Lorentz's theory and special relativity are

completely identical

László E. Szabó

Theoretical Physics Research Group of the Hungarian Academy of Sciences

Department of History and Philosophy of Science
Eötvös University, Budapest
E-mail: leszabo@hps.elte.hu

Abstract
It will be shown that special relativity and Lorentz's theory are completely identical in
both sense as physical theories and as theories of physical space-time. All statements of special
relativity about those features of reality that correspond to the traditional meaning of terms
space and time are identical with the statements of Lorentz's theory. On the other hand,
all statements of Lorentz's theory about those features of reality that are called space and
time by special relativity are identical with the statements of special relativity. The only
dierence between the two theories is terminological.

Key words: Lorentz's theory, special relativity, space-time, operationalism

PACS: 01.70.+w, 03.30.+p

Introduction

There are basically three ways how people think about the relationship between pre-relativistic
theory of physical spacetime and special relativity:

(A) The rst is the naive approach of some physics textbooks, according to which the emergence
of Einstein's special relativity isat least according to the rational reconstruction of the
storyone of the usual discoveries of new facts of nature : Certain experimental ndings
necessitate to draw the conclusion that the geometry of space-time is something dierent
from what we believed before.

(B) According to the second approach, the switch to the relativistic theory of space-time is a
convention, rather than an unambiguous theoretical conclusion drawn from the empirical
facts.
1 As we know, Lorentz's theory is capable to explain the null result of the Michelson
Morley experiment and other experimental ndings through the deformations of moving
material objects, at the same time, however, it remains completely within the framework of
the classical theory of space and time.
There are various views about the epistemic status of such a choice between Lorentz's theory
and Einstein's relativity, as well as there is a variety of (usually erroneous) justications of
why we should prefer relativity. Our concern here is the general logical scheme of the
conventionalist approach: There are two theories of space and time, and there are two
corresponding physical theories. In accordance with Poincare's general thesis that geometry
and physics only together can be compared with the empirical facts, the conventionalist
thesis asserts the following relationship between these theories:

(Classical space-time) + (Lorentz's theory) = (empirical facts)

(Minkowski's space-time) + (relativistic physics) = (empirical facts)

1 Inasmuch as such unambiguous conclusion can be drawn from empirical data at all. The point is, however,
that the known empirical data do not unambiguously imply the denial of the classical theory.

1
 (C) According to the third approach, the relation of Lorentz's theory to Einstein's special rela-
tivity is the relation of a constructive theory to a corresponding principal theory. J. S. Bell
characterizes this relation as follows:

If you are, for example, quite convinced of the second law of thermodynamics,
of the increase of entropy, there are many things that you can get directly from
the second law which are very dicult to get directly from a detailed study of
the kinetic theory of gases, but you have no excuse for not looking at the kinetic
theory of gases to see how the increase of entropy actually comes about. In the
same way, although Einstein's theory of special relativity would lead you to expect
the FitzGerald contraction, you are not excused from seeing how the detailed
dynamics of the system also leads to the FitzGerald contractions. (Bell 1992, p.
34)

It is to be mentioned that all of the above three approaches can nd support in Einstein's own
writings. Approach (A) can nd support in Einstein's famous popular book on special and general
relativity, for example. At the end of Section 13 Einstein regads Fizeau's experiment as an
experimentum crucis in favour of theory of relativity.
2 Then the next section begins with the
following summary:

Experienece has led to the conviction that, on the one hand, the principle of relativity
holds true and that on the other hand the velocity of transmission of light in vacuo has
to be considered equal to a constant c. By uniting these two postulates we obtained
the law of transformation for the rectangular co-ordinates x, y, z and the time t of
the events which constitute the process of nature. In this connection we did not ob-
tained the Galilei transformation, but, diering from classical mechanincs, the Lorentz
transformation. (Einstein 1961, p. 42)

In other writings Einstein seems to agree with the conventionalist approach (B):

Geometry (G) predicates nothing about the relations of real things, but only geometry
together with the purport (P) of physical laws can do so. Using symbols, we may say
that only the sum of (G) + (P) is subject to the control of experience. Thus (G) may
be chosen arbitrarily, and also parts of (P); all these laws are conventions. All that
is necessary to avoid contradictions is to choose the remainder of (P) so that (G) and
the whole of (P) are together in accord with experience. (Einstein 1983, p. 35)

Finally, in some other writings, as H. R. Brown and O. Pooly pointed out in a quite recent paper,
3
Einstein claims that special relativity is a principal theory, and that principle theories lose out to
constructive theories in terms of explanatory power:

The universal principle of the special theory of relativity [the relativity principle] ...
is a restricting principle for natural laws, comparable to the restricting principle of
the non-existence of the perpetum mobile which underlies thermodynamics. (Einstein
1969, p. 57)

... when we say we have succeeded in understanding a group of natural processes, we

invariably mean that a constructive theory has been found which covers the processes
in question. (Einstein 1982, p. 228)

My aim in this paper is to show that, contrary to the above three usual approaches, the truth is
that
2 Einstein 1920, p. 28.
3 Brown and Pooly 2001.

2
Thesis Lorentz's theory and Einstein's special relativity are completely identical, in both sense,
as physical theories and as theories about space and time.

