You are on page 1of 300





Managing Editor

Department of Philosophy, University of Western Ontario, Canada

Managing Editor 1980--1997

Late, Department of Philosophy, University ofWestem Ontario, Canada

Editorial Board
JOHN L. BELL, University of Western Ontario
JEFFREY BUB, University of Maryland
ROBERT CLIFTON, University of Pittsburgh
ROBERT DiSALLE, University of Western Ontario
MICHAEL FRIEDMAN, Indiana University
WILLIAM HARPER, University of Western Ontario
CLIFFORD A. HOOKER, University ofNewcastle
KEITH HUMPHREY, University of Western Ontario
AUSONIO MARRAS, University ofWestem Ontario
JORGEN MITTELSTRASS, Universitiit Konstanz
JOHN M. NICHOLAS, University of Western Ontario
ITAMARPITOWSKY, Hebrew University
GRAHAM SOLOMON, Wilfrid Laurier University

A Study in Natural Philosophy




A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 978-90-481-6444-8 ISBN 978-94-017-0175-4 (eBook)

DOI 10.1007/978-94-017-0175-4

Printed on acid-free paper

AII Rights Reserved

2003 Springer Science+Business Media Dord.recht
Originally published by Kluwer Academic Publishers in 2003
SoftcQver reprint of the hardcover 1st edition 2003

No part of this work may be reproduced, stored in a retrieval system, or transmitted

in any form or by any means, electronic, mechanical, photocopying, microfilming, recording
or otherwise, without written permission from the Publisher, with the exception
of any material supplied specifically for the purpose of being entered
and executed on a computer system, for exclusive use by the purchaser of the work.
For Michelle Wilson



Ernst Mach was a physicist, sense-physiologist, and philosopher and, after the death
of Helmholtz, probably the last individual to make significant professional
contributions to all three areas. Because of his broad training, Mach was also one of
the first scientists to suggest that the results of these special disciplines ought to
harmonize with one another in a metascientific natural philosophy, in which, for
example, the data of psychology would be valued equally with the data of physics. In
particular, Mach believed there should not be a gulf between a physical science of
objects and motions and the psychological science of sensations and thoughts.

In the seventeenth century, dualism had been proposed as a way out of the
dilemma, making sensations into subjective, secondary qualities which were caused
by interaction with a sensationless world of matter and motion, like that proposed by
Galileo and Descartes. Henceforth, the scientist could ignore the quality of
experience as such and concentrate on mathematical models of the primary qualities
of nature: bulk, number, shape, and motion.

By Mach's time, dualism had broken down and led either to idealism or
materialism. Berkeley's idealism, for example, proposed that the entire world
consisted of sense qualities, either ofhuman beings or of God, who has sensations of
objects even when human beings do not. By comparison, in materialism, sensations
and mental phenomena had no reality and it was hoped that particles and forces
would eventually explain them away. At the beginning of the nineteenth century in
Vienna, materialism was at its height. But Mach rejected both of these alternatives.
His goal was a natural philosophy that could bring the abstractness and idealization
of physics into harmony with the concreteness of sensations. This alternative
conception was later called "neutral monism" by Bertrand Russell.

Philosophers are familiar with Mach as a forerunner of the Vienna Circle, that
group of mid-twentieth century scientist-philosophers that included Rudolf Carnap,
Moritz Schlick, Otto Neurath, and others loosely called logical positivists. Positivists
were thought to hold to a "verification principle" according to which statements
unverifiable in principle were branded "metaphysical" and removed from scientific
discussion. It can easily be shown that Mach neither believed in this principle nor did
his own physical and philosophical speculations measure up to it. So if the
verification principle were the measuring stick, Mach was not a positivist, at least
not a logical positivist.

I became acquainted with Mach's writings while I was extremely hostile to

positivism of any sort, and I read his Analysis of Sensations several times before I
could think seriously about it. Indeed, it was really only after considering Mach's

development in the context of German philosophy and science from Kant, J.F.
Herbart, G.T. Fechner, and Johannes Muller that I realized the usual positivist
reading of Mach was in error, a result of emphasizing "Mach the positivist
philosopher" over Mach the scientist and natural philosopher.

Mach began his physics training as a straightforward realist: believing in a

world really consisting of matter moving in space and time, independent of human
sensory powers. Reading Kant as a fifteen-year-old forced Mach to doubt the
application of spatial and temporal concepts to the world beyond the conditions of
human perception, and Mach himself began to doubt the Ding an sich or a
permanence behind the appearance of matter.

In J.F. Herbart's Allgemeine Metaphysik, which he studied as a young man,

Mach found an example of a philosophical construction of space and matter out of
unextended elementary forces or energies. But it was Mach's own work in
psychophysics and the influence of G.T. Fechner that convinced him the real
constituents of the world were concrete qualities and functions. It was then but a
short step from Fechner and Herbart to his own Elementenlehre.

Such was the great axis on which Mach's thought moved: reconciling
psychophysics with physics. But whereas Fechner and others had tried to make
psychophysics look more like physics, Mach attempted a sense-physiological
critique of physical concepts. Mach thought the spatial and temporal form of
physical principles made concessions to the human need to visualize events in a kind
of sensory continuum, like the visual or auditory fields, which later could be dropped
in a more mature science. Mach's historico-conceptual studies of mechanics and
thermodynamics were attempts to set the most general and abstract results free from
their historical background as well as from visualizable picture-thinking.

I would like to thank the Deutscher Akademischer Austauschdienst (DAAD) for

a grant that allowed me to work on the Mach Nachlass. I would also like to thank
Arnold Koslow, Michael Levin, Joseph Dauben, Rob Deltete, and John Blackmore
for comments on various sections, Alice Calaprice for her help copyediting,
remaining errors being entirely mine, and Aris Noah, my undergraduate teacher, for
sparking my interest in naturalistic philosophy in our discussions of W.V. Quine. I
also want to thank my mother, Laurene Buckley, for her unstinting financial support
toward my education in times of need.

New York, 2002



(NL 174) Ernst Mach Nachlaft, Deutsches Museum, Munich

(AU) Ernst Mach. Blackmore, John and Hentschel, Klaus, eds., Ernst Mach als
Auftenseiter. Wien: Wilhelm Braumilller, 1985.
(WK.) Thiele, Joachim, ed., Wissenschaftliche Kommunikation: Die Korrespondenz
Ernst Machs. Kastellaun: A. Henn, 1978.


(C) Compendium der Physikfor Mediciner. Wien: Wilhelm Braumilller, 1863.

(V) "Vortriige iiber Psychophysik." Osterreichische Zeitschriftfor praktische Heilkunde


(R) "Bemerkungen iiber die Entwicklung der Raumvorstellungen." Fichtes Zeitschrift

for Philosophie undphilosophische Kritik 49 (1866): 227-232.

(DM) "Ober die Definition der Masse" Carls Repertorium fiir Physik und physikalische
Technik, 4, 1868 355-359.

(CE) The History and Root of the Principle of the Conservation of Energy. Translated
by P.E.B. Jourdain, Chicago: Open Court, 1911. First Edition: 1872.

(M) The Science of Mechanics. Sixth English Edition. Translated by Thomas

McCormack, Chicago: Open Court, 1960. First Edition: 1883.

(AS) The Analysis of Sensations. Translated from the First German edition by C.M.
Williams and from the Fifth German edition by Sidney Waterlow. New York: Dover,
1959. First Edition: 1886.

(H) Principles of the Theory of Heat. Translated by Thomas McCormack, P.E.B.

Jourdain and A.E. Heath. Dordrecht: D.Reidel, 1986. First Edition: 1886.

(PSL) Popular Scientific Lectures. (1898) Third Edition, Translated by Thomas

McCormack. Chicago: Open Court, 1898.

(PWV) Popular- wissenschaftliche Vorlesungen Fifth Edition Wien: Bohlau, 1987.

(SG) Space and Geometry. Translated by Thomas McCormack. La Salle, IL: Open
Court, 1906.

(KE) Knowledge and Error. Translated by Thomas McCormack and Paul Foulkes,
Dordrecht D. Reidel, 1976. First Edition: 1905.

(SW) Sitzungsberichte der kaiserlichen Akademie der Wissenschaften, Wien,

naturwissenschaftliche- mathematische Klasse.


(DL) Blackmore, John, ed., Ernst Mach: A Deeper Look: Documents and New
Perspectives. Dordrecht: Kluwer Academic Publishers, 1992.

(MB) Ratliff, Floyd. Mach Bands. San Francisco: Holden Day, 1965.


I make no pretension to the title of a philosopher. I only seek to adopt in physics a point of view
that need not be changed immediately on glancing over into the domain of another science; for,
ultimately, all must form one whole. (AS, 30)

Ernst Mach (1838-1916), a physicist, sense physiologist and philosopher malgni lui
was perhaps the last person to make significant professional contributions in all three
fields. Raised outside Vienna, Mach studied physics and sense physiology in the city
and taught, first at Graz, and then for most of his career (1867-1894) in Prague.
Through his historico-critical studies such as Die Mechanik in ihrer Entwicklung
(1883) and Beitriige zur Analyse der Empjindungen (1886), Mach wrote for a world
wide audience outside his Prague lecture room: physicists such as Albert Einstein,
Wolfgang Pauli, and Erwin Schrodinger; philosophers such as William James,
Bertrand Russell, Rudolf Camap, and Otto Neurath and a host of others, including
the economist Friedrich von Hayek and the novelist Robert Musil.'

In 1895, Mach returned to Vienna and took up a chair in the history and theory
of the inductive sciences, especially prepared for him. He taught only three years
until 1898, when he was paralyzed on one side by a stroke. Yet it was at this time,
while living in retirement in Vienna, that Mach exerted his greatest influence on the
next generation of intellectuals. Einstein even came to visit Mach in 1910, and the
two men sat down for a physico-philosophical discussion on the economical
expediency of assuming the existence of atoms, which Mach had always denied, but
which the leading historians now believe he accepted by that time.2

Mach finally moved in with his son Ludwig in Vaterstettin near Munich in 1911,
where he died in 1916, still unaware of the outbreak of World War I. That was also
the year Einstein published the article announcing his general theory of relativity, in
which Mach is mentioned prominently by name. In particular, Einstein
acknowledged Mach's criticism of Newton's bucket experiment as his motivation to
deal with the inertial effects induced by accelerated motions3 (even if Einstein's

principle of equivalence was not a solution even remotely suggested by Mach). 4 But
after correspondence with Einstein and several attempts to understand his approach
to the problems of space and time, 5 Mach had distanced himself from the theory of
relativity in the 1913 Preface to his own Optik, calling it an aperr;u, but not the
breakthrough he was hoping for. Because of the very strong affinities between Mach
and Einstein, one scholar (Gereon Wolters) refuses to take this denial at face value.
Einstein, however, did, and indicated that he believed Mach's philosophy to be too
positivistic: "a catalogue but not a system," as he said in his Paris lectures of 1922. 6

Mach's main contributions to philosophy were contained in his Analysis of

Sensations and Knowledge and Error and woven into his works on the history of the
natural sciences. Mach's History and Root of the Principle of the Conservation of
Energy (1872) and his Mechanics (1883) charted the rise of mechanical thought from
the machines of the ancient Greeks to the seventeenth century to analytical
mechanics. In his Principles of the Theory of Heat (1896) Mach gave the historical
development of the newer science of thermodynamics and extended analogies to the
postulates of mechanics.

The resulting unified conception of the different sciences was called

"comparative physics" by Mach. On this conception-to be considered in the body
of this book-the physicist would form a sort of qualitative catalog of all the
different fundamental forms of energy, which transformed into one another by a
conservation law, and which, Mach said, obeyed an analog to the law of least action
in mechanics, when the corresponding terms were substituted for equivalents in the
nonmechanical sciences. Mach seemed to think the path to a unified science of
energy was a historical destiny, implicitly followed by the great investigators, which
would become evident to anyone who viewed the history of science in an unbiased
fashion. However, Max Planck and Ludwig Boltzmann were two prominent
opponents of this "energetic" view who were not ignorant of the history of their

Mach's approach was called "phenomenological" since, for him, physics could
not really penetrate beyond the phenomena of energy transformations, even in
natural interactions at the smallest scale. Just as science had begun with the implicit
assumption of the conservation of energy in the impossibility of a perpetual motion,
so, for him, science would be expected to end with a catalog of fundamental energy
types and empirical laws of transformation, with no further scaffolding supporting it,
much the way phenomenological thermodynamics remains a science of postulates
about the permissible transformations of heat to work. Any attempt to introduce
mechanisms behind such transformations, due to atoms, an ether or what have you,
Mach regarded with extreme skepticism. The professed disbelief in atoms in effect
alienated Mach from the younger generation of physicists like Einstein, who could
well appreciate Mach's point that mechanics was not the undisputed basis of physics,
but who could not accept his inveterate skepticism about atoms. Einstein even sent

his own paper on Brownian motion to Mach in hopes of converting the older man.
And yet Mach's tendency to concentrate on the appearances influenced many
phenomenalistic scientists at the beginning of the twentieth century and could have
been an indirect influence on the quantum theory.


Mach's scientific views have often been attributed to his so-called sensationalist
philosophical position, encountered in the Analysis of Sensations and in his other
writings. In two of the most quoted passages, Mach indeed seemed to be claiming
that reality was a mass of individual (human) sensations:

The assertion .. .is correct that the world consists only of our sensations. In which case we have
knowledge only of sensations, and the assumption of nuclei ... or of a reciprocal action between
them turns out to be quite idle and superfluous. 7

Properly speaking the world is not composed of "things" as its elements, but of colors, tones,
pressures, spaces, times, in short what we ordinarily call individual sensations.'

Although I do not share it, "sensationalism" has been the favored interpretation
of Mach for two reasons: there is prima facie support for it in his writings and it
makes easy sense of his various positions on disparate--often difficult to master-

It is easy enough to write a book about Mach the protopositivist, for if Mach
believed only in the reality of human sensations then naturally he would reject
atoms, since they could not be seen. Even Mach's deep critique of Newtonian
absolute space can be likewise attributed to the fact that absolute space was
"unverifiable to the senses" (which Mach even says several times). And one can go
on from here. His advocacy of the unity of science and the elimination of
metaphysics can be interpreted as a call to adopt a universal basis for all sciences in a
base-set of sensations, and the elimination of any statement unverifiable in terms of
this evidence. The term "pseudoproblem" can be interpreted crudely to mean a
problem that does not admit of an empirical decision one way or another. And
finally, Mach's subtle doctrine of the economy of thought can be set on this
Procrustean bed by considering concepts to be suppressed lists, inductive groupings
of sense data that make their variations easier to remember and which ease in the
prediction and control of future events. Packaging it all together gives us a picture of
the "positivist" Mach. And indeed, whether this man existed or not, he was very
influential in twentieth-century science and philosophy. Some of his stock
expressions, such as "eliminate metaphysics," "unity of science," "pseudoproblem"
and "element" were eventually taken over and cited with approval by members of the

Vienna Circle. 1 For historical reasons, then, it is often convenient to see Mach's
philosophy as an early version of Vienna Circle empiricism.

Mach's actual philosophy, which will be treated in its historico-conceptual

development in this book, was somewhat different. His overall position is often
called "neutral monism," a name invented by Russell in 1914.

Neutral monism holds that minds and physical bodies are functionally
determined complexes, both of which consist of the same, very tiny physical qualia,
the elements. Mach said that these neutral elements were only sensations in their
psychical variations, i.e., when related to elements of memory images or to elements
of the human body such as the skin or the retina. An example of a psychical variation
is the formation of an after-image of the sun in the eye. However, when the
variations of the body and psychological data were held constant but the sense data
still underwent independent variations, Mach called them physical objects and, in
that case, called the data "elements":
A color is a physical object as soon as we consider its dependence, for instance, upon its
luminous source, upon other colors, upon temperature, upon spaces, and so forth. When we
consider however its dependence upon the retina .. .it is a psychological object, a sensation. Not
the subject matter but the direction of our investigation is different in the two domains. 11

An example of a physical variation would be an object falling in the visual field at

9.8 rnls 2, a law that is independent of the human sensory apparatus.

Because of the way he writes, it often seems that all of Mach's physical
elements can be interpreted as my sensations, by considering only their psychical
variations. 12 For example, Mach even says in the Analysis that "the whole system of
elements" undergoes a change if the psychical organs like the retina or skin are
interfered with. If so, then Mach's elements were severely limited to one
consciousness at a time, and he would be vulnerable to charges of solipsism and
idealism 13 despite his own insistence on their objectivity. V.I. Lenin (while he was
still overthrowing philosophies and not monarchies) made this point with brio in his
Materialism and Empirio-Criticism, directed against Russian Machists:

Mach and Avenarius secretly smuggle in materialism by means of their word "element," which
supposedly frees their theory of the "one-sidedness" of subjective idealism, supposedly permits
the assumption that the mental is dependent on the retina, nerves and so forth and the
assumption that the physical is independent of the human organism. In fact of course the trick
with the word "element" is a wretched sophistry, for a materialist who reads Mach and
Avenarius will immediately ask: what are the "elements"? Either the "element" is a sensation,
as all empirio-criticists, Mach, Avenarius, Petzoldt, etc. maintain-in which case your
philosophy, gentlemen, is idealism, vainly seeking to hide the nakedness of its solipsism under
the cloak of a more "objective" terminology; or the element is not a sensation-in which case
absolutely no thought whatever is attached to the "new" term; it is merely an empty bauble. 14

Mach is open to attack here because the point is so obliquely addressed in his
published writings, although on the subject of other minds in the Analysis of
Sensations, Mach did speak of sensations "added in thought" (hinzugedachte
Empjindungen ):

When I speak of the sensations of another person, those sensations are, of course, not exhibited
in my optical or physical space; they are mentally added to the brain observed or rather
functionally presented."

Moritz Schlick, who held a representative view of perception, pointed out

sharply that Mach could not have it both ways, sometimes believing only in the
givenness of his own sensations and sometimes implicitly adding those of others. 16
These and other issues will be dealt with later in the book, but instinctually what one
really wants from Mach is some clear statement that he believed in objects or
elements that were not human sensations under any variation at all: elements that
made up a mind-independent world of nature even when no human being is sensing
it, similar to what Russell later called sensibilia-but not as blatantly
anthropormorphic as Berkeley's sensations in the mind of God.

There is evidence that Mach did indeed embrace such "world elements," not as
philosophical dodgework-as Lenin claimed-but as part of the natural development
of his scientific theory of the world. This can be demonstrated conclusively.

The first mention of nonhuman sensations occurred in Mach's Vienna lectures

on psychophysics of 1863, where he stated that, like G.T. Fechner, he believed that
nature, in addition to its quantitatively described exterior properties, possessed also a
private "inner side," accessible only to the thing itself. This inner side was of the
nature of sensation, and by a (rather broad) analogy one could imagine animal
sensations, and then blur the boundary between simple animals and plants and finally
between living and inanimate matter:

We know of our own soul immediately. Through analogy we attribute to other men a soul
similar to our own, since we conclude from their similar behavior that they have similar
thoughts and inner states. With less certainty the manner of reasoning is applied to animals and
with still less certainty the lower we go in the series of animals. Where is the limit of this
reasoning, asks Fechner? The mere question must be recognized as a brilliant one and when we
reflect on it more precisely we can answer with Fechner: nowhere! We are not justified in
denying a soul to animals and plants. Granted, when we remember how essentially
organization is connected with psychical life, capability for motion with perception, for
example, we must admit that the psychical life of plants must be very different from that of
animals. 17

This theme was very pronounced in the young Mach thanks to the pervasive
influence of Fechner on his thought at this time, as extensively documented by
Michael Heidelberger. 18 Elsewhere in the lectures on psychophysics, he had even
attributed "sensations" to the interior of atoms, a Ia Leibniz, very tiny qualia at a
subatomic scale:

We cannot attribute to atoms an outer side. If we must think anything, we must attribute to
them an inner side, an inwardness analogous in some respects to our own soul. In fact, where
could the soul come from in a combination of atoms in the organism if the kernel did not
already lie in the individual atoms? 19

Nine years later, in lectures entitled "Uber einige Hauptfragen der Physik"
(1872), Mach was ready to give full reign to this hypothesis of sensation in matter, or
a matter made up out of sensation. Here, again, the desire to reconcile physics and
psychology comes to the fore:

Sensation is a general property of matter, more general than motion. Let us seek to set down
this proposition clearly. An organism is a system of molecules. Electrical currents run into the
interior and come back again into the muscles. Everything is physically explainable. But not
that the person should have sensations. What we can investigate physically is always merely
physical. We fmd no sensation. And yet the human being senses. The material flows forward
through and through him. The old depart. The new comes in. We have therefore the problem of
finding something fundamentally new in the whole that is not in the parts. We escape this
difficulty when we consider sensation as a general property of matter.

Thus, we have two sides, an inner and an outer side. According to analogy we assume these
also in animals, although on weaker evidence. In plants still weaker, but the boundary blurs.
With respect to the inorganic again the boundary blurs. Therefore no doubt everywhere an
inner side. One can only be convinced of it from one's own soul. The rest is conjectured.

What scientific value this assumption of a general sensation of matter has, this can only be
decided by how much better we can deduce and understand physical phenomena through them.
Rules for deduction of our sensations with the help of other sensations added in thought and in
causal relation to them. Thus, as the most immediate goal of science the construction of the
world out of sensations. The appearance of matter, in so far as it is sensation, built up similar to
the way physics has built up material added in thought (the atom).

Why do we assume a soul outside of our own? This assumption: to certain of our own, to add
in thought certain others, which make more comprehensible the triggering of our
own ... animals and plants. Still more of such sensations which we add in thought to make
others comprehensible. Finally we add them in thought everywhere?0

Mach was quite capable of postulating unobservable elements when it suited

him and even adding unobservables in causal relations to present sensations,
comparing his own procedure to the dreaded postulation of atoms behind
appearances. Hence, if he rejected the introduction of atoms behind appearances, as
he already had by 1872, it was probably some feature of the mechanical atom itself
that he rejected and not hypothetical reasoning as such. In 1910, near the end of
Mach's career, the world elements surfaced again in his correspondence with
Friedrich Adler. The exchange was provoked by Adler, who wrote:

I defend the standpoint: Elements that do not belong to any "I" do not occur and are
extraneous, not to say metaphysical for our world view. My question is now: Do you consider
necessary the assumption of elements that belong to an object, but not to a subject? Obviously
such elements are not directly given, but do you assume them hypothetically, as it appears to
me you do from the places cited in your Analyse der Empfindungen? 21

And Mach wrote back:

Because I only seek a standpoint on which physics can rationally build itself I can stop by the
double-dependence of elements, for these are experimentally demonstrable, and have nothing
to do with a philosophy or metaphysics. From such a standpoint, I may also assume analogous

elements as contents of another I and this approach can even extend to the lower animals,
plants and inorganic bodies. This hypothesis serves only to round off the world view
provisionally, and in hope of the future construction of biology ... Healthy biological research
must teach if this hypothesis has any worth, and, if so, what worth. Speculation cannot manage
this. Provisionally it appears to me that we completely overlook a side of our experience when
we overlook this hypothesis .. .I have not further cultivated all of these matters, for I always
feared the nearness of the metaphysical abyss, where there is no experiential foundation. 22

World elements were thus more than the creatures of Mach's night-thinking. In
his published writings, there are clear signs that he envisioned a future science, made
possible by advances in psychophysics, in which such interior sensations in matter as
he had hypothesized could become known to natural investigators. For example, in
the conclusion to his lecture "On the Economical Nature of Physical Inquiry" of
1882 he wrote:

We cannot draw the science of the future in sharp contour. We can only intimate that the hard
dividing wall between man and the world will gradually disappear; then that men will not only
approach one another, but also the whole organic and so-called life-less Nature with less self-
seeking and a warmer feeling? 3

The same prediction reappeared in the Mechanik of 1883:

Careful physical research will lead to an analysis of our sensations. We shall then discover that
hunger is not so essentially different from the tendency of sulfuric acid for zinc, and our will
not so different from the pressure of a stone as it now appears. We shall again feel ourselves
nearer nature without its being necessary that we should resolve ourselves into a nebulous and
mystical mass of molecules, or make nature a haunt ofhobgoblins.24

Mach thus believed that world elements could become experienced by observers
through advances in biology or psychophysics, and sometimes he wrote that the
greatest advances to be expected in science would come from the addition of a
scientific psychophysics to physics. In letters to Gabriele Rabel not published until
1921, Mach speculated on this possibility:

When the tunnel between the physical and the psychical is built through, or almost so, we will
not be so restricted. Then we will be permitted to ask how the animals sense. Admittedly, that
still lies far off. But no restrictions, no limits for those who will come after us and who will be
cleverer than us. The astronomers of today know that the ptolemaic and the copernican world
views are both practical conventional restrictions and that one can permit oneself a freer basis
for questioning. 25

From these, and other dated utterances, it is clear that Ernst Mach really did
believe in observer-independent elements in matter all the way from 1863 to 1916,
which formed what one perceptive critic called a "genuine metaphysical reality." 26
However, Mach was extremely loath to speculate about these world elements except
as a hypothesis to fill out his metascientific worldview, or where he could express
his hopes for a "future science"-two areas where scientists are traditionally granted
the luxury of waxing metaphysical without being branded as metaphysicians. Mach
also tell us that he began to suppress his views about elemental processes behind the

phenomena of matter and motion because of the Gespensterfurcht ("fear of ghosts")

they stirred up among his colleagues.27


Mach's works engendered a vigorous critical reaction shortly after 1900. Like Richard
Honigswald's Kritik der Machschen Philosophie (1903) and Herbert Buzello's Kritische
Untersuchungen von Ernst Machs Erkenntnistheorie (1911), some trenchant criticism
came from neo-Kantian philosophers anxious to defend Kant from Mach. By contrast,
Lenin's Materialism and Empirio-Criticism (1905) was written to stem the influence of
Machism on Russian philosophers such as Alexander Bogdonov and defend Marxian
materialism against what Lenin saw as Mach's ineffectual idealism.

Authors such as Friedrich Adler, Josef Popper, and Joseph Petzoldt published
books and articles with the goal of elucidating or improving upon Mach's theories. Also
in this class are Hugo Dingler's Die Grundgedanken der Mach'schen Philosophie
(1924), which included the first extracts from Mach's notebooks to appear in print. Also
appearing in this period, Hans Henning's Ernst Mach als Philosoph, Physiker und
Psycholog (1915) was an exploration of Mach's unity of science thesis; and Kurt
Gerhards' Machs Erkenntnistheorie und der Realismus (1914) was an intriguing attempt
at a realist construal of Mach's elements.

There have been several additional waves of publication in the last thirty years. The
second coincided roughly with the fiftieth anniversary of Mach's death in 1966, in
which three commemorative volumes appeared,28 the first from the Ernst Mach Institut
in Freiburg, the second from the Boston Studies in the Philosophy of Science, and the
third in a special symposium in Synthese. During this time, Floyd Ratliff translated
Mach's articles of the 1860s on neural inhibition, raising the Mach Bands to a central
concern of sense physiology, where they have remained among the most discussed and
analyzed sense illusions.

Meanwhile, much was accomplished on the historical front. The first complete
bibliography of Mach's works was published by Joachim Thiele in 1963,29 followed a
bit later by a selection of letters from the NachlaB under the title Wissenschaftliche
Kommunikation (1978). Friedrich Herneck published Einstein's letters to Mach and
revived the Mach-Einstein debate,30 which centers on Einstein's coinage of "Mach's
Principle." Some historical and scientific materials relevant to this debate have been
gathered in the volume Mach's Principle (1995). 31 Finally, two biographies topped the
crest of the second wave: K.D. Heller's Ernst Mach: Wegbereiter der modernen Physik
(1966) and John Blackmore's Ernst Mach: His Work, Life and Influence (1973), in
which much historical and biographical material appeared for the first time.

Some of the criticism of Mach in the so-called second wave32 came from scientific
realist scholars of the 1960s who were reacting against the dominance of logical
empiricism in America. As Gereon Wolters perceptively remarked in his review of the
literature in 1987, since Mach was so closely associated with philosophers such as
Rudolf Carnap and other members of the Wiener Kreis, Mach scholarship was
instrumentalized for a battle these scholars were waging against the institutionalized
positivism of the three previous decades.

It is to be regretted. ..that the "internal" arguments [by which Wolters means argwnents others put in
Mach's mouth], beginning with the second half of the 1960s, were presented exclusively by
philosophers and historians of science of the 'anti-positivist' persuasion. These colleagues had an
easy time of it, seeking to disavow positivism by hanging the sign 'Positivism' around Mach's neck,
and dismissing him as an ignorant simpleton.33

Fortunately, a third wave began around 1988, Mach's 150th birthday, when two
new collections appeared in German: Rudolf Haller and Friedrich Stadler's collection
Ernst Mach: Werk und Wirkung (1988), and Hubert Laitko and Dieter Hoffmann's Ernst
Mach: Studien und Dokumente zu Leben und Werk (1991 ), including the release of new
correspondence, biographical material, and excerpts from the notebooks. Some of this
material appears in English in John Blackmore's collection Ernst Mach: A Deeper Look
(1992). A most interesting collection of letters is contained in Blackmore and Klaus
Hentschel's Ernst Mach a/s Aufienseiter (1985), including some real gems from Mach's
later years.

Highly influential in the third wave was Paul Feyerabend's defense of Mach. As
one of the chief critics oflogical empiricism, Feyerabend distinguished between Vienna
Circle style philosophy of science and Mach's enterprise, favoring the latter. In his
interpretive article "Mach's Theory of Research and its Relation to Einstein, "34
Feyerabend argued that Mach held a principle-driven (not inductive) theory of inquiry
similar to Einstein and a theory of elements unlike that of logical positivism. He stressed
the fact that Mach's elements were only provisional in nature, and represented the tiniest
divisions of which science was currently capable, not a fundamental base set of givens
for all time. Feyerabend disdained Vienna Circle empiricism as a weakening of Mach's
critical stance toward science into a dogmatic acceptance of whatever "science says":
The first and most noticeable change is the transition from a critical philosophy to sense-data
dogmatism. Elements are replaced by sensations not just temporarily, and as a matter of hypothesis,
but once and for all... Secondly, criticism of science is replaced by logical reconstruction, which in
plain English is nothing but a highly sophisticated form of conformism As a result the idea of
logical reconstruction became conformist The task was now to correctly present rather than change

And then, Feyerabend backed this up with a damning criticism of twentieth-century

philosophy of science. Whereas Mach did history and philosophical analysis in order to
actually change science, he said, the logical empiricists had started a tradition of
philosophical work that fed only on itself. Beginning with logical empiricist theories of

explanation and theory confirmation,36 philosophers of science began to criticize each

other and misapply mathematical logic to scientific theories instead of trying to alter the
course of real inquiry. 37 I think Feyerabend would agree, ifhe were alive today, that the
contemporary field is no longer so confmed to methodological and linguistic issues but
deals more directly with science itself. But ifFeyerabend had an axe to grind against his
contemporaries-more than one perhaps-he did also set the tone for writers trying
hard to set Mach free from his positivist reputation, and from the 1960s scholars.

Another writer in this category is Gereon Wolters, who claimed, somewhat

fancifully, in his Mach I, Mach II, Einstein und Die Relativitiitstheorie (1987) that
Mach's "anti-relativity" preface to his Optik, as well as a passage of his Mechanik, were
forged by his son Ludwig. Wolters's claim suffers from a binary criterion of
"acceptance/rejection" and "favorable/unfavorable" inadequate to Mach's many grades
of theory acceptance and appreciation. 38 To call special relativity an "apen;ue," as Mach
did, is not to call it false, for example.

Wolters' claim that the Optik preface is a forgery has been rejected by Gerald
Holton, John Blackmore and Paul Feyerabend, and Wolters has responded to some of
these criticisms. 39 We are missing the one document that would settle the controversy:
the second half of Mach's Optik. It has never been located and remains something of a
mystery story in Mach scholarship. According to his preface, Mach intended to state his
objections to relativity in this second part of the book. Some of the mystery could be
cleared up by a diligent study of Mach's notebooks on optical experiments, currently
housed at the Deutsches Museum in Munich.

The most important result of the third wave was that the stereotype of Mach the
arch-positivist could be contradicted openly. Wolters, for example, thought Mach had
"little to do" with the Vienna Circle and that Schlick's original group adopted the name
Verein Ernst Mach to take advantage ofthe popularity of a local hero. 40 Rudolf Haller
sought to distance Mach from the logical sense-data construction of Rudolf Carnap's
Der logische Aujbau der Welt. 41 Another writer, Keiichi Noe, rightly stressed the
similarities between Mach and the movement against logical positivism led by
historically minded philosophers of science in the 1960s.42 Thus, the reevaluation of
Mach's thought has been slow in coming and is still not entirely complete.


Lingering comparisons are made between Mach's Elementenlehre and Carnap's

Aujbau,43 itself thought to be one of the best of the positivist attempts at a theory of
elements similar to Mach's. But under the surface, there are surprisingly few affinities
between these projects. There is no doubt about the influence of Mach on Carnap. In
some autobiographical remarks, Carnap does say he was influenced by "Ernst Mach and
phenomenalistic philosophers" in his choice of a sense-data basis for the Aujbau.

Meaning, I suppose, that he thought Mach had laid down such a basis in his own
writings-not the provisional theory44 Mach actually represented:

Under the influence of some philosophers, especially Mach and Russell, I regarded in the Logische
Aujbau a phenomenalistic language as the best for a philosophical analysis of knowledge. I believed
that the task of philosophy consists in reducing all knowledge to a basis of certainty. Since the most
certain knowledge is that of the immediately given, whereas knowledge of material things is
derivative and Jess certain, it seemed that the philosopher must employ a language which uses
sense-data as a basis. In the Vienna discussions my attitude changed gradually toward a preference
for the physicalistic language.45

Carnap said that he adopted a physicalistic basis when it suited him and that he was
"ontologically neutral" between the two.46 This neutrality again sounds like Mach's
neutral monism, but for Carnap it really came to mean something like the adoption of
one sort of thing-language as opposed to another and had nothing to do with an
underlying physico-psychical neutrality of real elements of the world. Unlike Mach,
who was trying to get at the nature of the elements through science, Carnap had no
expressed interest in the entities of the world, except as it might affect the way one
talked about them; his linguistic "neutrality" bordered on ontological indifference. It
was thus no accident that W.V. Quine chose to attack Carnap exactly here, effecting a
tum toward questions of existence and ontological commitment, which Carnap had
taken far too lightly.

As John Blackmore has recently argued,47 Carnap was preoccupied with questions
of logical form current in the early twentieth century, not with the concrete
psychophysical investigations that had been Mach's domain. Indeed when it came to a
choice between physicalism, realism about entities like atoms, fields or tensors, and
phenomenalism Carnap always seems to have chosen standard physicalism anyway. He
held to phenomenalism, it seems, for epistemological purposes only and was very far
from seeing the nature of the physical world in Machian elements. As a criticism of
himself, Carnap wrote in 1969:

The positivist thesis concerning the reducibility of thing concepts to auto-psychological concepts
remains valid, but the assertion that the former can be defined in terms of the latter must now be
given up and hence also the assertion that all statements about things can be translated into
statements about sense data.48

As Michael Friedman has shown,49 Carnap's true achievement in philosophy lay in

embellishing Russell's work on relations and formal structures. In particular, he
formulated a technique called "quasi-analysis," representing similarity between qualities
by a relation that is reflexive, symmetric but non-transitive. Whatever other problems
this approach may have, 5 and there are many, the formation of classes out of "similar"
sense data is a much lower philosophical point of view than Mach's and harkens back to
eighteenth-century British associationism.

In fact, this is a crucial philosophical difference. Mach's elements were not related
to one another by similarilf 1 in the eye of the beholder, but by real physical functions

that described reciprocal changes of intensity in the elements. For example, Mach's
theory of color data and sound52 broke up seemingly elemental magnitudes into two or
more "opposing processes" (hence the title of Mach's book, the Analysis of Sensations,
not the Sensations of Analysis). Colors are not simples but break up into white-black,
red-green and blue-yellow processes, so that two colors "next to each other" on the
spectrum are simply ordered by three independent proportions; they are no more "alike
or different" than any others. 53 This understanding of qualia as possessed of causal
powers derives from German psychology, in particular from J.F. Herbart and Ewald


Anti-atomism has been a signature topic in Mach scholarship, for obvious reasons, since
Mach's stand against atoms is widely regarded as his greatest scientific mistake. Erwin
Hieberf 4, Steven Brush, and Laurens Laudan have written about Mach's growing
dissatisfaction with the atomic theory of his time. And more recently, Wolters and
Blackmore have added to the story that Mach really did come around to accepting
experimental evidence for the existence of atoms by 1910.

Stephan Meyer had claimed that he had experimentally convinced Mach of the
existence of atoms in his laboratory in 1903, 55 although, as John Blackmore pointed out,
this revelation did not effect any immediate sea-change in Mach's ideas, until he finally
wrote in his notebook c. 1913-14: "Atome nicht occult?" (Atoms not occult?i6

Some think that Mach accepted atomism in his early writings but later only as an
hypothesis for organizing the appearances, as he purportedly told Einstein during their
meeting in 1910. 57 Or perhaps the mechanical atomism Mach attacked was a patently
false "billiard ball" or "Ding an sich" theory not believed in by anyone (as Feyerabend
claimed). 58 Mach sometimes said that he thought atoms were by defmition
transcendental Dinge an sich, utterly inaccessible to any possible experiment. In his
Mechanics he stated that "atoms cannot be perceived by the senses, like all substances
they are things of thought, "59 and in this passage of 1903 he said:

But the moment we begin to operate with mere things of thought like atoms and molecules, which
from their very nature can never be made the objects of sensuous contemplation, we are under no
obligation to think of them as standing in spatial relationships which are peculiar to the Euclidean
three-dimensional space of the senses (italics his). 60

Mach generally regarded substances as inalterable and hence not part of the real
world of change; he also tended to rain antimetaphysical criticism on the spatial
properties of atoms as invalid extensions of sensory properties of the macroworld to the
microworld beneath. It seems to have occurred to Mach around 191 0 that the atoms his
colleagues were talking about were not transcendental, occult and permanently beyond
experience, but just very small things, otherwise absolutely continuous with events
observed on a macroscopic scale.

