This article was downloaded by: [Moskow State Univ Bibliote]

On: 19 November 2013, At: 16:35

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number:
1072954 Registered office: Mortimer House, 37-41 Mortimer Street,
London W1T 3JH, UK

International Studies in the

Philosophy of Science
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/cisp20

The genesis and structure of

models in the modern theory
of gravity
a
R. M. Nugayev
a
Lecturer in the department of philosophy ,
University of Kazan , USSR
Published online: 09 Jun 2008.

To cite this article: R. M. Nugayev (1987) The genesis and structure of models in
the modern theory of gravity, International Studies in the Philosophy of Science,
2:1, 84-104, DOI: 10.1080/02698598708573304

To link to this article: http://dx.doi.org/10.1080/02698598708573304

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all
the information (the “Content”) contained in the publications on our
platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy,
completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of
the authors, and are not the views of or endorsed by Taylor & Francis.
The accuracy of the Content should not be relied upon and should be
independently verified with primary sources of information. Taylor and
Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in
relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
 reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access
and use can be found at http://www.tandfonline.com/page/terms-and-
conditions
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013
 The genesis and structure of
models in the modern theory
of gravity

R. M. Nugayev
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

Abstract

The analysis of theory-choice problems in modern theory of gravity

necessitates consideration of the genesis and structure of the systems of
gravitational abstract objects. My approach to physical theory structure
uses and develops the ideas of V. S. Stepin. The fundamental equations of
general relativity Einstein's equations - are shown to describe the relations
between the abstract objects of the basic model only. The theory of gravity
chain consists of general relativity, nonmetric theories of gravitation,
Newton's theory of gravitation and the special theory of relativity. All
these theories describe the relations between the abstract objects of partial
models subordinated to the basic one.

1 Introduction. Metric and nonmetric programs: the cold war or

peaceful coexistence?
As is well known, a situation in modern physics has arisen where all the
available experimental data are satisfactorily described by more than
twenty relativistic theories of gravitation (see, for instance, ref. 1 and the
papers cited therein). Well, what shall we do to eliminate the uncertainty?
'Whether 'tis nobler in the mind to suffer the slings and arrows of
outrageous fortune or to take arms against the sea of troubles and by
opposing end them?' (W. Shakespeare).
They usually say that one should make a choice, in favor of one of those
empirically-equivalent theories, in accordance with certain rules. That is
why such situations are usually called 'theory-choice situations'1 in philosophy
of science, and the problem of uncertainty elimination for the sake of
further unrestrained development - 'a theory-choice problem' (see, for
example, ref. 2 and papers cited therein).
International Studies in the Philosophy of Science Volume 2 Number 1 Autumn 1987
© R.K.P. 1987 0269-8595/87/0201-0000 $2

84
 The genesis and structure of models in the modern theory of gravity

All the relativistic theories of gravity (RTG) are proposed within metric
or nonmetric competing programs. The metric program (A. Einstein, H.
Cartan, R. Dicke et al.) is based on the assumption that the gravitational
field description necessitates the application of the non-Euclidean mathe-
matical calculus. According to the nonmetric program (H. Poincare, W.
Pauli, R. Feynman et al.), the gravitational field must be considered as a
usual physical field analogous to an electromagnetic one or to a field of
strong interactions. Gravity can be described with only half of the
Minkowski metric.
The number of gravitational theories significantly exceeds the number
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

of gravitational experiments, but there are only two principal programs

being carried out. Is it possible to make the choice between programs, not
theories first? If it is, the 'theory-choice problem' can be solved in
principle. For instance, the choice of an adequate theory within the metric
program is possible due to PPN experiments of Nordtvedt and Will ([3]).
In the 1960s A. Schiff made a conjecture that would allow a choice
between the two programs. In brief, the conjecture states that all theories
of gravity which satisfy the weak equivalence principle (WEP), (i.e. which
predict a unique composition-independent trajectory for any test body at a
given point of spacetime and with a given initial velocity through that
point), must satisfy the Einstein principle of equivalence (EEP), (i.e. must
show that the nongravitational laws of physics are the same in every freely
falling frame.) When specialized to relativistic theories of gravity, Schiff s
conjecture says that every theory satisfying WEP is necessarily a 'metric
theory', i.e. belongs to a metric program.
Plausibility arguments have frequently been given for the conjecture,
but there have been few detailed calculations that bear upon its validity or
invalidity. In 1973 Alan P. Lightman and David L. Lee (California
Institute of Technology, Thome's group) worked out the method of testing
a relativistic theory of gravity for WEP (well-founded by the experiments
of Eotvos, Dicke and Braginsky). Applying the method to all the
nonmetric theories they found in the published literature, Lightman and
Lee ruled out all of them (nonmetric theories of Belinfante and Swihart, as
well as those of Naida and Capella, previously were believed to agree with
all current experiments). But was the 'theory-choice problem' solved in
principle?
No. As Lightman and Lee point out, 'Schiffs conjecture' is so sweeping
that it will probably never be proved with complete generality. 'Such a
proof would require a moderately deep understanding of all gravitation
theories that satisfy WEP including theories not yet invented, and never
destined to be invented. Such understanding is well beyond one's grasp in
1973', ([4], p. 384). On the other hand, one can gain useful insights by
proving restricted versions of the conjecture, and by searching for the most
general versions that are provable. For example, one might first analyze

85
 R. M. Nugayev

test bodies with purely electromagnetic internal interactions and thereby

attempt to show that particles and electromagnetism must interact with
gravity in the manner of metric theories in order that WSP be satisfied.
Next one might analyze purely nuclear systems and attempt to show that
nuclear fields must couple to gravity metrically. But unfortunately, even a
general proof of Schiff's conjecture for bodies with internal electromagnetic
interactions is too much to expect. To make it manageable, one must
assume some restricted form for the interactions.
The aim of this paper is to develop the Schiff program further and to
strengthen arguments in favor of metric theories. (Yet, as a byproduct of
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

