3.1: A List of Possible Hypotheses
As I have shown elsewhere,1 all hypotheses about the duration of matter
that anyone has so far been able to put forward comprise a combination of
one of those in the following (A) list with one in the (B) list:
Hypotheses about the Duration of
Matter and Energy in the Past
(Al) All matter and energy have
existed for all time.
(A2) All matter and energy have
existed for approximately the same
length of time, i.e. from the date of
the Creation. According to this hypothesis no particle has existed for a
longer time than has elapsed since the
Creation began nor for a shorter time
than has elapsed since the process of
Creation was completed.
(A3) Any particle of matter or
quantum of energy may have existed
for any length of time. As will be
shown later, this is a way of saying
that matter and energy are originating
without cause, continuously, at random, and not as a result of anything in
the existing state of affairs.
Hypotheses about the Duration of
Matter and Energy in the Future
(Bl) All matter and energy will
continue to exist for all time.
(B2) All matter and energy will
continue to exist for approximately
the same length of time, which is the
time that will elapse until the End of
the World. According to this hypothesis no particle will last for a shorter
time than will elapse up to the beginning of the End of the World nor for a
longer time than will elapse until the
process of destruction is completed.
(B3) Any particle of matter or quantum of energy may cease to exist at any
time. As will be shown later, this is a
way of saying that matter and energy
are disappearing without cause, continuously and by extinction, at random,
and not as a result of anything in the
existing state of affairs.
There are in all nine possible combinations. Three of these are obtained
by combining respectively (Al) with (Bl), (A2) with (B2), and (A3) with
B3). They are symmetrical. The remaining six combinations are all
asymmetrical. As every one of the combinations is just as much an hypothesis
as any other, each has to be assessed by the criteria that are applied to
hypotheses when a choice has to be made between them. One of these is
the criterion of minimum assumption, Occam's razor. Let us first assess
the various possible hypotheses about the past and future duration of
matter by this criterion. Assessment by other criteria is, of course, also
necessary and will follow later.
3.2: The Hypothesis of Continuous Existence in the Past
In this chapter the hypotheses in the (A) list only will be discussed.
Those in the (B) list will be discussed in Chapter IV.
(Al) fails badly by the criterion of minimum assumption. As will be
shown below, it can be maintained only with the help of a number of
assumptions that are far from minimum ones.
A minimum assumption is recognized, it will be remembered, by use
of the word 'any' in its formulation. The statement that all matter and
energy have existed for all time precludes the use of this word and would
seem to bring (Al) into conflict with the criterion of minimum assumption.
But this is not the major objection. I should feel diffident about applying
the criterion thus rigorously. In spite of verbal appearances it is just possible that 'all time' is really less of an assumption than 'any time'. So I
should hesitate to dismiss (Al) simply because its formulation does not
contain the word 'any'.
Irreversible Processes: The true objection is deeper and stronger. It is
that many other very specific assumptions are implicit in (Al). This
becomes apparent as soon as one tries to reconcile this hypothesis with a
simple fact of observation. This is that the world we live in is a changing
one. Let me show why this presents us with a serious difficulty as soon as
we try to support (Al).
Certain observed processes are irreversible. The most familiar of them
is the process of falling. Things fall from a region of higher to one of lower
potential and not vice versa', in colloquial terms things fall downwards and
not upwards. One can describe the process as that of following a potential
gradient and its result is such that the total potential energy is reduced by
the process, it being converted into kinetic energy. As is well known, the
conversion of potential into kinetic energy is equivalent to an increase in
entropy and is a manifestation of the second law of thermo-dynamics.