I shall of course also explain why the two theories seem to be inequivalent on the level of
the narratives. As we will see, the only dierence is that the two theories express the same laws
of physics (including the laws of space-time) in dierent variables just as we may use polar
or Cartesian coordinates. The confusion is caused by the unfortunate fact that these dierent
variables have the same names in the two theories, namely space and time coordinates. Thus
the alleged choice between special relativity and Lorentz's theory is merely an inated special
case of a semantical banality holding for any and all linguistic signs or symbols, a banality wich
Grünbaum calls trivial semantical conventionalism.
4
It is to be noted that although my Thesis denitely contradicts to the spirit of the three
standard views, it is not in logical contradiction with them. Actually, the complete identity of
Lorentz's theory and special relativity is the sole case when approaches (A )(C) are logically
compatible with each other.

1 Lorentz's Theory

Lorentz's theory is usually not contained in the physics curricula, hence a brief digression on the
topic may be worthwhile. I shall follow J. S. Bell's reconstruction in his very clear and concise
review paper, How to teach special relativity.
5 Let us x at the beginning a reference frame K
in which the International Bureau of Weights and Measures (BIPM) in Paris, together with the
etalons (the standard measuring rod, the standard clock, etc.) are at rest. The laws of physics,
like the Maxwell equations are understood and valid (empirically tested) in this reference frame.
6
Consider now the electromagnetic eld of a point charge q. It is the familiar spherically
symmetric Coulomb eld when the particle is at rest (in K ). How does this eld changes when we
set the charge in motion? Maxwell's equations can also answer this question. Here is the result:
the electromagnetic eld of a point charge q moving with constant velocity v along the z axis is

− 32
Ez = qz 0 x2 + y 2 + z 02

− 3  2
− 12
Ex = qx x2 + y 2 + z 02 2 1 − vc2

− 3  2
− 12
Ey = qy x2 + y 2 + z 02 2 1 − vc2
Bx = − vc Ey
By = vc Ex
Bz = 0
where

z − zq (t)
z0 = q
2
1 − vc2

and zq (t) is the position of the charge at time t. The lines of electric eld are shown on the Fig. 1:
the original Coulomb eld (v = 0) attened in the direction of motion. So, the electric eld of a
charged particle changes if we set it in motion. What kind of similar deformations do we know,
on the basis of classical pre-relativistic physics?
Calculate, for example, how this deformation of the electromagnetic eld modies the orbit
of an electron in a single (classical) atom moving with constant velocity v along the z axes. To
answer this question, in this deformed eld we have to solve the equation of motion of the electron:
 
 
d  r m0 dr 
 = −eE (r) − e dr , B (r)
dt  ( dr ) dt
2  c dt
1 − dt
c 2

4 Grünbaum 1974, p. 27.

5 In Bell 1987, p. 67. For further details, see Jánossy 1971.
6 At this point my presentation diers from Bell's paper. Bellfollowing Lorentzstarts with a reference frame
at rest relative to the aether. From the later results it turns out however that we do not need to operate with the
aether.

3
 x

Electric field of and of source moving

source at rest in z direction

Figure 1: The lines of electric led are shown on the gure: the original Coulomb eld (v = 0)
attened in the direction of motion

r(t) is the electron position and we take into account

where the empirically known
7 mass formula
m = pm0 ṙ2 . One can solve this equation by computer: if the acceleration of the nucleus is
1− c2
suciently gradual, the initially circular orbit contracts in the direction of motion to a fraction
q
v2
1− c2 (Fig. 2).

q
v2
l0 l0 1− c2

Figure 2: The initially circular orbit of an electron in a single (classical) atom contracts in the
q
v2
direction of motion to a fraction 1− c2

One can calculate the period of the orbit:

T0
T = q
v2
1− c2

where T0 is the period of the system at rest. We obtain a similar result if we calculate how the
contraction modies the period of a light-clock also moving with velocity v along the z axes (see
Fig 3).
From these results we arrive at the conclusion that an arbitrary physical system suers a
deformation if we set it in motion, such that its original extension l0 in the direction of the motion
contracts,
8

r
v2
l0 7→ l0 1− (1)
c2
and all clock-like processes in the system slow down:

∆t0
∆t0 7→ q (2)
2
1 − vc2
7 It is an empirical formula also in relativity theory!
8 It is to be noted that the only experimental test of the Lorentz contraction is based on the deformation of the
electric eld shown in Fig. 1: One can observe that the track of a charged particle in a bubble chamber becomes
wider if the particle moves with high velocity.

4
 It is easy to verify that these deformations can completely explain the null result of the
MichelsonMorley experiment, and any other experimentally observed phenomena normally ex-
plained by relativity theory.

v
2l
At rest: T0 = c p 2
p 2
l 1− vc2 l 1− vc2
In motion: T = c−v + c+v = p T0 v2
1− c2

Figure 3: The calculation of the period of a light-clock moving with velocity v along the z axes

Consider the following new variables:

t− vz
x0 = x y0 = y z 0 = pz−vtv2 t0 = p 2 c
2 (3)
1− c2 1− vc2

One can easily verify that these variables are nothing but the space and time coordinates
9 mea-
sured by a co-moving observer blindly applying co-movingconsequently, deformedmeasuring
stick and clock, and blindly dening simultaneity in the standard way, as if he or she were at rest.
One can extend the set of these primed variables with others denied by the co-moving observer
0
in K simply repeating the operational denitions of the similar physical quantities in K , blindly
0 0 0 0
applying x , y , z , t as if they were the space and time coordinates. For example, the electric
eld strength E in K is denied as the force on a unite charge at rest:

d2 r
m0 = qE (4)
dt2
Therefore we dene E0 as the force on a unite charge at rest in K 0:

d2 r0
m0 = qE0 (5)
dt02
On the other hand, we know that
 
 
d  r m0 dr 
 = qE + q dr , B (6)
dt  ( dr 2 dt
 c dt
dt )
1 − c2

Applying (3), from (4)(6) one nds

Ex − c By
v
Ey − c Bx v
Ex0 = p v2
Ey = p v2
Ez0 = Ez
1− 1−
c2 c2

From a similar derivation we obtain

Bx + c Ey
v
By − c Ex v
Bx0 = p v2
By0 = p v2
Bz0 = Bz
1− 1−
c2 c2

Lorentz has observed that the motion of a physical system modies its behavior in a peculiar
way:

Lorentz principle: The laws describing the behavior of a system in motion can be obtained
in the following way: rst solve the problem relative to the same system at rest , then, in the
solution thus obtained, perform the substitution

x, y, z, t, E, B, etc. 7→ x0 , y 0 , z 0 , t0 , E0 , B0 , etc. (7)

9 The quotation marks are very importantat least from the point of view of the classical conception of space
and time!

5
In other words, the laws describing the behaviour of the moving system, expressed in the primed
variables, has the same form as the laws describing the same system at rest, expressed in the
original variables. Consequently, the Lorentz principle can be reformulated in the following way:

The laws of physics have the same form in every inertial frame, if they are, from frame to frame,
expressed in terms of dierent physical variables dened by measurements performed with the co-
moving measuring apparatuses, that is, with the same standard measuring apparatuses accelerated
from one inertial frame to the other.

Note that this principle is nothing but the principle of Lorentz covariance in relativistic physics.
Thus, until relativistic physics correctly describes the world, Lorentz's theory can do the job, too.
10
It is instructive to see in more detail how the things are described by a moving observer. Let
me quote a longer passage from Bell's above mentioned paper:

The important point to be made about moving observers is this, given Lorentz
invariance: the primed variables, introduced above ... are precisely those which would
naturally be adopted by an observer moving with constant velocity who imagines herself
to be at rest. Moreover, such an observer will nd that the laws of physics in these
terms are precisely those that she learned when at rest (if she was taught correctly).
Such an observer will naturally take for the origin of space coordinates a point at
rest with respect to herself. This accounts for the vt term in the relation

z − vt
z0 = q
2
1 − vc2
q
v2
The factor 1− c2 is account for by the Fitzgerald contraction of her metre sticks. But
will she not see that her metre sticks are contracted when laid out in the z direction
and even decontract when turned in the x direction? No, because the retina of her eye
will also be contracted, so that just the same cells receive the image of the metre stick
as if both stick and observer were at rest. In the same way she will not notice that her
clocks have slowed down, because she will herself be thinking more slowly. Moreover,
imagining herself to be at rest, she will not know that light overtakes her, or comes to
meet her, with dierent relative velocities c ± v. This will mislead her in synchronizing
clocks at dierent places, so that she is led to think that

t − vz2
t0 = q c
2
1 − vc2
is the real time, for with this choice light again seems to go with velocity c in all
directions. This can be checked directly, and is also a consequence of the prime Maxwell
equations. In measuring electric eld she will use a test charge at rest with respect to
her equipment, and so measure actually a combination of E and B. Dening both E
and B by requiring what looks like the familiar eects on moving charged particles,
0 0
she will led rather to E and B . Then she will be able to verify that all the laws of
physics are as she remembers, at the same time conrming her own good sense in the
denitions and procedures that she has adopted. If something does not come out right,
she will nd that her apparatus is in error (perhaps damaged during acceleration) and
repair it.
Our moving observer O0 , imagining herself to be at rest, will imagine that it is the
stationary observer O who moves. And it is as easy to express his variables in terms
of hers as vice versa
 
x0 = x  x = x0

 

  y0
y0 = y 
  y =

z 0 +vt0
z0 = pz−vt 2 ⇔ z = p 2 (8)
1− vc2   1− v2

 

c
0
t0 = p
t− vz
2 
 
 p
t0
+ vz
 t = c2
c
2 2
1− v2 1− vc2
c
10 Actually the situation is much more complex. All of Lorentz' principle, Einstein's special relativity principle
and the principle of Lorentz covariance are of restricted validity. See Szabó 2003.

6
 Only the sign of v changes. She will say that his metre sticks have contracted, that
his clocks run slow, and that he has not synchronized properly clocks at dierent
places. She will attribute his use of wrong variables to these FitzgeraldLarmor
LorentzPoincaré eects in his equipments. Her will be logically consistent and in
perfect accord with the observable facts. He will have no way of persuading her that
she is wrong.
11

It is sometimes objected that Lorentz's theory is unable to account for the Lorentz contraction of
a rod at rest, from the point of view of an observer in motion, since the contraction of a rod at
rest cannot be explainedthe objection saysby the deformations of moving objects, described
by Lorentz's theory. It is, however, clear from Bell's above analysis that the contraction of a
rod at rest, observed by an observer in motion, can be very well explained through the physical
deformations of the moving metre sticks and moving clocks of the moving observer.
Now we have completed the illustration of the fact that special relativity theory and Lorentz's
theory are equivalent with respect to the empirical facts.
12 Therefore, it seems, we have to make a
choice between Lorentz's theory and special relativity. There are many possible ideas behind the
choice of relativity theory.
13 Although it is not my concern here to valuate these justications, it
may be worthwhile reecting on a couple of common misunderstandings associated with Lorentz's
theory.

Remark 1 It is a common view that Lorentz's solution is based on  ad hoc  assumptions of the
existence of new phenomena:

So we are ultimately forced to choose between these two types of transformations and
to admit

1. either the continued validity of Newton's equations and of the Galilean trans-
formation insuring their invariance. In that case, we must, in electrodynamics,
assume the existence of new phenomena which nd expression in the Lorentz-
Poincaré equations, and which insure, by a sort of compensatory phenomenon,
the invariance of the Maxwell equations and the permanently elusive character of
ether.