it, it turns out that there are some nonmetric theories, but not the
nonmetric program itself, that must be ruled out. This program appears to
be a necessary link connecting the theoretical basis with the experimental
one).
This paper was initiated by the author in the course of an examination
of a nonmetric theory of gravity on the Weak Principle of Equivalence,
carried out under tenure of the Kazan State University Relativity Chair ([5]).
The theory in question considered a very contradicted WEP indeed.
But even its examination demanded some alterations of Lightman-and-
Lee's rules. These alterations appeared to be so considerable that it shook
my conviction in the omnipotence of empirical methods. One of Lightman-
and-Lee's restrictions appeared to be completely inapplicable to the whole
class of nonmetric theories.
In the course of further investigations it became more and more clear
that to develop Schiff s approach we must give up the traditional individual
comparison of each RTG with experiments. The problem should be
attacked in a broader methodological context than that proposed by the
Received View ([6]). Not theories themselves, but the programs are
decisive. To solve the theory-choice problem in principle we have to
consider the relative development of Einstein's and Poincare's programs
(see section 3 for details), i.e. apply Imre Lakatos' Methodology of
Scientific Research Programs [7].
The Methodology of Scientific Research Programs (SRP) develops
standards for the evaluation of scientific change that apply not to an
isolated theory or a conjunction of theories but rather to a research
program, with a conventionally accepted 'hard core' and with a 'positive
heuristic' which defines problems, outlines the construction of a belt of
auxiliary hypotheses, foresees anomalies and turns them into examples, all
according to a preconceived plan. Lakatos's system evaluates the evolution
of a program and not its shape at a definite time. Further, the evolution is
estimated only in comparison with the evolution of rivals, not by itself.
Research programs can be evaluated in terms of progressive and
degenerating problemshifts. The progress in one program P, is a vital
factor in the degeneration of its rival. If P| constantly produces 'novel

86
 The genesis and structure of models in the modern theory of gravity

facts' these, by definition, will be anomalies for the rival program P2. If a
research program progressively explains more than a rival, it 'supercedes,'
it, and the rival can be eliminated. But did the metric program supercede
the nonmetric one?
Before answering this question it is worth noting that the methodology
of SRP is not free of some drawbacks (see ref. 8 for details). First of all,
Lakatos's primary model of programs' competition is too rough to describe
the origin of theory-choice situations. What does Lakatos's model look
like?
'When two research programs compete, their first "ideal" models
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

usually deal with different aspects of the domain (for example, the first
model of Newton's semi-corpusular optics described light-refraction; the
first model of Huyghens's wave optics - light interference). As the rival
research programs expand, they gradually encroach on each other's
territory and the n-th version of the first will be blatantly, dramatically
inconsistent with the m-th version of the second' ([7], p. 158).
But unfortunately, neither in 'Falsification and the Methodology of
Scientific Research Programs', nor in his later works, does Lakatos explain
the following properties of the competition process.
(i) If the ideal models of the first program are dealing with one aspect
of the domain while the ideal models of the second program are dealing
with another, how can the theories in both programs lead to the same
empirical consequences? The existence of a theory-choice situation is
considered merely a fact of external history.
(ii) Lakatos's primary model admits the cases where N (N > 2) rival
programs compete. Though actual appraisals are always comparative in the
Methodology of SRP, the single criterion in terms of which such appraisals
are made is applicable to an individual research program alone. Choosing
between several programs, one first locates each individually on the
fruitfulness scale, and only then does one compare them ([9]). Since the
competing programs deal first with different aspects of the domain, we can
imagine a situation with N (N > 2) rival programs. Some of them
degenerate while the others keep successfully predicting novel facts, each
with respect to its own aspect. In such cases Lakatos's rules of SRP-
elimination seem to be insufficient.
(iii) All the Methodology of SRP case studies ([10]) consider two
competing programs. But why only two programs? In Lakatos's metho-
dology the facts about competition of two programs belong to the external
history.
But, according to Lakatos, each methodology (or theory of the
rationality of scientific progress) provides a theoretical framework for the
rational reconstruction of the history of science. All methodologies
function as historiographical (or meta-historical) research programs. The
normative-historiographical version of the methodology of SRP supplies a

87
 R. M. Nugayev

general theory of how to compare rival theories of rationality in which

history may be seen as a 'test' of its rational reconstructions. All
methodologies can be historiographically falsified. But 'falsified' cannot, of
course, be read in 'naive falsificationist' sense. It must be understood in a
more liberal sense developed for the methodology of SRP. Though
methodologies may not be falsified individually, they may be evaluated in
pairs, the first prize being given to the one that fits the larger range of
historical facts.
The methodology of SRP supercedes the others because it predicts (or
'retrodicts') novel historical facts, unexpected in the light of extant
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

historiographies, that were corroborated by a historical research. It

expands the range of internal, explanations reclaiming ground from
external history. Yet, according to Lakatos's own standards, each
methodology that expands the range of internal explanation in comparison
with the Methodology of SRP, supercedes it. For instance, the methodo-
logical model, that will be presented in section 2, turns the fact of the
competition of only two programs from an external (in Lakatos's
methodology) into an internal one (see [11] for details).
(iv) Real competition processes can only arise when the rival programs
are alternative. This means that the decision to accept one of them should
simultaneously be the decision to abandon the other. Therefore the hard
cores of alternative programs should be incompatible. And this is exactly
the case when each novel prediction of one program appears to be a vital
factor in the degeneration of its rival. All the 'novel facts' produced by one
program will be 'puzzling anomalies' for another only when their domains
of validity coincide. But Lakatos's model, with the different research
programs dealing with different aspects of the domain, admits the
existence of complementary programs (corpuscular optics, wave optics and
quantum theory of light).
So, in the course of its further development, the Methodology of SRP
should properly explain and theoretically reproduce the process of the
origin of theory-choice situations. Hence, having criticized the Methodology
of Lakatos, I would like to suggest a modified version of it and show how
this modified version can deal successfully with program competition in the
modern theory of gravity. Therefore, in the second section of my paper I
present a methodological model of the origin of theory-choice situations,
and in the third and fourth sections this is applied to gravity.

2 The methodological model of the occurrence and resolution of

theory-choice situations

An abstract theoretical object of a set of abstract theoretical objects of any

mature theory belongs either to a subset of basic theoretical objects or to a

88
 The genesis and structure of models in the modern theory of gravity

subset of derivative theoretical objects (see ref. [12]). According to their

definitions, the relations between basic objects are described by funda-
mental laws of the mature theory, and the relations between the derivative
objects are described by the consequence of the fundamental laws.
For instance (ref. [12]), 'the electric field at a point', 'the magnetic field
at a point', and 'current density' are the basic theoretical objects of
Maxwellian electrodynamics. The relations between them are described by
Maxwell's equations. 'The material point', 'the force', 'the inertial system
of reference' are the basic theoretical objects of Newtonian mechanics.
The relations between them are described by Newton's laws. The
derivative objects of Newtonian mechanics are: 'an absolutely rigid body',
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

