The Philosophy

of Sir Arthur Eddington

Thesis presented to the

Cardinal Mercier Institute of Louvain
for the obtention of
the doctorate in philosophy

‫ﱮﱭ‬
Charles De Koninck
1934
 pa r t o n e

Some Theses in the Philosophy of Science

 Introduction

Sir Arthur Eddington was born in Kendal, England, on 28 December 1882.

His father was the principal of the Quaker School at Stramongate, the school
in which John Dalton had been sub-master. Eddington himself is a fervent
Quaker and, as we shall see, his religion has had a profound influence on his
philosophy. He received his higher education in Manchester first, at Owen
College, and then at Trinity College, Cambridge. It was there that he was the
“senior wrangler” in 1904, a title given the student who excels in mathemat-
ics and literature equally. His writings have an exceptional literary value. His
knowledge of classical and modern literature is astonishing. In 1907 he was
given the Smith prize from his college and in the following year was elected
a Fellow of that celebrated institution.
His first scientific writings were published while he was assistant to the
head of the Greenwich Observatory (1906 ‒‒ 1913). To that point he seemed
destined to have a career as a professional astronomer. But in 1913 he was
named Professor of Astronomy in the Plumian Chair at the University of
Cambridge, a chair that he still occupies today.
In 1914, he was put in charge of the Cambridge Observatory and in that
same year was named a Fellow of the Royal Society. He is a member of many
scientific societies, in England and abroad. He has been awarded prizes and
medals everywhere, too many to enumerate here. In 1930, he was knighted
and acquired the title Sir.1
In 1920 another Eddington showed himself with the publication of a
work which marked an epoch in the philosophy of science: Space, Time, and
Gravitation. This work demarked for the first time the domains of science
and of philosophy in what concerns the theory of relativity. With this work
he became the leader of philosophers of science. And he inspired con-
fidence, because he was himself one of the most celebrated representatives
of modern science. With him, there could be no question of preconceived
ideas.

103
 104 | Charles De Koninck

It has been said that the work is merely one of popularization. It is much
more than that. It is first of all a philosophical contribution to the problem
of modern science. His contributions have not been merely negative, as has
been the case with many other modern scientists, Einstein included, who are
content to chase philosophers from their domain—very rightly, moreover.
In the work, The Analysis of Matter, Lord Bertrand Russell writes: “(Edding-
ton), better than Einstein or Weyl, has the theory (of relativity) in the form
best adapted to the needs of philosophers. . . . In philosophy, I have been
guided almost exclusively by Eddington.”2
The Nature of the Physical World, the Gifford Lectures of 1927, were an
even more important revelation than the preceding work. The demands of
these lectures in Edinburgh obliged him to deal with transcendental prob-
lems, something he did with extraordinary skill. This work assured him of
a place among the most distinguished contemporary philosophers. He has
been sharply criticized but today everyone must take him into account.
For us, Eddington has a double interest. First, he has carefully delimited
the philosophical problem of relativity and quantum indetermination. But
he has also given us a rather complete metaphysical system. He calls himself
an idealist, an interesting point, because he says that he was carried toward
idealism by his scientific work.3
The problem of the relation between scientific knowledge (in the re-
stricted sense) and metaphysical knowledge interests us profoundly, espe-
cially because this problem is presented in a rather embroiled way by our own
philosophers. Perhaps Eddington, who knows the sciences in depth, and who
is no less a philosopher, offers us a chance to make the relation precise.
 section one

Epistemology and Metaphysics

Let us review the chief original contributions of Eddington to science in the

experimental domain as well as in the domain of theory. This will enable us
better to situate his personality. For it must be acknowledged that Eddington
is above all a physicist who does philosophy only on the side—“a little,” in his
estimation! His competence in the scientific domain make his extra-scientific
speculations all the more interesting in that he is capable of envisaging these
problems from within physics. His assertion that we must go beyond physics
is therefore all the more grounded. We cannot overlook this point. He is, very
exceptionally, a great physicist and a great philosopher. These two qualities are
inseparably united in his very personality, and we would distort it if we ne-
glected either the one or the other.

