The British Society for the Philosophy of Science

A Note on Berkeley as Precursor of Mach

Author(s): K. R. Popper
Source: The British Journal for the Philosophy of Science, Vol. 4, No. 13, George Berkeley
Bicentenary (May, 1953), pp. 26-36
Published by: Oxford University Press on behalf of The British Society for the Philosophy of Science
Stable URL: http://www.jstor.org/stable/686043 .
Accessed: 26/03/2013 16:09

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Oxford University Press and The British Society for the Philosophy of Science are collaborating with JSTOR to
digitize, preserve and extend access to The British Journal for the Philosophy of Science.

http://www.jstor.org

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 A NOTE ON BERKELEYAS PRECURSOR
OF MACH
K. R. POPPER

THE purpose of this note is to give a list of those ideas of Berkeley's

in the field of the philosophy of physics which have a strikingly new
look. They are mainly ideas which were rediscovered and re-
introduced into the discussion of modern physics by Ernst Mach
and Heinrich Hertz, and by a number of philosophersand physicists,
some of them influenced by Mach, such as BertrandRussell, Philip
Frank,Richardvon Mises, Moritz Schlick,1Werner Heisenberg, and
others.
I may, perhaps,say at once that I do not agree with most of these
positivistic views. I admire Berkeley without agreeing with him.
But criticism of Berkeley is not the purpose of this note, and will
be confined to some very brief and incomplete remarksin section 5.
Berkeley wrote only one work, De Motu, devoted exclusively to
the philosophy of physical science, but there are passagesin many of
his other works 2 in which similar ideas, and supplementary ones, are
represented.
The core of Berkeley'sideas on the philosophy of science is in his
criticismofNewton'sdynamics. (Newton's mathematicswere criticised
by Berkeley in The Analyst and its two sequels.) Berkeley had a
great admirationfor Newton, and no doubt realisedthat there could
have been no worthier object for his criticism.

The following twenty-one theses are not always expressed in

Berkeley's terminology; their order is not connected with the order
1 Schlick, under the influence of Wittgenstein, suggested an instrumentalist.
interpretationof universal laws practicallyequivalent to Berkeley's 'mathematical
hypotheses'.
21 shall refer, apartfrom DM (= De Motu, 1721), to TV (= Essaytowardsa New
Theoryof Vision, 1709) ; Pr (= Treatiseconcerningthe Principlesof HumanKnow-
ledge, 17Io) ; HP (= Three DialoguesbetweenHylas and Philonous, 1713) ; Alc
(= Alciphron,1732) and S (= Siris, 1744).
26

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 BERKELEY AS PRECURSOR OF MACH
in which they appearin Berkeley'swritings, or in which they might
be presentedin a systematictreatmentof Berkeley'sthought.
For a motto, I open my list by a quotation from Berkeley (DM
29).
(i) ' To utter a word and mean nothing by it is unworthy of a philo-
sopher.'
(2) The words 'absolute space' and 'absolute time' are without
any empirical (or operational) meaning; Newton's doctrine of
absolute space and absolute time must therefore be discardedas a
physical theory. (Cf. Pr 97, 99, 116; DM 53, 55, 62; S 271:
'Concerning absolute space, that phantom of the mechanic and
geometrical philosophers,it may suffice to observe that it is neither
perceived by our sense, nor proved by our reason . . .'; DM
64: 'for. . . the purpose of the philosophers of
mechanics..
it suffices to replace their "absolute space" by a relative space
determined by the heavens of the fixed stars . . Motion and
rest defined by this relative space can be conveniently used instead
of the absolutes. .. .')
(3) The same holds for the word 'absolute motion'. The
principle that all motion is relative can be establishedby considering
the meaning of'motion', and also upon operationalistarguments.
(Cf. Pr as above, 58, 115 : 'To denominate a body " moved" it is
requisite . . . that it changes its distance or situation with regard to
some other body . . .'; DM 63 : 'No motion can be discerned or
measured,except with the help of sensible things'; DM 62: '...
the motion of a stone in a sling or of water in a whirled bucket cannot
be called truly circular motion . . . by those who define [motion]
with the help of absolute space. .. ')
(4) The word 'force' is misused in physics; to introduce force
as the cause or 'principle' of motion (or of an acceleration)is to
introduce an occult quality. (DM 1-4, and esp. 5, Io, II, 17, 22, 28 ;
Alc vii, 9; DM 6: 'It is plain, then, that it is useless to proffer
gravity or force as the principle of motion; for how can this prin-
ciple be any better known by appealing to what is [commonly]
called an occult quality ? That which is itself occult explainsnothing.
It goes without saying that an unknown acting causeis more properly
called a [metaphysical]substancerather than a quality.')
(5) In view of these considerations,Newton's theory cannot be
accepted as an explanationwhich is truly causal,i.e. based on natural
causes. The view that gravity causallyexplainsthe motion of bodies
27

