Você está na página 1de 18

DISKUSSION

Quantum Gravity and the Structure of Scientific Revolutions


JURGEN AUDRETSCH
Summary
In a case study Kuhn's morphology of scientific revolutions is put to the test in confronting it
with the contemporary developments in physics. It is shown in detail, that Kuhn's scheme is not
compatible with the situation in physics today.

1. I N T R O D U C T I O N

Scientists are human beings. They are within their scientific activities subject to
influences which can only properly be described by psychological and
sociological methods. There are not many attempts today to do this
adequately. The number of people working out how scientists should behave,
is still larger than the number of people undertaking the difficult task to study
how scientists really do behave. It is the great merit of Kuhn to have pointed
out psychological, sociological as well as logical aspects of this question in his
essay "The Structure of Scientific Revolutions" (Kuhn, 1962). Being a
historian, Kuhn restricted to a historical method. He tried to extract some
general traits from a collection of historical examples. Those philosophers of
science trained in logical-empiricist analysis, in the construction of axiomatics
and the like, turned his historical analysis into the logical reconstruction of the
dynamics of theories, thus obtaining a certai~ general scheme for the
development of science. In the following we will give a critical discussion of
this scheme.
To extract a scheme and claim general validity for it, is in complete
accordance with Kuhn's intentions: "I am not less concerned with rational
reconstruction, with the discovery of essentials, than are philosophers of
science. My objective, too, is an understanding of science, of the reasons for its
special efficacy, of the cognitive status of its theories. But unlike most
philosophers of science, I began as a historian of science, examining closely the
facts of scientific life" (Kuhn 1970 b, p. 236). It is this generality of Kuhn's
scheme which, taken seriously, causes severe problems when it is confronted
with historical examples not used by Kuhn, or even with what is happening in
physics today, as we will demonstrate below.
A second motivation for this essay originates in the following observation:
Usually, analytical philosophy of science is concerned with what has been
XII/2(1981)
@FranzSteinerVerlagGmbH,D-6200Wiesbaden

Zeitschrift fiir allgemeine Wissensehaftstheorie

Quantum Gravity and the Structure of Scientific Revolutions

323

called in a slightly exaggerated way the "autopsy of dead theories ''l. Similarly,
history of science prefers geocentric astronomy, phlogiston theory and case
studies like this. Examples are only classified post hoc. What is needed instead
is the logical, psychological and sociological analysis of physics today,
especially of those parts which are still "under construction". It is necessary to
finally reach a "contemporary history" of science, and to examine closely the
facts of contemporary scientific life.
This essay tries to make such an attempt. It confronts Kuhn's morphology
of scientific revolution with actual developments of modern physics. The
result will give an answer to the question, whether Kuhn's scheme can
adequately be related to the situation in physics today, and consequently,
whether it can be taken as a basis for a generally valid dynamics of physical
theories.
The reader who is familiar with modern physics may immediately turn to
chapters 2c, 4 and 5. For readers who are not, the other chapters may be
helpful.

2. POINTS OF DEPARTURE

Before classifying metric theory of gravitation and quantum theory with


regard to Kuhn's concepts, we briefly recall to the reader's mind some
characteristic physical traits of both schemes without trying to give a complete
or systematic description.

2a. Metric Theory of Gravitation


Maxwell's theory of electromagnetism is incompatible with a non-relativistic geometry of space and time. This was the starting point for the introduction
of special relativity. An appropriate description of electromagnetism must
incorporate the special relativistic theory of the behaviour of measuring rods
and clocks, which in turn makes essentially use of light rays. Special relativity
represents a framework for the construction of theories and a prescription for
the extraction of observer-dependent physical predictions. It consists of i) a
flat Riemann space, representing the unified concept of space-time, ii) a system
of worldlines, representing in a cinematical (!) way a field of classical
observers 2, iii) the physical substratum, e. g. a Maxwell field or the worldtine
of a point particle, which is mathematically described for example by tensor
fields embedded in the space-time.
Measuring results are obtained by interaction between the observer and the
observer-independently introduced physical system. This measurement is a
Shrader-Frechette (1977).
2 Usually considerations are based on canonical systems of observers representing inertial
systems, but this is not a fundamental point. Observer fields representing accelerated frames of
reference can be introduced in a corresponding way.

