for the ontology of QFT is either

* I would like to thank Fred Kronz, Peter Morgan, and an anonymous referee for their helpful
comments.
** James Madison University
particles or fields.
1
Many people have argued that a particle ontology is
incompatible with QFT, which implies that QFT is a theory about fields.
However, new philosophical works by Halvorson (Halvorson and Mueger
2007) and Baker (2009) challenge a field interpretation of QFT called the
wavefunctional interpretation. I will argue that a modified field ontology is
possible that is consistent with the mathematical structure and physical
assumptions of QFT. The modified field ontology to be developed is based on a
mathematically rigorous version of smeared quantum fields, which has not
been discussed very much in the philosophical literature on QFT. I then show
how this ontology is carried over to the framework of algebraic quantum field
theory (AQFT). AQFT generalizes this modified field ontology by assigning a
collection of determinables to open bounded regions of Minkowski spacetime.
The plan for this paper is as follows. Section two of this paper examines two
aspects of a particle: localizability and countability. It reviews various No Go
results that undermine both localizability and countability in QFT. These
results show that there is little hope for a particle ontology in QFT. Section
three examines how the field concept changes in QFT. The notion of a field as
assigning determinables to spacetime points cannot be correct mathematically
in QFT, at least if the field is understood as an operator. However, the field can
be smeared out so that the field, while no longer defined on spacetime
points, is defined on spacetime regions. In section four, I discuss how this field
ontology involving spacetime regions is generalized in AQFT. The appearance
of classical observables in AQFT introduces a new aspect to field
configurations in AQFT. Conclusions are given in section five.

2. THE PROBLEMS WITH PARTICLES
Many QFT books begin by constructing Fock space, which seems to suggest
that QFT has a particle interpretation.
2
The Fock space can be created using
the creation and annihilation operators, which satisfy the canonical

1
Alternative ontologies for QFT have been proposed; see the articles in Kuhlmann et al. 2002.
2
My discussion of QFT will necessarily be non-technical. For more details, the reader should
consult the references citied in the paper and especially Halvorson 2007.
Tracy Lupher Not Particles, Not Quite Fields 157
commutation relations (CCRs)
3
, and a vacuum state.
4
When the creation
operator acts on the vacuum state, a new state with one particle is produced.
5
If
the creation operator is applied n times on the vacuum state, a state containing
n particles is created. Roughly, the set of all such states created in this way can
be used to construct the Fock space, which is a (infinite) direct sum of the
Hilbert spaces for zero particles, one particle, two particles, etc. Using the
creation and annihilation operators, a number operator can be constructed that
counts the number of particles in a state. The Fock space construction might
suggest that a particle interpretation is the most natural ontology for QFT.
Particles seem to have existence independently of their properties, which
makes a substance ontology a natural fit. In particular, particles exhibit a kind
of primitive thisness, haecceity, or transcendental individuality.
6

However, there are numerous diverse and devastating criticisms against any
particle ontology in QFT. These attacks come from physicists such Weinberg
(Feynman and Weinberg 1987, pp. 78-79), and Wald (1994, pp. 51-52) as
well as from philosophers such as Malament (1996), Halvorson and Clifton
(2002), and Fraser (2008). The connection between the particle-states
used to build the Fock space and the particle concept are tenuous at best.
There are two crucial features that characterize particles: (1) particles are
discrete, localizable entities and (2) particles are countable or aggregated. (1)
is undermined by results in relativistic quantum theory and QFT, while (2) is
undermined in QFT.

3
Requiring the creation and annihilation operators to satisfy the CCRs generates a Fock space
for boson particles. We can also require that the creation and annihilation operators satisfy the
canonical anti-commutation relations, which would generate a Fock space for fermion particles.
4
The term vacuum state is a bit misleading. Though it is usually defined as the state with no
particles in it, the state has energy fluctuations, which seems to suggest that the picture of the vacuum
as a state where nothing is happening is incorrect. This already suggests that labeling the states of the
Fock space as particle states is misleading (Kuhlmann 2006).
5
Teller (1995) referred to such a state as having one quanta in it not a particle. The distinction
between particles and quanta is discussed below.
6
Primitive thisness faces serious problems dealing with indistinguishable particles (bosons) in
quantum mechanics, and particle indistinguishability is still an issue in QFT (Teller 1995). This is not
to say that the notion of a particle in classical physics is without conceptual difficulties. For example,
modeling electrically charged particles as point-particles would result in the particle having infinite
energy.
2.1. ARGUMENTS AGAINST LOCALIZABILITY
Malament (1996) proved a theorem that excludes the possibility of localized
particles in relativistic quantum theory. The idea behind the theorem is that a
strictly localized particle cannot be detected in two disjoint spatial regions at
the same time. Under certain mild assumptions, there is no position operator,
which transforms covariantly, of a strictly localized particle. The theorem
shows that there is a conflict between relativity and the localization condition
for a fixed number of particles. One might think that a weaker notion of locality
might save particles, but Halvorson and Clifton (2002) showed that there are
no unsharp localized position operators either. All of these results hold in
relativistic quantum mechanics, so a defender of particles might hope for more
success in QFT. However, Halvorson and Clifton (2002) proved another
result which shows that there are no localizable particles in any relativistic
theory including QFT.
7
A result in axiomatic QFT known as the Reeh-
Schlieder theorem shows that a strictly localized entity, such as a particle, in a
bounded spacetime region is incompatible with the axioms.

2.2. ARGUMENTS AGAINST COUNTABILITY
Teller (1995, p. 29) rejected a substance ontology (primitive thisness) for
quantum particles and argued that quantum particles are better characterized
as quanta, which can be counted or aggregated. The idea behind countability is
that there are states in which there is some definite number of quanta, so a state
that has four quanta and a state that has five quanta can be added together and
result in a state that has nine quanta. This property is undermined by the
Unruh effect, interacting quantum field models, and the Reeh-Schlieder
theorem.

2.2.1. THE UNRUH EFFECT
The Unruh effect involves a uniformly accelerating observer who will detect a
thermal bath of particles in the Minkowski vacuum state, while an inertial
observer will not detect any particles in the same region of spacetime. Wald

7
Approximately local particle detectors can be constructed in algebraic quantum field theory; see
Halvorson and Clifton 2002, pp. 21-22 and Halvorson and Mueger 2007, pp. 762-763 for details
and how to make some sense of the notion of a particle.
(1994, p. 116) has argued that this shows that the particle concept is observer-
dependent. Others have argued that since the particle concept has to be
observer-independent, there are no particles (Davies 1984). Using the Unruh
effect, an argument against particles being fundamental entities has been given
a concise formulation by Arageorgis, Earman and Ruetsche:
(A1) If the particle notion were fundamental to QFT, there would be a matter
of fact about the particle content of quantum field theoretic states.
(A2) The accelerating and inertial observers differ in their attributions of
particle content to quantum field theoretic states.
(A3) Nothing privileges one observers attributions over the others.
(C4) Therefore, there is no matter of fact about the particle content of
quantum field theoretic states (from (A2) and (A3)).
(C5) Therefore, the particle notion is not fundamental (from (A1) and (C4)).
(Arageorgis et al. 2002, p.166)
8

It has been proven (Clifton and Halvorson 2001) that the accelerated
observer will never count only a finite number of particles in the inertial
observers vacuum state and that the inertial observer will never count only a
finite number of particles in the accelerated observers vacuum state. Clifton
and Halvorson (2001) argued that the Unruh effect shows that the accelerated
observer and the inertial observer have different complementary particle
concepts. The different counting results for the two observers are due to the
fact that the accelerating observers Fock space and the inertial observers
Fock space are unitarily inequivalent to each other.
9
The topic of unitarily
inequivalent representations is one of the most important topics in the
philosophy of QFT, and we will discuss it in more detail later. If Clifton and
Halvorsons idea is to add a new complementary particle concept for each
unitarily inequivalent representation, then the existence of a continuum of
unitarily inequivalent representations would suggest that there would be a
continuum of different complementary particle concepts. Unless other
restrictions on the particle concept are imposed, the proliferation of different

8
Arageorgis, Earman, and Ruetsche (2002) do not think that the Unruh effect undermines the
particle notion in QFT and they attack this argument by challenging (A3). However, they do agree that
particles do not have fundamental status in QFT.
9
See Clifton and Halvorson 2001 for details.
complementary particle concepts might be another argument in favor of
abandoning the effort to shoehorn in a particle concept into QFT.

2.2.2. INTERACTING QUANTUM FIELD THEORIES
The concept of quanta as being countable entities has also come under attack
in the case of interacting quantum field theories (Fraser 2008). The Fock
space contains a number operator, which can count the number of quanta in a
state, but the Fock space can only be used for free (noninteracting) quantum
fields. However, mathematically rigorous interacting quantum field models,
such as the
4
and Yukawa interactions in two and three dimensions, cannot be
interpreted as having states containing aggregable particles or superpositions
of particles. The Fock space for the free field is unitarily inequivalent to the
interacting field a result sometimes referred to as Haags theorem.
10
While
the Fock space constructed for the free field has a total number operator, the
Fock space constructed for the interacting quantum field theories do not have a
total number operator that can count particles.

2.2.3. THE REEH-SCHLIEDER THEOREM
One of the consequences of the Reeh-Schlieder theorem is that local
measurements are not able to distinguish the vacuum state from an n particle
state. If particles or quanta are countable entities, then this result shows that
there is no way to locally distinguish between two states that have different
numbers of particles. The notions of localizability and countability associated
with particles are also jointly ruined by the Reeh-Schlieder theorem because
one of its corollaries is that no local number operators exist (Halvorson and
Mueger 2007, p. 762). One last problem for the notion of particles in QFT,
which is defined on a flat (non-curved) spacetime called Minkowski spacetime,
is that there does not appear to be any way to have a particle interpretation of
states in curved spacetimes (Wald 1994), which are crucial for any future
theory of quantum gravity.

10
See Fraser 2008 for details.
3. THE PROBLEMS WITH FIELDS
Both Malament (1996, p. 1) and Halvorson and Clifton (2002, p. 23)
conclude that talk about particles must be understood in terms of the
properties and interactions of quantum fields. How are fields and particles
different? Here are two key differences that are often cited.
Particle: localized / discrete, finite number of degrees of freedom
Field: non-localized / continuous, infinite number of degrees of freedom
While a particle is supposed to be a discrete, localized entity, a field is defined
on every point in spacetime; a field is a continuous entity. A field is not
contained in any particular region which is a proper subset of all of spacetime.
A point-particle can be described by its position and momentum in three
dimensions, which implies that a single point-particle has six degrees of
freedom. Since the field is defined on each spacetime point, the field has
properties at each point, such as the value of the field at that spacetime point.
Each property of the field at each spacetime point is a degree of freedom. Since
spacetime has a continuum of points, the field has an infinite number of
degrees of freedom. These properties or field values can be assigned a scalar,
vector, or tensor at each spacetime point. Fields often must also satisfy a field
equation(s) such as the Klein-Gordon equation or Maxwells equations. What
is most important for the purposes of this paper is that a field has been viewed
as essentially an assignment of properties to spacetime points.
The argument for a field ontology often has the following implicit form:
(A1) QFT can only be interpreted in terms of particles or fields.
(A2) The No Go theorems in the previous section show a particle
interpretation is not possible for QFT.
(C3) Therefore, QFT only admits an interpretation in terms of fields.
11

There are other alternative ontologies, which undermines (A1).
12
A more
compelling argument for why fields are the fundamental entities of QFT comes
from what is called field quantization. Heuristically, this involves taking a

11
See Huggett 2000 and Teller 1995.
12
Some of the proposed alternative ontologies for QFT include events, processes, or tropes; see
the articles collected in Kuhlmann et al. 2002. For the purposes of this paper, I will focus on what type
of field ontology is consistent with the mathematical structures and physical assumptions of QFT.
classical field theory, such as electromagnetism, and replacing the classical
field with a quantum field. The procedure for generating a quantum field from a
classical field is similar to the quantization of a classical theory: take the
classical observables of position and momentum, promote them to operators,
and impose the CCRs. Similarly, we take the classical field () and its
conjugate momentum (), promote them to operators
() and (), and

impose the (equal-time) CCRs.
13

However, there are important differences between a classical field and a
quantum field. Teller (1995, pp. 94-97) has argued that quantum fields,
unlike classical fields, do not assign definite values of physical properties to
each spacetime point. Rather, a quantum field assigns a determinable to each
spacetime point (Teller 1995, p. 95). For example, having mass is a
determinable property which has no specific value. The property of having a
mass of five kilograms is a determinant property that is, it is a specific value
of a determinable property. A field configuration for a determinable is a
specific assignment to each spacetime point of the value of a determinable. A
field configuration for the classical electromagnetic field assigns definite values
for the electric and magnetic field to spatial points as well as a direction of the
field at each spatial point.
While a field configuration can be specified for classical fields, Teller
(1995, p. 101) argues that quantum fields do not by themselves constitute a
field configuration. The reason that quantum fields only assign a
determinable(s) to each spacetime point is that they are operators. More
precisely, they are operator-valued fields.
14
Operators are mathematical
entities that do not represent definite values of a physical quantity. The
eigenvalue spectrum of an operator is a list of the possible specific values for
that property.
() is an observable. It has possible specific values, but no

specific value is assigned to each spacetime point by the field operator
()
alone.
does not by itself specify a field configuration. According to Teller

(1995, p. 101), a field configuration requires
() and states. Given a state

, an expectation value
() for all possible products of
(),

13
The CCRs are used for free bose fields. The canonical anti-commutation relations are imposed
for free fermion fields. For the purposes of this paper, I will focus on free neutral scalar Bose fields.
14
We will learn later that quantum fields cannot be thought of as operator-valued fields due to
various No Go results.
where
is evaluated at arbitrarily chosen spacetime points,

constitutes a field configuration (Teller 2002, p. 145). It is these expectation
values that assign specific values of the field determinable to spacetime
points.
15
The actual state is a contingent fact. Thus,
(), by itself, does not

encompass all of the ontology of QFT.
There are problems with Tellers notion of a field configuration.
Expectation values are an average expected value. No specific actual value of
the field is being assigned to the spacetime point x by
, which
Teller (1995, p. 101) acknowledges. If field configurations require the
assignment of actual values to each spacetime point, then

would have to be
an eigenvector of
() for every spacetime point! The problem is that while

the eigenvalue spectrum of the field operator is invariant, field operators at
different spacetime points such as
() and
() (where ) will have

different eigenvectors (Wayne 2002, p. 129), so one state by itself will not be
an eigenvector for the field operators at every spacetime point. We could
salvage a notion of a field configuration by defining an average field
configuration as the assignment to each spacetime point the expectation value
given a state

. When are two average field configurations
different? Two average field configurations are different when their
expectation values differ in at least one spacetime point. This will be important
later when we discuss unitarily inequivalent representations. It is the states and

that are the basis of the wavefunctional interpretation of QFT.
16

In the wavefunctional interpretation, the free quantum field operators,
which satisfy the CCRs, have states defined on a Hilbert space of
wavefunctionals (). The states can be interpreted as superpositions of
classical field configurations in the way that states in Fock space are
interpreted as superpositions of classical configurations of particles. Each state
is a probabilistic propensity of a certain classical field in the event of a
measurement. Expectation values of
for a particular state give the mean

expected value of the classical field strength at x. However, there are significant

15
The reason Teller includes all possible products of the field operators is due to a criticism of
Waynes (2002) that the content of the field operators can be reconstructed from n-point vacuum
expectation values, which involve n field operators defined at n different spacetime points. That
reconstruction theorem is proved in Wightman and Streater 2000, pp. 117-126.
16
My explanation of the wavefunctional interpretation below comes largely from Baker 2009 and
Halvorson and Mueger 2007.
obstacles for the wavefunctional interpretation. For example, interpreting the
states in the Hilbert space of wavefunctionals as probability distributions over
classical field configurations is ruled out because determinate field
configurations are identified with the zero vector in the Hilbert space
(Halvorson and Mueger 2007, p. 779). Thus, there is no state of the quantum
field which is in a specific configuration. Another criticism of the
wavefunctional interpretation given by Baker (2009) is based on the unitary
equivalence of the wavefunctional Hilbert space and the Fock space. Based on
that unitary equivalence, Baker argues that the Malament-Clifton-Halvorson
arguments used against particles being localized at or around points and
Frasers argument that rigorous forms of interactions in QFT cannot be given a
quanta interpretation can be used as arguments against the wavefunctional
interpretation. The appearance of unitarily inequivalent representations also
raises substantial problems for the wavefunctional interpretation, according to
Baker. However, a modified field ontology is possible without assuming the
wavefunctional interpretation.

3.1. NO GO RESULTS FOR

One problem for Tellers account of quantum fields is that no non-trivial field
operator

exists. There are a number of No Go theorems that show that no
non-trivial field operators can be defined on spacetime points.
17
To make the
field operators mathematically well-defined they must be smeared across
spacetime regions. The smearing works by convoluting the field operator
at a spacetime point x with a test function f defined on a finite spacetime

region O that includes the point x. The test function f is a smooth function
with compact support, which means that the function is zero outside the region
O and can be non-zero inside O. The quantum field is no longer an operator-
valued field, but an operator-valued (tempered) distribution:

17
See Halvorson and Mueger 2007 for details about the various No Go theorems. It is possible
to define a quantum field at spacetime points if the field is a sesquilinear form not an operator
(Halvorson and Mueger 2007, pp. 774-777). Roughly, every operator defines a sesquilinear form,
but it is not clear if a sesquilinear form admits a representation as an operator. There is a further
question of whether a sesquilinear form can represent a physical quantity or whether a sesquillinear
form can be thought of as a field defined at a spacetime point. A sesquilinear form can be defined in the
Wightman axiomatic formulation of QFT, but it is not clear whether these results hold in the algebraic
approach to QFT.
is a linear map from test functions to

operators and
represents the average field value in the region O.

18
Thus,
the quantum field is no longer an operator defined on spacetime points, but an
operator-valued defined on a finite spacetime region.
Using smeared fields requires other mathematical changes. The CCRs for
the smeared fields will have a slightly different form involving the inner product
of test functions. The expectation values for
() can also be arbitrarily large

for certain states (Streater and Wightman 2000, p. 97), which suggests that
() behaves like an unbounded operator. An unbounded operator is not

defined on every state in the Hilbert space. If the expectation value of
(), in
a certain state, is infinite, then the field operator cannot be defined on that
state. Infinite field strengths are supposed to be unphysical because they would
require an infinite amount of energy. Thus, even if the field operator is
smeared out around the spacetime point x by a test function and a state is
chosen to represent some physically contingent fact, the expectation value may
not yield an average value for the field because the expectation value is infinite.
In cases like that, no average field configuration is possible.
One way to deal with domain questions is to use the Weyl form of the CCRs
for the field operators. This makes the field operators bounded, which removes
the problem of specifying the domain on which the operators are defined.
Bounded operators are defined on all states in the Hilbert space. The Weyl
operators () are constructed by taking an exponential of the field operator
(): i.e., =
()
. These operators generate a C*- algebra called the
Weyl algebra, which allows us to use the powerful mathematical framework of
algebraic quantum field theory (AQFT). I will discuss AQFT more in section
four.

3.2. UNITARILY INEQUIVALENT REPRESENTATIONS
The CCRs are formal constraints. Many operators will satisfy them. For
example, we can create a new field operator
() by essentially adding a

18
We are smearing the fields in both space and time. For a free Bose field, it is only necessary to
smear the field over space alone by test functions. Some authors have suggested that the fields must be
smeared in space and time in the case of interacting fields, but there are no theorems that prove that
interacting fields must be smeared in space and time (Halvorson and Mueger 2007, pp. 779-780).
complex constant c multiplying the identity operator I:
+.
19

This new field
() will also satisfy the CCRs. Are these two fields
() and
() different determinables? In one sense, they are not. They both generate
isomorphic Weyl algebras that is, they both have the same abstract algebra,
but their form as operators acting on a Hilbert space are different. Does that
make any difference with respect to their expectation values? One way answer
this question precisely is to determine whether the representation
of
()
and the representation
of
() are unitarily equivalent. When two

representations are unitarily equivalent, there is a bijective mapping between
the set of observables belonging to both representations and a bijective
mapping between the set of states belonging to both representations. Another
consequence of unitary equivalence is that two unitarily equivalent
representations are empirically equivalent, i.e., they have the same expectation
values. However, if they are unitarily inequivalent representations, they do not
have the exact same expectation values.
20
Thus, they are not the same average
field configuration.
Each representation has a set of states defined on its associated Hilbert
space and each of those states can be used to define an average field
configuration. Thus, a representation is a collection of average field
configurations. If the representations
and
are unitarily inequivalent,

then the wavefunctional Hilbert spaces associated with them are also unitarily
inequivalent. Assuming that both representations are irreducible (i.e., that
there are no non-trivial subrepresentations), the wavefunctional spaces have no
states in common and have disjoint collections of possible average field
configurations.
21
Once we start using the Weyl algebra and discussing
representations, we have the tools of AQFT at our disposal. We shall see that
the field ontology developed thus far is further modified in AQFT.

19
The complex constant involves an integration of the test function f (Baker 2009, p. 597).
20
The expectation values for two unitarily inequivalent representations of C*-algebra can be
weakly equivalent to each other. Roughly, weak equivalence means that the expectation values in one
representation can be approximated in a unitarily inequivalent representation. For a full discussion of
the issue and the limitations of weak equivalence, see Lupher 2008.
21
The folium of an abstract state is the set of all abstract states which can be expressed as
density operators defined on s representation. Two irreducible unitarily inequivalent
representations have folia that are disjoint from each other, i.e., they have no abstract states in
common.
4. AN ONTOLOGY FOR QFT
Lets see where we are. We started with the idea that a field is an assignment of
properties, which can be represented by scalars, vectors, or tensors, to
spacetime points. Teller argued that the quantum field operators assign
determinables (not properties) to spacetime points. However, the No Go
theorems discussed in the last section show that the quantum field cannot be an
operator defined on spacetime points; the quantum field operator must be
smeared with a test function over a finite spacetime region. The appropriate
modification of the field ontology in QFT involving smeared quantum fields
is that a quantum field is the assignment of a determinable to a finite spacetime
region. This modified field ontology is consistent with Wightmans axiomatic
approach to QFT and the No Go results for quantum fields. However, the
unbounded smeared field operators are mathematically difficult to use.
Bounded versions of the smeared field operators can be used to construct the
Weyl algebra and that allows us to use the powerful mathematical tools of
AQFT to further explore the field ontology developed so far.
Algebraic quantum field theory (AQFT) also assigns determinables to
spacetime regions by mapping an algebra of observables

to an open
bounded region of spacetime O. The mapping () of these open
bounded regions O to an algebra of observables () generates a net of
algebras. Setting up a net of algebras satisfies the basic field ontology
discussed: it is an assignment of determinables to open bounded spacetime
regions. Different nets ()

can be constructed by using different
open bounded regions. They are different ways of carving up spacetime into
finite regions. The resemblance between AQFT and smeared fields is more
than superficial. Given a field smeared by test functions having support in the
open bounded region O, a von Neumann algebra on O can be generated.
22
If
we work with the Weyl form of the smeared fields, then a net of C*-algebras
over spacetime can be defined (Halvorson and Mueger 2007, p. 760).
To what extent does the notion of a field configuration change in the
algebraic approach? If we are interested in the assignment of definite values of
observables, then a partial field configuration is possible in AQFT. An algebra
of observables can contain a commutative subalgebra, which is a collection of

22
For some of the connections between smeared fields and AQFT and further references, see
Haag 1996, pp. 105-106.
observables that commute with every element of the entire algebra. These
observables in the commutative subalgebra could be considered the classical
observables in the algebra. It can be proven that every abstract state of the
algebra will have a definite value for each of these classical observables.
Thus, once the algebra of observables is mapped to a particular open
bounded region of a spacetime O, a partial field configuration can be given for
these classical observables. Call this a classical field configuration, which
is a kind of partial field configuration in that definite values for some of the
observables can be assigned to an open bounded spacetime region. States are
not necessary to have a classical field configuration.
There are many different types of classical observables such as electric
charge, chemical potential, mean magnetization, macroscopic order
parameter, chirality, the algebra of observables at infinity, which can be
examined in AQFT. Temperature plays a role in explaining the Unruh effect.
When the inertial observers vacuum state is restricted to the right or left
Rindler wedge, the state is a thermal state and has a temperature (Wald 1994,
p. 115). For a particular representation of the abstract algebra, there may be
no non-trivial classical observables.
23
If such classical observables do exist
and they are relevant for describing the system under investigation, they should
certainly be part of the field ontology just as temperature is part of the ontology
involved in the Unruh effect. Another advantage of having classical observables
as part of our field ontology is that they provide a way of distinguishing
between different unitarily inequivalent representations. Roughly, two
different unitarily inequivalent representations (more precisely, two disjoint
factor representations) will maximally differ in their expectation value for at
least one classical observable (Lupher 2008).
24
However, classical
observables are only a part of the field ontology discussed so far. Now we have

23
If the von Neumann algebra is a type III
1
factor, which is the predominant algebra for open
bounded spacetime regions, then the center of the algebra consists of scalar multiples of the identity
(Halvorson and Mueger 2007, p. 766). The construction of classical observables may involve using
the central projections of different unitarily inequivalent representations instead of using just one
representation.
24
The classical observables belong to a larger abstract algebra called the bidual, which is a
W*-algebra. Briefly, the observables belonging to the abstract C*-algebra are a subset of the
observables belonging to the bidual. A nontrivial classical observable in the bidual may take the form
of the zero operator in the Hilbert space of a particular representation while in another representation
the classical observable may be a non-zero operator in that representations associated Hilbert space.
to look at the quantum observables which belong to the non-commuting part of
the algebra of observables.
Once the algebra of observables has been specified for a particular open
bounded region of spacetime, a representation and a Hilbert space can be
constructed through the GNS theorem. This provides a realization of the
operators of the algebra as bounded operators acting on a Hilbert space and a
set of states in the Hilbert space. The state of the system will be an eigenvector
for some of the (non-commuting) observables in the Hilbert space, which gives
us a collection of (quantum) beables.
25
That provides a more complete field
configuration than the classical field configuration described above. A
different state in the Hilbert space will have a different set of quantum beables,
but the same classical beables. Every state in the Hilbert space will have the
same value for a particular classical beable. For example, a representation of an
equilibrium state at temperature T will have a Hilbert space in which each state
has a temperature T. Classical beables and the quantum beables for a particular
eigenstate give definite values. Thus, a beable field configuration assigns
specific values of classical and quantum beables to an open bounded region of
Minkowski spacetime. There are still a number of quantum observables for
which the Hilbert space state will not give definite values. For those quantum
observables and a particular state in the Hilbert space, we can only compute the
average expectation value. We would then have definite values or average
expectation values for all of the observables, which would provide the most
complete notion of a quantum field configuration assuming the net of algebras
has already been chosen.
There are a number of worries about underdetermination that can be raised
about this field ontology. (1) Is there more than one type of abstract algebra?
(2) Is there a preferred net of algebras? These are important philosophical
questions to which I will make a few brief comments. With respect to (1), as I
discussed in section three, all of the different concrete representations of the
Weyl relations give rise to the same abstract algebra: the Weyl algebra.
26
In

25
The term beable was originally introduced by Bell (1987). A beable is an observable that has
a determinate value. Though Bell used the term in the context of quantum observables, classical
observables also qualify as beables since they have determinate values.
26
AQFT typically assumes that the abstract algebra contains all bounded operators. The Weyl
algebra contains bounded versions of the field operators. One might worry that AQFT is leaving out
important unbounded operators such as position and momentum. Self-adjoint (possibly unbounded)
operators can be affiliated with a von Neumann algebra by considering the operators family of spectral
that case, the abstract algebra is not underdetermined.
27
With respect to (2),
while there is freedom to choose different nets, there are constraints on the
net. For example, the net must satisfy the axioms of isotony and microcausality.
The choice of a net describes a particular way of carving up spacetime
according to regions and assigning determinables. Haag (1996, p. 105) claims
that fields are ontologically dispensable since a field corresponds to a particular
coordinatization of the net of algebras, i.e., a particular mapping ().
Different nets () are possible, but there may be no physical
differences between () and (). Support for this point of
view comes from the concept of Borchers classes. Different nets may belong to
the same Borchers class and thus have the same S-matrix. If that is correct, then
choosing a particular net would lack any physical significance. It would be
similar to the lack of physical significance in the choice of Cartesian
coordinates to solve a particular problem instead of using polar coordinates.
However, that view may be incorrect because there may only be one possible
net in certain circumstances. The Doplicher-Roberts reconstruction theorem
shows that there is a unique field net and gauge group that are compatible with
the algebra of observables and its vacuum state (Halvorson and Mueger 2007,
p. 849).
28
One of the remarkable achievements of AQFT is that the net of
algebras is sufficient to uniquely reconstruct the fields and the gauge group
including items such as isospin, baryon number, and other observables.
29

Thus, there is no underdetermination of the net of algebras in terms of the
reconstruction of what is called the field algebra. The field algebra, which
includes the gauge group, is generated by local algebras whose elements
represent local fields that have excitations within a bounded spacetime region.
A field algebra and a gauge group acting on a Hilbert space give a preferred set
of representations, i.e., those that can be created from the vacuum state by the
action of local fields.
30

projections (Clifton and Halvorson 2001, p. 424). There still remains the question of whether to work
with abstract C*-algebras or W*-algebras. For reasons why an abstract W*-algebra should be
preferred, see Lupher 2008.
27
Unitarily inequivalent representations show that there is not a unique representation of the
abstract algebra.
28
For a discussion of when two field systems are theoretically equivalent, weakly observationally
equivalent, or observationally equivalent, see Halvorson and Mueger 2007, section 11.3.
29
See section 10 of Halvorson and Mueger 2007 for details.
30
The Doplicher-Roberts reconstruction theorem and the Doplicher-Haag-Roberts treatment of
superselection sectors are discussed in more detail by Halvorson (Halvorson and Mueger 2007).
5. CONCLUSIONS
While there are significant challenges in finding a suitable ontology for QFT, a
modified field ontology is consistent with mathematically rigorous versions of
QFT such as Wightmans axiomatic QFT, which uses smeared quantum
fields, and AQFT. The field ontology, which assigned properties to spacetime
points, has been modified in QFT. The new modified field ontology assigns
determinables to open bounded regions of spacetime rather than spacetime
points. It shows how Tellers account of a field is modified by the No Go results
involving quantum fields that are defined on spacetime points and how classical
and quantum observables impose changes on our understanding of a field
configuration. There remain many open questions. The ontology discussed so
far did not discuss the role of dynamics, superselection sectors, nor
interactions. The answers to these questions will further illuminate the nature
of the field ontology in QFT.