Another manifestation of the same law is the equipartition of energy
between colliding particles. When a particle that is at rest, but free to
move, is hit by a moving particle, the one at rest acquires kinetic energy
at the expense of the one that hits it. Again, when two particles of irregular
shape collide some of their kinetic energy of rectilinear motion is translated
into that of spin. It is easy to deduce from these facts that, given sufficient
collisions between moving particles, there will be a tendency for those that
began with little kinetic energy to acquire more and for those that began
with much to lose some of it. Eventually, therefore, the average kinetic
energy in any sample large enough to be typical will be the same as in any
other. The temperature of a body is proportional to the average kinetic
energy of its component particles. Two substances in which the particles
have different average kinetic energies are therefore at different temperatures. When they are brought into contact the kinetic energies come to be
shared equally and the temperature difference is smoothed out. This is but
another way of saying that heat flows by conduction from hotter to cooler
bodies and not vice versa.
There are many other irreversible processes. Electric charges tend to
neutralize each other. Chemical reactions tend to be such that the resultant
compounds are less reactionable than the substances from which they were
formed. Exothermic reactions, in which heat is liberated, are more probable and frequent than endothermic ones, in which it is absorbed; and so
Such irreversible processes tend towards a terminal condition. If things
keep on falling, the time must come eventually when everything that can
fall will have done so until it can fall no further. If bodies in motion keep on
interacting, the time must come when the total kinetic energy is shared
equally between all of them. If heat keeps on flowing from hotter to cooler
bodies, temperatures throughout the universe must eventually be equalized;
the world must reach the condition called by German philosophers the
'warmetod' (heat death). Similarly all electric charges must be neutralized
and all chemical compounds reach a state in which they cannot act chemically on each other.
All this can be expressed as a principle that I propose to call the
Principle of Stabilization. By this principle processes in a self-contained
system tend to reduce their causes. If such processes continue for long
enough all causes must be reduced so much that they cease to operate.
When this has happened there can be no more change.
But we do observe change around us. Things do continue to fall. The
ponderable contents of the universe have by no means reached positions
where they can fall no further. About one-half of these contents exist in
the form of a thinly diffused gas in interstellar space. All parts of this gas
are in some gravitational field, for such a field surrounds every star and
stretches into the outer distance, where, though very faint, it exerts a
finite pull on every molecule of the diffuse interstellar gas. So this gas
must be falling very slowly towards the nearest stars. The terminal
condition, in which all of the interstellar gas has completed its falling, is
still a long way off.
Similarly, great differences of temperature are still observed. Heat in
large amounts is flowing from hot to cool surfaces. Interacting bodies are
still far from sharing their kinetic energy equally. Particles in collision still
impart considerable increases in spin to each other. Electrical and magnetic
potential gradients abound. Highly reactionable chemical substances
are found throughout the universe. In short, what makes for change
is to be found in abundance. The hypothesis of continuous existence
in the past (Al) fails to explain this.
(A.I) Can be Saved only by Piling Hypothesis upon Hypothesis: When
an hypothesis fails to explain observed facts one has the choice between
two alternative courses. The first is to abandon the hypothesis in favour
of another one with superior explanatory power. The second is to retain
the hypothesis and to combine it with further ones in such a way that the
combination does explain the observed facts. One could, for instance,
attempt to reconcile the hypothesis of infinite past duration of matter with
the observation of present change by adding sundry additional hypotheses
about the laws of physics.
A collection of carefully reasoned papers on the age of the universe
has appeared in The British Journal for the Philosophy of Science.2 There,
several of the additional hypotheses are mentioned that might serve to
save (Al) and they are critically discussed. I do not want to cover more
of the same ground here than is absolutely necessary, but I do want to
emphasize that all attempts to save (Al) have to be rather desperate and
cannot be reconciled with the way a physicist regards the material
It could be thought that, perhaps, those processes that seem to us to
be irreversible are really cyclical. The expansion of space is one of these;
it has actually been suggested that it may alternate with contraction and
be analogous to the motion of a pendulum, which slows down as it
approaches the limit of its swing and then reverses. In the above-mentioned
collection of papers J. T. Davies says 'it is impossible to say whether an
expansion might not be merely a phase of a more complex phenomenon
such as pulsation'. Ernst J. Opik speaks of 'the collapsing Universe
rebounding from the elastic forces of the nuclear fluid at a state of
maximum compression'. On this hypothesis the moment when any given
irreversible process began is not a starting point but only a turning point.