'central field', 'harmonic oscillator', etc. The relations between them are
described by the particular laws of Newtonian mechanics: that is, by the
laws of a rigid rotation, movement in central field, etc.
The basic objects form the basis of a mature theory. This means that
each derivative object can join the system of theoretical objects only as a
result of construction from the basic objects according to certain rules. The
basic theoretical objects are constructively independent: none of them can
be constructed from the others.
So, the abstract objects of each mature theory are organized in a
complicated system, including the subsystems connected with each other
according to the level-hierarchy principle ([12], [13]). The subsystems of
the lower level are subordinated to a basic subsystem.
Completion in the creation of any mature theory (for example, T])
gives rise inevitably to questions about the relation of T,'s basis, (BO, to
the system of basic objects, (B2), of another mature physical theory, T2.
Are the theoretical objects B, K O and B2" (k,l=I,2, . . . ,n, . . . ,m)
constructively independent? Or is it likely that (B,) belongs to a subsystem
of derivative objects of T2?
It is impossible to answer these questions without taking into account
the following peculiarities of the derivative-object-construction rules.
(A) The rules for construction of the derivative objects from the basis
are not clearly and definitely formulated algorithms. They are vaguely
determined by the problem-solving examples or paradigms included in the
theory during the process of its genesis (ref. [14]).
(B) Application of these rules for reducing the basis to the subsystem of
the derivative objects presupposes that one should take into account the
peculiarities of empirical reality (ref. [12]). Those peculiarities vary from
one field of investigation to another.
The account, (A) to (B), demonstrates how exhausting it is to reveal
that Tj is subordinate to T2 (or vice versa). Therefore in everyday scientific
practice, simple minded conjunction of (Bi) and (B2) usually is assumed to
form a new basis.
The true necessity of analyzing the intricate interrelations of (Bx) and

89
 R. M. Nugayev

(B2) emerges in science only when the use of both theories together is
needed to explain certain experimental data. It is usually assumed that the
data can be described by a system of derivative objects constructed from
the basic objects of both theories. Utilizing V.P. Branksy's notion ([15]),
we will call such derivative objects 'crossbred objects' or simply "cross-
breeds'. The system of derivative crossbred objects will be the subsystem of
T( and simultaneously the subsystem of T2. The relations between the
crossbreeds will be described by the laws of both T, and T2.
The process of joint application of T| and T2 to solve a problem will be
called "theory cross', while T| and T2 will be named 'cross-theories'. The
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

set of statements describing the relations between the crossbreeds will be

given the name of a "crossbred theory'.
For instance, the completion of Maxwellian electrodynamics gave rise
to problems dealing with the relations of its basis to the system of basic
objects of Newtonian mechanics. The problems of a theoretical description
of the black-body radiation spectrum, the electromagnetic radiation
process, construction of a theoretical model of the atom and some others
necessitated the joint application of both theories (Poincare, 1890, note 16,
Planck, 1906, ref. 17, Einstein, 1909, ref. 18).
Consider one of these examples more thoroughly. While solving the
problems of the theoretical description of a black-body radiation spectrum,
J. Jeans (1905 [19]) and R. Rayleigh (1905) investigated the system of
standing electromagnetic waves in a closed cavity. The treatment of such
waves as a system of harmonic oscillators (the construction of crossbred
theoretical objects) enabled them to use a well-known law of statistical
mechanics (the equipartition theorem). That was the first time that the
temperature and frequency dependence of black-body radiation energy
were discovered in this way. The system of crossbred theoretical objects,
the correlation of which form a model of black-body radiation, is the
subsystem of classical electrodynamics (i.e. the system of standing
electromagnetic waves). On the other hand, it is this model that forms the
subsystem of derivative objects of classical mechanics (a mechanical system
with an infinite number of degrees of freedom).
The relations between the crossbred objects are described by T,
statements as well as by T2 statements. The crossbreeds belong to the
subsystems of both theories. Hence the operation of constructing crossbred
objects is identical to that of endowing the crossbred objects of each
crosstheory with new properties. The systems of derivative objects of each
cross-theory were constructed before they met. Each of these systems is a
peculiar generalization of the corresponding experimental studies carried
out independently of the investigations refering to another mature theory.
Therefore it is no wonder that there can appear theoretical objects with
incompatible properties resulting from the operation of crossbred construc-
tion in one and the same subsystem of derivative objects of one of the

90
 The genesis and structure of models in the modern theory of gravity

cross-theories. In the above example, the appearance of an object with

incompatible properties was characterized by physicists as the 'ultraviolet
catastrophe' (P. Ehrenfest).
Consider this trite paradox more carefully. As a result of constructing a
crossbreed system, the electromagnetic field appeared to possess an
additional property, transferred from mechanics - that is, 'to be a
mechanical system with an infinite number of degrees of freedom'.
Einstein, independently of Rayleigh and Jeans, and making use of classical
statistics, demonstrated that at an arbitrary but finite temperature, the
density of the electromagnetic field should be infinite. Indeed, at an
arbitrary finite temperature on each degree of freedom there falls one and
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

the same amount of energy proportional to temperature. However, the

infinity of the electromagnetic field density is incompatible with the second
principle of thermodynamics which is properly based from a statistical-
mechanical point of view. One can always extract energy from the cavity
containing such radiation and set the perpetuum mobile of the second kind
into motion. Therefore, the property of thermal radiation being a
mechanical system with an infinite number of degrees of freedom is
incompatible with its property 'to be a system of standing electromagnetic
waves'.
The system of theoretical statements is that of statements dealing with
the relations between abstract theoretical objects. Therefore, in the system
of derivative objects, the objects characterized by incompatible properties
should give rise to the mutually contradicting statements in both
crosstheories.
Bearing in mind Podgoretzky and Smorodinsky's notion (ref. [20]), I
would like to call the appearance of incompatible statements, when the
theories cross, 'cross-contradiction1. To give an example of cross-
contradiction, it suffices to list statements in the black-body radiation
theory. 'There exists heat equilibrium of radiation with matter' (this
theorem follows from the second law of thermodynamics; see Planck,
1906, ref. 17), and 'there does not exist heat equilibrium of radiation with
matter' (the consequence of Rayleigh-Jeans law; see Lorentz, 1909.
ref [21]).
The cross-contradiction results from the crossbred object construction.
To eliminate this cross-contradiction, one should, therefore, think of a
theory T3 satisfying the requirement that it comprise both cross-theories so
as to exclude the opportunity of constructing crossbreeds. Theory T3 will
be called a 'global' theory.
According to a methodological model I develop, two ways of global-
theory creation are logically admissible: 'reproductionisf and 'synthetic'.
(R): Application of a reductionist method of creating a global theory is
based on the assumption that the bases of both cross-theories refer to
different levels of theoretical object organization. Hence D,, the domain