REFERENCES
Arageorgis, A., Earman, J., & Ruetsche, L. (2002). Fulling Non-Uniqueness
and the Unruh Effect: A Primer on Some Aspects of Quantum Field
Theory. Philosophy of Science, 70(1), 164-202.
Baker, D. (2009). Against Field Interpretations of Quantum Field Theory.
British Journal for the Philosophy of Science, 60(3), 585-609.
Bell, J. S. (1987). Speakable and Unspeakable in Quantum Mechanics.
Clifton, R., & Halvorson, H. (2001). Are Rindler Quanta Real? Inequivalent
Particle Concepts in Quantum Field Theory. British Journal for the
Philosophy of Science 52(3), 417-470.
Davies, P. (1984). Particles Do Not Exist. In S. Christensen (Ed.), Quantum
Theory of Gravity (pp. 66-77). Bristol: Adam Hilger.
Feynman, R., & Weinberg, S. (1987). Elementary Particles and the Law of
Physics : The 1986 Dirac Memorial Lectures. Cambridge: Cambridge
University Press.
Fraser, D. (2008). The Fate of 'Particles' in Quantum Field Theories with
Interactions. Studies in the History and Philosophy of Modern Physics,
39(4), 841-859.
Haag, R. (1996). Local Quantum Physics. Berlin: Springer-Verlag. [1992]
Halvorson, H., & Clifton, R. (2002). No Place for Particles in Relativistic
Quantum Theories?. Philosophy of Science, 69(1), 1-28.
Halvorson, H., & Mueger, M. (2007). Algebraic Quantum Field Theory. In J.
Butterfield & J. Earman (Eds.), Philosophy of Physics (pp. 731-922).
Huggett, N. (2000). Philosophical Foundations of Quantum Field Theory.
British Journal for the Philosophy of Science, 51(4):617-637.
Kuhlmann, M. (2006). Quantum Field Theory. Stanford Encylopedia of
Philosophy.<http://plato.stanford.edu/entries/quantum-field-theory/>
Kuhlmann, M., Lyre, H., & Wayne, A. (Eds.) (2002). Ontological Aspects of
Quantum Field Theory. River Edge, NJ: World Scientific.
Lupher, T. (2008). The Philosophical Significance of Unitarily Inequivalent
Representations in Quantum Field Theory. Ph.D. Dissertation. Austin,
TX: University of Texas.
Malament, D. (1996). In Defense of Dogma: Why There Cannot Be a
Relativistic Quantum Mechanics of (Localizable) Particles. In R. Clifton
(Ed.), Perspectives on Quantum Reality (pp. 1-10). Dordrecht: Kluwer
Academic Publishers.
Streater, R., & Wightman, A. (2000). PCT, Spin and Statistics, and All That.
Princeton, NJ: Princeton University Press.
Teller, P. (1995). An Interpretive Introduction to Quantum Field Theory.
Princeton, NJ: Princeton University Press.
Teller, P. (2002). So What Is the Quantum Field?. In M.Kuhlmann, H. Lyre &
A. Wayne (Eds.), Ontological Aspects of Quantum Field Theory (pp.
145-162). River Edge, NJ: World Scientific.
Wald, R. M. (1994). Quantum Field Theory in Curved Spacetime. Chicago:
University of Chicago Press.
Wayne, A.(2002). A Naive View of the Quantum Field. In M. Kuhlmann, H.
Lyre & A. Wayne (Eds.), Ontological Aspects of Quantum Field Theory
(pp. 127-133). River Edge, NJ: World Scientific.


Relativistic Thermodynamics and the Passage of
Time
*

Friedel Weinert **
f.weinert@bradford.ac.uk

ABSTRACT
The debate about the passage of time is usually confined to Minkowskis
geometric interpretation of space-time. It infers the block universe from the
notion of relative simultaneity. But there are alternative interpretations of
space-time so-called axiomatic approaches , based on the existence of
optical facts, which have thermodynamic properties. It may therefore be
interesting to approach the afore-mentioned debate from the point of view of
relativistic thermodynamics, in which invariant parameters exist, which may
serve to indicate the passage of time. Of particular interest is the use of
entropic clocks, gas clocks and statistical thermometers, which suggest that
two observers in Minkowski space-time could agree on an objective passing of
time.

1. INTRODUCTION
The long running debate about the block universe, taken to be a conceptual
consequence of the Special theory of relativity, usually relies on the discovery
of relative simultaneity for support. As mechanical clocks run differently in
reference frames, moving inertially with respect to each other, the relativity of
simultaneity implies that there can be no universal Now in Minkowski space-
time. Hence the passage of time must be a human construction, in the Kantian
sense, whilst the physical universe is taken to be a four-dimensional block.
Gdel (1949) argued that the relativity theory provided proof of the idealist
view of time, due to the relativity of simultaneity, which implies that the

* The author would like thank an anonymous referee for helpful comments on an earlier draft of
this paper.
** University of Bradford
temporal succession of events loses its objective status. Those who argue
against the block universe, as a conceptual consequence of the Special theory,
usually point out that the space-time interval, ds, is invariant for all observers,
who consider time-like related events (Capek 1966, 1983). The invariance of
the line element, ds, implies that the order of time-like related events is the
same for all observers, even though they cannot agree on the simultaneity of
these events. This debate is usually confined to considerations of the geometric
structure of Minkowski space-time. Only Einstein, in his reply to Gdel,
brought thermodynamics into the debate. In characteristic style, Einstein
resorted to a thought experiment. An electromagnetic signal is sent, within a
light cone, from a past source, A, to a future receiver, B, through a present
point, P. Einstein concludes that the sequence of events is irreversible, since
the emission of the signal happens before its reception at B. This secures the
one-sided (asymmetric) character of time [], i.e., there exists no free choice
for the direction of the arrow (Einstein 1949, p. 687). Following Planck,
Einstein (1907) had established early in his career that entropy is frame-
invariant. Einsteins appeal to thermodynamics raises the question whether
relativistic thermodynamics offers invariant relationships, in addition to ds,
which would allow us to avoid Gdels conclusion.
The purpose of this paper is to consider whether it is possible to build a
thermodynamic clock, based on considerations of invariants in relativistic
thermodynamics, which would lead to a frame-invariant reading of time for two
observers who move inertially with respect to each other at relativistically
significant velocities. If such invariant relationships could be found, they may
be more appropriate for a discussion of the passage of time in Minkowski
space-time. The space-time interval ds is merely a geometric measure in
Minkowski space-time, whilst thermodynamic properties of signal propagation
are physical in nature, a fact to which Einstein alluded in his thought
experiment. The question of temporal becoming in relation to relativistic
thermodynamics has, according to the authors knowledge, not yet been
considered in the literature but it may throw new lights on an old debate.

2. LORENTZ-INVARIANT PARAMETERS
In a thermodynamic system, moving with velocity, v, several thermodynamic
parameters remain invariant. According to Max Planck (1907), the following
invariant relationships hold in relativistic thermodynamics:
Friedel Weinert Relativistic Thermodynamics and the Passage of Time 177
o o o
S S n n p p = = = , , .
Of particular interest is the Lorentz invariance of entropy, S, which can
easily be established by recalling the definition of entropy in statistical
mechanics: N k S log = . The number of microscopic states, N, which
correspond to a given macrostate does not depend on the velocity of the
thermodynamic system, so that
o
S S = . According to this equation two
observers in inertially moving systems would agree on a change in entropy of a
system under observation. Could the invariance of entropy serve to measure
the passage of time, objectively? In considering this question it is useful to
recall that mechanical clocks are not necessarily needed to measure time.
Galileo used simple water clocks in his fall experiments to establish that
distance, d, is proportional to time squared,
2
t . The frame invariance of
entropy means that in both systems, moving inertially with respect to each
other, the entropy of a body in thermodynamic equilibrium is constant and is
not dependent on the velocity of the body (Mller 1972, p. 237). The same
conclusion holds for increases in entropy, oS, for systems in relative motion
(Schlegel 1968, p. 148). Whether the invariance of a change in entropy of a
body as measured in two different reference systems could serve two observers
as a basis for the measurement of time depends on whether we consider
entropy is defined in classical thermodynamics or statistical mechanics.
Imagine that two observers agree to perform an experiment: observer A sits on
a train, which moves at relativistic speeds and observer B sits on the platform.
Let the experiment simply consist in computing the change of entropy when an
ice cube is put into a container with warm water. Lets say that the two
observers agree to perform a certain act when the ice cube has melted but
they would have to agree on the moment when they begin the experiment, for
which they would need a clock. But mechanical clocks are not frame-invariant.
Furthermore the change in entropy for such a system is defined, in classical
thermodynamics, as
=

but have they established that the temperature, , is frame-invariant?
Unfortunately, there is no agreement amongst physicists as to whether
temperature is Lorentz invariant. According to the classic papers by Planck and
Einstein, the proper transformation for temperature is:
=
0
, where = 1
2
,
with the consequence that a moving body would appear cooler (Pauli 1981,
.46). Whilst this view was generally accepted, it has also been proposed that
=
0
, with the consequence that a moving body would appear hotter.
1
In
either case temperature suffers a dilation or contraction effect in the moving
system, similar to the well-known relativistic effects. Such relativistic effects
seem to rule out an invariant clock reading across two inertial systems. But
does this mean that two observers should conclude that time is a human
illusion? Many authors jump to this conclusion:
Due to the absence of an observer-independent simultaneity relation, Special
relativity does not support the view that the world evolves in time. Time, in the
sense of an all-pervading now does not exist. The four-dimensional world
simply is, it does not evolve. (Ehlers 1997, p. 198)
Each observer has his own set of nows and none of the various systems of layers
can claim the prerogative of representing the objective lapse of time. (Gdel
1949, p. 558)
It has not been sufficiently appreciated that the situation, to which the
authors appeal, is not essentially different from many other situations, in which
two different observers need to adopt transformation rules to calculate their
respective perspectival observations. Consider, for instance, how two famous
Greeks, Eratosthenes (276-196 BC), located in Syene, and Archimedes (287-
212 BC), located in Alexandria, would determine midday in their respective
locations. They would use sundials but the distance between the two locations
means that they would not agree on when midday occurs. Nevertheless, they
could use some elementary facts to determine midday in each others location.
When the sun is at its highest point in Syene, it throws a shadow of 712' in
Alexandria, which is approximately 800km to the west of Syene. As the Greeks
knew the rotational speed of the sun around the Earth (on the geocentric
model of the universe) Archimedes and Eratosthenes could calculate midday in
each others location, using appropriate transformation rules. In order to do
so, it is essential that the motion of the sun across the sky is both regular and
invariant for the two observers. As this was the case they had no reason to
conclude that midday is an illusion.

1
See McCrea 1935, .54; Perrot 1998, p. 264; Dunkel and Hnggi 2009.
For the modern physicist the transformation rules are provided by the
Lorentz transformations for the time coordinates:

which indicates the time dilation in moving clocks:

.

It is not obvious why the observers should conclude that the passage of time is a
human illusion, as it is often claimed. The Lorentz transformations are
perfectly invariant and do not depend on the reference frame of the observer.
However, for the two observers to calculate their respective clock times, they
must know the velocity of their respective frames, just like Archimedes and
Eratosthenes must know the distance between their respective locations and
the velocity of the suns motion. The need to know the velocities of the
respective systems so that some transformation rule can be applied leaves a
loophole in the argument. If the observer A does not know the velocity of B
then the clocks tick differently for them but they cannot calculate each others
clock times. In this case time and its passage seem to be truly relative.

3. GEOMETRIC AND THERMAL TIME
It is therefore worth investigating whether there are invariant relationships in
the Special theory, which could serve as a common basis for time. It is
customary for defenders of temporal becoming to point to the existence of the
so-called space-time interval, ds, (
2 2 2 2 2 2
dt c dz dy dx ds + + = ),
according to which the space-time length between two observers, moving
inertially with respect to each other, is invariant (ds = ds'), whilst their clock
readings are subject to time-dilation effects, and relative simultaneity (Capek
1966, 1983, Nehrlich 1982). However, the space-time interval ds, although
invariant, is ill-suited for the measurement of time, because it measures the
space-time length between two events and lets the temporal parameters vary
according to the reference frame.
The velocity of light is also an invariant in Minkowski space-time. However,
we should note a systematic ambiguity in the treatment of c. In the standard
geometric interpretation of Minkowski space-time, c is not a physical signal,
which propagates through space-time at a constant velocity (approximately
s
m
x
8
10 3 ). Rather, c constitutes a geometric limit of the light cones
sometimes taken to represent the causal structure of Minkowski space-time
such that no subluminal particles, which trace their world lines through space-
time, are permitted to reach this limit because of the effect of relativistic mass:

=
0
1
.
According to this geometric representation of Minkowski space-time, time
stands still for light signals, c, which constitute the causal boundaries of the
light cones in Minkowski space-time. On the standard geometric
interpretation this reduces the measurement of time to the frame-dependent
behaviour of mechanical clocks. It is geometric time.
Einstein hints at an alternative reading of c: it is also a physical signal, which
propagates at a constant speed between two events in space-time. Although
Minkowskis geometric approach is the predominant view today, alternative
approaches were developed soon after the publication of the Special theory of
relativity.
2
They are based on the propagation of light signals in space-time. As
Einstein indicates, on these so-called axiomatic approaches, thermodynamic
aspects of optical facts have to be taken into account. These axiomatic attempts
reverse the usual tendency to spatialize time. For instance, Robb starts with
the thesis that spacial relations may be analyzed in terms of the time relations
before and after or, as he concludes, that the theory of space is really a part
of the theory of time (Robb 1914, Conclusion). Essential for this conception
is the notion of conical order, which is analyzed in terms of the relations of
before and after instants of time. An instant (an element of time) is the
fundamental concept, rather than the space-time event. Furthermore the

2
See Robb 1914, Cunningham 1915, Carathodory 1924, Schlick 1917, Reichenbach 1924.
before/after relation of two instants is an asymmetrical relation. In this way
Robb builds a system of geometry, in which we encounter the familiar light
cones of the Minkowski representation of space-time. Robb reverses the
Minkowski approach in terms of geometrical relations and starts from physical
facts, an approach, which is reflected in Einsteins later reservations about the
block universe.
If a flash of light is sent out from a particle P at A1, arriving directly at particle
Q at A2, then the instant A2 lies in the o-subset of instant A1, while the instant
A1 lies in the -subset of A2. Such a system of geometry will ultimately assume
a four-dimensional character or any element of it is determined by four
coordinates. [] It appears that the theory of space becomes absorbed in the
theory of time. (Robb 1914, pp. 8-9)
Here the o-subset is the future light cone of instant A1 and the -subset is
the past light cone of A2. (Figure I)

P(A1)
Figure I: Corresponding to any point in
space, there is an o-cone of the set
having that point as vertex, similarly
there is also a -cone of the set having
the point as vertex.
If A
1
be any point and o
1
the
corresponding o-cone, then any point
A
2
is after A
1
, provided A
1
= A
2
and A
2

lies either on or inside the cone o
1
.

(Robb 1914, 5-6)
o
1

Q(A2
)
The axiomatic approaches raise the question of frame-invariant
thermodynamic properties.
The propagation of c between space-time events has certain limitations for
the measurement of time. Whilst the velocity of light is the same for all
observers in Minkowski space-time, the direction of a light ray depends on the
relative angle between the frame and the light source. This phenomenon is
known as aberration. (Giulini 2005, .3.6) Furthermore if the two observers
exchange light signals between each other they will be subject to the relativistic
Doppler effect, which affects the frequency, v, and wavelength, , of the signal,
respectively.
Taking these relationships into account, observers can still calculate each
others signal properties. For instance, the twins in the famous twin paradox
use light signals as light clocks but as before, the Earth-bound twin needs to
know the velocity of the space ship of his twin brother. He is then able to make
corrections, which will allow him to compute how much time has elapsed for
this brother. The brothers can even send regular signals to each other and
using the formulae for the relativistic Doppler Effect, they can calculate how
many signals will arrive at each end. Once again, on the basis of these
calculations, the twins have no reason to conclude that time is an illusion.
Are there any other invariant relationships, which could be used for this
purpose? It is at this point that we should turn to relativistic thermodynamics,
for it offers a number of invariant relationships, which may be useful to two
observers moving inertially with respect to each other. If we free ourselves
from the assumption that time must be measured by mechanical clocks, which
are subject to relativistic effects, then it is possible to conceive of other ways of
measuring time. As mentioned above, Galileo did not use mechanical clocks to
measure time in his inclined plane experiments but used water clocks, i.e., he
measured the passage of time of the ball down the plane by the flow of water
from a bucket and then weighed the amount of water collected. Is it, for
instance, conceivable to build a gas clock, which uses the invariance of p? Such
a clock could be made of two chambers with a small connecting tube through
which gas can leak from one chamber to the other (Schlegel 1968, p. 137).
Observers in relativistically moving systems know the pressure is invariant
(where pressure is force per unit area), so that the relationship
0
p p = could
be used to coordinate activities in relativistically moving systems. For instance,
when the pressure is the same in both systems, the two observers can perform
certain actions simultaneously. Invariance means that two such gas clocks have
the same rate both at rest and when in uniform motion relative to each other
(Schlegel 1968, p. 137). This requires of course that the two gas clocks are in
the same state and that the beginning of the experiment in the two inertial
systems can be coordinated in a Lorentz-invariant manner, like the exchange of
light signals.
In a similar manner we can return to the Lorentz-invariance of entropy in
statistical mechanics. As the entropy increases in the two clocks, attached to
two inertial systems in relative motion with respect to each other, is invariant,
we must conclude that two such entropy clocks, in relative uniform motion,
will run at the same rate (Schlegel 1968, p. 148). The invariance of entropy
and pressure could therefore serve both to measure the passage of time (by
appropriate period units) and to coordinate events in relativistically moving
systems. Note that these parameters do not require knowledge of the relative
velocities of the systems involved. What about temperature? We have reason to
reconsider temperature, since according to Landsberg, there is no
experimental confirmation of the relationship
o
T T | = and it has to be taken
into account that temperature is a statistical concept (Landsberg 1966, p.
571).
Of particular interest is the question whether a thermodynamic
thermometer could be built, which would indicate the passage of time for two
observers, attached to different reference frames. The possibility of such a
Lorentz-invariant thermodynamic clock, in which the notion of temperature
plays a significant part, is based on the work of P. T. Landsberg. In a series of
publications, Landsberg (1966, 1967, 1978) critically discussed the
prevailing view, mentioned above, that temperature is not Lorentz-invariant in
relativistic thermodynamics. The crucial idea that the passage of time must be
linked to physical phenomena, in particular thermodynamic phenomena,
motivates much of the research in the area of the emergence of time from
fundamental symmetry. It gives us a notion of thermal time:
[] it is the statistical state that determines which variable is physical time, and
not any a priori hypothetical flow that drives the system to a preferred
statistical state. When we say that a certain variable is the time, we are not
making a statement concerning the fundamental mechanical structure of
reality. Rather, we are making a statement about the statistical distribution we
use to describe the macroscopic properties of the system that we describe
macroscopically. The thermal time hypothesis is the ideal that what we call
time is the thermal time of the statistical state in which the world happens to
be, when described in terms of the macroscopic parameters we have chosen.
(Rovelli 2008, p. 8)
Landsbergs hypothesis about the invariance of temperature has recently
received renewed attention, because of the importance of a relativistic
equilibrium velocity distribution in the description of several high-energy and
astrophysical events (Chacn-Acosta et al. 2009, Br and Vn 2010).

4. LORENTZ-INVARIANT TEMPERATURE
Going against the established views, Landsberg
3
argues that temperature is
invariant (T
o
= T) and some recent theoretical results and computer
simulations seem to confirm this view. To develop the argument, Landsberg
assumes a transformation law
o
a
T T = , according to which a body has
temperature T
o
in an inertial rest frame, I
o
, and an identical body v has
temperature
o
a
T in a moving inertial frame, I. If these bodies are allowed to
interact, one expects heat flows between them. If that is the case, then
observers in the respective frames will see the other body as hotter if a>0, and
cooler if a<0, respectively. For instance, if a<0, then in s rest frame, s
temperature should increase at the expense of vs temperature (Landsberg
1978, p. 348). In s rest frame, this heat flow would take the form:
+ |
o o
T T
v
,
and in vs rest frame, by the principle of relativity, it would take the form:
+ |
o o
T T
v
, .
The reason for this inconsistency is that, according to the Second law of
thermodynamics, heat flows are involved when two bodies are brought into
thermal contact, whereas no heat flows occur between moving clocks. In order
to avoid the inconsistency, Landsberg sets a=0, so that T=T
o
. If this argument

3
Landsberg 1978, ch. 18.5; Landsberg 1966, 1968; cf. C. Mller (1972, p. 233). In other,
more recent approaches, temperature has been used to connect geometric and thermal time: We now
interpret temperature as the ratio between the thermal time and the geometrical time, namely = .
Temperature is =
1
. Whilst the thermal time flow is state-dependent, the authors also

consider a state-independent notion of time flow (Connes and Rovelli 1994, pp. 20-21).
is correct then it should be possible in principle to construct a thermodynamic
clock, in which the temperature remains invariant since it is unaffected by the
relativistic motion of bodies. Then a thermostat, on Landsbergs hypothesis,
would read identical temperatures and temperature readings could be used as
primitive clocks. Then, if certain conditions are satisfied, the rise of the
temperature in the test bodies, placed in the system at rest and v in the
system in motion, could be used to measure the passage of time. Equally, when
the two bodies, and v, reach the same temperature in the two inertially
moving systems, the observers could perform some prearranged action
simultaneously. If temperature is indeed Lorentz-invariant, they would know
that they perform their actions simultaneously. Of course the thermostats have
to be coordinated at the beginning of the experiment.
As mentioned above no agreement has been reached about the existence of
a Lorentz-invariant temperature. The problem is that none of the formulations
proposed are covariant and in order to proceed we need a generalized
equipartition theorem. (Landsberg 1967, p. 904) Recent work has
emphasized that the different relativitistic temperature concepts are based on
different assumptions about thermodynamic quantities, like heat, energy
transfer etc., which lead to different transformation rules for temperature. One
way to identify these assumptions is to consider from which hyperplane in
Minkowski space-time thermodynamic parameters are defined (Dunkel and
Hnggi 2009; Br and Vn 2010). As this procedure gives rise to different
ways of identifying thermodynamic quantities with statistical averages, it is
desirable to obtain a covariant formulation of the equipartition theorem and the
Jttner distribution the relativistic generalization of the Maxwell distribution.
The aim of such a covariant formulation is to show how an invariant
temperature can be derived from it, thus avoiding the question of which
Lorentz transformation to adopt for temperature. Mathematically this
procedure involves the use of four vectors, in terms of which the invariant
temperature is defined (Chacn-Acosta et al. 2009; Br and Vn 2010).
What makes this work interesting is that recent computer simulations are
seen as an experimental confirmation of Landsbergs hypothesis (Cubero et al.
2007). Temperature, according to these findings, can be measured in a frame-
independent way. Thus moving bodies appear neither cooler nor hotter,
offering the possibility of constructing a thermodynamic clock, which would
tick at the same rate for all inertial observers. The authors first experimentally
confirm that the correct generalization of the Maxwell velocity distribution in
Special relativity is the Jttner distribution (1911). The Jttner distribution
revises the Maxwell distribution function to accommodate upper bounds on
the velocity of atoms. It also yields a well-defined concept of temperature in
Special relativity. As Landsberg had already indicated when two systems are
brought into contact they approach a thermodynamic equilibrium state; each
subsystem is described by the same velocity distribution function
(,
, )
and they only differ in their rest masses but share the distribution
parameter
, which is determined from the initial energy. This parameter,

which already appears in the Maxwell distribution function =
can
then be used to define a relativistic equilibrium temperature ( =
)
1
.
(This requires that the system be spatially confined.) In the experiments
j
|
had the value 0.702(
1
2
)
1
. Adopting = (
)
1
as a reasonable
definition of temperature, the question arises how the respective observers can
measure it. The authors suggest that a moving observer with rest frame '

can measure T by exploiting a Lorentz invariant form of the equipartition
theorem (Cubero et al. 2007), in which the velocities of the particles are
averages with respect to the '
frame. On this basis the authors conclude that
a Lorentz-invariant gas thermometer on a purely microscopic basis (Cubero
et al. 2007) has been defined.
Put differently, this intrinsic statistical thermometer determines the proper
temperature of the gas by making use of simultaneously measured particle
velocities only; thus, moving bodies appear neither hotter nor colder. (Cubero
2007, pp. 170601-4)
If these findings are correct what follows for the measurement of time from a
Lorentz-invariant gas thermometer or gas clock? If the temperature of a body is
not dependent on its state of motion coffee on a fast-moving train has the
same temperature as coffee on the platform then it is imaginable that two
observers moving inertially with respect to each other can employ the statistical
thermometers to measure objectively the passage of time. Equally entropic
clocks or pressure gas clocks are relativistically invariant. Philosophically,
these findings have serious implications for the static view of time, associated
with the block universe.