324

Jiirgen Audretsch

classical one in the sense that the result is obtained without disturbing the
system. It is assumed that the influence of the observer can in principle be
eliminated. Furthermore it is assumed that the measurement can be performed
pointwise. Mathematically, the measuring results are obtained as scalars by
essentially "projecting" (taking scalar products) the geometrical quantities,
describing the physical system in a point, on quantities describing the observer
in the same point.
Accordingly, the framework of special relativity consists out a flat space
time which in no way can be influenced but in turn influences everything else,
because it represents the geometry which governs the kinematics of observers
and of physical objects. This is completed by a theory of classical measurements and by an all comprising demand all parts of physics are to be described
in such a way that the description fits into the scheme sketched above (all
physical theories are to be described in a special-relativistic way). It is this allclaim which we will rediscover again in general relativity. Theories of space
and time are the basis of other physical theories.
Because Newton's theory of gravitation is based on a different space-time, it
is the immediate consequence of the all-claim above to ask for the construction
of a special relativistic theory of gravitation. In fact this step can be omitted. In
Newton's theory it is a property of matter called mass which reacts to gravity.
As experiments show, it is again exactly this property which is responsible for
the inertial forces in accelerated frames of references. On this equality Einstein
based his procedure of unifying gravity and inertia in eliminating the concept
of gravitational force and inertial force all together. This is done by replacing
the flat space-time of the scheme above by an appropriately curved one.
Curvature takes on the role of gravity, the other ingredients of the scheme
above are taken over. The result is called metric theory of gravitation. Frames
of reference are still to be represented by the congruence of worldlines of the
respective observers. And again all non-gravitational fields are to be embedded
in this now curved space-time. Measurement remains a local process.
Another concept of special relativity is also taken over. Rest-mass is only a
special form of energy. And accordingly all sorts of energy, pressure and so on
are taken as source of the curvature representing gravity. Consequently, and
this is reflected by the non-linearity of the respective equations, even gravity
can act as source of gravity. Remembering that every measurement and that the
way a physical system develops is influenced by the underlying space-time
geometry, we have: all physical processes influence geometry (are to be taken
as source of curvature), and on the other hand all physical processes are
influenced by geometry (develop within the respective space-time). Note that
the all-claim of special relativity has now even been strengthened, because to
determine the geometry of space-time no single non-geometrized part of a
physical system may be omitted.
Essential consequences of the theory can be demonstrated experimentally.
The main domains of application of the theory up to now are those showing
strong gravitational fields: cosmology of the early stages of the universe,
exterior and interior of very compact stars.

Quantum Gravity and the Structure of Scientific Revolutions

325

2b. Quantum Theory


An essential characteristic of quantum theory is that it exactly predicts the
amount of our possible knowledge or ignorance of a physical state. Uncertainty relations prevent that particular physical quantities, for example
position and momentum, can be measured simultaneously with any desired
accuracy. It is this non-classical measurement which plays a central role in the
theory. Measurement always contains an interaction of the system with the
classical measuring device. This interaction itself is usually not treated within
quantum mechanical dynamics. Instead, its outcome is made a part of the
underlying axiomatic structure of the theory. Complete knowledge about a
physical state can only be obtained by measuring a complete set of
commutating observables. This state then develops according to the respective
dynamics in a mathematically well defined way. But future measurements can
be predicted only in the sense of a probability, accordingly future knowledge
of the system will be restricted by uncertainties. Quantum theory takes
seriously the point that we only know something for certain about a system
immediately after having performed an appropriate measurement.
The standard formulation of the quantum mechanics of a system with a
finite number of degrees of freedom makes use of a Hilbert space description,
with the physical state described by a Hitbert vector and the physical variables
attributed to selfadjoint operators defined by commutator relations. They
contain Planck's constant ~as a characteristic indicator because the basic
commutator between position and momentum, on which a Hamiltonian
approach is based, is of the dimension of an action.
The quantum theory of systems with an infinite number of degrees of
freedom like fields (as for example the electromagnetic field) is based on
essentially the same ideas and formulated as a quantum field theory. For
special relativistic fields the whole quantum mechanical apparatus can be
formulated in accordance with the demands of special relativity. As part of
quantum field theory, interacting many-particle systems with the possibility of
creation and annihilation of particles can appropriately be described.
With regard to the formalism there is no principle distinction between
"matter" and "interaction". For electromagnetically interacting electrons for
example both the Dirac field and the Maxwell field have to be quantized.
Nowhere within the quantum theoretical scheme are there any interactions,
substances or physical systems in general which are excluded from quantization. This is again an all-claim, of different nature but of equal totality and
universality. Roughly speaking, the fact that there is no physics without
measurement implies that there is no physics without quantization, demanding
the quantum mechanical form of measurements and of dynamics.
Traditionally the domain of application of quantum theory is microphysics,
but there are also important macroscopic quantum phenomena of laboratory
scale like superconductivity and superfluidity.

326

J/irgen Audretsch

2c. The Structure of the Two Basic Schemes


As it is common in theoretical physics, we have used the word "theory" for
the metric theory of gravitation and for quantum theory. But as we have
pointed out above, both of them are not a physical theory like for example
electrodynamics or the theory of the Dirac field or strong interaction.
Both are instead:
- Collections of basic terms and procedures on which a physical explanation
has to be based, i. e. to which physical theories must submit.
-Collections of this sort which govern the work of two groups of
practitioners (sometimes not very precisely called "relativists" and "fieldtheorists").
- Frameworks or schemes for the cognition, comprising a mathematical and a
conceptual apparatus.
- Schemes describing the structure of physical knowledge.
Basic sets of beliefs and committments of the scientific community.
For good reasons both schemes must therefore be called a paradigm in the sense
of Kuhn (1962).
These two paradigms exist simultaneously. In Kuhn's scheme too, there is a
situation, when there are two simultaneous paradigms. To enable a comparison, we will characterize the contemporary situation in physics in more detail.
Kuhn's paradigm is first of all sociologically defined and in this sense prior
to theory. He says (Kuhn 1970a, p. 10): "Normal science means research
firmly based upon one or more past scientific achievements, achievements that
some particular scientific community acknowledges for a time as supplying the
foundations for its further practice." These achievements "are sufficiently
unprecedented to attract an enduring group of adherents away from competing
modes of scientific activity '~ and they are simultaneously "sufficiently openended to leave all sorts of problems for the redefined group of practitioners to
resolve". "Men whose research is based on shared paradigms are committed to
the same rules and standards for scientific practice" (p.11).
Are there today such habits which can be attributed to our two paradigms ?
The sociological situation is characterized by the fact that there are indeed, as
far as the research interests are concerned, the "'family" of "relativists (generalrelativists, geometers)" and the much larger "family" of "field-theorists",
which base their activity on quantum theory. But for about fifty years, both
groups coexisted without much interaction, both convinced that their work
has nothing to do with the work of the other group. Furthermore this living
together was and still is essentially a peaceful one, and accordingly at least.at
the better universities, students are educated in both paradigms only chosing
afterwards to which one their research will be committed.
In addition it has to be stressed, that the two contemporary paradigms have
a particular trait in common which most of the historical paradigms (as quoted
by Kuhn) do not have: their claim for totality, universality, i. e. their all-claim.
From the point of view of quantum theory this means : Because the role of the
observer and the corresponding statistical interpretation is only appropriately
described within the scheme of quantum theory, all physical theories have to
-