91
 R. M. Nugayev

of validity of Ti, is part of D 2 , the domain of validity of T2. The basis of T2

acquires the title of a 'true basis.' And T2 itself is declared a 'fundamental
theory', while T[ a 'phenomenological' one.
To make sure of the validity of a phenomenological theory, one has to
construct its basic objects from the basis of a fundamental theory and prove
that its main laws follow from those of a fundamental theory. Lastly, the
basis of a phenomenological theory should occupy the place of a derivative
system of the fundamental theory. The opportunity to construct a
phenomenological basis from the basis of a fundamental theory should also
be grounded. The problems proving such opportunities are of special
importance. They are called "the fundamental problems'1 (see ref. 22).
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

(C) The application of a synthetic way of creating a global theory is

based on the following assumption: Basic objects of both cross-theories are
supposed to be constructively independent of each other. Their bases
belong to one and the same object-organization level. Hence, cross-
contradiction must be resolved by creating such a system of global objects
as from which the bases of both cross-theories could be constructed. The
fundamental laws of both cross-theories should be deduced from those of
the global theory. Finally, the bases of T| and T2 should occupy the
positions of the derivative subsystems of the global theory.
Usage of the ways of eliminating cross-contradiction is based on two
assumptions, equally reasonable, but mutually incompatible. Therefore, in
order to realize these assumptions, alternative programs - two reductionist
and one synthetic - of the global theory construction should be carried out.
Each program is to create its own sequence of scientific theories on the
basis of one of the above assumptions. That is why, following Lakatos, we
can call these fundamental assertions the 'hard cores' of the reductionist
and synthetic programs.
Neither a single 'crucial' experiment nor a sequence of such experiments
can definitely point out which program reductionist or synthetic - is able to
resolve successfully the cross-contradiction. For example, one can conclude
that a reductionistic program is unable to resolve the contradiction only if it
becomes clear that this program is unable to solve the fundamental
problems.
Each (n + l)th version of a reductionist or synthetic sequence of
theories represents a more perfect realization of a program than the n-th
version. Each of these sequences tends to a certain limit or ideal of the
global theory. It is the ideal that determines the direction of the
development of each SRP type. The third feature of a program that
enables it to develop successfully is the so-called 'protective belt of
auxiliary hypotheses' around the core against which the 'modus
tollens' are redirected. It should be specially pointed out that the term
'protective belt' was introduced by Imre Lakatos to characterize its main
function: 'to bear the brunt of tests and get adjusted and re-adjusted,

92
 The genesis and structure of models in the modern theory of gravity

or even completely replaced to defend thus a hardened core' (ref. 7, p.

133). This notion is not completely appropriate for my model since
reductionist and synthetic hard cores are irrefutable in it not by
convention, but as different means of cross-contradiction elimination.
Nevertheless, I retain Lakatos's notion taking into consideration 'Occam's
razor'. The protection of the reductionist program consists of a number of
assertions describing the relations between the theoretical objects of a
fundamental theory. The protective belt of synthetic SRP 'does not
actually emerge fully armed like Athene from the head of Zeus. It
develops slowly by a long preliminary process of trial and error' (Lakatos).
If the program is finished successfully, and the global theory is created,
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

say, in a reductionist way, it cannot be created by synthetic means.

Otherwise it would be possible to construct one and the same basis of the
phenomenological theory for sufficiently different domains of reality from
one and the same basic system of the synthetic global theory.
Nevertheless, suppose that all three programs are successfully realized,
and three global theories (two reductionist and one synthetic) are created.
Let us compare them.
All the ideals contain both T, basic objects and basic theoretical objects
of T2. The domain of validity of each ideal contains the T, domain of
validity together with the domain of validity of T The subsystems of
theoretical objects of each ideal contain one and the same derivative
objects of both cross-theories. The relations between basic and derivative
theoretical objects are described in each ideal by the same equations, i.e.
the partial and fundamental laws of both cross-theories. Each ideal
describes, explains and predicts experimental results, using the languages of
the same theories belonging to both cross-theories.
Any verification (or 'refutation') of a reductionist global theory is
verification (or 'refutation') of a synthetic global theory also. Any
consequence of a reductionist global theory may also be obtained from a
synthetic global theory. On the other hand, any consequence of a synthetic
theory, referring to the domains of validity of Tr and T2, can also be
obtained from a reductionist global theory. Therefore with respect to the
domains of validity of both cross-theories, the limits of all alternative
programs are empirically equivalent. In general, the ideals of all alternative
programs are empirically-equivalent only homomorphically. That is, each
corroboration of a reductionist global theory is simultaneously the
corroboration of a synthetic global theory, but the opposite statement is
invalid. As a matter of fact, various global theories differ in their ways of
organizing the same objects of both cross-theories into a unified body.
But to achieve more than one ideal is impossible. Only one of three
alternative sequencies will tend finally to its limit. That is why in a real
'live' science (and history of science) we can register the simultaneous
existence of theories from the sequences that belong to unfinished,

93
 R. M. Nugayev

alternative programs. The theories from unfinished programs cannot be

empirically equivalent with respect to each other in the strict meaning of
this term. They are empirically equivalent only approximately, with the
accuracy to the limit approach.
The fact of the simultaneous coexistence of theories from unfinished
alternative programs became known in philosophy of science as the
'theory-choice situation'. Since the genesis of the competing theories was
usually underestimated, the methodologists falsely concluded that resolu-
tion of a theory-choice situation should consist in choosing a single theory.
The account of this process testifies to the fact that a theory-choice
situation should be resolved by choosing a program not a theory. Not
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

theories but research programs are decisive. A program must be chosen

that can provide successful resolution of cross-contradiction. Let us return
to our history-of-science example.
As is well-known, Einstein's special theory of relativity and Lorentz's
ether theory co-existed in the early twentieth century as empirically
equivalent ones. What are the reasons for Einstein's victory over Lorentz?
Revelation of the cross-contradiction between the laws of mechanics
and the laws of electrodynamics gave rise to the first program, which
attempted to reduce electrodynamics to mechanics. Newtonian mechanics
was acknowledged a 'fundamental theory', while the electrodynamics was
acknowledged a 'particular theory'. The failure of Maxwell's reductionistic
program forced Lorentz to complete the system of Maxwell's equations
with the system of Newton's equations and to incorporate the 'Lorentz's
force'. Basic theoretical objects of both theories were unified in a single
theory. But the unification appeared to be unsatisfactory.
However, Lorentz had an opportunity to realise the opposite reduc-
tionist program to Maxwell's. Maxwellian electrodynamics became the
fundamental theory, and Newtonian mechanics - a particular one. The
fundamental problem of Lorentz's program consisted in 'consideration of
the particles of matter as some local perturbations of the ether' (Lorentz,
1909), i.e. in the construction of an electromagnetic model of an electron.
But, as it was shown in 1905 by Poincare, the contractile electron could
be considered as a stable entity only if the definitely non-electromagnetic
counterpressures were invoked. The only opportunity of further realization
of the reductionist program lay in the constructing of a structureless
electron. But this failed in 1909, when Lorentz calculated the force with
which an electron acts on itself. If we try in the force-expression to
eliminate the structure-dependent terms by letting the radius of the
electron approach zero, the energy of the self interaction will diverge. This
is physically meaningless. The program of the reduction of mechanics to
electrodynamics failed.
Einstein realized the synthetic method of unification. But in order to
explain why the scientific community accepted the special theory of