5. CONCLUSION
As mentioned above, in these considerations the passage of time is intimately
related to physical, in particular thermodynamic processes. One conclusion to
be drawn from these results is that both invariance and regularity are essential
if relativistically moving observers are to agree on the passage of time, as
measured by some clock. In mechanical clocks, in a relativistic context, the
ticking is regular in each inertial system, but the rate of ticking is dependent on
the velocity of the system. Hence the invariance criterion fails. But invariance
does not fail in the gas clocks, considered above, which are based on the
Lorentz-invariant parameters of pressure and entropy. Hence their ticking
rates remain the same. In a Lorentz-invariant thermometer, based on velocity
averages, the temperature rises in a regular fashion in the same way in both
systems, hence temperature is Lorentz-invariant. In a similar way without
knowledge of the regular motion of the sun, Archimedes and Eratosthenes
could not compute the occurrence of midday at their respective locations. With
respect to their different locations, the regular motion of the sun is invariant.
According to the standard geometric interpretation of Minkowski space-time,
there is no universal Now and both the simultaneity of relativity and the time
dilation effect make clock time relative to the state of motion of the system. But
if time is what the clock tells the observers according to their respective frames,
the passage of time, so the argument runs, cannot be an objective feature of the
physical world. Space-time is a four-dimensional manifold and the passage of
time must be a human impression, due to the perspectival slicing of space-
time, according to particular coordinates of the respective observers. As was
pointed out above, it is hard to understand why this conclusion should follow,
given the existence of Lorentz transformation rules, which allow the respective
observers to compute their respective times. It is true that the mechanical
clocks indicate the time relative to the velocity of the inertial system, but from
this relativity it does not follow that time is merely a human illusion, only that
the ticking rate is not invariant. The velocities of the respective systems must
be known, if the observers want to compute the ticking of each others clocks.
If these quantities were not known, geometric time would be completely
relative. But as opponents of the block universe point out, there are other
Lorentz-invariant phenomena. The Lorentz-invariance of entropy, pressure
and temperature allows us to envisage the construction of clocks, whose
ticking rate remains the same, irrespective of the inertial motion of the system.
Such Lorentz-invariant clocks are not in contradiction to the well-known
results of the Special theory of relativity,
for there is nothing in the principle of relativity that specifies what particular
physical variables, or functions of variables, shall be Lorentz-invariant. We are
required to drop the prescription that any t variable, for no matter what
process, is subject to a Lorentz transformation with change of relative state of
motion. Time, instead of being a substratum entity which controls all physical
phenomena, must now be regarded as a concomitant or measure of physical
process []. (Schlegel 1977, p. 252; italics in original)
Traditionally, objections to the inference from the results of the Special theory
of relativity to the human illusion of the passage of time have appealed to
parameters like ds and c; however, these objections still remain within the
standard geometric interpretation of Minkowski space-time. But the
measurement of time must be based on physical phenomena, a conclusion,
which in the present context is emphasized by the so-called axiomatic
approaches. They are based on optical signals, which involve thermodynamic
properties. Users of a Lorentz-invariant thermodynamic clock could infer that
the passage of time is objective, since it is based on Lorentz-invariant
parameters, like p, S and T. But it remains true within the Special theory of
relativity that time is what the clock tells. They use thermal time. Whilst the
standard inference to the unreality of time in Minkowski space-time relies on
frame-dependent parameters, the inference to the reality of time in Minkowski
space-time relies on frame-independent properties. But it seems to be no less
legitimate than the first inference. This situation suggests that inferences as to
reality or unreality of the passage of time in Minkowski space-time are
conceptual in nature. Claims about a static block universe or a dynamic space-
time do not follow deductively from the principles of the Special theory of
relativity. The conclusions we draw seem to depend on the criteria we choose
for drawing these inferences. As the inferences are conceptual in nature, the
scientific theory, from which they are drawn, cannot conclusively falsify one
inference at the cost of its rival. But the certainty, with which proponents of the
block universe infer from the Special theory that time is a human illusion is
thrown into doubt by the existence of invariant thermodynamic clocks in
Minkowski space-time.
What is at issue is not the validity of the Lorentz time transformation, which has
been amply demonstrated, but whether its purview is an entire metaphysical
time, suffusing all of nature, or only an aspect, temporal in its properties, of
physical interactions. (Schlegel 1977, p. 252)
The possibility of Lorentz-invariant clocks, within the Special theory, is
incompatible with the inference to the block universe. Whilst a physical theory
does not have the resources to prove or disprove conceptual inferences, which
are drawn in its name, it remains entirely possible that one inference is more
compatible with the underlying theory than its rival if all the facts are taken into
account.

REFERENCES
Br, T. S., & Vn, P. (2010). About the Temperature of Moving Bodies.
Europhysics Letters 89(3), 30001, 1-6.
apek, M. (1966). Time in Relativity Theory: Arguments for a Philosophy of
Becoming. In J. T. Fraser (Ed.), The Voices of Time (pp. 434-454).
New York: George Braziller.
apek, M. (1983). Time-Space rather than Space-Time. Diogenes, 31(123),
30-48.
Carathodory, C. (1924). Zur Axiomatik der Relativittstheorie.
Sitzungsberichte der Preussischen Akademie der Wissenschaften,
Physikalisch-Mathematische Klasse 5, 12-27.
Chacn-Acosta, G., Dagdug, L., & Morales-Tecotl, H. A. (2009). On the
Manifestly Covariant Jttner Distribution and Equipartition Theorem.
<http://arxiv.org/abs/0910.1625>
Connes, A., & Rovelli, C. (1994). Von Neumann Algebra Automorphisms and
Time-Thermodynamics Relation in General Covariant Quantum
Theories. <http://arxiv.org/abs/gr-qc/9406019>
Cubero, D., Casado-Pascual, J., Dunkel, J., Talkner, P., & Hnggi, P. (2007).
Thermal Equilibrium and Statistical Thermometers in Special Relativity.
Physical Review Letters 99(17), 17601, 1-4.
Cunningham, E. (1915). Relativity and the Electron Theory. New York:
Longmans, Green & Co.
Dunkel, J., & Hnggi, P. (2009). Relativistic Brownian Motion.
<http://arxiv.org/abs/0812.1996>
Ehlers, J. (1997). Concepts of Time in Classical Physics. In A. Atmanspacher
& E. Ruhnau (Eds.), Time, Temporality, Now (pp. 192-200).
Heidenberg: Springer.
Einstein, A. (1907). ber das Relativittsprinzip und die aus demselben
gezogenen Folgerungen. Jahrbuch der Radioaktivitt und Elektronik,
4, 411-462.
Einstein, A. (1949). Reply to Criticisms. In P. A. Schilpp (Ed.), Albert
Einstein: Philosopher-Scientist, vol. II (pp. 665-688). La Salle, IL:
Open Court.
Gdel, K. (1949). A Remark about the Relationship between Relativity Theory
and Idealistic Philosophy. In A. Schilpp (Ed.), Albert Einstein:
Philosopher- Scientist, vol. II (pp. 557-562). La Salle, IL: Open Court.
Giulini, D. (2005). Special Relativity. Oxford: Oxford University Press.
Landsberg, P. T. (1966). Does a Moving Body appear Cool?. Nature,
212(5062), 571-572.
Landsberg, P. T. (1967). Does a Moving Body appear Cool?. Nature
214(5091), 903-904.
Landsberg, P. T. (1978). Thermodynamics and Statistical Mechanics. Oxford:
McCrea, W. H. (1935). Relativity Physics. London: Methuen & Co.
Mller, C. (1972): The Theory of Relativity. Oxford: Clarendon Press.
Nehrlich, G. (1982). Special Relativity is not Based on Causality. British
Journal for the Philosophy of Science 33(4), 361-388.
Pauli, W. (1981). Theory of Relativity. (tr. by G. Field). New York: Dover.
[1921]
Perrot, P. (1998). A to Z of Thermodynamics. Oxford: Oxford University
Press.
Planck, M. (1907). Zur Dynamik bewegter Systeme. Sitzungsberichte der
kniglichen Preuischen Akademie der Wissenschafte, Berlin Erster
Halbband, 29, 542570.
Reichenbach, H. (1969). Axiomatization of the Theory of Relativity. (tr. by M.
Reichenbach). Berkeley, CA: University of California Press. [1924]
Robb, A. A. (1914). A Theory of Time and Space. Cambridge: Cambridge
University Press.
Rovelli, C. (2008). Forget Time. <http://arxiv.org/abs/0903.3832>
Schilpp, P. A. (Ed.) (1949). Albert Einstein: Philosopher-Scientist, 2
Volumes. La Salle, IL: Open Court.
Schlegel, R. (1968). Time and the Physical World. New York: Dover.
Schlegel, R. (1977). A Lorentz-Invariant Clock. Foundations of Physics
7(3,4), 245-253.
Schlick, M. (1917). Raum und Zeit in der gegenwrtigen Physik. Die
Naturwissenschaft, 5(12), 177-186.
Tolman, R. C. (1934). Relativity, Thermodynamics and Cosmology. Oxford:
Clarendon Press.


Book Review
Scientific Representation: Paradoxes of Perspective
Bas van Fraassen
Oxford University Press, Oxford, 2008

Bradley Monton*
bradley.monton@colorado.edu

It probably goes without saying that the advent of a new book by Bas van
Fraassen is a major event in the world of philosophy of science. Indeed,
Scientific Representation does not disappoint. Its full of interesting and
erudite discussions, and presents controversial arguments that philosophers of
science will want to come to grips with.
There are two ways one could review a book like this. One could just go
through the book, step by step, and highlight the key discussions and
arguments. Or, one could look at the book through the lens of van Fraassens
previous work in philosophy of science, and focus on the things van Fraassen
says that directly relate to his previous work. Ill take the latter approach. Van
Fraassen is famous for promulgating a version of scientific anti-realism known
as constructive empiricism: the doctrine that science aims to give us theories
which are empirically adequate, and that acceptance of a theory involves as
belief only that it is empirically adequate (van Fraassen 1980, p. 12). That was
van Fraassens characterization in The Scientific Image of how an aspiring
empiricist like him should understand science. While van Fraassen is still a
constructive empiricist, he has more to say about how an empiricist should
understand science, and he does so in parts of Scientific Representation.
Before I focus my review on issues related to empiricism in science, a brief
overview of the whole book is in order. The book is divided into four parts. In
Part I, Representation, van Fraassen gives a high-level discussion of the nature
of representation. He points out that representation can happen with physical
or mathematical artifacts, and that distortion can sometimes be crucial to
accurate representation. He also argues that there is an essential indexical
element to representation.

* Department of Philosophy University of Colorado at Boulder
194 Humana.Mente Issue13 April 2010
Part II, Windows, Engines, and Measurement, focuses on measuring. Van
Fraassen argues that measuring, just as well as theorizing, is representing.
[] measuring locates the target in a theoretically constructed logical space
(p. 2). Van Fraassen looks at measurement two ways. First, he does so from
within the historical process, when measurement procedures are still being
established. A crucial point is that there is no independent epistemic access to
the parameters to be measured no access independent of measurement, that
is (p. 138). Second, he looks at measurement from a standpoint where
measurement procedures have already been established, and provides a
general theory of measurement, in part motivated by the work van Fraassen has
done on measurement in quantum mechanics.
Ill talk about Parts III and IV in more detail below, but briefly: in Part III,
van Fraassen argues for an empiricist version of structuralism in science. In
Part IV, van Fraassen argues that a theory need not provide a full account of
how measurement outcomes link to reality.
Scientific Representation is a somewhat long book over 400 pages total,
with the main body of the text a little over 300 pages. But if one is just
interested in the parts of the book that are especially relevant to constructive
empiricism and related issues, one can just read about 100 pages. Parts I and II
can be skipped (with the exception of about 15 pages, as Ill discuss below).
Parts III and IV comprise about 120 pages, but the first half of Part III is a
historical discussion of structuralist views in late 19th century/early 20th
century physics. While I found it fascinating, the less patient reader can skip it.
Thus, after looking at the Introduction (pp. 1-3), a reader would not be remiss
in starting at the discussion of Putnams Paradox on p. 229 and reading to the
end (p. 308). The appendix to the last chapter, provocatively titled Retreat(?)
from The Scientific Image, is also worth reading (pp. 317-319).
1

But between the Introduction and Putnams Paradox, there is one
fascinating discussion that directly addresses a key counterintuitive claim of
constructive empiricism, the claim that science does not aim to give us truths
about what is seen through an optical microscope. This discussion takes place
from pages 96 to 110. Here, van Fraassen argues that the optical microscope
should not be thought of as a window on the invisible world, but instead as an
engine creating new optical phenomena (p. 109). Specifically, he thinks that

1
It takes up van Fraassens changing views about probability, and about what it is for a
probabilistic theory to be empirically adequate
Book Review Scientific Representation: Paradoxes of Perspective 195
the images produced by an optical microscope should be thought of as a
public hallucination, akin to a rainbow. Saying that one sees a rainbow, for
van Fraassen, is like saying that (while looking through an optical microscope)
one sees a paramecium. He writes: As long as ordinary discourse is not
filtered through some theory it does not imply that [the rainbow and
paramecium] are objects (p. 110).
Interestingly, even though van Fraassen pushes hard for the view that what
one potentially sees through an optical microscope falls on the unobservable
side of the observable/unobservable distinction, he ends on a concessive note.
He writes:
I draw the [observable/unobservable] line this side of things only appearing in
optical microscope images, but wont really mind very much if you take this
option only, for example, for the electron microscope. After all, optical
microscopes dont reveal all that much of the cosmos, no matter how veridical
or accurate their images are. The empiricist point is not lost if the line is drawn
in a somewhat different way from the way I draw it. The point would be lost only
if no such line drawing was to be considered relevant to our understanding of
science. (p. 110, emphasis in original)
It would be worth thinking about how the arguments for constructive
empiricism from The Scientific Image would change were one to draw the
observable/unobservable line in such a way that things viewed through the
optical microscope count as observable objects. Van Fraassen does not discuss
this issue, beyond the paragraph I just quoted.
Moving on to Part III, we find an argument for an empiricist version of
structuralism in science. Van Fraassen starts by reviewing the history of
structuralism, and arguing that past structuralist views failed because they
didnt adequately take into account the crucial indexical role in scientific
representation. He holds that structuralism finds its proper articulation only
in an empiricist setting (p. 237). Empiricist structuralism is not a view of
what nature is like but of what science is (p. 239). Structuralists often say that
all we know is structure and van Fraassen takes that on board with a caveat:
all we know through science is structure (p. 238). What science does is
represent the empirical phenomenon as embeddable in certain mathematical
models, and these mathematical models are describable only up to structural
isomorphism. The empirical phenomena are represented by a data model, and
a successful theory will have the data model appropriately link to a theoretical
model. But theres more to our understanding of a scientific theory than that. A
crucial indexical aspect is needed, to provide the appropriate link between the
data models and reality (p. 257). Van Fraassen holds that we can know that the
indexical propositions which provide the link are true, and yet the propositions
cannot be part of a scientific theory (p. 261).
Lets move on to the final chapter of the book, which, to my mind, is the
most provocative. Van Fraassen calls it Rejecting the Appearance from Reality
Criterion, but as Ill explain, he could have called it Against Philosophy of
Quantum Mechanics, or Why the Measurement Problem is an Artifact of Bad
Philosophy. But before getting to the Criterion, I have to lay out some
terminology.
Appearances, for van Fraassen, are the contents of measurement
outcomes. He distinguishes them from phenomena, which are observable
entities of any sort (p. 283). In a interesting footnote, he writes: I have only
slowly come to see the importance of marking such a distinction. In The
Scientific Image I did not make this distinction either carefully or clearly (p.
391, n. 24).
Now, the Appearance from Reality Criterion holds that appearances must
clearly derive from what is really going on (p. 282). As van Fraassen explains,
this derivation sometimes happens in science. For example, Copernicus
derived the appearance of retrograde motion from (what his theory said was)
the real motions of planets around the sun. But van Fraassen does not think
that the Appearance from Reality Criterion is a completeness criterion for
science: a theory can be just fine even if it doesnt explain how the appearances
are derived from reality (p. 295).
Van Fraassen recognizes that its quite generally accepted at least amongst
philosophers that the Appearance from Reality Criterion is a completeness
criterion for science (p. 280). His argument against these philosophers
appeals to the norms of science, as exemplified by the actual practices of
physicists. The basic argument is that in the context of quantum mechanics,
physicists are generally perfectly happy with simply using the Born rule to
make predictions for measurement outcomes (where, roughly, the Born rule is
the rule that gives probabilities for the possible outcomes on the basis of the
quantum state of the system). But nothing in the theory of quantum mechanics
as typically used by physicists tells one how the particular appearance we
actually get links to reality. As van Fraassen writes, at the very end of the last
chapter of Scientific Representation:
Advocates of the Appearance from Reality Criterion will not be satisfied with
quantum mechanics. [] Conversely, irenic acceptance of the theory such as
that which we have seen prevalent in the physics community, throughout the
last century would seem to signal an attitude content without any sustained
attempt to satisfy that Criterion. This, it seems to me, should allow us to draw
the right moral about what are and what are not norms that govern scientific
practice. (p. 308)
Now, it would be open for a philosopher to argue in response to van
Fraassen that physicists are simply wrong or confused if they reject the
Appearance from Reality Criterion, or if they dont understand why a solution
to the measurement problem is needed to for quantum mechanics to be a
complete theory. But for a philosopher to give that response doesnt fit with
the deference to science that philosophers nowadays typically show.
2

I wish there were another appendix to the last chapter, called Retreat(?)
from Quantum Mechanics: An Empiricist View. Van Fraassen is putting forth a
position on quantum mechanics that does not come up in his 1991 Quantum
Mechanics book, but its not obvious to me whether the new position is meant
to be complementary to the 1991 position. My sense is that the views are in
tension, because in the 1991 book van Fraassen puts on the table a particular
solution to the measurement problem, known as the Copenhagen Variant of
the Modal Interpretation.
3
The CVMI solves the measurement problem by
postulating that systems sometimes have value states in addition to the
quantum state, and the probability that a system has a particular value state at
the end of a measurement is given by the standard Born Rule (and moreover,
what it is to be a measurement is given a purely physical characterization). This
sort of solution makes sense if one endorses the Appearance from Reality
Criterion: the reality is that the system has a certain value state at the end of
measurement, and that accounts for why the appearance the measurement
result is what it is. If van Fraassen is rejecting the Appearance from Reality
Criterion, and is hence saying that there is no measurement problem, then
theres no need to postulate an interpretation like the CVMI.
But perhaps van Fraassens 1991 view of quantum mechanics really is
compatible with his new position. In Scientific Representation, van Fraassen

2
It also doesnt fit with one of the motivating themes of van Fraassens work, which is to admire
science as a paradigm of rational inquiry.
3
Moreover, van Fraassen played a prominent role in the whole development of modal
interpretations.
does not say that we should never postulate links between appearances and
reality; he just says that we should reject the Appearance from Reality Criterion
as a completeness criterion for science. In other words, its permissible to seek
to link appearances with reality, but we shouldnt fault a scientific theory if it
doesnt provide the link. But what then would be the point of postulating an
interpretation of quantum mechanics that provides the link if the link is
unnecessary? Perhaps the answer is that postulating interpretations gives us a
better understanding of the theory. The idea is that each interpretation
provides an answer to the question how could the world possibly be the way
this theory says it is? (van Fraassen 1991, p. 4) Thus, its worthwhile to seek
many solutions to the measurement problem, because each solution helps us
to better understand the ways in which the theory could be true.
4

But the last paragraph of Scientific Representation provides a potential
challenge to that view. We would think that physicists would want to
understand quantum mechanics, but physicists for the most part irenically
accept the theory they do not concern themselves with trying to solve the
measurement problem. So perhaps scientific understanding goes just fine
without solutions to the measurement problem, and its only deep-seated
sympathies toward the Appearance from Reality Criterion that leads some to
think otherwise.
I explained above that the last chapter poses a provocative challenge to
those philosophers of science who endorse the Appearance from Reality
Criterion. We now see that the last chapter poses a provocative challenge to
those philosophers of quantum mechanics who think that there is a
measurement problem that needs to be solved. If there is no need to explain
how measurement outcomes link to reality, then what is the point of finding
solutions to the measurement problem? Perhaps the answer is just that the
solutions help give us a better understanding of the theory, but its not clear
whether that answer fits with actual scientific practice. Moreover, even if that is
the answer, it becomes a serious open question as to whether all work on the
measurement problem makes sense when seen in that guise. Its one of the
virtues of Scientific Representation that it puts these important questions on
the table van Fraassen is continuing his tradition of giving philosophers
something provocative to talk about.

4
And indeed, personal communication with van Fraassen confirms that he does still think that
these interpretations aid our understanding of the theory.
REFERENCES
Van Fraassen, B. (1980). The Scientific Image. Oxford: Oxford University
Press.
Van Fraassen, B. (1991). Quantum Mechanics: An Empiricist View. Oxford:


Book Review
Persistence and Spacetime
Yuri Balashov

Lorenzo Del Savio*
lorenzodelsavio@gmail.com

Yuri Balashovs Persistence and Spacetime addresses a traditional issue
concerning the persistence of material objects over time. Yet the approach is
far from traditional. Balashovs work is rather scientifically based than just
science aware. The debate about persistence is eventually brought home by
taking on board Einsteins special relativity, here presented as a theory of the
geometry of spacetime. Even though the text is quite technical, two
communities of very different size should be interested in Persistence and
Spacetime: metaphysicians specialized in persistence as well as scholars whose
researches focus on the boundaries between ontology and physics. More
generally, the Balashovs study is a great source of insights into the relation
among science, common sense and philosophy: these topics are explicitly
examined by several methodological observations throughout the book.
Material objects persist: they exist at different times, typically through an
interval t. At least three accounts of persistence have been recently developed
by metaphysicians, in a few words: endurance: an object persists through an
interval t being wholly present at each t belonging to t; perdurance: an
object persists through t having temporal parts at each t of t; exdurance: an
object persists in t having temporal counterparts at every t in t. Established
arguments for a, b, c derive from the problem of temporary intrinsics, the
paradox of material constitution and mereological universalism. These
conceptual predicaments stretch our pre-theoretical beliefs beyond their
standard use. Therefore, traditional discussions have been by and large matter
of a trade-off between pros and cons of each account of persistence from a
conceptual analysis standpoint. According to Balashov, such a debate
continues to be rooted in the manifest image of world and ignore important

* University of Milan

scientific developments, which have rendered many common-sense notions
untenable and obsolete. Despite this basic weakness, he believes that several
results of contemporary metaphysics dealing with persistence should be saved
in a physics-conscious context. He takes indeed as a rule to seek to minimize
the degree of the overall ontological revision while putting physics inspired
arguments ahead of many others. Since physics refers to a wide group of
theories if not to an ongoing enterprise, one might ask which theory if any
is relevant to persistence. Balashovs answer is special relativity (SR). The
theory published by Einstein in 1905 should be regarded as a good
approximation of the spacetime of our world. Consequently, SR is necessary
in order to assess the respective merits of theories intended to be accounts of
the life of bodies within our spacetime. Given these assumptions, the
Balashovs goals are the following. Firstly, he aims to state endurance,
perdurance and exdurance in a relativistic framework. Afterwards, he discusses
whether those statements render fitter one of the view. Being the first goal
achievable without contradictions, Balashov points out that there could not be
any deduction from SR to perdurance. Nevertheless he argues that his
discussion provides at least two good arguments supporting perdurance.
The book is introduced chapter by chapter below. Many topics which
appear in the work must have been omitted. More important, a review cannot
hope to give an idea of the accuracy that distinguishes Balashovs definitions
and arguments. As a result, the general structure of the study and the remarks
on metaontology that Balashov has spread around his book have been the only
focuses of the following introduction.
Some methodological clarifications open the book: the problem of
persistence is sketched, the goals of the research are outlined and many
assumptions defended. One of the latter is worth quoting, being paradigmatic.
Theories of persistence are correlated with other metaphysical theories which
deal with the reality of past, present and future objects. According to
presentism, only present objects are real, whereas past and future objects do
not exist. On the other hand, eternalist observes that the present does not have
any privileged ontological status and thus non-present objects are real as well.
Balashov hastens to get rid of presentism by means of a physics-inspired
argument. There is no concept of objective (frame independent) present
available in the best model of our spacetime, namely Minkowskis spacetime,
hence presentism is meaningless. Other assumptions include a substantivalist
attitude toward spacetime, atomism and an intermediate position on the
Book Review Persistence and Spacetime 203

nihilism vs. universalism debate on mereological composition. At least the last
hypothesis will play a significant role in the following discussion, while
atomism will be absolutely essential for simplicitys sake but harmless.
The next chapter focuses entirely on definitions. Endurance, exdurance
and perdurance are stated within a generic spacetime saving their intuitive
core. The main novelty here is a clear distinction drawn between perdurance
and exdurance. The latter had been generally regarded as a semantic alternative
to the former. Consider a spacetime worm and its temporal parts. It seems a
semantic matter whether you take the worm as a whole labelling it object
(perdurance) or choose one of its parts as the object dubbing the other parts
its temporal counterparts (exdurance). On the contrary, Balashov says that
the endurantists main idea is also preserved by exdurance: an object is wholly
present at each instant of its life in both analyses. Thus perdurance and
exdurance must be told apart. Full presence will be later the feature of both
exdurance and endurance whereby Balashov argues against these theories. The
definitions might be briefly summarized as follows. Being P the spacetime path
of an object O and S a generic achronal slice of P (S is achronal iff no point in S
temporally precedes any other point in S), O endures iff is located in each S of
P; O perdures iff is located in P and has a temporal part in each S of P; O
exdures iff is located in a S of P and has a temporal counterpart located in each
other S of P.
Chapter 3 introduces the special relativistic spacetime in the standard
Minkowskis way. It begins with a Newtonian spacetime, a structure where
events are separated by a definite Euclidean spatial distance and a definite
temporal interval, thus rendering meaningful statements about velocities of
objects. This structure is then replaced by a Galilean spacetime, which makes
meaningless all the absolute ascriptions of velocities by means of declaring that
there is no fact of the matter as to which objects are at rest and which are in
inertial motion (only accelerations have a frame-independent meaning in
classical mechanics). Therefore, being in the same position at different times is
no longer an allowed concept. Finally, as Galilean structure had abolished the
sameness of position, Minkowskis spacetime abolishes the concept of absolute
simultaneity, hence treating space and time in the same way. Balashov states
accurately the main features of such structure: length contraction, time dilation
and invariance of the interval I=c
2
t
2
-x
2
between two events. Some
consequences of the last characteristic are quite essential for the subsequent
discussion and a brief explanation is in order. Each event e splits the