2. or the universal validity of the Lorentz-Poincaré equations. The Lorentz trans-

formation entails the invariance of the Maxwell equations, but it requires the
construction of a kinematics and dynamics that will be in harmony with it. (Ton-
nelat 1966, pp. 127-128)

This is even more sharply formulated in the following quote from Cornelius Lanczos:

The negative eect induced Lorentz to assume that the motion relative to the aether
causes a contraction of lengths in the direction of motion (the FitzgeraldLorentz
contraction hypothesis), thus compensating for the eect which would otherwise occur.
(Lanczos 1970, p. 230)

This is however a misinterpretation not only of Lorentz's theory but also of the relativistic physics.
The contraction in Lorentz's theory is not assumed but rather derived from the ordinary pre-
relativistic laws of physics; it is not an extra hypothesis of Lorentz's theory that moving objects
suers deformations just nullifying the otherwise non-zero eect in the MichelsonMorley exper-
iment, as Lanczos and many others suppose it to be, but rather a simple consequence of the
standard pre-relativistic physical theories, like the Maxwellian electrodynamics.

Remark 2 To encounter these deformations, on the other hand, one does not need Lorentz's
theory. They follow also from the Einsteinian relativistic physics. Many believe, however, that
these deformations in relativity theory are of dierent nature from the similar deformations in
Lorentz's theory; according to this belief, the Lorentz contraction and the time dilatation in
11 Bell 1987, pp. 75-76.
12 For the extension of Lorentz theory for the general relativistic context, see Jánossy 1971.
13 Cf. Brush 1999.

7
relativity theory are not real physical processes, but they are just obtained from the comparison
of quantities dened in dierent reference frames. I will argue, on the contrary, that this is a
misinterpretation of the laws of relativistic physics; the deformation of a moving object in relativity
theory is as real as any other change of a physical system, associated with its motion. It is just as
real as, for example, the change of the electric eld of a point charge when we set it in motion.
Consider the following physical problem: Imagine a rod at rest in a reference frame K. What
kind of physical processes are going on, according to relativistic physics, when we set the rod
in motion? Does the length of the rod change, for example? According to the general rules

ore = l0 be the
K
of relativistic physics, one can solve this problem in the following way. Let lbef
length of the rod in the frame K before we set it in motion. One starts with the assumption
that the laws of physics determining the length of the rod are Lorentz covariant. Consequently,
0
the length of the moving rod in the co-moving reference frame K is equal to the length of the
K0
rod at rest in the original reference frame at rest, that is, laf ter = l0 . Now we perform a Lorentz
transformation back to the original frame and nd that the length of the moving rod in the original
q
v2
K
frame is laf ter = l0 1− K K
c2 . That is, laf ter < lbef ore , the rod has contracted. And this is a real
deformation of the rod. The critical point of misunderstanding is that such a question whether
or not the length of the rod has changed can be answered by comparing the earlier length of
the rod with its later length in the same inertial frame . One cannot argue that there is no real
deformation only because there exists some other reference frame K 0 (as it happens, the co-moving
one) such that
   
the length of the the length of the
=
deformed rod original rod
in K 0 in K
Arguing in this way would be as absurd as to say that a rod which is continuously at rest in K
0 K0
becomes deformed because there is another frame K such that l 6= lK .

Remark 3 Another source of confusion is the belief that the Lorentz contraction of a rod is only
a fact of kinematics in relativity theory, unlike Lorentz's theory, where it is a consequence of a
complex physical reasoning. This belief is, again, based on a misunderstanding of both theories.

(i) It is neither in relativity nor in Lorentz's theory true that the Lorentz contraction of a rod
simply derives from the rules of kinematics. To be sure, it is a fact of relativistic kinematics
q
0
v2
K
that laf ter = laf
K
ter 1−
c2 , as a simple consequence of the Lorentz transformation. But, it
is a contingent physical assertion, an empirically conrmed fact of nature, that the dynamics
of the constituents of the roddetermining its lengthis Lorentz covariant (the Lagrangian
K0
of the whole system is Lorentz invariant), that is, laf ter = lbef ore .
K

(ii) In Lorentz's theory, too, with the assumption that the physical laws determining the length
of the rod satisfy the Lorentz principle, one can calculate the contracted length of the rod in
the same simple way as in relativity theory.

(iii) On the other hand, also in relativistic physics one can perform the calculation in the reference
frame at rest, by directly applying the laws of (relativistic) physics to the moving object.
The procedure is analogous (and as complex as) the one within the framework of Lorentz's
theory.

Let me illustrate (ii) and (iii) by a more familiar textbook example: What is the electric eld of
a point charge moving with constant velocity? There are two methods to answer this question:

1. Consider the Coulomb eld of a charged point particle at rest. It follows from the Lorentz
(covariance) principle that the electric eld of the moving source in the co-moving reference
frame is also the Coulomb eld. One can then perform a Lorentz transformation from the
co-moving frame back to the reference frame at rest.
14

2. One can directly solve the Maxwell equations in case of a moving point charge. First one
solves the Maxwell equations for arbitrary time-depending sources. Then, from the retarded
potentials thus obtained, one derives the LienartWiechert potentials, from which one can
determine the electric eld.
15
14 Landau and Lifsic 1971.
15 Feynman, Leighton and Sands 1963, Vol. 2.

8
Both solutions lead to the same result: the electric eld of a charged point particle changes if we
set the particle in motion. Due to this change, for example, the track of a charged particle in a
bubble chamber becomes wider if the particle moves with higher velocity. The widening of the
track is a real, observable physical phenomenon. The change of the electric eld is a real physical
change.