I believe that Mach's opposition to atomism did not rest on empirical grounds, but
was a much more troublesome problem of which ultimate theory of matter went the
deepest. This is a philosophical question on which scientists permit themselves to
speculate, some of them favoring particulate matter bound by forces, others a field or
ether theory, and still others favoring more exotic possibilities. Machian elements were
an example of this last option.

As I will show later in the text, there was a specifically Machian theory of matter at
the subatomic level, in which extended matter was to be made up of his spaceless
elements and functions. So, in addition to his skepticism about macroatoms, Mach
might also have doubted the ultimacy of atoms in comparison with his own deeper
theory of elements. 61 Feyerabend also insisted on this reading: "Elements as envisaged
by Mach are more fundamental than atoms. "62 If, furthermore, elements represented
something like very small physical energies, Mach's Elementenlehre would dovetail
nicely with his phenomenalism in physics, i.e., that at the smallest scale there is nothing
to seek behind the mechanisms of energy transfer besides more elements, as finely-
grained as science is able to divide them at its historical stage of development.

In fact, as early as the 1860s Mach had declared his support for a "new theory of
matter" replacing spatial atoms with aspatial elements. In the 1863 Lectures on
Psychophysics, Mach proposed holding onto macroatoms as spatial shells, replacing
their interiors with complexes of elements. Then he suggested eliminating these shells
altogether. Later in the text, I will develop the thesis that Mach's new theory of matter
was inspired by his reading of the German philosopher J.F. Herbart and enhanced by his
study of space sensations and the new geometry of Riemann and Helmholtz in the
1860s. Many years later, Mach proposed a theory he called "chemical space"--a
generalized Riemannian manifold for physics with a material volume element as its
invariant. In this theory, spatial extension was replaced by an association and
dissociation of Machian elements at a fundamental level of description far smaller than
atoms and spatial volume itself was replaced by what he called a generalized "capacity
factor" across the different energies the elements represented.

As Mach's actual views come to light, we will see a sophisticated natural

philosopher at work. In light of the trajectory from his new theory of matter to his
chemical manifold, I believe that Mach really wanted to eliminate certain properties of
spatiality and extended matter from physical science. This philosophico-scientific
program for the elimination of space was at the root of his inveterate opposition to
mechanical conceptions generally, and source of the real problem that obsessed him
under the name of atomism. As he explained much later in 1911 (after his 1910

Even if the kinetic world-picture, which in any case I consider hypothetical without intending
thereby to denigrate it, could "explain" all physical appearances, I would still hold that the
manifold- ness of the world had not been exhausted, because for me matter, space and time are still
problems which the physicists (Lorentz, Einstein, Minkowski) are also approaching. 63

Mach had held out hopes for a nonmechanical foundation of physics to come from
thermodynamics or electromagnetism. Unfortunately for him, the period when he
worked, c.l863-1900, was during the flowering of the mechanical worldview, when the
scientific imagination had been seized once again by the program of reducing all
phenomena to the quantitative matter and motion of particles and forces. Since the rise
of the mechanical view, the qualitative features of nature favored by Mach, such as
sensations and now his worldelements in matter, had been considered metaphysically
dispensable "secondary qualities," secondary to the properties of extension and motion
described by mathematics. Entirely in this spirit Galileo once said:

If ears, tongues and noses were removed, I am of the opinion that shape, quantity and motion would
remain, but there would be an end of smells, tastes and sounds, which abstracted from the living
creature I take to be mere words. 64

The most influential representative of the mechanical view in the nineteenth

century was Hermann von Helmholtz, who had used the law of the conservation of
energy in 1847 to argue that nature was composed solely of particles in empty space
with forces emanating from their centers and dependent on spatial distance. This
implied that all energy manifestations, whether electrical, thermal or gravitational, were
essentially mechanical phenomena and could be exactly described in quantitative
dimensions of matter, space, and time.

What Mach is saying in his 1911 quotation above is that this reduction, even if
successful, still explained nothing philosophically, since matter, space, and time were
simply the same three great metaphysical problems bequeathed by the seventeenth-
century natural philosophers. Mach's critique influenced many in the succeeding
generation. Einstein, for one, acknowledged that, while a student, Mach's Mechanics
had shaken his "dogmatic faith" in the certainty of mechanical conceptions as the basis
for the rest of physics. 65 In fact, reading Einstein's later "self-obituary," in the
commemorative volume, Albert Einstein Philosopher-Scientist, there emerge strong
similarities between Mach's belief in the value of thought experiments based upon broad
extramechanical principles of experience such as the excluded perpetuum mobile,
showing the independence of energy exchanges from the particular constitution of a
mechanical or thermodynamic system, and Einstein's principle of special relativity,
showing the independence of the velocity of light from the velocity of the emitting
source, which Einstein himself compared to the excluded perpetuum mobile.66


Recent historical scholarship has sought to characterize Mach as a thinker with his own
ideas who still belongs to the context of twentieth-century positivism, and many other
historical contexts besides. Two recent examples to relate Mach more soundly to the
Vienna Circle are Friedrich Stadler's Studien zum Wiener Kreis (1997) and Rudolf
Haller's Neopositivismus: Eine Historische Einfiihrung in die Philosophie des Wiener
Kreises (1993). No one today seriously denies some affinity in temperament between

Mach and philosophers like Camap who seemed to be influenced by him-however

different the products of their thought ultimately became. Slogans associated with
positivism such as "the elimination of metaphysics," "pseudo-problem," and "unity
of science" all do stem from Mach, although he meant something very different by
them, (I think having to do with Mach's unique physiological critique of physical
concepts which was not shared by any of the logical positivists). The Vienna Circle
philosophers were not the "heirs of Mach," nor did they really see themselves that
way. They were the heirs of Russell and Frege and forerunners of their own strand of
analytic philosophy heavily invested in the philosophy of language, it seems to me,
to the detriment of what little natural philosophy they contributed.

I must point out however, despite my own negative assessment, that the history of
the unity of science movement is itself currently being rewritten, as significant
differences emerge among the members of the Vienna Circle and later Unity of Science
Movement. There is also underway a strong advance in historical studies of Mach's
influence on thinkers of the early twentieth century. John Blackmore has released two
new books: one on Mach's Vienna (2001), another on Mach's Prague (2003). Thus, in a
relatively short time, the literature on Mach has reached an advanced state, with the
availability in print of much useful material that can be employed in a comprehensive


As much as Mach delighted in recounting the scientific revolution of the seventeenth

century, he felt that his own century required a different natural philosophy, one that
had added the knowledge of chemistry, thermodynamics, and electrodynamics to the
fundamental sciences and, on top of that, one that was charged with integrating the
facts of psychophysics and perception into one scientific worldview. The greater the
tendency of scientific knowledge to expand, it seems, the greater the inward pressure
toward a metascientific natural philosophy to hold all of the disparate parts together,
although interestingly enough the twentieth century contributed greatly to science
proper without contributing nearly as much toward natural philosophy as previous

Mach believed the integration of psychical data would force any sufficiently
generalized physical science to confront the inner qualitative aspects of nature that
had been masked by the modem, mathematical emphasis on external properties
dominating science since the seventeenth century. Mach thought that qualia in nature
really existed and were the proper object of analysis and study. He even speculated
that there would come a time when such properties could come to be experienced
directly by observers, if biology and psychophysics advanced far enough toward
greater concreteness. This has not been the trend since Mach's death. If anything, the
sciences have tended toward far greater formalism than anyone would have thought

Mach once said he was not a philosopher and cared little what philosophers
thought of his work, and yet it is in philosophy where his ideas about a future science
still seem interesting and capable of further development in close partnership with
the sciences. Perhaps, then, the best title for Mach is the epithet so often given to
Galileo and Newton: natural philosopher.


Ernst Mach was born in Chirlitz, Moravia, in 1838 into a cultured, freethinking family.
Practically for all of the scientist's long life, Austria-Hungary was ruled by Emperor
Franz-Joseph, who reigned from 1848 to 1916. Mach's father, Johann, was an eccentric
who had studied science and philosophy in Prague and who was, for a time, a private
tutor to the sons of the Baron Brethon in Vienna. He preferred farming and his own
schemes, including the surprisingly half-successful idea to raise silkworms in Austria. 1
Mach owed his first education in science to his father's garden-physics demonstrations.

The Mach family was sympathetic to the calls for reform and self-determination that
followed the Revolution of 1848. They even harbored a fugitive professor who had
supported the Hungarian uprising. 2 The new regime of eighteen-year-old Franz-Joseph
cracked down hard on the Hungarian revolt and its friends and got off to an
authoritarian start; Erwin Hiebert writes of the effect of these times on Mach's distrust
of authority:

Parental sympathies in the family home were certainly on the side of the Htn1garian revolution and
critical of the autocracy of the Hapsburg monarchy. Accordingly Mach was brought up in an
environment that nurtured skepticism and =trained critical and stubborn inquisitiveness about the
natural world, politics, religion and the status quo in genera1. 3

In a surprising turn, the latter half of the reign of the bewhiskered and medal-
bedraped Franz-Joseph was marked by increasing liberalization, in which the different
ethnic groups within the empire (Czech, Jew, Hungarian, Romanian, and Serbian) were
knit together by a benignly incompetent bureaucratic empire, familiar from the novels
of Kafka, in which dormant nationalisms emerged with rancor. Mach himself, as rector
of the Prague university, was plunged into a conflict that divided the institution into
Czech and German-speaking halves. 4

Franz-Joseph's reign was also marked by a cultural enlightenment, tolerance of

political criticism (from a spectrum that included Karl Kraus and Austro-Marxists such
as Friedrich Adler), and a flourishing of art and science that is hard to overestimate. To
some degree, the openness of the Austro-Hungarian Empire was due simply to its
ineffective efforts to suppress opposition, as William Johnston remarks in his excellent
book, the Austrian Mind:

Nothing illustrated so well the sloppiness of bureaucracy as the manner in which it handled
censorship of the press. Each morning preliminary copies of every paper were rushed to the
censor, who might order any story confiscated. In its place would appear an empty space, bearing
the word konfisziert. Because papers were read so hastily, frequently a story confiscated in one
paper would be overlooked in another. In such cases, every paper was allowed to reprint the story,
citing the unconfiscated version as its source. 5

Vienna, city of dreams, exhibited a manifest content of authority, rationality, and

science coexisting with a latent reality of the anti-authoritarian, instinctual and peculiar,
a juxtaposition that would reemerge in Mach's work, as it did in that of many of his
famous contemporaries. Sigmund Freud, Ludwig Boltzmann, Franz Brentano, Robert
Musil, Ludwig Wittgenstein, Gustav Klimt, and even James Joyce in Trieste were all at
some point under the wing of this empire "Kak:anien" of Musil's novel, der Mann ohne
Eigenschaften. As Thomas Szasz writes in his introduction to the Analysis of

Parallel to the political and social changes, an atmosphere developed which was very encouraging
to the arts and sciences...The years between 1848 and the first world war in Austria-Hungary were
characterized by a widespread belief in the stability and security of the government and the State
coupled with a belief that forces promoting slow but steady changes toward ever better social
conditions were constantly at work. This at least was the social Weltanschauung of the middle,
upper and still higher classes of this society. Almost all of Mach's life (he died in 1916) was spent
during this epoch which could be called the "Age of Security" (in Central Europe).

Among the postrevolutionary movements that affected Mach directll the most
significant was the Exner-Bonitz educational reform of the 1850s. These two ministers
of education were disciples of the philosopher and pedagogue J.F. Herbart (1776-
1841), who proposed ending the rote learning of the cramschools, pitching education
directly at the native critical and reasoning powers of the student, giving him time to
absorb and fight his way through to new ideas. Given Mach's supreme critical powers
and the strong influence of Herbart on Mach's thought, this influence is not without

The most important science in Vienna at mid-centmy was no doubt medicine. Not
coincidentally, it was the site of the largest centralized hospital in Europe, which, as
Wolfram Swoboda remarks, 7 permitted physicians to observe the progress of diseases
and to perform autopsies in large numbers of cases, since all the patients flowed through
the same institution. The University of Vienna was geared toward the education of
physicians and physiologists, which is evident from the high caliber of talent educated
there, such as the young Freud and Breuer, to name two who worked with Mach's
teacher Ernst Briicke. Mach's first job, in fact, was to teach a course of physics for
medical students.

Swoboda has thoroughly researched the forces that led to a specific brand of
Viennese empiricism and physicalism one generation before Mach. The Viennese
doctors had rejected the vogue of Naturphilosophie from Germany and the various
forms of vitalism that accompanied it. Vitalism was the idea that living organisms
possess within a "vital force" (Lebenskraft) unexplainable by material means, while
Naturphilosophie was a principle-driven amalgam of science and philosophy, loosely
based on the theories of Schelling and Hegel, and a downright bane to medical research
since its abstract "polarities" and "oppositions" sanctioned no specific treatrnents. 8 The
Viennese reaction was a race to the opposite pole of extreme empiricism on the model
of English philosophy, as Hugo Dingler describes:

Mach's first beginnings were made during the time when the struggle against the grmmdless
speculations of the Schelling- and Hegelian school still echoed faintly in the distance. In reaction
against that groundless over-stretching of a completely unproven principle, a most extreme
Empiricist cast of mind, jnst as unfounded in the last anslysis, had established itself: and was
supported by the influence of English philosophy. 9

As Swoboda relates, Mach's two most important teachers, Ernst Briicke and Carl
Ludwig, were students of Johannes MUller, who had founded an institute in Berlin for
the study of human physiology. MUller, the author of the famous Handbuch der
menschlichen Physiologie (1838), was an intriguingly complex personality: a hard-
headed physiologist and experimenter who remained a vitalist and a phenomenologist
on the model of Goethe. As we will see later, MUller even proposed reviving Aristotle's
theory of direct perception in connection with his famous specific-energies doctrine to
replace Locke's representative causal theory. Ludwig and Briicke apparently retained
some of the complexity of their teacher. Mach relates of Ludwig, for example, that he
believed physiological processes were only the "outside" of sensational processes. 10
Briicke was also the teacher of Freud and held a Herbartian doctrine of psychical
"forces" and mental energies that is similar in Freud and Mach. 11

But there were other students of Muller, such as Hermann von Helmholtz and Emil
DuBois-Reymond, who believed that even a good physiological account should be
further reduced to the mechanical atomic theory. First materialism and then
reductionism (in the formula "Physiology as applied physics") was the creed of these
men. The view was very influential in the medical community, as Michael Heidelberger
writes: "In the later fifties and in the sixties materialism enjoyed its greatest popularity
and among doctors the materialistic Weltanschauung...was especially current. 12"

Dingler says that anti-metaphysics (so often associated with Mach himself) was
actually the rallying cry of the generation before Mach in their crusade against
Naturphilosophie and vitalism. According to Dingler, anti-metaphysics was an idee fixe
for Helmholtz, who often used the word "metaphysics" in the crudest sense to mean
anything unverifiable with the human sense organs. "With Mach," Dingler added. "the
concept of metaphysics is no longer so purely superficial." 13

Helmholtz had no compunctions about the conflict between atomism (believing

that qualityless atoms are the ultimate constituents of matter) 14 and phenomenalism (the
position that sensations provide the immediate data of science against which theories
are tested). He believed that sensory qualities and their elements, despite their
importance in providing sensory evidence, were superficial for physics and could be
explained by recourse to the purely quantitative behavior of large numbers of atoms,
similar to the way that a pointelliste painting is a mixture of colored dots. 15 The conflict
seems to have disquieted Mach greatly. Although his early works echoed the standard
view, 16 the young Mach realized early, and with more penetration than the preceding
generation, the ontological conflict between sensations (or their elements) and
sensationless atoms. How could Helmholtz hold to the epistemological primacy of

sensory evidence and yet believe at the same time that this evidence was merely a
superficial property of the natural world, perhaps not a real property at all?

For his part, Mach said later in the "Vortriige" that sensory qualities could not be
just "confused" properties emerging from a large collection of atoms that had no quality
whatever in its parts. "How," he asked, "could the soul suddenly come into a
combination of atoms in the organism, if the kernel did not lie in the individual
atoms?" 17 It seems to have been this conflict that set young Mach on the path of finding
a common scientific foundation for the three sciences he learned at Vienna: sense
psychology, physiology, and physics. 18 As Swoboda has documented, while Mach was
officially a physics student he went out of his way to immerse himself in the new sense
physiology with Brlicke, chemistry, anatomy and even six hours per week in the
dissecting room during the winter semester of 1859-1860. Failing to push through an
application for his Habilitation as a "physicist-physiologist" in 1860, he had to settle for
physics in an unrestricted sense, receiving his venia /egendi on February 24, 1861. 19


Mach's publications for the degree included a coauthored paper on electrical discharge
and induction, the first of a series of papers on the Doppler effect and his first paper on
psychophysics?0 In particular, the Doppler controversy with Josef Petzval, one of his
teachers at the Physical Institute, helped make Mach's early reputation.

Swoboda treated this controversy in some detaif 1 so I will merely summarize the
main points. Everyone is familiar with the Doppler effect from daily experience. When
a whistling locomotive goes by, the tone of the whistle changes; as it is speeding toward
the observer the whistle is high pitched, and as it passes and speeds away the pitch
suddenly drops. The same phenomenon is responsible for the astronomical red-shift of
the spectra of galaxies receding away from the observer. (Mach was one of the first to
propose that the speed of astronomical objects be measured by their light's Doppler

Petzval had argued against Doppler that, among other things, a wave motion
propagates through a medium by the flexing and relaxing of the internal connections
holding the material together. Thus, moving the medium will not get the wave through it
any faster. As Mach added, a sound that reaches us through the air on a windy day does
not alter in pitch. Doppler's effect does not depend on the speed of the wave in a
moving medium but on the movement of the source and the destination with respect to
each another. In the case of a speeding siren, for example, the waves all move off at the
same velocity, but they are emitted from different points each time. As the first wave is
getting off, the source moves-and catches up with it a little before emitting the second

Hence, they tend to "overlap" and superimpose to create a wave that seems to
arrive with a higher frequency. The same effect is created by a motion of the observer
toward the source, because then, conversely, the waves are encountered with greater
apparent frequency. Mach constructed an apparatus to separate the two independent
factors: a tone is heard at the same pitch in still air and moving air, but alters in pitch
with any relative motion of the source and the observer. His article announcing the test
was the conclusion of the controversy.

Mach's debut paper on psychophysics, "Ober das Sehen von Lagen und Winkeln
durch die Bewegung des Auges"23 was also of lasting importance. Here he showed
himself to be a disciple of Gustav Theodor Fechner, whose Elemente der Psychophysik
had just appeared in 1860. Mach had written Fechner on January 14, 1861:

For a long time I have busied myself with mathematical psychology; my goal being to seek out
methods of experiment and observation for psychology similar to those long well-known in
physics. I could find no correct basis for my ideas until, at last, your Psychophysik appeared. I read
the book with affection, and found my expectations greatly exceeded.24

He goes on to say that he has extended these methods in experiments on spatial

vision and the time sense, indicating already the drift of Mach's psychological
researches toward the experience of space and time. In his paper, Mach argued that
small, just noticeable differences in the lengths of lines and opening of angles were
estimated by the induced strain of the eye muscles needed to scan them. It is already
assumed implicitly that sensations of length are not direct retinal perceptions like color,
but due rather to the muscular action of the eye muscles moving in their sockets.
Fechner's law of psychophysics, which will be discussed in more detail later, had been
thoroughly tested for various types of muscular sensations-for example lifting
weights-so Mach logically assumed it would also apply to the eye muscles and to
sensations of small lengths (changes of position) and arcs: "It appears that the
hypothesis is in harmony with experience that the position (Lage) of a line is made
known to us through the sensation of the eye muscles produced as the eye moves along
this line. "25

Mach then pointed out that the exertions of the system of eye muscles were not the
same in every direction. Strain along the diagonal was greater, for example, as it
required the combined work of two sets of ocular muscles, those that move the eyes
from side to side and those that move them up and down. A diagonal was harder to scan
and so produced a sensation of greater apparent length. Mach showed that, indeed,
depending on the situation of a line, its apparent length shrank or grew, and angles
opened and closed. He then appealed to the familiar illusion of the congruent square
and diamond as evidence that the sides of the square looked shorter than the diagonally
situated sides of the diamond.

It cannot be overestimated how important such psychophysiological investigations

were for Mach's physical thought about space and geometry. His experiments with
visual space would eventually demonstrate to him that it was possible to conceive of
manifolds in which properties of length depended explicitly on position and direction,
cases that were in the forefront of investigation in mathematics. In the 1860s, Mach
would extend his investigations in spatial vision to the phenomena of physiologically
similar figures, symmetry sensations (independent "sensations of form" due to the
vertically symmetrical arrangement of the eye), and the physiological (as opposed to
geometrical) significance of straight lines and curves. Many of these sense-
physiological investigations, as well as those into different sensory manifolds, with their
own dimensions and symmetries, eventually prompted Mach to reconsider the way the
properties of physical space were determined.


After his degree, from 1861 until his departure for Graz in 1864, Mach taught as a
Privatdozent in Vienna, where he was an active participant in the brilliant cafe
discussion society. John Blackmore quotes a second hand reminiscence by Ludwig
K.arpath about one of Mach's favorite haunts:

A close relative has told me a lot about a Vienna coffeehouse which I can no longer remember, the
Cafe Elefant which was located in a narrow passage between Stefansplatz and the Graben. Every day
scholars, artists and doctors of medicine and law would gather together. The regulars included
such later famous people as Professor Mach, Lynkeus (Popper), a group of Wagner-oriented
musicians ...and many others. People wandered in about 2pm. and stayed until 2 in the morning,
that is, some were always leaving while others were arriving. Uobroken wit and argument on
philosophical, scientific and artistic matters kept the discussion sharp and stimulating. To a
certain extent the young Dozent Ernst Mach presided over the gathering. His profound
understanding and reflective manner impressed everyone. According to my relative he was one of
the first to occupy himself with the recently published work of Helmholtz on tone perceptions
about which he formed many interesting and instructive conclusions?6

Such was Mach's typically Viennese love of music that, even while on the rocks,
he always managed to afford piano lessons. 27 His daughter-in-law, Frau Anna Karma
Mach, relates that despite pecuniary difficulties he remembered this time with fondness:

After his Habilitation at the university began a time of great deprivation. His father came into
financial troubles, his mother died and this sapped his father's energy. He had to earn his way by
giving lessons and lectures. Yet he remembered this time with great joy because he presided over a
merry company of writers, musicians, artists and painters. He was himself an excellent piano and
organ player. Incidentally, I should like to mention that a large painting hangs in the Prague
Musewn, painted by a friend of his youth Prof. Lauffer. It represents the scene at Siegfried's
fimerary procession, and in the middle is Ernst Mach in the guise ofHaagens.28

Mach was also prominent in the circle of students surrounding Briicke and Ludwig.
In a letter to the Austrian Academy of Sciences of April28, 1913, Mach wrote that he
was "less a student at the university than an apprentice of several members of the
Academy. "29 In fact, Mach "inherited" his first scientific problems from his teachers.
For example, it was Andreas Ritter von Ettingshausen who suggested he construct his
Doppler apparatus. And through Ludwig's attempt to construct a machine to measure
blood pressure, Mach became involved in the analysis of fluid oscillating under
resistance in a tube/ 0 and the mechanics of oscillating systems under resistance that he
later applied to the physiological acoustics of the ear. 31

As a Privatdozent, Mach was not paid to lecture at the university but had to earn
his way by giving private lectures: in 1862 and 1863 on psychophysics and in 1863 and
1864 on Helmholtz's theory of music. At the university, beginning in the fall of 1861, he
taught his course, "Physics for Medical Students," along with "Higher Physiological
Physics" and "Methods of Physical Investigation." In the summer of 1862 he taught
"Mechanical Principles and Mechanistic Physics in its Historical Development."32 As
the title suggests, Mach had already been led to the historical presentation of material. 33
Swoboda points out, however, that historical presentation was not that unusual in the
science of Central Europe. 34 Even up to the time of Max Planck, a working scientist
would often have a strong command of the history ofhis subject.


From the dates of his lectures in the Nachlass it appears that Mach's working method
was to develop his ideas in the lecture room through his teaching, several years before
publication. In his lectures and notebooks, he expanded on ideas that he later worded
far more cautiously in published work for a professional audience. Mach said that he
often experienced disappointment in public discussions, and that, as a consequence, he
concealed his views from his physicist colleagues for years. 35 It is a shame that we do
not possess lecture notes and notebooks from the 1860s (the Nachlass notebooks begin
in 1871) and that instead we must glean his views from his more cautiously worded
publications. More the shame, because the 1860s were the time when Mach developed
practically all of his major ideas: his elements, his redefmition of mass, his critique of
Newton's law of inertia/ 6 analogies between the energies 37 and the extramechanical
significance of the excluded perpetuum mobile principle.

What we have instead are later remarks about his remembered intellectual
development. These remarks describe an intense intellectual "struggle" during the
1860s that led Mach to his theory of elements ten years before the publication of his
best-known works, the Mechanik (1883) and the Beitriige zur Analyse der
Empfindungen (1886). In the preface to the fourth edition (1902) of the Analysis of
Sensations, Mach wrote:

About thirty-five years ago I succeeded by overcoming my own prejudices, in firmly establishing
my present position and in setting myself free from the greatest intellectual discomfort of my life.

This passage dates the end of the struggle to about 1867, during his Graz and early
Prague tenure. I cannot believe that the whole philosophical, physical, and
psychophysical tumult was quieted in the Graz years from 1864 to 1867. Elsewhere, for
example, Mach said 1871 was the year of his "greatest intellectual ferment" 38 when he
"attained stability in his views." 39 I believe the composition of History and Root of the
Principle of the Conservation of Energy in 1871 was the official end of the struggle.
Remarkably, Mach could still refer back to this short, brilliant book unaltered in a
second edition of 1909 as "containing the fundamental ideas" of all of his major
works. 40

Several of the turning points of the struggle, including an intermediate

''psychomonadology" period, are described in the following long footnote to the
Analysis ofSensations:

I have always felt it as a stroke of good fortune that early in life, at about the age of fifteen, I
lighted, in my father's library, on a copy of Kant's Prolegomena to any Future Metaphysics. The
book made at the time a powerful and inextinguishable impression on me, the like of which I
never afterwards experienced in any of my philosophical reading. Some two or three years later the
superfluity of the thing in itself abruptly dawned upon me. On a bright summer day in the open
air, the world with my ego suddenly appeared to me as one coherent mass of sensations, only more
strongly coherent in the ego. Although the actual working out of this thought did not occur until a
later period, yet the moment was decisive for my whole view. I still had to struggle long and hard
before I was able to retain the new conception in my special subject. With the valuable parts of
physical theories we necessarily assume a good dose of false metaphysics, which it is very difficult
to sift out from what should be preserved, especially when those theories have become very
familiar to us. At times, too, the traditional, instinctive views would arise with great power and
place impediments in my way. Only by alternate studies in physics and the physiology of the
senses, and by historico-physical investigations (since about 1863) and after having endeavored in
vain to settle the conflict by a physico-psychological monadology (in my lectures on
psychophysics... ) have I attained to any considerable stability in my views.41

Hugo Dingler speculated that Kant set Mach on the path of studying questions of
space, time, causality, and conservation laws in science. It is easy to imagine Mach
reading Kant's famous examples in the Prolegomena of the enantiomorphic left and
right hands42 and the inverse square law of force under the consideration of a
mechanism communicated in space over three-dimensional shells. Such conceptions of
a priori science would later become the target of Mach's harshest criticism, but that does
not rule out a strong Kantian influence in youth. Mach once claimed that all of his
critical thinking began with Kant,43 and in the Conservation of Energy of 1872, he still
credited "only Kant and Herbart" as philosophical influences on his scientific thought.44

Kant had an enormous impact on the scientists of the ninteenth century, of course.
But it is often difficult to sort out Kant's influence on scientists because the great
Konigsberger's ideas about physics were seen in two ways. Either he was the author of a
mathematical reine Naturwissenschaft of appearances, holding that principles of
conservation, inertia, and the like could be deduced from spatial and temporal

considerations alone, i.e., the possibility of a spatial and temporal ordering of events
demanded that certain a priori synthetic principles be true, like the conservation of
substance. Or else, Kant was the one who declared space and time to be mere forms of
appearance and not the true face of nature at all. And this left open the way for
speculation about "intensive" natural processes at a more fundamental level beneath
spatial and temporal description, as had been the case in the Leibniz-Wolff
metaphysics. Actually Kant meant both,45 but which one is taken to be "Kant's view"
often depends on whether the transcendental aesthetic or the transcendental analytic
looms larger in one's mind.

Both views probably held some sway with Mach. In an autobiographical fragment,
Mach claimed that he owed to Kant the destruction of his "naive realism," by which he
probably meant Kant's sharp refusal to extend . spatial properties to things in
themselves. 46 At this point, Mach was probably a "naive realist" about an external world
of matter, space, and time encountered in his physics reading, otherwise Kant's claim
would scarcely have shocked him so. Mach also claimed that he "with the help of Kant
the metaphysician eliminated all tendency toward metaphysics within himself." By this
Mach probably meant his rejection of the thing in itself, referred to above in the
experience of the summer day, which Mach later described in a more detailed way as
the rejection of a certain type of inference in which a reified sensory complex or object
of appearance is reintroduced in thought behind appearances.47 Appearances are
adduced in support of appearances. 48 Mach's criticism of the inference to the thing-in-
itself through such a turnabout maneuver was indeed one that he would later apply to
every one ofhis antimetaphysical criticisms.

After Kant, Mach acknowledged a shift to idealism, which he believed was latent
in Kant and which he later found in Berkeley and Hume, although it turns out he really
meant Herbartian psychology. 49 For example, he says that up to the point when he
developed stability in his views that he was acquainted only with Kant and Herbart. 50
Mach later bitterly denied being a Berkeleyian and it turns out that he did not read
Hume's Treatise ofHuman Nature until at least 1907-8, long after his struggle of ideas
was over. 51 Rather, Mach's true influences of this time were G.T. Fechner and
Herbart, 52 both of whom represented sophisticated post-Kantian idealisms.53

Elsewhere, Mach recounted some of the steps from his intermediate stage of the
Herbartian-Fechnerian "psychomonadology" in the Vortriige iiber Psychophysik to his
mature Elementenlehre. For example, he said that after he rejected the Kantian thing-in-
itself, he also eliminated the unchanging I, the "simple subject" of Herbart. 54 In this he
was influenced by the German aphorist Georg Christoph Lichtenberg, who had
suggested that instead of saying "I think" like Descartes, one should say Es denkt, there
is a thought, to indicate that the thought simply is, in an unrestricted sense, without
being confmed to a subject who thinks it. The same would later be true of Machian
sensations, which do not presuppose an ego complex in which to be contained, but
actually comprise this ego and need no further substrate. Mach followed the same
eliminative procedure in getting rid of material objects by breaking them up into
functions of elements:

The older generation, especially the physicists and chemists, will be alarmed by this proposal not to
treat matter as something absolutely constant, but to take as constant instead a fixed law of
connexion among elements which in themselves seem extremely unstable. Even younger minds
may find this conception difficult; but the view is inevitable, though I myself at one time went
through a great struggle in order to arrive at it. 55

By breaking up both complexes of the thing and the I, Mach could now deal
directly with the elements, building up objects and egos from various functional
dependencies. But there was still a lingering dualism to be dissolved between the
psychological elements, liberated from the I, and the physical elements, liberated from
the thing. How could they be the same type of element if one was psychological and the
other physical? Mach said that he advanced slowly to his neutral monistic conception of
the elements by overcoming "dualist and panpsychist" tendencies:

Of all the approaches to my standpoint, the one by way of idealism seems to me the easiest and most
natural. And connected with this is the fear of pan-psychism, which at the same time seizes
my readers. Many are the victims that fall prey to pan-psychism in the desperate struggle between
a monistic conception of the universe and instinctive dualistic prejudices. In my early youth I had
to work through these tendencies myself. 56

A second line of attack started from his work in physics and his conviction that a "new
theory of matter" was needed to replace the image of spatially extended atoms common
during the rise of the mechanical worldview. Both of these starting points will be
examined in depth in the next two chapters.


From 1863 to 1872 Mach was preparing a philosophical "new theory of matter,"
with some important parallels to his work in physics. He felt a new theory was
needed after conducting a series of fault-finding arguments with spatial "billiard
ball" type atoms. Mach advanced several types of arguments during this period,
some physical, some philosophical, and some sense physiological observations.

Mach's first attacks on atomism occurred in two early papers on emission

spectra and in his textbook Compendium der Physik for Mediciner (1863). At this
time, Mach perceived difficulties with contemporaneous atomic views, but had not
yet proposed an alternative. As he said, Mach used the Compendium, a text for his
course of physics for medical students, as an organ to communicate his views on
mechanical atomism:

In the year 1862 I drew up a compendium of physics for medical men in which because I strove
after a certain philosophical satisfaction, I carried out rigorously the mechanical atomic theory.
This work frrst made me conscious of the insufficiency of this theory and this was clearly
expressed in the preface and at the end of the book, where I spoke of a total reformation of our
views on the foundations of physics.'

In the book, Mach approached atomism cautiously, laying out the indispensable
features of discreteness, impenetrability, and obedience to the law of inertia? His
atoms exerted on one another a force of some nature falling off with the square of the
distance. Atoms were also surrounded by cloud like Aetheratome, which repelled one
another by forces of contact. Intra-atomic forces were not discussed. In fact, Mach
wasn't even sure whether the atoms were chemically different among themselves or
whether chemical differences emerged from varying combinations of identical
atoms. 3 Such was the very rough nature of atomism of the time.