This hypothesis can sound plausible when it is applied to the expansion
of space. But this expansion is only one of a great number of processes
that are found to be irreversible. If we are to develop a sound cosmology
we dare not overlook any of the others, and when it is applied to other
processes the hypothesis of a pulsating universe can sound plausible only
so long as one thinks of these processes in a vague and abstract way.
If we think of entropy only as an algebraic symbol in a mathematical
expression, we shall find no difficulty in believing that in a self-contained
system its rate of change may sometimes be positive and sometimes
negative. Perhaps, we might then say, the second law of thermo-dynamics
defines cyclical and not unidirectional processes; perhaps during infinite
time a tendency for the entropy of a self-contained system to increase
alternates with a tendency for it to decrease. Perhaps a tendency for
processes to reduce their causes alternates with a tendency for them to
increase their causes.
But when one thinks of the concrete reality that finds its generalization
in the second law, those 'perhapses' fade to an unconvincing pallor. The
law expresses, among other things, the fact that colliding bodies tend
with time to share their kinetic energies equally. Perhaps, one is led to
assume, there have been times when a body that was at rest and was hit
by a moving one imparted kinetic energy to the body that hit it and lost
some of its own (non-existent) kinetic energy. Perhaps spinning bodies
once-upon-a-time gave up their spins on collision and converted their
spin energies into energy of rectilinear motion. Perhaps at one time gravitation was reversed and things fell up instead of down. Perhaps inert
chemical substances used to act on each other and cinders converted
themselves into coal. Perhaps heat used to flow from the cooler to the
hotter surface. Perhaps electrical potential gradients used to grow steeper
when charges were neutralized. Perhaps the laws of physics are not
universal but change from time to time. Perhaps the Cosmic Statute Book
comes up for periodic revision. Perhaps the last occasion was the precise
moment when a pulsating universe had reached its maximum concentration
and began to expand. Perhaps all the present irreversible processes started
together just then. If we attempt to save (Al) we shall run the risk of
drowning in an ocean of strange hypotheses.
Most, if not all, of them are untenable anyhow. That objects of
irregular shape tend to spin when they collide is not the consequence of
a transcendental law but of simple mechanics. It can be easily shown that
there are innumerable ways of colliding that cause spin for one way that
does not. The laws of physics allow the objects to collide in 'any' way and
so nearly all collisions are of the spin-causing kind. The irreversible process
by which rectilinear motion tends to be converted to spin does not arise
from a specific law but from the absence of a specific law. To say that
'perhaps' it was the other way about at one time is to say that perhaps
there was a specific law at that time to control collisions, a law that has
since been struck off the Statute Book.
It can, moreover, be shown that many of the so-called laws that seem
to show processes to be irreversible are merely true by definition. One
defines a field of force in terms of the direction in which things fall in it.
To say that they fall in the opposite direction is to contradict the definition.
One defines temperature gradient similarly in terms of the direction in
which heat flows. To say that it flows from the hotter to the cooler
surface is implicit in the definition. This is why, if we seek to offend
against the rule of Economy of Hypotheses in the forlorn hope of
saving (Al), we shall do so in vain. There would rarely be the need to
reproach us for piling one hypothesis on another. One could refute us by
For the foregoing reasons the notion that the whole existing contents
of the material universe have existed for all time seems, today, to have
but few sponsors among serious scientists, whether the notion be of a
universe in a process of unidirectional change or of pulsation. It fails badly
by any acceptable scientific criterion and certainly by that of minimum
3.3: The Hypothesis of Past Finite Existence
One can explain the fact that there is still change in the universe if
one adopts the hypothesis that all the contents of the physical universe
have existed for a definable time. This is (A2), and at the time of writing it
has considerable vogue. Sundry estimates have even been made concerning
the age of the universe.