Remark 4 Sometimes in the philosophy of science literature, the choice between Lorentz's the-
ory and special relativity is regarded as a typical example for the choice, governed by external
factors, between empirically under-determined scientic theories. For example, Michael Friedman,
admitting that the two theories are empirically equivalent, claims that there is a stage in the evo-
lution of theory ... at which methodological criteria do play a role, namely the elimination of
the LorentzFitzGerald-type 'aether' theory. So, on the basis of the methodological principle of
parsimony, he ends up with special relativity.
16 We must however emphasize that neither special
relativity nor Lorentz's theory needs to suppose the existence of aether, and none of them excludes
its existence. As we have seen, Lorentz's theory can be formulated without even mentioning the
aether.

Remark 5 With regard to approach (C), it is to be mentioned that, in nal analysis, there is
no dierence between the two theories whether they are used as principle theories or as more
detailed constructive descriptions. So the analogy with phenomenological thermodynamics does
23
not hold: It is, indeed, impossible to describe the detailed dynamics of the 10 molecules of a gas
in the terms of phenomenological thermodynamics. But, relativistic physics is capable to describe
the behaviour of the constituents of a solid body, describing in this way how the detailed dynamics
of the system leads to the LorentzFitzGerald contractions.

After this digression on how people usually justify the choice between Lorentz's theory and
special relativity, return to my main Thesis that there is actually no such a choice : The two
theories are completely identical.

2 Identity of special relativity and Lorentz's theory

Let me rst formulate an erroneous version of Poincaré's conventionalist schema for relativity:
 
  physics with  
classical empirical
+  deformations  =
space-time facts
of moving objects

 
  physics without  
Minkowski's empirical
+  deformations  =
space-time fact
of moving objects

However, as I pointed out in Remark 2, there is no such a benet of switching from the classical
space-time to Minkowski's space-time. All physical deformations of moving objects are still exist
in relativity theory, too. So the conventionalist schema does not express the correct relation of
special relativity and Lorentz's theory! The correct relationship would be this:
 
  physics with  
classical empirical
+  deformations  =
space-time facts
of moving objects
(9)
 
  physics with  
Minkowski's empirical
+  deformations  =
space-time fact
of moving objects

How is it possible? No doubtmany argue, special relativity and Lorentz's theory describe the
same observable physical phenomena, but they are dierent theories , because they account for the
geometry of space-time dierently. That is, the same phenomena are described by them in two
dierent ways. For example,
16 Friedman 1983, p. 293.

9
(S1) Velocity is an additive quantity in the Lorentz theory, but this is not true in relativity theory.

(S2) The speed of light is the same in dierent inertial frames in relativity, but this is not so in
Lorentz's theory.

(S3) Simultaneity is the same in all reference frame, according to Lorentz's theory, but this is not
true in relativity theory.

However, that is not the case. The three above sentences, for example, are false. As we will see,
they are false in the same trivial sense as if someone were confused with the British and American
usage of the word billion. The truth isas my main Thesis claimsthat the two theories are
completely identical in both sense, as physical theories and as theories about space and time.
When can we say that two theories are identical? Two physical theories are denitely identical
if they assert the same things about all physical quantities, that is, if they assign the same numbers
to all physical quantities with respect to all physical objects, and they claim the same functional
relations between these values. So, if sentence (S1) were true, the two theories could not be
identical. However, to compare two propositions of two dierent theories about physical quantities,
one has to clarify, rst of all, whether the propositions are about the same quantities, or not. And
here we arrive at an essential point: the meaning of the terms space coordinate, time coordinate,
velocity, etc. are dierent in Lorentz's theory and relativity theory. In sentence (S1) the term
velocity refers to dierent physical quantities in Lorentz's theory and in special relativity.
17
So let us start with clarifying these empirical denitions. We are interested in the space and
time coordinates dened in a reference frame at rest relative to the International Bureau of Weights
0
and Measures (BIPM) in Paris K , and the same quantities dened in a frame K moving with
velocity v relative to K according to Lorentz's theory and special relativity. For the sake of
simplicity consider only one space dimension and assume that the origin of both K and K 0 is at
BIPM at the initial moment of time.

Empirical denitions
(D1) Time at the origin in K according to Lorentz's theory

tK := τ

where τ is the reading of the standard clock at rest in the BIPM.

(D2) Space and time coordinates in K according to Lorentz's theory18

Assume we sent a light signal at time tK
1 from the origin of K (i. e., from the BIPM)
to the location of event A such that the signal arrived just when A occured. Then, at
the moment of A, we sent back a light signal to the origin, which arrived at time tK2 .
We dene the time and space coordinates of the event A as follows:

1 + t2
tK K
tK (A) :=
2
2 − t1
tK K
xK (A) := c
2
(The value of c, say 299792458 m
s is a convention in this approachotherwise distance
should be dened through a standard meter stick or the like.)

17 When I claim that velocity in the two theories is not the same physical quantity, I mean something entirely
dierent from the incommensurability thesis of the relativist philosophy of science (see Kuhn 1970, Chapter X;
Feyerabend 1970). As we will see later, both velocityLorentz and velocityrelativity are meaningful physical quantities
in both theories, and they are commensurable.
18 Throughout this paper I use the standard  ε = 1 -synchronization. I do not want to enter now into the
2
question of the conventionality of simultaneity, which is a hotly discussed separate problem. (See Reichenbach
1956; Grünbaum 1974; Salmon 1977; Malament 1977; Friedman 1983.)