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 K. R. POPPER

(that of the planets, of free-falling bodies, etc.), or that Newton

discovered that gravity or attraction is 'an essential quality' (Pr 106)
whose inherence in the essence or nature of bodies explains the
laws of their motion, must be discarded (S 234; see also S 246, last
sentence). But it must be admitted that Newton's theory leads to the
correctresults(DM 39, 4I). To understandthis, 'it is of the greatest
importance . . . to distinguish between mathematicalhypotheses and
the natures [or essences] of things . . . If we observe this distinction,
then all the famous theorems of mechanicalphilosophy which . . .
make it possibleto subjectthe world system [i.e. the solar system] to
human calculations,may be preserved; and at the same time, the
study of motion will be freed from a thousand pointless trivialities
and subtleties,and from [meaningless]abstractideas' (DM 66).
(6) In physics (mechanicalphilosophy), there is no causalexplana-
tion (cf. S 23I), i.e. no explanationbased upon the discovery of the
hidden nature or essence of things (Pr 25). '. . . real efficient
causes of the motion . . . of bodies do not in any way belong to
the field of mechanics or of experimental science. Nor can they
throw any light on these . . .' (DM 41).
(7) The reason is, simply, that physical things have no secret
or hidden, 'true or real nature', no 'real essence', no 'internal
qualities' (Pr IoI).
(8) There is nothing physicalbehindthe physicalbodies, no occult
physical reality. Everythingis surface,as it were ; physical bodies are
nothing but their qualities. Theirappearance is theirreality(Pr 87, 88).
(9) The province of the scientist (of the 'mechanical philo-
sopher') is the discovery, 'by experimentand reasoning' (S 234), of
Laws of Nature,that is to say, of the regularitiesand uniformitiesof
naturalphenomena.
(Io) The Laws of Nature are, in fact, regularitiesor similaritiesor
analogies (Pr io5) in the perceived motions of physical bodies (S
234) '. . . these we learn from
experience' (Pr 30) ; they are ob-
served, or inferred from observations (Pr 30, 62; S 228, 264).
(I I) 'Once the Laws of Nature have been formed, it becomes
the task of the philosopherto show of each phenomenon that it is in
conformity with these laws, that is, necessarilyfollows from these
principles.' (DM 37; cf. Pr Io7; and S 231: 'their [i.e. the
" mechanical philosophers'"] province being . . . to account for
particularphenomena by reducing them under, and shewing their
conformity to, such general rules.')
28

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 BERKELEY AS PRECURSOR OF MACH