Minkowskis spacetime in three distinct regions according to the value of I
(e,p). These are the absolute elsewhere or topological present (I<0), the
absolute past and the absolute future (both with I0). The last is the set of
points reachable by a signal sent from e. So the boundaries of the future are the
light-like separated points accessible only by a signal which is travelling at the
light speed. The absolute past is the set of points which have e in their absolute
future. In the end, the topological present is causally cut-off from e, being the
set of points too far apart even for a light-speed signal and, a fortiori, for causal
influence. An event belonging to the absolute elsewhere of e is said to be
spacelike separated from e.
The definitions gathered in chapter 2 are being now translated from the
generic spacetime firstly to the Galilean and then to the Minkowskis structure.
No wonder if the first translation strictly resembles the current state of the art
of the debate on persistence, thus proving that metaphysicians have been
moving in the classical framework. Sparse exceptions do exist: Quine and
Smart are responsible of an early attempt to draw conclusions about
persistence from SR. However, nobody had previously tempted such a broad
account of the topic. Balashov reconstructs the whole discussion well beyond
the sketched suggestions of his precursors. Chapter 4 and 5 should be
considered the very core of the project. Here Balashov succeeds in bringing
the theories of persistence into the scientific treatment of spacetime. As a side
work, the argument from vagueness to perdurance is assessed at length.
Unrestricted mereological diachronic composition inherits his likelihood from
unrestricted synchronic composition, thus implying the existence of temporal
parts. Balashov rejects the deduction by bringing a diachronic counterexample
where we do have a strong reason to restrict composition, that is violation of
the law of conservation of matter and energy. Again, a physical observation
puts a severe constraint on the analysis, therefore narrowing the range of
possibilities. Balashov adds interestingly that these restrictions simply follow
the joints of nature.
Another metaontological dilemma rises later, when the translation of the
ordinary concepts in the relativistic framework get harder. The dilemma might
be suitably dubbed adjustment or replacement, here is an instance. The
familiar notion of moment of time is replaced by the relativistic-inspired
achronal hypersurface indexed to an instant in an inertial frame reference.
According to SR, inertial frames do not have any superior metaphysical status.
Nevertheless, descriptions of phenomena within such kind of frames are

convenient because of geometrical reasons: laws keep their form if translated
from an inertial frame to another. Therefore, the concept of an achronal
hypersurface indexed to an instant in an inertial frame reference does not have
any metaphysical priority. Why should we adjust the concept of a moment of
time rather than replacing it with a neutral scientific description? According to
our best theory, we had better get rid of moments of time. Why must a
metaphysician do the effort of translating the familiar notion in a scientifically
respectable concept? One might indeed wonder whether the whole agenda of
physics-oriented metaphysics is worthwhile. Balashov has a general answer
here, the already mentioned principle which invites to avoid large ontological
revisions. If conceptual revolutions lead too far from our common sense, they
might be ineffective. Science-oriented metaphysicians bring our everyday
concepts into a new environment, thereby providing landmarks in yet
unexplored lands.
The end of the fifth chapter contains several refutations of many objections
on behalf of endurance. Balashov especially insists that no deduction is
possible from SR to perdurance. The temptation to reify spacetime paths must
be resisted in order to avoid the main gate to perdurance. Multi-location offers
a ground for resistance: since objects might be multi-located in several slices of
an objects path, as it is the case according to endurantists and exdurantist
alike, paths are just path. They are not worm-like temporal extended objects.
Before putting to work his definitions, Balashov needs one more piece of
analysis. Any account of persistence should not turn out odd if conjuncted with
a theory of coexistence. What exactly is coexistence? Given eternalism,
everything trivially coexists with everything. Yet there is a sense according to
which dinosaurs have never coexisted with human beings. Chapter 6 deals with
the second remarkable meaning of coexistence and an associated restricted
meaning of existence. We would like to say that something no longer exists,
has already existed or has not existed yet even if everything trivially exists
without adverbial qualifications. Balashov explains the point with an analogy.
Lewis argued that we need a restricted existential quantifier there is at least
one x in w in order to express the non-trivial fact that there is not any possible-
non-philosopher-Lewis (in our world) even though there trivially is a possible-
non-philosopher-Lewis (in some world). In other words, analysis of existence
is univocal but sometimes exists-at-t would be as useful as exists-at-w in
order to describe non trivial facts. Since time has became tricky, an analysis of
exists-at-t and coexists-at-t must be provided. The one Balashov prefers is

Coexistence as Sharing a Hyperplane of simultaneity (CASH) because it ties
coexistence to an invariant structure of Minkowski spacetime without going
too far apart from our ordinary beliefs. For our purposes, we might
approximate CASH as follows. Two objects located respectively in p
1
and p
2

coexist iff p
1
and p
2
are spacelike separated. CASH could be shown satisfactory
because it agrees with everyday intuitions within classic limits. Furthermore,
Balashov rejects what he calls the Alexandrov-Stein coexistence, an analysis
based on several beliefs about coexistence that were originally exploited in
order to analyse tensed utterances within an eternalist framework. The
speakers present may be defined as the set of the locations of the objects with
which she can interact during a short interval. The length of the interval is
determined by the minimum time needed for an act of thinking. A coexistence
relation among objects could then be defined as substantial overlapping of
their presents. Consequently, if two objects AS-coexist, they could be causally
related. As Balashov notes, no theory might be go farther from CASH. CASH-
coexistent objects are spacelike separated and hence causally cut off while AS-
coexistents are would-be interactor. Hence Balashovs basic objection against
AS: AS misses our beliefs about causation. Temporal precedence of causes is
quoted ever since Hume as an essential feature of causation. Even in a classical
framework, interacting takes time. Therefore, coexistent objects cannot
interact. Furthermore, AS-coexistence relies on the concepts of a speakers
present, not exactly an objective feature. What about AS-coexistent stars and
planets?
Chapter 7 pursues the second goal of the Balashovs research, assessing
whether SR supports one of the accounts at stake. Balashov claims it does. The
basic idea is the following. Imagine that Balashov is now wholly present 250
light-years apart. Then he still CASH-coexists with Napoleon and already
CASH-coexists with Putin. In a sense, Putin and Napoleon are temporally
together in the Balashovs present although their life spans never overlap.
Balashov declares the verdict unacceptable. More precisely, he claims that the
situation is trivially perspectival and hence unproblematic for the perdurantist
while interestingly locative and thus dooming for exdurance and endurance
alike. Reasons should be sought in the common feature of the latter theories,
that is full presence. Two statements must be told apart and independently
defended: (i) exdurance and endurance are, but perdurance is not, committed
to the locative reading of the above situation (asymmetry thesis), (ii) the
locative reading is threatening (absurdity thesis). The Balashovs defence of (i)

is simple: temporally-laden qualifications such as already or still are
weakly attributed to coexistence facts among temporal parts while strongly
attached to existence proposition about wholly present objects. A modal
analogy may help again. Obama might have been the 45
th
President of the USA.
In a Kripkean-standard account, we are directly speaking of Obama. According
to Lewis, we are certainly talking of him but less straightforwardly. Obama
might have been the 45
th
instead of the 44
th
because there is an Obamas
counterpart who is the 45
th
. Lewis worlds cannot overlap, while Kripkes
might and in fact do. Overlapping is the modal equivalent of full presence.
Temporally laden qualifications on existence are thus ascribed weakly from the
perdurantists point of view very much like modal properties are indirectly
ascribed in a Lewisian account. The absurdity thesis (ii) is less beyond dispute
because of the subtle argumentative structure of the chapter. One might
indeed argue as follows. If a revisionist analysis of coexistence is given, our
pre-theoretical opinions about coexistence facts are no longer a safe guide. It is
really strange that Putin and Napoleon are somehow temporally together.
However, we are moving in an unexplored land. A radical endurantist may
declare that something has gone wrong with CASH instead of shed endurance
because it makes stranger a strangeness. Otherwise, she might say that oddities
are expected and should be accepted whenever a revisionist account is given.
In any case, scientific facts alone could not settle the matter. Only a careful
balance of the consequences could help.
The last chapter presents further evidence supporting perdurance: the
argument from perspectival phenomena in spacetime. According to SR, bodies
cannot keep invariant their 3D shape because of the length contraction.
Balashov argues that there is an invariant 4D shape that stands behind the
different 3D perspective. The 4D shape of a body is indeed frame independent
and thus objective. The endurantist cannot explain why 3D perspectives of the
same object fit very well together, forming a neat 4D shape. Quite the
opposite, perdurance is itself a good rationalization of the phenomena. Thus,
by inference to the best explanation, objects are 4D extended and hence they
perdure. In the same way, how could it be that many 2D shapes represent the
same object? If we come up with a 3D object which fits with all of them, we are
allowed to infer plainly that we are dealing with a 3D entity. Among the refuted
objections, the one put forward by Sider is remarkable and so it is the
Balashovs reply. Sider suggests endurantists to provide an account of the
phenomena based on micro-facts about the locations of the mereological atoms

of a 3D object. The description of the paths of the atoms together with the
rules of 3D composition suffices as an explanation. Balashovs answer is that
the latter account is inferior because too detailed. Perdurance discovers a
general 4D pattern behind the raw phenomena whereas a micro-explanation is
yet another description of them.
Several Balashovs remarks on metaontology have been listed previously.
Perhaps the most important opens his book. He states that the discussion is
obviously limited to the realm of the physically possible and I can implement
this program without considering physically impossible scenarios, therefore
the modal force of my conclusions is limited. SR does put several constraints
on theoretical choice. The simplest example involves presentism: there is not
any privileged framework which sets simultaneity and hence no absolute
present. Consequently, while being logical possible, presentism is not
physically possible. Presentism is an eligible theory in some Newtonian
universes and might be the correct one in some of them. Nonetheless, this is
not the case here. Why does this eligibility matter? More generally, what
should we ask to the modal force of our analysis? For some purposes, taking
into account broadest range of possibility may be appropriate. Yet this is hardly
the case for the majority of philosophical enterprises. Our concepts are
designed to cope with our world, that is why scientific constraints should be
put ahead. Nevertheless, science is not a constraint on analyses at all. The
discovery that an abstract space might be non-Euclidean (a mathematical
achievement) and that one of this odd objects actually represents our space
better than the Euclidean geometry (a physical theory) has been a great source
of insights into new possibilities. Metaphysician have gone farther beyond their
a priori constraints, thanks to science. No doubts that adding scientific facts
restricts the modal force of our conclusions. De facto, it helps us to imagine
new possibilities.
For similar reasons, Balashovs Persistence and Spacetime casts new light
on an old subject. Overlooked prejudices are brought to the surface. Common-
sense constraints are replaced by up-to-date scientific theories. Eventually,
perdurance wins a fresh ally: special relativity. Although several conclusions
might and will be disputed, Balashov has set the agenda of the persistence
debate. The significance of spacetime theories for metaphysicians could no
longer be neglected.

Book Review
Identity in Physics:
A Historical, Philosophical and Formal Analysis
Steven French and Dcio Krause
Clarendon Press, Oxford, 2006

Giovanni Casini*
giovanni.casini@humana-mente.it

The development of quantum physics has undermined some of the main
assumptions on which both classical physics and the contemporary
commonsense view of the world rest.
In this book, the effort of the authors is primarily devoted to the analysis of
some metaphysical/ontological issues raised by contemporary physics, and in
particular they investigate the relation between three main notions:
individuality, distinguishability, and trans-temporal identity. Such an
investigation, that is primarily of a philosophical kind, is carried forward by
coordinating both philosophical and historical issues, and it ends up with the
proposal of a logical system that results appropriate for the formalization of
both quantum and non-quantum objects.
The central problem concerns the impact of quantum theory on the
classical metaphysical theory of identity, based on Leibnizs Principle of the
Identity of Indiscernibles (PII) (if two objects share all their properties, then
they are identical, they are the same object). In fact, quantum particles are
characterized by a kind of behavior that that is not compatible with a notion of
identity based on the properties instantiated by items, in particular on their
spatiotemporal properties.
The book is organized in 9 chapters. Chapters 1-4 deepen the metaphysical
issues about identity end individuality with respect to the history of
contemporary physics, underlying how the status of particles has changed
moving from classical to quantum physics. In chapters 5-9 a set theory and a
logical system are defined in order to provide an appropriate formal framework

* University of Pisa

for non-individuals, that is, objects which individuation is not compatible with
PII.
In the introduction (chapter 1) the authors delineate the main metaphysical
options about identity. On one hand we have the traditional notion of
individuality, based on PII, that states that individuality is conferred by the set
of the qualitative properties that an entity instantiates: two objects that are
indistinguishable (that is, that share the same properties) cannot be two
different objects; following this approach there is a perfect correspondence
between the notions of individuality and distinguishability. On the other hand
there is a transcendental notion of individuality, where every individual is
defined by something else, transcending ordinary properties (this approach
has taken many form in the history of metaphysics, and the authors do not
commit to any particular one); in this case the fact that two entities are two
distinct individuals does not imply that they are distinguishable with respect to
the properties they instantiate.
While classical physics is compatible with an approach based on PII
(chapter 2), by referring ultimately to the spatiotemporal properties of the
particles, quantum physics cannot share such an approach (chapter 3), for
more than one reason. First, quantum entities cannot be distinguished on the
basis of their properties, even the spatiotemporal ones, since the failure of the
assumption of impenetrability in quantum physics prevents any reference to
the spatiotemporal properties for determining the individuality of particles.
Moreover, when quantum particles become entangled with one another, their
collective state is not reducible to their separate, individual states, and so the
relations between particles acquire a kind of ontological value, respectful of the
holistic aspect of quantum reality. Finally, there is a difference in the statistical
approach: on one hand classical statistical mechanics are based on a model that
treats every particle as a particular distinct individual, and so all the aggregative
configurations which are structurally equivalent but differ for a permutation of
the particles are considered distinct states of a system; on the other hand in
quantum statistics some types of particles cannot be labeled, that is, they
cannot be treated as distinct individuals, and so the different states are
distinguished only with respect to their configurations, regardless of the
possibility of permuting the particles.
Hence, the quantum particles violate PII, and this takes to the main
question about their individuality (chapter 4): can they be seen as characterized
by a kind of transcendental individuality, or are to be considered as non-
Book Review Identity in Physics 211

individuals? The authors conclude that the results of quantum physics are not
determinant for a choice between these two options, and they can be both
embraced, even if the non-individual approach is preferable since it meshes
better with the framework of quantum field theory. The main point is that the
metaphysical issues are not univocally determined by the physics, leaving the
space for different options. Notwithstanding, it is possible to develop a logical
formalization that results appropriate as a model of quantum objects,
independently of the particular metaphysical option chosen.
The rest of the book is devoted to the formulation of such a formal theory
for non-individuals.
Formally, to be an individual means first of all to satisfy the law of identity
(for every x, xx, that is, every object is identical to itself). Hence, the main
idea is to define a semantics and a logic such that the law of identity is not
satisfied, that is, the formula For every x (xx) is not valid anymore. Since
such a law is ineradicable from classical set theory, French and Krause point to
the definition of an alternative theory that avoids the validity of the law of
identity. They take as a basis Dalla Chiara and Toraldo di Francias quaset
theory (chapter 5), and Da Costas System S (chapter 8).
Semantically, they present the Quasi-set theory (chapter 7). Quasi-set
theory is a two-sorted set theory, with two types of elements: the M-atoms are
classical set-theoretic elements, while the m-atoms are quasi-objects, non-
individuals. If we restrict our attention to the former type of atoms, we remain
in the field of the classical set-theory, while if we consider also the m-atoms, we
obtain a system that distinguishes between a relation of indistinguishability
and a relation of identity ; in the theory we have that, while a law of
indistinguishability is always satisfied (for every element x, xx), the law of
identity is not valid anymore (it is not true anymore that for every x, xx).
Then, starting from Da Costas System S, French and Krause define their
Schrdinger Logics (chapter 8), a class of non-reflexive (that is, not satisfying
the law of identity) logical systems that can be characterized by a semantics
based on quasi-set theory.
The last chapter is dedicated to recap and take stock of the situation.
The work presented by French and Krause is impressive. The book is not an
easy reading: there is a continuous reference to the main results in quantum
physics, and a good background in the modern history of physics is also
required; moreover a good familiarity with formal logic is necessary to cope
with the last chapters of the book. Indeed, people with an appropriate

background for appreciating this book from all the angles are presumably quite
a few, and I am surely not one of those. However, also a partial estimation of the
value of the various investigated issues can be extremely fruitful, and the
authors are always mindful in presenting the general problems and the main
results of their endeavour in an accessible way. In the end, for the multiplicity
and the richness of the topics involved, this book is not only an important
contribution to a particular issue, that is, the notion of identity and its relation
with contemporary physics, but it reveals itself as a significant text for the
philosophy of physics in general.

Book Review
The Metaphysics Within Physics
Tim Maudlin

Emanuele Coppola*
ec.4975@gmail.com

Written with a lively and unaffected style, this book by Tim Maudlin weaves a
robust web of analyses, which should be of pressing interest both for
philosophers of science and metaphysicians, as well as for philosophers tout
court, since, on the one hand, it upholds a general methodological claim, to the
effect that, in doing philosophy, we ought to seriously evaluate the significant
implications carried by scientific practice as to the conceptual health of certain
notions or principles (such as the notion of metaphysical possibility or the
principle of Ockhams razor), while, on the other, it engages in subtle and
provocative attacks against metaphysical doctrines, most notably Humean
supervenience and traditional accounts of universals, and against philosophical
reconstructions of scientific laws, such as Armstrongs (1983), Vallentynes
(1988) and van Fraassens (1989).
A central place in this web is occupied by what in Chapter 6 is labeled as
Maudlins Non-Humean Package, i.e. the idea that both the laws of physics
and the direction of time are ontologically primitive (p. 182). Such a twofold
ontological primitivism is recognized to be an unpopular view, at odd with
influential conceptions of metaphysics and science (especially Lewiss
Humeanism, van Fraassens constructive empiricism and the main
interpretations of general relativity), but it develops through sophisticated
arguments, with abundant examples and case studies mainly drown from
contemporary physics (quantum mechanics in a strategic way), which give the
overall impression that the plan has some big cards to play.
The book divides into six chapters, conceived as independent essays
between an introduction and an epilogue. The first chapter, packed with
particular discussions that would be hard to even hint at, outlines a modest (in

*

University of Florence

Swifts ironical sense) proposal about laws: Maudlin begins with a conceptual
analysis of the systematic connections between beliefs about laws of nature and
beliefs about possibilities, counterfactuals, and explanations, putting off the
problem of ontological commitment, but a decisive metaphysical move has
already been made: he radically shifts laws from analysandum to analysans (p.
45); once we take laws as ontological primitives, we have no reason to accept
Armstrongs universals approach (1978 and 1983), at the same time being
able (unlike Lewis 1973 and 1986) to distinguish between true and false
counterfactuals. For this purpose a three-step recipe is offered (pp. 22-23) as a
guide for the evaluation of counterfactuals about the future evolution of
physical states. The recipe, based on fundamental laws of temporal evolution
(FLOTEs) and their adjunct principles, together with Cauchys surfaces and
boundary conditions, is supplemented with a sketched description of how we
should conceive of the evolution of stochastic events (i.e. taking into account
both infected and uninfected physical magnitudes: pp. 30-31). The
counterfactual is taken to be a function whose first argument is not a
proposition, but a command (p. 23) providing instructions to generate an
altered description of a physical state (in step 2 of the recipe). This threefold
procedure is thought to be closer to our psychological processes through
which we construct counterfactual situations by means of laws, than some other
abstract semantic technique (like Lewiss appeal to judgments of overall
similarity between possible worlds). Only FLOTEs (Schrdingers equation,
for instance) are really primitive and basic: other kinds of laws the simply and
the special laws of temporal evolution, respectively LOTEs and SLOTEs are
parasitic on them. Among the results of ontological primitivism about laws
there is also the fact that notions like physical necessity and objective
propensity derive from FLOTEs (pp. 19-20).
The second chapter is an exposition of the reasons why one should reject
Humean supervenience. Lewiss ideas are dwelled upon in some detail;
according to Maudlin, Lewiss Humeanism comprises two logically distinct
theses: 1) Separability (we can crumble the whole space-time in pointlike
objects, each of which having its own intrinsic physical state, while all the rest
supervenes on these material bits laid out in space and time); 2) Physical
Statism (the total physical state of a world does entirely determine modal and
nomological facts about that world). Concerning Separability, Maudlin
maintains that the radical metaphysical innovations stemming from quantum
mechanics pose a weighty challenge to the sort of metaphysical
Book Review The Metaphysics Within Physics 215

compositionality of properties postulated in a separable universe:
notwithstanding Einsteins worries and Lewiss caution, the superposition of
the entangled states forces us to give up sticking with the belief in a mosaic of
independent objects and events. Concerning Physical Statism, it is a distinctive
claim of this book that laws precisely the FLOTEs cannot supervene on the
total physical state of the world: nothing in the current scientific practice
which Maudlin assumes as a starting point and as a touchstone for a sound
philosophy suggests the possibility of such a reduction. More generally,
contemporary physics seems to provide us with new insights in our universe,
which run definitely afoul of the desertified picture of reality favoured by many
analytic metaphysicians. Not always less is more and Ockhams razor does hide
an ideological and counterintuitive use, which we ought to deprive it of: If the
ontology that arises most naturally from reection on physics is too rich for
Ockham or Hume or Lewis, then so much the worse for them. Let others
subsist on the thin gruel of minimalist metaphysics: Ill take my ontology mit
Schlag. (p. 4).
In chapter 3 we come across a couple of specific tools of contemporary
mathematical physics (besides Cauchys surfaces in chapter 1): gauge theories
and fiber bundles. Given the technicality of the issue, suffice it to say here that
the main result of this essay is the alleged collapse of the age-old doctrine of
substance and attribute together with the traditional theory of universals.
Relying on the unnoticed ontological import of fiber bundles, Maudlin holds
three strong metaphysical theses (p. 86):
(i) there are no intrinsic (or metaphysically pure) external relations;
(ii) there are no intrinsic internal relations;
(iii) there are no intrinsic properties.
According to contemporary physics there can be no fact of the matter
concerning whether any two quarks are the same color or different
(chromodynamics is introduced for the sake of exemplification: pp. 94, 96).
Here one finds the more controversial side of Maudlins project: a physicalistic
elimination of ordinary universals via the application of a mathematical
structure. But, to use with a slight modification Maudlins own words about
laws (p. 12), parochial universals (red, blue etc. as metaphysically pure
properties occurring in human discourse) are still universals, whose abiding
resistance in the face of their absence in mathematical manifolds or
electromagnetic fields cannot be dismissed as a short-sighted trust in the folk

ontology delivered by our middle-sized-thing-language let alone the fact
that the Author himself (pp. 98, 102) acknowledges the existence of at least
one universal: the geometrical structure associated to each fiber.
In chapter 4 both the intuitive idea of the passage of time and the
theoretical concept of an intrinsic asymmetry in the temporal structure of the
world are defended against the commonly held view, according to which
relativity theory states the impossibility of an objective flow of time (Gdels
view is mentioned by way of example: pp. 115-16). Believing in the reality of
both past and future events (thus accepting the metaphor of the block universe)
is not at all incompatible with the thesis that time passes. Such a thesis has to
cope with three types of objections (logical, scientific, and epistemological), to
each of which a separate and careful discussion is devoted. Even granted that
the physical laws are Time Reversal Invariant (p. 117: but this is not the case,
in the Authors opinion: see the phenomenon of the decay of the neutral K
meson), such an invariance just presupposes a temporal direction;
furthermore, it is difficult to see how a hypothetical time-reversed human
Doppelgnger could be conscious and even intelligible in its anti-
thermodinamic behaviour (p. 125). Among the strategies carried on by the
deniers of the passage of time there is the ontological reduction of the
direction of time to the entropy gradient; Maudlin contends that the entropy
gradient is, on the contrary, explained through the asymmetrical treatment of
the boundary conditions of the universe (recall the recipe in Chapter 1), an
asymmetry which in its turn is a reflection of the fact that time passes (p.
131): there is a preferred direction (an arrow) of time, and there is an entropy
increase, only because time intrinsically passes such a primitive truth
pertains both to our scientific image and, less problematically, to the manifest
one.
Chapter 5 is a thorough critique of the counterfactual analysis of causation.
In all the attempts to secure the causal arrow through counterfactual
dependencies there is a third fundamental factor, linking causes and
conditionals, to be traced: the law of nature. The need for a third factor arises
when we consider, on the one hand, the possibility to have knowledge of causal
connections without knowing any Humean counterfactual (If C had not
occurred, E would not have occurred, when both C and E did actually
occurred: p. 143), as the example of a Newtonian particle collision (with
monitoring devices and potential back-up particles) shows. On the other, we
consider also the inverse possibility: having knowledge of counterfactuals
Book Review The Metaphysics Within Physics 217

without knowing causal connections. In the first possibility the key point is that
laws in our world suffice to determine causes without the aid of Humean
counterfactuals, and it is because he has implicitly picked up the real cause of
an event that the counterfactual analyst can keep improving his theory. As to
the second possibility, Maudlin sketches a modified version of Conways game
of Life (p. 149) and argues convincingly that in this hypothetical world even a
detailed knowledge of all counterfactuals cannot settle the question of its
causal structure, as well as the question of what are the laws operating there. A
complete about-turn is needed: an alternative account in which laws will play a
key role in the identification of causes, thereby figuring in the very truth
conditions for causal claims. It is in this crucial passage that we find the notion
a particular set of laws, called quasi-Newtonian (p. 156), which represent the
way we ordinarily think of the world and the objects therein. These laws consist
in inertial laws (describing the initial, unaltered behaviour of some entities),
extended with laws of deviation (describing how the inertial behaviour
undergoes changes); in contexts of such a sort it is quite natural to identify
what, as a perturbing factor, causes what. Our cognitive performances,
however, are not entirely based on the assumption of quasi-Newtonian laws
and, a fortiori, on FLOTEs: in our special sciences we usually employ lawlike
generalizations (although we expect these too to be quasi-Newtonian),
contriving taxonomies that could hardly be explained using the vocabulary of
physics, in their carving the world at its joints (here comes an instructive
treatment of an example from McDermott 1995: pp. 161ff.).
Chapter 6 lingers over other critical aspects of Humean supervenience and
focuses by contrast on further merits of the non-Humean package: in particular
on the notion of a productive explanation (p. 175), which does justice to the
asymmetrical order of physical events (the meaning of Brombergers flagpole is
remembered for the adequacy of the syntactic structure of D-N explanations).
But, were not for the primitive asymmetry engendered by the passage of time,
the very idea of a production would be completely out of range: it is not by
chance that only fundamental (non pragmatic) productive explanations involve
a FLOTE (p. 178).
Readers looking for the metaphysical import of physical theories like
general relativity and quantum mechanics can gain precious insights from this
book: its suggestions are innovative, though non always immediately clear in all
of their implications. And readers persuaded that there is, beside physics,

much more than mere stamp-collecting, will find in it stimulating challenges
to face.