Quantum Gravity and the Structureof ScientificRevolutions

327

be formulated in a quantum mechanical manner. According to the usual


procedure, certain physical properties of finite or infinite degrees of freedom
have to be represented by self-adjoint operators obeying commutator or anticommutator relation and so on. The all-claim of the paradigm "metric theory
of gravitation" on the other hand is : The gravitational interaction is a universal
one, it is to be represented by curved space-time, i. e. by geometry. Therefore
alltheories describing physical interactions are to be formulated in such a way
tfiat the influence of the omni-present gravity is encorporated in a geometrical
way, which means that the influence is to be formulated as the influence of a
nontrivial space-time geometry. And furthermore, that for all processes it is to
be taken into account, that they influence geometry (back-reaction). In the
standard procedure, space-time is represented as a manifold and its geometrical
properties are formulated in every point in the sense of a classical field theory.
Accordingly, to incorporate the interaction with geometry in this way, all
physics is to be described in a localized way.
As a third important fact it should be noted that the two paradigms
described above are incompatible. This should be distinguished from Kuhn's
opinion that paradigms are incomparable or incommensurable. One possibility
to demonstrate their incompatibility is to make use of the fact, that an essential
part of the metric gravitation paradigm is the concept of the classical observer
which enables pointlike measurements without disturbances. This is opposed
to the quantum mechanical process of measurement according to which
between two measurements only the probability for the future outcomes of the
subsequent measurement is "propagating". But of course the physical system
itself does not disappear in the meantime. Therefore, following the first
paradigm, the system is affected by curvature and it gravitates itself. Now
simply take this influence on the geometry and measure the respective change
of the geometry according to the rules of the metric gravitation paradigm, to
obtain in every point a rather complete information (no probability) about the
system between the two quantum mechanical measurements. A corresponding
physical situation, where this contradiction can be demonstrated, is for
example an electron passing a double-slit. Therefore, applying both paradigms
successively in the sense of a thought experiment, may lead to contradictions.
What we have briefly sketched above is, that there can be no submission of one
paradigm to the other.
Why then do the two groups of supporters of the respective paradigms live
nevertheless peacefully together ? Simply because despite of the all-claim and
despite of the incompatibility the two paradigms are traditionally applied to
different domains, namely macrophysics and microphysics. But today this
situations has changed considerably. There is an elaborate scheme to treat at
least certain aspects of the influence of metrically described gravity on
quantum field-theoretically described systems3. This can be done in a tolerably
consistent way only in the sense of approximations.
3 Thereis also an experimentdemonstratingthe influenceof gravityon macroscopicquantum
systems. But gravityis therebytreated on the Newtonianlevel,so that this experimentis of no
importancefor our discussion.

328

Jiirgen Audretsch
3. THE PRESENT SITUATION4

3a. Exterior Field Approximation


In the following we briefly discuss such an approximation in order to show
firstly that restricted physical questions can be answered already within the
approximation scheme, secondly that these approximations have only a limited
domain of application and finally that the importance of the intended physical
domain of application makes it highly desirable to have a complete theory
(with regard to which the approximation mentioned above will become at all
proper approximations). This last point then will in turn make understandable
the psychological situation of those working hard to find the new over-all
scheme.
It seems heuristically plausible that strong gravitational fields like for
example strong electric fields can create particles. We restrict our discussion to
cosmological particle creation. A reader not familiar with theoretical physics
may find the Appendix A helpful, which contains some arguments based
essentially on a dimensional analysis.
In the exterior field approximation the particle creating field enters the
calculation as an unquantized background field. It influences the quantized
field, but is not influenced itself. Accordingly, in our case the cosmological
space-time acts as a classical background. Geometry remains unquantized as it
is in a metric theory of gravitation. Its particle creating influence is represented
by the fact that quantum field theory is embedded in the given curved spacetime. Feasible as an in-out-approach to the process by which the incoming
physical state defined in an in-region in the distant past (preparating quantum
mechanical measurement) is compared with the outgoing state registered in the
out-region in the distant future (registrating quantum mechanical measurement). For the interval in between it is assumed that the interaction of the state
with the curved background develops without additional external disturbances. This implies that during this time no measurements of what sort so ever
may happen, and correspondingly nothing is known about the state in
between. Correlated to this is the fact that there exists no particle concept for
the region of interaction where, according to the naive picture, the particles are
produced. Only in the in-and-out-region, provided that the interaction
disappears appropriately (i. e. the space-time becomes flat), it is possible to
introduce a meaningful particle concept.
Referring to the dynamics in the Heisenberg picture, particle creation means
that if the system starts as empty space (i. e. as a vacuum state) in the in-region
then a measurement in the out-region will register particles. The mean-value of
the particle number operator attributed to the definition of particles in the outregion and taken with regard to the invacuum state is unequal to zero
i < vac,in Ta+utaout I vac,in > 12 =k 0
4 For a more detailed study of the topics mentioned in this chapter, see the review articles in:
(Hawking and Israel, 1979), (Held, 1980), (Isham et al., 1975), (L6vy and Deser, 1979).