Irreversible Processes have been Claimed to Provide Evidence for (A2):
For this purpose a few of the observed irreversible processes have been
used as clocks. It is assumed that the universe started in some specific
condition and that it has been departing at a known rate from this condition ever since. The degree of departure is taken as a measure of the time
that has been occupied by the process.
One of the clocks is provided by the expansion of space. It is assumed
by supporters of (A2) that the whole contents of the universe started in a
small volume of very high mass density and have been spreading out
more and more ever since. By measuring the rate of expansion and the
distance between discrete objects one arrives at the time assumed to have
passed since the expansion began. One of the other clocks is the ratio
of helium to hydrogen in stars; it is assumed, with considerable justification, that all ponderable matter has begun its existence in the form of
hydrogen and that its slow conversion into helium is among the irreversible
processes that give a reliable time scale. Another clock is provided by the
partition of kinetic energy between stars, and yet another one by the
proportion of certain isotopes found in rocks. These and several other
clocks give roughly the same age, namely between four and seven thousand
I do not propose to discuss in detail the observations on which the
various estimates are based. They have been fully described by others
and there is no reason to doubt that the estimated time-interval of some
thousands of millions of years has a cosmic significance of some sort.
It does not therefore seem likely that hypothesis (A2) will have to be
rejected on the grounds that the various clocks provided by irreversible
processes are unreliable; the agreement between them is too good for
their readings to be lightly dismissed.
But we must take care to draw the correct conclusions from those
readings. One may safely conclude that since the clocks were, so to speak,
wound up the part of the material universe where we are has been behaving
as it does today. But this conclusion is equally compatible with (Al),
with (A2), and with (A3), as has been pointed out by sundry writers of the
papers that appeared in the above-mentioned 'Age of the Universe'
number of The British Journal for the Philosophy of Science.
Supporters of the hypothesis that a pulsating universe has been
existing for infinite past time, (Al), could point out that the moment
when the clocks started need not be interpreted as a moment of origin;
it could equally well be a moment of change in a universe that had existed
in a different form before then and had obeyed a different set of laws. The
strange collection of hypotheses that would have to be designed in order
to save (Al), and that has been mentioned already, may be refuted by
various arguments, but it is only fair to add that it cannot be refuted by
the evidence of the clocks. Michael Scriven, one of the contributors to
the papers on the age of the universe referred to above, speaks of 'phases
in the Universe prior to the present "expansion",' as a description of a
possible view, though not one to which he seems personally to commit
himself. He also asks the pertinent question: 'What is the difference
between saying that the Universe has an origin in time and saying that
each of the processes of change within the Universe has an origin in
time ?' The implication is, of course, that the evidence of the clocks might
equally well be used as evidence for the pulsating universe that, according
to (Al), would have existed for all time.
It will later become apparent that the evidence of the clocks could
also be used to support (A3), the hypothesis of continuous origin. But
that can be easily understood only after (A3) has been discussed. Suffice
it at the present moment to refer again to the above-mentioned papers,
where J. T. Davies shows this cogently when he says 'thus the theory of
continuous creation provides us with a picture rather like that of a living
population; individuals are born, age and die, but the population may
remain in a steady state'. What, in other words, the clocks record is not
necessarily the age of the universe as a whole. Each clock may only record
the age of the particular part of the universe in which it is placed.
It will appear in due course that one should expect matter that was
originating continuously, in accordance with (A3), periodically to form
nebulae, like our galaxy. One should expect the various clocks placed in
any one of them to record its age and to give consistent readings between
In short, claims that there is observational evidence for (A2) have, to
say the least, been greatly exaggerated. It is doubtful whether there is any
evidence at all for it that is not also evidence for either of the other
(A2) Fails by the Criterion of Economy of Assumption: Perhaps more
scientific considerations will be found some day in favour of the hypothesis
of a definable time for the past existence of matter than have been' found
hitherto, but while we wait for these we must also assess the hypothesis
by the criterion of minimum assumption. By this it fails at least as badly
as the hypothesis of past infinite existence. This must cause disquiet to any
scientist who is contemplating the hypothesis. To emphasize this disquiet
I have on a previous occasion referred to (A2) as the 'once-upon-a-time
theory'.3 My reason was that, like a fairy-tale, it fails conspicuously to
conform to Occam's razor.