(12) This process may be called, if we like, 'explanation' (even

causal explanation'), so long as we distinguishit clearly from the
truly causal (i.e. metaphysical) explanation based upon the true
nature or essence of things. (S 231; DM 37: 'A thing may be
said to be mechanically explained if it is reduced to those most simple
and universal principles' (i.e. 'the primary laws of motion which
have been proved by experiments .. .' DM 36) 'and proved, by
accuratereasoning,to be in agreementand connection with them..
This means to explainand solve the phenomena, and to assign them
their cause .. .' This terminology is admissible (cf. DM 71) but
it must not mislead us. We must always clearly distinguish (cf.
DM 72) between an 'essentialist' explanation with appealsto the
nature of things and a 'descriptive' explanation which appeals to
a Law of Nature, i.e. to the description of an observed regularity.
Of these two kinds of explanation, only the latter is admissible in
physical science.
(13) From both of these, we must now distinguisha third kind of
'explanation '--an explanation which appeals to mathematical hypo-
theses. A mathematicalhypothesis may be describedas a procedure
for calculatingcertainresults. It is a mere formalism,a mathematical
tool or instrument, comparable to a calculating machine. It is
judged merely by its efficiency. It is not only admissible,but it may
be admirablyuseful. But, as opposed to the explanationby essences
(which, in mechanics,are simply false) and to that by laws of nature
(which are, if the laws 'have been proved by experiment', simply
true), the question of the truthof a mathematicalhypothesisdoes not
arise-only that of its usefulnessas a calculating
tool.
(14) Now, those principles of the Newtonian theory which 'have
been proved by experiment'-those of the laws of motion which
simply describethe observableregularitiesof the motion of bodies-
are true. But the part of the theory involving the concepts which
have been criticised above-absolute space, absolute motion, force,
attraction,gravity-is not true, since these are 'mathematical hypo-
theses'. As such, they should not, however, be rejected, if they
work well (as in the case of force, attraction, gravity). Absolute
space and absolute motion have to be rejected because they do not
work (they are to be replaced by the system of fixed stars, and
1 The term 'essentialist'(and 'essentialism')is not Berkeley'sbut was intro-
ducedby me in 'The Povertyof Historicism',Economica, N.S. II, 1944, and in
my OpenSocietyandIts Enemies.
29

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 K. R. POPPER

motion relative to it). '" Force ", " gravity ", "attraction ",1 and
words such as these are useful for purposesof reasoningand for com-
putationsof motions and of moving bodies; but they do not help us
to understandthe simple nature of motion itself, nor do they serve
to designate so many distinct qualities .... As far as attraction is
concernedit is clear that it was not introducedby Newton as a true
physical quality but merely as a mathematical hypothesis' (DM 17).1
'But what is said of forces residing in bodies, whether attractingor
repelling, is to be regarded only as a mathematicalhypothesis, and
not as anything really existing in nature.' (S 234; cf. DM 18, 39
and esp. Alc vii, 9.)
(I5) Properly understood, a mathematical hypothesis does not
claim that anything exists in naturewhich correspondsto it-neither
to the words or terms with which it operates,nor to the functional
dependencieswhich it appearsto assert. It erects, as it were, a fic-
titious mathematicalworld behind that of appearance,but without
the claim that this world exists. It claims only that from its assump-
tions, the correct consequencescan be drawn. But it can be easily
misinterpretedas claiming more; that is to say, as claiming to de-
scribe a real world behind the world of appearance. But such a
world cannotbe described; for such a descriptionwould be meaning-
less.
(16) It can be seen from this that the same appearancesmay be
successfullycalculatedfrom more than one mathematicalhypothesis,
and that two mathematicalhypotheses which yield the same results
concerning the calculatedappearancesmay not only differ,but even
contradict each other (especially if they are misinterpretedas de-
scribing a world of essences behind the world of appearances);
nevertheless,there may be nothing to choose between them. 'The
foremost of men proffer . . . many different doctrines, and even
opposite doctrines, and yet their conclusions [i.e. their calculated
results] attain the truth . . . Newton and Torricelli seem to disagree
with one another, . . but the thing is well enough explained by
both. For all forces attributed to bodies are merely mathematical
hypotheses . ..; thus the same thing may be explainedin different
ways' (DM 67).
(17) The analysis of Newton's theory thus yields the following
results:

1 The italics in the Latin original appearto function here as quotes.

30

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 BERKELEY AS PRECURSOR OF MACH
We must distinguish
(a) Observationsof concrete,particularthings.
(b) Laws of Nature which are either observationsof regularities
or (DM 36) proved (comprobatae) by experiments,or (Pr Io7)
discovered 'by a diligent observation of the phenomena '
(c) Mathematicalhypotheses which are not based on observation
but whose consequencesagree with the phenomena (or 'save
the phenomena', as the Platonistssaid).
(d) Essentialistor metaphysicalcausalexplanationswhich have no
place in physicalscience.
Of these four, (a) and (b) are based on observation,and can be,
from experience,known to be true; (c) is not based on observation
and has only an instrumentalsignificance-thus more than one instru-
ment may do the trick; and (d) is known to be false whenever it
constructsa world of essencesbehind the world of appearances. Thus
(c) is known to be falsewhenever it is claimed to have the statusof (d).
(18) These results,clearly, apply to cases other than Newtonian
theory, for example, to atomism (corpusculartheory). In so far
as this theory attempts to explain the world of appearancesby
constructingan invisible world of'inward essences' (Pr 102oz) behind
the world of appearances,it must be rejected. (Cf. Pr 50; S 232,
235.)
(19) The work of the scientist leads to something that may be
called 'explanation', but it is hardly of great value for understanding
the thing explained, since the attainableexplanationis not one based
upon an insight into the nature of things. But it is of practicalim-
portance. It enables us to make both applicationsand predictions.
laws of nature or motions direct us how to act, and teach us
..
what to expect' (S 234; cf. Pr 62). Predictionis basedmerely upon
regularsequence (not upon causalsequence-at least not in the essen-
tialist sense). A sudden darknessat noon may be a 'prognostic
indicator, a warning 'sign', a 'mark' of the coming downpour:
nobody takes it as its cause. Now all observed regularitiesare of
this nature, even though 'prognostics' or 'signs' are usually mis-
taken for true causes. (TV 147; Pr 44, 65, io8; S 252-254; Alc iv,
14, IS.)
(zo20)A general practical result of the analysis of physics may
be called 'Berkeley's razor'. It allows us a priori to eliminate
from physical science all essentialist explanations. If they have a
mathematical and a predictive content, they may be admitted qua
3'

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 K. R. POPPER

mathematical hypotheses (while their essentialist interpretation is

eliminated). If not, they may be ruled out altogether. This razor
is sharper than Ockham's: all entities are ruled out except those
which are perceived.
(21) The ultimate argument for these views, the reason why
occult substancesand qualities,physicalforces,structuresof corpuscles,
etc., absolutespace, and absolutemotion, are eliminated,is this : we
know that there are no entities such as these because we know
that the words professedly designating them must be meaningless.
To have a meaning, a word must standfor an 'idea' (or a 'notion ' ;
but words belonging to physical science cannot stand for 'notions ').
Now the words here in question do not stand for ideas. 'Those
who assertthat active force, action, and the principle of motion are
really in the bodies, maintaina doctrine that is basedupon no experi-
ence, and supportit by obscureand general terms, and do not them-
selves understandwhat they wish to say' (DM 3I).

Everybody who reads this list of twenty-one theses must be

struck by its modernity. It is astonishinglysimilar, especially in its
criticismof Newton, to the philosophy of physics which Ernst Mach
taught, for many years, in the conviction that it was new and revolu-
tionary; in which he was followed, for example, by Joseph Petzold;
and which had an immense influence on modern physics, especially
on the Theory of Relativity. There is only one difference: Mach's
'principle of the economy of thought' (Denkoekonomie) goes, no
doubt, beyond what I have called 'Berkeley's razor' in so far as it
allows us not only to discard certain 'metaphysical elements', but
also to distinguishin some cases between various competing hypo-
theses (of the kind called by Berkeley 'mathematical'), with respect
to their simplicity. Very striking also is the similaritywith Hertz's
Principlesof Mechanics(1894), in which he tried to eliminatethe con-
cept of'force', and with Wittgenstein's Tractatus.
What is perhaps most striking is that Berkeley and Mach, both
greatadmirersof Newton, criticisethe ideas of absolutetime, absolute
space, and absolute motion, on very similarlines. Mach's criticism,
exactly like Berkeley's, culminates in the suggestion that Newton's
argumentsin favour of his absolute space (Foucault'spendulum, the
rotating bucket of water, the effect of centrifugal forces upon the
32