Quantum Gravity and the Structureof ScientificRevolutions

329

To work this out mathematically the creation operators in the out-region a+o,,t
have to be expressed by the creation and annihilation operators in the inregion. It can be done by solving the particle field equation (e. g. the Dirac
equation for electrons and positrons) embedded into the respective space-time.
This is the way geometry comes in. Along these lines it has been worked out
for several cosmological space-times how the cosmological curvature can
create particles, thus filling the universe with matter.
What are the deficiencies of the procedure? Although feasible, it is of a
limited applicability: i) Because space-time itself remained unquantized,
although quantization seems to be necessary for the high curvature near the big
bang. ii) Because at the beginning of the universe there is a big bang and not a
flat space-time region, therefore the procedure cannot be applied for universe
with big bang. iii) Because the in-out-approach allows no statement about
what happens in between. This would be necessary for treating the backreaction of the created particles on the curvature. The general opinion is, that
these are deficiencies of an approximation which can be overcome in the
complete framework which encorporates geometry in a quantized way.
The approximation indicates already that there is a quantum mechanical
interaction between non-gravitational fields (matter fields) and geometry, and
that this interaction will be of a decisive importance for example during the
early stages of our universe. There are further examples for overlapping
domains of application, which cannot be treated appropriately with today's
methods.
Already at this stage of the development of the theoretical physical
procedure it is worthwhile to analyse the logical, conceptual and systematical
structure of the theory using the tools of modern philosophy of science to
obtain new answers to old philosophical questions concerning space, time,
matter and the way they are related. In order not to disturb our case study of
the structure of a contemporary scientific revolution, some comments are
given in Appendix B.
3b. Today's Attempts
It is characteristic for today's situation that those working in the field are
aware that something new is necessary, in fact that a new paradigm 5 in the
sense described in chapter 2 must be built up, and that this can only be done in a
paradigm unification absorbing the two momentary paradigms. To prepare the
appearance of a new paradigm and to prepare oneself for the time after its
appearance when it is to be applied, there is only one reasonable strategy which
can systematically be persued by a larger group of people: push as far as
possible one paradigm into the prevailing domain of application of the other in
integrating as many of its characteristics as possible, and at the same time try to
learn as much as possible about the "interaction" between the two paradigms.
This will at least enable you to appreciate the new paradigm when it comes up.
5 Of course a physicist would not cai1it a paradigm.

330

Jiirgen Audretsch

According to this strategy, the problems are attacked today along several
lines. We mention some very briefly. For details we refer the reader to the
literature cited above. The whole area of research is usually called Quantum

Gravity.
Quantum field theory in a given curved space-time (external field approximation) as sketched above is an attempt to incorporate the unquantized gravity
described as curved space-time into quantum field theory. Another attempt is
the so-called covariant approach to quantize geometry in quantizing the
metric. It is based on a separation of the metric according to
g~t~ = ~ht~ + h~
into a classical background vla(usually the one of the flat space-time of special
relativity) and a disturbance h~ which is to be quantized. In a metric theory of
gravitation the metric represents the gravitational dynamics in the form of a
non-trivial kinematics thus abolishing both concepts. The covariant approach
on the other hand artificially reintroduces both concepts in splitting the metric
into a kinematical part ~1,~ and a dynamical gravitational potential h~f3.This is
against the basic concept of the curved geometry paradigm, but has the
advantage that everything is reduced to field theory in Minkowski space. Now
the elaborated methods of the particle physics paradigm, learned in connection
with Yang-Mills theories respectively non-abelian gauge theories, can tentatively be applied. Thereby gravity comes out as the theory of gravitons and
gravitinos.
Another attempt is the canonical approach. In this case one takes gravity as
geometry seriously and tries to extract from the metric tensor the true
dynamical degrees of freedom, which then are taken as canonically conjugate
variables on which commutation relations are imposed.
4. QUANTUM GRAVITY AND THE ROLE OF THE CRISIS

Having discussed the characteristic traits of today's situation, we are now


prepared to compare it with Kuhn's view of the structure of scientific
revolutions. To do so, we briefly recapitulate Kuhn's scheme and then contrast
it with what is happening instead.