(A2) postulates, for instance, two (and not only one) significant and
quite specific moments in all time. The first is the moment at which it has
to be assumed that the very first element of the universe came into existence. It is the moment before which it is assumed that there was nothing,
The second specific and significant moment of time is the one at which the
very last element of the material universe came into its correct position.
It is the moment before which is is assumed that the universe was incom-
plete, and after which it is assumed that it was complete. These two
moments, it is assumed, have never been repeated in the whole of time,
Two more-specific assumptions are hard to imagine.
It is alien to the spirit of physics to contemplate unrepeated and
unrepeatable events. It is part of the physicist's faith that events are all
repeatable whenever similar conditions are reproduced, and that given
sufficient time similar conditions will always be reproduced. To postulate
two specific moments in time is to reject this. Physicists will be unwilling
to do so without a great deal of evidence.
But this is not even the whole of the objection. Hypothesis (A2)
implies that the laws of physics, as we know them, did not come into
operation until the second of the two significant moments. The implication
of the hypothesis is that there was once in the past a period of finite duration between those two moments, which may have lasted for seconds only ,
or for millions of years, but was in any case finite. From a specific moment
in the year t1 B.C. until another specific moment in the year t2 B.C. it is
implied that the laws of conservation of energy and of matter were suspended, that they came into force precisely at the second of the two
moments, and that they never can be suspended again.
For all that can be proved to the contrary it may have been so. But an
hypothesis that implies it cannot be called a minimum one. The belief
that the laws of physics hold, and have held, and will hold, everywhere
and at all times is a strong one. This is why it demands a substantial
intellectual effort to be persuaded of (A2). But the effort has not to be
directed at understanding something; it has to be directed at ignoring
something, namely the implications of the hypothesis. The period between
the beginning of the Creation and the completion of the universe is one
that does not bear thinking about. To make (A2) acceptable one must
prevent this period from intruding on one's attention. An hypothesis that
calls for this ignoble kind of intellectual effort must be suspect. That it
fails conspicuously by the criterion known as Occam's razor would not
alone be a sufficient reason for rejection of (A2). But it seems also to fail
by all other scientific criteria. The evidence for it is slight and of doubtful
validity. It has but little unifying and explanatory power. If it is to be
justified at all, it can only be on non-scientific grounds.
3.4: The Hypothesis of Continuous Origin
I doubt whether the hypothesis of continuous origin (A3) would ever
have suggested itself if both the other two hypotheses had not proved so
unsatisfactory. It is fair to say that I did not arrive at (A3) by a direct
approach to the problem but by a process of elimination consisting of the
rejection of (Al) and (A2). To be justified, (A3) will have to be free from
the glaring defects of the other two hypotheses. It will have to meet the
Principle of Minimum Assumption adequately, it will have to manifest
greater explanatory power than the other two, and it will need better
support from the evidence of observed facts.
The justification of hypothesis (A3) on the grounds that it involves a
minimum assumption was presented when I first advocated this hypothesis
in 1940. To emphasize this justification I called it the 'at-any-time' theory.
Others have advocated the same hypothesis since and found sundry other
reasons for its justification.4,5,6 That it deserves serious consideration is
therefore by no means a new claim. Its great explanatory and unifying power
will appear in subsequent pages, together with the evidence that supports
it. For the moment I shall simply discuss its main features in a little detail.