10
(D3) Time at the origin in K according to special relativity

e
tK := τ

where τ is the reading of the standard clock at rest in the BIPM.

(D4) Space and time coordinates in K according to special relativity

Assume we sent a light signal at time e
tK
1 from the origin of K (i. e., from the BIPM)
to the location of event A such that the signal arrived just when A occured. Then, at
the moment of A, we sent back a light signal to the origin, which arrived at time e
tK2 .
We dene the time and space coordinates of the event A as follows:

e eK
1 + t2
tK
e
tK (A) :=
2
e
t2 − e
K
tK
1
eK (A)
x := c
2

(D5) Space and time coordinates of an event in K 0 according to Lorentz's theory

0 K0
The space coordinate of event A relative to the frame K is x (A) := xK (A)−vtK (A),
0 0
where v = v (K ) is the velocity of K relative to K in the sense of denition (D8).
K
0 K0
The time coordinate of event A relative to the frame K is t (A) := tK (A)

(D6) Time at the origin in K 0 according to special relativity

0
e
tK := τ
0
where τis the reading of the standard clock co-moving with the origin of K , such that
0
the clock was set to show τ = 0 when the origins of K and K coincided.

(D7) Space and time coordinates in K 0 according to special relativity

0
Assume we sent a light signal at time e
0
tK
1 from the origin of K to the location of event
A such that the signal arrived just when A occured. Then, at the moment of A, we
sent back a light signal to the origin, which arrived at time e
0
tK2 . We dene the time
and space coordinates of the event A as follows:

0
0 e1 + t2
tK eK 0
e
tK (A) :=
2
0 e
tK
2
0
− e
tK
1
0

eK (A)
x := c
2

(D8) Velocities in the dierent cases

Velocity is a quantity derived from the above dened space and time coordinates:

∆xK
vK =
∆tK
∆e
xK
veK =
∆e
tK
0
0 ∆xK
vK = 0
∆tK
0
0 ∆e
xK
veK =
∆e
tK 0

11
Remark 6 With these empirical denitions we dened eight quantities for each event, such that

xK (A) ≡ x
eK (A) (10)

tK (A) ≡ e
tK (A) (11)
K0 K0
x (A) 6≡ e (A)
x
K0 e
0
t (A) 6 ≡ tK (A)

where ≡ denotes the identical operational denition. Moreover, it is a contingent fact of nature
that

0 0
xK (A) 6= eK (A)
x (12)
K0 e
0
t (A) 6 = tK (A) (13)

Therefore, dierent physical quantities are called space coordinate, and similarly, dierent phys-
ical quantities are called time coordinate by Lorentz's theory and by special relativity and that
causes much confusion.
19
0 0
We also have to realize that xK (A) and tK (A) are meaningful physical quantities for special rela-
0 0
tivity, and, on the other eK (A) and e
hand, x tK (A) are intelligible physical quantities for Lorentz's
theory. In other words, assuming that both theories are suciently complete accounts of physical
reality, we can legally query the values of all four quantities in both theories. So, special relativity
and Lorentz's theory are dierent theories of space and time if they are dierent accounts of quan-
K0 0 0 0
tities x eK (A) and e
(A), tK (A), x tK (A). From the mere fact, however, that special relativity and
K0 K0 K0 0
Lorentz's theory call x , t e and e
, x tK dierently, it does not follow that they are dierent
theoretical descriptions of space and time. On the contrary, we will see that they are identical
descriptions.

Remark 7 Notice that the operations in denitions (D6) and (D7) are the blind repetitions of
the operations in denitions (D1) and (D2), simply ignoring the fact that a clock suers a loss
of phase when moving. This is obvious in case of (D6). In addition, the light signal denition
of simultaneity in (D7) has the same background: Let us calculate the reading of a clock slowly
0
transported in K from the origin to the locus of an event A. The clock is moving with a varying
velocity
20
(t ) = v + wK (tK )
K K
vC
0
where w (t ) is the velocity of the clock relative to K , that is, w (0) = 0 when it starts at
K K K
0
xC (0) = 0 (we assume that the origins of K and K coincide at t = 0) and wK (tK
K K
1 ) = 0 when
K
the clock arrives at the place of A. The reading of the clock at the time t1 will be

s
Z tK 2
1 (v + wK (t))
T = 1− dt (14)
0 c2

Since wK is small we may develop in powers of wK , and we nd from (14) when neglecting terms
of second and higher order
 R tK 
tK 1 w K (t) dt v
1 v+
tK (A) − x c(A)v
0 K

1 −
tK c2 2
T = q = q (15)
v2 v2
1− c2 1 − c2

Thus the reading of the clock slowly transported to the place of event A diers from tK (A) because
of the loss of phase accumulated by the clock during its journey. From the comparison of (15) and
(17) we can see that e 0
tK (A), dened in (D7), is equal to the reading T . In other words, denition
(D7) assigns a time tag to event A equal to the time measured with the slowly transported clock,
19 This was rst recognized by Bridgeman (Bridgman 1927, p. 12), although he did not investigated the further
consequences of this fact.
20 For the sake of simplicity we continue to restrict our calculation to the case of one space dimension. For the
general calculation of the phase shift suered by moving clocks, see Jánossy 1971, pp. 142147.

12
neglecting the fact that the clock in question is not only slowed down but also having a phase shift,
relative to the standard clock at rest in the BIPM.
The denition of the space tag in (D7),

0
0 e2 − t1
tK eK 0
eK (A) :=
x c
2
is actually equivalent with the ignorance of the contraction suered by a meter stick when we set
0
it in motion. Indeed, one can easily show that x eK (A) is nothing but the distance of the locus
0
of A from the origin of K , measured by a co-moving, therefore contracted, meter stick. And,
0
of course, if we measure the distance and time in K using the contracted meter stick and the
distorted clock, but disregard these distortions of the equipments, then we nd the speed of light
to be equal to c.