Mach then gave several arguments as to why discrete particles were a more
favorable view of matter than the competing theory of a material plenum. For
example, the conduction and radiation of heat through matter in wave motions
implied that bodies consisted of strata of discrete particulate matter capable of
vibration. (As Swoboda has shown, Mach took over many of these pro-atomic
arguments verbatim from G.T. Fechner's Atomenlehre, whose atomic conception was
the model for Mach's Compendium treatment.t Such was Mach's skill in rehearsing
the conventional arguments and explanations of the theory that Erwin Hiebert writes:
He had mastered the argwnents of the atomic theory very well ... although he later denounced
the atomic theory vehemently and resolutely, he manifestly was not engaging in such polemics
because of blindness or ignorance of the strengths and weaknesses of these conceptions. 5

However, Mach went on to complain that the proposed reduction of all physical
processes to the atomic level had not gone through. He then considered the possibility
that atomism had outlived its usefulness, and that this hypothesis would give way to a
deeper "metaphysical" theory of matter:

Here I have let the atomic theory step into the foreground everywhere, not indeed in the belief
that it is the last and highest, and doesn't itself need further support, rather because it brings the
appearances into a simple and visualizable association. The atomic theory, when one may
express it thus, can be considered as a fonnula, which has already led to many results and will
lead further to still more. In fact, whatever kind of metaphysical opinion may arise in the future,
the results won according to the atomic theory will be translatable into them, as one may
express fonnulas in polar coordinates or in parallel coordinates.6

He inclined toward such a new theory at the end of the book:

We have now become acquainted with a series of physical phenomena and Jaws, which we, so
far as it was possible by elementary means, have embraced under the point of view of
mechanics and the atomic theory. A complete and strict tracing back of physical Jaws to a few
exact principles is no longer possible today. The gaps are still too great. For heat and light
phenomena we have at least found the most general contour of a theory and at least this contour
rests on fixed foundations. We have not yet won so much insight into the phenomena of
electricity and magnetism. In these areas are rather empirical rules, to find one's way about in a
large quantity of facts, than a real theory. However in this lies a stimulus to further research,
and we hope that this will attain the goal, granted perhaps after a total refonnulation of some of
our fundamental physical views. At least the peculiarities of electrical and magnetic forces seem to
point to the necessity of such a transformation .7

In the Compendium Mach's main objection to the mechanical view concerned

Weber's law of electrical force between two moving, charged particles, which was
also puzzling the world of physics at the time. The law said that, like gravity, the
electrical force of attraction between two particles in motion was proportional to the
charges and inversely proportional to the square of the distance. But, unlike gravity, it
was also proportional to the velocity and the acceleration of the two charges.
Presumably, if both gravity and electrical forces were mechanical phenomena at
bottom, such asymmetries would not exist: "This dependence of forces on velocity is
still a riddle at present; it never happens when two bodies exert gravitational forces
on one another for example. "8

Mach held the view that thermal or electrical phenomena might be more
foundational sciences than mechanics and said in 1909 that "this thought seems to
be becoming an actuality." But Mach's early objections came to focus on a
different philosophical issue: the conception that atoms were extended in space. He
said that his work on the spectra of elements convinced him that the underlying
atomic processes were not three dimensional:

My attempts to explain mechanically the spectra of the chemical elements and the divergence of
the theory with experience strengthened my view that we must not represent to ourselves the
chemical elements in a space of three dimensions. I did not, however, speak of this candidly
before orthodox physicists. My notices in Schlomilch's Zeitschrift of 1862 and 1864 contained
only an indication of it!
While still an upholder of the atomic theory, I sought to explain the line spectra of gases by
the vibrations of the atomic constituents of a gas molecule with respect to each other. The
difficulties which I encountered suggested to me (1863) the idea that non sensuous things did not
necessarily have to be pictured in our sensuous space of three dimensions. 10

In the earlier paper of the two, "Uber die Spektra verschiedener Korper," Mach
adopted the premise that when the atoms of a gas or heated substance were shaken,
oscillations in the distances between the atoms suspended in a molecule produced the
bright lines visible in their emission spectra. There was no explicit skepticism about
the existence of atoms in this article; indeed, he seemed to need the atomic theory to
preface his argument, but Mach limited himself in a tell tale way to the observable
facts of spectral lines and the distances, without relying on the atoms as such. It was as
if he proposed to consider distances without answering the question, "distances
between what and what?" Perhaps for this reason, among others, Mach's article struck
physicists as "zu naturphilosophisch." 11 In response to his critics, Mach followed up
with an 1864 article, "Vorlaufige Bemerkungen iiber das Licht gliihender Gase" in
which he said: "I must remind you that I by no means consider the atomic theory as
something established in itself, but rather as a useful temporary empirical formula; as a
kind of regula falsi to be used in drawing closer to the truth." 12

He then reprised his argument from the Compendium that the results of the
atomic theory could be "translated into a different 'metaphysical' view of matter which,
henceforth, would prove to be more cogent." But now he explicitly declared his
support for an alternative: "I personally am now already very much inclined toward
such a translation. And I shortly hope to show that my remarks by no means hang
together so necessarily with theories, but can be understood very readily as an
expression of the facts." 13

Like many physicists of his time, Mach had wanted to investigate the constitution
of matter and saw the phenomena of light and heat radiation emanating from bodies as
the key to uncovering it. For example, on September 26, 1860, Mach sent a letter to
G.R. Kirchhoff describing a method for isolating the spectrum of a distant
phosphorescing body from the spectrum of its reflected light by using a polarizing
filter, which got a gruff answer and not the promise of financial support Mach
anticipated. 14 Unlike most physicists of his time, Mach came to see the wavelike
transmissions of energy not merely as a sign of the constitution of matter, but as
comprising a more fundamental level of description. In his notebook of 1874, for
example, Mach wrote that he considered the passage of matter through space to be a
wave, or a propagating "potential difference": "The movement of matter is a wave. A
difference that progresses forward ... Light as periodic potential." 15

In itself, a wave theory of matter does not imply a deeper level of description
under matter, but this is not the only indication that Mach came to favor a subspatial
theory involving waves, even this early. In his important 1866 paper, "Uber die
Entwicklung der Raurnvorstellungen," he wrote that he considered light to be a
chemical separation and combination, because light interacted chemically with bodies
and solutions. 16 His idea here was not only to compare light to a chemical process,
but also to conceptualize chemical processes within matter as similar to light waves.
In particular, he pointed out that chemical changes in matter might not be due to the
motion of bodies lying in space "next to one another" but that the phenomena of
motion might really be qualitative "chemical" changes. By "chemistry" Mach meant a
general natural science in which every sort of change of physical state (whether
mechanical, thermodynamic, or electrical) could be considered under one roof. Had
he not been frustrated in his attempt to study physics with Franz Neumann in
Konigsberg, 17 it is likely that he would have pursued these interests in a physical
framework. Instead, he pursued his studies in sense physiology and developed his
unusual ideas further in philosophy.


A very significant influence on Mach's theory of matter and space came via the
German philosopher Johann Friedrich Herbart (1776-1841), whose psychology and
metaphysics were enormously influential in Germany and in Mach's Vienna. Mach
said that in 1862 he was studying Herbart's "intelligible space," a construction in the
philosopher's Allgemeine Metaphysik:

I was busied at the same time with psychophysics and with Herbart's works and so I became
convinced that the intuition of space is bound up with the organization of the senses, and,
consequently that we are not justified in ascribing spatial properties to things which are not
perceived with the senses .. .At the same time the quite arbitrary and, on this account, faulty
!imitation of the nmnber of dimensions in Herbart's derivation of "intelligible" space struck me.
By that, now, it became clear to me that, for the understanding, relations like those of space,
and of any nmnber of dimensions, are thinkable. 18

I will consider Herbart's views and their influence in the detail they deserve in
the next chapter. For now, suffice it to say that Herbart was part of a realist reaction
against Kant. He adopted a Leibnizian19 monadology and came to believe that
simpler intensive properties and their functional causal relationships could replace the
more complex property of extended magnitudes. In Herbart's Metaphysik, space did
in fact result from a level beneath, occupied by spaceless monads called Wesen
(beings, essences) and their qualities. Herbart described the Wesen as "pressing" one
another through their qualities, as if these were analogous to forces. Thus qualities,
for Herbart, were not passive sensations but individualized causal powers. Unlike
Hurnean qualia, they were not merely associated but bound by intrinsic causal
relations inseparable from the qualities themselves. A sensation that does not react
upon or cause a change in another is not a sensation, for Herbart, as would be the

case for Mach also. These confmed and confining pressures, overcoming and giving
way to one another, ultimately generated quasi-spatial waves that moved abstractly
through the community of Wesen as the underlying intensities changed. It is very
likely that when Mach referred to matter as a wave in his notebook, passing over
differences of potential level, he was drawing directly on Herbart's space

Herbart's Wesen also provided a good transitional entity for Mach, since he did
not have to give up on atoms altogether. Herbartian Wesen were spaceless and really
little more than nodal points in the flux of intensities. The stability of the Wesen
depended on the equilibration of the pressures exerted by their qualities; hence, it was
really only the relations of the qualities that mattered because these determined both
the stable forms of objects and the tracing out of extended regions of space.


Mach's first attempt at his own theory of matter was presented in his 1863
"Vortrage tiber Psychophysik," and shows a strong Herbartian influence. Here Mach
even declared that he wished to eliminate the hypothesis of "spatial atoms" 20 in favor
of Wesen und Kriiften, i.e., a Herbartian manifold, possibly multidimensional in
nature. He expressed his opinion that space resulted "very probably from a mediate
interaction of a plurality of Wesen," 21 and then cited his own paper on gaseous
spectra as a possible means of verifYing this construction:

In ... (Schomilch's Zeitschrifl for Mathematik VII. Jahrg. Ill. Heft) I have shown how one may
deduce the nature and constitution of gas molecules from the results of the latest spectral
investigations of light from glowing gases. From the oscillations that shake a system, one may
draw conclusions about its constitution .. .If we wish to mix in as few hypotheses as possible,
then we must say that physics leads always and everywhere to the idea that it is a finite number of
Wesen and forces that are active in the phenomena. 22

The "Lectures on Psychophysics" were addressed to Viennese doctors interested

in the latest sense physiology of Fechner, Wundt, and Helmholtz. Mach used the
occasion to call attention to problems in the unity of the sciences raised by the new
psychology. For instance: If psychology employed a basic unit of sensation which
could be made ever smaller how could these ever be reduced to the matter-and-
motion of quality less atoms? Again Mach found himself in conflict with the
mechanical philosophy. In response, he proposed a theory capable of accommodating
the sensations of psychology and the inanimate matter of physics in one system, but
only by doing away with atoms in favor of Herbartian Wesen and their intensive
qualia. Thus, instead of eliminating qualities in favor of a swarm of mechanical
atoms, Mach allowed that even inanimate matter was made up of monadic qualities of
its own, an "inner side" of Nature:

How do we have to think of these Atoms? Colored, lighted, sounding, hard?

These are sensory properties which atoms only have in community with one another, for all
physical phenomena result from a plurality of atoms. We cannot attribute these properties to
one atom alone. We cannot even think of atoms as spatially extended. For as we have seen
space is nothing original and results very probably from a mediate interaction of a plurality of
Wesen.--Physicists, too, have already felt the difficulty of imagining atoms concretely, and for that
reason some consider the atoms as mere centers of force. But a center of force for itself is actually
nothing. And what does it mean anyway to say one center of force acts on another? Let us simply
admit it! We cannot reasonably succeed in giving any kind of outside to atoms; should we think
anything at all we must attribute an inner side to it, an inner side in some respects analogous to our
ownsoue 3

By enlivening matter with qualities similar to sensations, Mach was following his
second, no less influential, mentor of those years, G.T. Fechner. In his Psychophysik,
Fechner had proposed a theory of mind in which mental properties were simply the
appearance of the brain to itself, a monadic perspective available only to the
possessor, of what appeared to the outside as impulses and brain tissue. 24 Mach's
Monadenlehre in the "Vortriige" was, as he said, an attempt to find a Fechnerian
solution to the problem of inanimate matter and its relation to mind; he also said he
shared Fechner's opinion that all of nature possessed an inner side analogous to
mind. 25 Thus, Mach's critics among his physicist colleagues had correctly smelled out
in his earlier paper a naturphilosophisch construction of matter out of physical qualia
brewing behind the positivistic proposal to examine only the "observable
phenomena" ofthe spectral lines.


Although the "Vortriige" theory of matter was thoroughly Herbartian in nature, Mach
quickly realized that Herbart's Wesen were little more than a formal pattern in the
flux, or a function entirely determinable by qualitative pressures and their intensive
relations. Thus, by eliminating these last vestiges of matter Mach quickly arrived
at his mature theory of elements and functions. In fact, as Andreas Laass notes, Mach's
philosophical move of eliminating the Wesen paralleled exactly his proposal to ignore
atoms and concentrate on their potential to emit spectrallines.26 These developments
even occurred at roughly the same time, from 1862 to 1863.

Mach also rejected Herbart's claim that his space construction was limited to a
three-dimensional manifold, a criticism that may also have occurred to two other
prominent readers of the Metaphysik: Bernhard Riemann and Hermann Grassmann.
Mach pointed out that this limit was psychological, not logical or even physical, and
hoisted Herbart with his own petard, as Herbart himself had often pointed out with
irritation that conceivability and Anschaulichkeit had nothing to do with what was
physically or logically possible.

In Mach's Conservation of Energy of 1872 and in his 1871 article, "Uber die
physikalische Bedeutung der Gesetze der Symmetrie," he parlayed this critique of
Herbart's intelligible space into an argument against three-dimensional molecules,
writing that the "physiologically based" assumption of 3-space for physics artificially
limited the number of possible distances between atoms. This "metaphysical" error
arose by assuming those atoms were already embedded in 3-space, rather than letting
space arise from their physical interactions.Z7 In order to do the latter, Mach drew up
a table of the number of atoms in a molecule and the distances possible between them
in three dimensions. It could be expected that given a certain number of distances, the
rest would be determined. For example, in a molecule of five atoms in 3-space, nine
distances determine the tenth. However, experimentally, if those distances were
given, or rather if physical magnitudes were measured that varied with the distances,
and the rest were not determined, Mach could argue that space in the small was not
three dimensional. Mach also suggested, in line with his idea that chemical elements
not be represented spatially, that n distances or pairings between atoms actually be
replaced with some physical property instead. He used the example of heats of

The heat of combustion generated by a combination gives a clearer picture than any pictorial
representation. If, then, it were possible, in any molecule composed of n parts, to determine the
n(n-1 )/1.2 heats of combination of every two parts, the nature of the combination would be
characterized thereby.28

Since he attributed the spectral lines to oscillations in the n parts of a molecule,

Mach implied that he had found more lines due to independent oscillations than
possible between atoms suspended in three-dimensional space. Thus, extending the
physiological property of three-dimensionality to physical space could be exposed as a
metaphysical error. 29

That threat, although it is only expressed fully in the 1871 "Symmetrie" paper
and the 1872 Conservation of Energy, seemed already to hang over Mach's earlier
articles on emission spectra addressed to physicists. Of course, Mach hardly bothered
to consider the idea that chemical differences were intra-atomic, suggesting that the
form of atomism under attack by him was really the a-tomos, indivisible
homogeneous matter, with no internal structure.

In his "Symmetrie" paper, Mach made good on his threat, applying his
physiological critique to the "billiard ball" atomic theory, and saying "those theories
which derive all phenomena from the motion and equilibrium of the smallest parts,
the so-called molecular theories, are already set tottering due to the theory of the
senses and of space, and one may say that their days are numbered. "30 He then
proceeded to show that a space of one dimension was given for the sensations of
sound, a Tonraum, or tone space, analogous to 3-space but possessing fewer
dimensions. Tone space had certain properties that could be discovered by
experiment; for example, melodies could be recognizably preserved through a

transfonnation of major into minor keys, just as certain figures in visual space
retained properties preserved through transfonnations in place, such as sliding and
rotation. Mach then imagined creatures in tone space trying to formulate a physical
theory by imagining that the world, too, is a one-dimensional manifold:

If someone who could only hear wanted to try and develop a world-view in his linear space, he
would come up considerably short, in that his space is not equal to the many-sidedness of the
real relationship. It is no more justified when we think we can press the entire world, even the
unseen parts, into the space of our eye. But all molecular theories fall under this case. Yet we
possess a sense which, in respect to the many-sidedness of the relations it is capable of
contemplating, is richer than any other. It is our reason. It stands above our senses. It alone is in
the position to establish a lasting and complete worldview. The mechanical worldview has
achieved enormously much since Galilei. But it must make way for a freer perspective. 31

He then added in a footnote (deleted in some editions): "It follows from this that
the dependence of natural phenomena be expressed through relations of number, not
spatially or temporally." 32 Here we can see immediately how Mach's physiological
researches on the properties of sensory manifolds such as the Tonraum directly
influenced his view of physical space. But, as his footnote indicates, Mach was also
looking toward the elimination of spatial properties altogether.

In his Conservation of Energy, Mach put forward an element-and-function

Elementenlehre, capable of representing physical phenomena without the inclusion of
spatial and temporal properties in the fundamental description. Even Mach's
conception of functional dependence did not smuggle in either space or time, he

I think I must add, and have already added in another publication, that the express drawing of
space and time into consideration in the law of causality is at least superfluous. Since we only
recognize what we call time and space by certain phenomena, spatial and temporal determinations
are only determinations by means of other phenomena. 33

Mach demonstrated considerable mental flexibility in contemplating aspatial and

atemporal functional connections between events, criticizing the idea, which he
attributed to Wilhelm Wundt, that locomotion, or change of place, was the only kind
of change admissible in physical science. Qualitative changes were equally
admissible, he maintained, if "a thing is destroyed in one place and in another an
exactly similar thing is created. "34 Kevin Mulligan and Barry Smith consider Mach's
abstract functional dependencies as a forerunner to those aspatial, atemporal concepts
of determination that became current in phenomenology, for example I suppose
intentional relations between minds and objects. 35

Already here in his early work, one can see that Mach was more than a mere
relationist about spatial properties, if by "relationism" we mean that spatial properties
are shorthand for relative spaces or physical relations that themselves may be spatial
in nature. If the physical relations themselves involve measurements or other
operations that presuppose space, relationism really does not deserve the name of an

elimination of space at all. Mach really ought to be called a constructive eliminativist

of space, in the sense that, on his view, space and spatial relations (perhaps also
temporal ones) should be replaced with a construction of those properties and
relations in terms that do not involve space at all, i.e., the non spatial elements and
functions of the Elementenlehre.


Mach's attack on metaphysics is one of the best-known elements of his philosophy.

And as much as that critique was against physically isolated Dinge an sich or a priori
necessary principles, it is clear from the "Symmetrie" paper that the original sense of
antimetaphysical critique during his struggle harkens back to Mach's training as a
sense physiologist. Anti-metaphysics, in this case an attack on metaphysical atoms,
was originally an attempt to separate the pure conceptual content of physics (given by
the elements and functions) from the residue of physiological properties that helped
human beings (physicists included) to visualize physical events. 36 A typical, if
elementary, metaphysical error is committed by the investigator who finds that
images of objects on the retina are oriented "upside down" and wonders how the
person sees them "right-side up." 37 "Up" and "down" have different meanings in the
manifold of sight from the meaning used when speaking of the physical orientation of
objects outside. The error is one of interpolating properties due to the human
apparatus of sensibility to events and objects when they are outside the range of that
sensibility. For example, the property of three-dimensionality is one that
physiological space exhibits because of the arrangement of the sense organs. In
nature, however, three-dimensionality ultimately depends on how many physical
conditions may be chosen independently of others, a criterion that may also be met
but not by pointing to the fact that visual space, or the space of human imagination, is
three dimensional.

Thus, the concepts within physics that Mach tended to oppose were those in
which he felt that sensory properties were uncritically present. (A parallel
development was already underway in mathematics where uncritical intuition was
discarded in favor of abstract methods of proof and logic.) Mach thought absolute
space in physics was a vestigial image of a real sense-physiological experience,
which does orient the body in a semiabsolute sense by means of the otoliths and the
semicircular canals of the ear, as he himself showed in his Sensations of Movement.
The problem is that such experiences then come to be attached to (what should be)
more abstract thinking. Mach said this later in his critique of absolute space in the
Mechanics: "I have a faint suspicion that the persons who imagine they have
conceptions of absolute motions, in the majority of cases have in mind the memory
pictures of some actually experienced relative motion." 38

No doubt this is what Mach meant by his autobiographical remark that "with true
physical theories we often absorb a dose of false metaphysics," which must then be

purged from the sciences by considering the sense-physiological source of memory

images and visualizations. Hence also Mach's claim that studies in the physiology of
the senses actually cured him of many of the metaphysical ideas he had learned
through his physics training and led him to the idea of a sense-physiological critique
of physical concepts.

But what about metaphysics in the standard philosophical sense of extra-

empirical or "beyond physical" speculation? If Mach had been an opponent of all
metaphysical speculation, he would never have been open to the space constructions
of Herbart, nor would he have taken those speculations as important to physical
science and geometry. The idea that Mach was drawing certain of his scientific ideas
of matter and space from a German metaphysician may seem strange given Mach's
supposed antipathy toward metaphysics generally, but I do not see any hostility at all
on Mach's part to metaphysics in the Aristotelian sense of "the most general natural
science," although Mach strenuously objected to using the term that way? 9 Yet as
early as the Compendium, Mach had acknowledged the role of philosophy in natural
science, alerting his readers that philosophical work continuous with science was
necessary to see physical problems from another side:

Where philosophical questions are touched on, which was not completely to be avoided, I
restricted myself to the ground of natural science, in part because this standpoint has a certain
authority, in part also in the conviction that one must have tasted of this outlook for a time to
feel the necessity that the investigation be pursued from another side; some never get this far. 40

Mach himself frequently considered philosophical questions in the foundations

of science from the "other side": questions such as discreteness versus continuity,
locomotion versus qualitative change, symmetry versus asymmetry. In his
"Definition of Mass" article of 1868 and the 1872 Conservation of Energy, Mach
even referred to the ostensibly "metaphysical" principles of causality and sufficient
reason as the logical root for the excluded perpetuum mobile principle, although he
indicated that such principles as "every event has a cause" were empty unless
experience informs us as to what to put in for "causes." Mach also applauded Julius
Robert Mayer's use of heuristic metaphysical principles of monism in seeking a
concept of energy.41 And Mach's later reception of speculative endeavors in his
Knowledge and Error really does endorse what seems to me a broadly Aristotelian
view of the relation between science and natural philosophy:

The ultimate end of all research is the same. This shows itself also in the fact that the greatest
philosophers such as Plato, Aristotle, Descartes, Leibniz and others have opened up new ways
of specialist inquiry while scientists like Galileo, Newton, Darwin and others have finnished
philosophic thought without being called philosophers. Yet it is true that what the philosopher
regards as a possible starting point appears to the scientist as the distant goal of his
work. .. Through its attempt to summarize the most general features of large areas, philosophy
has gained ample experience in this line, even learning gradually to recognize and avoid some
of its own mistakes that the philosophically untrained scientist is almost bound to commit even
today. However philosophy has finnished science with some positive notions of value too, for
example ideas of conservation.42


Now let us return to the development of Mach's own natural philosophy. In the
next step, after discarding aspects of the "vain" Monadenlehre of the Lectures on
Psychophysics, Mach presented his new theory in the "Bemerkungen tiber die
Entwicklung der Raumvorstellungen" (1866), and it is, as expected, an attack on
spatial representations in physics. Here numerous aspects of Mach's mature
Elementenlehre made their first appearance. Objects were no longer point like
monads but were now completely broken up into qualities and their functional
relations. Indeed, Mach began by analyzing spatially extended matter into its
constituent pressures:

It cannot be my intention here to criticize our conceptions of matter whose insufficiency is,
indeed, generally felt. I will merely make my thoughts clear. Under matter, let us imagine,
then, a something in which different states can occur, say for simplicity a pressure in it,
which can become greater or smaller. 43

He then attacked the representation of spatial distances, claiming that the law of
force in nature holding between pieces of matter need not be thought of spatially, but
rather perhaps distances themselves could be derived from the intensity of forces:

Physics has long been busied in expressing the mutual action, the mutual attraction (opposite
accelerations, opposite pressures) of two material particles as a function of their distance from
each other--therefore of a spatial relation. Forces are functions of the distance. But now, the
spatial relations of material particles can, indeed, only be recognized by the forces which they
exert on each other. Physics, then, does not strive, in the first place, after the discovery of the
fundamental relations of the various pieces of matter, but after the derivation of relations from
other, already given, ones. Now it seems to me that the fundamental law of force in nature need
not contain the spatial relations of the pieces of matter, but must only state a dependence
between the states of the pieces of matter 44

This proposal is somewhat similar to his attempt to use heats of combination to

substitute for pairwise properties or "distances" within molecules in the Conservation
of Energy. What is striking about Mach's view of matter and distance is how both
conceptions are to be reduced to pressures and nothing but pressures. Both pressures
of contact (due, ultimately, to its inertial resistance and proportional to mass) and
pressures due to fundamental forces (due to gravitational attraction and proportional
both to mass and to distances between attracting masses) could both be set down as
the same kind of pressure, magnitudes of homogeneous dimensions that could be
compared with one another, without intermediary conceptions. Mach then goes even
further and concludes his synopsis by imagining a physics in which all of the pressure
states could be set down as mutually dependent JUnctions of one another, a
completely abstract dependence more basic than, and underlying, the forms of space,
time, and ponderable body:

If the positions in space of the material parts of the whole universe and their forces as functions
of these positions were once known, mechanics could give their motions completely,45 that is to
say, it could make all the positions discoverable at any time, or put down all positions as
functions of time.

But, what does time mean when we consider the universe? This or that "is a function of time"
means that it depends on the position of the swaying pendulum, on the position of the rotating
earth, and so on. Thus "all positions are functions of time" means, for the universe that all
positions depend upon one another.

But since the positions in space of the material parts can be recognized only by their states, we
can also say that all the states of the material universe depend upon one another.

The physical space which I have in mind--and which at the same time contains time in itself--is
thus nothing other than dependence ofphenomena on one another. A complete physics which
would know this fundamental dependence, would have no more need of special considerations
of space and time, for the latter considerations would be included in the former knowledge.

Mach thus considered space and time to be concepts of measurement, not

primary qualities or basic concepts of science. Whatever it is that space and time
are measurements of are the basic entities. And these seem to be Mach's elements
if, by an element, Mach meant something like a pressure of sorts. If Mach meant
pressure to stand for force, then elements were forces; but if Mach meant pressures
to represent forces that act over elements of time and which are reciprocally
opposed by restraining forces, as I think more likely given their reciprocal
functional relations, then elements were energies, or even better, elements of
action, like quanta. In this "Raumvorstellungen" program, the relations of matter
across spaces can be cashed out in terms of elements and functional relations, with
the explicit goal of eliminating space and time from the foundations of physics.


The last passage of the "Raurnvorstellungen" paper introduces Mach's important

principle of the functional dependence of the elements on one another. He clearly
considered the concept of function to be much more solid than any Hurnean principle
of "loose association," for example. Machian elements were transitory impermanent
states, but the fact that the intensities of the elements were bound together and
brought to zero in a function, Mach believed satisfied any desideratum of a more
permanent substrate in which passing qualities inhere. 46 Mach felt that expressing the
permanence of body as a system of equations (as analytical mechanics does) had
distinct advantages vis-a-vis the visualization ofbrute matter and bulk:

Indeed one is used to regarding the conservation of weight or mass as a direct proof of the
endurance of matter. Only this proof evaporates, when we get to the bottom of it, into such a
mass of instrumental and intellectual operations, that in a certain sense it only constitutes an
equation which our representations picturing facts in thought must satisfy. The dark clwnp that
we involuntarily add in thought, we seek in vain outside our own thinking.47

Hence in the Conservation ofEnergy Mach felt the need to assume only elements and
functions as the essential features of nature. He wrote down an expression:

F(a,~,y,8, ... ,m)=0.

a=(~,y,8, ... ,m).
~=(a,y,8, ... ,m).
y=(a,~,8, ... ,m).

These equations illustrate the fact that if one group of elements is to undergo
variations, other compensating changes must occur elsewhere. If not, then it would
be possible to imagine a source of change that could be imagined to operate, from
which further changes could be obtained without putting anything into the process,
a perpetuum mobile. The variations of the elements Mach refers to seem to express
changes of energy, because all of the examples he gives in the Conservation of
Energy are processes that involve the uncompensated generation of work. 48 At
least one aspect of Mach's elements, then, besides being characterized obscurely as
pressures in the "Raumvorstellungen" paper, is the fact that their variations can
involve or induce energy changes, and that they act and react upon one another in
patterns of dependence that can be set down in functions. 49 But this functional role
is not something superadded to the changes of intensity the elements undergo
rather, the elements seem embedded in various sorts of "oppositions" intrinsically
as part of their nature. It is likely Mach simply took over Herbart's view of
qualities as always occuring in pairwise oppositions of equal and opposite
magnitudes a Ia Newtonian action and reaction. As we will see, "quality" for
Herbart included an embedded sense of direction or opposition against other
qualities. Imagining red as not being directed against its opposite green would be,
for Herbart, like imagining left not opposed by right.

Furthermore, because functional variations were not thought of as spatial or

temporal in nature, but rather as what Mulligan and Smith call non causal
dependencies, I believe that it is best to represent them as changes in the intensities
of those elements that are the arguments of a given function. The matters that are
dependent on one another are the intensive magnitudes of the elements. These
dependencies were not totally deterministic, however. Mach emphasizes in the
Conservation ofEnergy, by way of a novel argument, that the mere fact that nature is
capable of change or variation at all indicates that the dependence of elements on one
another cannot be totally deterministic or overdetermined:

The law of causality is identical with the supposition that between the natural phenomena
a,p,y,li, ... ,co certain equations subsist. The law of causality says nothing about the number or
form of these equations; it is the problem of positive natural investigation to determine this; but
it is clear that if the number of equations were greater than or equal to the number of the
a,p,y,o, ...,co all the a,p,y,o, ...,co would be thereby overdetermined or at least completely
determined. The fact of the varying of nature therefore proves that the number of equatioos is
less than the number of the a,p,y,o, ... ,co.'0

Mach seems to have two sorts of cases of indeterminism in mind. First, he admits
the possibility of physically closed systems of variations in which the changes are
internal to this group and shut out from the rest of the variation of nature around

Groups of appearances can be given, which are engaged in continual variation without being
sources of continual variation of other groups of phenomena. These are groups shut up in
themselves. How such groups can be divided, that is to say, which phenomena depend on one
another and in what manner, and which do not, can be taught only by experience, and the law
of causality says nothing about it. 51

The other case concerns variations of the whole nexus of natural elements, which
Mach considered a mistaken way of speaking (even if he himself wrote out such a
universal set of elements and set them equal to zero). For Mach, variations could only
be applied to events within the whole, not to the whole itself He attacked Lord
Kelvin's and Rudolf Clausius's extension of the second law of thermodynamics to
describe a perpetual increase in the entropy of the universe, the "heat death, "52 and
the attribution of a time to the evolution of the cosmos. 53 "For the universe," Mach
countered, "there is no time:" 54

If we say of a thing in the universe that, after the lapse of a certain time, it undergoes the
variation A, we posit it as dependent on another part of the universe, which we consider as a
clock. But if we assert such a theorem for the universe itself, we have deceived ourselves in that
we have nothing over to which we could refer the universe as to a clock ... Scientific statements
like the one mentioned seem to me worse than the worst philosophical ones. 55

But what factors do determine the functional variations of the elements? Mach
tells us relatively little in the Conservation of Energy about that. But he does point
at his own, and others', cultivation of analogies between the different forms of
energy known to him at the time (heat, mechanical work, electricity, and

S. Carnot found that whenever heat performs work, a certain quantity of heat goes from a
higher temperature level to a lower one. He supposed in this that the quantity of heat
remains constant. A simple analogy is this: if water (say by means of a water mill) is to
perform work, a certain quantity of it must flow from a higher to a lower level; the quantity
of water remains constant during the process ... Electricity can perform work when it flows
from a body of higher potential to one of lower potential; the quantity of electricity remains
constant. A body in motion can perform work if it transfers some of its vis viva to a body
moving more slowly. Vis viva can perform work by passing from a higher velocity-level to
a lower one; the vis viva then decreases. 56

This will be discussed in more detail later, but suffice it here to point out that
Mach had published a brief journal note in 1871 describing the analogy between
the behavior of heat and other energies. Just as heat energy was divisible into a
capacity or quantity of heat at a given absolute temperature (Q/T), so Mach had
already extended this analogy to several other cases: electricity, the impact of
masses, and the attraction of masses. Somewhat earlier, Kelvin and Maxwell had

also extended mathematical analogies between heat, streamlines and continuity in

fluid mechanics and electric and magnetic lines of force and potential; these
analogies seem to have played a major role in Maxwell's own framing of his theory
of electromagnetism. 57

Mach emphasized in these analogies that it was not the constancy or substance
of energy (as a stuff) that mattered but the fact that all energies could be
represented as a certain combination between a natural potential to cause change
and the intensity at which that potential is capable of acting. Mach called the
intensity level (the temperature in the case of heat) the "potential function" because
the potential to act depends directly on the intensity level: a body can only
exchange heat spontaneously with another of lower temperature, a body can only
attain a velocity by passing from a greater to a lesser height 'above the earth, and so

In his Prague years, beginning in 1867, Mach began to work through these
analogies much more thoroughly with his students. He would also have to confront
various imperfections and disanalogies, especially those between heat and other
energies. Mach was joined in this task by the so-called energeticists, Helm and
Ostwald, even though there was no agreement among these three men about how
the analogies should be construed. But in this earlier period, Mach believed he had
more or less hit on a "law of happenings" in nature in which all sources of natural
change that involved exchanges of energy could be represented as drops of
potential from a higher to a lower potential-level, or intensity level. Natural
happenings, at least those he could account for, could be represented as
equalizations of level differences, similar to fluids seeking their lowest point:

Equalization of differences.
Only where there are differences, is a process possible, since here determination is not lacking.
With the equalization of differences, necessarily equilibrium
Increase of differences. An increase determinative of a process? Change depends always upon a
clearing away of the cause ofchange.
A state will not pass over into another in which greater causes of change are included.
What does time have to do with the latter point.
Diminution of a difference can draw an increase in another after it. But all in all a diminution
must occur...58

In his account of the natural processes, Mach considered only equalizations of

level, not their emergence. When processes create a difference it is always on the way
to making another smaller. In the continuation of the same passage, Mach says that an
actual increase in the differences could have no determinate rule: "Equalities cannot
carry in themselves any ground for inequalities. That would indicate no rule
according to which it goes on. Inequalities can only become smaller. "59

Mach returned to the problem of the generation of differences of level in a late

edition of the Analysis of Sensations after he had been accused of being an

indeterminist. Here he admitted plainly that differences must enter into our
environment through "certain independent variables which are inaccessible to us": 60

The only assumption compatible with a general representation of our own limited environment
is that of a tendency, on the whole, to a diminution of differences. But if circumstances that set
up differences did not make themselves felt by forcing their way into our environment a time
wonld soon come when nothing more would happen at all.61

He also wrote in a letter to Josef Popper:

I agree that potential differences not only disappear but must also come into being. The latter
point is uncommonly hard to understand. Do all potential-differences come from the outside,
e.g. from the sun to the earth? Or do they arise from other quantities already on earth? The
strangest case for me has always been the cooling of a liquid beneath the environing
temperature through evaporation.62

Thus, for all the claims of functional dependence of all on all, there were
aspects of natural variability Mach couldn't explain and he was very up front about
it. Certain empirical laws not derivable from the principle of energy or the "law of
differences" had to be found by special investigation. Mach did not attribute the
existence of differences either to a brute asymmetry in initial conditions of energy
changes or to the local neighborhood of the earth or the solar system, but to
"certain independent variables."


Philosophically speaking, Mach's function concept was a crucial feature separating

his empiricism from that of Hume. For Mach, every element occurred in some
complex as some function of others, even if we cannot specify these others or the
form of the function. Mach certainly did not consider functional relations as
necessary connections that could be known a priori; they were all empirically
determined. But the overall fact of some, yet to be specified, dependency was not
to be empirically contested. Mach was concerned to rule out the extreme empiricist
conception of independent existences. For example, the empiricist argues that if
elements are thinkably independent in every case, because we do not know which
functions they obey except by experience, then for all we know they might be
physically independent from one another, too. To take an extreme case, there
might be individual elements segmented or isolated from the rest of nature, for
example an isolated variation from which excess energy might be periodically
siphoned off. This was obviously a case Mach wanted to exclude because it
suggested the Ding an sich, that is, a thing with not even a conceivable connection
to the rest of natural phenomena (perhaps Ding bei sich might be a better
In his Knowledge and Error Mach emphasized that his elements did not go
roaming by themselves in the world but always belonged to functionally

determined complexes: "Even a body, a lump of lead, the crudest item known to us
always belongs to a complex and so to the world; nothing exists in isolation. "63

In the Mechanics he pointed this out sharply:

Nature does not begin with elements as we are obliged to begin with them. It is certainly
fortunate for us, that we can, from time to time, tum aside our eyes from the overpowering
unity of the All, and allow them to rest on individual details. But we should not omit,
ultimately to complete and correct our views by a thorough consideration of the things
which for the time being we left out of account. 64

Thus, although Mach did agree in some ways that nature was a heraclitean
flux of individual elements, he also held the monist intuition that the individual
processes of nature always occur as part of a whole which is also given as part of
every experience. Hence what really appears as the given is a fusion of elements-
in-functions, never a handful of segmented existences. By contrast, Hume is often
read as calling for a doctrine of logically separable sensations, or separable
existences, in which the existence of each sensation and the intensity and quality it
has becomes independent of the existence of any other. Hence, not only is it
possible that any sensation might vary independent of any other, it is likewise
possible that any sensation might occur alone in the universe without the existence
of anything else. To my knowledge, Hume himself never went this far. His
metaphysical views are extremely obscure, compared to his straightforward
epistemic views, but he appeared to believe in a flux of elements always associated
at least in general space and time orderings. Even failing some obvious causal
connection, or any knowledge of it, Hume says, "Were ideas entirely loose and
unconnected, chance alone wou'djoin them." 65

Mach went further than Hume in allowing that relations between elements
might be given to the observer just as much as the elements themselves. At the
time when he was composing the Mechanics, Mach wrote several passages in his
notebook indicating just that idea:

Connection also given66

The elements and the connection. The connection is also a fact. What sort of fact? An
inner? 67

In the body of the Mechanics, Mach stated his belief in "big facts" 68 of nature:
general empirical principles of connection, such as the excluded perpetual motion
principle, which could be assumed to be true and then used to derive other facts by
contradiction. Mach showed in the Conservation of Energy that Stevinus derived
principles of mechanics and hydrostatics from the excluded perpetuum mobile and
Lagrange used it to derive the principle of virtual velocities. For Mach the
principle was thoroughly empirical, although it did seem to derive from a general
"relatedness of nature" as a closed system of elements. We just never know the
particular relations or functions they will follow.