94
 The genesis and structure of models in the modern theory of gravity

relativity in 1910-12, we must renounce the traditional comparison of

Lorentz's 'ether program' solely with Einstein's 'On the Electrodynamics
of Moving Bodies' and the other 'pure relativity' papers. We must assume
that the three papers of 1905 were all parts of a single program. To
determine its 'hard core' we must address ourselves to Einstein's 'On a
Heuristical Point of View Concerning the Processes of Occurence and
Transformation of Light', concerning the photoeffect. The 1905 paper on
special relativity was published three months after this paper on the
photoeffect and is only a part of the unification program.
An ascertainment of 'deep formal difference between the theoretical
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

notions of the physicists of gases and other weighty bodies and Maxwell's
theory of electromagnetic processes' brings Einstein to his hypothesis on
discrete energy distribution of free radiation. The hypothesis explains the
photoluminescence, the occurrence of the cathode rays, etc. But 'are there
not the laws of occurrence and transformation of light such as if light
consists of similar energy quanta?' This is the question put by Einstein at
the end of his paper. But the concept of the ether prevents a positive
answer. We need the electromagnetic fields as independent formations
that can be emitted by the sources 'as well as in Newton's emitting theory'
(i.e. the energy transmitted in the process of eminence should not be
dissipated in space, but should be completely preserved until an elementary
act of absorbtion). But within the ether theory the electromagnetic field is
considered as a specific state of the ether — a state of a medium which is
continuously distributed in space. An elementary act of radiation is
connected in such a medium only with a spherical wave. While an outgoing
spherical wave is radiated by a single oscilating ion, the realization of the
in-going spherical wave needs an infinite number of radiating centres. The
process of radiation is irreversible in the ether theory.
But if the ether does not exist and the electromagnetic field is an
independent formation that propagates in vacuo, the velocity of the field
must depend on the velocity of the sources. The velocities of light and of
source must add up in accordance with Galileo's addition formula. But
that contradicts with the known experiments as well as with the astronomic
observations. In Lorentz's theory this problem did not even exist because
the velocity of light was determined there as the velocity of ether waves.
The velocity of waves could not depend on the velocity of their sources.
Consequently, if we want to consider the processes of occurrence and
absorbtion of light 'just as in Newton's emission theory' we ought to
replace all the classical kinematics by other laws that yield:
(1) The same velocity of light in any inertial system of reference;
(2) Galileo's addition formula for small velocities;
(3) Lorentz's transformations for space coordinates.
Namely, just what was done in the 1905 paper 'On the Electrodynamics

95
 R. M. Nugayev

of Moving Bodies' which was published three months after the work on the
photoeffect. Einstein disclosed that acceptance of (l)-(3) is equivalent to
the modification of the simultaneity concept and to clock delay in the
moving systems of reference.

3 A rational reconstruction of the origin of theory-choice situation in

the theory of gravity

The completion of the special theory of relativity (SR) gave rise to

Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

problems dealing with the relations of its basis (B)) to the system of basic
objects (B2) of Newton's theory of gravity. The description of fast-moving
gravitating mass interactions, of interaction of gravity and electro-
magnetism necessitated the joint application of both theories. It was quite
natural for solving these problems to suppose that a simple conjunction of
(Bi) and (B2) itself forms a new basis. In the cases of the theories's cross,
the experimental data had to be described by the objects constructed from
(Si) and (B2). The interpretation of gravitating mass as a 'gravitational
charge' made it possible to impose on the SR basis the restrictions that
corresponded to the gravitational-processes' peculiarities and to transform
it to the system of crossbred objects. It was in this way that the basic
objects of the first nonmetric theories of gravity were constructed
(Lorentz, Poincare, Einstein and Fokker, et al.).
The crossbreeds belong to derivative subsystems of both crosstheories.
The relations between these objects are described both by the laws of
special relativity and Newton's theory of gravity. But the bases of these
theories were created before they met. Each basis is apeculiar generalization
of the corresponding experimental studies carried out independently of the
investigations referring to another mature theory. Therefore it is of no
wonder that theoretical objects with incompatible properties (resulting
from the operation of crossbred construction in one and the same
subsystem of derivative objects of one of the cross-theories) can appear
(the situation of cross-contradiction). Just that happened when SR and
Newton's theory of gravity met at the beginning of the 20th century.
Each nonmetric theory of gravity is created in full analogy with
electrodynamics. Hence the gravitational field in such a theory obeys the
Superposition Principle. This means that the field f3 produced by two
'gravitating charges' m, and m2 is a simple sum of the fields f, and f2
produced by one 'charge' independently of the other. But, being applied to
the gravitational field, the Superposition Principle contradicts the SR
Principle of Equivalence. According to the latter Principle, the energy of
interaction can also be the source of a gravitational field. That is why the
field f3 should be less than (fi+f2) owing to the mass defect, i.e. due to the
interaction between mi and m2. The interaction diminishes their energy

96
 The genesis and structure of models in the modern theory of gravity

and, consequently, the whole mass of the system ([23]). The same
contradiction was pointed out by Misner, Thorne and Wheeler ([24]),
when analysing Pauli-Fiertz theory. In this nonmetric theory the field
equations contradict the equations of motion.
According to the methodological model developed, the cross-contradic-
tion between the SR and Newton's theory of gravity can be eliminated by
the realization of the reductionist as well as synthetic programs. From a
logical point of view, all the programs of an adequate RTG-construction
are equally admissible. But, as far as I know, there are only two synthetic
programs - metric and nonmetric - being carried out in real 'live' physics.
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