4a. Kuhn's Scheme


According to Kuhn's paradigm view of science the development of science
does not go by analytic continuation but shows the following steps (all
quotations refer to Kuhn, 1970a)
normal science ("paradigm-based research" (p. 25), "puzzle-solving activity" (p. 52), "a highly cummulative enterprise" (p. 52))
- awareness of anomalies (p. 52) ("opens a period in which conceptual
categories are adjusted until the initially anomalous has become the
anticipated" (p. 64))

Quantum Gravity and the Structureof ScientificRevolutions

331

crisis ( " w h e n . . . an anomaly comes to seem more than just another puzzle
of normal science, the transition to crisis and to extraordinary science has
begun" (p. 82))
of a new paradigm as the response to crisis ("the new
p a r a d i g m . . , emerges all at once, sometimes in the middle of the night, in
the mind of a man deeply immersed in crisis" (p. 89), "a relatively sudden
and unstructured event like the gestalt switch" (p. 122) "scales are falling
from the eyes" (p. 122))
- competition between the old and the new paradigm, succession of the new
one, ("that decision must be based less on past achievement than on future
promise" (p. 157), "a decision of that kind can only be made on faith" (p.
158), "a new scientific trfith does not triumph by convincing its opponents
and making them see the light, but rather because its opponents eventually
die, and a new generation grows up that is familiar with it" (quoted from
Max Plan&) (p. 151) "The transfer of allegiance from paradigm to paradigm
is a conversion experience that cannot be forced" (p. 151))
normal science ruled by the new paradigm.
Scientific revolutions are thereby called "those non-cummulative developmental episodes in which an older paradigm is replaced in whole or in part by in
imcompatible one" (p. 92). Extraordinary science as opposed to normal science
appears when normal science is faced with a crisis.
For further use we have to describe in more detail the two basic concepts
"anomaly" and "crisis" which play a central role in Kuhn's scheme:
According to Kuhn, scientific revolutions are initiated by anomalies and
because a paradigm will not too easily surrender, only those anomalies which
"penetrate existing knowledge to the core" will lead to paradigm change (p.
65). Anomalies are thereby seen predominantly as "anomalies in relation of an
existing theory to nature" (p. 97) or "anomaly in the fit between theory and
nature" (p. 81). "Anomaly appears only against the background provided by
the paradigm" (p. 65). Being engaged in a discussion with falsificationists, for
Kuhn anomalies were related to experiments which cause a failure of existing
rules. Technical breakdown is the core of the crisis (p. 69). Proliferation of
competing versions of a theory is a very usual symptom of a crisis (p. 71).
There is a state of a crisis when it is recognized that the paradigm is "failing in
application to its own traditional problems" (p. 69).
Studying a particular collection of historical examples, Kuhn based his
morphology of scientific revolutions implicitly on the following fundamental
assumption: For a domain of mature science there is always only one paradigm
A which is then replaced in a revolutionary process by a new paradigm B,
which again will be replaced in the same way by a third paradigm C and so on.

- e m e r g e n c e

A --~ B --~ C --~


When Kuhn speaks of competing paradigms, he has successive paradigms in
mind. During extraordinary science a paradigm and its predecessor are both
struggling for the mastery. Correspondingly during the period of normal

332

Jiirgen Audretsch

science, mature science is seen as a single paradigm science (Masterman 1970).


It is this assumption of a "one-dimensional" scheme for the paradigm change
which makes the crucial role of anomalies and crises plausible at all. Also the
sociological and psychological aspects of extraordinary science go back to this.
Furthermore, this assumption is responsible for the "rationality gaps", which
are sometimes attributed to Kuhn's scheme (compare for example Stegmiiller,
1976). This will become apparent when the assumption is given up.

4b. What is Happening Instead


Physics is an advanced mature science. Nevertheless it is not singleparadigm science, as the situation today demonstrates most drastically. Kuhn
seems not to be aware of the fact that a multi-paradigm status may occur in
mature science6. This will become a central point, when Kuhn's scheme is now
confronted with contemporary research in physics.
As we have described above in chapt. 2 and 3, physics today is characterized
by the following facts:
a) Contemporary physics is multi-paradigm science7 with two paradigms
having an all-claim and overlapping domains of application. But there is no
competition between the two paradigms. It is not assumed that one
paradigm will ever replace the other. There are neither anomalies nor is
there a crisis.
b) Nevertheless, apart from the elaboration of the twoparadigms in the sense of
normal science, there are evergrowing attempts to unify the two paradigms
A..,~

/,C
B

There are psychological, sociological and logical reasons for these attempts.
Physicists are thereby aware of the fact that the task of paradigm unification
necessarily means abolition of the two oM paradigms A and B in favour of a
new one C. A rational strategy exists to do systematically the first step towards the new paradigm.
In the following we will describe these points in more detail.
Both paradigms do not show any anomalies. There is not even one single
experiment related to the intended domain of application of the unified
paradigms. Accordingly there cannot be any experimental anomalies or
6 The idea that in normal science there is always only one paradigm in each field of science has
as well been criticized by Popper (1970, p. 55): "Although I find Kuhn's discovery of what he calls
'normal science most important, I do not agree that the history of science supports his doctrine
(essential for his theory of rational communication) that 'normally' we have one dominant theory
- a 'paradigm' in each scientific d o m a i n . . . "
7 Of course there are within each paradigm rivalling theories, for example based on different
field equations (Brans-Dicke eqs. versus Einstein eqs.). But this is of no importance for our
arguments.