By being assumed at the outset of an investigation, as Stevinus and Lagrange did,

the excluded perpetual motion principle may take on a postulational status relative
to other facts, but it is not thereby turned into an a priori truth; it only plays that
role with regard to the specific investigation undertaken. Mach thought that skilled
natural investigators, such as Newton, Fourier and J.R. Mayer, could actually "see"
general formal relations in the phenomena:

The considerations here developed will convince us that we can dispose by the Newtonian
principles of every phenomenon of a mechanical kind which may arise, provided we only
take pains to enter far enough into details. We literally see through the cases of equilibrium
and motion which here occur and behold the masses actually impressed with the
accelerations they determine in one another. It is the same big fact, which we recognize in
the most various phenomena, or at least can recognize if we make a point of so doing.69

Mach thus stood closer to radical empiricists, such as William James, who
have gone on to espouse the reality of relations alongside the elements of the
world. 70 James puts the point rather well when he says:

The parts of experience hold together from next to next be relations that are themselves
parts of experience. The directly apprehended universe needs, in short, no extraneous trans-
empirical connective support, but possesses in its own right a concatenated or continuous
structure. 71

Near contemporaries Hans Henning and Kurt Gerhards remarked on the fact
that Mach's concept of function not only separated him from Hume, as James
insisted for himself, but also constituted a new kind of entity, a fusion of a quality
element and its functional role:

Even if differences exist between Mach and Hume, the point of departure is still the same.
Yet Mach is not a critic of knowledge in the sense of Hume because he introduces the
function concept into the theory of knowledge. This is something fundamentally new,
which burns all bridges to earlier authors. 72

An element or sensation is not a sensation an sich, like the sensation of blue or warmth, but
rather a function ... One seeks behind the element nothing other than a function, behind the
complex only a Riemannean manifold. 73

Gerhards discussed the idea that what is given in every experience is more
than just the elements, but also the connectedness of the environment:

When I succeed in completely diverting my attention from the table, the window, the light
streaming in through the window and reflecting off the table, the inkwell... and finally from
myself, there still remain colors, brightnesses, darknesses and all their shapes .. .left over, and
I can take note that all of these stand in a certain connection with one another. This
connectedness, just as much as its constituents, is simply there. I have not added it in
thought, but merely noticed it. 74

It is likely that Mach had absorbed such holist views from Herbart, who had
also insisted that the forms of experience were "given" and that only groups of

sensations, and not individualized atoms, could be said to occur in experience: "In
that only groups of sensations are given, and series of ideas that arise from them,
there remains no psychological possibility to individualize sensations; rather it is
only a scientific abstraction when we consider them as standing individually."75

I think perhaps that Gerhards might have been trying to forge a link between
Mach and Machism and phenomenology, which was rapidly gaining power in
German universities. It was one of the primary breakthroughs of phenomenology to
pay as much attention to properties of the environment as to sensation qualities.
One not only sees individualized experiences but various invisibly structured
worlds around objects. Hence, there may be similarities among Mach, James, and
the phenomenologists, which some industrious researcher might want to explore,
but it must be recalled that Mach would have rejected a priorism in the form it took
for phenomenology. Forms of experience were empirical facts for him, no more,
no less. One could make them a priori for the purposes of an investigation by
assuming them as given at the outset, but they could always be overthrown by


I will consider for the sake of completeness, the only real candidate for an a
priori posit in Mach's functional variations and forms of experience, which is that
of a "whole of nature," or the idea that nature must form a related, interconnected
nexus-both the laws of the separate sciences and the system of existing things.
This seems to be Mach's only real article of faith, i.e, that there is "at most one
world," as exemplified by his youthful experience of the sununer's day in the field,
when he experienced the merging of the universe with his own ego, or when he
said cryptically that "nature has but an individual existence." Nor is this
particularly unusual, as scientists often express the idea rhetorically that nature is
an organically connected whole in which one part harmonizes with another.

Mach's "whole of nature" principle, for example, is extremely indeterminate in

the sense that it predicts nothing but the mere general fact of relatedness among the
things of nature. The principle has some content in ruling out elements isolated
from the rest of the universe, and the fact of relatedness in general, one would
hope, suggests the existence of some particular relations among the elements; but
it certainly doesn't say what they are, or what elements to expect bound up with
others as a rule. Indeed, the relations among the elements might be extremely
chaotic and unpredictable, with changing elements and changing functions, but still
consistent with the idea of a serene nexus overarching the whole.

The status of the "whole of nature" principle for philosophers is likewise a bit
of a mystery. Kant, in the Kritik der reinen Vernunft, allowed it only as a
regulative precept of reason and as the empirical unity of an intuited manifold of
space and time. Kant said there had to be a unity of intuition for the sake of
possible experience, but not as a determinate ground or discursive concept, from
which real facts could be deduced top-down. Kant had also maintained in pre-
critical works, against the Leibnizians, that the principle of monism was not an a
priori intuition and could well be contradicted by a plurality of worlds.

Kant's position in the Kritik der Urteilskraft was that a teleological principle
of organization for the whole of nature could only be gathered inductively or
judged to exist reflectively from the mass of empirical data, but never assumed as a
fully developed concept, as one could assume space for doing geometry. Kant
feared that if such a determinate ground of being were to be given, beyond the pure
intuitions of one whole of appearances in space and time, then super-scientific laws
of nature or organization from whole to part could be deduced from it, as Spinoza
had suggested his one substance could include all other causal relations
immanently within it.

In fact, there is some evidence that Mach found Spinoza's one substance
congenial to his own monistic views, 76 and this would not be a surprise at all.
Despite F.H. Jacobi's famous assertion that Spinozoism barred free will and led
ultimately to atheism and materialism, Spinoza was enormously influential in
ninteenth century Germany.

But Mach's nexus is unlike Spinoza's in that knowing the whole, or that there
is a whole, would not entail any knowledge of the arrangement of the parts. Nor do
the parts give knowledge of the whole. There is in fact a gap for Mach between the
properties defined within a causal nexus of the world and the properties said to
hold of the nexus itself. Mach says, for example, that we cannot attribute a time, an
entropy or other properties such as mass to the universe, for these terms apply only
to comparisons of processes within. So as a general relatedness, the whole of
nature was indeed something of a Kantian ideal of reason for Mach without
playing a role in any actual investigations. Probably this ideal was methodological
in nature and connected with Mach's notion that all of the sciences ultimately have
one and the same world for their subject matter, and that the sciences themselves
"must all form one whole," a feature of his monism that backed the idea of a
unifying science of elements.


Mach's Elementenlehre as a theory of physical elements found much of its inspiration in

the burgeoning field of German psychology, in which he was intimately involved as a
researcher. In the works of Herbart, Lotze, Wundt, and others, Mach was exposed to a
very sophisticated philosophico-psychological tradition devoted to a problem that also
faced his theory of elements. Given that the world consisted for him of extensionless
elements and spaceless-timeless functional relationships, how could one explain the fact
that the world appears to be spread out in space and time before us, both in physics
(where brute extension is assumed as given) and in psychology (where an explanation
of the extended sensory manifolds is sought)? Could one actually somehow construct
extension logically from simpler notions of intensity or quality similar to color or

The empiricist school of German psychology offered several ingenious attempts to

do just this. And although Mach soon broke with this point of view in psychology and
declared spatial representation to be an evolutionary adaptation innate at birth, the
empiricist ideas of these seminal psychologists eventually found concrete application in
his conceptions of physics and geometry.

Wolfram Swoboda relates that Herbart's ideas probably reached Mach through
Franz Lott, a family friend and professor of philosophy in Vienna from 1849 to 1872.
Lott had studied with Herbart at Gottingen, and Mach in turn studied psychology with
Lott in the summer of 1858. 1 Mach said himself that he was busy with matters of
Herbartian psychology and metaphysics in 1862, and he and Lott continued to
correspond on matters ofHerbartian psychology at least untill864. Letters from Gustav
Fechner of that year confirm that, in Fechner's eyes, Mach remained a convinced

Johann Friedrich Herbart (1776-1841) succeeded to Kant's chair at Konigsberg in

1809. In 1824-25 he published the two volumes of his Psychologie als Wissenschaft and
in 1828-29, his Allgemeine Metaphysik. He is known today as a philosopher of
education and as one of the founders of psychology; but in the nineteenth century,
Herbart's metaphysics and psychology had an extraordinary influence on science and
mathematics. Bernhard Riemann, in his celebrated Probevorlesung "Uber die
Hypothese welche der Geometrie zurn Grunde liegen," explicitly credits "some
philosophical investigations of Herbart" 3 and Hermann Grassmann's Ausdehnungslehre
was also influenced by this thinker. 4 Later, the Herbartian psychology was a significant
influence even beyond Germany and was widely discussed in Britain and America. 5

Herbart considered himself a scientific philosopher and thought metaphysics was

merely the continuance of science by other means, using the same methods but at a
higher level of generality. Yet he denied that such great generality gave metaphysics
more certainty:

It is not required that metaphysics be more certain and penetrate deeper than it can in following
experience. It rests on this, as on its own proper [eigentUmlichen] hypotheses. Should human
experience be found too limited, too incomplete, not enough in accord with hopes and wishes on
some points to ground a completely satisfactory conviction, the fault cannot be unjustly foisted on
metaphysics, which does not light with its own light but merely renders up what it has received. 6

Herbart also firmly separated psychological spatial representation from the

properties of physical space. He proposed a Kantian theory of spatial representation by
means of memory and imagination in the Psychologie als Wissenschafi, and a
Leibnizian construction of intelligible space in the Allgemeine Metaphysik. Herbart's
distinction became crucial for scientific progress later in the century, especially after
Gauss and Riemann, when it could not be maintained that the human psychological
capacity to represent space as Euclidean or even three dimensional indicated synthetic a
priori knowledge about space's physical properties.


Herbart's metaphysics, of which his Psychologie was a chapter, began as an

Elementenlehre of matter and form. 7 The matter of the world was made up of sensations
with their qualities and magnitudes considered as primitive properties. The forms of
space, time, and extended objects indicated abstract orderings of those elements
together. Herbart was at least partly a realist about forms of combination; his elements
always occurred in a causal nexus of some sort, even if it was impossible to know which
elements in particular combined with which:

Are the forms of experience given? Yes indeed they are given, although only as detenninations of
the manner in which sensations are bound up together. Were they not given, we could not only
sunder them from sensation in such a way that the sensed could occur completely isolated, without
any connection; rather we could also, at pleasure, see different shapes, hear other time intervals;
similarly we could put things together arbitrarily out of properties and change them. 8

The problem, for Herbart, was that the forms of experience contained
"contradictions" (for example, elements that could be either discrete or continuous,
containing gaps between them or no gaps) to be resolved by penetrating behind the
appearances and discovering the true relations between unextended monads and
qualities that were for him the things in themselves. Herbart did not abide by Kant's
stricture to stay within the bounds of experience and its preconditions, but believed
instead that analyzing the structure of human experience was simply a matter for a
scientific psychology, whose job it was to strip away confused and contradictory forms
of apprehension until what remained was the true underlying form and material of
things in themselves, i.e., the point like Wesen and their changing states:
Complexes and fusions, shaded off in inexhaustible variety, interwoven and roused to activity, give
our representations in part invented, in part empirical fonns. The Mechanics of mind, which not
only gets into the representations but also the states of the representing process, displays the
possible forms and modes of influence [Wirkungsarten] of complexes and fusions. It thus teaches
the conditions under which spatial shapes, time intervals, and series of changes are represented. The
fulfillment of this condition is nature's affair; we thus possess a knowledge of nature, subject indeed
to error and improvement, but of which we cannot be robbed, and which emerges victorious from
all difficulties. For in the connections of our ideas, in so far as they are formed by experience
anyway, are reflected the connections of things to one another and with us; and this connection
between that which is in us and that which is outside of us, becomes clear through psychology in
such a way that redounds not insignificantly to the confirmation of the true realistic metaphysics.9

Of particular importance for understanding the nature of Mach's elements is

Herbart's doctrine of qualities, originally presented in his Psychologie als Wissenscha.ft,
which considered sensations and ideas as analogous to Newtonian forces. The quality of
an element was like the direction of the force. Where two elements of opposing quality
met, they pressed against each other by a kind of intensive action and reaction. If the
two qualities were independent, they passed through each other harmlessly with no
effect, like forces situated at right angles. When two qualities opposed each other, the
degree or the power with which the stronger overcame the weaker gave their respective
magnitudes. 10 The reciprocal action of qualities for Herbart was what grounded an
association between them, and what made it possible to handle them mathematically.
According to David Leary, in his overview of the antecedents to Herbartian psychology:

Mental phenomena constitute a mechanical system in which any increase in the intensity of one
presentation is reciprocally related to a proportionate decrease in the intensity of another
presentation. With this premise the foundation of Herbart' s mathematical psychology is complete,
for if increases and decreases of intensity are exactly proportionate, as Herbart maintained, then
they can be mathematically represented. 11

Leary also makes the case that both the doctrine of intensive magnitudes and their
pairwise relations of action and reaction can be traced back to Leibniz.

In another development of great importance in understanding Mach's conceptions

of color space and tone space, Herbart believed that all intensities could be grouped into
qualitative continua by considering each place on the continuum as the outcome of an
opposition by the qualities (forces) at either extreme, red versus green, right versus left,
etc. He says that any object that presses another through its qualities will exhibit the
structure of a continuum12 as the force ofthe one gradually overcomes the force ofthe
other, giving the impression of a series of oppositional states that "shade off stepwise"
over time (Abstufong des Verschmelzens). As sensory manifolds similar to space,
Herbart gave examples of a tone-space, "a straight line in which intervals can be
measured with mathematical precision" and a color manifold in the shape of a triangle
with red, blue and yellow at the comers and all interior points represented as mixtures of
these extremes. 13 In a particular meeting of two opposed color ideas, if one quality is
more forceful and gradually overcomes the other, the process will appear as a colored
point "moving" through the continuum as the proportions of the stronger and weaker
component exchange in any of three directions.

It may well be that Herbartian psychology presented Mach with a fresh point of
view on his own profession of physics. Mach called Herbartian psychology "an ideal
close to that of physics" and it is easy to see why. If the qualities of mental presentations
are forces and capable of expression in those terms, then analogs of mechanical
properties can be extended to purely intensive entities, the way Mach said he wanted to
do physics with his pressure or action elements in the "Raumvorstellungen" paper.
Instead of representing energy units as the action of a force across space, one can
represent it in intensive terms as the action of one force against a resisting or
overcoming force in time, a somewhat purer level of analysis. As evident in Mach's
notebooks of the period, he understood D'Alembert's principle, and other variational
principles of mechanics, as describing the action of impressed forces in overcoming
other resisting forces and thus as expressing elements of "work and counter work." In
his "Lectures on Psychophysics," Mach also defended Herbart from the charge that all
of his magnitudes were interior and space-less and hence unable to be measured, by
arguing that even in mechanics terms were interdefined by one another and not by
comparing every isolated quantity with experience. All that was required for a scientific
treatment, Mach said, was that "inward states be of different intensities."14

When Herbart's qualities exactly canceled one another in magnitude and direction
they were in spaceless equilibrium and could group together into stable, point like
objects, which he called Wesen ("beings," "objects"). A Wesen was unextended and
simple, a node held in position by the mutually constraining forces acting there. 15 In
fact, since they were little more than functions of their qualities, Mach eventually
saw fit to do away with them and let the qualities, or elements, abut upon one
another in functional groupings. Andreas Laass comments on this change in relation
to Mach's contemporaneous views on atomism in physics:

There is a probable correlation between the atomic theory and the Herbartian, between that which is
established as real in philosophy and in psychology, but these already contain the seeds of the
destruction of [Mach's] monadology: both serve to ground empirically established facts, without
being ernpiricially established thernselves. 16

Herbart claimed that each Wesen conserved itself against the others by "pressing
back" through its qualities. In fact, he described two sorts of pressures. A Wesen could,
for example, press outward against others in its environment, in which case Herbart said
the Wesen was Together (Zusammen) with the others. Or its qualities could be pressed
inward against one another and away from the outside, in which case Herbart said the
Wesen was Not-Together (Nicht-Zusammen) with the others. Despite the fact that both
inward pressures and outward pressures existed, alternating over times, all the qualities
in the community pressed one another in general causal relations of action and reaction.
Indeed, were the forces in the community not generated in those mutually opposing
pairs, the Wesen, too, would not exist:

We could say through a sensory comparison what they [Wesens] do. They press one another. For in
the world of the senses we find resistance in pressure, where nothing gives way although each is
supposed to move. Pressure is rest, through reciprocal endurance against another. But sensory
comparison is dangerous here. We are not talking about spatial relations ...Here we are only talking

about a change in quality that each should suffer from the other, however against which it conserves
itself as that which it is. Disturbance should result; self-conservation cancels the disturbance in such
a way that it does not occur at all. 17

None of Herbart's metaphysical ideas was unprecedented, of course. Leibniz, and

the pre-critical Kant, had adumbrated such a community of unextended substances,
changing their states qualitatively, either by pre-established harmony emanating from
their insides (Leibniz) or by external forces like gravitation acting at a distance (Kant). 18
(Herbart seems to have combined these mechanisms.) Even in the first Critique, Kant
was still convinced that a community of substances undergoing dynamical, reciprocal
changes among themselves would be sufficient for a construction of space within, if
only "pure understanding could penetrate to things in themselves, and if they admitted
of space and time determinations." 19

Herbart's instantaneous, spaceless sensations were extrapolated from another of

Kant's positions: his doctrine of intensity. Possibly drawing on an "intensive"
conception of the infmitesimal in the analytic mechanics of his time, Kant had
emphasized that sensations, and even perhaps instantaneous moments of gravity and
other physical forces, 20 exemplified intensive magnitudes in nature:

Since sensation an sich is no objective representation, and in it neither the intuition of space, nor of
time, is met with, no extended magnitude is attributed to it, but to be sure a magnitude (and one in
which empirical consciousness grows from nothing = 0 to a given measure in apprehending it in a
certain time), that is an intensive magnitude. Correspondingly intensive magnitude must be
attributed to all objects of perception, in so far as they contain sensation, i.e., a degree of influence
on the senses. 21

Kant did not hold that the pure flux of sensations could be consciously experienced,
for, as he stated, an empirical consciousness had to apprehend a sensation rising or
falling in time, and his axioms of intuition demanded that experience always be of
extended magnitudes. That rule applied to every part of experience as well; space
consisted of ever smaller spaces, time of ever smaller times. What was non spatial, for
Kant, were sensations an sich when regarded as mere qualities with intensity, abstracted
from their attachment to an object, which Kant also called Eindriicke, or impressions.
He nearly always qualified such hints by adding that a sensation without a form, not
belonging to a spatia temporal object, was not to be exhibited in intuition and belonged
to a preconscious state of sensing before awareness.Z 2 The aspatiality or atemporality of
raw sensations or impressions is a point that has been bandied about by Kant scholars.
Norman Kemp Smith classes sensation with unextended magnitudes, 23 while others
hold that sensations were subject to the receptive conditions of the transcendental
aesthetic, and thus already spatial and temporal. One cannot thus claim consensus on the
aspatiality thesis, but many of Kant's immediate successors held it, including Fichte and
Solomon Maimon.

As part of his freedom to speculate on the basic metaphysical conceptions

underlying science, Herbart permitted himself to consider the pure underlying flux of

sensations an sich, before considering their confmernent in spatial and temporal objects.
Under the heading of sensations, Herbart also seems to have included the physical
forces that made up matter at a fundamental level.

Given his own pre-Critical Leibnizian allegiances, it is probable that Kant at least,
entertained the idea that experiences in space and time were constructed by the mind
out of a primordial array of spaceless, timeless sensations.Z4 For example, in the A
edition of the Transcendental Deduction, Kant ventured the hypothesis conditionally
that if an observer had impressions raw, they would appear in disjoint, strobe-like bursts
of intensity. Each state would vanish when the next occurred. If that were so, we could
not draw a line or collect up a series of points, because the points just drawn would
constantly fall away.

It is apparent that when I draw a line in thought, or think the time from one noon to another, or
merely imagine a certain number, I must necessarily first conceive these manifold representations
one after another in thought. Should I leave out and not reproduce the preceding (first parts of the
line, the preceding parts of time or the imagined units after one another) in passing to what follows,
then never would a whole representation, or the aforementioned thoughts, or even the purest and
first fundamental representations of space and time be able to arise?5

In the A deduction, at least, Kant seemed to have reasoned that the very possibility
of experience in space and time depended on the ability to reproduce past impressions in
memory and imagination that were no longer present. In experience, as opposed to
passive sensing, the productive Einbildungskraft produced and adjoined imagined past
and future impressions to present impressions and created an imaginary, abiding
framework of past, present, and future objects out of raw sensations that simply
appeared and vanished.


Herbart turned some of these hints into a theory of psychological space in his
Psychologie als Wissenschaft. His theory of Raumvorstellung was called the
"reproduction series," a term that mirrored Kant's "synthesis of reproduction." Like most
post-Kantian thinkers, Herbart ventured beyond the limits of experience to consider the
transcendental processes underneath.

In Herbartian psychology, the mind, or "apperceptive mass," was an agglomeration

of mutually resisting ideas all fighting to maintain their strength and clearness.
According to Herbart, one never has a single thought but always many competing ones,
some above the limin of consciousness, others pushed below it. In the preconscious
flux, every sensation vanished immediately after it occurred, but all sensations left
behind memory traces or Vorstellungen. Vorstellung has the connotation in German of a
product of the imagination, that is something that one sets before one's mind rather than
something perceived or experienced. These Vorstellungen inhibited one other until the
stronger ideas had pressed the weaker beneath the limin of consciousness. Mach showed

himself in the Lectures on Psychophysics to be very willing to accept such Herbartian

assumptions even without direct experience of them:

The cases that Herbart treats are simpler than ever occur in reality; yet this is necessary for the
beginning of an investigation. Mechanics proceeeds from similar simple cases. The mathematical
consequences of Herbart's assumptions however ... agree so often with experience that it is
impossible to believe the fundamental assumptions should be far from the truth. For a good
practitioner of introspection it is more than a mere image that the ideas reciprocally press one

From the assumptions of reciprocal inhibition, fusion, and complication, and a few
quantitative principles besides, Herbart derived a Mechanik des Geistes, the
undergirding rules of transcendental apperception. One of the primitive mechanisms of
this Mechanik was to reproduce past impressions in memory in the order of their
occurrence. For this to happen, the after-images of impressions must endure after the
disappearance of the sensations that caused them. Herbart assumed that sensations
vanished in a time "smaller than any assignable time" after they occurred, so that the
duration of sensational data would not be extended at all.27 However, the sensation's
after image does not vanish immediately but continues to persist and oppose other ideas
until it, too, is inhibited below the limin of consciousness.

Secondly, those enduring residues of the sensational data that had passed away had
to be affixed to present sensations and to their residues. 28 For example, if there are three
serially occurring qualities A, B, C, once A occurs and vanishes, it leaves behind
memory images, gradually sinking in strength, which Herbart denoted with primed
letters, A', A", A"', and so forth; B leaves behind its images B', B", B"', and so on. These
images are in turn fused with newly appearing impressions. So as B appears to take A's
place, A's after-images become fused with B. As a newly appearing C takes the place of
B, A's and B's residues will both become fused with C, and so on through the course of
experience. In this mechanism, the dying residues of a previous impression became
attached to every impression that succeeds it: 29

A' B'

But suppose, Herbart asked, that an impression similar enough to A should recur at
a later point in time and be recognized as a repetition of the same, by comparing it with
the memory trace left behind by the original A. At this point, he said, the new sensation
would enliven the memory images of the original A and actually put energy into it
leading it to rise again into consciousness along with all of its annexed residues. A's
strongest lingering memory trace (A') was first inhibited by B, so as A' pushed its way
back up above the limin of consciousness, B's image was pushed up ahead of it. A's
second strongest image was inhibited by C, so C was the next image reproduced, and so
on. Hence, with the recurrence of A, the images of B and C are also reproduced in the
exact serial order with which they once appeared after A. The "reproduction" series was

so called because it exactly reproduced the original order of inhibition, and, with it, a
facsimile of the original series A, B, C. Thus when we see A again, we immediately
anticipate B and C as well.

Because vanished impressions were still imagined as present after their

disappearance, the overall psychological effect of the reproductive imagination was
that a sequence of disjoint impressions A,B,C, appeared to be part of an extended
stretch from A to C, in which A shaded off continuously as B and C arose. If the
series could be run in reverse order as well, from C to B to A, the impression
produced by the two reproduction series would be of a moving point progressing in
two directions along a seemingly fixed continuum of qualities.

What should be clear is that since the qualities themselves are instantaneous, it
is the endurance of the memory images that provides the temporal extension of the
series. A good comparison might be the exploration of an Egyptian tomb with a
matchbook. Each time a match is lighted, a particular section of the tomb appears for
a very short time and can be represented in memory. As the next section is
illuminated, the memory images of the first are still present and can be related
serially to the next section. As the explorer moves in the tomb, he may find that
certain impressions tend to recur, although in reverse order, and he concludes that he
is walking up and down the same series of points. Nothing in the impressions
themselves, or the memories, would make this certain however. With Herbart, we
have, in essence, the process of serial or pure temporal representation pared down to
its most elementary parts: sensations, memory and an ability to recognized the
recurrence of the same, by comparing a present sensation to a memory trace taken
when it last appeared. This ability carries no certainty, either. It is possible to
imagine the explorer being tricked into recognizing new and different sections of the
tomb as the recurrence of older ones merely because they appear very similar.
Herbart's probable inspiration for interpreting extension as the idealized three-step
process of sensation, memory, and recognition, was with very little doubt Kant's
explanation of the drawing of a line.

In his Psychologie als Wissenschaft, Part 2, 1.3., Herbart applied his

reproductive series to the psychological problems of spatial and temporal
representation in all sensory manifolds, such as the hearing of a melody or the
generating of a monocular field of vision from the retinal mosaic. When a melody
was played, only one note at a time was heard. Therefore, the manifestation of the
whole was an object of the imagination and memory, except for the parts that
happened to be present from moment to moment. Similarly, in interpreting a mosaic
of disjoint light sensations on the retina as a two-dimensional extended manifold,
Herbart believed that each retinal point had to possess its own separate reproduction
series. The aggregate of these series, extending radially from the center, made up a
spatial field, reproduced constantly in time as the eye moved about refreshing the
retinal points. As Mach described in his Lectures on Psychophysics:

If the eye has run through the series a,b,c several times and in the opposite direction, the stimulation
of one member of the series will have the consequence of stimulating the running through of the
series in both directions from this member. From this member on, the members to each side of it
will appear simultaneously, with descending clarity. If we consider that from an a, the eye can run
through not just one but infinitely many series in different directions, and, if we apply the law of
reproduction in series, we attain something at least very similar to spatial perception. 30

As Mach also remarked, in any reproduction series where spatial extension was to
be intuited, Herbart assumed multiple series running at once along independent
directions. One series, for example that of a melody, at most ordered the impressions
temporally but not spatially. The extension was not directly intuited but simply
imagined as duration. For Herbart, a truly intuitable linear extension required two
overlapping and independent reproduction series, durchgekreuzt or literally "crossed
over" one another. 31 A thin wire seen head-on, for example, looks like a point. Only
when it is turned through an independent, dissociating, direction does it actually appear
to be a line. Thus, by interposing a linear series independent of the first, to be run
through as the original series is run through, the original series will be seen to trace out
an extended line.

This is a fairly subtle point, to which Herbart was quite sensitive. A better metaphor
for a simple linear reproduction series like A,B,C above would be a point that changes
qualitatively in hue, or in pitch. With the help of memory we can tell which qualities
have come up in what order and experience a duration through the persistence of after-
images. Once we get used to the pattern, we may even come to recognize the point as
traveling up and down in a fixed range of qualities.

The problem is accounting for the spaces in between the qualities. If Herbart was
serious about taking qualities as vanishing short in duration, like infmitesimals, then he
either needed an infmite number of them in a continuous series to make up any fmite
duration, or to explain how a series of vanishingly small items can come to take up time
when none of its elements do. Otherwise, the transitions between qualities are as
extensionless as the qualities themselves without spaces in between, or as Herbart calls
them Zwischenriiume. I suppose one explanation could have been found in the idea of
sensory qualities' resemblance to forces; the free and uncompensated communication of
force may be imagined to take no time. But if there is resistance the communication will
be slowed and cannot occur all at once. Something like this is probably true ofHerbart's
metaphysics and may have been intended to carry over into his theory of sensation.

Herbart, however, solves the problem of creating spaces between the members of a
series by interposing a second series at an independent direction to the first. By
superimposing on the first series an independent set of qualities, another linear series as
it were sitting perpendicular to the frrst, the original point's progress may be seen as
tracing out an extended line. This use of a double row to mark extension is an absolutely
essential feature of Herbart's theory of space because it establishes the necessary
dissociation of elements that is found in all extended magnitudes. Thus, for the
construction to work we must imagine that every quality covered by the point in the
original series is also a member of another series, intersecting it at that point (and
associated with it there) but also running perpendicular in a direction that is not

dependent on the progress of A,B,C. For simplicity, and to stick with Herbart's idea of a
spatial visual field, imagine four reproduction series, each point changing in two
directions through its own range of qualities. Series (l) will run straight up and down
the page, while (2), (3) and (4) run perpendicular to it left and right. 32

(2) A B' C' ...
(3) A" B C" ...
(4) A"' B"' c ...
The first series (1) A,B,C includes the hues of our original qualitatively changing
extensionless point. The hues of the second series A, B', C' are associated with the first
series only at A, and are otherwise independent in their changes. The third series meets
the first only at B and is otherwise independent, and the same for the fourth at C.

Now for the individual associations. A is associated with B (and through B

mediately with C), but A also calls up images of B' and C' and it is to this extent
causally dissociated from B and C. Likewise, B is associated with A and C, but
dissociated from them by its connections to A" and B". Finally, C is associated with B
and mediately with A, but dissociated from them by its associations with A"' and B"'.

If we set the machine working and think of all these series running at once, then in
passing from hues A to B to C, A is both associated with B and dissociated from it,
while B is associated with C and dissociated from it. We may think of these
independent dissociations as "gaps" of extension between the members setting A apart
from B apart from C. If we take any one of the series, in fact, its members will be set
apart and extended against the background of laterally intervening series. Herbart's
application of this model to the eye has obviously become somewhat fanciful. We
would have to imagine that the eye had a practically infinite number of elements
arranged in rows and an infinite number of reproduction series running in independent
directions through each element at all times.

Now the effectiveness of this construction was by no means universally granted,

even in Herbart's time. 33 "Heaven knows how!" exclaims William James in his
Principles of Psychology, p. 631. One recent author reviewing the reproduction series,
Gary Hatfield, objects that even if the various mechanisms are granted, even if it is
granted that Herbart's construction reproduces the relevant information conveyed by
spatial representation, the result still does not look like space:

Why should reversible chains in themselves be sufficient to produce a representation with the
phenomenal character of space? Such chains could easily represent spatial relations, just as a series
of tones--or any series whose members can be brought into one-to-one correspondence with a set
of spatial locations-might convey infonnation about spatial relations by, say, pitch and sequence.
But although the pitch and sequence of the tones would bear infonnation about spatial location, and
in that sense "represent" spatial configurations, the experience of these tones would not in itself
reproduce the experience of space. Why should the members of Herbart's reversing series be any

Why should they be perceived as not only temporally contiguous but also as being simultaneously
next to one another (as seems to be required of a properly spatial representation)?34

What often escapes notice is Herbart's prominent use of dissociation, which was
his major breakthrough in spatial representation and in the logical analysis of extended
magnitudes. Herbart acknowledges that a simple, single reversible reproduction series
can at best be a temporal representation of the permanent-although I would even doubt
the temporal extension of such a series. It is the addition of the durchgekreuzt
independent series both associated with and dissociated from the first that makes the
intuition of extended magnitudes possible. So in answer to Hatfield's question, Herbart
could have pointed out that the dissociated series need not be spatially next to, or at a
right angle to, the original reproduction series; it might also be causally independent just
so long as the dissociation between the two series is established. Again, a physical
metaphor may help. Forces that act at right angles to one another are independent and
cannot affect one another, but forces may also exhibit the same independence in a
physical way by acting as if the other is not there, gravitation and electricity for


With his theory of psychological space completed by 1825, Herbart turned to his
construction of "intelligible space" in the Allgemeine Metaphysik. Here, too, there were
several other Kantian Denkwege that Herbart followed up, especially concerning
spaceless sensations. For example, Kant had held that physical intensities were logically
prior in some sense to extension, since in one and the same instant of time or space, the
real could communicate itself to any intensive degree:
Since every reality has a degree, which can diminish to nothing (the void) through infinite
gradations without in any way altering the extensive magnitode of the appearance, there must be
infinitely many degrees in which space and time may be filled. Intensive magnitode can in different
appearances be smaller or greater, though the extensive magnitude of the intuition remain the

Conversely, if the magnitudes of intensities depended on their extension then ten

times the force would always require ten times longer, or ten times the space, to
communicate; a greater quantity of matter would fill a greater volume since each unit
quantity would have to be set aside in a separate space. 36

On the aspatiality thesis, outlined above, Kant had implied that the real was
intensively communicated in a sensation, in which space or time were not to be met
with. In a famous passage, in the schematism of the pure principles of the
understanding, Kant ventured to describe how extended temporal intervals were
obtained by schematizing instantaneous magnitudes that filled only an instant:

Reality is that in a pure concept of the understanding which corresponds to a sensation, that,
therefore, whose concept in itself points to a being (in time); negation is that whose concept
represents non-being (in time). The opposition between the two happens in the distinguishing of the

same time as a filled or empty time. Just as time is only the form of intuition, therefore of objects as
appearances, so is that which corresponds to sensation in them the transcendental material of all
objects as things in themselves (thing-hood, reality).

Now every sensation has a degree or magnitude, through which it fills the same time, in respect to
the self-same representation of an object, to a greater or lesser degree, until it ends in nothing (=0
negatio). Hence there is a relation and connection, or rather a transition [Obergang] from reality to
negation, which makes every reality representable as a quantum; and the schema of a reality as the
quantity of something in so far as it fills time is just its continuous and uniform production in time,
in that one goes, in time, from a sensation with a certain degree downwards to its disappearance, or
gradually climbs up from negation to the same magnitude.''

The passage, a justly famous one, begins with Kant's assumption that even all of
the sensations that appeared in a single instant manifested a ratio, or, as he says, an
opposition [Entgegensetzung] between their intensive reality (a maximum of intensity)
and negation (the zero). When Kant spoke of sensations "filling" time in this passage, he
seemed to be speaking of two things:

l. The filling of an instantaneous moment by means of a constant reality to negation

ratio, or unit measure such as R+N= 1.

2. The filling of an temporally extended series of moments by means of a schema, i.e.,

a stringing together of the many instantaneous moments of 1. into a time over
which an apprehended sensation could be represented as rising or falling.

The job of the schematism of the real (i.e., via the imagination) seems to have been
to pass from 1. to 2., that is, "from intensity to extension," from a timeless,
instantaneous ratio to the extended experience of sensations increasing or decreasing:

tl t2 t3 t4 t5

Kant did guarantee that every quality could be anticipated to fall somewhere
between a certain maximum, R, and a minimum of intensity, N, which not only set the
scale for comparing all intensities, but also manifested a constant quantum, R+N in
which any intensive magnitude was contained. This would happen if the reality and the
negation content always added to a fixed quantity or to zero. Kant did not say how he
could be so sure that all qualities could be compared along the same scale, unless he,
too, like Herbart, regarded sensations as qualities possessed of opposing directions.

Once the instants had been marked off as intensive quanta, the schematization of
timeless moments into stretches of time could have proceeded with the help of the
imagination. Past appearances of a quality could be imagined with their intensive ratios

in a former instant, and the imagination could then string the instants together by
picturing the quality rising or falling over the remembered or imagined stretch of time.

Within a community of qualities, such as that considered by Herbart, assuming that

the RIN ratios of all the sensations were mutually comparable, or assuming that
whatever some A gains, some B loses, the set of present and remembered reality-to-
negation ratios, (N+R), (N'+R'), (N"+R") ... ~nJ+R(nJ), provide a kind of procedure for
arranging all intensive magnitudes into the moments of a universal time order, assuming
that the opposing relations of those magnitudes also manifest such an order among

Such an interpretation, while it stretches the limits of what Kant was willing to
assume as necessary for experience, seemed clear enough to Herbart, who took it for
granted that the schematizing of rising and falling intensities provided the essentials of
time determination. His chapter of the Metaphysik "Von der Zeit" begins in fact with a
discussion of the various types of intensive magnitude and the ways in which they might
fall, rise, or remain the same. Herbart called special attention to the annihilation of past
impressions and the appearance of new ones, claiming that these positings and
cancelations were really what was numbered by time. 38

Overall, Herbart's guiding interpretation of Kant seems to have been that time
schematization completed present sensations into ein Bild, by adding to them the images
of non present past sensations and future ones, linking them together with the present
impression. The result is an ideal, permanently extended time experience, the smallest
stretch of which always contained a past, a present, and a future. If so, the real, the
sensations an sich, were rightly considered timeless, as well as spaceless, and did not yet
form objects of experience for Kant.

Herbart also took it upon himself to correct what he saw as Kant's omissions. For
example, Kant had assigned every quality its instantaneous magnitude of R+N, which
could always be assumed to be a fixed quantum. This was not justified unless the
relations of the qualities of the world to one another were always such that whatever one
quality gained another would lose, i.e., that "action and reaction" was be observed in
their changes. Again guidance may have come from Leibniz on this score. For Fichte,
attempting his own reconstruction of Kantian metaphysics at the transcendental level,
every quality was matched against an opposite, so that the reality content of one was
always the negation content of the other and vice versa. What was for Kant a ratio of
reality to nothing39 became for Fichte, and for Herbart, a ratio of reals to other reals.
That, at least, would certainly have fit with Kant's view of the empirical world (qua
realitatis phaenomena) as consisting of an equilibrium of opposing forces.

Action and reaction also provided a sufficient condition for the qualities to form
into one community, the existence of which both Fichte and Herbart deduced
immediately from the fact that opposite qualities depended upon and pressed against

one another. 40 In fact, once he had assumed such a community as fact, Herbart stated
that the opposition of qualities to one another contained the needed time determination
in the numbering of their changes by a constant quantum. It could easily be imagined
that as the qualities pressed against one another, they changed one another's magnitudes.
They filled time through their action, which could be represented instantaneously by
ratios, or over imaginary stretches of time and space. The actions were all in step with
one another, in a Wechselwirkung:

The togetherness and separateness of substances is subject to an alternation [Wechsel], which

encloses immediately a time-detennination. One easily sees that motion and space are also assumed
within;... Provisionally we call that space, which we unavoidably think in addition with the coming
and going of substance, intelligible space.41

Herbart did not give a measure of the time instant in terms of the instantaneous-
perhaps he thought Kant had already done so with his reality-to-negation quantum-but
it is not hard to see how it might have been done. If two forces, a known test force A
and a force B whose intensity is also known, are set against each another, the
instantaneous ratio (A+B) they fall into could give a measure of the intensity with which
the overcoming occurs. 42 If a force is twice as strong as a test force, we would expect
them to fall into an instantaneous ratio of two parts of the one to one part of the other.
No matter how small the instant might be taken, the same 2:1 ratio would be observed.
This ratio could then be transported from one experimental setup to another and used to
measure the ratios of other oppositions. One could then say that, all else being equal, the
overcoming of a force A by a force B occurs to twice the degree that C overcomes D,
without involving extended times.