Poincare's nonmetric program began with the theories in which the

gravitational potential was described by scalar quantities. The attempts to
bring those theories into agreement with the results of three general
relativistic 'critical' or 'crucial' experiments (the bending of light in the
gravitational field of the sun, the red shift of spectral lines and the
anomalous behaviour of the Mercury perihelion) led to the creation of
mathematically sophisticated vector, tensor, scalar-vector, scalar-tensor
and scalar-vector-tensor non-metric theories. The fact of the simultaneous
existence of empirically-equivalent theories relative to the famous 'critical'
effects of General Relativity is described as a 'theory-choice situation in
theory of gravity'.
But almost all the nonmetric theories are fallacious ones. For instance,
many of them predict a null redshift for the particle version of theory of
light and a nonzero red shift for the wave version (see ref. 3). Almost all
the nonmetric theories contradict the Weak Principle of Equivalence.
Nevertheless, even the most strict and meticulous 'static' analysis of
ready-made, finished theories with the theory-of-gravity empirical basis is
insufficient. Let us suppose, for example, that we have shown that all the
nonmetric theories, created up to now, contradict the results of a certain
experiment. But who guarantees that in future there would not be created
such a nonmetric theory as could successfully explain all the difficulties
away?
That's why to find an effective method capable to resolve the theory-
choice situation we must compare, not theories, but the research programs
that create them. What is necessary is a history-of-science analysis of each
program: of the process of its genesis, development, modification, the
process of theory-change on the basis of the 'hard core' and negative and
positive heuristic. In the liberal limits of the latter, 'dynamical' approach
the following ways of theory-choice situation resolution have been
proposed.
A. To resolve the theory-choice situation an analysis of the individual
development of one program - that of Poincare - is sufficient. Poincare's
program is unacceptable since the additional hypotheses that provide the
creation of an (n+l)-th theory on the basis of the n-th are the ad hoc

97
 R. M. Nugayev

hypotheses (i = 1,2,3. see ref. [25] for detailed discussion of 'ad hoc
hypotheses' types).
Ai(i=l). The subsequent theories predict nothing new in comparison with
the previous ones.
A 2 (i=2). Even if they predict something new, none of their predictions was
experimentally confirmed.
A 3 (i=3). The additional hypothesis proposed contradict the positive
heuristic of Poincare program.
But an analysis of the Poincare program's individual development is
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

insufficient for the following reasons:

A]. The Tensor nonmetric theory of Belinfante and Swihart quite
definitely predicts the gyroscope precession in Earth's gravitational field
(see ref 26). This effect cannot be predicted by nonmetric scalar theories in
principle.
A2. Bending of light rays in the gravitational field was predicted first by the
nonmetric theory of Einstein and Fokker in 1911 [27].
A 3 . It is quite possible that some 'empirically progressive problem-shifts'
were performed within Poincare's program owing to radical changes in the
'positive heuristic'. But there are also important inconsistencies in the
realization of the metric program. The most obvious example of this is the
history of the gravitational waves problem. At first, in full accordance with
the Covariance Principle, the researchers tried to construct the covariant
stress-tensor of gravitational field. But it appeared that all the components
of this tensor must vanish identically owing to the conservation laws. Then
Einstein turned to the pseudo-tensor approach, but it roughly broke the
positive heuristic of the metric program. Introduction into the 'tissue' of
the program of such strange elements led to various contradictions. For
instance, we can always choose the system of reference in which the
components of the gravitational-energy pseudo-tensor vanish ([28]).
So, if the individual evolution of a program is insufficient, we must
consider the relative development of two programs. According to the
arguments proposed.
(B) comparison of Einstein's and Poincare's rival programs shows that the
first of them supersedes the second. Indeed,
(Bi) nonmetric theories predict nothing new in comparison with the metric
ones.
(B2) Theories from Einstein's and Poincare's programs are empirically-
equivalent only relative to the so-called 'critical' effects of GR. But, in
addition, to the consequencies of general relativity belongs the nonstation-
ary cosmological Freidman's solution, i.e. the Hubble shift of the far-
moving objects ([29]).
But the arguments (B!)-(B2) are insufficient also. (Bi) There exists a
field of investigations in the theory of gravity in which the nonmetric-

98
 The genesis and structure of models in the modern theory of gravity

theories' predictions deviate from those of metric ones. This field is the
domain of gravitational field quantization. Quantization of the gravitational
field within nonmetric programs causes no substantial difficulties at all"
Moreover, many of nonmetric theories are specially proposed for it. The
majority of nonmetric authors are the specialists in quantum-field theory
(Pauli, Thirring, Deser, Feynman, Weinberg, Belinfante, Logunov et al.).
The abstracts of many papers on nonmetric theories began like this: 'the
lozenz-invariant linear field theory constructed that is easily quantized'
(ref. 30, p. 168).
But within GR it is unclear up to now whether it has even been possible
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

to speak about the existence of gravitational-waves. Moreover in metric

theories the gravitational field is satisfactory quantized only in the so-called
'linear-approximation', i.e. in the weak field approximation. But they are
the processes taking place in strong fields where the quantum-gravity
effects are significant (see ref. 31, for example).
(B2) The explanation of cosmological observations is not the privilege of
metric theories alone. For instance, in A.A. Logunov and V.N. Folo-
meshkin's work 'the quasilinear theory of gravitation with linear equations
for free gravitational field is constructed and an opportunity of non-
stationary universe models is considered. The theory satisfactory describes
the red-shift effect' (abstract of ref. 32).
Thus the arguments (A)-(B) only reflect the general dissatisfaction
with nonmetric theories. They are insufficient to induce abandonment of
the Poincare program as a whole. The theory-choice problem in the theory
of gravity is the problem of a choice of an adequate programme of gravity -
investigation. But neither an evaluation of the Poincare program's
individual development, nor a relative comparison of the program-
competition according to I. Lakatos's criteria are sufficient until we find an
answer to the question, why the theories from both programs are
empirically-equivalent ones? But consideration of the theory-choice-
situation genesis shows that the cause of the empirical equivalence of
metric theories with nonmetric ones consists in the fact that Einstein's and
Poincare's programs are in fact the programs of unification of the same
crosstheories. Metric theories, as well as nonmetric ones, lead their way to
describe the results of experiments 'through' the language of both SR and
Newton's theory of gravity. That's why it is namely the terms of both
crosstheories that form the vocabulary of a !neutral language' for rival
theories comparison ([33]).
Einstein's and Poincare's programs should be compared as the different
programs of cross-contradiction elimination. In the course of comparison
such programs must be chosen as can successfully resolve this contradiction.
Let us consider the efficiency of the cross-contradiction elimination within
the Poincare program.
The field-theoretical World Picture constitutes the basis of this