REFERENCES
Armstrong, D. M. (1978). Universals and Scientific Realism. Cambridge:
Armstrong, D. M. (1983). What is a Law of Nature?. Cambridge: Cambridge
University Press.
Lewis, D. (1973). Counterfactuals. Cambridge, MA: Harvard University
Press.
Lewis, D. (1986). Philosophical Papers, vol. II. Oxford: Oxford University
Press.
McDermott, M. (1995). Redundant Causation. British Journal for the
Vallentyne, P. (1988). Explicating Lawhood. Philosophy of Science, 55(4),
598-613.
Van Fraassen, B. (1989). Laws and Symmetry. Oxford: Clarendon Press.

Book Review
Anaximander
Carlo Rovelli
Forthcoming, Dunod

Umberto Maionchi
umberto.maionchi@humana-mente.it

The interest of Carlo Rovelli, a brilliant contemporary physicist known for his
fundamental contributions to the so called loop quantum gravity, for
Anaximander, is motivated purely by scientific considerations. His only true
aim is to evaluate, from the preferred perspective of contemporary physics, the
first great conceptual revolution that stands probably as the birthmark of the
scientific thought. In other words, his interest is the birth of that research spirit
that nowadays we call the scientific investigation of nature and of natural
phenomena.
Anaximenes, Thales and Anaximander constitute a first group of thinkers
that Aristotle himself called naturalistic philosophers, because of their attempt
to provide an explanation of the natural world solely in terms of natural causes.
Thales was concerned about a very general theme: what is the fundamental
origin of nature and all of its parts, what is that fundamental substance that all
of the things are composed of? Thaless own idea comes from the empirical
observation that wherever there is water there you find life. Thus he advanced
the hypothesis that water is the fundamental element that constitutes somehow
the origin of matter in all of its forms. This was just an hypothesis, but, as
Rovelli points out, it was the first one about the constitution of reality that
neither appealed to mythical explanations nor required divine interventions.
Anaximander, Thaless pupil, learned the very spirit of this first lesson. He
founded however his teachers answer unsatisfactory and thus he set on to
provide a different and better one. Probably his argument was something like
the following. If there is a fundamental substance out of which all other
substances somehow derive their being, it cannot be any of those visible
substances that are immediately available to sensible experience. That
fundamental substance cannot have constant properties but it must be
something indefinite in its own nature, always capable of assuming different

forms without thereby changing into a particular of those forms.
Moreover he takes a further step. The natural world, natural events and
their intrinsic dynamics have to be regulated by Necessity that manifests itself
in laws that govern the passage from one state of affair to another. These laws
are exactly natural laws, in the sense that they regulate natural events according
to a necessary temporal order. Every natural phenomenon is at the same time
both cause and effect of other different natural phenomena and thus the natural
world is causally closed.
We know that Anaximander wrote a treatise, On Nature, but all that is left
from this work is an obscure fragment. What we know about him and his work
comes from the testimony of later authors such as Aristotle and Theophrastus.
Therefore a faithful reconstruction of Anaximanders work is not an easy task.
However in recent times modern scholars have been reading again all of the
existing materials and have been able to find new material as well. On the basis
of these new developments Rovelli is able to sum up Anaximanders thought
and highlight the revolutionary character of many of his thesis. The result is
simply amazing for audacity and depth. Anaximander was the first one to hold
that:
1) Meteorological phenomena have natural causes.
2) Earth is a body of finite dimension that is suspended in empty space
and does not fall since there is no privileged direction of motion.
3) Sun, Moon and Stars revolve around the Earth dragged by invisible
wheels.
4) Every animal, and men among them, comes from the sea and they have
evolved from different forms of life.
5) All of the things in their multiplicity comes from a unique origin or
principle, called apeiron, that can be translated as indefinite and
limitless and
6) things change according to necessity, i.e., in accordance with universal
laws.
Philosophical historiography has focused solely on the philological or
metaphysical aspects of the word Apeiron and on its origin. There has been a
long discussion on how to interpret this terms that can have so many different
meanings. Rovelli, on this underestimation of Anaximanders contribution to
the birth of the scientific thought, has a precise idea that boils down to the
difficulty, for many intellectuals that have a historical or philosophical
Book Review Anaximander 221

formation, in evaluating the measure of those contributions that have an
intimate scientific nature.
The most flagrant example of this underestimation, and the central core of
Rovellis marvelous book, is the genial intuition of the fact that the Earth is
somehow suspended in the empty space. Rovelli points out an interesting
analogy. The fact that the earth was round and not flat is already found in one of
the most widely discussed of Platos dialogue, namely the Phaedo. This
suggestion is another staggering contribution to the scientific thought that is
always forgotten. For those who read and study Phaedo do focus only on the
question of the immortality of the soul. This is a clear example, according to
Rovelli, of the distance that separates two kinds of cultures, scientific culture
on one hand and humanistic culture on the other.
Scientific thought is certainly a historical product of human civilization.
And because it is a historical product is destined to undergo the influence of
historical events and to share their destiny. The development of science is not a
linear, cumulative process that does not know any stops or involutions. Rovelli,
on this very point, refers to the wonderful essay, La Rivoluzione Dimenticata
(The Forgotten Revolution) by Lucio Russo (1996). This is to remind us that
science and scientific spirit even more, are fragile conquests that are always in
danger, always exposed to the attacks of numerous and unsuspected enemies.
More often than not underestimation is a weapon in the hands of those
enemies.
Then today a new reflection on the nature of scientific thought is necessary.
We should examine again its definition, its fundamental characters, its peculiar
methods, its scope and its aim. And, as it is usually the case, one of the
privileged way to understand the essence of a cultural phenomenon, is to go
back to its roots and origins and focus more carefully our attentions to those
first stages of its development. This is exactly what Rovelli has done in this
wonderful book.
The figure of Anaximander seems to sum up all of the most salient aspects
of scientific enquiry, its guiding principles and methods. According to Rovelli,
Anaximander was the first one to understand e put in practice what can be
thought as the fundamental credo of every modern scientist, that is that we
should study great Teachers, comprehend their lesson, and on the very basis of
this lesson, reveal their mistakes, correct them and promote an always better
and always perfectible understanding of the world.
The cultural basis of the birth of the science is the same one on which

democracy is built upon, that is the discovery of the effectiveness of criticism
and of dialogue among peers. Anaximander that proposes an insightful
criticism of his own teacher Thales puts forward again, on an intellectual level,
so to say, what was standard practice at the social and political level. And that
was the fact that the authority of any political power should not be accepted
unconditionally, for authoritys sake, but the proposal of the city magistrate,
should stand critical scrutiny, in a shred awareness that there is always a better
proposal.
Rovelli draws attention to this analogy explicitly. It is in a certain sense the
discovery of the scientific method. Someone proposes an idea, a thesis. It is
considered carefully, it is criticized, it is improved upon. Then other theses are
advanced. They are compared. The extraordinary discovery is that this whole
process sometimes converges.
Science and scientific research are public, in the widest sense possible.
Everyone can participate and everyone can criticize and even refute everyones
thesis. There are no absolute truths, nor untouchable authorities. The only
source of authority is not a given name but the strength of the argument that is
being put forward. And a proposed thesis is more convincing if it stands more
attempts to be refuted. But it always remains provisional, refutable. It is not
difficult to see, even in this very rough sketch of the nature of scientific
enterprise, the essential traits of democracy as a form of government.
Uncertainty and doubt are the strength of any truly scientific enquiry. This may
sound paradoxical, but only to those that do not know the history of scientific
development.
This topic inevitably leads to the question that Rovelli deals with in the last
part of his work: the relationship between science and religion, both from a
theoretical and an ethical standpoint. Rovelli, with subtle sensibility, asks
himself two difficult questions. The first one is why science seems to have lost
much of its fascination and seems more distant from the concerns and
problems of most of the people. The second one is what is the real nature of
religious thought and what are the reasons of why such a thought is so deep-
seated in human nature. Almost all of the second part is devoted to possible
answers to these two questions.
Rovelli observes that in the last decades the activity of practicing scientists
has become more difficult and almost esoteric for those who are not in the
field. And furthermore it seems to have lost its essential and fundamental
capacity of being a visionary discipline, an immense producer of images of the
Book Review Anaximander 223

world. And along with that it has lost its appeal as a creative activity and its
fascination of being capable of true human emancipation.
And these observations are, unluckily, true. Rovellis book then becomes
even more important for it gives us back an image of science that is but the
great image that another major physicist of our time, John Bell has left us. The
Enterprise is to understand the world and we should never betray the
Enterprise.


Commentary
A Structural Interpretation of Pure Wave Mechanics

Jeffrey A. Barrett*
jabarret@uci.edu

In the long version of his Ph.D. thesis, Hugh Everett III developed pure wave
mechanics as a way of solving the quantum measurement problem faced by the
standard von Neumann-Dirac collapse formulation of quantum mechanics.
1

Pure wave mechanics, however, encounters problems of its own. I will brief
review Everetts description in his thesis of the standard measurement
problem, how pure wave mechanics solves it, and the problems pure wave
mechanics itself faces; then I will explain how one might nevertheless
understand pure wave mechanics as a successful physical theory given the
notion of faithfulness that Everett presents at the end of his long thesis. The
result is a structural interpretation of pure wave mechanics as an empirically
faithful theory.
The standard collapse formulation of quantum mechanics has two
dynamical laws:
Process 1: The discontinuous change brought about by the observation of a
quantity with eigenstates
1
,
2
, , in which the state will be changed to
the state
j
with probability (,
j
).
2

Process 2: The continuous, deterministic change of state of the (isolated)
system with time according to a wave equation
= , where is a linear
operator.
The rule for when each applies is, on first pass simple: a physical system
always evolves according to the deterministic Process 2 unless a measurement

* University of California at Irvine
1
Since Everetts advisor, John Wheeler, was uncomfortable with Everett formulating his thesis as
a direct attack on Bohr and the standard formulation of quantum mechanics, Everett ultimately
defended a much shorter version of the thesis (1957a), which was essentially the same as the journal
paper (1957b). While completed before the short version, the long version of Everetts thesis was first
published in DeWitt & Graham (Eds.) (1973). References herein refer to that publication.
is made; in which case, it evolves in according to the random collapse Process 1
(1973, pp. 3-4).
The quantum measurement problem, however, arises as a result of the
conflict between these two dynamical laws. If we suppose that measuring
devices are physical systems like any other, then the standard collapse theory is
inconsistent because the incompatible laws might be applied to the same
evolution; on the other hand, if measuring devices are somehow special, the
standard theory is incomplete since it does not tell us what interactions should
count as measurements.
Everett describes the problem with the standard theory as follows:
The question of the consistency of the scheme arises if one contemplates
regarding the observer and his object-system as a single (composite) physical
system. Indeed, the situation becomes quite paradoxical if we allow for the
existence of more than one observer. Let us consider the case of one observer
A, who is performing measurements upon a system S, the totality (A + S) in
turn forming the object-system for another observer, B. [] If we are to deny
the possibility of Bs use of a quantum mechanical description (wave function
obeying wave equation) for A+S, then we must be supplied with some
alternative description for systems which contain observers (or measuring
apparatus). Furthermore, we would have to have a criterion for telling precisely
what type of systems would have the preferred positions of measuring
apparatus or observer and be subject to the alternate description. Such a
criterion is probably not capable of rigorous formulation. (Everett 1973, p. 4)
Everett proceeds to tell his own version of a story that has come to be
known as the Wigners Friend story, then concludes from the story that
It is now clear that the interpretation of quantum mechanics with which we
began [the standard von Neumann-Dirac collapse theory] is untenable if we are
to consider a universe containing more than one observer. We must therefore
seek a suitable modification of this scheme, or an entirely different system of
interpretation. (Everett 1973, p. 6)
2

2
Everett here tells the Wigners Friend story some four years before Wigner (1961) himself tells
the story in his famous paper. That this story is being passed around illustrates part of the history of
worrying over the foundations of quantum mechanics at Princeton during the 1950s. It is likely that
the story was originally Wigners and that Everett picked it up as a student; perhaps in the seminar on
mathematical physics Everett that took from Wigner. See Wigner 1961, Albert 1992 and Barrett
1999 for discussions of the story and how it illustrates the measurement problem in the standard
theory.
Jeffrey A. Barrett A Structural Interpretation of Pure Wave Mechanics 227
After briefly considering a handful of options for resolving the
measurement problem, Everett describes his own proposal. His stated goal is
simply to drop the collapse postulate, Process 1, from the standard formulation
of quantum mechanics, then deduce the empirical predictions of the standard
collapse theory as the subjective experiences of observers who are themselves
treated as physical systems described by the theory. He calls his theory without
Process 1 pure wave mechanics.
Dropping Process 1 from the standard theory clearly solves the
measurement problem in that it removes the possibility of a conflict between
the two dynamical laws. But in dropping the collapse dynamics, one gives up
the standard explanation for why we get determinate measurement records and
the standard explanation for why these records are randomly distributed with
the usual quantum statistics. The determinate-record and probability
problems, respectively, are the problems of providing replacement
explanations for determinate measurement records and quantum statistics in
pure wave mechanics without appeal to Process 1.
3

Consider an object system S and an observer A such that if S is initially in an
eigenstate
of the measured quantity, then the state of the composite system
will evolve to
over the course of a measurement interaction; in

other words, suppose that the system S is undisturbed and the observer As
state is changed to
, which might represent determinately recording the

measurement result i in a notebook. If the initial state of S is not an eigenstate
of the observable being measured, but, rather
then, by the linearity of

Process 2, the evolution from the initial state to the final state of the composite
system in pure wave mechanics will be given by:

which, as Everett himself points out, is not a state that describes an observer
with any particular measurement result:
Thus in general after a measurement has been performed there will be no
definite system state, even though there is a correlation. It seems as though
nothing can ever be settled by such a measurement. Furthermore this result is
independent of the size of the apparatus, and remains true for apparatus of

3
The determinate-record problem is often taken to involve a further problem of choosing a
particular physical quantity with which to provide determinate values. This is the preferred basis
problem. See Barrett 2008 for a discussion of these three problems in pure wave mechanics.
quite macroscopic dimensions. [] This behavior seems to be quite at variance
with our observations, since macroscopic objects always appear to us to have
definite positions. Can we reconcile this prediction of the purely wave
mechanical theory with experience or must we abandon it as untenable?
(Everett 1973, pp. 61-62)
Further, regarding quantum statistics, is not at all clear how one is to get
the standard statistical predictions of Process 1 for the measurement records
(i) when one apparently has no determinate measurement records to which the
statistics might apply and (ii) when, as Everett explains, nothing resembling
Process 1 can take place (Everett 1973, p. 61).
Everetts goal then was to explain both determinate measurement records
and the statistical predictions of quantum mechanics in pure wave mechanics.
More specifically, he said that his strategy for providing this explanation would
be to
deduce the probabilistic assertions of Process 1 as subjective appearances []
thus placing the theory in correspondence with experience. We are then led to
the novel situation in which the formal theory is objectively continuous and
causal, while subjectively discontinuous and probabilistic. (Everett 1973, p. 9)
That said, it has never been entirely clear how Everett intended to resolve
either the determinate-record or the probability problems. It is not that Everett
had nothing to say about these problems; indeed, as we have just seen, he
shows that he clearly understood both in the very statement of his goal. The
difficulty in interpreting Everett arises from the fact that Everett had several
suggestive things to say in response to each problem, none of these suggestive
things do quite what Everett seems to be describing himself as doing, at least in
his strongest statements of his project, and it is unclear that his various
considerations can be put together into a single account of how one is to
understand the theory as predicting determinant records distributed according
to the standard quantum statistics.
4

Everetts discussion of the goals of theoretical physics near the end of the
long thesis, however, suggests a conservative strategy for how one might
understand his deduction of determinate measurement records and quantum
statistics. This strategy provides a concrete sense in which pure wave
mechanics can explain both determinate measurement records and quantum
statistics. While the sort of explanation provided is relatively weak, (i) it can be

4
See Barrett 2008 for a brief description of the considerations involved.
made perfectly clear, (ii) there is textual evidence that something like this is
what Everett had in mind as the proper standard for the empirical acceptability
of physical theories more generally, and (iii) it is closely related to a type of
explanation that has a long history of being taken seriously by both physicists
and philosophers of science.
In the second appendix to the long thesis Everett explains that an essential
goal of theoretical physics is to produce faithful physical theories. A faithful
theory is one that can be put into a close structural correspondence with the
elements of the perceived world:
Every theory can be divided into two separate parts, the formal part, and the
interpretive part. The formal part consists of a purely logic-mathematical
structure, i.e., a collection of symbols together with rules for their
manipulations, while the interpretive part consists of a set of associations,
which are rules which put some of the elements of the formal part into
correspondence with the perceived world. The essential point of a theory, then,
is that it is a mathematical model, together with an isomorphism between the
model and the world of experience (i.e., the sense perceptions of the individual,
or the real world depending upon ones choice of epistemology). (Everett
1973, p. 133)
And, in an associated footnote, he explains that:
By isomorphism we mean a mapping of some elements of the model into
elements of the perceived world which has the property that the model is
faithful, that is, if in the model a symbol A implies a symbol B, and A
corresponds to the happening of an event in the perceived world, then the
event corresponding to B must also obtain. The word homomorphism would be
technically more correct, since there may not be a one-one correspondence
between the model and the external world. (Everett 1973, p. 133)
To begin, note that here, in his most careful description of the aims of
theoretical physics, Everett adopts a broadly empiricist position; and, in this
spirit, he is careful to argue that it is a mistake to require a successful physical
theory to be descriptive of the ontology of the world. Indeed, he argues for the
stronger line that it is a mistake to understand theories as descriptive of
metaphysics at all:
[W]hen a theory is highly successful and becomes firmly established, the model
tends to become identified with reality itself, and the model nature of the
theory becomes obscured. The rise of classical physics offers and excellent
example of this process. The constructs of classical physics are just as much
fictions of our own minds as those of any other theory we simply have a great
deal more confidence in them. It must be deemed a mistake, therefore, to
attribute any more reality here than elsewhere. (Everett 1973, p. 134)
He then uses this stronger line to characterize his version of empiricism more
precisely.
Once we have granted that any physical theory is essentially only a model for
the world of experience, we must renounce all hope of finding anything like
the correct theory. There is nothing which prevents any number of quite
distinct models from being in correspondence with experience (i.e., all
correct), and furthermore no of ever verifying that any model is completely
correct, simply because the totality of all experience is never accessible to us.
(Everett 1973, p. 134)
Given the strong formulation of empiricism Everett develops here and the
associated metaphysical ambivalence that we find here and throughout the long
thesis, it is difficult to imagine that he might ever have held that any particular
set of commitments concerning the metaphysical structure of the world was
required for a proper understanding of pure wave mechanics.
5
The suggestion
is that his metaphysical ambivalence might explain both why Everett did not
make the careful distinctions that would have selected one set of metaphysical
commitments over another for the interpretation of pure wave mechanics and,
consequently, why readers can find room in his description of pure wave
mechanics for talk of such diverse ontologies as those suggested by splitting
worlds, many minds, many histories, and such.
Rather than describing the metaphysical structure of the world, a successful
physical theory for Everett is supposed to be somehow isomorphic to the world
of experience. But in the quotations above Everett conflates what one might
take as the theory itself and a formal model of the theory in his description of
what a physical theory is. Adopting the standard distinction between theory
and formal model, a theory on Everetts view is faithful when one can find some
elements of the model of the theory that are in fact isomorphic to elements of
the perceived world. It is arguably a short step from this reconstruction of
faithfulness to something like the constructive empiricists description of what

5
Everetts pervasive ambivalence regarding metaphysical issues has been most compellingly put
to me both in conversation and in forthcoming work by Brett Bevers.
it is for a theory to be empirically adequate.
6
Taking Everetts goal to be to
show that pure wave mechanics is empirical faithful might then be thought of as
a structural empiricist interpretation of Everett.
While empirical faithfulness is an essential virtue of a successful physical
theory, Everett argues that it is also desirable to have a theory that is
comprehensive and simple. While he allows for yet other theoretical virtues, he
takes these two virtues, in particular, to be important to inquiry. His argument
is that if ones current theory is comprehensive and simple, then it is more
likely to provide a suitable context for engineering future theories that might
exhibit a yet higher standard of empirical faithfulness. He concludes that
it may be impossible to give a total ordering of [rival physical] theories
according to goodness since different ones may rate highest according to the
different criteria. (Everett 1973, p. 136)
Nevertheless, it is in the context of a sort of cost-benefit analysis of faithful
theories for the purpose of engineering future faithful theories that Everett
considers the comparative virtues of pure wave mechanics.
One might grant that pure wave mechanics is comprehensive in that it treats
all physical interactions in precisely the same way and simple in that it involves
only one dynamical law and this law is deterministic and unitary, but it remains
to show that pure wave mechanics has the essential virtue of faithfulness. While
Everett provides sufficient material for several different approaches for
reconstructing the details of an argument for the faithfulness of pure wave
mechanics, I will briefly sketch just one such argument here with the aim of
showing that the model of pure wave mechanics does indeed have enough
structure that one can find a substructure understand systems as having states
relative to each other simply by dint of the precise way in which their
observable properties are correlated. More specifically, one can think of
relative states as what one gets when one chooses a physical system S and a
state of that system
then ignores all components of the entangled global

state of the world that characterizes S as being in any state other than
. If the

6
The constructive empiricist would say that an empirically adequate physical theory is one that
has a model with a substructure that is isomorphic to the phenomena as one chooses to represent
them. This statement is meant to capture some degree of flexibility in what one takes as the relevant
empirical substructure of the model and to capture van Fraassens most recent formulation (e.g.,
2008, p. 253) where he also allows for a corresponding pragmatic degree of freedom in how one
chooses to represent empirical phenomena.
single remaining component also characterizes the system R as being in state
, then we say that the state of R is
relative to the state of S being
.
Consider again an ideal observer A beginning in a state corresponding to
being ready to make a measurement of a system S that is initially in a
superposition of states corresponding to different values of the observable
being measured. Given the linear dynamics and the perfect correlations
produced by an ideal observer, the postmeasurement state of the observer A
and her object system will be the entangled state
above. Since the

observers notebook record of the measurement outcome is perfectly
correlated to the observable being measured (that is, since every term in the
representation of the global state in the determinate-record--determinate-
value basis has the form
), Everett would say that As notebook is in a

state where she determinately recorded the result k relative to the system S
being in the state
. But since the global state of the world will typically be a

complex entangled state, however, it is likely that no such simple
measurements ever actually occur; nevertheless relative to one having
performed a simple measurement, one may end up in a postmeasurement
relative state like
. If so, then relative to there in the model that is

isomorphic to the quantum-mechanical expectations supported by experience.
One might think of the model of pure wave mechanics is a structure of
complex-valued weighted correlations between the observable properties of
different physical systems at each time. On the unitary dynamics, when one
physical system interacts with another, the properties of the two systems
typically become correlated, and the composite system ends up in a
nonseparable state. It is enough to characterize an interaction to say which
properties become correlated and how and to what degree they become
correlated.
Everett developed several ways of talking about the correlations between
observable properties that fully characterize the correlation model of pure wave
mechanics. The most important of these for understanding how he thought of
the theory was his notion of a relative state. The basic idea is this: while the
global state of the world may be a complicated entangled state involving most
every physical system, one can always being a record that the measurement was
performed and relative to A recording the particular result k, S is in the
corresponding relative state
.
Consequently, if the concern is empirical faithfulness as characterized
earlier, the determinate-record problem is solved simply by noting that the
values of determinate measurement records for those measurements we take
ourselves to have performed can be found the correlation structure of pure
wave mechanics as relative states, relative to there being a record of the
measurements being performed and relative to the properties we take the
measured systems to have. Given a clear picture of such nested relative states in
the correlation model, addressing the probability problem requires only
slightly more subtlety.
There is a parameter determined by the correlation model that covaries with
our standard quantum statistical expectations. It is not the norm-squared of the
coefficients on the global state, but it is closely related. Consider the state of
the observer and her object system relative to the observer having a record that
the measurement was performed but not relative to any particular record of the
result.
One might renormalize this relative state, then think of it as a state
describing the superposition of possible measurement records that would
result from a simple measurement interaction on the linear dynamics. Suppose
that the coefficient associated with the term characterizing the notebook as
recording the result k is
. Our quantum expectations for result k covary with

the parameter
2
; in other words, the parameter
2
can be taken as
representing the degree to which the result k is expected given the usual
quantum statistics. The faithfulness of pure wave mechanics with respect to the
usual quantum statistics simply consists in one being able to find such a
parameter in the correlation model. And, taken together, that pure wave
mechanics is faithful to our determinate measurement records and the
statistical distribution of these records simply amounts to the fact that one can
find an isomorphic substructure to our statistical experience with determinate
records in the model of pure wave mechanics.
7

7
Note that the preferred basis problem does not even arise on this standard of empirical
acceptability since faithfulness requires only that one be able to find our determinate records in the
model. If one required rather that there be a substructure in the model pure wave mechanics that one
might on theoretical grounds alone identify as the empirically relevant substructure and if one
required that this substructure be isomorphic to our experience, as a constructive empiricist might,
then one might argue that pure wave mechanics fails to account for our concrete experience because
there is much more in the correlation than our concrete experience since the correlation model
contains something more like all physically possible experience. Even so, one might reply that there is
a sense in which the entire structure of pure wave mechanics is isomorphic to the general statistical
structure of our experience. Indeed, van Fraassen himself makes a move like this in his discussion of
quantum mechanics at the end of van Fraassen 2008. But just as with Everetts notion faithfulness,
one should wonder whether this is all one should want from a successful physical theory.
While there is a sense in which showing that pure wave mechanics is
empirically faithful involves deducing determinate records and the standard
probabilities from pure wave mechanics, there is also a sense in which this is
not a deduction determinate records or of the standard quantum probabilities
at all. We have not deduced that one should expect measurement records to be
determinate in a world described by pure wave mechanics nor have we deduced
quantum probabilities as probabilities in such a world nor have we shown that
there is something like a canonical way to rationally assign expectations given
the theory; rather, we have shown that that we can find a parameter in the
correlation model that covaries with the standard quantum expectations for
determinate records that we have from empirical experience and just that. Pure
wave mechanics nowhere tells us that relative states have the metaphysical
virtues of determinate physical records nor that the quantity
2
represents
an objective probability or a constraint on rational choice given the nature of
the physical world. Not only would such conclusions require just the sort of
metaphysical commitments that Everett seeks to avoid, but he also repeatedly
insists that pure wave mechanics makes no assertions concerning the
probabilities of any sort.
8
Rather than claim that we have somehow deduced
determinate measurement records distributed with the standard quantum
probabilities from pure wave mechanics alone, the procedure of showing that
pure wave mechanics is empirically faithful is better characterized as one where
we start with our empirically informed expectations, then find something in the
correlation model that covaries with our empirically grounded expectations.
The point is just that the determinate record and probability problems are
solved here in just the sense that one can find our actual empirical records and
our empirically supported quantum expectations in the correlation model of
pure wave mechanics. That there is much more than just our actual determinate
records in the correlation model is something that Everett can, and does,
embrace.
The remaining question is whether a theory being empirically faithful in
this sense is enough for it to be considered empirically acceptable. While one
might worry that faithfulness is a relatively weak standard for the acceptability
of physical theory, it is clearly not empty since not every physical theory has a
model with a parameter associated with representations of possible

8
The centrality of this point is reflected in the fact that Everett originally titled his long thesis
Quantum Mechanics without Probability. He also repeatedly insists on the stronger point that pure
wave mechanics itself makes no statistical assertions (e.g., 1973, p. 8).
measurement records that covaries with the quantum expectations we find in
experience. Further, pure wave mechanics has significant virtues beyond
empirical faithfulness it is, after all, difficult to disagree with Everetts claim
that pure wave mechanics is both comprehensive and simple. That said, one
clearly might want more than even this from a successful physical theory, but
what more and why are questions for another occasion.