Quantum Gravity and the Structure of Scientific Revolutions

333

falsifications. There are only within each paradigm the usual difficulties, in
connection with the standard puzzle solving of normal science. None of these
experimental "anomalies" is penetrating existing knowledge to the core. There
is no technical breakdown. Neither are there within each paradigm theoretical
anomalies. Infinities occur within quantum field theory, but there is as part of
the theory a canonical procedure to remove them (renormalization, regularization) 8. In General Relativity singularities occur. But all those which have been
found up to today are hidden behind horizons 9.
Although there are no anomalies, is there perhaps nevertheless a state of
crisis ? The contrary is the case; both paradigms are extremely efficient and
both are highly respected today. The quantum field theoretical paradigm has
just shown to be very successful. Within its framework one has been able to
unify electromagnetic and weak interaction. The scientific community has
recognized this in awarding the 1979 Nobel Prizes in physics. Furthermore
people are hopeful to include strong interaction in a similar way (grand
unification), because the only plausible theory of strong interaction is also a
gauge theory. Among physicists there is no widespread belief that one is
approaching a critical point or, even more dramatic, that one is pushed into a
CFISIS.

A third trait which characterizes today's multi-paradigm status in physics is


that there is no competition between the two paradigms. One does not find
two groups of supporters of different paradigms, but instead only scientists
temporarily engaged in normal science activities within one paradigm. The two

8 Shrader-Frechette (1977) has argued that high energy physics is passing to a crisis because the
elementary particle paradigm is to be replaced. From our point of view the elementary particle
concept is only one of the more or less elaborated concepts within the quantum field theory
paradigm. It is not a paradigm itself. The arguments given by Shrader-Freehette (1977) like
"absense of clear criteria", "particles versus excited states" and so on could be taken to
demonstrate this.
9 Kanitscheider (1977) has discussed the philosophical implications of the singularity theorem
of Hawking and Penrose in detail. Singularities (in black holes as big bang and so on), horizons
and similar concepts form an important part of today's metric theory of gravitation. Physics as we
know it from elsewhere breaks down when space-time horizons and singularkies come in. Not
only "usual" physics, but also philosophical concepts introduced in relation to this physics can
break down (e. g. the traditional causal structure in connection with horizons). Furthermore, it is a
characteristic trait that philosophical arguments can strongly interfere a discussion about for
example the relevance of singularities inside black holes. Based on this, Kanitscheider has
concluded that physics has reached a state of crisis in the sense of Kuhn.
The objection to this conclusion is the following: If a physical situation (e. g. occurrence of
horizons) can be called anomalous with regard to certain philosophical concepts or if certain
philosophical positions are assumed, this may well have little or no influence on physics. From a
singularity which is hidden behind a horizon, we can on principle never obtain informations. It
therefore depends on the respective philosophical position, if the occurrence of a singularity which
is hidden empirically (not theoretically) should be called a crisis. There is in fact no technical
breakdown as far as the mathematical treatment of space-time singularities is concerned.
Accordingly, it is the general opinion among physicists, that as long as the theory does not lead to
naked singularities (where the singularity would be empirically visible), there is no crisis. This is
important, because crisis in the sense of Kuhn has a large sociological component.
27

334

Jiirgen Audretsch

paradigms are well established. Accordingly many physicists are educated


parallely in both paradigms; and an ever growing number is changing
paradigm from time to time.
Although there are neither anomalies nor the sense of a crisis, a group of
physicists is engaged in what Kuhn would call extraordinary science. Why are
physicists trying to unify the two paradigms hoping to initiate the uprise of a
new paradigm? There are the following psychological, sociological and, last
not least, logical reasons for this:
There are some physicists, who, although trained this way, have an aversion
to paradigm pluralism. For them mature science should be single-paradigm
science. In many cases, this may be nothing more but an unreflected
preconception absorbed from introductory pages of physics textbooks or by
reading memoirs of famous physicists. But, on the other hand, the history of
physics shows that unification is a very successful method, what may justify
the preconception. In any case, aversion to pluralism is one of the driving
psychological motives for paradigm unification among physicists.
A second psychological motive is, as in Kuhn's case, the hope that the new
paradigm will show surplus meaning. There is in the first place the hope that
unification may remove infinities in both initial paradigms, so that they do not
even appear mathematically. Perhaps space-time singularities may be avoided
when the source becomes genuine quantum mechanical matter. Perhaps
inclusion of gravity may render regularization unnecessary. Another hope is
that incompleteness will be abolished, that for example the cosmic initial
conditions may be fixed or may become unnecessary. Inclusion of gravity may
fix the mass spectrum of the elementary particles. Again it is not a sense of
crisis which pushes the physicist, but the promisses of the forthcoming unified
paradigm attract them.
A very significant motive, which can only sociologically be explained, is that
unification has become fashionable again. From the point of view of
interactions we have four interactions: electromagnetic, weak, strong and
gravitational. The electroweak synthesis has been successful. The inclusion of
the third interaction looks promising. Why not attack gravity along the same
lines (for example make use of the gauge structure)? We have briefly
mentioned in chapt. 3b how this is done today.
Let us now turn to the logical reasons for paradigm unification. In this
connexion it is important to note that there is in fact a rational approach of the
sort which philosophers of science, trying to close "rationality gaps", would
like to find - and had difficulties to find when Kuhn's scheme was taken as a
basis. If instead the structure is multi-paradigm science demanding for
paradigm unification, a rational approach emerges quite naturally.
In contrast to an accumulation of anomalies, which according to Kuhn raises
the sense of a crisis, the situation today is characterized by the fact that existing
paradigms can properly be declared inadequate: The intended domains of
application of the two paradigms overlap. Because one agrees with the allclaims of both of them, it is the generally accepted opinion that situations, as
for example the early stage of the universe, can impossibly be treated using one