In his construction of intelligible space, however, Herbart was less concerned with
working out the variety of relations possible at the level of qualities and more concerned
with the waves of action that undulated through the community and traced out lines,
surfaces, and solids in time. Of course, the Wechsel of qualities still drove the whole
construction since the Wesen were really no more than patterns formed by the play of
qualities beneath, as in the following figure:


Herbart applied his ideas to a construction of the natural number line, which he
imagined to be constructed by a generating process that could progress only in two
directions, left and right in discrete steps. To exhibit the structure of the natural number
line, Herbart thought that the construction of a Nacheinander of points had to create
both points and gaps, setting one point apart from another over time. It was essential for
him that the separateness of one point from another be established, as well as their

Herbart accounted for these properties of points and gaps by building them into the
generating process that traced out the line. The points were generated by the meeting of
two Wesen together (Zusammen) and the gaps by their not being together (Nicht-
Zusammen). He understood two Wesen to be together, for example, when their qualities
were in direct mutual influence, without a causal intermediary. He understood two
Wesen formerly together to separate when one suffered a change in its qualities and the
other remained in equilibrium. When this happened, an intermediary Wesen, or rather
Bild, represented by primed letters A' A" A"' took the place of the Wesen that had since
changed and represented the continuing presence of this past Wesen, as the images of
past impressions had done in Herbart's psychological reproduction series, see figure

As in the right hand side of the figure, a notation involving the symbols {A, B, +, -}
can be employed to show all of the possible directed paths Herbart could generate with
such a process. Let A and B be two Wesen. When they are together with each other the

qualities are in complete dependence, which means that when a quality of one rises in
intensity, a quality in the other falls. The ratio (R+N) is upheld as the time instant
demands. Suppose, however, that the qualities of B suffer a change and that the qualities
of A remain of the same magnitude. In that case, as Herbart represented in words, the
situation is written (A-B), writing the fixed Wesen first and the changing Wesen second.
When A changes and B remains the same, Herbart could represent this as (B-A). If, on
the other hand, A and B are two formerly independent Wesen just passing into
dependence on one another, they are represented as Together. (A+B) indicates that an
independent B has passed into dependence on a fixed A, and (B+A) indicates that an
independent A has passed into dependence on a fixed B. Taking all of the possibilities in
hand, the process of producing a set of points in either direction, is an alternation that
could be represented by "right" and "left" sequences, below:

Right: (A-B),(B+A),(A-B),(B+A)... Left: (A+B),(B-A),(A+B),(B-A)...

A +B -

11:1 ~

"Lett .,. ..,
... <
A +B
.;: 11:1

"' <


~ 11:1
m < 'Rl ght"

DA -

<[ ...
- -- BA


One of the problems with the metaphysical conception of a community of rigidly

opposed qualities is how to account for change if every action always has its equal and

opposite reaction. Kant, Fichte and Herbart were all agreed that the community of
substances, or Wesen, was subject to constant change, an Entstehen durch ein Vergehen.
But at the level of the qualities, how could changes ever occur among a set of rigidly
fixed oppositions and cancelations? When one quality acts, another must act against it.
How could there then be a rise or fall in intensity?

Herbart, once again to his credit, was the fust to realize that for one instant to give
way to the next in a rigidly connected community, the given ratio of reality and negation
between pairs of qualities had to be suspended, as for example when two Wesen were
said to dissociate. But that could only happen if the opposing moments of reality and
negation could dissociate in between instants. Herbart claimed that a time sequence,
natural number line, or series of degrees climbing and falling was not a series of points
where the qualities rigidly coincide everywhere, but a constructed set of points and
dissociating gaps. In the gaps the qualities of Wesen that begin together suddenly
became independent of one another (Nicht-Zusammen), until they were equilibrated at
the next instant, this time perhaps with a different Wesen as partner.
This process is worth reconstructing in some detail, employing the same notation, but
this time applied to the qualities R, N within the Wesen.

Let the figure below represent a rising intensity on Herbart's scheme. That means
the reality content (R) of the quality increases at the expense of the negation content
(N). At first R and N are rigidly connected to each another.




Then they are allowed to slip out of dependence as a gap occurs. In the dissociation
of the two qualities R and N, they are free to change independently of each other and
come into dependence instead on two further qualities N' and R' since every quality's
magnitude always depends on some other. Let N, in fact, depend on some R' that is
greater than R, so that the intensity of N is now suddenly lower. Meanwhile, let R drop
out of connection with N, and into dependence on anN' whose magnitude is less than
N, meaning that R's intensity is now greater. After these switches have been made, and
assuming they still add to the same fixed quantum, we can join R and N together again
at their new intensities. After joining them again, R has a higher intensity and N a lower.


Herbart claimed that his construction of Wesen and images in Figure 4 gave every
"thinkable distancing" that could occur between Wesens A and Bin both directions. 43
Every directed path was representable. For example, a process could move forward any
number of spaces and then backwards retracing these spaces, or it could move infmitely
in either direction, or infmitely back and forth over the same three points. What is
important is that Herbart thought of space as being continually drawn and redrawn by
wave-like processes of association and dissociation, not as a co-present solid or
container. To represent a space to oneself one must imagine the process of drawing it
out, just as in the generation of reproduction series.

An infmite set of possible directed paths evolving over time was not yet a natural
number line, however. As Herbart seems to have been aware, a natural number line was
the set of all possible combinations of directed paths that satisfied certain additional
assumptions. Among these was the property that tracing three spaces forward can be
exactly reversed by another process that traces those spaces in the opposite order. For
example, 3,2,1 is the opposite of 1,2,3. Herbart thus postulated that his generating
process remained similar to itself in both directions forward and backward. The process
A+B (B coming to a fixed A) is the opposite in sense of B+A (A coming to a fixed B).
Likewise A-B is the reverse of B-A. 44 We can treat these as cases where the process
revisits a position it has previously occupied (which we know already to be something
of an idealization). The property of isotropy or indifference to direction is thereby
established through such an assumption.

By adding the property that tracing three spaces is the same process no matter
where it occurs on the line, (i.e., that the spaces 1,2,3 are generated by exactly the same
process as spaces 17, 18, 19, and that translating the place of the process makes no
difference to the spaces produced), the property of homogeneity, or indifference to
place, is established.45 The more properties, or more accurately indifferences, are added
to Herbart's directed paths, the more a traveling process of joining and separation
appears to be a moving point, revisiting the same positions on an enduring line in space.

Compared to the vanishing exchanges of qualities underneath, Herbart's line was, in

the last analysis, an idealization, a space of intelligibilia. As he himself pointed out, the
images A',A", A"', ACnl and B', B", B"', sCnl, used in the actual collecting up of previously
reproduced points were only abstractions (leere Bilder). They did not subsist after the
constructing process moved on, because all previous places touched upon changed in
the interim; thus "intelligible space" depended on whether one remembered, or chose to
remember, revisited points as part of "the same" line. This repetition of the same is the
abstraction that supports both basic properties of isotropy and homogeneity. For
Herbart, if there were certain remembered regularities in the causal structure of the
community, the tracing process could be imagined to return the way it had come,
revisiting, more or less, the same places. The regularity of these causal relations makes
possible the apparent tracing out of uniform lines in one space over and above the
particulars in it.

Herbart went on to describe the tracing of surfaces and spherically propagating

waves by adding independent Wesens C, D that could undergo processes of Together
and Not-Together with A and B. For the four processes A+B, B+A, A-B, B-A there
were sixteen further combinations the process can undergo with a Wesen C.

Herbart also called explicit attention to the fact that there was no possible
correspondence between a simple Nacheinander and a continuous process such as the
tracing of a hypotenuse or a circular arc (Kreislinie). He regarded these magnitudes as
produced by functions in which the independent variables could be simple
Nacheinanders, in step with one another, without the dependent variables being so
representable. Thus, the only way he could represent a continuous magnitude was as "a
flowing of points into one another" to any degree of density. 46 Herbart was aware of the
problems presented by continuity, but could not resolve them except by positing a
different kind of line altogether, the stetige Linie, which he placed in a different logical
category, the set of functions.

Going back to the community on which the entire construction was based, Herbart
had resolved the lingering metaphysical problem of fixed oppositions by introducing
gaps of dissociation. For the qualities of a community of Wesen, there was still a
universal dependence of all on all, without a particular dependence that always held
between any two individual qualities, or Wesens. These were permitted to slip in and
out of dependence, supporting the waves of action that traced out stretches of space.
Such general dependence without particular dependence was what allowed alternating,
independent gaps to subsist and, with them, the concept of extended magnitudes.
Herbart's assumption of standing dissociations in a community otherwise made up of
rigidly associated qualities, or energies, is interesting and it is possible that this step may
have influenced Mach's decision to admit standing differences in the potential
differences of natural causes of change, referred to in the last chapter.


In the "Vortriige uber Psychophysik," Mach was quite favorably disposed toward
Herbart's space construction. He endorsed the idea that space was "nothing original"
and that it ''results from a multiplicity of Wesen,'"'7 like Herbart's intelligible space.
Mach had, of course, by this time made his own addition of multiple dimensions and the
extension to the necessary multidirectional generating processes which he really
believed could be found at the atomic scale. Mach also accepted the mechanism of the
reproduction series, which I don't think he or other German psychologists really ever
gave up, but he thought it failed to explain why every sense is not spatial, since sounds,
tastes, and smells could all conceivably give rise to similar series.48 This shortcoming
reflects Herbart's contempt for physiology and his disposition to keep his analysis
strictly phenomenal. But as a sense-physiologist already with some experience of the
nature and structure of organs like the ear and eye, Mach demanded that the constitution
of the organ of sense must also have something to do with the phenomenal structure of
what is sensed there. To this end Mach turned his "Vortriige" discussion to the theories
of Hermann Lotze's theory of local signs, and thence to Wundt's modification of this
theory for full spatial vision.49


Hermann Lotze was a metaphysician-psychologist like Herbart, whose ideas on space

had import for both the conceptual nature of space as well as its psychological
manifestations. His great breakthrough came in the doctrine of Lokalzeichen, or "local
signs." Lotze observed that when different parts of an organ like the skin are stimulated,
in addition to the quality of the sensation felt there (heat, cold, pain, displeasure), the
mind also received a distinctively different local feeling. We learn from experience that
the same local sign may be associated with any of a range of qualities, and that more or
less any quality can be associated with each of the different local signs. (This is not
precisely true since quality and local sign are always slightly dependent, for example
color sensations vary with their distance from the fovea.)

According to Lotze, local signs were unextended feelings different only in their
quality. As such they formed a mosaic of partitioned sensations but not a connected
manifold of points. A mere collection has no dimensions, distances or directions from
one point to another. 50 As William James later remarked, local signs are feelings of
place without space.

Local signs and qualities always occur together; every quality has a locality and
every locality is sensed along with a quality that occurs there. Without a specific
psychological organization to dissociate them, they would remain fused and
indistinguishable. Lotze believed that the mind was innately organized to construct
spatial appearances by separating out this data. With repeated experience and
recollection, it will be noticed (if not consciously then unconsciously) that the same

quality Q is associated at one time or another with any of a series of local signs A,B,C.
In memory we can dissociate the quality from the local sign, since the appearance of Q
doesn't trigger the recollection of any of the A,B,C uniquely. In like fashion, any local
sign A may be dissociated from its qualities P,Q,R.

However, qualities always strongly recall past qualities associated with them, and
local signs strongly recall past local signs. So even if a quality Q changes its local sign,
the mind interprets this as an identical quality in each new location, moving against an
independent background of local signs A,B,C. The other, completely equivalent,
relative motion is that Q be sensed as fixed and A,B,C (affixed to the body of the
creature) be sensed as in motion. Mach pointed out in the Analysis twenty years later
that it is very easy to get into confusion about which case is which, especially if there
are no peripheral sensations to tell us whether our bodies are in motion or not. If we
stand on a bridge and stare fixedly at the water below, soon the water is seen at rest and
we and the bridge are felt to move. 5 1

When motion does occur, instead of a mosaic of fused quality-and-local sign tiles
together we sense groupings of qualities detached and moving against a fixed
background of local signs on our bodies. The qualities that move together are sensed as
a connected object, not a mosaic oflittle separable dots, and the background is sensed as
a region of connected points. Both series first become connected manifolds by motion
across the other.

The explanation draws on the reproduction series. Say that qualities, not yet
connected, A,B,C,D,E, fall on three local signs K,L,M and move from left to right. At
first, in (1 ), A is associated with K and they will recall each other on their reappearance;
as B will recall L, and C, M.

(l) ABCDE (2) BCDE (3)CDE


Thus in (2), K reappears and recalls A, but now with K's present partner B. In
Herbartian terms the fading after-image of A on the spot K becomes inhibited by the
new B. Similarly, L recalls B's after-image, but now conjoined with L's new partner C,
while M recalls C's after-image and conjoins it with D. A,B,C thus come into
association with one another through their after-images, while gradually dissociating
from K,L,M in the same motion. In (3) the process continues: Know recalls B's after-
image with C, L recalls C's after-image with D, and M recalls D's after-image with E.
A,B,C,D,E all end up associated with one another and connected, as they are
progressively dissociated from KLM.

In the opposite direction, the qualities A,B,C,D,E also recall the local signs K,L,M
and associate them with one another, while dissociating them from the quality row. If
we add motion in the vertical, the fused mosaic comes to look like two connected
surfaces sliding over each other, one invisible and fixed due to the arrangement of local
signs on our bodies, the other containing the world of colored and solid objects.

To account for the almost constant motion required by the theory, Lotze devoted
special attention to the eye movements, especially a reflex that fixates all distal objects
at the center of the visual field, as the mechanism for keeping the associative and
dissociative processes going all the time. As Mach pointed out, Lotze's construction of
the monocular visual field is thus more realistic than Herbart's radially extending
reproduction series because of the inclusion of physiologically based local signs. In fact,
the orientation of German psychology was to start with general mathematical and
metaphysical constructions and to embody those constructions more and more
concretely with the help of the latest physiological research.


Nowhere is this direction away from metaphysics and toward the empirical more
pronounced than in the work of Wilhelm Wundt, who modifed Lotze's local sign theory
and adjoined further anatomical and psychological details. His theory of spatial vision is
called the "complex local sign theory." 52 Like the others, Wundt held that visual space
was a construction out of non spatial elements, associated and dissociated by the eye
movements. Long after Mach and many other psychologists had become nativists
Wundt wrote:
Spatial ideas of sight cannot be considered original and immediately given... any more than can the
spatial ideas of touch. The spatial order, here, too, is developed from the combination of certain
sensational components, which, taken separately, have no spatial attributes. 53

Wundt thought that Lotze had achieved half of the construction of monocular
visual space by turning a fused mosaic into two connected manifolds of points sliding
over each other. However, Lotze's construction did not explain other spatial properties
of distances traced between points and visual directions. We have, with Lotze, only a
diffuse hang-togetherness of points, extension without metric properties: more than a
mosaic of places, but less than total spatial perception.

According to Wundt, the muscles of the eye accounted for what we would now call
the the metrical properties of the visual field. He graded the muscular feelings from the
center of the eye in direction and intensity. These feelings formed a second local-sign
system associated with the retinal local signs. Whenever an image falls on a peripheral
point of the visual field, a reflex moves the eye to fixate that point at the fovea, the
center of the retina. In this way, Wundt explained, all of the local signs on the retina
come to possess a direction, and an intensity of effort needed to fixate the given point at
the center of the visual field. Through practice, these reflexes remain lively even when

we try to stare dead-ahead. Direction and distance are thus interpreted, respectively, as
the type and intensity of effort the ocular muscles would have to expend to fixate that

As Wundt said, and Mach repeated, the retinal mosaic and the system of the ocular
muscles form a fixed register into which sensation qualities are entered with respect to
place, direction, and distances from one another. We bring this semi-innate spatial
representation to all visual experiences, and are so used to it that we think we see these
metric properties fully formed solely by retinal images.

In support of his theory, Wundt called attention to a whole class of illusions of

direction and length of lines due to the eye's musculature: for example, because it is
easier for the eyes to look down than up, vertical lines look longer, making true squares
look like vertically stretched rectangles. Mach had already discovered some of the
experimental proof in his 1861 article "Ober das Sehen von Lagen und Winkeln durch
die Bewegung des Auges." 54 As Wundt reported, patients with partly paralyzed eye
muscles overestimate the distances to objects in the direction of the most difficult ocular
movements. 55

Wundt also had physiological evidence that distance in the visual field is not
estimated on the surface of the retina. The retina itself looks like a mosaic of tiles, but
unlike a true mosaic they are unevenly spaced: tightly packed receptors at the fovea and
spaced out more widely and irregularly toward the periphery. Yet we rate the distances
between two spots the same, whether in the center of vision or the sides, even when the
actual retinal arc between points is twice as great. 56 This does argue that distance is
indeed a property determined independent of locality, a fact that makes it possible to
separate conceptually two sorts of manifold extensions: those that exhibit metric
properties such as distances that are independent of place and properties that differ from
one locality to another. Moreover in a manifold extension that does have a metric, it is
possible to imagine that the metric property and the properties used to determine
coordinate locations are different.


A reading of Mach's views of psychological space must begin with Wundt's complex
local sign theory, which he soon expanded to include sensations of series, symmetry,
and shape. But as we will see in the next chapter, Mach veered quickly away from
Wundt's empiricism and toward a nativism about spatial representation in humans and
animals, similar to that of his Prague colleague Ewald Hering. Mach the psychologist
came to reject the idea of building up psychological space from unextended elements, as
he said in his 1902-3 Monist articles later gathered as "Space and Geometry": "The
attempt which is frequently made of deriving tactual space psychologically ... by the aid
of the concept of time and on the assumption of space-less sensations is an altogether
futile one. "57 This did not mean however that the philosophical idea of deriving
extension from intensity had been lost on him, and certainly his "Raumvorstellungen"

paper indicated that his physical theory of elements would carry forward such a

Mach the physicist was a thorough going empiricist about space, by which I mean
that he did not consider space to be a fundamental concept of physical science, but
rather a complex conception that needed to be analyzed and reduced to its empirical
components. Mach's empiricist German predecessors in psychology had attempted to
show, for example, that a perception of space was already quite complex, containing
many separable experiences: those of place, extension, direction, and distance, all
separted. The psychologists also showed that the various philosophical conceptions
involved in extension could be built up conceptually out of certain functional
associations and dissociations of intensive content. Mach's view of physical space
remained close to those of the empiricist psychologists, especially to Herbart. This was
nowhere more evident than in Mach's idea of a chemical manifold, a general physical
manifold in which all forms of energy transfer could be mapped, from gravitation to
heat and even chemical and physiological reactions, which he had already mentioned in
his "Raumvorstellungen" paper, and which he developed further.

Mach also described the characteristics of this same physical manifold further in a
section of"Space and Geometry" devoted to multidimensional spaces:

Physics would be justified in considering an extended material continuum, to each point of which a
temperature, a magnetic, electric and gravitational potential were ascribed, as a portion or section of
a multiple manifold. 58

The multiple manifold was one that was enlarged by the addition of physiological
and chemical data. Is it therefore so difficult to imagine that Mach seriously considered
coming forward with a physical analog to Herbart's construction, for example one in
which potentials took the place of the qualities and in which many details of the space
construction could be brought over into physics, such as the Herbartian concept of
extension by means of opposition and independent dissociations and action among the
qualities, as well as the space-tracing waves of action as elements of the extended
manifold were drawn out? I think it very likely that Herbart's construction remained in
Mach's thoughts, and, as we shall see with further evidence from the Nachlass brought
to bear, Mach indeed imagined something like intelligible space as a pattern for a
physical construal of space. Thus, the influence of Herbart on Mach may be cited as a
first-class example, and by no means the only one, of the rampant night trading in ideas
bei Nacht und Nebel between metaphysicians and scientists in German-speaking


In 1864, after his lean years as a Viennese Privatdozent, Mach obtained a position in
Graz as professor of mathematics. He later claimed that Graz was something of a
neglected institution, but that there he was able to improve his self-acquired
mathematical education and to pursue experiments in sense physiology. Mach's lectures
there included a reprise of his Vienna lectures on psychophysics in the winter of 1864-
65, and lectures on mechanics in the summer of 1864 in which he developed most of his
views on physical space and time and on multidimensional spaces, without, as he
pointly remarks, having read the famous Probevorlesung of Riemann which was
published in 1867. 1

It was during Mach's time in Graz that he discovered what have come to be known
as the Mach Bands? These phenomena occurred when a light distribution, on a wall or
on a rotating disk, exhibited a flexion, i.e., a point either above or below the average
brightness of immediately neighboring points. Mach argued that because the eye senses
the brightness of a point only in relation to its environment, this had the result that at
such flexions, where the illuminance suddenly turned upward or downward, an
exaggerated high spiking appeared in light or dark bands. Mach's physiological
explanation in terms of processes of lateral inhibition in the retina became standard and
was later extended by the Nobel Prize winner Georg von Bekesy to other senses, calling
it "Mach's law of contrasts." In an episode I have discussed elsewhere, Mach attributed
his law of inhibition to sensations of depth. 3

Mach soon expanded his researches on space sensations of the eye to the
phenomena of symmetry. He began with Wundt's empiricist complex local sign theory,
according to which the retina was represented as a place-mosaic of local signs while the
property of distance from one point to another was attributed to the eye movements and
the muscular feelings associated with them. However, Mach interpreted Wundt's mosaic
and musculature of the eye as an innate, extended framework of space sensations,
waiting to be filled in by experience.

In his article of 1865, "Bemerkungen tiber die Lehre vom raumlichen Sehen," 4
Mach asked how we perceive similarities of spatial form even when the individual
qualities are different: for instance, we recognize the "circularity" of both a blue circle
and a yellow circle. Herbartian reproduction series could not explain this because their
inhibition and reproduction all depended on the qualities involved and their degree of
opposition. Mach pointed out that the yellow and blue circles would reproduce each
other if they possessed identical "form-qualities" over and above their different colors.

These form-qualities were in fact Wundt's register of muscular sensations:

If two differently colored figures of equal size reproduce each other and are recognized as equal,
the result can be due to nothing but the existence in both series of perceptions of an idea or percept
that is qualitatively the same. The colors are different Consequently, like or equal percepts must
be connected with the colors which. are yet independent of the muscular feelings of the eye when
confronted by the two figures. We might say that we reach the vision of space by the registering of
light-sensations in a schedule of graduated mnscle sensations.5

He added that sensations of form were also to be found in ornamental motifs and in
music, where the same melody is recognized through different absolute pitches of the
notes and the same rhythm is recognized independent of the absolute durations of the
beats.6 Mach thought that whenever two experiences are judged similar, there must be
elements within those experiences that are exactly the same. Hence his formula:
similarity is partial identity. Yet the mind must not be capable of inferring all possible
patterns of similarity, for then everything would be similar to everything else in some
respect. Thus, Mach held that we are physiologically predisposed to sense only certain
identities, in sound, rhythm, spatial form; the rest we ignore. He described the strategy
in the Analysis:

If I see figures which are the same in size and shape but differently colored, I seek, in connexion
with the different color sensations, certain identical space sensations and correspondingly identical
nerve processes. If two figures are similar (that is, if they yield partly identical space sensations)
then the corresponding nerve processes also contain part identical components. If two different
melodies have the same rhythm, then, side by side with the different tone-sensations there exists in
both cases an identical time-sensation with identical corresponding nerve processes. If two
melodies of different pitch are identical then the tone-sensations as well as their physiological
conditions have, in spite of their different pitches, identical constituents. 7

In his 1865 article "Untersuchungen fiber den Zeitsinn des Ohres," Mach wrote that
time sensations, which accompanied identical or similar rhythms, were also a "register"
in the body or a "contour outline" of experience in which the individual qualities are

One can completely liberate the rhythm of a melody from the sensations of tone, like the contour
drawing of a painting, in that one can tap out the melody without tones. This would not be
possible if there were not, to some extent, self-standing series of rhythm sensations.8

As Mach pointed out in his next major article, "Bemerkungen fiber die
Entwicklung der Raumvorstellungen," this innate register of space sensations also had
certain physiological peculiarities that show up prominently in the phenomena of
symmetry. Repetitions of an ornamental motif, even in various sizes, tends to condition
a repetition of the muscular sensations over the shapes, and a pleasurable aesthetic
feeling of "relie"

Psychological spatial symmetry, for Mach, was also to be explained as a repetition

of identical form sensations. For example, vertical symmetry is immediately noticed and

occurs when the plane of symmetry of objects divides the head on either side of the
nose. However, horizontal symmetry-the symmetry of a house and its upside-down
image in a lake-is not immediately noticed and requires a special intellectual effort.9
Hence, Mach claimed that geometrically similar figures were only psychologically
similar if they were situated such that they occasioned the same muscular sensations. He
got into deep water here, since there seem to be too many similarities of form to keep up
with. We sense the similarity of two straight lines of different sizes, inclined at different
angles to each another, although none of these scanning motions measures out identical
distances or muscular strains. But on the other hand, there are many opportunities to
infer optical similarities based on identical scanning motions and muscular strains
where we do not.

To keep the similarities from proliferating, Mach had to postulate certain innate
preferences for the eye. For example, he thought that straight lines, when they are
symmetrically (horizontally or even vertically) situated, produce special physiological
effects, which help us to recognize them as similar later when they are situated

The straight line is defined geometrically as the shortest distance between two points, which
already for that reason is peculiar since all measurements of length presuppose the straight line.
This definition certainly has no psychological meaning. I have already mentioned earlier that the
straight line can be laid down (vertically or horizontally) so that both of its retinal images fall on
identical or symmetrical points. This is what gives the straight line a preference before other
curves and this psychological preference is not without significance for the history of geometry. 10

Straight lines are all similar because identical sensations of direction are felt when
the eye moves from one point on a line to another, a pleasurable sensation. Further, the
sensations produced by moving to a position on a straight line is always the average of
the intensity of sensations of moving to neighboring points, which is not true of curves.
Thus, a physiological preference for straightness emerges. A curve may cast the image
of a straight line when it is seen edge-on, and a long straight line tilted out of the plane
of vision may cast a short retinal image, but these possibilities are almost always
disregarded. Whenever a retinal image can be viewed as a straight line, with all its
points at the same depth, it will be. 11

Mach held as a general physiological principle that sensory systems are geared to
notice any above-average deviations of sensations from a resting position. The resting
position is where the intensity of any given sensation intensity (whether of length, depth,
light, or muscular strain) is the average of the intensity of its neighbors and no special
contrasts occur. 12 This resting position is always preferred unless the sensory system is
forced out of it. Physiologically speaking, curves and their first and second derivatives
are perceived as forced deviations from the resting position of the straight line. 13
According to Mach, this is why we can perceive directly both the slope of a curve (its
first derivative) and its curvature (its second derivative), but not higher-order
derivatives. He believed we simply have no organs for sensing those higher-order

Mach was so taken by Wundt's register of "muscular space" that he added to it a

sensation of distance not associated just with the performed ocular motions, but with the
mere will to move the eyes to a certain position. He said quite bluntly: "The will to
perform movements of the eyes, or the innervation to the act, is itself the space
sensation." 14 Now, Mach was certainly no believer in a Kantian noumenal freedom by
"willed" eye motions, so what could he have meant by this remark? Some reviewers of
the first edition of the Analysis of Sensations, such as Carl Stumpf, were simply baffled
. 15
by It.

In fact Mach introduced these central innervations of will for the mundane purpose
of distinguishing between cases of relative motion, where impressions move over our
fixed bodies, and cases where the body moves against a seemingly fixed background of
objects (as in the illusion of the bridge and the river). 16 In addition, by 1871, after his
move to Prague where advanced equipment was made available to him, Mach
discovered a series of involuntary eye movements triggered by conditions of
disequilibrium in the semicircular canals and the otolith apparatus of the inner ear, a
mechanism discovered simultaneously by Freud's friend Josef Breuer and by Crume
Browne. The presence of such a mechanism, however, was anticipated much earlier by
the remarkable William Charles Wells. 17 These compensatory motions are often
unnoticed in everyday life. For example, most people are not aware that their eyes are
capable of rolling motions about their centers. 18 One only needs to stand in front of a
mirror and tilt one's head to observe the eyes rotating in their sockets in the opposite
direction. Mach discovered while riding in a train that as the train rounded a curve and
the centrifugal forces pulled to the side, his eyes involuntarily rotated as if the side of
the train were in the position of being straight down and he saw the inside of the train
suddenly aslant. In work that he conducted using a special rotating chair in Prague,
Mach discovered several other involuntary movements of the eyes indicating
considerable spatial sophistication. As the body rotates or the eyes follow a moving
object like the cars of a train, the eyes follow the first car for a time and then suddenly
snap back to the next car, without creating the sensation of a jagged discontinuity. Mach
found that by inducing angular accelerations he could give rise to these eye movements
in a subject completely enclosed in his special rotating chair.

The effect of these eye movements is, in short, to compensate for motions of the
head and to keep images of objects fixed on the retinal surfaces, without a special act of
attention. Of course, as in the illusion of the bridge and the river, the same relative
motions are produced when we are moving our eyes to fixate objects moving across the
retinas. According to Mach, the way we tell the difference between such equivalent
relative motions is that the voluntary innervations to our eye muscles tell us that we are
moving our eyes; hence in these cases the world will appear fixed. 19

Now getting back to the question of will and the space sensations, Mach allowed
for the voluntary motions of the eyes to leave memory traces left behind after they had
pointed the eyes to a certain place in visual space. Thus, even when the eyes are not
currently pointed to that place, the disposition to do so persists. According to Mach,
memory images of these early reflexes endure and occasion the new, so-called
voluntary motions.20 In the case of the eye-musculature, this memory image is of a
previously performed motion that led to that desired result, a rather delicate problem of
associating external points in space with points on the body and its complex register of
local feelings, musculature, and eye movements. The set of memory images of
previously performed fixations constitutes a memory space of possible directions and
distances. However, once the reflexes are trained up, and the manifold of place
sensations firmly fixed in body memory, all of those complex motions are coalesced
into a simple will to move to a certain position, and the mind only needs to be conscious
of the places themselves and not the complex procedures of reaching them.

Like Wundt, Mach had an experiment to support his view that the will to move is
the crucial thing in a perception of space: if the eyes are fixed in place with putty and
we merely will to turn them to the right, objects in the visual field suffer a strong
rightward displacement, although the retinal images have not moved. 21 I suppose that
Mach's biological view culminates in the conception that the many spatial manifolds of
the body, including the locative properties of sound, become fused into a single,
imaginary space with a manifold of memories for places and the will to fixate them.
Perhaps because of the prominent role for the will, Mach's central innervation theory
for distinguishing relative motions came under attack from William James-hostile to
any notion of will, even in Mach's watered-down sense--and Hugo Munsterberg, who
both argued that the relative motion of the body is really sensed by peripheral
kinaesthetic sensations of the skin and muscles and not through central innervations?2
Mach came back with the response that the innervations need not be sensed
consciously. 23 As some commentators have observered, Mach held a remarkably beefy
notion of will as a directly experienced sensation of force or pressure, and he was not
averse even to Schopenhauer' s conception that something similar to the human will
existed in nature, der Wille in der Natur. 24

One could describe Mach's position on biological space representation as

"evolutionary empiricist" and "individually nativist." He came to regard the system of
involuntary movements necessary to fixate points as innately wired together in the
organism, based on the accumulated experience of untold previous generations of
organisms moving about in space. In addition to the sophisticated eye-movements in
humans, one example that impressed him was observing that newly hatched chicks
could peck accurately at what they saw with practically no experience. Mach marveled
at the fact that they were simply born with neck and beak movements perfectly
coordinated with fixated spots on their retinas, which it takes human babies some time
to master.

In a footnote to this development, Mach also adopted the extremely anti-empiricist

position of Hering and James of "intensive extensity," i.e., that sensations of color,
sound, or touch possess extensity, not incidentally or in association with one another but
as part and parcel of the qualitative experience. Hering for example described colors as
having "roominess" and James described pains and aches as being small and sharp or
large and voluminous?5 Mach gives the example of high pitched tones as possessing
sharpness and tiny extension whereas lower tones seemed to "fill the head."26 Here the
antipode of empiricist notions of psychological space is reached. But whether the three
were onto a real phenomenon here, perhaps common to all qualities, physical and
psychical, or whether they made a common introspective mistake of failing to
disaggregate complex properties, it is hard to say. Certainly nothing logically rules out
the idea that qualities may possess volume or that psychological extension could be a
directly perceived property similar to hue or pitch. Indeed a certain "inverse
relationship" is suggested by the examples of the Hering-Mach-James triumvirate,
between those qualitative characteristics indicating high intensity (high pitches, sharp
pains) and feelings of small extension as well as characteristics of low intensity (low
pitches and dull, aching pains, roomy colors) and feelings of great extension.


By no overestimation, the 1860s in Germany saw a revolution in geometry. Riemann's

famous Habilitationsschri.ft and Helmholtz's mathematical treatment of differential
geometry showed that the three-dimensional flat space of Euclid was but one of many
possible spaces of constant curvature. These new spaces might have any number of
dimensions, they might have various degrees of curvature (positive, zero, and negative),
or they might be simple extended manifolds of place without the property of distance or

To a sense psychologist of the 1860s such as Mach, the new differential geometry
of Gauss, Riemann and Helmholtz would have appeared a quite reasonable
development, with much in common with the ideas of Lotze and Wundt on visual space.
That different senses exhibited different types of manifold extension with different
numbers of dimensions and perhaps a different geometry from the Euclidian was more
or less accepted as established. Helmholtz knew that lengths in the visual field shrink
and grow from one place to another and that the same was true on the skin, and he even
wrote that he was led to investigate metric properties from his researches in sense-

In the [Handbook of] Physiological Optics I had before me two examples of other spatially
representable manifolds capable of change in several senses, namely the system of colors, which
Riemann also mentions, and the measuring out [AILI'messung] of the visual field with the glance
[AugenmaaYs]. Both show fundamental differences from the measurement system of geometry and
provide the stimulus for a comparison. 1

Sense physiologists knew it could not be taken for granted in all manifolds that
length was independent of position or inclination, as was customarily the case in
geometry. In vision, and on the skin, the length of a line changed from one locality to
another and, when its situation changed, from one angle to another. Mach himself had
investigated such cases in one of his earliest publications. 2

Mach also came by himself to the notion that spatial determinations were, in part,
determinations of measure--one of Riemann's signal contributions. In his 1866
"Raumvorstellungen" paper, Mach called geometry the science of "metric space"
because he thought geometrical axioms rested in the last analysis upon facts about
measurement. He said that geometry laid down a standard measure-a fixed length,
area, or volume- and instituted procedures for carrying this measure into congruence
with every other length, area, or volume through motion. If these relations are upheld, it
follows that every spatial magnitude can be compared with every other through their
relations to the standard length. 3 Riemann had written in his Probevor/esung that the

critical empirical hypothesis to ensure the congruence of lines and figures is "that the
length of lines be independent of their situation, that every line be measurable by every

Now, it may well be, as Mach said, that he did not know of Riemann's ideas until
1867. The earliest mention of Riemann among Mach's papers occurred in 1865 while he
was at Graz, administrating a hiring committee for the philosophical faculty. One of the
prospectives was Hermann Hankel, a mathematician who had studied at Gottingen and
had been present on the few occasions when Riemann had treated of his new points of
view exhaustively. 5 The report does not mention geometry per se. Instead, Mach wrote
in his report on Hankel that: "it should be especially emphasized his familiarity with the
Riemannean theory of complex functions and with the so-called new geometry as
founded by Mobius and Steiner. "6

The common denominator between Mach and Riemann was Herbart, as both
men had studied Herbart's Psychology and Metaphysics and become familiar with
the notions of extension employed there. 7 Recall that Herbart had considered
qualitative continua of color or tone in which the intermediate places could be
determined by proportions of the two extremes, literally "transitions," as the one
quality overcomes the other by force.