Remark 8 In denition (D5), on the contrary, the distortions suered by the etalon measuring
equipments when they are set in motion are taken into account. That is why we dene the space
0
and time tags in K through the original space and time data, measured by the original distortion
free equipments, being at rest relative to the BIPM. It is easy to see, that one would nd the same
xK (A) and tK (A) if the space and time tags were measured with the co-moving equipments, but,
with compensations of the distortionsaccording to our classical intuition. So, the basic dierence
between the denitions according to the two theories is that in Lorentz's theory the deformations
of the moving measuring apparatuses are taken into account, while these deformations are ignored
in the denitions according to special relativity.

Remark 9 We note that if v = 0 in (15) then T = tK (A), that conrms the denition (D2)
with the help of light signals in the reference frame at rest relative to the BIPM.

Remark 10 Due to the popular/textbook literature on relativity theory, there is a widespread

aversion to a privileged reference frame. However, like it or not, there exists a privileged reference
frame in both special relativity and Lorentz's theory. It is the frame of reference in which the
International Bureau of Weights and Measures is at rest. To be sure, it is not privileged by
nature. But it is privileged by the trivial semantical convention providing meanings for the terms
distance and time, by the fact that from all possible measure sticks and clocks of the universe
we have chosen as the etalons the ones oating together with the International Bureau of Weights
and Measures in Paris.
Many believe that one can avoid the reference to the etalons of a privileged frame by dening,
0
for example, the unit of time for an arbitrary (moving) frame of reference K in the sense of
0
denition (D6)through a cesium clock, or the like, co-moving with K . That is not the case,
however. Such a denition has several tacit assumptions like that the dierent cesium clocks go
uniformly and that the laws governing the behaviour of the cesium clocks are Lorentz covariant,
etc. The validity of such contingent statements cannot be empirically tested without comparing
the readings of the dierent cesium clocks with one etalon clock.

Now we are ready to prove that Lorentz's theory and special relativity are identical descriptions
K0 K0
of space and time. In order to see this identity, consider how special relativity describes x , t
eK and e
0 0
and, on the other hand, how Lorentz's theory accounts for x tK . In such a comparison we
can utilize the operational identities (10) and (11), and we can express everything through, say,
xK and tK . Let A be an arbitrary event.
According to the empirical denition (D5),

h 0 i h 0 i
xK (A) =x ve
eK (A) − etK (A) = xK (A) − vtK (A) = xK (A)
relativity Lorentz

where v = v K (K 0 ) = ve = veK (K 0 ). Similarly,

h 0 i h 0 i
tK (A) =e
tK (A) = tK (A) = tK (A)
relativity Lorentz

13
 tK

C
F
J
A
D H
O
xK
G
B

E I

0 0
Figure 4: Calculation of the coordinates eK (A)
x and e
tK (A), according to the denition (D7)

0
From the Lorentz transformation one can express eK (A)
x through the coordinates in K:
h 0 i eK (A) − ve e
x tK (A) xK (A) − v tK (A)
eK (A)
x = q = q (16)
1− e
2
relativity v2 1 − vc2
c 2

Similarly,
h 0 i ve
tK (A) − e x (A)
tK (A) − v x c2(A)
K K
e
e
tK (A) = q c2
= q (17)
1− e
2
relativity v2 1 − vc2
c 2

0
In Lorentz's theory, on the other hand, one can directly calculate the coordinates eK (A)
x and
e 0
tK (A), following denition (D7). The straight line EF (Fig. 1) corresponds to the trajectory of
0
the origin of K in the frame K . DA = x (A) and OD = t (A). We have the following equations:
K K

tK (A) − tK (E) c = xK (A) + tK (E)v (18)


tK (F ) − tK (A) c = x (A) − t (F )v
K K
(19)

Taking into account the slowing down of the standard clock moving along trajectory EF ,
r
e
0 v2
tK (E) = tK (E) 1 − 2
c
r
e
0 v2
tK (F ) = tK (F ) 1 − 2
c
from (18) and (19) we have

h 0 i e 0
tK (E) + e
0
tK (F ) tK (A) − v x c2(A)
K

e
tK (A) = = q
Lorentz 2 1− v
2
c2
h 0 i eK 0
t (F ) − t eK 0 (E) x (A) − v tK (A)
K
eK (A)
x = c= q
Lorentz 2 2
1 − vc2

We see thus that

h 0 i 0
h i
eK (A)
x = eK (A)
x
relativity Lorentz
h 0 i h 0 i
e
t (A)
K e
= t (A)
K
relativity Lorentz

14
 Consequently, there is no dierence in how the two theories account for the velocities:
h 0i h 0
i
vK = vK
relativity Lorentz
h 0i h i
K0
veK
= e
v
relativity Lorentz

Both Lorentz's theory and special relativity agree that what we traditionally call velocity is an
0 00 000
additive quantity. For arbitrary three frames K , K , K :

0 0 00
v K (K 000 ) = v K (K 00 ) + v K (K 000 )

For example,
0 0 00
v K (light signal) = v K (K 00 ) + v K (light signal)
At the same time, both special relativity and Lorentz's theory agree that the quantity vewhich
is called velocity by the relativity theoryis not additive:

0 00
0 veK (K 00 ) + veK (K 000 )
veK (K 000 ) =
1+ ev K 0 (K 00 )e
v K 00 (K 000 )
c2

There is no disagreement between the two theories that the (x, t)-map of the world (more
exactly the (x, y, z, t)-map
if we return to the 3-dimensional space), that is, the space-time in
1 3
the traditional sense can be conveniently described through a geometrical structure like E × E ,
1 3
where E is a one-dimensional Euclidean space for time, and E is a three-dimensional Euclidean
space for space, with two independent invariant metrics corresponding to the time and space
intervals.