99
 R. M. Nugayev

program. It presupposes the consideration of gravity as a usual physical

field in flat spacetime in full accordance with the ordinary electromagnetic
field. According to this World Picture, the cross-contradiction between SR
and Newton's theory of gravity is caused by the field equations'
independence from the equations of motion of the field's sources. In
Lorentz's Theory of Electrons this independence forced the addition to
Maxwell's equations of the equations of Newton. In the physics of the first
half of the 20th century the cross-contradiction between Maxwell's
electrodynamics and Newton's mechanics was eliminated (at least
partially) in quantum theory. This theory contains classical mechanics and
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

electrodynamics as its own partial theories. From a point of view of

classical electrodynamics, Maxwell's linear equations are only an approxi-
mation that describes the weak-field processes. The nonlinear character of
the charge-field interaction should be taken into account by Feynman's
diagram method.
That's why, in complete analogy with electrodynamics, nonmetric
theories are linear approximations only. The nonlinear character of
gravitational interactions should be taken into account with the help of
quantum-field methods (see, for example, Richard Feynman's report
'Quantum Theory of Gravity' at Warsaw International Congress, ref. [34],
pp. 697-698).
Success in Poincare program realization thus depends on the opportunity
to incorporate the theory of gravity into a more general quantum-field
theory scheme. Hence the problem of gravitational-field quantization is
crucial for the nonmetric program. If the gravitational field cannot be
quantized it means that the gravity basis cannot be constructed from the
basis of quantum electrodynamics. In this case we can (in principle)
construct a gravitational analogue of Heisenberg's y-microscope capable
to measure coordinates and velocities of elementary particles with any
degree of accuracy.

4. The resolution of theory-choice problem in the theory of gravity

The possibility of any field quantization is defined by the Bohr-Rosenfeld

analysis of ideal-measurement procedures. However, the application of
this analysis to linear gravitational field equations leads to the following
result ([35]). The gravitational-field quantities in the spacetime regions
connected by a light signal can be measured simultaneously with any
degree of accuracy. That is physically meaningless.
In general, the Bohr-Rosenfeld difficulties exist for the interpretation
of commutation relations for all bose-fields with spin higher than 1. In this
sense quantum electrodynamics is a physically privileged theory. Hence,

100
 The genesis and structure of models in the modern theory of gravity

what is good for quantum electrodynamics, is inapplicable to theory of

gravity. We can conclude that Poincare's program fails.
General Relativity and the metric theories realize another way of
unifying SR and Newton's theory of gravity. In the sequence of Einstein's
and his disciple's papers finished in 1939 (as well as in V.A. Fock's and his
disciple's investigations) it was shown that in GR the equations of motion
of sources are determined by field equations. Any text-book on gravity
contains the proof that SR and Newton's theory of gravity are the partial
theories of GR. The transition from GR to SR is governed by the Strong
Principle of Equivalence. It states that all physical laws should reduce to
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

that of SR for any freely falling body in the gravitational field system of
reference. The transition to Newton's theory of gravity is governed by the
so-called 'weak-field approximation in GR'.
Thus, the metric program seems to be the single program capable of
eliminating the cross-contradiction between SR and Newton's theory of
gravity. But does it means that we must refuse nonmetric theories?
Let us consider the transition from GR to Newton's theory of gravity
carefully. Almost all the sections of 'the weak-field approximation in GR'
begin like this.
Consider the weak gravitational field. In this case it is obvious that we
can choose such a system of reference in which all the metric-tensor
components slightly differ from their Minkowski values: g = TI + h (see,
for instance, ref [36]). The demand to ignore the squares and other
multiples of h is important for the transition to Newton's theory of gravity.
But it means nothing else than that index rising operation is carried out by
TI the metric tensor of flat spacetime. In the 'weak-field approximation'
the gravitational field equations take the form of usual wave equations in
flat spacetime. Consequently,
(1) In the approximation considered the GR equations transform to the
equations of one of the nonmetric theories (Pauli-Fiertz theory).
(2) The basic theoretical objects of the nonmetric theory appear to be
constructed on the GR basis. This is quite natural since the weak field is
considered as a tensor in flat spacetime. (ref. [29]).
So, one of the nonmetric theories appears to be, together with the SR,
a necessary link connecting the basis with the system of theoretical objects
of Newton's theory of gravity! This fact was usually ignored, that can be
explained by the GR ways of creation. GR's basis was constructed not by
the subsequent generalization of bases of less degree of generality (as
Maxwell's theory), but by the method of mathematical hypothesis (see, for
instance, [37]). In the last case the theory construction begins with the
search for an adequate mathematical formalism. The physical interpretation
is given later, when the theoretical model is justified as an idealized scheme
of real interactions (see [12]). The GR construction began with the 'high-
floors' - with Einstein's equations. But these equations describe the

I 101
 R. M. Nugayev

relations between GR basic objects only. But the whole theory-of-gravity

chain consists of the following links: GR, the nonmetric theory, SR and
Newton's theory of gravity.
The Einstein and Poincare programs are alternatives only as programs of
cross-contradiction between SR and NTG elimination. In all the other
aspects these programs are complementary since the experimental gravity
description necessarily involves the languages of nonmetric theories as well
as metric ones ([38], [39]).
The part of my investigation that is devoted to the necessity of
nonmetric theories is a realization of the ideas of A.Z. Petrov - the
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

founder of the Kazan University Relativity Chair. Petrov frequently

pointed out that the notion of a Riemann space-time continuum common
for all metric theories considerably complicates all the main notions of
theory of gravity and hampers the analogies with other physical theories.
Since the ambiguity is characteristic for all the main notions of general
relativism, the approach to this definition should be determined not by
analogies and incidental facts, but by general considerations, connected
with experiment, i.e. with the theory of physical measurements . . . 'No
matter how far the events lie out of the frames of classical physical
explanations all the experimental data should be described by classical
notions' ([40], pp. 59, 66).
It is a pleasure to thank the members of Kazan University Relativity
Chair for criticism and helpful advice.

Summary
A methodological model of the origin and settlement of theory-choice
situations (previously tested on the theories of Einstein and Lorentz in
electrodynamics) is applied to the theory of the Origin of Theory-Choice in
the Modern Theory of gravity. The process of origin and development of
empirically-equivalent relativistic theories of gravitation is theoretically
reproduced. It is argued that all of them are created within the two rival
programs - metric (Einstein et al.) and nonmetric (Poincare et al.). Each
program aims to eliminate the cross-contradiction between special theory
of relativity and Newton's theory of gravitation. New arguments in favor of
Einstein's program are given. But this does not mean that it is necessary to
rule out nonmetric theories, since Einstein's and Poincare's programs are
alternatives only as different tools of cross-contradiction elimination. In all
the other aspects these programs are complementary: description, explana-
tion and prediction of experimental results necessarily involves the usage
of the languages of nonmetric theories as well as of metric ones.