What is it that Originates? To the claim that matter is originating
continuously it is natural to ask In what form?' At the present stage of
knowledge concerning the basic constitution of the material universe it is
not possible to answer this question with any precision. One would
naturally not assume that complex structures come into spontaneous
existence completely synthesized. To do so would be to assume that the
process of originating was also coupled with a process of ordering to
specific forms and that would not conform to the Principle of Minimum
Assumption. For this reason none of the advocates of (A3) has suggested
that complete molecules of chemical compounds are among the originating
bodies. Nor has it been suggested that the nuclei of any of the heavier
elements are among them. These are known to be composed of particles
that are more elementary than they, such as protons and neutrons, and
the minimum assumption is that the originating particles are the most
elementary of all. It is, however, impossible to say in the present state of
knowledge what the uttermost elementary particles are.
If all so-called elementary particles had the same mass one might
regard them as truly elementary. But indivisible particles are known with
various inertial masses. It may be that each of them is in some way composite and that the respective masses of the electron, the proton and the
sundry mesons are each the consequence of a synthesis from some even
more elementary components. One should hesitate to regard any of these
particles as the uttermost elementary components of the material universe.
The same objection does not apply to unit charge, which has always
precisely the same value on the electron with the negative sign and on the
proton with the positive sign. There is, however, convertibility between
various elementary particles, such as between neutrons, protons, electrons,
positrons and others. A proton may, for instance, convert into a neutron,
a positron and a neutrino. One is thereby led to doubt whether unit charge
is a quite basic constituent of matter.
It is doubtful even whether the most basic constituents ought to be
called particles. For there is also convertibility between particles and
photons, as there is between mass and energy. In the sense in which protons
and neutrons are common constituents of atomic nuclei some as yet undefined components of matter may be common constituents of positive and
negative charges, elementary masses, and photons. If so, these undefined
components will be no more like the things of which they are components
than neutrons and protons are like the nuclei into which they synthesize.
Bearing in mind that matter is not so much in space as of space, the
most accurate description, and a non-committal one, of the uttermost.
elementary constitutent of matter may be a bit of differentiated space. But
the inseparability of matter and space has not yet become a familiar
concept, so I propose to speak here, even more non-committally, of the
continuous origin of elementary components of the material universe.
That I have to speak thus non-committally proves that the hypothesis
of continuous origin (A3) is incomplete in the form in which I have been
able to state it. In the next chapter I shall develop the hypothesis of the
continuous extinction of matter and this will be equally incomplete. I have
no answer to the important question: 'What is the basic unit that originates
and what is the basic unit that becomes extinct?' The question abounds in
difficulties. Some of these will be discussed in Appendices G and H.
One of the conclusions to be reached in Appendix H is worth mentioning now, for it will help towards understanding what is said in the
main parts of this book. The conclusion to which more than one line of
reasoning leads is that a complete atomic nucleus, together with its electric
charges, is a basic unit and becomes extinct as a whole.
This conclusion is important. If a part of a nucleus were to become
extinct while the remainder continued to exist, this remainder would not
only be radio-active: it would also be the nucleus of another element. But
observation shows that it is not so. Hence the hypothesis of continuous
extinction would have been severely shaken if the line of reasoning followed
in Appendix H had led to the conclusion that only a component part of
a nucleus becomes extinct at any particular moment.
Even if the suggestions put forward tentatively in Appendix H are
corroborated the hypothesis will, however, still be incomplete. Many
difficult questions will remain unanswered. In drawing attention to this.
I should also draw attention to what I have said on page 16 of the Preface.
To say that a theory is incomplete is not the same as to say that it has been
Top of Page
1 The British Journal for the Philosophy of Science, November, 1955, Volume 6, No 23
2The British Journal for the Philosophy of Science, November, 1954, Vol. 5, No. 19.
3R. 0. Kapp, Science versus Materialism, London, 1940, Chapter XXIV
4 V. Hoyle, Stellar Evolution and the Expanding Universe, Nature, 1949, 163, 196
5 H. Bondi and T. Gold. The Steady State Theory of the Expanding Universe.
Monthly Notes Royal Astronomical Society, 1948, 108, 252
6 W. H. McCrea, The Steady State Theory of the Expanding Universe, Endeavour,
1,1950, 9, 3; Relativity Theory and the Creation of Matter, Proceedings of the Royal