When
the Scientist turns Philosopher.
Friedel
Weinert
Abstract
This paper examines how such
fundamental notions as causality and determinism have undergone changes as a
direct result of empirical discoveries. Although such notions are often regarded
as metaphysical, a priori concepts, experimental discoveries at the beginning of
this century - radioactive decay, blackbody radiation and spontaneous emission -
led to a direct questioning of the notions of causality and determinism. The
experimental evidence suggested that these two notions must be separated.
Causality and indeterminism were compatible with the behaviour of
quantum-mechanical systems. The argument also sheds some light on the
Duhem-Quine thesis, since experimental results at the periphery of the
conceptual scheme directly affected conceptions at its very
core.
I.
Ever since Thomas S. Kuhn
pointed out the importance of the history of science for the philosophy of
science, it has become customary for philosophers of science to support their
philosophical considerations by appeal to real-life science. The philosopher
seeks evidence for some general principles about the nature of science often
from some historical case studies. If there is a common territory between
science and philosophy, as many writers have affirmed,[1]
it must also be possible to go from science to philosophy. This is indeed what
some of the greatest scientific minds throughout the centuries have attempted to
do. Their reflections fall into the oldest branches of philosophical thinking:
ontology or the question of what the basic constituents of nature are;
epistemology or the question by which tools the human mind can acquire knowledge
about the external world; ethics or the question of what moral responsibility
scientists have with respect to their discoveries.
In such
contributions, scientists, prompted by the most recent discoveries in their
respective fields, provide interpretations of science and the natural world and
thereby contribute to their understanding. The heartbeat of science is at its
most philosophical rhythm when major conceptual revisions or revolutions are
afoot and scientists feel the need to go beyond the mathematical expressions of
natural processes to reach a level of understanding which assigns some physical
meaning to the mathematical comprehension of the natural world or offers a
re-interpretation of the nature of the scientific enterprise. What is
interesting in this process from a philosophical point of view is that empirical
facts filter through to the conceptual level and bring about changes in the way
the world is conceptualised. ‘Old notions are discarded by new experiences’, as
Max Born once said. The common territory between science and philosophy lies in
this interaction between facts and concepts. In re-interpreting the natural
world in the light of new experiences the scientist becomes an active
participant in the shaping of human views about the surrounding world.
Physicists like Max Planck perceived the ‘confusing amount of new evidence’ as a
threat to the established mechanistic physical worldview. Such reactions reveal
the role of the philosopher-scientist of which scientists are fully aware:
‘History
has shown that science has played a leading part in the development of human
thought.’[2]
In this paper, I shall
briefly consider the role of understanding in the natural sciences and then
investigate how old notions have indeed been discarded by new experiences in a
specific case in the history of science: the impact of quantum theory on the
notions of causality and determinism. The concern of this paper is less with
deep-rooted metaphysical assumptions or presuppositions which may have guided or
misguided the research of scientists and much more with the impact of scientific
discoveries on general conceptions of nature.
II.
Concepts like understanding
and meaning are usually associated with particular aspects of the social
sciences: Social life produces and reproduces symbolic meaning and the social
scientist needs to acquire an understanding of the inherent symbolic meaning
invested in social life by adopting the viewpoint of a passive participant. In
the natural sciences understanding usually means the assignment of representable
physical mechanisms and causal processes to the formal, mathematical aspects of
physical theories as they may be expressed in scientific laws. This is achieved
via the introduction of various kinds of models. The formalism of statistical
mechanics, for instance, is given an interpretation in the kinetic model of
gases. After some hesitation Planck interpreted the quantum of action, h, -
which he had introduced into his distribution law for blackbody radiation- not
as a fictitious entity but as a real physical constant. This interpretation led
to a radical rethinking of the physical worldview and, as we shall see, to a
re-interpretation of fundamental notions with which scientists described the
natural world.
It is
sometimes affirmed that although the metaphysical principles of science can
change they ‘are assumed to be true independently of any scientific
experience.’[3]
The present considerations present a counterexample: Specific scientific
discoveries led to a reshaping of some of the fundamental notions with which
scientists construct a physical worldview. By implication this means that no
scientific experience prior to the discovery had led to a questioning of these
fundamental notions. Hence the status of these notions changes as a function of
scientific experience. It is the contention of this paper that some fundamental,
even metaphysical notions of science have undergone changes as a direct result
of scientific discoveries. This claim runs counter to one part of the
Duhem-Quine thesis, which holds that empirical results only affect the periphery
of the conceptual scheme and do not touch core conceptions at its very centre.