REFERENCES
Albert, D. Z. (1992). Quantum Mechanics and Experience. Cambridge, MA:
Harvard University Press.
Barrett, J. A. (1999). The Quantum Mechanics of Minds and Worlds. Oxford:
Barrett, J. A. (2008). Everetts Relative-State Formulation of Quantum
Mechanics. In Stanford Encyclopedia of Philosophy.
<http://plato.stanford.edu/entries/qm-everett/>
DeWitt, B. S., & Graham, R. N. (Eds.) (1973). The Many-Worlds
Interpretation of Quantum Mechanics. Princeton, NJ: Princeton
University Press.
Everett, H. (1957a). On the Foundations of Quantum Mechanics. Ph.D.
Thesis, Princeton, NJ: Princeton University, Department of Physics.
Everett, H. (1957b). Relative State Formulation of Quantum Mechanics.
Reviews of Modern Physics, 29(3), 454-462.
Everett, H. (1973). The Theory of the Universal Wave Function. In B. S.
DeWitt & R. N. Graham (Eds.) (1973), The Many-Worlds
Interpretation of Quantum Mechanics. Princeton, NJ: Princeton
University Press.
Van Fraassen, B. (2008). Scientific Representation: Paradoxes of Perspective.
Oxford: Oxford University Press.
Wigner, E. (1961). Remarks on the Mind-Body Problem. In I. J. Good (Ed.),
The Scientist Speculates (pp. 284-302). London: Basic Books.


Commentary
Metaphysical Language, Ordinary Language and
Peter van Inwagens Material Beings
*

Peter van Inwagen
Cornell University Press, Ithaca, 1990

Daniel Nolan**
daniel.nolan@nottingham.ac.uk

Material Beings is an immensely important work in contemporary metaphysics,
even though hardly any metaphysicians accept its central conclusion. (Many
works of contemporary metaphysics are like this: metaphysicians are a
disputatious lot). There is a lot of value in Material Beings about the
metaphysics of parts and wholes, material addressing the metaphysics of
existence over time, puzzles about existence of people over time, and a
surprising defence, in the final two chapters of the book, of abandoning
classical logic in metaphysics in favour of a three valued logic. In this
commentary, however, I will focus on what I take to be the main conclusion of
the book, what van Inwagen says to sugar the pill of this conclusion, and a new
problem that arises for van Inwagens theory which is very similar to the sort of
problem he is at such pains to solve. Finally, I suggest that reflection on this
new problem raises an epistemic challenge to van Inwagens position.
The main conclusion of Material Beings is perhaps the second-most
surprising claim in that book. It is, to put it baldly, that the only material
objects that exist are either ultimate material particles, or living beings. No
other kinds of material objects exist: no cups, no clouds, no clothing, no
mountains, no benzene molecules, no dead bodies, no planets or stars. This
conclusion is useful for answering a number of traditional puzzles about
material objects: in dealing with a tricky case of personal identity over time
involving brain removal, for example, van Inwagen can say, as he does, [t]he

* Thanks to Carrie Jenkins, Shieva Kleinschmidt, Ted Sider, Robbie Williams and the
metaphysics group at NYU for helpful discussion.
** Department of Philosophy University of Nottingham
solution to this paradox is simply that ones brain does not exist (p. 172).
Since the only things that exist are living creatures or ultimate particles, the
only parts that living beings have are themselves either living beings or ultimate
particles: I may have electrons or maybe cells as parts, but not things like hands
or a brain.
The theory that there are no material objects besides living things and
ultimate particles is unpopular, in my view deservedly unpopular. But there are
important theoretical pressures pushing us towards van Inwagens position,
and it is the genius of Material Beings that van Inwagen marshals his case for
the view in such an intriguing way.
Van Inwagen begins by posing what has become a central question in the
metaphysics of parts and wholes: the Special Composition Question (pp. 30-
31), hereafter SCQ. It is the question of what the necessary and sufficient
conditions are for some objects (the xs) to compose another object (y): where
for the xs to compose y is for all the xs to be parts of y, distinct xs to not
overlap, and for every part of y to overlap one of the xs. That is, when do some
things make up something? Van Inwagen argues against some representative
answers to the SCQ, and proposes his own solution: that the only way some xs
can compose something is either if 1) there is only one of the xs, and it is an
ultimate particle with no parts other than itself, or 2) the activity of the xs
constitute a life (p. 115): the xs make up a living thing. Much of the last part
of van Inwagens book consists of examining how different problems about
parts and wholes are resolved in the light of his answer to the SCQ.
I believe van Inwagen has another motivation for his preferred answer
besides giving a good answer to the SCQ, which surfaces at a number of points
in his discussion.
1
There are a host of paradoxes about parts and wholes:
arguments with apparently plausible premises that yield contradictory
conclusions. Consider, for example, the ancient paradox of Dion and Theon.
Dion is a man, and Theon is the large part of him which includes everything
except his left foot. Suppose Dion has his foot amputated. Plausibly, he and
Theon become the same entity, since Theon underwent no change but now all
Dions parts are Theons parts. But on the other hand it seems that they cannot
be the same entity: at the later time, it is true that Dion used to have two feet,
for example, while it is not true of Theon that it used to have two feet.

1
See e.g., pp. 69-71, 78, 179.
Commentary Material Beings 239
Or consider a case involving a valuable antique car which at the beginning
of the story only has three wheels, having lost one long ago. Call the car at the
beginning Ridge (contracted from Original). Suppose I attach a garish
new wheel, complete with shiny hubcap, to Ridge. It seems I will then have a
four-wheeled car: let me dub the four-wheeled car after the attachment
Hercules. Hercules, it seems, has a large part with old components: all of it
except the new wheel (and screws). Call that large part of Hercules Hercules
Minus. What happened to Ridge? It seems I did not destroy that car, so it still
exists. Cars can gain wheels, so maybe Ridge is now Hercules. But Hercules
Minus also has a good claim to be Ridge it is entirely antique, like Ridge was,
and is made of the same parts as Ridge. But it seems Ridge cannot be both,
since Ridge did not have the features about to gain a fourth wheel and about
to continue to have exactly three-wheels at the same time. All the consistent
ways to resolve this puzzle seem initially unattractive. But the puzzle does not
arise if there are no cars to begin with. Heavily restricting what material objects
his theory is committed to seems to enable van Inwagen to sidestep these
vexing puzzles.
I said, above, that I thought van Inwagens conclusion that no material
objects except ultimate particles and living creatures was the second most
surprising claim in Material Beings. The most surprising claim in Material
Beings is this: that nothing that ordinary people say or think is in conflict with
van Inwagens solutions to these puzzles (pp. 98-102). Despite what you may
have thought, you probably have never said or believed anything that implies
that some people own cars, or that chairs can be found in your home, or
anything else that entails that objects such as cars or chairs or seas or stars
exist. (Except perhaps if you have engaged in metaphysics). Van Inwagen does
not spell out in detail how his claims in Material Beings are consistent with
ordinary beliefs and utterances: he says I do not propose to defend my
philosophy of language in the present work (p. 102), and as far as I know he
never gives an entirely systematic treatment of this issue elsewhere. Instead, he
illustrates his thesis with analogies (pp. 101-102): just as someone who, when
asked whether its raining, can sometimes properly reply It is and it isnt, or
someone who accepts Copernicanism can still, in the ordinary course of
events, talk about the sun moving during the course of an afternoon, ordinary
people can say There are two chairs in the next room and say something true
even though, by van Inwagens lights, there are no chairs anywhere.
Here is a suggestion that I hope is true to van Inwagens intentions here.
Let us suppose that there are two ways of talking about what material objects
there are. There is a strict and careful way, which philosophers typically
engage in when discussing metaphysics or ontology, and a more loose or less
constrained way which is what we all use in the ordinary course of events in
describing the world and also describing our beliefs about the world. When
asking, in the strict way, what material objects exist, van Inwagen thinks the
answer is that every material object is either a simple or a living creature. For
most claims apparently involving material objects to be true in this strict sense,
according to van Inwagen, those objects must exist and be the way they are
characterised: under the strict interpretation, there are two red chairs in my
office requires, for its truth, at least the existence of two chairs which are red.
However, that sentence also has a loose or popular reading on which it can be
true even if two chairs exist is false in the strict sense. If I wish to say, strictly
speaking, what it takes for the sentence there are two red chairs in my office
to be true when interpreted loosely, all it requires is that there are tiny ultimate
particles arranged in a certain way. Van Inwagens hypothesis, then, is one that
he offers in the strict mode of talking: speaking strictly, for example, no chairs
exist. Van Inwagen takes himself to be disagreeing with other metaphysicians
who he takes to be speaking strictly and say that chairs do exist when they
speak strictly. Van Inwagen takes himself to have no direct dispute with people
speaking only loosely about this issue: indeed, he agrees that loosely speaking,
chairs exist. (And he will agree, loosely speaking, that people own cars, or that
chairs can be found in your home, or anything else that entails that objects such
as cars or chairs or seas or stars exist. So he might complain that I put the
surprising feature of his view in a misleading way, above. I think he is only
committed to the view that what is meant by people own cars, for example,
when interpreted strictly is something that virtually no ordinary speaker or
thinker says or believes).
Interpreted this way, van Inwagens claim that his thesis does not conflict
with much of what we ordinarily say and believe does not seem as incredible.
Indeed, it can make his central claim seem far less incredible as well. Many of
us are very reluctant to think that a clever philosopher could show us that, after
all, there were no brains or chairs or cars. But if it turns out that he only intends
to show us that no brains exist in a special sense, furthermore one which we
do not ordinarily use when discussing anatomy, then he is perhaps not
disagreeing with our ordinary opinion about the world, at least not to the
extent it might have first appeared.
Let us grant, for the sake of the argument, that metaphysics is carried out in
a special strict jargon, and that when van Inwagen says that there are no
tables or chairs or brains he does not say anything in conflict with what we
ordinarily say when we tell our children that there are brains inside our skulls,
or when we tell a colleague that there is a spare chair in our office they can
borrow. Let us also grant, for the sake of the argument (though only for the
sake of the argument!) that van Inwagen is right that, speaking strictly, there
are no material objects besides simple ultimate particles and living beings.
Speaking with the metaphysicians, then, there is no puzzle about a car that
gains a new wheel, because there are no cars and no wheels.
But what happens if we re-ask our puzzle about the car and its new wheel in
ordinary loose language? Suppose we make clear that we are not talking in
any special strict way, and then tell the story of a certain car, Ridge, and the fact
that a wheel was attached to it, that the four-wheeled car was dubbed Hercules,
and then we ask whether Ridge is the same thing as Hercules or not. Now,
since we are speaking loosely, van Inwagen should agree, on pain of changing
the subject, that there are cars (in the sense in play), and that sometimes they
continue to exist and get renamed, and sometimes they go out of existence.
When I ask the loose question whether e.g., Hercules used to have three
wheels, or whether Ridge now has four tyres or only has three (but is attached
to a wheel with another tyre), it looks like we face a challenge very similar to the
one described above. We face the same problems with each answer. Loosely
speaking, cars can lose a wheel and later regain another one. Loosely speaking,
the new wheel is attached to an old object which is not destroyed by fixing a
wheel to its exterior. We feel the temptation to say that Ridge is identical to
Hercules to say that they are the same car, which has just grown by a wheel,
and we also feel the temptation to think that Ridge is still in existence (and still
has three wheels), as a large part of Hercules but differing from Hercules by a
wheel. It is not even loosely true, at the later time, that there is something that
is both identical to Hercules and only has three wheels, when Hercules has four
wheels.
When we speak loosely, we are tempted to accept all the premises stated for
the argument, and deny the apparently contradictory conclusion, even when it
is stated in our ordinary idiom. Being told that there is another way of speaking
where we would not talk that way seems only of limited help: of course the
problem might not come up in conversation if we no longer talked about cars
or wheels, but that would be avoiding thinking about the problem rather than
solving it.
If we accept van Inwagens central contention, and then accept his thesis
that it does not conflict with what we ordinarily say, then it seems we can re-ask
the puzzle that motivated us in an ordinary idiom, and we seem to have a very
similar puzzle back again. Indeed, it may even be the original puzzle: van
Inwagen does not tell us when people started to speak in his strict metaphysical
idiom, but it may be that those thinkers who originally posed this sort of
challenge had not yet shifted to the rarefied form of speech van Inwagen
attributes to metaphysicians today.
Without an answer to the loose question, van Inwagens picture seems
incomplete. Were we to say to him Look, in the sense that there are cars and
wheels, is Hercules identical to Ridge?, what is the best reply he has available?
He could argue that by using expressions like is identical to, we have,
despite ourselves, slipped into the strict idiom, and so must be answered
neither Hercules nor Ridge exist (or are identical to anything). But that
insistence seems forced, especially since e.g., holding up two before and
after photos and asking is this the same one as this seems a pretty ordinary
thing to ask (and might be asked for a non-philosophical purpose, such as
working out whether our attempt to find a stolen car is successful). So he
should not adopt this reply.
Van Inwagen does have some things to say about some puzzles about
identity through change of parts. When he describes the Ship of Theseus, a
paradox about a ship gaining and losing parts, he admits we can speak loosely
of ships, but says that after his speaking about the relevant events in his strict
vocabulary there is no philosophical question to be asked about the events I
have described (p. 129). Later, he says more generally, we shall be able to
formulate no philosophical questions about the identities of artifacts at all (p.
130, his emphasis). Perhaps by philosophical questions he means questions
posed in the strict way of talking, in which case his insistence that these
problems do not arise when speaking that way is entirely understandable. But
perhaps he is suggesting that questions asked in loose speech (such as the
question of whether, loosely speaking, Hercules was Ridge) are not questions
that philosophy should notice or address. If I granted that, I would then want to
ask the questions nonetheless, philosophical questions or not. And it does
seem a sociological error to think that these questions are not ones that
philosophers are interested in asking and answering, even if van Inwagen
would prefer not to do so. So I think it would be better for van Inwagen (or at
any rate some proponent of his view) to indicate what answers he thinks are the
correct ones to these questions, when the questions and answers are spoken in
the ordinary loose way of talking.
Van Inwagen has a range of other options here: most have been thrashed
out in the literature on identity across time and identity through change of
parts, which for the most part does assume that there are things like cars and
wheels. Perhaps, loosely speaking, Ridge at the earlier time has become both
Hercules and Hercules-minus-a-wheel: if we say this, we will probably want to
resist inferring, by the transitivity of identity, that since Hercules=Ridge and
Ridge=Hercules-minus, then Hercules is identical to Hercules-minus: but
perhaps, when we speak loosely, we can say that identity is not transitive, or is
only transitive at a time but not across times. (Van Inwagen would not be
prepared to admit counterexamples to the transitivity of identity when we
speak strictly, of course, but saying identity is not transitive in a loose
context may not mean anything inconsistent with the principle we endorse in
the strict context). Or we may want to allow, speaking loosely, that while
Hercules-minus used to be Ridge, it was never true that Ridge was going to be
Hercules-minus. Or we might want to say, speaking loosely, that Hercules and
Hercules-minus were never the same object, but both used to exist exactly
where Ridge existed, with the same shape, colour, and so on. Again, van
Inwagen objects to thinking that, strictly speaking, there are ever two material
objects in the same place at the same time, but perhaps it is okay to talk loosely
in this way, just as we might loosely talk about the double life of an
accountant-by-day, DJ-by-night, or even speak of Jenna-the-accountant and
Jenna-the-DJ as if they were two women, even though strictly speaking they are
the same person.
There are other options as well: van Inwagen may even be prepared to let us
talk loosely of temporal parts of objects. He objects to a temporal parts
metaphysics as any part of the truth, strictly speaking, of material objects, but I
do not know of anything he has said against the idea that we could help
ourselves to that sort of way of talking when we are talking loosely. In the case
of Jenna, above, perhaps van Inwagen would allow it does no harm in ordinary
talk to speak of Jenna-the-accountant being around for nine hours, then being
replaced by the other Jenna for the night: provided we do not take the talk with
metaphysical seriousness.
As well as selecting from the standard menu of options, van Inwagen could
also allow that any of these ways of talking are all right when it comes to
ordinary, loose talk about cars and their wheels. (Though we might want to
avoid mixing several ways of talking on one occasion presumably we should
not say in the same breath that Hercules and Hercules-minus were always
distinct, and that they used to both be identical to Ridge). Or that it is an
indeterminate matter which way is the way we ought to speak when we speak
loosely. Or maybe all the ways of speaking loosely are somehow defective, so
none is how we ought to talk, but one or more of them is good enough for
practical purposes. This last option is less friendly to ordinary talk than van
Inwagen seems to be: when van Inwagen tells us the metaphysical truth does
not contradict our ordinary beliefs (p. 98), and affirms that people very often
say true things with sentences about chairs or stars (p. 100), presumably the
ordinary talk is not so defective that it stops us saying the truth and surely
expressing the truth about subjects at issue is pretty good!
The puzzle about what to say when speaking loosely about what happens
when Ridge gains a wheel seems to call out for a solution. As I have suggested,
a number of responses seem available to van Inwagen: though by the same coin
this suggests that what he has told us so far does not deliver an answer about
which response is correct (or which responses are correct).
Once van Inwagen allows that even one of these responses are acceptable
when we speak loosely (two things becoming one, dividing things into
temporal parts, saying that there are two things in the same place at the same
time), another issue arises. If van Inwagen allows that one (or more) of these
ways of talking can be used to express truths when we say, loosely speaking,
that cars sometimes gain and lose wheels, presumably whatever judgements
van Inwagen relies upon to yield his strict answer to questions about gaining
and losing parts are not incompatible with what we express when we talk in
these loose ways. But we might wonder about the epistemic credentials of the
premises that van Inwagen insists are strictly true.
We might start off convinced that (speaking strictly) if a=b, then anything
true about how a will be in the future is true about how b will be in the future,
for example. But suppose we are then forced, by reflection on the case of
Ridge, to allow that if a=b, then anything true about how a will be in the future
is true about how b will be in the future was not universally true when
interpreted as loose talk. Or suppose we started by thinking that no two
material objects occupy exactly the same place at the same time is strictly
speaking true, but then concede that the sentence no two material objects
occupy exactly the same place at the same time expresses a generalisation with
exceptions when understood loosely. (e.g., we do not allow that one object
becomes two, but rather that, loosely speaking, Hercules and Hercules-minus
used to occupy the same place and there were two objects there all along). One
principle, stated strictly, being true while another principle, said loosely, being
false is of course quite possible, even if the two principles sound the same
sound is not an infallible guide to meaning. But it may make us wonder whether
we really have good reason to believe the principles when stated strictly. After
all, our pre-theoretic beliefs about these matters may not have been very
sensitive to the difference between the two claims we can now distinguish. And
once we concede that those principles, interpreted loosely, are untrue, we may
wonder whether any certainty we initially had about the strict-speaking version
of those principles is still warranted. If someone speaking loosely speaks truly
when he says two things can become one thing or two material objects can
be in the same place at the same time, why should we be so sure that someone
speaking strictly could not be speaking the truth when she says two things can
become one thing or two material objects can be in the same place at the
same time? Especially since those people can coherently say, speaking
strictly, what van Inwagen cannot: that there are tables, chairs, cars, planets,
and all the other things which many of us believe in, even when we are speaking
with the metaphysicians.
Nothing I have said is intended as a knock-down objection to van Inwagen:
as far as I can tell, his is an internally coherent position to hold onto, and I
expect once one gets used to saying and thinking that, strictly speaking, all that
exists are ultimate particles or living creatures, it can even come to seem an
intellectually comfortable position. And Material Beings raises important
challenges for those who do not wish to follow van Inwagen, which we would do
well to pay serious attention to.
But for those of us who think we would need a compelling reason to accept
the conclusion that e.g., strictly speaking I have never worn any clothes (for
there are no clothes to wear), van Inwagens position often fails to convince. In
this note, though, I have tried to focus attention on a set of questions which van
Inwagen does not give detailed answers to. Speaking loosely, when do some
objects make up a whole that contains only parts that overlap those objects?
Speaking loosely, in the case of Ridge and Hercules, is Ridge Hercules, or
Hercules-minus (or both, or neither)? And so on for the many other questions
about parts and wholes which we can ask in our ordinary loose way of talking. I
suspect once a philosopher who holds van Inwagens views answers these
questions in loose talk, the answers might sound just like answers given by
some or others of van Inwagens apparent opponents. And the answers will
involve allowing that premises that van Inwagen relies upon, when he speaks
strictly, sound just like sentences which, when said loosely, are mistaken.
Perhaps van Inwagen or his allies will be able to explain to us how we can be
sure that the claims made with the strict sentences are true while the claims
made with the loose sentences which sound very like them are false. But this is
work that remains to be done.

Commentary
Lost in The Labyrinth of Time
Michael Lockwood

Giuliano Torrengo*
giuliano.torrengo@gmail.com

Lockwoods book is a all but complete guide to the concepts and theories that
you have to understand in order not to get lost in the labyrinth of time. Many
threads pass through the whole work. The first six chapters of the book are an
introduction to Special and General Relativity, the revolutionary notion of time
that emerges from them, and the cosmological hypothesis based on them.
Chapter seven introduces the philosophical problems that the idea of time
travel raises. Chapters eight to thirteen are focused on the problem of the
arrow of time, entropy, and the emergence of order. The following three
chapters deal with the problem of the interpretation of quantum physics. The
final chapter is on the psychology of time perception. In this commentary, I will
mainly focus on the issue of time travel, from the classical paradoxes to their
approach in a quantum setting.
The ordinary notion of time seems to carry along the idea that time passes,
and that the passage of time is something out there objectively in the world.
However, in Special Relativity the temporal dimension can be separated from
the three spatial ones only relative to a frame of reference. This situation gives
rise not only to the counterintuitive notion of the relativity of simultaneity, but
also to the idea that the reality of time is nothing over and above that of its unity
with space in space-time. Within each frame of reference, temporal and a
spatial distance between events are constant quantities, but normally only
spatiotemporal intervals (a certain relation between the two) remain constant
through variations of frame. If an objective time flow requires that the distance
between events be frame independent, then nothing seems to be left of the idea
that the passage of time is a genuine, or even essential, feature of reality.

* University of Torino

In General Relativity, space-time reveals further disturbing properties,
with respect to our ordinary notions of (space and) time. The geometry of
different regions of space-time varies depending on the distribution of matter
and energy that we find in the corresponding region. Space-time is not flat
although locally we can consider it as flat it is curved by matter and energy,
i.e., matter and energy stretch and shrink space-time intervals. This allows us
to interpret gravity not as a force, but as an effect of the curvature of space-
time. Besides, how precisely to shape up the spatiotemporal metric within a
frame of reference is by large a matter of practical convenience, given that the
interesting physical quantities remain constant through modifications of the
metric, that is many radically different metrics are indistinguishable with
respect to the physical laws governing the interaction between energy-matter
and space-time (the Einstein field equations). In particular, the space-time
manifold can be foliated in many (reciprocally incompatible) ways into three-
dimensional layers, and none of them has any physical or metaphysical
significance over the other. If an objective flow of time requires a global and
absolute succession of now, again such a notion seems to receive a fatal blow
form physics.
However, many cosmological models based on General Relativity such as
the Friedmann-Robertson-Walker models and the inflationary models depict
the universe as starting at a singularity, the Big Bang, and continuing in its life
with distinct, successive, phases. It is therefore tempting to use certain
cosmological characteristics, such as the mean distribution of matter and
energy, or the constant curvature of certain hypersurfaces (i.e., three-
dimensional slices of space-time) to signal out a preferred foliation and thus
a cosmic time to be identified with the objective flow of time that we all
experience.
1
This project, as Lockwood makes clear, has proved to be difficult
to carry out in a convincing manner. Any way to spot a preferred foliation
remains at bottom unjustified, and what is worst, General Relativity gives us
reasons to doubt even the possibility of providing any global foliation such as
required by a cosmic time. The Einstein field equations are compatible with the
presence of closed time-like curves (CTC), that is path in space-time such that
(1) can be followed by an object at a sub-luminal speed, and (2) with respect to

1
See for instance Lucas 1999. For further critics to the project see Bourne 2006 who argues
that even if the project were sound, cosmic time would not make do to play the part of tensed time
i.e., time with objective time flow.
Commentary The Labyrinth of Time 249

reference frames anchored to objects that do not follow a CTC, an object that
follows an almost complete CTC will arrive at a time that is earlier than its
departure time. Thus, many space-time manifolds that contain CTC cannot be
foliated on a global scale although it is still possible to define locally a
temporal order.
2
For instance, we can distinguish between the public time of
ordinary people who do not follow CTC in living their lives, and the personal
time of someone following a CTC, i.e., a time traveller.
CTC figures prominently in Gdels argument that General Relativity, by
allowing CTC, forces us to renounce to an objective notion of lapse of time
and even to embrace an idealistic view of time to the effect that temporal order
is not objective. Lockwood grants to Gdel that the physical possibility alone of
CTC menaces the objectivity of the temporal passage and order
3
, but he rightly
maintains that simple compatibility with Einstein field equations does not
necessarily boil down to physical possibility.
The possibility of having CTC in a spatiotemporal manifold clearly is
related to the possibility of having time travel. However, as also Gdel seemed
to think, one can accept CTC in a manifold while refusing the possibility that
such CTC could be followed by ordinary object or persons long enough to
allow a time travel situation. The grandfather paradox is the standard
objection to the possibility of time travel. Suppose that time travel is possible
and that twenty years old Tim embarks on a time machine for the fifties. It
seems as Tim could reach his grandfather and kill him before he generates
Tims father, thus preventing his conception to happen. Since it is not possible
that both Tim is and he is not born, we have to give up the hypothesis that time
travel is possible. Philosophers have questioned this line of reasoning in
various ways, after the seminal articles by Harrison 1971, Lewis 1976 (and
Horwich 1975, who is more critical towards the idea though). The standard
view, which in the first part of this chapter Lockwood provisionally defends, is
that self-consistent time travel does not require miracles or extraordinary
forces to prevent paradoxes, but only unusual (for non-time travellers)
coincidences. The situation is not substantially different from cases of