Quantum Gravity and the Structure of Scientific Revolutions

335

paradigm only. O n the other hand, paradigms are incompatible, what can be
elucidated by thought experiments (while proper experiments or observations
play no role at all in our case !). From this situation inevitably follows the need
to reconcile the all-claims in unifying the paradigms, what will lead to one allembracing paradigm. Therefore, in contradiction to Kuhn's scheme, today's
normal science can find its own limitations, not because there is a falsificating
experiment, but because normal science is multi-paradigmatic science and the
paradigms have an all-claim. The regions where the two paradigms overlap are
the sources of tension which leads to innovation.
As a second aspect of rationality we have shown in chapt. 3, that there is a
strategy to do the first step towards unification, which is again completely
rational Try to absorb as many elements of one paradigm into the other and
take the resulting theory as one particular limiting case of the yet unknown
new paradigm (compare the example of the external field approximation
sketched above). In doing so, there is a rational discussion between the two
groups of practitioners which are trained in the respective paradigms.
As a last comment we mention an aspect, which has also a sociological
component. If a new paradigm originating from paradigm unification appears,
it will have no difficulties to be accepted as the replacement of the two old
ones. Being a unification it comprises at least the two domains of application
where the previous ones have in fact been successful. From the very beginning
therefore this new paradigm is obviously superior to each of its two
predecessors. As compared with Kuhn's scheme of two succeeding paradigms.
there is a great deal of rationality also in this part of extraordinary science.
5. CONCLUSION

Our intention was not to criticize Kuhn's ideas from the point of view of
other elaborated programs and concepts of philosophy of science. This has
vastly been done already (compare for example the review in Stegmiiller,
1976). Neither have-we been interested in proposing an additional alternative
scheme and claiming general validity for it. Our aim was instead to put Kuhn's
scheme to the test in applying it to the developments in physics today.

We have shown above in detail that this contemporary situation in physics is


not compatible with Kuhn's scheme for the structure of scientific revolutions.
Accordingly, the minimal coclusion which has to be made is, that Kuhn's
scheme is at least incomplete and has to be adjusted in one way or the other.
Because there are indications that today's situation is not a singular one, that
there have been similar situations in the past as well l, our critique is in fact
potentially more far-reaching.
I0 It would be worthwhile to study in detail the following conjecture: Non-relativistic
quantum mechanics and special relativity were both coexisting paradigms with an all-claim.
Finally these all-claims were convincinglyconciliated in relativistic quantum field theory. As in the
case discussed above, there has been a phase in the meantime, when the deficiencies of certain
"approximations" to the new theory, which was still to be created, were discussed (Dirac equation
and Klein-Gordon equation in the framework of first quantization).
22*

336

Jiirgen Audretsch

Let us, at the end of this essay, prevent a misunderstanding. It has not been
written to contribute to the debate between Kuhn, Popper, Lakatos, Feyerabend and the respective followers which dominate today's discussion on
rationality and progress in science. We have criticized the work of Kuhn in
particular, because in this case the underlying structure of a "one-dimensional" dynamics of paradigms (a paradigm has only one predecessor) and its
consequences for the over-all scheme is most evident. But our objections apply
mutatis mutandis also to other approaches to theory dynamics, to the extent as
the are based on "one-dimensional" schemes which neglect the importance or
the necessity of unifications.

APPENDIX A: HEURISTIC APPROACH TO COSMOLOGICAL PARTICLE CREATION

A rather simple curved cosmological space-time is the 3-flat RobertsonWalker universe. It describes galaxies moving away from each other (expansion
of the universe) and is characterized by the fact that the 3-space between the
galaxies is 3-flat. Attributing fixed coordinate values x, y, z to particular
galaxies, the measured distance between them is given by the line-element
(1)

ds 2 = c2dt2 - R2(t) {dx2 + dy 2 + dz 2}

where the common factor R(t) determines the time dependence of the spacelike
distances and c is the velocity of light. For a radiation filled universe R(t) is
characteristically of the form
(2)

R ~ t 1/2

and for a matter filled universe


(3)

R ~ t 2/3

both showing a big bang at t = 0.


The gravitational interaction between the galaxies causes attraction. Accordingly, the velocity with which the galaxies move away from each other is
permanently diminished (decelerated expansion). There is a relative acceleration of the galaxies towards each other
In this universe not only two galaxies but any two test particles separated at
a small distance A1 experience an acceleration b towards each other of the form
(4)

b = Q-AI

where for the expansion laws above we have


(s)

Q - 1/.

Quantum Gravity and the Structure of Scientific Revolutions

337

In a very heuristic way the field theoretical vacuum may be described as


containing virtual pairs of particles and antiparticles of mass m, which are
continuously created and immediately afterwards annihilated. This creation
"out of nothing" violates the conservation of energy by an amount of

(6)

AE

2mc 2.