If we take a direct analogy with spatial continua, the opposed qualities are like
the directions of left and right. One cannot, I think, imagine the hues or tones all
spread out and given at once, like the keys of a piano, or the number line from -oo, 0,
+oo. Rather, every process of opposition, or transition from the one hue to the other,
is a particular path traced out against the background of all possible transitions
between the extremes. Each path of shading off for one particular opposition is thus
representable as a function of the exchange of these two processes acting at that
point. Herbart's conception of these processes as forces or pressures acting in
opposing directions, and with different intensities, might be a good representation of
how he thought of these hues. And if neither one is able to overcome the other, the
point is situated equally between them at a kind of "zero point." By then imagining
all possible paths as simultaneously produced, or produceable, in both directions it is
possible to manifest the numerous dynamical relations between qualities as if they
were located along a single dimensional continuum such as a continuous line from -
oo, 0, +oo. Although only one state is manifested at a time, previous and future states
can be substituted in imagination so that the impression of a progression through a
series is retained. Herbart himself makes the comparison between motion in these
continua and motion in space. 8


As we shall see now, Herbart seems to have influenced Riemann on some

fundamental matters relating to the mathematician's concept of a manifold.
Riemann's acknowledgment in the famous 1854 Probevorlesung on geometry reads:

While I now attempt to solve the ... problem of the development of the concept of multiply
extended manifolds, I think myself the more entitled to ask considerate judgment inasmuch as
I have had little practice in such matters of a philosophical nature, where the difficulty lies
more in the concepts than in the construction, and because I have not been able to make use of
any preliminary studies whatever, aside from some very brief hints which privy Counsellor
Gauss has given on the subject in his second essay on biquadratic residues and in his Jubliee
booklet, and some philosophical investigations of Herbart. 9

Riemann may have come into contact with Herbart's ideas in Berlin and Gottingen,
where he studied philosophical subjects, among other things Naturphilosophie.
Herbart of course had a great influence in both cities at the time and had finished his
career in Gottingen. It is not therefore surprising that Riemann's published Nachlass
contains a number of notes about Herbart's metaphysics and psychology, including
the following remarks:

My major work concerns a new conception of the known natural laws-the expression of them
by means of other fundamental concepts-through which it becomes possible to use
experimental data concerning the reciprocal action between heat, light, magnetism and
electricity for the exploration of their interconnection. I was led to it in the main by the study
of the works of Newton, Euler and, from another side, Herbart. As far as the last, I adhere
almost completely to the earliest investigations of Herbart, the results of which are expressed
in his Promotions and Habilitations-theses (22 and 23 1804), but must diverge from the course
of his later speculations in one essential point, which is determined by a difference in relation
to his Naturphilosophie and those laws of psychology which connect it to the

The author is a Herbartian in Psychology and in the theory of knowledge (Methodology and
Eidolologie), but for the most part he cannot own himself a follower of Herbart's natural
philosophy and the metaphysical disciplines related to it (Ontology and Synechologie ). 10

Ontologie and Synechologie are the sections in Herbart's Metaphysik devoted to his
Wesen and to intelligible space, and a number of authors, including Bertrand Russell,
have used this partial dismissal to conclude that it was Herbartian psychology, in
specific the reproduction series and quality continua such as color manifolds, not his
metaphysics, that made an impression on Riemann. According to Russell:

Herbart's actual views on Geometry which are to be found chiefly in his Synechologie, are not
of any great value, and have borne no great fruit in the development of the subject. But his
psychological theory of space, his construction of extension out of series of points, his
comparison of space with the tone and color series, his general preference for the discrete over
the continuous, and finally his belief in the great importance of classifying space with other
forms of series (Reihenformen) gave rise to many of Riemann's epoch-making speculations

and encouraged the attempt to explain the nature of space by its analytical and quantitative
aspect alone. 11

Not every writer is willing to give Herbart even this much credit. 12 One
exception is Erhard Scholz, who published several extracts from Riemann's
Nachlass, demonstrating, as we will see, a clear connection between Herbart's
qualitative continua and Riemann's notion of manifold extension. 13 Another writer,
Luciano Boi, who considers the relationship between Riemann's geometry and
Naturphilosophie, writes that Riemann "considers relations of extension by
comparison with a continuum of qualitatively different properties (quantitatively
undifferentiable), characterized by a kind of natural topology." 14 This seems to me
both the right view and the one that Mach took of Riemann's work in his writings on
the subject.

In the 1854 Probevorlesung, Riemann developed the concept of multiply

extended magnitude and its curvilinear coordinate system in the first part. Then, in
the second part, he laid down the requirements for relations of measure on this
manifold and a suitably generalized defmition of the intrinsic curvature of the space.
What concerns us here is only the very beginning of Riemann's essay, the definition
of a multiply extended manifold. There were two stages of this construction: (1 ). The
conceptual defmition of 'multiply extended quantity' (mehrfach ausgedehnte
Grosse), which was a philosophical task, and (2). The determination of a curvilinear
system of coordinates within this extension (an empirical task known as geodesy).
The second was due to Gauss but the first was probably due, in part, to Herbart.

In his "Plan der Untersuchung," Riemann said that "space and the first
fundamental notions for constructions in space" could not be assumed as given a
priori, as Euclid and other geometers had done. 15 He remarked that the basic notions
of geometry had been neglected for so long because "the general concept of multiply
extended magnitudes, in which spatial magnitudes are comprehended, has not been
elaborated at all." This is already a sign that Riemann did not intend to begin with
space but with simpler concepts, leading him to investigate the meaning of raw
extendedness. Riemann offered such a definition of extension in the following
passage by saying that a concept of quantity must first admit of two modes of
determination (Bestimmungsweise) and that as there existed a discrete or a
continuous transition (Ubergang) from one mode to the other, there would then exist
an extended manifold of either points or elements:

Notions of quantity are possible only where there exists already a general concept which
allows of various modes of determination. According as there is or is not found among these
modes of determination a continuous transition from one to the other, they form a continuous or
a discrete manifold; the individual modes are called in the first case points, in the second
elements of the manifold. 16

Riemann claimed that while discrete manifolds were more abundant and
exemplified by words in natural language, "in common life ... the positions of objects
of sense and the colors are probably the only simple notions whose modes of
determination form a multiply extended manifold." 17 By mentioning color so
prominently alongside continuous position in space, and by using the word
"transition" between modes of determination, Riemann seems to refer directly to a
construction of qualitative continua such as Herbart's tone manifold and a
construction of extended magnitude by means of a continuous transition from one
quality to another opposing quality. In his discussion of the Tonlinie, Herbart had
also emphasized that the tones were one dimensional because only a single transition
was possible between the two exchanging states ofhigh and low. 18

Riemann also believed it was necessary to have two modes to determine an

extension, not one, even in the case of discrete manifolds. These modes of
determination are not themselves points, but rather seem to be Herbartian-style
qualities. A point or element of a manifold first arises by means of the transition
between two qualitative means of fixing an individual, as a position on the tone row
can be determined both from low to high or from high to low. In a Nachlass
fragment published by Scholz, Riemann is more explicit:

If a continuous transition is possible from one mode of determination of a changeable object to

every other mode, so do all of the modes of determination (if a continuous transition is
possible from every one of a set of different modes of determination to every other, so the
totality of these forms a) continuous extended manifold; every individual of this manifold is
called a point.

A continuous extended manifold is a non-measureable quantity and thus may not be

represented as a designated variable number. (A variable section of a manifold is thus a
variable magnitude whose values are not designated numbers). 19

The requirement of two qualitative modes for determinations of quantity baffled

Russell, who asked: "What is meant, to begin with, by a general conception capable
of various determinations? Does not this property belong to all conceptions?" Surely
it does, and surely the fact that we possess independent means of determining objects
(color, figure, size, species) helps in isolating an individual from many others. A long
string of adjectives describing independent properties quickly narrows down the
extension of a term to a specific individual. But if those adjectives were dependent in
meaning, or so ambiguous that every two of them picked out the same individuals,
then even discrete manifolds would not be possible.

If modes of determination do have a geometrical analogy, it is to directions and

not points. As in Herbart's one-dimensional Tonlinie, Riemann points out that the
essential mark of a simple manifold is to contain only two directions of forward and

In a concept whose various modes of determination form a continuous manifold, if one passes
in a definite way from one mode of determination to another, the modes of determination
traversed constitute a simply extended manifold and its essential mark is this, that in it progress
is possible from any point only in two directions forwards or backwards.20

The color manifold Riemann had in mind in his remarks was probably one in
which the hues were arranged in transitions like right and left or high and low along
the tone series. But it is impossible to tell from the example exactly which colors
would pair off as "modes of determination." Nor does it really matter. For example,
hues could be classed in pairs of {white, black}, {red, green}, and {blue, yellow}, as
was later the case in the Mach-Hering classification of colors. The red-green
dimension, for example, was a simply extended continuum from the reds to the
greens, with gradually changing proportions as the extremes were approached. By
adding dimensions of blue and yellow and black and white, the manifold had three
separate one-dimensional manifolds, each with two modes of determination separated
from all the others. Thus we can fix an individual in a color space as we might fix a
point two ways in regular space, from the right and from the left.

Now, the next step in the construction of space is to find some means of laying
out a coordinate system on the manifold. In the case of regular space, the coordinate
system of a manifold is determined by a system of determinations carried out within
the manifold itself by means of a procedure known as geodesy. One fmds a function
in which two of the modes of determination are constant and do not transition, while
the others do undergo variations. One then has, as Riemann says, a manifold of one
fewer dimension and a constant function that behaves like a coordinate axis. In
Cartesian coordinates the y axis can be found by determining all of those values for
which x holds constant, and vice versa. But in Gaussian coordinates the functions
may be curves. Furthermore, the coordinate system is simply a classification of
localities and directions and does not imply any distance-information. To use Edna
Kramer's apt example, city blocks are a coordinate grid suitable for finding all of the
houses by their numbers and cross-streets, but no information is given on their
distances from each other.Z 1 It is possible to fmd one's way from any point to any
other simply by means of directions and point-landmarks; for example, start at the
New York Public Library and walk north until you are at Central Park, then turn
right, and so forth.

A color coordinate system could be developed analogous to Riemann's Gaussian

coordinate system by fmding functions in which there is variation along the red and
green dimension, say, while the others remain the same, and so forth for the other
directions. What we have in the case of the colors then is a manifold in which there
are systematically identified places and multiple extension but no conception of a
distance from one place to another. As Riemann says, this requires a metric property,
which is independent of place and capable of moving freely through the manifold
without alteration. A rigid body such as a meter stick retains its shape and length
through motion from one coordinate position to another, as do analogous shapes such
as spherical and concave triangles in spaces of constant curvature. Color coordinates,

like city blocks, can be laid down without the need for a metric and the equivalent of
a rigid meter stick must be sought "outside the manifold. "22

Thus, Russell's criticism of Riemann's analogy of colors with space, viz., that he
had given no means of comparing one color with another by motion or superposition,
is quite unfair.Z3 The metric property, if there is one, need not be the same property
by which the coordinates are determined. In fact, it is better to pull apart the stages of
manifold determination and metric considerations as Riemann does. Russell-who
apparently did not have sense-physiological cases in mind-believed it was
"impossible to set up a coordinate system in a manifold in which free mobility or the
preservation of distance from one position to another did not also hold." But this is
only true if the coordinate system is laid down with the help of spatial distance
measurement, and then a property of distance surreptitiously derived by means of
those coordinates (then, indeed, Riemannian geometry would be guilty of question
begging, as Russell accuses).Z4 But the whole point of comparisons with the color
manifold, or with visual space, is to show that the properties of locality and direction
are prior to distance. Russell also does not see that his own main desideratum of a
philosophy of space, namely that spatial properties be abstracted from more
fundamental qualitative determinations/5 is the very thing that Riemann sought to

If there is a criticism to make from a philosophical point of view, it is that

Riemann's conception of constructing point or element manifolds from two
associated modes of determination does not seem to include Herbart's fundamental
idea that these modes, or directions, should both associate and dissociate from one
another in the tracing out of extended magnitudes. This problem even affects the
color and tone modes, for their directions remain rigidly set against one another and
hence fully dependent. Hence to get some motion out of them requires a second
series of modes of determination or a way for the directions to come apart, change,
and come together again. Consider, for example, the passage, or change, from a certain
qualitative coordinate (B,Y) to a certain other with a different proportion of blue to
yellow (B',Y~ close to it but different. Start with B andY dependent on each other:


Now imagine that B and Y dissociate and change their values to B' and Y' as they come
into dependence on two qualities X' and Z' from the background:

{{B,Y}, {{B',X'}, {Y',Z'} }}.


Let us then say that the new values B' and Y' are in a new proportion-whereby
whatever B' has lost or gained from B, Y' has gained or lost from Y-and thus can join
back together at those new values, B' andY':

{{{B,Y}, {{B',X'}, {Y',Z'}} }, {B',Y'} }.

That completes the passage from the coordinates (B,Y) to (B',Y'). The collection of the
last stage would then be the simplest concept of quantity or extension, a progressive
sequencing of associations and dissociations of two different modes B and Y with the
help of certain background qualities (X, Z).


Mach had read Riemann on geometry by 1872 as evidenced by his reference to the
work in his notes to the Conservation of Energy, if only to remark that he did not
know of Riemann's Probevorlesung before 1866. In any case, Mach did make the
comparison of space to other sensory manifolds as early as 1862, as part of his
investigations of the inner ear and of the proper phenomenological description of the
manifold of tone sensations. Certainly the similarities between Mach's and
Riemann's views could have been due to their common influence by Herbart and
encountering qualitative continua in his writings.

For his tone row, Mach took as his qualitative "coordinates" a combination of two
qualities, or modes of determination: Dull [Dumpj] and Bright [He/1]. 26 Each tone
sensation on the row could be located immediately by its different proportion of the two
qualities, more Dull in the mix of a point near the low end, more Bright near the high
end of the scale?7 A tone sensation on the row thus has two qualitative modes of
determination and coordinates constructed according to the simplest rule,


or if H goes up, D goes down, and vice versa. Apart from the properties and processes
that set the coordinates, Mach claimed that a separate property determined an invariant
interval in the tone row/ 8 analogous to a distance. His expression for the interval (in
terms of the rate of vibration) was

log (n'/n)= Const. 29

which expresses "the constancy of the length on displacement" 30 just like a rigid meter
stick moved about through space: "The unalterable, substantial physical object which
we sense as an interval is for the ear temporally determined, whereas the analogous
object for the senses of sight and touch is spatially determined. "31

Thus, although the tone row does not have complete symmetry in the directions of high
and low, there is a clue that a tone invariant exists, since the same "notes" repeat one
octave higher in pitch-a kind of symmetry. In fact, every tone has "partial tones" that
occur at other intervals-the third, the fifth and so forth. When two tones have
overlapping partials, the ear hears out the "distance" between them. Mach also
remarked in his symmetry paper that when a melody is switched from major to minor
key it is still recognizeable as the same melody. However, Mach's example suffers from
the fact that the invariant is not a function of the coordinates, but of rates of vibration n'/n
(which I suppose could be correlated to sensation qualities).


In his three Monist articles on "Space and Geometry" (1901-3) Mach treated the color
manifold according to the Riemannian view, and in a way similar to his treatment of
tone. The system of color coordinates is achieved by pairing every two opposing color
processes, {White, Black} {Red, Green}, {Blue, Yellow}, along its own independent
dimension. 32

Comparing sensations of space to sensations of color we discover that to the continuous series
"above and below," "right and left," "near and far" correspond the three sensational series of mixed
colors, black-white, red-green, blue-yellow. The system of sensed (seen) places is a triple
continuous manifold like the system of color sensations. 33

Along each axis, we can judge where a hue is from a fixed standard hue by successively
constructing all hues in between according to a rule, such as (H)+(D) = 1. These are
once again the alternating transitions between two modes, as mentioned by Riemann.
Because there are six qualities in all, a point can alter its hue in each of three different
dimensions simultaneously. The actual psychological details suggest a different story,
but the differences are unimportant for an abstract conceptual treatment of space such as
Mach was pursuing here.

In a rather illuminating section called "Another View of Riemann's Manifold,"

Mach approached the problem conceptually by represented all spaces (sensory and
physical) as combinations of elemental qualities:

Let u,,u2,UJ,Il4, ... ,Un+I be any elements whatsoever (sensational qualities, substances, etc). If we
conceive these elements mingled in all of their possible relations, then each single composite will
be represented by the expression

where the coefficients a satisfy the equation

at,az,UJ,U4, ... ,Un+l = 1.


In as much, therefore, as n of these coefficients a may be selected at pleasure, the totality of the
composites of the n+ 1 elements will represent an n-fold continuous manifold. (If the six
fundamental color-sensations were totally independent of one another, the system of color
sensations would represent a five-fold manifold. Since they are contrasted in pairs, the system
corresponds to a three-fold manifold.) As coordinates of a point of this manifold we may regard
expressions of the form:

The coordinate expression (ada1) is a function relating the fixed point (a1) to a
point to be located (a.n) according to some rule for constructing all of the coordinates in
between. The coordinate value will thus be a result of the fixed point chosen as origin
and the outcome of the procedure changing from the origin to a place. There are no
metric properties to a coordinate manifold, just manifold extension and a rule for
constructing every position from every other. The sensation qualities (the non-bolded
alpha letters) are the Riemannian modes, a usage faithful to Riemann's own examples of
tones and colors. For example, a possible interpretation for:

is (t) units in the green direction, (a2) units blue, (3) units white, (4) units red, (as)
units yellow, (II(;) units black. Assuming only two opposing colors, the expression
reduces to:

If one quality in one direction is picked the other is determined, hence the manifold
is one-fold extended.35 The coefficients of this two-color manifold (the bold alpha
letters) describe the qualitative mixture of one color process (in normal alpha letters)
and its opposing color process, adding, it says, to 1. If the mixture contains a1 of a~, it
contains (1-t) or 2 of a2.


One of the groundbreaking features of Helmholtz's and Riemann's papers on geometry

at the time was their effort to eliminate spatial intuition and to take a thoroughly
analytical approach to geometric problems.36 Differential and integral calculus had
arisen with the help of geometric comparisons (the derivative as the slope of a tangent
to a curve, the integral as the area beneath), but by the nineteenth century, geometric
notions such as curvature could in turn be defined using functions and their analytical
properties. These analytical breakthroughs had an effect on approaching other
seemingly given properties of space from an abstract point of view. For example, if we
cannot appeal a priori to the fact that space has three dimensions, we must consider
instead how many fixed points and distances can be picked independently without
determining the rest. As Helmholtz pointed out, intuition had to be questioned and not
assumed naively:

The analytical treatment of the question, how space is differentiated from other measurable
[abmessbar] multiply extended continuous manifolds, is to be recommended... for in just these
circumstances visual intuition [Anschau/ichkeit] is lacking, and is therefore not subject to illusions
that are so difficult to avoid because of the special limitation of our intuitions. 37

Helmholtz had also expressed caution that psychological representations of space

not be carried over into physics or geometry. Rather, he wanted spatial concepts to arise
out of underlying physical laws. Thus, Mach's argument that sensory concepts were not
to be extended uncritically to physical occurrences outside the range of the senses, like
little upside-down images on the retina, certainly had a basis in Helmholtz's warning
about "illusions that are difficult to avoid."

It also seems to me that Helmholtz's paper "Uber die tatsachlichen Grundlagen der
Geometrie" may have influenced one of Mach's breakthroughs during the struggle of
the 1860s, namely the replacement of the notion of material objects with invariant
functions. 38 Helmholtz had defined a rigid body as a system of simultaneous equations
holding between each pair of points in it. 39 Taking all possible combinations in hand,
there will be m(m-1 )/2 simultaneous equations for all pairs. When the rigid body is
rotated or slid, the coordinates of one of its points (X.Y,Z) change to new coordinates (x',
y', z), but the equations are recognized as the same pattern after a transformation, and
they are what is permanent about the body. Helmholtz is thus pointing out a symmetry,
or the preservation of a certain property of rigid form through a set of coordinate
transformations, i.e., continuous rigid motions.

Likewise, Mach's physical and sensory objects were the invariants of a physical or

sensory manifold composed of elements. 4 Functional connections could survive the
changing of the underlying element qualities, just as the rigid body preserves its shape
and integrity on a transformation of its coordinates. Again, sense psychology provided
Mach with an example. In his Mach Band papers, he had observed that masses of light
and spatial properties were connected with one another in the perception of an object. 41
Changes in the spatial properties, such as perspective reversal, tended to occasion
concomitant changes in the properties of the lighting so that the overall perception of a
solid object remained, just as if the object were a function of light and space sensations
that remained invariant as its component elements changed.42

If Helmholtz's spatial coordinates are replaced by Mach's qualitative modes of

determining points in a manifold, or his elements, the Helmholtzian view of a body as a
set of equations practically becomes Mach's view of matter as a function. 43 The main
difference between them concerned the properties used to set up a manifold extension
and a system of coordinates. Helmholtz appeared to use the property of distance for both
tasks: employing distances between one coordinate position and another in a rigid
body, and as the invariant property conserved in the manifold for rigid bodies that did
not deform on coordinate transformations such as displacement and rotation.

This indeed is something of a petitio principii of the sort Russell accused Riemann of
committing,44 unless Helmholtz really meant by the principle of physical "rigidity"
and distances between points no more than a convenient geometrical stand-in for
some physical property such as the strength of molecular or chemical forces.

Going back to 1872, Mach was determined to figure out some way to avoid using
spatial properties as coordinates for physical manifolds, as he had done for the tone
space. As we have seen, in his "Symmetry" paper Mach even anticipated a replacement
of "Galilean science" along these lines of replacing spatial variables by other numerical
methods. Now, ordinarily, to set up a coordinate system in 3-space, each individual x
must be specified by giving three arbitrary fixed points, (a,b,c) and the three distances
from the fixed points to x.45 For example, if I want to locate any point in my room, I
have to give its distances from a fixed point on the floor, and two on the walls. If the
space has n dimensions, we may pick n distances arbitrarily with which the n+ lth is
thereby determined. When a rigid body is given as a set of simultaneous equations, if
some of the coordinates of points on the body are given, others will be determined. The
determining factors are the dimensions of the space (n) and the number of points on the
body. In the Conservation of Energy46 Mach worked out a table of the analytical
relationships between pairs of points in a rigid body, such as a molecule, and the
number of fixed points and distances needed to solve all of the equations in a manifold
of an arbitrary number of dimensions. For example, if we have a molecule of five points
in 3-space we only need to know nine distances for the tenth to be determined. Now, as
we have already seen, Mach wished to interpret distances between the points of a
molecule as the qualitatively different heats of combination of each pair of chemical
elements, which he said "gives us a clearer idea of the stability and manner of
combination than any pictorial representation. "47 The replacement he suggested sounded
reminiscent of the way color hues are arranged:

Perhaps, too, a more rational manner of writing chemical combinations can be foilllded on this.
We would write the components in a circle, draw a line from each to each, and write on the latter the
respective heat of combination. Perhaps the reason why, hitherto, people have not succeeded in
establishing a satisfactory theory of electricity is because they wished to explain electrical
phenomena by means of molecular events in a space of three dimensions. 48

The variety of different heats gives the number of equations that hold between pairs
of atoms in the molecule. A physical body thus became a set of simultaneous equations,
but defmed over qualitative coordinates. The differences between these coordinates
could rather be like the qualitative differences between hues of colors, or degrees of


As Mach's Nachlass amply attests, he continued to press forward in the attempt to

construe space qualitatively. In his notebooks, Mach worked through doubts about
Riemann's "modes" and asked how a solid extension could ever really be composed of
an alternating mixture of qualities:

How does it happen that the diagonal is longer than the two sides of the square, if the length is a
difference of properties of the elements of the manifold: m=aA+~B. The third side applied
perpendicularly must be the sum of both of the others. Is the distance the sum of differences?
Homogeneous function fi dx2+di+di'] w. But dx dy dz is itself distance.49

Mixture of several elements. A member corresponds to a point of a space of several dimensions.

But where is the volume? Where is the derived distance? 50

Coordinates already presuppose distance. Distance. Straightness. Straightness rigid bodies. H in

the plane. 51

By 1907, Mach came to regard extension as a Wechselbeziehung (a relation of two

alternating terms) of two qualities situated in independent directions, along the "two
legs" of the right triangle:

The hypotenuse through the exchanging relationship of two qualities. [He adjoins a picture of a right
triangle]Inverse quadratic with the distance. Foundation of distance measurement 52

Riemann had written that the invariant length ds could be expressed as a function
of both legs of an infinitesimal right triangle dx and dy. The question that disturbed
Mach was: if this length ds is the first extended quantity, what are dx and dy? Riemann's
use of the Pythagorean theorem for infinitely small quantities was confusing when taken
as a definition of raw "extendedness" since the legs that he assigns the "lengths" dx and
dy are still extended. True, they are limits of extended quantities, but they are never
allowed actually to be of zero length, otherwise the limiting expressions lose their
meaning and give expressions such as a/0 or 0/0. Mach, on the other hand, was
concerned with representing the differentials, dx and dy, as qualities or perhaps

It appears, then, that Mach imagined the "hypotenuse" of the qualitative right
triangle to be produced by an exchange of the two modes, and by constructing elements
of a more logically complex, extended magnitude. These conceptions would indeed
come to play a role in Mach's chemical manifold. Hence it is proper to locate him in the
Herbartian-Riemannian project of analyzing extension into simpler qualitative modes.
The basic kernel of these attempts is to see extension as an outcome of an associative
and dissociative transition "Ubergang" between concepts of direction or quality modes.


Now, the "modes of detennination" are supposed to combine in a proportion and yet
always be mutually opposed, a seeming contradiction that raised an eyebrow with more
than one of Mach's critics. The psychologist Carl Stumpf, in his 1886 review of the
Beitriige zur Analyse der Empfindungen, 53 complained that if a tone is made up of
mutually opposing processes of Hell and Dump/, related like + and -, why don't they
cancel each other out, leaving a residue of only one of them? In a way this is what they
do, since the one quality recedes to the degree that the other is present according to
Mach's equation (H+D= 1).

Franz Brentano, Stumpfs teacher and an opponent of Riemann, had a similar

objection to the Mach-Hering analysis of colors; if they are opposed, and if a bluish
yellow or a reddish green are supposedly impossible, how could mixtures subsist? 54
Mach could have backed off to a Herbartian conception of force qualities (since
opposing forces cancel one another or produce a resultant without completely
annihilating each other's reality), but instead he called attention to chemical processes
in organic tissue, saying that black and white sensations can both be present at the same
time in gray while reds and greens and blues and yellows each take from one another's
magnitudes and leave only the residues. It appears, then, that Mach had given up on
oppositions for white and black, but really, he says, they are still opposed, reversed
processes. They just reverse along different paths through the organism than the paths
along which they go forward, while the other color oppositions reverse and go forward
along the same paths and hence cancel one another.

On another front, Mach was criticized for his "metric" conception of space by Franz
Brentano and Herbert Buzello, 55 who pointed out against him, and Riemann too,
that eliminations of space and time in favor of systems of measurement still did not
eliminate the intuition of space, since it was still presupposed that the standard meter
stick was capable of moving about in space, and that the chronometer is capable of
measuring intervals of time over different times. But these critics confounded the
invariant with the properties used to construct the coordinates. If those coordinates are
qualitatively detennined, or detennined by certain physical properties, no such petitio
arises. They had in mind a construction like Helmholtz's, in which the property of
distance plays both parts but neither Mach nor Riemann was guilty of that.


As we have seen in broad outline, the motivation for Mach's early theory of
"psychophysical monadology" was to establish the new psychology of the 1860s on a
rigorous basis by bridging the ontological gap between psychology and physics, a very
ambitious subject for a young scientist who was just establishing a solid reputation.

Mach began these dense lectures by expressing his views on how psychology could
become a science, and how the sciences of physics, psychology, and physiology could
all be unified. 12 He claimed that the psychology of his time was divided between the
phenomenological school of Herbart and that of the physiological psychologists
(Fechner, Wundt, and Helmholtz). It seemed to Mach that Herbart had concentrated on
the phenomenal relations between ideas to the exclusion of their physiological basis,
while Fechner and Wundt pinned down a physiological correlate for each individual
sensation, letting dangle the remaining phenomenal relations of sensations to one
another, which were not exhausted by the physiological account.

The newest physiological researches, namely those of Fechner, Helmholtz, Wundt and others have
now already penetrated so deep that they leave off just where the Hemartian psychology picks up.
This science does not just stand there isolated, for the physiologist its study is not in vain. 3

Mach suggested that introspective psychology and physiology work in a "pincer

movement." Phenomenological analysis would isolate the elements of perception, and
physiology would locate the corresponding structures in the body. Such a parallelism is
already at work in his article "Zur Theorie des Gehororgans" (1863), where Mach
determines the phenomenal tone sensations as fixed "places" in a one-dimensional
manifold and then works outward to match them to a manifold structure in the ear. By
1865 Mach articulated his principle of psychophysical parallelism:

Every psychical event corresponds to a physical event and vice versa. Equal psychical processes
correspond to equal physical processes, unequal to unequal ones. When a psychical process is
analyzed in a purely psychological way, into a number of qualities a,b,c, then there corresponds to
them just as great a number of physical processes a, ~. y. To all of the details of psychological
events correspond details of the physical events.4

This was his working principle of psychophysical research, to which he later added
principles of continuity and sufficient determination in the Analysis ofSensations to rule
out chance correlations between sensations and physiological structures that seemed to
match but were really unrelated. 5

Mach was aware, of course, that so long as sensations and their physiological
counterparts remained heterogeneous (ungleichartige) magnitudes, no real
psychophysical law or function could be said to exist. According to a principle
attributed to the physicist Fourier, all formally adequate natural laws must exhibit
homogeneous dimensions on both sides.67 As a physicist Mach knew that if one side of
an equation has dimensions, fo example, of mff2 so must the other. 8 This ensures that
both sides are scaled to the same units, can be added, canceled, multiplied, and so forth.
A psychophysical law thus assumes a fundamental likeness in kind between the
quantities to be compared, before they can even be correlated.

A theory of homogeneous dimensions for physics and psychology is a vital concern

of the "Vortrage," since without it psychology could hardly stand next to physics as a
science. Even while Mach defended Fechner's use of heterogeneous dimensions of
sensation and stimulus, he said that psychophysics may have to do its own dimensional
analysis, and not necessarily with the space, time, and mass units of physical science:

Of course we measure stimulus and sensation by different units, stimulus by stimulus, sensation
by sensation, for both are heterogeneous magnitudes. This is no more swprising than that physics
measures time and space by different means. And more! We must ourselves lay down different
measures for different sensations, different for light, different for sound, different for vision.
Physics currently measures everything with three buckets: mass, time, and space. We haven't
gotten so far in psychophysics, although not all prospects are lacking for an approach to a similar
goal. 9

The idea of basing the unified science of physics and psychology on a new set of
dimensions seems to point to the need to find a common element which both physical
science and psychology share. Herbart had already proposed intensive dimensions of
quality and force, which Mach likened to physics and believed represented the essential
facts of psychology, albeit schematically. In a lecture of 1872, Mach stated what must
have long been his view, that Herbart's elements were such psychophysical magnitudes:
"Herbart assumed certain fundamental phenomena, and we find these upheld by all of
his reals, thus both by the soul as well as by the rest of the monads which compose so
called lifeless matter. 10"

Herbart also had the view that the same qualities were both apprehended in
psychology and made up matter in itself, although in different formal combinations. 11
Hence the matter of sensation was a neutral ingredient of both the physical and the
mental worlds, waiting to be combined into different forms or variations which then set
apart a psychological investigation from a physical one. That is quite close to Mach's
"neutral monist" view that sensations under their mental variations are the proper
subject matter of psychology and, under their physical variations, the elements of
physical objects. 12


Mach now turns to the issue of measurement of the sensations. He originally thought
that Fechner's 1860 psychophysical1aw could succeed where Herbart had failed to
provide a direct measurement of sensation-units. Although Mach soon changed his
mind after a closer analysis, his work on the Fechner law does reveal the philosophical
conditions he thought psychophysical laws had to meet, particularly the homogeneity
requirement. The Fechner-Weber law is written:

y =a log(xlb) 13

which means that the sensation y is proportional to the logarithm of x, the stimulus. The
formula involves several philosophical assumptions that would drive Mach away from

Weber's law, from which the above formula is derived, is empirical and describes
the familiar fact that our sensitivity to a stimulus-whether or not we notice a small
increase or decrease-depends on the total amount of the stimulus present. For
example, we notice when a pound is added to a pound, but not when it is added to one
hundred pounds. 14 Each sense organ has its own peculiar constant of sensitivity (a),
which expresses the ratio of(~), the stimulus increase needed to produce a "just-
noticeable difference" G.n.d.) to (x) the total stimulus present:


A very sensitive organ like the eye has a small (dimensionless) constant of about
1/100, whereas muscular sensations in the arm are grosser and have a larger constant of
about 1117. Intuitively the law says that the more stimulus is present, the greater the
amount of stimulus necessary to produce a just noticeable difference.

What Fechner added was the idea that this ratio can be interpreted as a direct
measure of the sensation's intensity (y). For example, when the sensation intensity
increases by one j .n.d., the stimulus needed increases by the Weber ratio for that sense
organ. And as the amount present increases, the stimulus needed to get to the next j .n.d.
increases in geometric ratio.

The correlation hangs on the idea that "adding up" a number of same-sized units of
j.n.d. arithmetically yields up the total sensation intensity. But we don't know that the
j.n.d.s are all the same sized unit. It is often argued that a scaling mistake is made in
arguing from Weber's law to Fechner's formula. William James 15 (who took a very dim
view of Fechner's work) argued that the j.n.d. was not even a sensation-unit but a
complex judgment that a sensation has occurred, and it has since been shown thatj.n.d.s

are indeed sensitive to "higher" processes of attentiveness, distraction, and memory,

among other psychological factors.

In 1865 Mach officially retracted his "Vortrage" defense of Fechner's law. In the
interim, the German physiologist Fick had demonstrated that a sensation's strength was
directly proportional to the electrical stimulus induced in the nerve. 16 Mach continued
to rate Weber's law as genuine but limited, and he thought he could deduce it from his
own law of neural inhibition. If nerve energies inhibit one another, only activity that
originally stands out above the average of surrounding activity will remain; the rest will
be leveled off by inhibition. Thus, the eye senses only ratios of contrast above the
average intensities. This indicates the sense organ's evolutionary adaptation to react to
ratios of intensity rather than the absolute strengths:

The remark, for example, that a visible object under varying intensity or illumination can be
recognized as the same only when the sensation excited depends on the ratio of the illumination-
intensities of object and surroundings, makes intelligible a whole train of organic properties of the
eye. In this way we understand how the organism, in the interest of its survival, was obliged to
adjust itself to feel the ratios of light-intensity. The so-called law of Weber, or the fundamental
psychophysical fomrula of Fechner, thus appears not as something fundamental, but the explicable
result of organic adjustments. The belief in the universal validity of the law is naturally herewith
relinquished. I have given the arguments on this point in various papers. (SW 52 (1865);
Vierteljahresschrift 1868; SW 57 ( 1868)). 17

In his 1872 lectures, "Uber einige Hauptfragen der Physik," Mach also challenged
Fechner's view that sensations were "sums" of unit j.n.d.s. Here he says that different
intensities of sensation are actually different qualities, and, as heterogeneous
magnitudes, cannot be added together:

The sensations are different each according to its locality and therefore heterogeneous.
Heterogeneous [magnitudes] are not to be added together... when the sensation goes over from gray
to white, we do not sense more strongly, but rather completely differently. There are not
homogeneous elements that compose a stronger here, rather completely heterogeneous ones. These
cannot be added.

It will not do to analyze an intensive sensation into such sensations that can be numbered. Thus
Fechner's way of measuring intensity has no sense psychologically.
Measure and number too is only an exact description, which is inapplicable to the psychical. Here
things must be descnbed in an essentially different way. Thus the measure of sensations appears to
me as something fairly illusory. 18

Mach said later in the Analysis that Fechner's technique is only a conventional
proportion between sensations on one side and stimuli on the other:

Strictly speaking I consider the expression "proportionality" also to be inappropriate, since there
can be no question of an actual measurement of the sensations; all that can be done is to
characterize them by numerical means. Compare what I have said about states of temperature
(Prinzipien der Wiirmelehre p. 56). 19

It seems to me Mach was applying Fourier's principle to Fechner's law. While

stimuli admit of ordinal scaling, sums and differences, sensations do not. A seemingly
more intense sensation does not include every smaller, as every stimulus indeed does.
Moreover, Mach calls attention to the arbitrariness of setting equal units of j .n.d. against
the log of equal stimulus units. In his Theory of Heat he asserts that the proportionality
between temperature and changes of volume (in a thermometer, for example) is of the
same conventional nature.

Concomitantly with the thermal sensation which a body provokes in us, other properties of the
body also undergo alteration--, as for example, its electric resistance, its dielectric constant, its
thermoelectric motive force, its index of refraction, etc. And not only might these properties be
employed as indices of the thermal state, but they actually have found such employment. In the
preferment of volume therefore as a test of states of heat, there is involved ... a certain caprice; and
in the general adoption of this choice a convention. 20

Mach argues further that should some collection of, say, three mutually
independent properties be chosen to represent the thermal state of a body, the state can
be represented as a threefold manifoldness and not a simple manifold of numbers:

The properties of the system of symbols we employ are not decisive of the properties of the states
symbolized. If we were to take, for example, as our criterion of the state of a body K the pull
exerted by K on an iron ball suspended from a balance, these pulls, the aggregate of which as
symbols constitute a simple manifold, could be determined indifferently by the electric, magnetic
and gravitational properties of K, and would be the symbolic correspondent consequently of a
threefold manifold. Inquiry must determine in each case whether the symbolic system chosen is
the appropriate one. 21

As we saw, when Mach actually analyzed sensations of color and tone he held to
the oppositional theory of Herbart rather than the j.n.d. of Fechner. An intermediary
color between red and green is not obtained by "adding up" units of red, but is an
opposition between elements of green and red, where green is not so much the absence
of red but a different opposing sensation process. 22 The mixture of proportions of the
red process to the green process gives a "numerical" catalog of these intermediary
states, all of which are different in quality, but none greater than (or inclusive of)
another. In Mach's theory of tone, the same framework is proposed. He observes that
every tone is a qualitatively different sensation, and yet they do form a series in which
we know which are higher and which lower. Mach was not against the classification of
sensations per se, but only insisted that the qualitative dimensional analysis precede the
quantitative manipulation of symbols. And, to be sure, the dimensions of Machian
sensations looked much more like Herbart's than like Fechner's. This disagreement and
others were soon to come to a head between the two men.