More specifically, the aim of this paper is to show how fundamental scientific
discoveries had an immediate impact on the way scientists interpreted the
structure of the natural world by reference to such fundamental notions as
causality and determinism.
III.
These notions came under
scrutiny after the first experimental successes of the budding quantum theory.
Often these two notions are used interchangeably. This identification of strict
causality with predictive determinism is a feature of classical physics as
eternalised in Laplace’s spirit. For this superhuman demon the whole past,
present and future state of the physical world is stretched out before his very
eyes like a filmstrip. There exists no novelty, no genuine becoming. From the
present frame, all other frames are predictable or retrodictable, given the
knowledge the demon possesses of the laws of physics and the boundary conditions
of the physical universe. Furthermore, one frame causes the next and is caused
by the previous frame. Hence causality and determinism are identical for the
Laplacean demon. However, a central conceptual change which occurred as a result
of experimental evidence from quantum mechanics was the separation of these two
notions. Central figures like Born, de Broglie and to a certain sense Planck
came to hold that causality and indeterminism were
compatible.
To
understand this separation it is important to recall that three new physical
experiences - the phenomenon of radioactivity, the need for quantum
discontinuity and the experience of spontaneous emissions and absorptions -
jointly led to a reflection on quantum behaviour and the appropriateness of the
classical notions of causality and determinism.
(a) The
discovery of the phenomenon of radioactivity (Becquerel, Curie) around the turn
of the century and the establishment of the decay law - - by
Rutherford and Soddy (1900), with its characteristic half-life decay curve, were
amongst the first indications that the causality concept of classical physics
may have to come under scrutiny. But Rutherford was not prone to philosophical
reflections so that it was left to a later generation of physicists to point out
that the discovery of the decay law had wide-ranging philosophical consequences.
The impossibility to determine which particular atom in N will disintegrate
‘seemed to remove causality from a large part of our picture of the physical
world,’ as James Jeans observed. Soon after the discovery of the decay law
further events began to cast serious doubts on the adequacy of the classical
notions of causality and determinism. The major event of these years was (b) the
emergence of the quantum concept in the work of Max Planck. Planck’s
introduction of the constant h into physics - which meant the insertion of
discreteness and discontinuity - was at first an ad hoc manoeuvre to make his
energy distribution law in blackbody radiation compatible with the experimental
evidence (avoidance of ultraviolet catastrophe). To drive home the inevitability
of this new physical constant Planck uses an analogy: Water whipped up by the
wind will slowly be transformed from ordered waves to an unordered calm sea,
from ordered molar energy to unordered molecular energy, when the wind dies
away. It may be suspected that radiant energy in a cavity will also experience a
slow transformation from ordered long infrared waves to unordered short
ultraviolet waves. On the analogy with the water waves the infrared waves are
expected to disappear and be transformed into ultraviolet waves. This is however
not the case: the energy transformation reaches a maximum in the region of
visible light (), and then rapidly falls off towards shorter and longer
wavelengths. In order to account for this behaviour Planck postulated the
hypothesis that energy can only be absorbed and re-emitted in discrete bundles
of energy: quanta or multiples of the constant h. Experiments soon showed that h
was a new fundamental physical constant and this interpretation provided it with
a physical meaning. For a theoretical justification for the distribution law
Planck turned to Boltzmann’s connection between entropy and probability, as
expressed in the equation:
(where W is the
thermodynamic probability and k is the Boltzmann constant; more precisely, W is
a measure of the number of micro-conditions, which are compatible with a given
macro-condition of a physical system in a certain thermodynamic state). In his search for a justification Planck
remained in the realm of statistical mechanics, in which the concept of the
statistical law was of paramount importance. He was struck by the existence of
two different types of laws in physics: the dynamical laws of classical
mechanics which he took to be deterministic causal laws and statistical laws
which reflected lawlike tendencies of ensembles of particles (as for instance in
the decay law). Planck always lamented the co-existence of these two types of
laws in physics because they led to two different kinds of causal connections
between physical states: on the one hand the necessary link between cause and
effect familiar from classical physics, on the other hand the merely probable
link between an ensemble of thermodynamic systems and their evolution towards a
maximum state of entropy.
From the introduction of the
quantum in blackbody radiation Planck was naturally led, through its connection
to statistical laws, to a consideration of causality. As we shall see his
solution was to differ from that of Einstein whose work on statistical mechanics
and quantum theory provided the third new experience (c), which led to a
re-questioning of the concept of causality. Einstein’s deep worry about the fate
of causality in quantum mechanics had its root in a process known as spontaneous
emission. Both the absorption and emission of photons in atomic processes which
govern whether atoms are in a ground state or excited state have a quantum
nature. Although Einstein was able to write down equations which govern these
processes, he was deeply worried by their philosophical implications. He found
it unacceptable, as he wrote to Max Born, to do without ‘complete causality’
even in quantum-like processes like absorption and emission of photons. He found
the idea unbearable that an electron which had received a pulse of light energy
should freely choose the moment and the direction of its escape from the atom.