Special relativity and Lorentz's theory agree that the e, ye, ze, e
x t -map of the world can be
0
conveniently described through a Minkowski geometry, such that the e tK -simultaneity can be
described through the orthogonality with respect of the 4-metric of the Minkowski space, etc.
0
Finally, since in an arbitrary inertial frame K for every event A the tags


(x(A), y(A), z(A), t(A)) can be expressed in terms of e(A), ye(A), ze(A), e
x t(A) and vice versa, the


laws of physics can be equally well expressed in terms of both (x, y, z, t) and xe, ye, ze, e
t .
Thus, we have completed the proof that there is no real choice between Lorentz's theory and
special relativity, because they are identical in both sense, they are identical theoretical descriptions
of physical space-time and they formulate identical laws of physics.

Concluding remarks

'What is time and what is space?' is a metaphysical question par excellence. There is no denite
answer to the question 'What kind of physical quantities can adequately represent the various
aspects of time and space perceptions?' For physics, however, time and space (distance) are
ordinaryalthough fundamentalphysical quantities with denite empirical meaning. As we
have seen the meaning of the terms space coordinate, time coordinate, velocity, etc. are
dierent in Lorentz's theory and special relativity.
Having claried the meaning of the terms, it turned out that the two theories are identical. In
other words, it is not the case that special relativity claims something new about space-timein
comparison with the pre-relativistic Galileo-invariant conceptions of space and time, but simply
calls something else space-time, and that something else has dierent properties. All statements
of special relativity about those features of reality that correspond to the traditional meaning of
the terms space and time are identical with the traditional pre-relativistic statements. On
the other hand, all statements of the pre-relativistic theory about those features of reality that
are called space and time by special relativity are identical with the statements of special
relativity. Thus the birth of special relativity was a terminological turn, rather than a revolution
in our conception of space and time.

Acknowledgements
The research was supported by the OTKA Foundation, No. T 037575 and No. T 032771.

15
References

Bell, J. S. (1987): Speakable and unspeakable in quantum mechanics, Cambridge University Press,
Cambridge.

Bell, J. S. (1992): George Francis FitzGerald, Physics World 5, pp. 31-35.

Bridgman, P. (1927): The Logic of Modern Physics, MacMillan, New York.

Brown, H. R. and Pooley, O. (2001): The origin of space-time metric: Bell's 'Lorentzian pedagogy'
and its signicance in general relativity, in Physics meets philosophy at the Planck scale.
Contemporary theories in quantum gravity, C. Calleander and N. Huggett (eds.), Cambridge
University Press, Cambridge.

Brush, S. G. (1999): Why was Relativity Accepted?, Physics in Perspective 1, pp. 184214.

Einstein, A. (1920): Über die spezielle und die allgemeine Relativitätstheorie (Gemeinver-
ständlich), Friedr. Vieweg & Sohn, Braunschweig.

Einstein, A. (1961): Relativity, the special and the general theory: a popular exposition, Crown
Publishers, New York.

Einstein, A. (1969): Autobiographical Notes, in Albert Einstein: Philosopher-Scientist, Vol. 1.,

P. A. Schilpp (ed.), Open Court, Illionis.

Einstein, A. (1982): Ideas and Opinions, Crown Publishers, New York.

Einstein, A. (1983): Sidelights on relativity, Dover, New York.

Feyerabend, P. K. (1970): Consolation for the Specialist, in Criticism and the Growth of Knowl-
edge, I. Lakatos and A. Musgrave (eds.), Cambridge University Press, Cambridge, pp. 197
230.

Feynman, R. P., Leighton, R. B. and Sands, M. (1963): The Feynman lectures on physics,
Addison-Wesley Pub. Co., Reading, Mass.

Friedman, M. (1983): Foundations of Space-Time Theories  Relativistic Physics and Philosophy

of Science, Princeton University Press, Princeton.

Grünbaum, A. (1974): Philosophical Problems of Space and Time, Boston Studies in the Philos-
ophy of Science, Vol. XII. (R. S. Cohen and M. W. Wartofsky, eds.) D. Reidel, Dordrecht.

Jánossy, L. (1971): Theory of relativity based on physical reality, Akadémiai Kiadó, Budapest.

Kuhn, T. S. (1970): The Structure of Scientic Revolution, University of Chicago Press, Chicago.

Lanczos, C. (1970): Space through the Ages. The Evolution of Geometrical Ideas from Pythagoras
to Hilbert and Einstein, Academic Press, London and New York.

Landau, L. D. and Lifsic, E. M. (1971): The classical theory of elds, Pergamon Press, Oxford,
New York.

Malament, D. (1977): Causal Theories of Time and the Conventionality of Simultaneity, Noûs
11, p. 293.

Reichenbach, H. (1956): The Direction of Time, University of California Press, Berkeley.

Salmon, W. C. (1977): The Philosophical Signicance of the One-Way Speed of Light, Noûs 11,
p. 253.

Szabó, L. E. (2003): On the meaning of Lorentz covariance, arXiv:physics/0308036.

Tonnelat, M.-A. (1966): The Principles of Electromagnetic Theory and of Relativity, D. Reidel
Publishing Company, Dordrecht, Holland.

16