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 BERKELEY AS PRECURSOR OF MACH

shape of the earth) fail because these movements are relative to the
system of the fixed stars.
To gauge the significanceof this anticipationof Mach's criticism,
I shall produce two quotations, one from Mach and one from
Einstein. Mach wrote (in the 7th edition of the Mechanics,1912, Ch.
2, sect. 6, II) of the reception of his criticismof absolutemotion,pro-
pounded in earlier editions of his Mechanics: 'Thirty years ago, the
view that the notion of " absolute motion" is meaningless,without
any empirical content, and scientificallywithout use, was generally
felt to be very strange. Today, this view is upheld by many and by
well-known investigators.' And Einstein said, in his obituary notice
for Mach (' Nachruf auf Mach', Physikalische Zeitschr.,1916), with
referenceto this view of Mach's : 'It is not improbable that Mach
would have found the Theory of Relativity if, at a time when his
mind was still young, the problem of the constancy of velocity of
light had agitated the physicists.' This remark of Einstein's is no
doubt more than generous. Of the bright light it throws upon
Mach, some reflectionmust fall upon Berkeley.x

A few words may be said about the relation of Berkeley's philo-

sophy of science to his metaphysics. This is in many respectsvery
differentindeed from Mach's.
While the positivist Mach was an enemy of all traditional,that is
non-positivistic,metaphysics,and especiallyof all theology, Berkeley
was a Christian theologian, and intensely interested in Christian
apologetics. While Mach and Berkeley agreed that such words as
'absolute time', 'absolute space' and 'absolute motion' are mean-
ingless and therefore to be eliminated from science, Mach surely
would not have agreed with Berkeley on his reasons why physics
cannot treat of real causes. Berkeley believed in causes, even in
'true' or 'real' causes; but all true or real causeswere, to him,
'efficient or final causes' (S 231) and therefore spiritual,and utterly
beyond physics (cf. HP ii). He also believed in true or real causal
explanation(S 231) or, as I may perhapscall it, in 'ultimate explana-
tion'. This, for him, was God.
All appearancesare truly caused by God, and explained through
1 This is not the place to discuss other predecessorsof Mach, such as Leibniz.
c 33

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 K. R. POPPER
God's intervention. This, for Berkeley, is the simple reason why
physics can only describeregularities,not find true causes.
It would be a mistake, however, to think that the similarity
between Berkeley and Mach is, by these differences,shown to be only
superficial. On the contrary, Berkeley and Mach are both con-
vinced that there is no physical world (of primary qualities, or of
atoms; cf. Pr 50; S 232, 235) behind the world of physical appear-
ances. Both believed in a form of the doctrine nowadays called
phenomenalism-the view that physical things are bundles, or com-
plexes, or constructsof phenomenalqualities,of particularexperienced
colours, noises, etc. Mach calls them 'complexes of elements'.
The differenceis that for Berkeley, these are directly causedby God.
For Mach, they are just there. While Berkeley says that there can
be nothing physical behind the physical phenomena, Mach suggests
that there is nothing at all behind them.