I. The Double Stellar Current [Le double courant stellaire]

It was in 1906 that he attracted the attention of the astronomical world on

the occasion of his first presentation to the Royal Society, entitled “The Sys-
tematic Motions of the Stars.”4 In this study he takes up a hypothesis ad-
vanced by the celebrated Dutch astronomer Kapteyn in 1904, an hypothesis
which held that the movements of the ensemble of stars is not fortuitous,
as the current opinion would have it, but that there are two favored move-
ments. Until then, this hypothesis had had no success. Here Eddington al-
ready displays “the feeling of being on the right path before any proof,” of
which he often speaks.5 He allows himself to be first guided by the aesthetic
appearance of a theory.
This is how this problem should be situated. Since the time of Herschel
the hypothesis has been suggested that the galactic system has the form of
an extra-galactic nebula. This hypothesis extends only to the static form of
the system. The known movements were considered to be proper to each

105
 106 | Charles De Koninck

star taken individually, whereas the movements of the whole were fortuitous.
It was then that Kapteyn showed that the stars seemed to follow two opposed
directions in the galactic plan: the movement of stars in the neighborhood of
the sun would follow two opposed courses: there are two systems, one in a
movement opposed to the other.
Eddington wanted to verify this hypothesis by an independent investi-
gation. The rotary movement of the nebula was known. The demonstration
of this hypothesis would show that the galactic system is itself a dynamic
system, just like the extra-galactic nebula. This demonstration would pro-
vide a point of reference for new studies of the origin of this system and of
its situation in the whole universe.
His investigation took from eight to nine years. In 1914, we find these
results synthesized in a volume titled Stellar Movements and the Structure
of the Universe. The conclusion was affirmative. It was based first of all on a
minute analysis of the Catalogue of Professor Boss which gave the proper
movements of about six thousand stars. Applying the statistical method,
Eddington showed that it is necessary to divide them into two systems. He
found a second confirmation of the hypothesis of Kapteyn in the spectro-
scopic speed of stars.
By this quantitative demonstration of the Hollendale hypothesis, the
name of Eddington would remain associated with this branch of sidereal
astronomy. This work taken in its entirety marked, according to the expres-
sion of de Sitter, an advance in the history of astronomy.6

II. The Internal Constitution of Stars

The first attempt to classify stars according to their spectra was made by the
Italian Jesuit Secchi (1818 ‒ 1878). He divided the life of a star into four stages
which marked a greater and greater decrease of temperature. Then Lane (in
1878) showed that when a gaseous body contracts in losing heat, its temper-
ature increases. It was therefore equally possible that stars having a reddish
[rougentre] spectrum, and thus belonging to a lower stage, increased in tem-
perature while losing heat. Therefore they could be, contrary to the view of
Secchi, in their youth. In 1913 H.N. Russell and E. Hertzsprung showed that
it was necessary to place stars in two categories: giants and dwarfs. The first
 The Philosophy of Sir Arthur Eddington | 107

are of great size, but of weak density, whereas the second are shrunk, more
dense, and losing temperature. Thus each star passes twice through the same
degree of temperature.
Eddington’s study, “On the Radiative Equilibrium of the Stars,”7 gave a
theoretical explanation of these observations. The investigations were con-
ducted on the basis of the laws of a perfect gas. They were thus restricted
to the study of giant stars, the dwarfs being thought to be too dense to obey
these laws.
There are therefore in a star two opposed tendencies: the upper strata
cooling by contracting, whereas the lower strata increase in pressure. But,
observation shows us that the structure of stars is relatively constant: so two
tendencies of equilibrium must be recognized.
The mass and superficial temperature of the sun are two known magni-
tudes. Eddington asked himself if there were not a way to explain this equi-
librium by a theory based on the known laws of the internal structure of stars.
Direct calculations made on known magnitudes lead to impossible results.

III. The Theory of Relativity Generalized

Even before taking up the problem of stellar equilibrium, Eddington had

been much interested in the theory of relativity. Perhaps it was because of this
that he advanced his theory of the pressure of radiation with such confidence.
Already in February 1915, therefore before he knew the developments
achieved by Einstein, he had written,

It is scarcely too much to say that the nature of gravitation remains as

much a mystery to-day as when the law was first formulated by Sir Isaac
Newton. In the meantime, theories of matter, of aether, and of elec-
tricity have arisen, have held their vogue, and have been superseded by
others; but gravitation stands apart from these changing views. No
experiments have as yet shown any relation between it and the other
phenomena of nature; the simple law, unconditional and universal,
has been all-sufficient hitherto. We have grown accustomed to regard-
ing gravitation as something outside the scope of ordinary physical the-
ories. If a new model of the atom is put forward, we ask if it accounts for
 108 | Charles De Koninck

the Zeeman effect, for chemical affinity, for the dispersion of light, and
a host of incidental phenomena; but it would be considered unfair to
suggest that it ought to account for the one fundamental and universal
property of matter—gravitation.8