Admittedly, the knowledge of initial conditions and general laws failed to lead
to precise predictions - a fundamental assumption of classical physics. But this
pragmatic failure of determinism did not particularly worry Einstein. He held
instead that the notion of ‘complete causality’ had a clear sense and he took it
to mean ‘predictive determinism’, which is closely associated with the
differential equations of physics.
Thus there
were at least three new physical experiences which led to a questioning of old
notions. Einstein’s refusal to abandon the idea of strict causality, even though
in practice it may not be possible to adhere to the classical programme of
strict predictability gives a hint of how some of the central figures reacted to
this situation. It is at this juncture that the concepts of causality and
determinism part and go their separate ways. Einstein himself, as Max Born
reminded him, had spoken of the need to constantly check the usefulness and
justification of concepts against experience. Despite popular impressions that
quantum mechanics has spelt the end of causality, quantum mechanics does not
abandon causality. But the classical identification of determinism and
causality, so well captured in the Laplacean spirit, comes to an end. Quantum
mechanical evidence led to a careful distinction between determinism and
causality. It is significant to note that neither Planck, Born, de Broglie nor
even Heisenberg or Bohr gave up the notion of causality. But these central
notions shifted ground in response to the new experimental
evidence.
Determinism is the one
concept which has to be abandoned in quantum mechanics due to the statistical
nature of quantum mechanical experiments and measurements. Determinism even in
classical physics describes an ideal situation: from the knowledge of boundary
conditions and the presence of universal laws it should be possible to make
precise predictions about the future spatio-temporal location of the system
under consideration. Conversely, from the existence of present conditions and
the knowledge of general laws it should be possible to retrodict past initial
conditions. Heisenberg describes this strict causal determinism as the principle
of Newtonian physics. It implies strict predictability and
retrodictability:
Astronomy provides the
paradigmatic example. But even in this case idealising conditions need to be
introduced. The planetary laws, formulated by Kepler and Newton, describe the
behaviour of one planet under the gravitational influence of a central sun, and
abstract from the existence of other planets.
Indeterminism becomes the
important notion for quantum-mechanical systems: it does not stand for a model
of random events at the root of all physical processes. Indeterminism refers
only to the incomplete determination of boundary conditions, as expressed in
Heisenberg’s Indeterminacy Relations for micro-systems. Hence indeterminism does
not mean the absence of lawful regularities, even in micro-systems and hence
does not mean complete unpredictability. It means that the present initial
conditions, either in terms of time and energy or in terms of location and
momentum, of individual particles in micro-systems cannot be sufficiently
determined to predict their precise space-time trajectories. Note that
Heisenberg’s indeterminacy relations express a fundamental indeterminacy at the
level of the behaviour of quantum systems and not merely at the level of human
knowledge about them.
Causality comes to depict,
in the words of Planck, a lawful connection in the temporal succession of
events. Born defines causality as ‘the belief in the existence of mutual
physical dependence of observable situations.’[4]
However, this general characterisation of causality takes us away from a
seemingly sounder classical model of a causal mechanism between two
macro-objects: a bullet penetrates a block of wood and causes the block to move
forward. If the mass and velocity of the bullet and the mass of the wood block
are known, it is perfectly possible to determine the velocity of the wood block
after the impact and to calculate the distance covered, taking into account the
effects of friction. Given the initial conditions of the material objects and
the knowledge of the law of the conservation of momenta, it is possible to tell
a causal story. But this model of mechanistic causality is quite different from
the model of conditional causality to which quantum mechanics appeals. Quantum
mechanics cannot trace the space-time trajectories of atomic particles with
complete determination in terms of classical parameters. The model of
mechanistic causality may have inspired the classical identification of
causality with determinism. But even in the classical world of macro-particles
and mechanical forces, mechanistic causality and determinism did not always
coincide. For instance, using Kepler’s third law,
, it is possible to determine, from the knowledge of the
orbital period of a planet, its average distance from the sun. Yet this is a
functional law: neither of its terms can serve as an antecedent, contiguous
cause on the model of classical causality. What experiments in quantum mechanics
typically show is that a quantum mechanical system can be prepared in a desired
state; the effect of interference with this state can then be measured; but the
measurement only records a statistical result.