Such a violation can be tolerated, but only if it does not last very long. The
respective duration At is related to the energy uncertainty AE according to the
uncertainty relation
(7)

AE " At ~ h

The maximum distance a pair of particles can separate from each other during
At is
(8)

A1 -< c a t

where c is the velocity of light. This gives with (6) and (7)
(9)
AI~<h/mc = Compton wavelength = 4 - 1 0 - 1 1 c m
positrons).

(for electrons,

The energy E the particle can accumulate because of the gravitational


acceleration when travelling the distance 1, is because of (4) and (5)
(10)

E = force distance ~ m b - A1 ~ ~

(A1)2.

O n the other hand, the energy, a virtual particle has to accumulate to be able to
become real, is at least its rest energy E = mc 2. This implies with (10) taking
into account (9) the following condition.
(11)

E ~-

h2
--Z-c2 > mc 2

which is a condition for the age t of the universe. According to (11) it is only
during the time
(12)

t ~<~h/mc? = 1,3 lO-21sec (for electrons, positrons)

after the big bang, that the gravitational forces are strong enough to produce
particles.
What can be seen from the very simple picture is, that there is a cosmological
particle creation and that it is relevant only for the very early stages of the
universe. H o w efficient the process is, i. e. how much matter is produced, can
only be worked out by a more serious calculation.

338

Jiirgen Audretsch

APPENDIX B: SOME NEW COMMENTS ON OLD PHILOSOPHICAL QUESTIONS

The study of the exterior field approximation already leads to some new
comments on the old philosophical discussion of the relation between space,
time and matter. We add the following comments here in order to demonstrate
that contemporary physics may be of interest as well for those philosophers,
who are not predominantly interested in schemes for the dynamics of scientific
theories.
Geometry on one hand and matter/energy on the other, are they dependent
in a causal sense ? Are they dependent in an ontological sense (one is of a prior
reality, the other is derived)? Is there a predominance of one over the other in
the structure of the theory? (Is "Mach's Principle" incorporated in Einstein's
theory?)
Before the inclusion of quantum theory the arguments have been: According to Einstein's field equations
G=~ = T=~
the existence of matter/energy causes curvature. But solutions of
G~ = 0
for empty space-time exist. Because curvature without matter/energy is
possible, the conclusion has been, that geometry is the more basic concept. The
geometry is more primary because it may exist independently.
The external field approximation shows that in general space-time curvature
creates matter. Does this now imply a total symmetry between geometry and
matter content because every empty space-time immediately fills itself with
particles ? There is at least one counterexample that this is not the case. For
massive particles without spin (Klein-Gordon equation) it has been shown in
the exterior field approximation, that a plane sandwich gravitational wave leads
exactly to no creation of particles (Gibbons, 1975) Accordingly there are still
empty space-time solutions possible which remain empty. Nevertheless it is
obvious that the relation between matter/energy and curvature has become
more symmetric because each one may create the other.
As a second comment we would like to draw attention to the fact that empty
space-time has got a new quality. Instead of being describable as what is
relating objects or events, empty space-time is the place of gravitational forces,
and now as well of the quantum field theoretical vacuum which may decay.

REFERENCES
Gibbons, G. W. 1975: "Quantized Fields Propagating in Plane-Wave space-times" Comm. math.
Phys. 45, 191.
Hawking, S. W. and Israel, W. (eds.) 1979: "General Relativity", Cambridge.
Held, A. (ed.) t980: "General Relativity and Gravitation", VoL I, New York.
Isham, C. J., Penrose, R., and D. W. Sciama (eds.) 1975: "Quantum Gravity", Oxford.

Quantum Gravity and the Structure of Scientific Revolutions

339

Kanitscheider, B. : "Singularit~iten, Horizonte und das Ende der Zeit", Philosophia naturalis 16,
480.
Kuhn, T. S. 1962: "The Structure of Scientific Revolutions", Chicago.
Kuhn, T. S., 1970a: "The Structure of Scientific Revolutions" (second edition, enlarged)
International Encyclopedia of Unified Science Vol. 2, No. 2, Chicago.
Kuhn, T. S., 1970b: "Reflections on my Critics" in Criticism and the Growth of Knowledge,
I. Lakatos, A. Musgrave (eds.) Cambridge.
L~vy, M. and Deser, S., 1979: "Recent Developments in Gravitation", New York.
Masterman, M., 1970: "The Nature of Paradigm" in Criticism and the Growth of Knowledge,
I. Lakatos, A. Musgrave (eds.), Cambridge.
Popper, K., 1970: "Normal Science and its Dangers" in Criticism and the Growth of Knowledge,
I. Lakatos, A. Musgrave (eds.), Cambridge.
Shrader-Frechette, K, 1977: "Atomism in Crisis: An Analysis of the Current High Energy
Paradigm", Philosophy of Science 44, 409.
Stegmiiller, W., 1976: "The Structure and Dynamics of Theories", New York 1976. Translation of
"Theorienstrukturen und Theoriendynamik" originally published as V.2, pt. 2 of "Probleme
und Resultate der Wissenschaftstheorie und analytischen Philosophic".
Adresse des Autors:
Prof. Dr. Jiirgen Audretsch, Universit~it Konstanz, Postfach 5560, D-7750 Konstanz

Você também pode gostar