2
CTC does not imply that the manifold possessing them cannot be globally foliated. If a
spacetime is closed by possessing a cylindrical topology, then it both possesses CTC and can be
globally foliated.
3
This is not the most common attitude; see Dorato 2002, Bourne 2006. Even the defence of
Gdels argument in Yourgrau 1999 seems to suppose that the main objection to Gdels argument
lies there. See also Calosi 2009.

foreknowledge. If a infallible foreteller predicts that I will not go to Mongolia
within the next five years, then no matter how hard I try to reach Mongolia in
the meanwhile, each time I try to go to Mongolia something will happen
preventing me to reach it: the plane is compelled to land somewhere before
Ulan Bator, the train derails before getting to the border, a snow storm stops
my expedition from Siberia to the Gobi Desert, and so forth and so on.
Self-consistent time travel requires that any attempt by a time traveller to
prevent her or his own birth is bound to fail as in general any bilking
attempt, i.e., any attempt to change the past or prevent an unavoidable event.
The particular reason why it is going to fail can be very different from one time
to another, but what is the general and deep reason for there being each time a
failure? If we discharge miracles and other non-ordinary forces, the only
answer seems to be that there is a conspiracy in nature that prevent the initial
conditions of any bilking attempt to be such to lead to contradiction. In
insisting that this situation is unacceptable, both for common sense and for
standard scientific practices, Lockwood parts company with the standard view
on the issue. It is important to notice that the constraints on the initial
conditions that are imposed by many time travel scenarios arise even in absence
of free will agency. In order to show it, Lockwood provides an intriguing
example (akin to a example in Earman 1972). Consider a train track that leads
to a wormhole tunnel with the following characteristics: if a train enters it at
a certain time, it exits from it ten minutes before it has entered it. Now, a
computer that controls a train T can be programmed in a way that if a train exits
from the tunnel at a certain time, then T stays at rest in the station, while if no
train exits from the tunnel, then T enters the tunnel. Assuming there is only
one train around, this situation leads us to a paradox a purified form of the
grandfather paradox, freed of any reference to human free agency.
If conspiracies on the initial conditions are required in order to avoid
contradictory situations of this kind, then time travel although not
contradictory seems to imply a violation of what Deutsch has called the
Autonomy Principle.
4
It is a matter both of common sense and of ordinary
scientific practice to assume that if locally the laws of physics admit of a certain
configuration of matter, then the global situation of the universe cannot
constraint the possibility of such a configuration of matter. If we can construct

4
For a critique of the idea that those cases are substantially different from cases implying free
agency, see Sider 1997.

a train and a computer with the physical properties that we have just described,
then we should be able to do it even in the proximity of a wormhole tunnel.
Global consistency, however, seems to require that this will not be possible
unless external factors interfere with the train or the computer: a meteorite
hitting the train before it enters the tunnel, or a electromagnetic field
interfering with the computer, say.
However, physicist Kip Thorne and his collaborators
5
has shown that, if the
laws governing a system are continuous, then it is possible in presence of
CTC to find out consistent continuations of initial conditions that do not
require conspiracy. Indeed, the problem turned out to be quite the opposite.
Many initial conditions show the following feature: there is more than one
consistent evolution of the system, and none of them is more likely to happen.
Imagine a billiard table in which the two central holes have the following
feature: whatever enters in the right hole exits from the left hole three seconds
before it has entered in the right hole. Now, roll a ball between the two holes:
what is going to happen? A consistent continuation of the initial conditions
(the launch of the ball in a straight trajectory between the two holes) is that the
ball passes between the two holes as it would on a normal billiard table.
However, it is also perfectly consistent to suppose that while the ball passes
through the holes it is hit by a ball emerging from the left hole, which deviates it
in such a way that it enters the right hole with a momentum such that it exits
from the left hole three seconds before and hits itself as it did. Many other
continuations are all possible and none of them is more probable than any
other to happen. In this example, as in many other settings, they are infinite!
Even a Laplacian demon possessing a perfect knowledge of all natural laws and
all former states of the universe would not be in a position to know what lies in
the future if it contains CTC, or to tell what is more likely to happen.
Lockwood does not discuss the issue whether this solution to the problem
can be generalized, and thus time travel never requires conspiracy on initial
conditions, or there are physically plausible cases that would require a failure
of the Autonomy Principle.
6
However, he does take into account the failure of
determinism at a macroscopic level implied by time travel, and acknowledges

5
See, for instance Echeverria et al. (1991). Their work follows previous seminal investigations
by Feyman and Wheeler 1949.
6
For a discussion of this problem see Artzenius and Maudlin 2002, and also Earman and
Wuthrich 2004.

that such a failure of determinism is more dramatic than that due to
probabilistic causation of quantum phenomena, in which we can usually weight
the probability that an effect is brought about against that of other effects.
Indeed, the solution he proposes to avoid failure of the Autonomy Principle is
intended to apply also to the underdetermined cases.
Lockwood introduces his many minds interpretation of quantum mechanics
with a lengthy and intriguing discussion of the measurement problem, focusing
on Einstein and Schrdinger criticism of the Copenhagen interpretation and
the failure of the local hidden variables solution provided by the Bell theorem.
While Einstein was mainly concerned with the lack of determinism,
Schrdinger was more concerned with the lack of continuity implied by the
quantization of the states of quantum systems. The Schrdinger equation,
indeed, describes in a continuous manner the evolution of the wave function of
micro-physical systems, by attributing to them superposed states. Now, how
should we construe the superposed states of a system that the continuous
Schrdinger equation implies given that attributing them simultaneously to
the same system is clearly contradictory? According to Schrdinger,
measurement does not collapse the wave function, turning the evolution of the
system into a discontinuous process, but only makes the macro-system of the
measurer entangled with the micro-system of the observer. The idea of
construing superposed quantum states as alternatives comes from a particle
interpretation of the reality that quantum theory is talking about, and from the
fact that when we measure we only observe one of these states. But in a way this
interpretation is just a consequence of our limited point of view, due to our
being entangled with the measured system.
If we reject the idea of altering the continuous dynamic of the Schrdinger
equation (as in collapse-Copenhagen, De Broglie-Bohm, GRW
interpretations), then the only alternative seems to be that of accepting that
superposed quantum states all simultaneously exist in parallel realities.
Schrdinger was the first who realized that if we take quantum theory at face
value, it predicts the existence of parallel realities, not only at the micro, but
also at a macro level. However, he also never managed to come to terms with
this conclusion, which is the core idea of Everetts interpretation of quantum
mechanics. Lockwood is quite careful in linking his position to that hinted at in
Schrdingers last lessons, rather than to the many-worlds interpretation that
Everetts work pioneered. Every interpretation of quantum mechanics that
accepts that superposed states all simultaneously exist has to reshaping our

concept of the macroscopic world so as to accommodate the idea of parallel
realities. The idea behind the many mind interpretation is that there is no
objective division of parallel realities. Rather, our minds which are
subsystems of our brains form, along with all the states that are possible
object of detection by them, the basis for such a division. Therefore, speaking
of parallel worlds composing what is sometime called a multiverse although
harmless in many occasions may be misleading with respect to such
interpretation.
The fact that the division between the many distinct realities is subjective
in this sense does not mean that according to Lockwood what exists depends
on the subjects. The main point of Lockwood here is that what in classical
relativity is thought to be a space-time manifold is indeed a space-time-actuality
manifold. Time, even more than space, paves the way for diversity in unity,
but space-time [may not be] the only arena within which Nature is able to
spread herself (p. 314). If the actual state of the world is simply the state in
which we happen to be in, then we can think at the dimension of actuality as a
dimension comprising simultaneous diversity within reality. If the hypothesis
of the space-time-actuality manifold is correct, then we can have diversity in
unity not only along the temporal dimension (today I have coffee as breakfast,
tomorrow I have tea), and along the spatial dimension (my head is warm, my
toes are chilled), but also along the dimension of actuality (today, in the
kitchen, in a superposed state, I am drinking coffee, in another state I am
drinking tea).
If the arena of reality is that of a space-time-actuality manifold, then it seems
that the problems raised by the possibility of time travel can be solved without
renouncing to the Autonomy Principle. Consider again the case of the
paradoxical train. According to this interpretation of quantum mechanics,
the train ends up in a mixed state encompassing two states (with equal
weights
7
). Both such states exist in two parallel realities: in one the train enters
never to be seen around, while in the other it emerges from the tunnel although
it never entered it. The same goes for underdetermined cases. Keeping the
previous example of the billiard balls in mind: in one reality the launched ball
goes through the holes, in another it is hit by a previous self, and so on. In

7
How do we make sense of states possessing different weights, as often happens in quantum
mechanics? According to Lockwood, the weight of a state s corresponds to the size of the regions of
actuality in which we find s.

general, objects and persons, by moving through the dimension of actuality,
can go back in time and reach, indeed create in a sense, different realities
from the one they come form.
Lockwood does not seem to be worried by a philosophical problem that has
been raised for the multiverse solution to the grandfather paradox, and which
still seems to stand for his interpretation.
8
Is travelling through a space-time-
actuality manifold a genuine case of time travel? Consider the grandfather
paradox again. By entering a parallel universe, Tim can indeed achieve his
murderous purposes and kill his grandfather. However, did he really manage to
reach the past? A first problem here is that speaking of the past in a space-
time-actuality manifold is ambiguous: the term can refer either to that area of
actuality from which Tim comes from (the reality from which he comes from) or
to that that he reaches after his travel (the reality in which he arrives). Of
course, if he had reached the past in the first sense, he could not have managed
to kill his grandfather, and thus he has reached the past only in the second
sense. But then in so far as in a space-time-actuality manifold is also possible to
reach the past in the first sense, should not we regard as genuine time travel
only the first kind of travel?
A supporter of the space-time-actuality manifold here could simply answer
that in so far as Tim has also travelled backwards along the time dimension, and
not only sideways through actuality, this is as good as time travel as we can
demand. Actually, she can even insist that the objection rests on a confusion
with respect to the core problem here. Forget about the grandfather paradox
and the idea of time travel. If the whole point with CTCs was that they seem to
lead to a violation of the Autonomy Principle (and to underdetermined cases of
causation), then, regardless of how we label the strategy, it does work. By
moving in a space-time-actuality manifold, Tim would not be constrained by
conspiracies to fail in carrying out his project to kill his own grandfather. This
response would be, I maintain, a good piece of reasoning. However, there is a
more substantial philosophical problem with the space-time-actuality
approach, which cannot be so easily dismissed.
What is right about the feeling that a movement in space-time-actuality is
not a genuine case of time travel, even if it follows a CTC with respect to the
temporal dimension
9
, is that it is not clear what the relation between the

8
See, for instance, Abruzzese 2001.
9
In a space-time-actuality a CTC is a line whose projection on the space-time hypersurface is

individuals that we find along the dimension of actuality is. Lockwood seems to
defend the idea that individuals located in different region of actuality can be
strictly identical to each other. At one point he says:
[t]he idea is that, just as you can be in different states at different times (relative
to your current motion), so also you can be in different states at the same time
at different points in actuality. (p. 316)
Speaking of you being at the same time in different point of actuality suggests
that individuals can move through actuality as they move in space. Indeed, the
strict identity with respect to the participants of the various superposed states
could be seen as another point of distinction between the many minds
interpretation of quantum mechanics and the many worlds one.
10

However, the strict identity thesis between individuals located in different
region of actuality is problematic. The problems bear similarities to that of
identification of individuals through possible worlds in modal logic. Although
no discussion of such topic is to be found in Lockwoods book, the
consequences of not accepting a strict identity thesis for the solution to the
problem of the apparent violation of the Autonomy Principle that the many
minds interpretation of quantum mechanics purports to give can be seen even
without entering a sophisticated discussion. Consider the paradoxical train
case, and assume that there is not strict identity between the train that enters
the tunnel in the region of actuality in which the computer detects no previous
train entering the tunnel and the train that emerges from the tunnel in the
region of actuality in which the computer detects it. In the reality in which a
train arrives from a CTC that partly lies in other region of actuality, the
Autonomy Principle is not violated. However, in the reality in which the train,
following the command of the computer governing it, enters the tunnel and
disappears the Autonomy Principle does seem to be violated. That the train
literally goes nowhere is as good a constraint on the behaviour of the system
programmed to behave in a certain way and positioned near a CTC as an

closed. Therefore, it may not be topologically closed.
10
Lockwood does not stick to a strict identity vocabulary across the board, and often speaks of
copies of oneself as dwelling different zones of actuality (e.g., p. 325). Indeed, he seems to
acknowledge that a refusal of the strict identity thesis has psychological plausibility: But whereas
memory gives you access to your own states at other times, there is no counterpart of memory that
gives you access to your own states at other location in actuality: states that you can think of as
belonging to your alter egos (p. 316).

asteroid arriving from a far region of space hitting the computer and destroying
it for good.
A possible solution of this predicament could be endorsing a rather non
standard view on the individuation of individuals, modelled after the view that
individuals persist in time by having temporal parts that are located at different
points in time. Individuals moving through space-time-actuality may be
identified with entities possessing not just spatial and temporal parts, but also
modal parts.
11
There would be, thus, strict identity of an individual that is
spread through different regions of actuality, while its modal parts would not
be identical with each other. Whether Lockwood idea of a space-time-actuality
manifold can indeed be put to use to solve such problems as the Autonomy
Principle failure, then, can be judged only on the background of a more
detailed analysis of identity within such a manifold.
In starting this note I have claimed that Lockwoods book is a almost
complete guide of what it takes to understand the most distinctive aspects of
time that our scientific theories allow us to discover. I have finished it by
pointing out something that Lockwood seems not to have taken fully into
account. It may seem that I have changed my mind along the way, but this
impression should be dismissed. Because Lockwoods book provides us with a
deep, interesting and fully worked out insight in the different aspects of time
that emerges from contemporary physics, it is also apt to prompt new
discussions on the pure philosophical side. And this is a characteristic that
good books, irrespective of whether they deserve the all but complete label
or not, have.

REFERENCES
Abruzzese, J. (2001). On Using the Multiverse to Avoid the Paradoxes of Time
Travel. Analysis, 61(1), 36-38.
Arntzenius, F., & Maudlin, T. (2002). Time Travel and Modern Physics. In C.
Cambridge: Cambridge University Press. An updated version (2009) is
in Stanford Encyclopedia of Philosophy.
<http://plato.stanford.edu/entries/time-travel-phys/>

11
Varzi 2001 fiddles with the idea in a more traditional possible worlds context.

Bourne, C. (2006). A Future for Presentism. Oxford: Oxford University Press.
Calosi, C. (2009). Dust and Time: On Relativity Theory and the Reality of
Time. Humana.Mente, 8.
<www.humanamente.eu/PDF/Paper_Dust%20and%20Time%20_iss
ue%208.pdf>
Deutsch, D. (1996). Comment on Lockwood. The British Journal for the
Dorato, M. (2002). On Becoming, Cosmic Time and Rotating Universes. In
C. Callender (Ed.), Time, Reality and Experience (pp. 253-276).
Earman, J. (1972). Implications of Causal Propagation Outside the Null Cone.
Australasian Journal of Philosophy, 50(3), 222-237.
Earman, J., & Wuthrich, C. (2004). Time Machines. In Stanford Encyclopedia
of Philosophy. <http://plato.stanford.edu/entries/time-machine/>
Echeverria, F., Klinkhammer, G., & Thorne, K. (1991). Billiard Ball in
Wormhole Spacetimes with Closed Timelike Curves: Classical Theory.
Physical Review D, 44(4), 1077-1099.
Feynman, R., & Wheeler, J. (1949). Classical Electrodynamics in Terms of
Direct Interparticle Action. Reviews of Modern Physics, 21(3), 425-
434.
Harrison, J. (1971). Dr. Who and the Philosophers or Time Travel for
Beginners. Aristotelian Society Supplementary, 45(2), 1-24.
Horwich, P. (1975). On Some Alleged Paradoxes of Time Travel. Journal of
Philosophy, 72(14), 435-436.
Lewis, D. (1976). The Paradoxes of Time Travel. American Philosophical
Quarterly, 13(2), 145-152.
Lucas, J. R. (1999). A Century of Time. In J. N. Butterfield (Ed.), The
Arguments of Time (pp. 1-20). Oxford: Oxford University Press.
Sider, T. (1997). A New Grandfather Paradox?. Philosophy and
Phenomenological Research, 57(1), 139-144.

Varzi, A. (2001). Parts, Counterparts and Modal Occurrents. Travaux de
logique, 14, 151171.
Yourgrau, P. (1999). Gdel Meets Einstein. Time Travel in the Gdel
Universe. Chicago and La Salle, IL: Open Court.

Commentary
General Relativity from A to B
*

Robert Geroch
University of Chicago Press, Chicago, 1978

James Owen Weatherall **
weatherj@uci.edu

Bob Gerochs General Relativity from A to B is, as its title suggests, an
elementary book the first word, rather than the last, on General Relativity
(GR). This, I take it, is what makes the richness of the book so remarkable. As
Geroch says in his preface, this book is not a view from below [] of a tower
shrouded in mystery. Instead, it shows GR as it actually works (p. VII).
Geroch reveals the nuts and bolts of the theory without getting bogged down
in, or even introducing, the often complicated formalism of differential
geometry. The book is a proof of concept: Geroch ably demonstrates that a
detailed, precise, and yet fully accessible introduction to an advanced topic in
physics is possible after all. One often hears authors and physicists, especially
in the popular press, note the beauty or elegance of GR, and there is much in
the structure of the theory to support such judgments. But what Geroch
reminds us is that GR is also a simple theory. An advanced high school student
could walk away from this book fully equipped to make predictions about the
most exotic space-times one might think of.
General Relativity from A to B is not the kind of book that one responds to,
in the sense of argue with. There is little to agree with or disagree with, here. It
is also a recent enough book that little is to be gleaned from studying its
historical context. Instead, my focus in this commentary will be on the feature
of Gerochs discussion that permits him to say so very much, so precisely, but
with essentially no technical formalism: the space-time diagram. There are
three modest remarks that I want to make about Gerochs use of space-time

*
Thank you to Jeff Barrett, Bob Geroch, David Malament, and John Manchak for comments on a
previous draft. Thank you especially to David Malament for suggesting the connection to J. S. Bells
essay suggested in footnote 6 here.
**
Department of Logic and Philosophy of Science University of California

diagrams, which I will approach in turn. I do not take any of these remarks to
be shocking or groundbreaking. I simply want to draw attention to how
powerful these diagrams can be, both as computational tools and as a way of
probing at least some foundational issues in gravitational theories.
The first remark is methodological, concerning how one should approach
problems in GR. In solving such problems, Geroch writes:
The pattern in every case is the same. One first elicits a detailed statement of
the actual physical experiment to be performed, complete with the
measurements to be taken. One then represents the experiment by a space-
time diagram. (pp. 155-156)
One then goes on to solve the problem, using the space-time diagram thus
constructed. It is easy to take this suggestion in some limited way, perhaps as
follows: when first approaching elementary problems, one should construct a
space-time diagram. More difficult problems (naturally) should be approached
using more sophisticated methods. But to take the suggestion in this way
seems to me to be a mistake. I think Geroch has something more general in
mind.
Geroch uses space-time diagrams in the fashion common to elementary
treatments of relativity theory: he uses them to describe simple experiments in
which different observers measure lengths, or durations, or send signals to
each other. But he also uses them, even more effectively, in addressing the most
advanced topic in the book: the physics of black holes. Here he proceeds
exactly as he recommends in the quoted passage. He begins by giving a general
account of the space-time under consideration, by describing
(diagrammatically) the light cone structure of the Schwarzschild solution. He
then describes a number of experiments that one might perform: he imagines
an observer traveling towards the event horizon of the black hole and then
turning back; an observer passing through the event horizon; an observer
sending and receiving signals as she approaches the event horizon; etc.
Robert Geroch General Relativity from A to B 261

Illustration 1: An example of a space-time diagram from the black hole chapter of
General Relativity from A to B. The boundary of the cylinder represents the event-
horizon of the singularity, represented by the line at the center of the cylinder. At
point u, the observer with wordline C passes the event horizon. One can see the
light cones along C lean increasingly towards the singularity so that at u, there is no
future directed timelike curve that leaves the cylinder. Geroch uses such diagrams
to explain the physics of complicated space-times without introducing detailed
formalism.


In each case, rather than proceed via tensor analysis, he simply draws a space-
time diagram, from which (with no hand waving at all), he describes the
predicted results of the experiment in detail. He does not produce detailed
numerical predictions in this chapter, though as he points out he easily could
by including a little more metrical information in the diagrams.
My point, here, is not to suggest that space-time diagrams are in any way
obscure or little-known tools of GR. They arent. Everyone who has ever taken
a course in GR has drawn a space-time diagram. What I am trying to
emphasize, rather, is how broadly useful these diagrams can be, beyond the
elementary and expository purposes to which they are often limited. These
diagrams capture an immense amount of physics, in any physical situation to
which it is appropriate to apply GR. And unlike many other visualization
techniques in physics, such as Feynman diagrams or atomic level diagrams,
space-time diagrams capture the essential geometrical properties of the
physical configuration they represent. In other words, space-time diagrams
actually support the kind of geometrical reasoning that they invite.
The present point is particularly useful in the context of some (old
1
)
foundational questions in GR. One of the most striking features of Gerochs
book, especially for a popular account of GR, is what expressions do not
appear. (This is the second remark alluded to above.) The word paradox does
not appear in the book; paradoxical appears once, on p. 146, in the context
of insisting that there is nothing shocking associated with disagreements
between observers concerning judgments of distance or length. Other
expressions that one might expect to find in such a book, but which do not
appear here, are length contraction and time dilation. Geroch addresses
the determinations of elapsed time, simultaneity, and length made by different
observers in full calculational detail (here he does include numerical
treatments). But he does so entirely in terms of space-time diagrams, from
which perspective it simply does not make sense to talk about things like length
contraction or time dilation as concrete physical phenomena. It is a mistake to
approach GR by thinking of rods stretching or contracting, or of the gear

1
Such foundational problems are old mostly because the space-time perspective that Geroch
advocates (or, perhaps, embodies) has been effective in resolving them. But one should not forget that
many so-called paradoxes of GR greatly vexed physicists and philosophers in the early decades of the
theory. They remain enshrined in most introductory texts on the subject.

wheels of a watch turning more or less slowly.
2

The alternative view that Geroch recommends, the view suggested by
taking space-time diagrams seriously, is one in which different observers make
different determinations of certain quantities or relations, such as length or
simultaneity, by virtue of the structure of space-time and the measurement
devices available to them. Take the well-known twin paradox, wherein an
astronaut leaves earth, travels for some distance at a high speed, and then
returns to earth. When he lands, he is younger than his twin, who remained on
earth the whole time.
3
Beginning with just a space-time diagram, one can easily
determine how many ticks an earth-bound observer would attribute to an
observer on a rocketships clock (say), and vice-versa, without ever mentioning
time dilation or proper time or anything of the sort. The two observers are
represented by two different timelike curves in spacetime and they make
determinations using various instruments. Using a space-time diagram, one
can predict what these determinations will be. Often these observers
determinations will differ. But that is the end of the story: it should be no
surprise, once the details of the measurements are spelled out
diagrammatically, that they yield different results. There is simply no occasion
for paradox to enter in. In fact, once one is accustomed to thinking about such
problems geometrically, it is difficult to reconstruct what the paradox was
supposed to be.
4, 5

2
Emphasizing this way of thinking about the physics of GR is particularly salient, as Gerochs
perspective is not entirely uncontroversial these days. Harvey Brown, in his recent book Physical
Relativity (2005), seems to suggest that the dynamics of (for instance) rods stretching and contracting
are crucial for understanding relativity theory.
3
Geroch does not treat this example by name; I bring it up only to show how on the geometric
way of thinking, the paradox never has time to arise.
4
Indeed, I do not think I have reconstructed the paradox here. I take it the difficulty is supposed
to be something like as follows: from the perspective of the astronaut, earth recedes at a high speed for
some time, and then changes directions and begins to approach again. And so, by some sort of
symmetry principle, one is supposed to reason that the twin on earth ought to be younger. But they
cannot both be younger than the other, and thus the paradox. Treating the problem geometrically, it is
clear that there can be no such symmetry principle. The curves representing the two twins are not
equivalent: they have different lengths.
5
J. S. Bell (1987), in an essay called How to teach special relativity suggests another paradox
of special relativity (often called the Bell Spaceship Paradox in his honor). One considers two
spaceships initially drifting freely without any relative motion. The spaceships are assumed to be
connected by a fragile string. At some mutually agreed upon time (this makes sense, since the
spaceships initially have parallel inertial worldlines), both ships begin to accelerate uniformly and
identically. The question is whether the string breaks. Bell reports an argument with a distinguished

The final remark I want to make concerning space-time diagrams and
Gerochs presentation is this. One often sees space-time diagrams in
introductory treatments of GR, but Geroch begins by developing the space-
time diagram as a way of understanding classical space-times.
6
Treating
classical theories from the perspective of space-time is much less common; as
Geroch puts it, in the Galilean view, space-time is a luxury; in relativity, a
necessity (p. 220). Yet this luxury is worth the indulgence. For one, treating
classical theories, such as Newtonian theory (which has a Galilean space-time
structure), in terms of four-dimensional space-time and space-time diagrams
allows one to directly compare the mathematical structures of Newtonian and
relativistic physics. Geroch develops classical space-times for just this purpose:
he introduces relativistic space-times only after helping the reader to develop
her everyday, classical intuitions in terms of space-time. Throughout the
second part of the book, where he introduces GR, he reminds the reader of
how to understand the differences between relativistic and classical space-
times.
It is possible to go considerably further in developing Newtonian physics in
terms of the geometrical structure of space-time than Geroch does, on account
of the level at which he presents the material. In the early 1920s, in a lecture
series at cole Normale Suprieure, lie Cartan recast Newtonian gravitation
in the language of differential geometry. The resulting theory, now known as
Newton-Cartan theory or geometrized Newtonian gravitation, is strikingly
similar to GR: once again, the geometrical structure of spacetime depends on

experimental physicist (Bell 1987, p. 68) at the Swiss accelerator laboratory CERN who believed that
the string would not break. (Bell argued that it would.) To arbitrate, Bell and the experimentalist
informally canvased the CERN theory division and discovered a consensus, at least initially, that the
experimentalist was correct. Once the theorists spent time with the problem, however, they came to
agree with Bell. Bells own interpretation of these events is that many physicists have not recognized
the physical importance of Lorentz-Fitzgerald length contraction. But an alternative interpretation,
indirectly suggested by Gerochs book, is that Bells colleagues at CERN would have done well to
begin with a space-time diagram! Once one draws the appropriate diagram and considers the length of
the string as determined by an observer co-moving with either of the strings ends, it is easy to see that
the string stretches (and thus breaks).
6
Geroch considers two kinds of classical space-times: Aristotelian space-time and Galilean
space-time. An Aristotelian space-time, for Geroch, is one in which space-time has an absolute, fixed
standard of rest and in which space has a fixed origin; a Galilean space-time is one in which all
observers agree on determinations of simultaneity, but where there is no fixed standard of rest and
thus no origin. For a more technical treatment of these and a variety of other classical space-times, see
Earman (1989, ch. 2).

the distribution of mass within spacetime; conversely, gravitational effects are
seen to be manifestations of the resulting geometry. It is possible to show that
there is a rigorous sense in which Newton-Cartan theory is a limiting case of
GR (where the limit consists in allowing the lightcone at every point to expand
maximally). With the full Newton-Cartan theory in hand, one can extend
Gerochs comparative project and say precisely, in a wide variety of cases, how
GR and classical physics relate to one another.
7

Even without the fully geometrized gravitational structure, however, there
is much to recommend looking at classical physics from the space-time
perspective, especially to philosophers. Howard Stein (1967), for instance, has
brought a geometrical, space-time understanding of classical physics to bear
on historical questions concerning Newton, Leibniz, and Huygens
interpretations of Newtons theory. Earman (1989), meanwhile, surveys
historical debates on absolute and relational theories of space from a firmly
space-time perspective. Stein and Earmans work shows, I think, that the
space-time perspective is as helpful in understanding and even resolving
foundational problems in classical physics as it is in relativistic physics.
General Relativity from A to B is not intended as a philosophical work; nor
is it meant as a text for specialists. And yet it serves as a potent reminder to both
the physicist and the philosopher of physics of the power of a certain way of
thinking about GRand even classical physicsat both the calculational and
foundational levels.