At Graz, Mach was in correspondence and personal contact with Fechner. Mach
had also prepared a first draft of the Analysis of Sensations, which he wanted to
dedicate to Fechner. However, the older man very politely refused the dedication of the

whole work on "psychophysical relations," writing Mach on 18 Aprill864, graciously,

that he would rather be recognized instead as the impetus that sent Mach on his own
"self-standing way." He could not endorse this way personally:

I can only take as an honor your intention to dedicate to me your soon to appear book on
psychophysical relations and I sincerely take pleasure over this sign of recognition of beginnings,
in which respect I, too, have moved on. You might perl!aps think it over whether the self-standing
path of inquiry you are travelling might not take you further from these beginnings than you might
find convenient to acknowledge later, at least to the degree you would express the
acknowledgement. In the meantime it can only be a stimnlus that you need to acknowledge, and
from this point of view I would have the least reservations of accepting a dedication, which you
have so kindly offered?3

As Michael Heidelberger relates, Mach confided at seventy-three that Fechner's

rejection still pursued him "in his dreams." 24 What was this disagreement about that
Fechner's courtly refusal became so painful to Mach? The evidence points to Mach's
Herbartianism and Mach's new philosophical position, perhaps on psychophysical
relations, which Fechner elsewhere calls an "abweichender philosophischer
Standpunkt. "25 But was this standpoint the later "neutral monism" of the Analysis or
rather the "psychophysical monadology" of the "Vortriige"? Joachim Thiele speculated
that the early draft of the Analysis would have resembled the final metaphysical section
of the "Vortriige" where Mach discusses Fechner's view and then his own theory.Z 6
Thus that section may hold some clues.


Fechner believed that mental and physical were two aspects of a single underlying
substance that had both inner "mental" attributes and outer "physical" attributes. We
have an "interior" view of our own thoughts, feelings, and intentions, and others
observing the brain from the outside have an "exterior" view of the physiochemical
processes, although even the external aspects of the observed brain are internal to the
observers. 27 According to Fechner, we may break up all of these different observations,
both interior and exterior types, and pronounce them all phenomenal modes of
appearance functionally related to one another and comprising a single underlying

The reason why mind and body seem mutually exclusive, Fechner explained, was
that substance can appear to itself in its inner mental aspect or appear to others in its
physical aspect, but never both at once. The subject has his own idea of green, and the
observer has his own idea of the brain processes he observes. These cannot be
experienced simultaneously, just as one cannot experience both aspects of "convexity"
and "concavity" by standing inside a circle and outside it at the same time:
At the moment when we are conscious of our feelings and thoughts, we are unable to perceive the
activity of the brain that is associated with them and with which they are in tum associated--the
material side is then hidden by the mental. Similarly although we are able to examine the bodies of
other people, animals and the whole of nature directly in anatomical, physiological, physical and
chemical terms, we are not able to know anything directly about the minds that belong to the
former ... for the spiritual side is here hidden by the material...And what can be the reason for this
singular condition in which body and mind can be observed each for itself but never together, in
spite of the fact that they belong to each other? Usually we can best observe things which belong
together when they can occur together. The inviolability of this aspect of the relationship between
the mental and material worlds makes us suspect that it is rooted in their basic natures.28

Heidelberger has argued very convincingly that Fechner did not believe in a further
underlying Spinozist tertium quid over and above the functional relation of the aspects,
as was often supposed?9 Rather, an object for Fechner was the result of linking together
all of its phenomenal aspects (interior and exterior) by functions. Apart from the
functions there is no further substance. Hence this single object is half composed of
mental aspects, the other half of physical aspects, just as a penny is a functional unity of
a circular image, an ellipse, a rectangle, and so forth.

The "interior aspect" of all natural processes also explains why all aspects are
mutually exclusive, even when they are inextricably united in substances.3 Fechner's
sophistical cleverness is admirable in showing this: invincible duality proves identity,
since nothing would be so exclusive in its aspects unless they were interior and exterior
aspects of the same thing; and identity demands duality because nothing can be both
itself and observe itself at the same time from the outside.

As the title of Heidelberger's book aptly illustrates, Fechner was a panpsychist, and
something of a spiritualist, attributing an "inner side" to plants and inanimate matter-
although presumably those inner sides were of a peculiar nature radically different from
human experience. Mach seems to have adopted many of his own panpsychist views of
this time directly from Fechner.

Importantly, the inner side is only accessible for sich, by being the thing itself, and
all natural things (not just minds) have this privileged access to themselves that is not
granted to anything else. Hence it is impossible that this interior aspect should be
reflected in the brain's physical relations with the outside world. Mach attacked the
"simplistic" theory that identifies a particular green color with a particularly correlated
brain process:

It cannot be said that thoughts are physical processes in the brain. When we investigate the
processes in the brain more precisely, we fmd molecular motions but no thoughts. The electric
currents in the brain are rather something essentially different from the idea of a green color,
which perhaps only the investigated subject has. It is more likely our ideas that we find in
investigating the brain and not those of the investigated brain. 31

Mach also declared in the "Vort:rage" that sensations were the interior look of the
nerve processes themselves, while their chemical electrolytes and electrical pulses
represented external relations with measuring instruments or with other physical
materials. 32 Mach hoped until the end of his career to account for differences in
sensation quality by finding different chemical electrolytes carrying the electrical
impulse, but it was just not so. The sensory nerves do not seem to differ enough
chemically to account for differences in the quality of sensations, and the electrolytes
are sodium and potassium ions all over, not five types for five senses. There is not even
much of a difference between the nerve cells of different species. However, the facts do
cut the other way, too: quality does depend closely on the specific nerve activated, and
not, for example, on the quantitative intensity or frequency of the pulse it carries, which
E.D. Adrian first confirmed in 1928? 3 It is still not known what it is about the specific
locality in the brain, or the connections to local patterns of cells, that induces specific
sensation qualities, but it is thought that the same abstract pattern of firings in different
cells would occasion a different set of sensations. Even if one could prove otherwise,
the question would still remain why certain patterns occasion certain sensations and not
others. 34


Both Mach's and Fechner's notion of Innerlichkeit derived from the physiological
doctrine of specific energies proposed by Sir Charles Bell and Johannes Muller. 35
According to Edwin Boring, the doctrine states that we are aware not of external
objects, but of the states of our nerves themselves. An external object is not even
needed to cause a sensation, for we can cause colors and sounds by interfering with the
nerves themselves: pressing the optic nerve produces sensations of light, pressing the
auditory nerve sensations of sound, and so forth. 36 But there is also a more substantive
philosophical thesis to be found in Muller's formulation, which is that sensation
qualities are what physical nerve energies really look like on the inside? 7 As Milller
wrote in his work on vision: "Light and color are not outer things apprehended by the
senses, rather always apprehensions of the sensory energies themselves. "38 And, in his
famous Handbuch: "We are required to attribute to every sensory nerve certain energies
in the sense of Aristotle, which are its vital qualities, as contraction is the vital property
of muscles. "39

Sensations, in short, give us a direct look at the nature of a certain small class of
physical processes in us, suggesting of course the possibility that if we had an interior
view of other physical processes, these too would reveal their own interior aspects, as
well as the external aspects under which we know them.40 Mach did not feel that his
contemporaries had paid enough attention to Milller and wrote in a notebook entry of
1896, for probable inclusion in a new edition of the Analysis of Sensations,41 that he
considered himself and Ewald Hering the only two genuine heirs of the Berlin

I believe now that the thought expressed here, according to which just so many different kinds of
physiochemical nerve processes are suspected as there are sensation qualities to be distinguished,
likewise has heuristic value, and that this thought can hope for some support from the
physiological-chemical side. This thought which is merely a logical consequence of a monistic
conception of Muller's principle of specific energies, corresponds little indeed to modern ideas.
Following the lead of a great authority, one has gotten used to pushing away the explanation of
different sensation qualities into as yet unknown regions, and regarding all nerve processes as
completely homogeneous in quality and only different in quantity. Already in 1863 (Lectures on
Psychophysics) I had occasion to show how little this conception can lead to a deeper
understanding of our sensations, and how little justified it is to hold that all nerve currents are
qualitatively homogeneous physical processes. If I am not mistaken, only Hering still upholds
Muller's original teaching 42

The "great authority" is Hermann von Helmholtz, 43 who had extended Muller's
specific energies principle, without the Innerlichkeit of course, by localizing particular
sensations not only to the same sensory nerve but even to different fibers within the
nerve. 44 Helmholtz associated a different sensation of tone to each of the 2,500 different
fibers in the auditory nerve, or rather each row of them. Likewise, the three primary
colors (red, green and blue) of the Young-Helmholtz theory of color correspond to three
types of fibre postulated to exist within the optic nerve. Helmholtz broke with Muller,
however, by claiming that the different sensation qualities were reducible to quantitative
differences in the intensity and frequency of pulses carried by each nerve. The theory
came to be known as the "telegraph-wire" model, according to which the sensory nerves
were all of a kind and only conduits for electrical vis viva that varies in signal, or in the
number and type of connections between cells. The quantitative aspects of the message
moving through the nerve is what matters, and not the internal properties or locality of
the nerves themselves. Muller's theory seemed to Mach to dictate that where there was a
difference in the quality of sensation, there had to be some difference in the physical
energy of the nerve.

We fmd in the nerves only electrical currents, which can be thought of quite easily as being of
different intensity, but not of different quality. The nerves may accordingly be similar in their role
to telegraph wires, which release only certain signals corresponding to sensations ... That for
different sensations different nerves correspond is not to be disputed However, what if the
electrical processes in the nerves should be too simple to suffice as an explanation? Should it then
be necessary to push explanation back into unfamiliar regions? How then when after more precise
and thorough research of the whole brain, only electrical currents are found everywhere?

My humble opinion is this. The electrical investigations of the nerves are certainly of a very fine
sort. In certain respects however they are also very crude. An electrical current of a given intensity
tells us nothing, other than that a certain quantity of vis viva passes through a cross section of the
current in unit time. Which processes and molecular motions convey the vis viva, we don't know.
The most diverse processes could underlie the self-same intensity of current. Ludwig once spoke
to this point directly: when the chemical elements of a nerve are given us, we will know nothing
more of their constitution than we know of a telescope when someone tells us it consists of glass,
metal and lampblack. The electrical current, too, shows us merely the outside of nerve processes 45

Mach favored a much closer examination into the interior nature of physical processes,
even unto the subatomic level. The electrical side of these nerve processes is the side
that is accessible through physical interference with other phenomena and other

observers and is thus still too crude. Hence, Mach had come to believe that every
natural event, in addition to its quantitative external aspects, also possessed an internal
monadic aspect. (By "monadic," I mean a property that does not transfer to other
experimental points of view, like vis viva and nwnber of impulses, and cannot be
proven to exist independent of experiencing the property directly.) Monadic aspects or
properties could not translate into external relations, and could not be directly
represented by them. The external aspect, on the other hand, seems to have been a
stable quantitative and spatial pattern of interference that ranged broadly across natural
processes, giving them a uniform exterior and a uniform spatial platform for such
comparisons. Mach thought extension and motion belonged to the set of external
relations, including beats of electrical current moving through different media. Mach
declared his views clearly in a somewhat later lecture "Uber einige Hauptfragen der
Physik" (1870-71) where he stated that all matter possessed a monadic aspect similar to
that of sensations:

Sensation is a general property of matter, more general than motion. Let us seek to set down this
proposition clearly. An organism is a system of molecules. Electrical currents run into the interior
and come back again into the muscles. Everything is physically explicable. But not that the person
should have sensations. What we can investigate physically is always merely physical. We find no
sensation. And yet the human being senses. The material flows forward through and through him
The old depart. The new comes in. We therefore have the problem of finding something
fundamentally new in the whole that is not in the parts. We escape this difficulty when we
consider sensation as a general property of matter.
We have therefore two sides, an inner and an outer side. According to analogy we assume
these also in animals, though on weaker evidence. In plants still weaker, but the boundary blurs.
With respect to the inorganic again the boundary blurs. Therefore no doubt everywhere an inner
side. One can only be convinced of it from one's own soul. The rest is only conjectured.46

In case there was any doubt, Mach indicated that these world elements "sensations
in matter" were not just our sensations considered in their physical variations, but new
sensations not given to us but added in thought (hinzugedacht):

Why do we assume a soul outside of our own?... This assumption: to certain of our own, to add in
thought certain others, which make more comprehensible the triggering of our own...Animals and
plants. Still more of such sensations which we add in thought to make others comprehensible.
Finally we add them in thought everywhere.47

Furthermore these sensations of matter are not Dinge an sich, but stand in causal
relations to our own and are connected with them continuously, requiring no Kantian
break between the world of phenomena and nownena.

Now, as to the actual theory of psychophysical relations, how are the two orders of
physical magnitudes and sensations to come into contact with each another? Mach's
new physical theory of matter attributed an interior of pressures and qualities to atoms
similar to sensations.48 Such qualities were thus of infinitely tiny size, even beneath all
applicable measure of size apparently, and of extremely small energies, especially
compared with the rather large and messy energies produced in the human nervous

Mach claimed that human sensations were only attributable to atoms in large
groupings, and thus for him there could be no question of thinking a sensation of color
or sound into an atom. 49 Rather, what seems to exist there are extremely tiny, internal
elements, which perhaps in very complex cells or other physical arrangements might
superimpose to a whole sensation. I suspect that these internal qualities in the atoms
(similar to Leibniz's petites perceptions) were probably the part of Mach's
selbststandiger Weg that Fechner rejected.

As we now know, central sensation energies and atomic-scale energies, though

comparable in kind, are of vastly different orders of magnitude. Even if we were to
grant that sensations are the actual look of energy, perhaps the most fundamental
property of the physical world, Mach would still have to say something about how small
elements combine into much larger ones, on the interior side, in parallel with the
exterior agglomeration of small bits of matter into larger chunks studied by physics. But
this whole side of qualitative science is missing. Mach's principle of psychophysical
parallelism demanded that all psychical elements had their physical correlates that vary
in the same dimensions. Hence, it would seem that the tiny elemental sensations of
matter would have to combine in structures at least somewhat parallel with their
physical counterparts to form sensations, if such a principle were to be followed through
seriously. (Actually, I will show later that there is no such need for the parallel to be so
exact.) One such attempt to introduce such a compositional structure for the psychical
parallels of physical events was W.K. Clifford's theory of a descending series of mental
events, following the descending series of organisms into inorganic matter. 50

Whatever such compositional principles might be, the scales of magnitude

involved are so different that the interior appearance of a nerve energy probably is not
similar to the interior appearance of small physical energies. Rather, it is entirely
possible that Leibniz was right in considering human sensation merely a flooded,
confused apprehension of many true qualitative features of the world at once.

Mach was certainly wrong that different sensory nerves must contain different
electrolytes? But he was right about the localization of sensations to very specific
structures. Perhaps he could have thought up some kind of mechanism to explain the
quality of sensations by their localization. Such a mechanism of emergence was
suggested in his later analysis of tone sensations, where he postulated small
"supplementary colorings" to tone sensations, called Z. 52 These are indistinguishable in
single tones, but become conspicuous in combinations where they add up "just as the
contrasts of faintly colored, almost white lights become vivid when such lights are
brought together." 53 Mach was also aware, through his work on neural inhibition, that
the internal states of single cells could be affected by the complex environment of
neighboring cells. In his standard formula: no process in nature ever occurs in isolation
from others. He probably also held that the local environment of a particular nerve
process could conversely "drive its way inward" to alter the states of each individual in
the group.

In any case, the small, practically infinitesimal, world elements, which may seem
the most unrealistic feature of his "vain" monadology, were not among the aspects that
Mach changed in later work. The only features he did alter were the existence of the
monads or atoms, and his abiding tendency to think sensation elements spatially "into"
atoms or the physical cross-section of nerve processes. It thus seems that a different type
of causal relation or nexus of elements is responsible for the relations of tiny world
elements to macro sensation elements, perhaps a very complex compositional relation.
When Mach spoke again about psychophysical relations in the Analysis of Sensations,
these spatial metaphors of interiority had been purged, and all notions of relatedness,
such as relations between the "inside" and the "outside" of nerve processes were
replaced by the more abstract concept of function.


Mach's Elementenlehre of neutral elements and functions was complete on the physical
side by his "Raumvorstellungen" paper of 1866, and his first extension of that theory to
psychophysics and psychology was made in 1872 in the Conservation ofEnergy, where
his early thinking on the unity of science had developed into a program.

General Rtllllliiks
We very soon learn to distinguish our presentations [Vorstellungen] from our sensations
(perceptions). Now the problem of science can be split into three parts:

I. The detennination of the connexion of presentations. This is psychology.

2. The discovery of the laws of the connexion of sensations (perceptions). This is physics.
3. The clear establishment of the laws of the connexion of sensations and presentations. This is
psychophysics. 1

As we can see there is one kind of subject matter in all domains, the study of the
elements in their functional dependence on one another. The elements make up the
range of mental phenomena-colors, sounds, and the like-but they also make up the
objects of the world as little chunks of energy arrayed in physical functions and
governed by the principle that all variations of the elements of the world are set at zero,
as well as in any closed system. Fechnerian psychophysics falls into place as the
meeting ground between the elements of matter and the elements that make up mental
phenomena. Although Mach realized the complexity of such relations between energies
of such vastly unequal magnitudes, he felt the philosophical problem underlying
psychophysical relations was moot, as the types of the magnitudes concerned were no
longer heterogeneous. Hence, with this scaling requirement satisfied, it remained only
to find the functions relating their vastly different orders of magnitude, a difficult but
ultimately empirical task involving no further conceptual difficulties:

Now if we resolve the whole material world into elements which at the same time are also
elements of the psychical world and, as such, are commonly called sensations; if, further, we
regard it as the sole task of science to inquire into the connection and combination of these
elements, which are of the same nature in all departments, and into their mutual dependence on
one another; we may then reasonably expect to build a unified monistic structure upon this
conception, and thus to get rid of the distressing confusions of dualism. Indeed it is by regarding
matter as something absolutely stable and innnutable that we actually destroy the connection
between physics and psychology?

I think one encounters clear views in Mach's physical writings about the overall
relations among the various disciplines. One begins with a very large monistic set of
elements and variations and then obtains from this universal set the customary divisions
between physics, psychology, and psychophysics by economically specifying relations.

Here it is the monist Mach who comes to the fore, the one who begins with the
unlimited All and obtains parts of it by division into elements and domains. However, in
the Analysis of Sensations, a book addressed "primarily to biologists,"3 Mach's
conception of world elements, so prominent in his earlier physical work, receded into
the background. This gave the contrary impression that Mach did not begin with any
universal set but only with such elements that are the elements of one consciousness,
which he then pretends to expand only by hypothesis and analogy. Mach gave a clue in
his lecture Ober einige Hauptfragen der Physik why he had done so, where, after he
stated that sensation was a "general property of matter," he added: "Mir selbst, da ich es
ausgesprochen, iibel genommen." (I have borne ill-will for saying it). He adjoined later
that he did not want to excite the old Gespensterfurcht (fear of ghosts) among his
colleagues. Michael Heidelberger concludes that their objections to his psycho-
monadology had caused him to conceal his views: "It should probably be assumed that
the doctors before whom Mach held his lectures, were not exactly enthusiastic about
Fechner's world-soul fantasies and that Mach had to swallow a few insults for this. "4

Hence, having already experienced the unpopularity of his Fechnerian

speculations, Mach seems to have confined himself in the Analysis to what was strictly
observable, making a kind of allowance for unobservable elements only as an inference
by analogy from that data, to account for elements of other minds and the elements of
unobserved objects. Thus, vis-a-vis the materialist he could claim to be the one who was
really being scientific. In the process, Mach both laid hands on a powerful cudgel with
which to beat the materialists as well as the philosophical high ground, as he claimed to
be eliminating metaphysical constructions of unchangeable egos and Dinge an sich as
well as any ontological difference between the stuff that made up objects and that which
made up minds. Always in the back of his mind was the rampant materialism of the
period that dominated in Vienna. One can imagine Mach, with his own hylozoist
published utterances in tow, continually up against the view most strongly supported by
the medical community that sensations were nothing but confused arrangements of
atoms and that any philosophical problem of the origin of sensations, ideas or behavior
would be solved when "the paths of atoms in the brain" were known, after which
everything else would follow of itself. One of Mach's physicist colleagues in fact
proposed to him just such a view. 5

In the Analysis, bearing in mind the intellectual climate of the times, we must
imagine Mach under pressure to produce an antimaterialistic, rigorously scientific view
of the world, as well as one that could unify physics and psychology in a natural way.
This demanded, I think, two rather difficult conceptual steps: first, viewing the objects
of physics as sensation-like elements; and second, viewing mentalistic sensations as the
actual parts of physical objects extending fully from the external things into
consciousness, and not as mere representations or pictures. Mach's solution to this
double puzzle was quite ingenious and allowed him to hold both of these positions
simultaneously. If there was a problem with his view it was that he won his scorched-

earth battle with materialism a little too well, leaving many with the impression that he
wanted only a universe ofhwnan sensations and nothing else.


Mach's Analysis is in one way the biography of hwnan experience as it rises from a
primordial soup of elements and slowly delimits itself from the environment into its
own boundaries of an ego, with a strong command of its body, good association
between its different sensory modalities, and a relative permanence provided by
memory and time sensation. To some degree the organism is helped by the immense
native apparatus within its own body, called into action by experieBce but representing,
as it were, a built-in species memory, the accumulated evolved experiences of previous
generations of organisms.6

Although most organisms can rely on the existence of numerous reflexes at birth,
according to Mach, conscious experience begins as a flooded agglomeration of elements,
in some preconscious or unconscious state. It is very important in order to progress
beyond such a state that we possess an ability to remember our experiences
and compare them with one another. Otherwise, Mach says, we would be permanently
lost in primordial unawareness in which states occur and instantaneously vanish from the
mind without lingering there: "When a sensation is forgotten the moment after it has
vanished, the only possible result is a disconnected mosaic and series of psychic states
such as we have to suppose in the case of the lowest animals ... "7

Thus for Mach there has never been any such thing as mere awareness. Rather,
awareness, far from being the innate mark of the animal rationale, is a learned
activity--and one that is learned rather slowly. Awareness is mostly dependent upon
experiencing the same events again and remembering them to be the same. Hence, a
structure of repeating events and sensations of stable recurrent objects of the
environment are required for getting a foothold in the history of one's own experience
and attributing them to some stable ego complex who has them. This ego also changes,
in Mach's view, and is full of endlessly variable sensations and experiences; but by
isolating the recurrent feelings, and comparing present with remembered states, it too
comes to have a relative enduring solidity.

Now, with the aid of remembrances, and by trial and error, we learn to divide our
experiences into three classes. The first class, labeled A,B,C ...refers to our sensations of
physical objects with color and shape and sensed as if external to our bodily surfaces.
The second, labeled K,L,M ...refers to sensations we have of our own bodies (such as
local signs). Finally in the third class labeled a, ~, y... are the phenomena of after-
images, memory representations, imagination, and the like. Unlike Hume, Mach does
not give us the ability to tell the difference between these classes prima facie, nor does
he even consider certain experiences sense illusions and others veridical: "The
expression 'sense illusion' proves that we are not yet fully conscious ... that the senses
represent things neither wrongly nor correctly.

All that can be said of the sense organs is that, under different circumstances they
produce different sensations and perceptions. "8

Thus, for Mach every phenomenon has an equal reality taken as It Is. Even
Macbeth's delusional dagger of the mind represents the occurrence of something, even
if that manifestation proves to have closer functional ties with a fevered brain than with
the other objects in a room. 9 What reveals whether an item is a member of the class of
external objects or of the imagination for Mach is simply its variation with the other
elements. 10

Mach explained the difference between sensations and elements of the world by
pointing at their different variations. For example, quite often the A,B,C will vary with
the elements of our own bodies K,L,M. If the retina suffers a chemical change, the
colors of objects will also change. However, the A,B,C may also undergo their own
independent variations, for exantple when an object moves across the stationary retina.
When this latter case occurs, Mach calls the A,B,C physical objects or elements:

A color is a physical object as soon as we consider its dependence for instance upon its luminous
source, upon other colors, upon temperature, upon spaces, and so forth. When we consider,
however, its dependence upon the retina (the elements K,L,M), it is a psychological object, a
sensation. Not the subject matter but the direction of our investigation is different in the two
domams. . 11
The elements A,B,C ... are not only connected with one another, but also with K,L,M. To this
extent and to this extent only do we call A,B,C...sensations and regard A,B,C as belonging to the
ego .. .it is only in the connexion and relation in questions, only in their functional dependence,
that the elements are sensations. In another functional relation they are at the same time physical
objects. We only use the additional term "sensations" to describe the elements because most
people are more familiar with the elements in question as sensations (colors, sounds, pressures,
spaces, times). 12

The elements themselves are neither psychical nor physical but neutral, separated
only by the different types of functional variations they undergo. The variations are not
absolutely separate either. A thing may enter into physical variations while still being
subject to a whole range of psychological ones. The sensation of the yellow sun is
codetermined by physical events, but also by the condition of the retina. If the retina
were damaged, nothing would appear, and even the physical variations of shape,
extension, and radiant energy would be blocked to sight. Thus, instead of hard and fast
divisions between psychical and physical functions, it is more likely Mach thought the
two different orderings of elements were drawn from a larger class of natural variations,
neutral functions as well as neutral elements:

Our body, like every other, is part of the world of sense; the boundary line between the physical
and the psychical is solely practical and conventional. If, for the higher purposes of science, we
erase this dividing-line, and consider all connections as equivalent, new paths of investigation
cannot fail to be opened up. 13

In so far as they are elements and not in the specialized class of sensations, Mach
thought such qualities were actual parts of physical objects. Besides other neutral
monists like Russell and James, few commentators knew how literally he meant this to
be taken. For example, Paul Carns addressed the point in an 1893 article in the Monist,
in which he reported on a meeting with Mach earlier that year in which philosophical
matters were discussed:

When several months ago I met Professor Mach in Prague ... he assented to my speaking of
scientific terms as abstracts ... But when I proposed that the term "sensation" also was
according to my terminology an abstract term presenting one feature of reality only and
excluding other features, Professor Mach took exception to it, saying that he understands by
sensation reality itself. 14

Hans Kleinpeter also gave an apt description ofMach's new "realism":

Mach explained as the object of the naturalist's occupation "sense perceptions," a name which he
chose under compulsion in lieu of a more suitable one, and which he later replaced by "elements";
for the sense in which he wished to have the word taken deviated to some extent from the
customary usage. By the sound of the word we are inclined to put too much emphasis on the
perceiving subject. But it was far from Mach's intention to emphasize this connection from only
one point of view; on the contrary he saw at the time in sensations the material of the actual world.
In this his fundamental views are essentially different from those of idealistic philosophers,
Berkeley's among others. We may even call them realistic, but their realism is essentially different
from so-called philosophical realism. 15

I think the crucial point to realize is that Mach thought objects with their
boundaries were no more or less than functions, or variations of elements. An "object"
usually stands only for an immediate neighborhood of its effects on other things. In
truth, however, no system may be regarded as physically isolated as the elements of the
"whole world are ultimately interconnected with one another." 16 So as one expands the
area of variation, one can also expand the boundary of the object at pleasure, to make it
include more distant spatia temporal effects. The scope of these variations, for Mach,
seems to have been relatively arbitrary and dictated simply by the requirements of
forming a locally "closed system" of elements. One could therefore say (in a manner of
speaking more suited to a simpler theory of perception than Mach's) that the physical
object of perception reaches all the way from the outer thing into consciousness, if we
consider its "physical" variations to include all intermediate contact with the sense
organs and even the specific energies excited in the brain, with which the sensations
A,B,C are in turn physically connected. If I see a person waving his hand at me to
attract my attention, in some sense the waving hand is a foreign object different from
me that reaches into my mind and distracts my attention. When we draw a circle around
such a pattern of elements and declare it an object, we may draw it to include those
elements that extend all the way into consciousness. On a different variation we may
stress instead the connections of sensational qualities of the A,B,C to the K,L,M of the
body and to memory images. Hence the A,B,C, are common property to be annexed at
pleasure by either side. The immanent object is thus partly a mind-dependent sensation
(s) and partly a mind-independent world element (e):

s e

Sensations and Elements

But must every element be both a sensation and a world element, or are there such
elements that never end up linked to human consciousness, but make up objects in
themselves? On balance, the evidence I have cited in the Introduction points to the fact
that Mach did indeed believe that such elements made up the remainder of the physical
world beyond immediate observation, but in more or less remote causal interconnection
with our sensations. The true neutral monist picture that emerges is of a kind of "cosmic
soup" of world elements (including animal sensations and all of the physical qualities of
nature) causally linked together in functions. There is no stated restriction to consider
only the elements of the human ego. In fact, the ego complex itself is really nothing but
a complex of variations provisionally and economically segmented from the rest. Mach
said, for example, that the elements of egos existed together with the other elements in a
prior sense to our limiting them to individuals:

From the standpoint which I here take up for pUIJlOSCS of general orientation, I no more draw an
essential distinction between my sensations and the sensations of another person than I regard red
or green as belonging to an individual body. The same elements are connected at different points
of attachment, namely the egos. But these points of attachment are not anything constant. They
arise, they perish, and are incessantly being modified. But where there is no connexion at a given
moment, there is also no perceptible reciprocal influence. Whether it may or may not prove
possible to transfer someone else's sensation to me by means of nervous connexions, my view is
not affected one way or the other. 17

Mach thus speaks as if he has a master array of elements before him, from which he
is able to separate out the objects and the egos of various agents. He described this
broadened view of the sensational contents of minds as a 'functional presentation' to
distinguish it from a spatial representation:

When I speak of the sensations of another person, those sensations are, of course, not exhibited in my
optical or physical space; they are mentally added and I conceive them causally, not spatially,
attached to the brain observed or rather fimctionally presented.18

A spatial representation would have to include, as it were, two spaces, those of the
outside and the sensory manifolds of the ego, or an awkward inner space within an outer
space. However, if the elements are represented strictly by their ftmctional points of
attachment only, both spaces could be seen as arising from the same richer set of non
spatial ftmctional variations. Since Mach had already outlined this plan for his physics,
it was a small matter to extend the idea to psychology. In fact, both sides of the view
had developed together. The crucial step forward was taken when Mach extended
Fechner's doctrine of the inner side of nature and MUller's view of specific energies,
beyond their stated aims of making sensation the interior look of energies and physical
processes, as if the sensations were spatially lodged within nerves and their electrolytes,
or even inside their atoms as the young Mach, in his unripeness, had attempted. Instead,
Mach made his elements both the inner and the outer nature of the world and thus
eliminated the meaning of the distinction. I think the best way to represent egos, objects,
and their various interrelations on his mature view is by means of a ftmctional map,
where sensations, sensation elements, and world elements are all represented side by
side, where the closeness of one element to another indicates their causal proximity or
ability to affect one another's magnitudes. For example, a map of a person observing a
candle would include all of the physical elements of the candle, the elements of light
emitted, the interaction of the sensory surfaces with the light, the intermediary nerves,
their reciprocal actions, their terminal branches in the brain, and finally the
phenomenological sensations annexed causally alongside the brain tissue. Simplifying
some of the detail, and drawing boundaries around the relations of interest, determines
the objects to be considered, although these are by no means the only boundaries that
might have been drawn.

ee ss
ee ss
ee s s
ee s s .__......__ __,

From these heights, and tacitly considering the complete array of world-elements
and ftmctions as given to us all the way, Mach was able to attack many conventional
views of mind, body, and perception as "metaphysical errors," primary among them the
ego itself, as he famously declared "Das Ich ist unrettbar" (the ego cannot be saved).

Mach's analysis of the ego is one of the main philosophical tasks of the Analysis of
Sensations. For this he adopted what he called a "completely naive view," (prior to any
developed epistemology) refusing to acknowledge any but the experiences he regarded

as actually given to a human individual. He wrote excruciatingly about the very first
stages of differentiation of the body from the environment of external objects, showing
that we only gradually learn through specific experiences to regard the ABC as objects
external to our bodily surfaces KLM:

We see an object having a pointS. If we touch S, that is bring it into connection with our body, we
receive a prick. We can see S without feeling the prick but as soon as we feel the prick we find S
on the skin. The visible point, therefore, is a pennanent nucleus, to which the prick is annexed,
according to circwnstances, as something accidental. From the frequency of analogous
occurrences we ultimately accustom ourselves to regard all properties of bodies as "effects"
proceeding from pennanent nuclei and conveyed to the ego through the medium of the body;
which effects we call sensations. 19

Thus the view that our sensations are caused by objects external to our bodies is
one that we arrive at by means of a simpler, more primitive view in which both inner
and outer elements are regarded as essentially the same. Mach claimed that he wrote the
Analysis of Sensations from this naive standpoint, before the distinction between the
ego and world could be drawn. He wrote to Gabriele Rabel for example:

The kernel of my exposition lies in the elucidation of Fig I. of the Analysis [the famous "headless
body" picture20 E.B.]. We arrive at it by observing children and putting ourselves back into our
early childhoods, as we were just learning to differentiate our bodies from the environment. At that
point we know nothing of matter and soul, of physical and psychical, of object and subject,
stimulus and sensation. Everything still consists of homogeneous ingredients ABCD ... , which are
in themselves neither psychical nor physical, but rather indifferently neutral. They first become
physical or psychical through the particular type of dependency we have in view? 1

And in a letter to Friedrich Adler on the same subject:

The point of departure of my natural world view when I wrote the Analysis of Sensations was
more primitive. I placed myself artificially back in the position of a child, who has just begun to
separate his body from the environment. There for the first time results the two-fold type of
dependence of the elements on one another; not until then do these lead to the concept of a
subject, an object...22

Consciousness thus emerges at the end of a historical development in early life.

Mach expanded this philosophical conception into his own early form of
"developmental psychology," claiming that consciousness was in fact a humble, learned
ability to separate our bodies and psychical elements from the environment of the
physical A,B,Cs. What is meant is the environment of immanent objects, which may be
regarded either as sensations or physical objects indifferently. Until we can differentiate
self from object with the help of experience and memory we do not achieve the
restricted empirical awareness of the ego as its own kind of object and contents. Before
that what reigns is the pure experience of an undifferentiated totality of elements, which
cannot be called awareness but merely passive sensing. Only by limiting this originally
given totality to the immanent objects of the environment and an ego does an empirical
awareness become possible. Indeed the awareness of the self and the awareness of the
environing objects occur necessarily together.

The process is reminiscent of the differentiation of the local sign system of the retina
from the manifold of colors and shapes. By slow separation, the impression arises from a
fused mosaic of the objects of a fixed world sliding over the fixed sensory surfaces.

In Mach's view it is of course exactly these simple primitive inferences, so useful at

the beginning, that eventually lead to the worst philosophical excesses, the thing in itself
isolated from all perception for physical objects and the absolutely fixed noumenal ego
on its side. 23 Once these entities are set up as absolute, inalterable, and not even made up
of the same elements, then the unity of the nai've view of perception and its simple
variations are both lost. Mach finds as the common root of these problems of the ego and
the Ding an sich the need to posit objects as the bearers of sensations or elements, rather
than what seemed more sensible in his view: simply adopting the idea that elements are
functionally dependent on each other and thus for that reason tending to form solid,
functionally dependent complexes.

Bodies do not produce sensations, but complexes of elements (complexes of sensations) make up
bodies ... For us the world does not consist of mysterious entities, which by their ioteraction with
another mysterious entity, the ego, produce sensations, which alone are accessible. For us, colors,
sounds, spaces, times ... are provisionally the ultimate elements, whose given connection it is our
busioess to iovestigate.24

In Knowledge and Error Mach attacked the idea that sensations must necessarily
occur in a consciousness, relying again on the view that the elements are primary and not
the ego complexes in which they may occur:

A sensation will iodeed always occur io a complex, but that this latter should always be a complete
and wakeful human ego is doubtful; after all there is consciousness io dreams, hypnosis, ecstacy
and there is animal consciousness, all io degrees. Even a body, a lump of lead, the crudest item
known to us always belongs to a complex and so to the world; nothiog exists io isolation. 25

He also attacked the notion that sensations are necessarily conscious, holding
instead that consciousness consisted in a special relationship of memory and
reproduction among sensations:

Consciousness is not a special mental quality or class of qualities different from physical ones; nor
is it a special quality that would have to be added to physical ones io order to make the
unconscious conscious. Introspection as well as observation of other liviog things to which we
have to ascribe consciousness similar to our own shows that consciousness has its roots io
reproduction and association: their wealth, ease, speed, vivacity and order detennine its level.
Consciousness consists not io a special quality but io a special connection between qualities ... A
single sensation is neither conscious nor unconscious: it becomes conscious by beiog ranged
among the experiences of the present.26

For Mach the view of an ego as an object with its own special contents is an artifact
of our instinctual need for permanence, an urge that leads us to fixate any inquiry on
object after object and to ignore the elements that change and do not seem solid enough
to investigate.

The only reason why our sensations seem private to us and inaccessible to others is the
lack of causal relations between the sensations of one ego and the sensations of another,
"two ideas must come close enough to interact, rather like bodies in physics," Mach
said.Z7 So again our instincts, so valuable in getting to a state of practical life that makes
science possible, seem inevitably to lead us in the wrong direction when it comes to a
higher stage of inquiry. And it seems that scientists are among the greatest offenders
when they fail to recognize the influence of these instinctual views on their own
thought. Hence, the need for Mach's special, anti-metaphysical form of critique in both
physical science and psychology.