Then, after rejecting explanations given up to that time, he suggested: “Does

gravitation conform to the Principle of Relativity?”
There is an interesting remark in this same article. The possibility of
finding an experimental phenomenon in favor of an application of this prin-
ciple to gravity catches his attention. “It would be extremely difficult to de-
tect this deflection [of light] even during a total eclipse. . . . But a positive re-
sult would mean that gravitation has been pulled down from its pedestal,
and ceases to stand aloof from the other interrelated forces of nature.”9
It was Eddington himself who, on 1 February 1918, drew the attention of
the members of the Royal Society to the possibility of exploiting the total
eclipse of 29 May 1919. In articles and conferences he spread the idea. And
all this during the war, for which, being a Quaker, he had a supreme disdain.
It was under his direction that the two expeditions were made, one to So-
bral, and the other to the Ile de Prince. The result was positive. This was an
experimental proof in favor of the theory of Einstein. The first communica-
tion of these results was made to the Royal Society in the Philosophical Trans-
actions of the Royal Society of London (220, series A [1920]: 291 ‒‒ 333).
It was from the enthusiasm of this success that was born the work Space,
Time, and Gravitation. He no longer hesitated. The whole basis of physics
must be profoundly rethought. The philosophical conceptions of physi-
cists are altered by these demands. There remain only philosophical preju-
dices and Eddington gave himself the task of reversing them.
Up until now, his part in the development of the theory of relativity was
purely experimental. It was Eddington, and Eddington alone, who presented
Einstein to the English public. But he also played a role in the theoretical
development itself. This contribution appeared in 1921 under the title “A
Generalization of Weyl’s Theory of the Electro-magnetic and Gravitational
Fields,”10 and was incorporated into the synthesis, Mathematical Theory of
Relativity.11 For Einstein, mass and the quantity of motion were equal to cer-
tain characteristics of the geometry of space-time. But Eddington had shown
that mass and quantity of motion are not the names under which we recog-
 The Philosophy of Sir Arthur Eddington | 109

nize these geometrical characteristics in physical experiments. Einstein in-

voked a law of nature which is cause of this equality, whereas Eddington
reduces this putative law to a tautology. It is here that he showed his skill
in scientific methodology and made a rigorous application of this theory of
physical theory, which always contains two aspects.12 The mathematical
theory of relativity is above all an analysis which operates on certain sym-
bols. When we apply this theory to the physical world, we should find the
connection by the identification of properly physical entities with the enti-
ties of the mathematical analysis. These constitute there two parallel sys-
tems which ought to be identified at each point. But there must be also some
identities which are fundamental and from which one can deduce the others.
Without that, we will have no point of reference. It is on the basis of the iden-
tity of the mass and of the quantity of motion that one can identify the other
parallel entities.
There are then three domains in which he excelled. It was Eddington
who was among the first to be captivated by the theory of quanta, and who
made the first applications of it in the domain of astrophysics. It was Ed-
dington who discovered Abbé Lemaitre, who moreover had been one of his
students. To trace a precise and adequate portrait of Eddington the physicist
is quite outside my competence, but these few remarks perhaps suffice to con-
vince us that we have to do with a flexible spirit, youthful, who has an excep-
tional flair for the truth, and to introduce the physicist whose philosophy
we will study.
 s e c t i o n t wo

The Philosophy of Exact Science

In the second and third sections we propose to give as objective as possible

an exposition of the doctrine contained in the different philosophical works
of our author. Is this possible? Let us from the outset introduce necessary
reservations. We have, first of all, our own philosophy, and I think it is diffi-
cult to abstract from it. When a non-scholastic author speaks of “reality,”“ac-
tuality,” and of “substance,” and so on, and does not give us a sufficiently pre-
cise definition of these terms, we are inclined to assimilate the meaning of
these expressions to our own. And in the case of Eddington who does not
pretend to give us a well-defined and achieved system, the danger is consid-
erable. It is indeed possible that here and there I force the matter. Sometimes
he comes remarkably close to a metaphysics akin to ours and then goes in a
direction in which he would refuse to follow us. The doctrine treated here
was dispersed almost everywhere in a sketchy manner. The very wish to sys-
tematize ideas not always firmly delineated runs serious risks, and we will
not be able to systematize them without tightening them. All the same, we
cannot simply juxtapose his ideas and make a dictionary. But we will refrain
from tightening them too much. By seeking to establish an explicit conti-
nuity among these elements, we are doing something that Eddington has
not done for us. One further remark. We cannot study all the points that Ed-
dington calls “philosophical.” The meaning of this term is very ambiguous
among the English, as we shall soon see. Even problems to which scholastics
have given a philosophical value will not occupy us. Such as the problem of
the end of the universe based on the second fundamental law of thermo-
dynamics.13 This physical law, whatever its real value, cannot be made into a
philosophical truth. The philosopher can have his own reasons for predict-
ing the end of the universe, but these will never be assimilable by the physicist.
A theory can be scientifically true, but that does not give it a value, or rather
a meaning, that is philosophical.