Thus we see
emerge two conceptions of causality only the latter one of which, in terms of a
physical dependence of a set of posterior conditions on a set of antecedent
prior conditions, is adequate for quantum-mechanical systems. We have seen that
the mechanistic model of causality fails in the world of quantum mechanics
because of the ubiquity of the indeterminacy relations. But strictly speaking,
inexactness in the measurement of initial conditions and idealisations of laws
is also a feature of classical physics. There must therefore be a deeper reason
for the inapplicability of the model of mechanistic causality to the quantum
world. At this point Max Planck‘s philosophical instinct pointed to the deeper
level of ontology to fathom the reason why mechanical causality fails in quantum
mechanics: the quantum physicist needs to question the very ontology of the
classical world. The blame is to be laid at the door of the model of the
classical particle. Planck and other quantum physicists called for the
abandonment of the notion of the material point particle. It is this fundamental
Newtonian ontology - the corpuscular nature of the world, governed by
mechanistic laws of motion in a homogeneous space - which inspired mechanical
causality. But when quantum mechanics showed that the tracing of causal
mechanisms between events, required by mechanistic causality, failed for quantum
events, the classical identification of determinism and causality inspired the
erroneous picture of an indeterministic and acausal microworld. In fact it is
the very ontology of classical mechanics which fails to be supported by the
evidence of quantum mechanics. And for the physicist the nature of reality is
determined by the nature of the evidence.
The
decisive input to the new world picture consistent with quantum mechanics was
provided by Max Born with his probabilistic interpretation of quantum mechanics:
The motion
of particles follows the laws of probability, but the probability itself
propagates in accordance with the causal law. (...) This means that the
knowledge of the state in all respects in a particular moment determines the
distribution of the state in all later times.[5]
This is just the notion of
conditional causality understood as physical dependence between physical
parameters. The dependence sought now is not that between single causes and
effects but between a set of antecedent conditions of a whole quantum system and
the probability of occurrence of consequent conditions. The probability of the
consequent conditions of a quantum system is expressed in the Born rule: . Note that this conditional notion of causality is
compatible with the indeterminism of the quantum world. In the famous two-slit
experiment it is not possible to determine through which slit an electron passes
without disturbing its momentum; but this disturbance destroys its
characteristic coherence pattern. Any interference with the electron destroys
information needed for the precise determination of its trajectory, and hence
destroys its coherence effects. By contrast non-interference with the electron
in the two-slit experiment results in its characteristic coherence patterns but
at the cost of knowledge about its trajectory. The point is, however, that the
very acts of interference and non-interference, which either destroy or create
the intensity patterns, are causal processes, which lead to observable effects.
Hence in quantum mechanics we still have conditional causality, understood as
physical dependence of effect events on cause events, but no longer classical
determinism. The task of philosophy will be to overcome these two different
conceptions of causality and find a model of causality, which fits both
classical and quantum mechanics.
IV.
This paper has attempted a
brief discussion of the complex connection between science and philosophy.[6]
The principal theme has been that fundamental notions like causality and
determinism, with which scientists construct their worldviews persist, under the
tension of ongoing experimental research and undergo change as a direct result
of new empirical discoveries.
Department of Social
Sciences and Humanities
University of
Bradford
West Yorkshire BD7 1DP
UK
Email:
f.weinert@brad.ac.uk
[1] This is reflected in many books titles: J. Herschel, On the Study of Natural Philosophy (1830); J.H.Fr. Papillon, Histoire de la Philosophie moderne dans ses rapports avec le développement des sciences de la Nature (1876); H. Dingle, Through Science to Philosophy (1937); A.S. Eddington, The Philosophy of Physical Science (1938); J. Jeans, Physics and Philosophy (1943); M. Born, Natural Philosophy of Cause and Chance (1949); W. Heisenberg, Physik und Philosophie (1959). For the sake of brevity only references to direct quotes are given.
[2] M. Born, Natural Philosophy of Cause and Chance (Oxford 1949), p. 2
[3] C. Dilworth, The Metaphysics of Science (Dordrecht 1996), p. 71; italics in original.
[4] Natural Philosophy, p. 124; cf. p. 9
[5] ‘Quantenmechanik der Stobvorgänge.’ Zeitschrift für Physik 38 (1926), p. 54. This characterization is repeated almost verbatim in his Natural Philosophy of Cause and Chance, p. 103
[6] This essay is a revised and updated version of a paper first presented at the 20th World Congress of Philosophy, Boston 1998, available at http://www.bu.edu/wcp/Papers/Scie/ScieWein.htm. The ideas formulated in this paper are discussed at much greater depth in the author’s full-length study The Scientist as Philosopher (Heidelberg/Berlin/New York: Springer 2004).