102
 The genesis and structure of models in the modern theory of gravity

REFERENCES
1 Thorne Kip S., Will Clifford M., Ni Wei-Tou, The Theoretical Frameworks for
Testing Relativistic Gravity—A Review, Pasadena, California, 1973.
2 Mamchur, E.A. The Theory-Choice Problem, Moscow, Nauka, 1975.
3 Nordtvedt K., Jr. Parametrized post-newtonian framework, The Physical
Review, 769-1968, pp. 1017-1031.
4 Lightman A.P., Lee David L., 'Restricted Proof that the Weak Equivalence
Principles Implies the Einstein Equivalence Principle', The Physical Review, D,
8-2- 1973, pp. 364-377.
5 Nugayev R.M., 'An Examination of the Nonmetric Theory of Gravity on the
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

Weak Principle of Equivalence, Relativity and Gravitation, Kazan, Kazan State

University Press, 13-1976, pp. 91-98 (in Russian).
6 Supper F., 'What is Wrong with the Received View on Scientific Theories',
Philosophy of Science, East Lansing, 39 -1-1972, pp. 1-19.
7 Lakatos I. 'Falsification and the Methodology of Scientific Research Pro-
grammes.' In: Criticism and the Growth of Knowledge, Cambridge University
Press, 1970.
8 Nugayev R.M., 'The History of Quantum Mechanics as a Decisive Argument
Favoring Einstein over Lorentz,' Philosophy of Science, East Lansing, 52 -1-
1985, pp. 44-63.
9 Kuhn T.S. 'The Halt and the Blind: Philosophy and History of Science', British
Journal for the Philosophy of Science, 31 - 1980, pp. 181-192.
10 Howson C. (ed.) Method and Appraisal in the Physical Sciences: The Critical
Background to Modern Science, 1800-1905, Cambridge University Press, 1976.
11 Nugayev R.M., 'A Study of Theory-Unification', British Journal for the
Philosophy of Science, 56-1985, pp. 159-173.
12 Stijopin V.S. The Formation of Scientific Theory, Minsk, Belorussian University
Press, 1976 (in Russian).
13 Margenau H., The Nature of Physical Reality, N.Y.-L., 1950.
14 Kuhn T.S. The Structure of Scientific Revolutions, 2-nd edition, Chicago,
University of Chicago Press, 1969.
15 Bransky V.P. The Problem of Synthesis of Relativistic and Quantum Principles,
Leningrad, Leningrad University Press, 1973 (in Russian).
16 Poincare H., Eletricite et Optique: cours de physique mathematique, Paris, L.
Carre, 1890.
17 Planck M., Vorlesungen uber die Theorie der Warmestrahlung, Leipzig, J.A.
Barth, 1906.
18 Einstein, A., 'Uber die Entwicklung unserer Anschauungen uber das
Wesen und die Konstitution der Strahlung', Physikalische Zeitschrift, 10- 1909,
pp. 817-825.
19 Jeans J., On the Partition of Energy Between Matter and Aether', Philosophical
Magaazine, 10 - 1905, pp. 91-98.
20 Podgoretzky M.I., Smorodinsky J.A. On the Axiomatic Structure of Physical
Theories. In: The Physical Theory, Moscow, Nauka, 1980, pp. 53-61 (in
Russian).
21 Lorentz H.A., The Theory of Electrons and Its Applications to the Phenomena
of Light and Radiant Heat, New York: L.F. Stechart, 1909.
22 Tisza L., 'The Logical Structure of Physics,' Reviews of Modern Physics, 35 -
1949, pp. 151-163
23 Novikov I.D., Zeldovich Ja.B., 'International Conference on Relativistic
Theories of Gravitation', I London, 1965.
103
 R. M. Nugayev

24 Misner C , Thorne K.S., Wheeler J.A. Gravitation, San Francisco, W.H.G.

Freeman and Company, 1973.
25 Zahar E. 'Why did Einstein's Programme Supersede Lorentz's?' British Journal
for the Philosophy of Science, 24 -1973, pp. 95-123, 232-262.
26 Lightman Alan P., Lee David L., 'Analysis of Belinfante-Swihart Theory of
Gravity', The Physical Review, D, 7 - 12-1973, pp. 3578-3586.
27 Einstein A., 'Grundlagen einer Theorie die Gravitation', Annalen der Physik,
35 -1911, pp. 898-910.
28 Infeld Leopold and Plebansky Jerzy, Motion and Relativism, New-York,
Pergamon Press, 1960.
29 Zeldovich Ja.B., Novikov I.D. Stars and Relativity, London, Pergamon Press,
1973.
Downloaded by [Moskow State Univ Bibliote] at 16:36 19 November 2013

30 Belinfante F., Swihart J., Phenomenological Linear Theory of Gravity, Annals

of Physics, 1. - 1957, pp. 168-182.
31 Hawking S.W., 'Particle Creation by Black Holes', Communications in the
Mathematical Physics, 43 - 1975, pp. 199-220.
32 Logunov A.A., Folomeshkin V.M. 'The Stress-Tensor Problem and the Theory
of Gravity', Theoretical and Mathematical Physics, 32-2 -1977, pp. 3-42 (in
Russian).
33 Thorne K.S., Lightman A.P., Lee D.L., 'Foundations for a Theory of
Gravitation Theories', The Physical Review, D, 7-12-1973, pp. 3563-3579.
34 Feynman R.P., 'Quantum Theory of Gravitation', Meta Physica Polonica, 24-
1973, pp. 697-702.
35 Potupa A.S., Idealized-Measurement Procedures in Relativistic Quantum Field
Theory, Minsk, Belorussian University Press, 1970 (in Russian).
36 Landan L.D., Lifshitz E.M., Theory of Fields, New-York, Addison-Wesley,
1971.
37 Dyahlevy P.S., 'Introduction.' In: The Philosophical problems of Einstein's
Theory of Gravity and Relativistic Cosmology, Kijev, Naukova Lumka, 1965 (in
Russian).
38 Wignor, E.P., 'On invariance properties', Proceedings of the National Academy
of Sciences of USA, 28 -1949, pp. 314-318.
39 Arshinov, V.I., 'Levels of the structure of physical knowledge', Philosophical
Sciences, 6 - 1973, pp. 74-85.
40 Petrov A.Z., 'Creation of Gravitational-Waves Theory and Physical Measure-
ment'. In: The Methodological Problems of Theory of Measurements, Kijev,
Naukova Dumks, 1966 (in Russian).

104