110
 The Philosophy of Sir Arthur Eddington | 111

The restriction that we impose on ourselves is not arbitrary. We find its

basis in distinctions made by Eddington himself. Thus, the distinction he
makes between “the point of view” of relativity and the “principle” of rela-
tivity. The first is clearly philosophical. But Eddington does not explicitly
characterize it as such. And the same is true for the problem of indeterminism.

Chapter I. The Sources

Space, Time, and Gravitation is the first philosophical work of Eddington.

He wrote it as a result of the famous expedition to Sobral and to the Isle of
Prince that he himself organized. Einstein’s theories had been received with a
great deal of skepticism by English physicists and philosophers. Eddington
perceived that the prejudices were philosophical and not of the scientific
order. And it seems to us that he succeeded wonderfully in dissipating them.
The Nature of the Physical World is the series of Gifford Lectures that he
delivered in Edinburgh in 1927. Their subject was dictated by the founda-
tion. It is thanks to these requirements that we have a view of the whole on
his philosophy. The conferences that he gave before these appeared contain
nothing that is not assimilated in them.
Science and the Unseen World is a lecture as well, one he gave to a reli-
gious society, the Society of Friends, better known as the Quakers. They con-
tain important precisions on the problem of God and on the religious prob-
lem generally.
Since then we have from him two important philosophical articles. The
first, “Physics and Philosophy,” appeared in Philosophy. This was a lecture
given before the British Philosophical Association. The second appeared in
the Collection des Actualités Scientifiques et Industrielles. Both are above all
precisions on the problem of indeterminism. Two other works, less philo-
sophical, but which contain many important precisions, have been used:
above all, Mathematical Theory of Relativity (1923), an elaboration of the
appendix to the French translation of Space, Time, and Gravitation, which
was exclusively mathematical;14 second, the little book, The Expanding Uni-
verse (1933), which contains interesting remarks on physical theory.
I do not think that one can find any evolution in Eddington’s ideas. The
extensions and clarifications of the more recent contributions are all in
 112 | Charles De Koninck

direct continuity with the first. I will not hesitate, accordingly, to assemble
texts from different works without regard to their dates.
As for the historical and ideological sources of Eddington’s philosophy
itself, it seems to me that it would be difficult to trace them. He certainly
knows Hegel and Berkeley.15 And yet his idealism (we will soon see how it
is with his idealism) has nothing specifically in common with these two
philosophers. He even has a rather naive conception of the subjectivism of
German idealist philosophers. He also knows Bergson. We find in him dis-
tinctions to be found in Bergson as well, such as the distinction between re-
alism and idealism.16 His terminology and the way he poses transcendental
problems recalls the English philosopher Bradley and his disciple Bosanquet.
But the indications given will not suffice to establish a true dependence.
Moreover, the conclusions differ radically.
The one source we can determine with certitude is his religion. Ed-
dington is above all a religious man, and of the species Quaker. We have a
special chapter on this subject in which we will try to determine to what
degree it exercised an influence on his philosophy. His colleague, Sir James
Jeans, defends ideas very analogous to those of Eddington, but much less
precisely. In any case, the philosophical works of Jeans are later than those
of Eddington.17

Chapter II

Article 1. The Notion of Philosophy

Before raising the special problem of his philosophy of science, it will be in-
teresting to know what Eddington himself means by “philosophy.” Unhap-
pily, he does not give us any sufficient indication. Occasionally, he mentions
some of its properties or functions.
There is first of all the text in Space, Time, and Gravitation which attrib-
utes a proper point of departure to philosophy, “physical theory starts with
the simplest constituents, philosophical theory with the most familiar con-
stituents” (166 [205]). The distinction is clear enough. But one must read
it in its context. It is a question notably of a distinction introduced when
speaking of the materialist philosophy of Mach. But the thesis envisaged has