5
The great historicalimportance of Berkeley lies, I believe, in his
protest against essentialistexplanationsin science. Newton himself
did not interprethis theory in the sense of essentialism; he did not
believe himself to have found that physical bodies, by their nature,
are not only extended but endowed with a force of attraction (radi-
ating from them, and proportionateto the amount of matterin them).
But the essentialistinterpretationof his theory soon became the ruling
one, and remainedso up to the days of Mach.
In our own day, essentialismhas been dethroned; a Berkeleyan
or Machian positivism or instrumentalismhas, after all these years,
become fashionable.
Yet, clearly, there is a third possibility.
Essentialismis, I believe, untenable. It implies the idea of an
ultimateexplanation, for an essentialistexplanationis neither in need
of, nor capable of, further explanation. (If it is in the nature of a
body to attractothers, then there is no need to ask for an explanation
of this fact, and no possibility of finding such an explanation.) Yet
we know, at least since Einstein, that explanation may be pushed,
unexpectedly,furtherand further.
But although we must reject essentialism,this does not mean that
we have to accept positivism.
I shall not here discuss the positivist dogma of meaning, since I
34

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 BERKELEY AS PRECURSOR OF MACH
have done so elsewhere. I shall make only six observations. (a)
One can work with something like a world 'behind' the world of
appearancewithout committing oneself to essentialism (especiallyif
one assumesthat we can never know whether there is not a further
world behind that world). To put it less vaguely, one can work
with the idea of hierarchicallevels of explanatoryhypotheses. There
are comparativelylow level ones (somewhat like what Berkeley had
in mind when he spoke of'Laws of Nature ') ; higher ones such as
Kepler'slaws, still higher ones such as Newton's theory, and, next,
Relativity. (b) These theories are not mathematicalhypotheses, that
is nothingbut instrumentsfor the prediction of appearances. Their
function goes very much further; for (c) there is no pure appearance
or pure observation: what Berkeley had in mind when he spoke of
these things is always the result of interpretation,and (d) has therefore
a theoretical or hypothetical admixture. (e) New theories, more-
over, may lead to re-interpretationof old appearances,and in this
way change the world of appearances. (f) The multiplicity of ex-
planatorytheories which Berkeley noted (see Section 2, 5 I6, above)
is used, wherever possible, to construct, for any two competing
theories, conditions in which they yield different observable results,
so that we can make a crucial test to decide between them, winning
in this way new experience.
A main point of this third view is that science aims at truetheories,
even though we can never be certain, about any particulartheory,
that it is true; and that science may progress (and know that it does
so), inventing theories which, compared with earlier ones, may be
describedas better approximationsto what is true.
Such a view can, without becoming essentialist,admit that we
always attempt, in science, to explain the known by the unknown,
the observed (and observable)by the unobserved (and, perhaps, un-
observable). At the same time, it can fully admit, without becoming
instrumentalist,what Berkeley said of the riature of hypotheses in
the following passage (S 228), which shows the weakness of his
analysis (its failure to realise the hypothetical characterof what he
calls 'laws of nature') and its strength (its admirableunderstanding
of the logical structureof hypotheticalexplanation):
It is one thing to arrive at general laws of nature from a contem-
plation of the phenomena; and another to frame an hypothesis, and
from thence deduce the phenomena. Those who suppose epicycles,
and by them explain the motions and appearancesof the planets, may
35

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions
 K. R. POPPER
not therefore be thought to have discovered principlestrue in fact and
nature. And, albeit we may from the premises infer a conclusion, it
will not follow that we can argue reciprocally,and from the conclusion
infer the premises. For instance, supposing an elastic fluid, whose
constituent minute particles are equidistant from each other, and of
equal densities and diameters, and recede one from another with a
centrifugalforce which is inversely as the distanceof the centres ; and
admitting that from such supposition it must follow that the density
and elastic force of such fluid are in the inverse proportion of the
space it occupies when compressed by any force; yet we cannot
reciprocallyinfer that a fluid endued with this property must therefore
consist of such supposed equal particles.

The LondonSchoolof EconomicsandPoliticalScience

HoughtonStreet,Aldwych,London,WC2

36

This content downloaded from 171.67.34.205 on Tue, 26 Mar 2013 16:09:21 PM

All use subject to JSTOR Terms and Conditions