REFERENCES
Bell, J. S. (1987). How to Teach Special Relativity. In Speakable and
Unspeakable in Quantum Mechanics. Cambridge: Cambridge
University Press.
Brown, H. (2005). Physical Relativity. New York: Oxford University Press.
Earman, J. (1989). World Enough and Space-Time. Cambridge, MA: MIT
Press.

7
For more on Newton-Cartan Theory, the most systematic exposition available is Malament
2010. See also references therein for a sense of how the theory developed after Cartans original
presentation.

Geroch, R. (1978). General Relativity from A to B. Chicago: University of
Chicago Press.
Malament, D. (2010). Relativistic and Newtonian Spacetime Structure.
Unpublished lecture, notes available at <www.lps.uci.edu/malament/>
Stein, H. (1967). Newtonian Space-Time. The Texas Quarterly, 10(3), 174-
200. Also in R. Palter (Ed.) (1970), The Annus Mirabilis of Sir Isaac
Newton (pp. 258-284). Cambridge, MA: MIT Press.

Interview
Adolf Grnbaum
Florence, 18 November 2009

Edited by Duccio Manetti and Silvano Zipoli Caiani

LIFE

1. Your life is entirely characterized by a continuous interaction between
philosophy and science. Why did you develop such twofold interest? Was it
natural for you to reconcile these two different approaches to knowledge?
I grew up in Cologne in Germany; I was there till 1938, when the Nazis came
to town. And then my family and I, I was just a boy, left for the United States. I
picked up a book in a German series which was called Culture of the Present
Time (I am translating, of course) and it was on philosophy, a general book on
philosophy. It started to tell about Aristotle and his idea of four different kinds
of causes and so on. And this interested me very much. So I very soon realized
that you cannot deal with these questions just by being intelligent; you have to
have scientific knowledge, so I decided that I would study both science and
philosophy, because you couldnt do the philosophical work that I was
interested in doing without it.
In the United States, as you know, you study first for your Bachelors
Degree and then sometimes for a Masters degree and then for a PhD. And I
did exactly that. I majored as an undergraduate in both mathematics and
philosophy, because I knew that mathematics was important for physics, and I
knew that physics was important to understand the structure of the universe.
So I took the Bachelors Degree (Wesleyan University, Middletown,
Connecticut) in mathematics and philosophy, the Masters Degree (Yale
University) in physics, and then en route I took the PhD (Yale University) in
philosophy of science. I wrote a PhD thesis on what are known as Zenos
paradoxes. It was published as a book called Modern Science and Zenos
Paradoxes. When I was a graduate student I concentrated of course on
philosophy of science, and my PhD supervisor was Carl Gustav Hempel, who

was a wonderful and eminent teacher. And this was in brief my educational
career.

2. You never speak of some conflicts between a humanistic approach that is
typical of philosophy and a more rigorous character that is typical of
science
Well, I didnt feel this difficulty because my sympathies were entirely with the
scientific approach. I mean I have nothing against what might be called a
humanistic approach, provided that good arguments are given for it. And not
just what somebody would like to believe.

3. Your work started in the wake of generalized excitement for astonishing
scientific discoveries in physics, such as Einsteins theories and quantum
mechanics. Certainly, the twentieth century will be remembered for the
close contact and interconnection between science and philosophy. Do you
think that philosophy still has the same possibility, as it had in the past
during scientific revolutions, to give its contribution to the advancement of
the scientific knowledge?
Yes I think the job of philosophy in part is to provide careful logical analysis of
issues. And that task never stops, because as scientific knowledge develops,
new issues arise. So I think the task of philosophy remains to be a kind of
logical watchdog, offering critical scrutiny, and that is an ongoing task.

PITTSBURGH

4. You are the main founder of the Center for Philosophy of Science at the
University of Pittsburgh, one of the most important academic research
institutions where to study and do research in the field of philosophy of
science. Could you tell us how you came up with the idea of creating such a
place?
I had a very good inspiration. When I was just a beginner in the academic
world, there was the first center for philosophy of science in the United States,
founded by Herbert Feigl, who was a Viennese originally. He was a wonderful
Interview Adolf Grnbaum 269

human being in addition to being a good organizer. He invited me as a
beginner to come to the center, which was at the University of Minnesota in
Minneapolis, and to participate in discussions, workshops and so on. And so I
got the idea that I would like to run such a center, not in competition to Feigls,
but as a kind of extension. So when I was invited by the University of Pittsburgh
(Pitt) to create, from scratch, a philosophical enterprise, I told the chancellor
and the vice-chancellor involved that I would very much like to create a center
for philosophy of science. And they said: Thats wonderful, we would be happy
for you to do it. And they supported me and they said that if I acted as a magnet
to attract other people, they would be very happy to appoint them. And I did
that, because all kinds of famous people then decided to come to Pittsburgh. It
was very exciting and I loved it. I started the Center in 1960. But I also
energized the Department of Philosophy at Pitt by recruiting people in several
philosophic specialties.

5. What was the philosophical climate before the creation of the Pittsburgh
Center for Philosophy of Science in the United States?
It varied. In some places philosophy of science was already appreciated if not
cultivated, but at least appreciated. In other places it was almost a cemetery,
and so I decided it was a good place to begin. Then of course we started to
attract people. And people wanted to become what we call visiting fellows of
our Center. We have a program to bring people from all over the world
(scholars in philosophy, philosophy of science, psychology, cognitive science)
who come and spend a term or a year working with us and connecting up with
other fields at the university, such as physics. We have a very strong group in
relativity theory in physics, which is one of my main interests in the philosophy
of space and time. We have published a series of volumes called the University
of Pittsburgh Series in the Philosophy of Science.

6. Do you think that there is a genuine interest of scientists for philosophy?
That varies greatly. For people who have had a European exposure generally
the answer is yes. Others are either indifferent or hostile. And the reasons why
some are hostile are understandable: they have encountered philosophers who
knew nothing about science but told everybody what made the world go round.

And they didnt like that. And so these are people who will understandably stay
away from philosophers.

7. In your opinion, what does Pittsburgh represent today, what is its role in
the contemporary philosophical world?
Well, there is more in philosophy than philosophy of science. After I had been
at Pitt for a decade, by 1971, I created a special department, the department of
history and philosophy of science, because the history of science is also very
important for philosophy of science. But there is a general philosophy
department (for people who are interested in ethics and other fields of
philosophy), and they are, of course, perfectly respectable. They are not part of
our center, but we collaborate, especially in the training of students. We have
lectures and visitors of all sorts, so its a very active give-and-take between
different areas of philosophy, philosophy of science and the general university
environment. Its very good.

PHILOSOPHY OF PHYSICS

8. You said that its important for philosophy to have a connection with the
history of science. Do you think that this could be useful for science as
well? For physics, for example.
Very important. For this reason: we know, as has been said, that the history of
science is the history of abandoned theories. We know superseded theories,
such as Aristotles biology, the phlogiston theory of chemical combustion, and
so on.
These developments were very important because you learn from the history of
science that theories are replaced by other theories, we hope by better ones,
but they are replaced. And some of the ideas that were leading in physics, for
example early in the nineteenth century, are not known to most physicists now:
they learn contemporary physics. But students of the history of science and of
the history of philosophy of science of course study these developments,
because it shows them different conceptualizations of similar observational
material.


9. First of all, your work was related to questions of philosophy of physics.
What was the role of Poincars conventionalism in developing your
conception concerning the epistemological status of geometrical
conventions in physics?
It was very important. Ill say this about Poincar: he was a great mathematician
and he had a very clear mind, but when he wrote on philosophical topics, he
sometimes wrote in a very undisciplined manner. For example he wrote a paper
on conventionalism, which I read many times over; it was so awfully unclear. It
made me really unhappy that this brilliant mathematician was writing like this. I
must refer you to my careful clarifying account of Poincars geometric
conventionalism in my 1973 treatise Philosophical Problems of Space and
Time, 2
nd
edition, chapters 1C and 4B. But of course I read his Science and
Method and the three books of his The Foundations of Science. Then he
introduced a lot of problems in physics and mathematics, so in that sense he
was a pioneer also in the philosophy of science. But as I said, you have to be
very careful because he is often not very clear.

10. Certainly, within the neo-positivist approach to philosophy of physics, it
was Reichenbachs work about the nature of space and time that
emphasized the empirical character of the metric.
No, he did not emphasize the empirical character of the metric; but the reason
for that was complicated because he thought that congruence is a matter of
stipulation. Then he said that, once you have defined a standard for
congruence, for equality of length and for time intervals, it becomes an
empirical question what the actual geometry is. So it was a sophisticated
notion, and I was tremendously impressed by Reichenbach. Very early in my
studies, when I first studied physics and philosophy, I read his 1928 German
book Philosophie der Raum-Zeit-Lehre, and later (1957) Maria Reichenbach
translated it into English as Philosophy of Space and Time. Reichenbach was
for me a lodestar, a guiding inspiration.
With the rebuttal of the Synthetic a Priori (typical of the neo-positivists),
Reichenbach assumed that the choice concerning a spatial metric is
specified by a coordinative definition of congruence.

Yes, that was another way of saying that the statement that such and such will
be the standard is a stipulation, its not discovered but its conventional.
And he upheld this thesis also in the context of his distinction between
universal and differential forces.
Yet with respect to the so-called universal and differential forces, Reichenbach
was not very clear. But first lets say this: gravitation is a universal force in the
sense that it affects all things alike. But, differential forces are forces that affect
different materials to a different degree. For example heat would expand a
piece of metal more than a piece of wood, or than a stone. So these are so far
the differential forces, but Reichenbach unfortunately also used the notion of
universal force as a metaphor, as a pictorial device for discussing different
standards of congruence. For example if you take a rod, thats one criterion of
congruence. But if you say that when the rod is transported, it does not remain
self-congruent, Reichenbach describes that by using the term universal forces
metaphorically to say that all sorts of solid rods are being deformed alike under
transport.
But thats just metaphorical talk, and so I suggest to get rid of it, and to talk
about universal forces only when we mean the literal sense like gravitation.
Now, gravitation is not only a universal force, but also a differential force. For
example if you have a bookshelf made of wood and the books are very heavy,
youll see that it will bend. But if you have a bookshelf made of steel, its almost
not going to bend. And thats all because of gravity; so gravity also is a
differential force and not only a universal force.

11. Your position is, in some respects, coherent with that of Reichenbach.
What are the affinities and the differences between you and the neo-
positivist tradition?
I think that logical empiricism, and neo-positivism sometimes, had a good
influence in the sense that it called for clarity in discussing issues in
philosophy, and from that point of view it was very good. But in other ways I
think it was not very good. There is what I call phenomenalist positivism, which
essentially says Ernst Mach said it that the world is just a bunch of sense-
data. But the world is the world and we, because the world affects our bodies,
have sense-data, but the world doesnt consist of sense-data. So there has been

a lot of confusion about this, and that is so-called phenomenalist positivism,
which I never accepted.

12. In some pages of your book Geometry and Chronometry in Philosophical
Perspective (e.g., p. 250) you try to explain the differences between your
sort of geometric conventionalism and what you called trivial semantic
conventionalism. What exactly are the differences between your
conception of a coordinative definition of congruence and a mere
semantic stance? If this is not a position concerning only semantic values,
what motivates the choice of one definition of metric instead of another?
Everybody knows that words, which are noises, can be used to refer to things.
And thats true for words in English, in German, and so on. Its trivial because
everybody knows it. But in theory of measurement, the question arose (and was
asked already by Ernst Mach, the Austrian physicist): what evidence would you
give that, if you applied the meter stick here and if you then applied it there, it
remained congruent to itself under transport? And that is not a matter of fact,
but this is a matter of convention, though not in a very obvious or trivial sense.
This requires a little reflection.
Do you mean that a coordinative definition of congruence when we talk
about a stick involves a reflection about physical theory, so self-
congruence under transport is not a trivial semantic definition? If this is
not a position concerning only semantic values, what motivates the choice
of one definition of metric instead of another?
The famous German mathematician Bernhard Riemann, when he became a
professor, wrote his Habilitation Lecture On the Hypotheses which Lie at the
Foundation of Geometry. There he discussed the question of congruence
under transport, and here is what he said: The definition that the meter stick
remains congruent to itself depends for its consistency on a physical fact that is
not a matter of definition.
And that fact is the following: If you have two coinciding meter sticks and
you bring them to another place via different routes, they will still remain equal
to each other. If they didnt do that, it would become inconsistent to use them,
depending upon which one you use. It would be chaos. We would never have
developed the theory of solid bodies the way we did.

So that was Riemanns point, and it was a recognition that a convention is
made possible by certain facts of the behavior of bodies. It was very insightful,
and I discussed it at the beginning of my treatise Philosophical Problems of
Space and Time.

13. Besides philosophy of space and time, your interest has also been
concerned with broad epistemological questions, such as the Duhem-
Quine thesis about the limits of confirmation and discreditation of
individual scientific hypotheses. Contrasting Duhem and Quine, in one of
your famous articles, The Falsifiability of Theories: Total or Partial? A
Contemporary Evaluation of the Duhem-Quine Thesis, you claim that it
is possible to realize genuine crucial experiments, that is, to produce a
direct falsification also of individual hypotheses. This position seems to
support a kind of falsificationist epistemology as proposed by Popper.
What distinguishes your view from that of the Austrian philosopher?
He just talked about it in general terms. He didnt discuss these concrete
questions, and when he did, it was very unclear. Lets start with the very simple
example of modus tollens: If A then B, not B therefore not A. But Duhem
looked at the confirmation and at the refutation of scientific theories. He noted
that when you take a hypothesis and you want to see whether its well
supported by the evidence or refuted by the evidence, you do not test only this
hypothesis in isolation from everything else, but this hypothesis in conjunction
with other auxiliary hypotheses. The combination of them gives you a
prediction. And then you check that prediction, whether its correct or not, and
if it isnt, then it is not clear where the trouble is. And that makes it difficult;
therefore Duhem said that there are no crucial falsifying experiments and also
no crucial verifying experiments.
Well, what I did was look at some examples in physical geometry and I
found an example but that didnt prove anything in general in which you
could isolate a hypothesis epistemologically.
So thats when I got involved with Quine: I tried to make use of the idea of
what he called the holistic character of knowledge in his general philosophy.
But I wrote to him (and I published the letter), and he acknowledged that he
didnt mean to say something as strong as he had been taken to claim.

Well, do you think that a scientist has a rational criterion to define what is
the hypothesis that is responsible of the failure of a theory? Or is it
something like an intuition?
Much more the latter, because they focus psychologically on particular
hypotheses. Those are the ones that they worry about, the others they take
along as baggage. And thats how they work, but thats ok. Duhem was the one
who said: just wait a minute, think about this more carefully. And he certainly
did.

PHILOSOPHY OF PSYCHOANALYSIS

14. The reference to Popper gives us the opportunity to introduce another
aspect of your work. Popper, like you, has dedicated part of his efforts to
criticize the methodology adopted in the field of psychoanalysis. Even if
both you and Popper are not admirers of the psychoanalytical framework,
your opinions diverge. Can you clarify in what consists the difference
between your critique to psychoanalysis and Poppers?
I learned a great deal from thinking about what Popper had said. But I dont
think he was a very good philosopher of science. A lot of what he wrote was
very poor, and he is greatly overrated I believe. First, he knew almost nothing
substantive about Freuds psychoanalysis. But he wrote about it. And he grew
up in Vienna. After all, thats where Freud was working. Poppers work on
psychoanalysis was extremely uninformed. For example his statement that its
an unfalsifiable theory. Well, I have shown easily that its falsifiable by
reference to Freuds etiology of paranoia as set forth in his paper on a case of
paranoia, which runs counter to the psychoanalytic theory of the disease. There
Freud discusses what sort of evidence would have shown that his theory of
paranoia was incorrect. And Freud had published this in 1915, two years
before Popper said that psychoanalysis is unfalsifiable, which is wrong.
Secondly, its wrong to say that falsifiability is the criterion for the rational
acceptability of scientific theory. Its certainly relevant to it, but to say that
falsifiability is scientificity, thats sloppy and wrong. I wrote an essay in which I
show in detail all the mistakes Popper made re psychoanalysis, but he got me
interested in psychoanalytic theory, so I owe him that I started to study it,
thinking that there must be something wrong with what he had said.

And then the psychoanalysts started to get very nervous about what I was
doing because they saw that I was raising serious questions, and some of them
were very encouraging, others thought that I was a dangerous man.

15. Do you think it is possible and desirable to find, for Freuds theory, more
extra-clinical confirmations that are different from those obtained
through the normal psychoanalytical setting?
Yes I do. You see, the trouble is, psychoanalysis began as a clinical theory
based on working with patients. Thats totally understandable. But the problem
was that Freud and his followers then decided that its from the clinical
situation between a therapist and a patient that all the evidential material comes
from, and has to come from there. So the first thing they did not consider was
running experiments which are much better controlled than to sit there and
talk to their patients. Therefore the epistemology of the theory is much more
complicated than they recognized.
And also Freud himself recognized that in looking at the clinical responses
of patients, you have to realize what the power of suggestion is. If I am your
therapist and you are my patient, and I lead you to say that certain things bother
you and other things mean a great deal to you emotionally and so on, I am very
often leading you by suggestion to produce compliant responses, responses in
which you will do what I expect, trying to please your therapist, because you
depend on me for emotional health and support. And that could be very
misleading.
This is what happens of course, I believe, in religious counseling when
people go to priests or rabbis or ministers or whatever: They are their
telephone wire to god. So I said to the psychoanalysts: How do you know that
what your patient is producing is a real response or just an attempt to please
you? And you know, Freud had a very important colleague, Wilhelm Fliess,
who made him very angry because he asked Freud: How do you know that
youre reading your patients mind instead of reading your own mind?

16. More than a rebuttal of psychoanalytic theory, your work is also a critique
of the hermeneutical interpretation given to it by authors such as
Habermas, Ricoeur and Klein. Do you think it is possible that

psychoanalysis regains a naturalistic stance as hoped by Freud in an early
version of his thought?
Freud said: psychoanalysis is a natural science, and asked what else could it be?
Yes, he thought that, certainly. And he thought that the method of free
association can be used to discover the causes of peoples behavior. But, in my
writings, I have offered a careful critique of the claim that free association is a
causally probative method of investigation.
So do you think that psychoanalysis can work in the manner of natural
sciences? How?
Yes, but I think it has to be done carefully, and people have to use good criteria
of evidence, as I have tried to show in my writings.

17. Do you think that the development of a scientific theory of mind and of
the unconscious is possible, without however subscribing to a physicalist
reductionism?
Yes. I think its a prejudice that if a claim is scientific, meaning that its based
on evidence and good reasoning, then it has to be reductive physically or
biologically. Thats just a sloppy mistake. Not primarily a mistake by those who
believe in the scientific approach to cognition, but in the minds of those who
are opposed to it.

18. Do you think that today it is still reasonable to work at defining a sharp
distinction between what has to be conceived as pure science and what
doesnt?
No I dont, and Ill tell you why. Because such a sharp separation would provide
a criterion of demarcation, but there is no general criterion of demarcation.
There is a very valuable paper by my former colleague Larry Laudan called
The Demise of the Demarcation Problem (1983, Physics, Philosophy and
Psychoanalysis; Essays in Honor of Adolf Grnbaum, R.S. Cohen and L.
Laudan Eds.). There he discusses the different attempts to provide a general
formula of demarcation. I dont think there is a general formula.


THEISM

19. Over the last twenty years your works have also been dedicated to an
epistemological analysis of the theistic stance. In one of your most
relevant articles about this topic you consider Leibnizs critical question
Why is there something contingent at all, rather than nothing contingent?
with the aim of showing that no theistic stance is able to provide an
epistemologically sound answer.
But why should there be just nothing? That has to be shown, and they havent
shown that. So they assume something just to get the question going and then
they demand an answer to it. But that procedure begged the question.

20. What kind of reader did you have in mind when you started writing about
it?
Well, for one thing I believe that all claims to knowledge have to be examined,
and that includes of course stances about the whole world and what caused it
and so on. Some people think scientific skeptics like me believe that they know
it all. Thats not so. Its the religionists who claim that they know how the world
came into being and what makes it go around. They are the ones who think they
know it all. And then I ask them: show me. I think the arguments for the
existence of god are very poor.
Do you think that your analytic argumentation could somehow matter to
believers?
Yes, i.e., to those who dont think that its just a matter of faith, but who think
that there are arguments in so-called natural theology. They believe that
arguments can be given. They and I agree that the matter has to be discussed in
terms of arguments. Where we part company is when they think the theistic
argument is good but I think the argument is very bad.
What feedbacks have you received from them?
All kinds. There is Richard Swinburne with whom I debated publicly, and there
is Craig who has written about this. I have replied polemically to both of them.


21. Do you consider your work, in some respects, a continuation of the neo-
positivist battle against metaphysics?
Well, I would say that its too general. I dont think there is a sharp divide
between metaphysics and philosophy of science at all. I think there is
continuity. I think that the great merit of the logical empiricist movement was
to raise carefully questions such as: What do you mean by this?, How do you
know its true?, Why should we think its true?. These are all constructive
questions, and I think any good epistemologist would welcome this. Kant was
obviously interested in epistemology and he would have thought of the
positivists as continuing his tradition. So I think to say that metaphysics is bad
is a mistake. Thats just name calling and it seems to me to be unhelpful. The
thing to do is to ask What do you mean? and How do you know?. And those are
good questions. And if positivists ask them, I say good for them! And if
religious people ask them, I say good for them!

22. In which way do you think science can influence a religious way of
thought? We know that a direct conflict between science and religion is
not the best the way to deal with the situation. So maybe science needs a
different strategy, maybe it should work, like religion, from inside the
society?
I think that we should, not only in scientific pursuits, but also in discussing
religious beliefs, do our best to ask ourselves, What is the evidence?, What do
you mean?. And if its clear what they mean, then: How do you know? These
are the questions that I think should always be asked.


HUMANA.MENTE - ISSUE 13, APRIL 2010
Editor: Claudio Calosi
It is a widely recognized fact that metaphysics and physics have been
rather self isolated enterprises, even in the analytical community.
On one hand, metaphysical issues about identity, location, persistence
through time, material composition, causation and so on have rarely been
discussed within the framework of physical theories.
On the other hand, physics and philosophers of physics have been skeptic
about whether metaphysical issues are really capable of posing genuine
problems and about the possibility for metaphysics to provide a consistent
and valuable view of how the world is.
In recent years however there has been a tendency to bridge the gap
between physics and metaphysics. Considerations drawn from physical
theories have played a major role in metaphysical disputes like the
ontology of time, nature of persistence, theory of identity and even
mereology, to name just a few.
This raises interesting general questions about the relationship between
physics and metaphysics and more particular questions about the
possibility of importing physical considerations to solve specic
metaphysical problems.
For more information about the journal
visit our website at:
www.humanamente.eu
HUMANA.MENTE
Journal of Philosophical Studies
ISSN: 1972 -1293
www.humanamente.eu
Humana.Mente Journal of Philosophical Studies was
founded in Florence in 2007. It is a peer-reviewed international journal
that publishes 4 issues a year. Each issue focuses on a specic theme,
selected from among critical topics in the contemporary philosophical
debate, and is edited by a specialist on the subject, usually an emerging
researcher with both philosophical and scientic competence.
Humana.Mente wants to be a place for exploring the most recent
trends in the international philosophical discussion and wants to give the
opportunity to the international community of young researchers to
confront each other, and to discuss, control and verify their theories. An
analytic perspective is favored, and particular attention is given to the
relationship between philosophy and science, without however
neglecting the historical aspects of the philosophical topics.
In this issue, papers and works by:
Mauro Dorato, University of Roma 3 - Gabriele Veneziano, Collge de
France, Paris - John Norton, University of Pittsburgh - Sam Baron, Peter
Evans, Kristie Miller, University of Sydney - Vincent Lam, University of
Queensland - Claudio Garola, Sandro Sozzo, University of Salento -
Giovanni Macchia, Urbino University -- Adriano Angelucci, Verona
University - Vincenzo Fano, Urbino University - Tracy Lupher, James
Madison University - Friedel Weinert, University of Bradford - Daniel
Nolan, University of Nottingham - Jeffrey Barrett, University of
California at Irvine - Giuliano Torrengo, University of Torino - James
Weatherall, University of California at Irvine

Humana - Mente 13 Physics and Metaphysics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Humana - Mente 13 Physics and Metaphysics

Uploaded by

Copyright:

Available Formats

H

ISSUE 13 - APRIL 2010

. Of course, such a wavelength is not a universal constant since it

, is a truly universal length scale. We can say that the three

represent, respectively, the three basic

, the only length scale available. Thus quantum strings acquire,

> in which M is the spacetime manifold,

is the following FLRW metric

has two components: the recession velocity

and the peculiar velocity

(), and from = () one has

= (). Thus, whereas in the

= 0, with negative spatial curvature and = for which a

= 0, that is, if the universe were not expanding at the times of

), namely, that some expansion had occurred during the

= 0 does not imply

0.3 (Davis et al.

the cosmological redshift,

is the matrix representing the rotation of

could always change point by point. For this

, that is the metric peculiar to that point.

given the distribution of masses and energy

() and (), and

() is an observable. It has possible specific values, but no

does not by itself specify a field configuration. According to Teller

() and states. Given a state

() for all possible products of

is evaluated at arbitrarily chosen spacetime points,

(), by itself, does not

() for every spacetime point! The problem is that while

() (where ) will have

for a particular state give the mean

at a spacetime point x with a test function f defined on a finite spacetime

is a linear map from test functions to

represents the average field value in the region O.

() can also be arbitrarily large

() behaves like an unbounded operator. An unbounded operator is not

() will also satisfy the CCRs. Are these two fields

() are unitarily equivalent. When two

are unitarily inequivalent,

. Whilst the thermal time flow is state-dependent, the authors also

, which is determined from the initial energy. This parameter,

of the measured quantity, then the state of the composite system

over the course of a measurement interaction; in

, which might represent determinately recording the

then, by the linearity of

then ignores all components of the entangled global

, then we say that the state of R is

relative to the state of S being

above. Since the

), Everett would say that As notebook is in a

. But since the global state of the world will typically be a

. If so, then relative to there in the model that is

. Our quantum expectations for result k covary with

You might also like