Just as there are various patterns of elements, which may be chosen as objects, so
too there are various graded senses of the ego for Mach. "The ego," he says28 , "is not
sharply marked off, its limits are very indefinite and arbitrarily displaceable." He spoke
in particular of an ego in the "restricted sense" and an ego in the "unrestricted sense." 29
The ego in the restricted sense includes only the K,L,M, a, ~. y and the psychical
variations of the A,B,C, and is delimited against the objects of its environment, or the
physical variations of the A,B,C. But the ego too may expand out and appropriate the
elements of the environment to itself In this case, Mach spoke of a more impersonal
Absolute ego, including within itself the physical variations of the A,B,C. Mach also
gave examples of ego expansion where the influence of the ego over its environment is
such that the objects are causally linked with it as much as with each other:

When I say tbat the table, the tree, and so forth, are my sensations, the statement, as contrasted
with the mode of representation of the ordinary man, involves a real extension of my ego. On the
emotional side also such extensions occur, as in the case of the virtuoso, who possesses as perfect
a mastery ofhis instrument as he does ofhis own body... 30

In this sense, by allowing the ego to have a real effect on its environment,
Mach considers the elements of the ego as forces completely on a par with physical
elements and capable of influencing them causally.31 In both the Analysis of
Sensations and Knowledge and Error Mach admitted that one could speak of the
"ego in the unrestricted sense" by including the environment within the ego so that
"my ego includes the world as its idea and representation. "32 But in this case,
including the environment within the ego means it cannot be separated from the
world as a whole, as those elements are already immanent parts of an external
world. Mach insisted that such an absolute awareness, in the sense of a restricted
ego, would only be possible if we then restricted the Absolute I against the external
world of mind-independent elements and other minds at its boundaries:

If now I call the swn of my mental aspect, sensations included, my ego in the widest sense (in
contrast with the restricted ego), then in this sense I could say tbat my ego contains the world (as
sensation and idea). Still we must not overlook tbat this conception does not exclude others
equally legitimate. This solipsist position seems to abolish the world as independent, blurring the
contrast between it and the ego. The boundary nevertheless remains, only it no longer nms round
the restricted ego but through the extended one, that is through "consciousness." Indeed we could
not have derived the solipsist position without observing the boundary between my own and
others' egos. Those who say we cannot go beyond the ego therefore mean the extended ego, which
already contains a recognition of the world and other minds. 33

Mach backs the solipsist into a comer: accept an unconscious, unlimited set of
elements (which, because no explicit limit is drawn surreptitiously includes any other
elements that may exist), or accept an Absolute I delimited against other elements and
minds, whose existence we must then admit, anyway. Thus, in either case it is
impossible to rule out mind-independent elements, or to separate the ego from the world
as a whole.

I fully grant, of course, that Mach made numerous remarks that sounded very
solipsistic, as if the elements of the Absolute I were the only ones that existed. For
example, he said when considering the difference between physical and psychical
variations, that "the whole system of elements may be set in motion by the severing of a
nerve. "34 What could that mean but that every single element that exists is a human
sensation, at least in its psychical variations? I believe such remarks are the result of
Mach's having adopted the standpoint of naive consciousness in the Analysis, while
taking "for the purposes of general orientation" the magisterial view of beginning with a
complete set of world elements and dividing them up into egos and objects.

I don't think that Mach succeeded in completely limiting himself to the naive
standpoint as he promised, as he attempted to account for a knowledge of objects and
other minds. For example, he spoke only of a limited objectivity that sensation-elements
could achieve by artificially isolating the variations of the A,B,C from the K,L,M:

The ascertainment of the dependence of the elements ABC on one another, KLM being
disregarded, is the task of natural science, or of physics in the broadest sense. But in reality the
ABC are always dependent on the KLM. There are always equations of the fonn
f(ABC ...KLM)=O. Now since many different observers KLM, K'L'M', K"L''M" are treated like
physical instruments, each with its peculiarities, its special constants, and so forth, from which the
results, as finally indicated have to be set free, we succeed in eliminating the accidental influence
of the variation of KLM, and we thus obtain only the element that can be stated as common
property, namely the pure dependence of the ABC's on one another. 35

He said further that the pure variations of the ABC will lack a relation to any
particular qualities, which their contact with the sense organs of the human observer
gives them, and will assume the form of "spatial identities."36 This is of course no
conception of a mind-independent external object at all, not even an abstract spatial
identity, since for all we know the spatial properties may also be due to their causal
contact with the human sense apparatus. The method of variations only gives a kind of
standardized conception of the A,B,C in relation to the average human observer. This
purely immanent objectivity says nothing about what the relations of the A,B,C would
be without any observer present, which makes the A,B,C more dependent than ever on
the human body. Mach seemed to realize this in the above passage ("But in reality the
ABC are always dependent on KLM") and in his notebook:

Holding constant, excluding of the body

freeing it from chance incidentals. Not absolute, quant and qualit...
KLM cannot be excluded, but held constant. Then conceptual relations of the ABC. By
conceptually fixed ABC the pure influence ofKLM on ABC. 37

By the method of variations we can isolate abstract "spatial identities" which

represent the pure causal relations of the external object to human sensibility. The
human sense qualities may or may not be abstracted away from this object, but because
it is linked through causal relations to an I its form is still indirectly determined by
human sensibilities, not the object an sich. This is because in the drawing of boundaries
around groups of elements we do not begin with the objects an sich but with the
complex separated out in the human ego. By beginning with this boundary between the
ego and objects as a standard, the space-time forms of the objects of the environment
are reciprocally determined by that interaction. Hence, if the I is suddenly removed
from that complex of relations the form of objects as codetermined by their relation to it
also falls to pieces and they become susceptible to an infinity of other relationships.

Mach also spoke rather unsatisfactorily about the analogy to the sensations of other
minds, or even to the sensations of animals, by saying that we "add in thought" the
sensations of others that are not given to us. This suggests a traditional argument from
analogy used by idealist philosophers in which new, foreign elements have to be added
to the ones already assumed as given. But this is unnecessary, especially considering the
fact that both the restricted and the absolute ego are obtained by making narrower and
narrower restrictions from a set of elements that already includes these foreign elements
to begin with. The situation would be more accurately portrayed if we begin with the
unrestricted elements of the cosmic soup, including the elements of other minds,
animals, and the world elements of matter. Experience would then help us draw
boundaries restricting the elements of the soup to the community of egos and the world
of external objects. The inference to my own ego apart from others would then be the
same inference used to separate egos from the environment of other elements. Mach
suggests something of the sort when he says:

One who thinks that because he has recognized his own ego as the medium of all knowledge he
may no longer infer the ego of others by analogy thus falls victim to a strange though widespread
systematic superstition. For the same analogy serves the exploration of one's own ego. 38

Eventually, of course, Mach has to acknowledge that he tacitly begins with a larger
set of elements and functional relations when he says that "the sensations of another
person are causally attached to the brain observed" and that different egos are merely
"different points of attachment" drawn around a given array of sensations. Mach's
contemporaries, the philosophers Richard Avenarius and Wilhelm Schuppe, dealt more
directly with the inference from analogy in their writings. They adopted a similar
strategy of beginning with the sensations of "consciousness in general," meaning all
egos, and then dividing them up into the contents of individual awareness. 39

In Knowledge and Error, Mach finally confirms what I think was his real view all
along, that every element exists in a primordial "cosmic stream." As it turns out-and
indeed this was inevitable-the elements are put forward as the starting point of
scientific inquiry, not the ego and the given which are the terms of the idealist:

If the ego is not a monad isolated from the world but a part of it, in the midst of the cosmic stream
from which it has emerged and into which it is ready to dissolve back again, then we shalt no longer
be inclined to regard the world as an unknowable something and we are then close enough to
ourselves and in sufficient affinity to other parts of the world to hope for real knowledge. 40

Thus, in addition to what I consider miscarried (and otiose) epistemology, Mach

believed in the scientific possibility of what he calls "real knowledge" of the external
world, a direct acquaintance with the world elements of nature beyond the limits of even
the Absolute ego and its immanent objects. This meant a direct experiential knowledge
of nature an sich and not the quantitative chasing down of spatial identities that at most
could only give us knowledge of the exterior structure of the world without revealing its
interior nature. The future of science would be to discover and break through the link
between human sensations and the "cosmic stream" of world elements.

In Mach's correspondence with Gabriele Rabel, which she quotes in full in her
1921 Physikalische Zeitschrifi article, Mach wrote about what would happen when the
psychophysical tunnel was completed through:

Now we must investigate everything from either the physical or the psychical side. When,
however, the Tunnel between the physical and the psychical is completed through or almost
through, no longer will we be restricted to this. Then it will be granted us to ask how animals
sense. Of course, this still lies quite far off. But for those who come after us and are cleverer there
will be no restrictions, no Ignorabimus -- The astronomers of today know that the Ptolemaic and
the Copernican world-views are both practical restrictions and that one is allowed freer room for

Only direct experience of the world would qualify as Machian "real knowledge,"
not just the quantitative structure of quality-less relations. Mach thought the pursuit of
quantitative relations gave only half the naturalistic story, which is why physical science
so badly needed sense physiology and psychophysics in order to progress beyond its
quantitative, Galilean formulation. With the psychophysical tunnel built through, we
will know nature first hand. We will know what nature really looks and feels like an
sich, experience we already have in our narrow but direct access to the specific energies
of our nervous systems.

The existence of the Weltelemente-a seemingly speculative, philosophical issue-

thus transformed in Mach's capable hands into a practical question of scientific, perhaps
medical, possibility, not a theoretical one to be decided a priori. In fact, I think that
Mach's investigations could reach their goal with the empirical demonstration, or rather
experience, of the first world element (non human sensation) in the laboratory. One
could imagine experiments starting on both the physical and the psychical ends of the
tunnel, experimentation with chemical materials exhibiting lifelike properties likely to
be accompanied by psychical elements similar to human or animal sensations. On the
other end one could imagine subtle temporary alterations of human sensibilities-
sharing nerve endings and sensations of others (as Mach suggested). One might see
experiments made with substitutions into a full human awareness, fusing an ordinary
awareness with one part foreign sensations from some other source, whether animal,

chemical, or still more exotic substitutes. All such possibilities seem well within the
range of Mach's powers of contemplation.


Mach used his causal treatment of the elements to solve metaphysical "pseudo
problems" in psychology. Examples of these include the problem of upside-down
images on the retina, or, as he says, the discussion of how "a large perception of a tree
could fit in the little head of a man," confusions that concern the ambiguous placement
of the elements in space and the difference between psychological and physical spaces.
Mach also believed that psychologists had committed the "absurdity" of thinking
sensations spatially into the brain, i.e., imagining that sensations are inside, on top of, or
in some other spatial relationship to brain tissue, whereas he himself held that sensations
were functionally, but not necessarily spatially, attached to brain processes.42 This raises
the interesting question, suggested by Mach's subtitle of the Analysis, as to what the
relationship actually is between the elements of the physical and the sensations of the

Clearly for Mach, brain tissue is, like everything else, made up of elements, so
there is no need for a special psychophysical relation over and above the way one
sensation acts on another, or the way one element acts on another, to explain how a
sensation might act on a physical element. So dualism and its interactionist helps of
occasionalism, preestablished harmony, and "other monstrosities" are completely out of
the question. Mach also had little interest in other traditional approaches to the problem,
such as dual-aspect theory which attributes mental and physical "aspects" to some
tertium quid,43 or epiphenomenalism in which sensations may be caused by certain
physical events but not vice versa. Against the latter, Mach pointed out against Ribot
that sensations were just like physical elements in possessing causal powers: "[Ribot]
goes too far when he maintains that everything psychical is merely surajoute to the
physical, and that it is only the physical that produces effects. For us this distinction is
non existent."44

Likewise it is futile to search, as Lovejoy does in Russellian neutral monism, for a

damning property dualism or a dualism of psychological and physical predicates in
Mach's two sorts of variations of the elements. For one thing, psychological and
physical divisions of variations are only temporary based on the current state of
knowledge and are both taken from a set of general natural variations, which are neither
the purely quantitative abstractions of physics nor the more specialized and individual
variations of sensation, memory, and association of psychology. Rather just as there is
one type of element, so, too, there is but one type of variation. The fact that physics
tends to ignore the individual qualities of events and focus on quantitative features that
range across many individuals, and the fact that psychology handles seemingly
individual events not yet predictable by general laws, is only a "temporary" state of
affairs for Mach.

The more general merger of the methods of physics and psychology together into a
general natural science would obliterate such distinctions. Physical processes could be
understood intimately and individually, and the equivalent of psychical processes such
as memory and sensation could be found more generally in organized matter. 45

Mach's specific utterances about relations between sensations and elements are
surprisingly brief and not completely unambiguous. He distinguishes, for example,
between the distinct particulars of a green leaf and the nerve process that is correlated in
parallel with it, citing his old principle of the parallelism of the physical and the
psychical which he first stated in 1865 and which guided so much of his work in

Every psychical event corresponds to a physical event and vice versa. Equal psychical processes
correspond to equal physical processes, unequal to unequal ones. When a psychical process is
analyzed in a purely psychological way into a number of qualities a,b,c then there corresponds to
them just as great a number of physical processes a,p,y. To all the details of psychological events
correspond details of the physical events. 46

Mach also speaks against the identity of particular sensations and elements of brain
processes, as if the elements of the brain process and the sensations of the green leaf
exist side by side and are correlated by an external relation to one another:

When I see a green leaf (an event which is conditioned by certain brain-processes) the leaf is of
course different in its form and color from the forms, colors etc, which I discover in investigating a
brain, although all forms, colors, etc. are of a like nature in themselves, being in themselves
neither physical nor psychical. The leaf which I see, considered as dependent on the brain process, is
something psychical, while this brain-process itself represents, in the connection of its elements,
something physical.47

One might ask: What has become of Muller's notion that the green color is the
interior look of the specific nerve energy? Did that not suggest the ultimate identity of
particular sensational and physical energies? But the "absurdity of thinking sensations
spatially into the brain," 48 seems to forbid thinking of sensations as the spatial interiors
to brain processes. Moreover, the interior placement of sensations is false for
psychological space, for, as Mach points out, my sensations of objects are plainly "on
the outside" from the sensations I have of my head, which already "shares the same
spatial field with them. "49

Neither can we look at a friend's brain tissue while he is making an observation and
expect to fmd sensations of green or other sense qualities intermingled with his brain
processes. According to Mach, when I observe a person's brain I may functionally add
his sensations to the brain processes I observe, but these sensations are causally, not
spatially, annexed to the brain in front of me. Of course the sensations and the physical
elements being homogeneous can be causally linked into an object of which one might
say, after a fashion, that the sensations are energies of the nerve that are sensed by being

the nerve itself and not something linked to it externally. But because the sensations of
the brain itself are different particulars from the sensations observers will have of that
brain, or those induced in measuring instruments that are influenced by brain activity,
all of these different particulars must be set aside from one another in a Machian
"functional presentation."

For example, taking the case of the friend's brain under observation, let e-elements
be the physical elements of his brain, let e-elements be the elements of the object an
sich, let s-elements be the friend's sensations, and let s'-elements be the observer's
sensations, including both a perception of the object and a perception of the friend's
brain tissue. Then the causal map will look like this:

e s s s/e e e e e e e e s'/e s' s' e
e s s s/e e e e e e e e s'/e s' s' e
e s s s/e e e e e e e e s'/e s' s' e
e s s s/e e e e e e e e s'/e s' s' e
e s s s/e e e e e e e e s'/e s' s' e
e s s s/e e e e e e e e s'/e s' s' e

Now we can use our knowledge, and memory, of the functional relations to
separate out the different factors and delineate objects, such as the ego, the peripheral
stuff of the brain tissue, and the elements of the objects being perceived. For example,
the object being observed, say a candlestick, divides into: (1) the e elements of the
object an sich; (2) the e-elements of the friend's brain tissue; (3) the s elements of the
friend's sensations; (4) the s' elements of the observer when looking at the friend's brain
tissue; and (5) the s' elements of the observer looking at the candlestick.50 We may, if
we wish, unite numerous of these individual aspects functionally into objects of "the
brain tissue" or "the candlestick" as in Fechner's approach to the problem.

What Mach has really done is tum the world practically inside out, by making all
elements, even physical ones, individual existences an sich, each with their own
qualities as well as their external causal relations to other elements. It is as if Mach had
eliminated all exteriority and seen the qualitative interior everywhere-MUller's specific
energies writ large throughout nature. Every element thus has the look of its own energy
an sich and not just the interior energies of nerves. When we observe another's brain
tissue with our sense organs or with instruments, we are interfering causally with the
sensations an sich and producing changes in further elements annexed to them. These
patterns of interference finally show up in variations of second and third hand elements
as the external relations of the brain tissue "carrying" the sensation process. But all of
these elements are themselves existences an sich; they are not the exteriors of processes
in which sensations form the interiors. Thus, despite their causal linkage, physically

observed processes, no matter what their detail of correlation, never completely disclose
the nature of other elements an sich by means of external causal relations, for these
elements to be observed are always merely annexed to them extrinsically. The problem
is that we are always taking these annexed elements as directly representative or
identical to further elements merely related to them. In two Notebook fragments the
question is addressed-in passages probably about the sensations of movement and our
inability to be conscious of "sensations of innervation." Mach remarked that here
sensations that come to consciousness are only the "released" energy of the nerve
annexed externally to the energy that physically (externally) affects other nerves and
which remains unconscious:

Sensations, ideas, only the releasing, not acting, energy of the organism However both are of the
same nature. Consciousness of the connection. It [the connection] is to be sure not physio-
chemically "accessible." Consciousness a connection of energies. The size unimportant.
Ridiculously small. But the consequences priceless. 51

Mach thought sensations hung together causally with the other physical elements of
the nerve, and the external physical energies released by a nerve process, which, unlike
sensations themselves, are recorded by physio-chemical means and become the physical
correlates of sensations:

Physio-chemical until the central nervous system is reached a part is sensation, but not such that
they are deliverable to our sense-organs, the microscope, galvanometer, or reagents. Sensations
release other analogous ones in the nerve and muscle. There they appear as physico-chemical. 52

So when we observe brain tissues with instruments or with our own senses, the
qualities of the elements an sich are not retained, but those external properties that
survive a change from one place to another, of form, consistency, shape, and physical
properties like mass or charge, remain unchanged and preserved through the set of
transformations that the functions represent. It is because we are accustomed to call the
stable functions "objects" and ignore their constituent elements that we have so much
difficulty with psychophysical relations, for in the analysis of sensations individuality
matters and not just functional patterns that may be taken for "the same" object no
matter which individual elements make them up.

If, to the physicist, bodies appear the real, abiding existences, whilst the elements are regarded
merely as their evanescent, transitory appearance, the physicist forgets, in the assumption of such
a view, that all bodies are but thought-symbols for complexes of elements (complexes of
sensations) ... By the recognition of this fact, many points of physiology and physics assume more
distinct and more economical forms and many spurious problems are disposed of. 53

To the philosopher, such observations may remind one of the classroom exercise in
which everyone is encouraged to realize that the same object of the coffee mug on the
table is really a different individual manifestation altogether in the minds of the different
people seated around the table. Some see the handle as more circular, some do

not see it at all, some see the lip and the inside while others do not. In this way one
comes to realize that the object is really more of a construction, a functionally
determined social object and not a true individual, which is much more difficult to grasp
in its nature. The fact that these individuals are also in a constant flux and must be
recognized as recurrences only with the help of memory makes them even harder to
handle in any scientific manner to which we are accustomed.

One may well wonder what is seen of the brain in a PET scan or similar techniques,
especially when the similarity of form between the sensations of the subject and the
pattern of firing of the cells in the brain can be so close that one wants to say they are
identical. This certainly saves time for many a popular science journalist and even
trained researchers. Mappings of visual experiences such as bar-shaped light and dark
patterns on the retina onto tiny bar-shaped positively and negatively charged cells seem
unbearably close, as indeed Mach's psychophysical parallelism principle says they
should be. But the reason why we never see the sensations of the brain itself is because
elements are individual existences that may come into causal contact with other
elements, but whose individual quality is intrinsic and can never be directly represented
by another. One could never say a sensation of green was identical with a brain process
observed with some instruments, only that it (the sensation) was closely related to
certain physical elements, the physical substrates of the nervous system, and the
observer's sensations or interferences that are closely related to those substrates. But the
so-called privacy of mental states has nothing to do with the sensations being mental,
for calling them mental and not physical is again to point to their extrinsic relations. The
external "inaccessibility" of sensations is not a hallmark of the mental but is a property
of every particular element, such as the "private" physical elements making up a
hammer. The reason why we do not notice this is because we are so used to ignoring the
qualitative uniqueness of elements in material objects and to concentrating on their
quantitative relations in physics. Thus, without exactly agreeing with Mach, one can
well imagine why he so despised atoms, as these, to him, were nothing but
indistinguishable bundles of external quantitative properties of mass, charge, force, etc.
with supposedly no individual nature. 54


For the philosopher, Mach's idea of a natural science that would span psychophysics
and physics is both exciting and frustrating because, like all science, Mach's proposal is
open-ended and does not settle a priori the many empirical questions of the relations
between elements and sensations. For example, the nature of the causal connections
between the elemental qualities is left undefmed. We know as much and as little about
these actions as we do about the transformations of physical energies like heat, vis viva,
electrical and magnetic forces and gravitation into one another. The initial states and the
fmal states of such transformations can be known to a degree of closeness (limited by

the uncertainty relations, as we know now) but nothing about the "how" of the
transformation is thereby revealed-indeed not even physical science can do this.

Still, one might ask about the causal force that Mach attributed to sensations to act
on other elements, and vice versa. He differs from epiphenomenalists, whom he
criticized for failing to recognize mental to physical causation and the concomitant
extensions of the ego out into the environment. The Notebook passage cited above has
Mach speaking of the tiny "energy" of the ego and of consciousness as "a connection of
energies." We might compare this with what he says to the Analysis of Sensations,
acknowledging the fact that the conservation of energy could neither rule in or rule out
such sensation energies. Still, he suggested that the psyche might still play a causal role
in events by intervening between the equal initial and final states of physical processes:

I cannot refrain from. .. expressing my surprise that the principle of the conservation of energy has
so often been dragged in in connex.ion with the question of whether there is a special psychical
agent. On the assumption that energy is constant the course of physical processes is limited but
not necessarily determined with perfect uniqueness. That the principle of conservation of energy is
satisfied in all physiological cases merely tells us that the soul neither uses up work nor performs
it. For all that, the soul may still be a partly determinant factor. 55

What he has in mind is that sensation energies can fill intermediary links between
physical elements and leave the total energy entering and leaving the sensation complex
the same. The conservation law of course says nothing about how many intermediary
processes there might be. In the Conservation of Energy Mach admitted little islands
(closed systems) of elements shut up in their own reciprocal interactions, and it is very
tempting to think that some of these islets were egos. 56 Their existence opens up
contingencies in the causal map that cannot be predicted a priori.

The existence of the peculiar qualities of the ego is a contingent matter, which
cannot be completely determined by all of the environing facts of elements annexed to
them. Mach's principle of parallelism demands only that the elements annexed to the
particular sensation qualities be sufficiently rich and variable in their dimensions so as
to be able to annex causally to the richness and variability exhibited by the
phenomenology of sensation.

If the particular qualities of sensations are contingent, this does not mean that there
could be a brain with all of the normal physical processes and no interior nature
whatever. It has been suggested, for example, in a recent book by David Chalmers, The
Conscious Mind, with some neutral monist overtones, that there could be (as a matter of
logical or conceptual possibility) "zombies" physically identical with ordinary
functioning human beings but without any interior sensations or mental life. While
certainly a Machian could accept the idea that two physically identical brain states
might possess different sensational interiors for the same extrinsic physical states, the
elements themselves would remain, which are indeed just as much phenomenal
qualitative data as any sensations. If one accepts the idea as a conceptual truth that

individual quality is an essential property even of physical events, as, for example,
Russell held that all physical events have an intrinsic, individual quality or character, a
"physical zombie" has no meaning, even for natural objects. It is impossible given this
conception of the world to find just nothing corresponding the interior nature of
physical elements an sich, and merely the set of external relations with no interior
nature at all, which would really be tantamount to asserting their non-existence, and
with them the nonexistence of physical realations among these entities.

One of the traditional questions any theory of psychophysical parallelism is

expected to answer is the following: do identical sensations always accompany identical
brain processes, and do identical brain processes always accompany identical
sensations? Mach's 1865 statement of his principle of parallelism suggested there was a
double-sided implication, but in light of the fact that the relation between elements was
an external one, there do not seem to be any special conditions for one to accompany
another besides the requirements of ordinary causal relations, whatever those turn out to
be. Of course we must expect that not just any bundle of physical elements could
causally annex to the complex and rich human phenomenology with its sensations of
blue or the timbre of a violin, for they might lack the equality of dimensions demanded
by the principle of psychophysical parallelism. But, broadly speaking, there is no reason
why those sensation-elements should not also be annexed to a colloid of other
sufficiently rich physical elements besides brain tissue. Since it may also be considered
a real mind-independent physical quality, a human sensation of blue might certainly
occur in complexes other than a human ego, as Mach hinted. 57

Another fascinating development would be if the same brain states could be

annexed to completely different mental states at different times. Could a person be in the
same brain state A at two different times and be in mental state B the first time and B' the
second time? Could a person be in the same mental state B at two different times and be
in two different brain states A and A'? As unlikely as this is, and a gross affront to the
regularity of nature, as far as I can See nothing whatever prevents this from happening on
a Machian view. The double-sided parallel between a specific complex of sensation
elements and a specific complex of physical elements might be upset on either the side
of elements or of sensations if either of them undergo reciprocal actions among
themselves and then causally annex to the other side. Or the annexed sensations
(annexed elements) might simply be different at different times. Polyadic many-to-one
interactions between elements and sensations open up still further possibilities of
practically infinite complexity. This is one true sign that the true causal map of the
world, including the psychophysical relations, can only really be filled in by the hard
work of natural science. All that can be done a priori is to remove the instinctively
tough mental barriers to such research, otherwise known as the elimination of


Mach's doctrine of the economy of thought, or Denkokonomie, has long been

recognized as one of his most important philosophical achievements. He claimed to
have held it as early as 1861, when he began his Privat-Dozent lectures, and
expressed it first in the Conservation of Energy. 1 Subsequent treatments occurred in
his lecture "On the Economical Function of Physical Theory" (1882), the Mechanics,
and chapters in his later writings, the Theory ofHeat and Knowledge and Error.

Generally speaking, Denkokonomie plays the much needed role of a formal

principle in Mach's philosophy. Mach generally neglected the need for form,
insisting that all that existed for him were elements and their instantaneous
functional connections. Hence, the role played by extended objects was a derivative
one in his scheme. Objects for Mach were always changing complexes grouped
together in a single history for the sake of convenience because it was more
economical and less of a strain on memory to do so. There is thus a general
skepticism of objects as realities in their own right for Mach, since the temptation, he
believed, was to reify them into causally isolated things in themselves or private
egos. However the stable reality offered by objects and concepts was not one Mach
could ignore, and the economy of thought was his attempt to come to terms with it.

In all, there seem to have been two forms in which Mach presented the economy
of thought and which he used for different purposes: first as a theory of scientific
explanation and second as an ontological theory of the genesis of objects and laws
from his elements.


Mach's theory of explanation appeared in his Conservation of Energy, where he was

considering the deduction of laws from more general ones. The most universal of all
the laws of physics, Mach claimed, was the principle of the excluded perpetual
motion, or the idea that energy cannot be created nor destroyed, either once or as the
result of a repeating cyclical process. From here various special laws could be shown
to follow, such as Lagrange's principle of virtual velocities and Camot's principle
(that the capacity of a substance to perform work from heat is dependent solely on its
temperature and not its chemical composition). They follow simply by virtue of the
fact that if they were not true, then a perpetuum mobile could be constructed.
Deducing so many results from such a simple idea seems to be enlightening, as if the
various principles of science could all be explained by this one principle. But Mach
quickly puts a stopper on the enthusiasm. For him, "scientific explanation" is a

misnomer. We make special laws intelligible only by deducing them from more
general "unintelligibilities" that remain unexplained. Thus, the more specific law is
gathered under the more general, but we do not gain thereby in knowledge, he said.
In order to know that a particular problem is susceptible to application of the
excluded perpetual motion principle, we must first recognize the particular
phenomenon (heat energy in Carnot' s cycle, statical moment in the principle of the
lever, virtual work in the case of the inclined plane and pullies) as manifestations.
But in this knowledge the rest is already contained, so all we know ultimately
redounds to knowledge by experiment. Where we do gain from deductive
explanation as such, Mach says, is in the economical power of organizing our special
laws under more general ones:

Besides the collection of as many facts as possible in a synoptical form, natural science has yet
another problem which is also economical in nature. It has to resolve the more complicated
facts into as few and as simple ones as possible. This we call explaining. These simple facts, to
which we reduce the more complicated ones, are always unintelligible in themselves, that is to
say, they are not further resolvable ... Now it is only on the one hand an economical question,
and, on the other, a question oftaste, at what intelligibilities we stop. 2

Mach says that history plays a leading role in the acceptance of

unintelligibilities, and pointed out that the principle of the excluded perpetual motion
predated mechanics and seemed natural because it was already familiar. 3 Of course,
which arrangements of laws are the most economical depends a great deal on need.
Different features (fewer primitives, greater observational content, easier
visualization) may be economized (or maximized) by different arrangements, with
the choice among alternatives often being nothing more than judgments of
convenience. As Mach says, what is economized is memory. 4 By deducing the many
particular laws from a central principle, the mind is not taxed with the remembrance
of every case. But Mach had not shown in any way that the material to be ordered,
the laws or the facts themselves, were the creations of need or memory or of some
deeper teleological or normative requirement. This explanatory version of the
economy of thought might thus be said to concern the arrangement of laws already
framed; it would not concern the search for basic natural law itself. One could be
pluralistic, for example, about which laws are more fundamental than others but still
hold that the objects and laws of nature must already be present before they are
ordered by some teleological principle.

Historically, many of the criticisms of Machian economy were directed at this

"explanatory" version. The American editor of the Monist, Paul Carus, under whose
auspices many of Mach's books were translated and published in English, wrote that
Mach had only explained the arrangement of laws under one another and not the
existence of real laws in nature:

While I have no objection to Mach's description of science as economy of thought, I would hesitate
to say tbat it is sufficient as a definition, because science is a correct (or adequate) description of
facts in their essential features. Exactness, correctness, adequacy, truth, or whatever you may call it,
is the main thing and comprehensiveness comes in second as a natural conseuence whenever the

essential features have been rightly understood .. .I try to explain why economy of thought is
possible at all, and my explanation is based on the idea that theory is not a purely subjective device
to deal with experience but that there is a feature in experience itself that justifies the formulation of
theories ... the purely formal is not merely a matter of method, it is not purely subjective, but must be
a feature of the objective world. It explains why there are uniformities. 5

Herbert Buzello, a Kantian who produced one of the first systematic critiques of
Mach's philosophy, stated the objection more succinctly:

A teleological consideration in a science has only the value of a method or presentation, or of

inquiry. It serves to order the individual facts for a given goal, in so far as they are suited to realize
that goal. But it is continually assumed that the facts as such are already given. 6

Finally, Edmund Husserl, the founder of phenomenology, wrote a criticism of

Mach's Denkokonomie in the Logische Untersuchungen. He argued that the
psychological details of conserving memory and strain had little to do with the
underlying logical norms of pure thought driving any scientific investigator to collect all
special laws under a general principle. This rational norm seemed to Husserl purely
logicat,? and the use of thought-economy merely an imperfect psychological
approximation to it:

Before any economy of thought, we must already know the ideal of what science ideally strives for,
what lawlike dependencies, what fundamental laws and derived laws and the rest are and can
manage ideally, before we can determine and estimate the thought-economical function of their
knowledge. To be sure we have certain vague concepts of these ideas before scientifically
investigating them, and thus there may be a discussion of thought economy before the building of a
science of pure logic. But the basic lay of things is not changed, in itself pure logic goes ahead of all
economy of thought and it is against all sense to base the former on the latter.'

Husserl often writes in a Platonic vein as if the housebuilder must consult the
perfected idea before beginning to build. But Mach's answer to Husserl emphasized
their differences of approach: Mach starting with the specific history of science and the
historical development of concepts, and Husserl beginning "at the other end" with a
rational ideal to which these stages seem to approximate. Mach also emphasized that a
normative principle might well underlie his own theory of economy:

If my theory of mental economy be conceived merely as a teleological and provisional theme for
guidance, such a conception does not exclude its being based on deeper foundations, but goes
toward making it so 9

And to leave no doubt, Mach attaches a footnote to the Analysis of Sensations,

where he discusses a biological and evolutionary foundation for thought very much
at odds with Husserl's logical norms. There is thus a basic difference between Mach
and Husser! over the meaning of science. Husserl essentially saw all science as a
derivation from the ideal afforded by logic and ultimately perhaps from a supreme
principle of reason. In Mach's case, the approach is a genetic and historical one and
the focus is less on the validity of scientific law than on how it came to be. Mach
also emphasized that the unintelligibilities we stop with are historically and
instinctually conditioned. Mach insisted that there is no rationale for preferring one
fundamental set of principles to another if the same set of specific laws are deducible

from both of them. Indeed, often the same particular laws will be deducible from
different initial assumptions and one can adopt this or that basis at pleasure. Husserl,
on the other hand, was trying to discover principles of representation and logic so
general as to be used in the deduction of anything from anything, which is the power
of his technique. Such conflicts have been handled by separating the "context of
discovery" from the "context of justification." Husserl accepted such a distinction
already, but I doubt that Mach would have granted that the context of discovery has
nothing to do with what principles or procedures are ultimately accepted as justified.
That is why Mach insisted that historical studies were often the only way to break up
the rigidity of established sciences, making progress along alternative routes possible
that simply could not be seen within the science itself.


We now pass to Mach's use of economy as a principle of form in his Elementenlehre

to account for objects and laws. Machian science was essentially a two-floor
building. On the first floor were elements and functions, where the business of
science was to describe the functional dependence of the elements on each another.
Mach believed that at the most fundamental level, the world consisted of neutral
elemental qualities (colors, sounds, tastes), some of which were human and animal
sensations and some of which were purely physical magnitudes (such as pressures and
point-potentials). Mach was convinced that the most realistic portrayal of the world was
in terms of these unique, irrepealable states, a proposition summed up in his famous
maxim that "Nature is but once there." 10

Because of the basic uniqueness of every experience, Mach described his view as a
"mosaic" of individual qualities making up the world at any instant, appearing suddenly
and then vanishing. However, like a mosaic, in which the placement of the tiles depends
in a general way on the placement of those that bound it, there is also an instantaneous
dependence of the elemental magnitudes on one another, determined by their
intensities. 11 He expressed this dependence in a general equation by holding that the
intensities of the elements had to sum to zero, in closed systems and in the universe as a

F(a,~,y,o, ... ,ro) =0. 12

Alternatively, each element's variations in intensity could be expressed as a function

of the others':

a =f1 (~,y,o, ... ,ro).

~ =fz(a,y,o, ... ,ro).
y =f3( a,~,o, ... ,ro).

Mach considered the fundamental equation as the exclusion of a perpetuum mobile, a

perpetual motion being defmed as a variation of elements not compensated for by a
variation elsewhere in the opposite sense. One could say that the changes of intensity of
these elements--energy changes being one such manifestation-are more or less related
like action and reaction. In the 1910 essay "Sensory Elements and Scientific Concepts" 13
Mach even said that he regarded the instantaneous elements as opposing "reactions,"
against one another, which explained why the elements so readily compound
themselves into functions and complexes, and why Mach could assert that elements
always occurred in complexes.

Thus there was a Machian notion of functional complex or object that emerged
directly from the opposing intensities themselves. To the degree that a certain class of
closed systems of elements can be identified within a snapshot of time, Mach was
willing to admit that there were real objects of experience, real forms and relations. But
these forms were completely immanent in the exchanges of intensity the elements
underwent. Thus, Mach's functional objects were neither enduring over times, nor were
they even uniquely determined within times by the criterion of"closed system." Rather,
the whole system of elements is closed and there are various islets within the whole
system that are also closed unto themselves; that is the energy that enters, if not zero, is
equal to the energy that leaves, if not zero. The boundaries of closed systems are such
that any composition of closed systems is also closed.

Barry Smith and Kevin Mulligan 14 call Mach's functionality principle "non-causal
dependence" a notion they fmd employed by other Austrian philosophers such as Franz
Brentano in place of spatia-temporal connection. Perhaps the most important feature of
Mach's version of dependence for them is that the functional variations are
instantaneous relationships between qualities and the intensive dimensions in which
they vary. Indeed, Mach had said this himself, insisting that space and time were merely
indications of more general aspatial, atemporal relationships among the elements so
abstract that mathematical functions alone could describe them.

On arriving in Mach's world of elements and functions described above, we

would not recognize the bewildering flux of intensities appearing in one instant,
opposed by others, and then vanishing in the next, leaving no trace behind. There
would be no way to orient oneself in such a world, no stable forms to hold on to
from one moment to the next. Even the fiction of a Kantian self-observer preserving
and recognizing himself as identical against the flux would not survive. What saves
us from flailing away in such a dismal state, according to Mach, is not our intellects,
which are exceptionally feeble at this point, but rather our memories and our ability
to keep records. Mach affirmed, here in his notebook for example, that it was part of
experience that we continually reproduce memories along with our present sensations:

Representation of time
The images of earlier times are tied to the images of present states. States in the memory field are
bound to other states in the perceptual field. That which was, we see at the same time with that
whichis. 15

While individual sensations themselves never repeat, nature being but once given,
we do carry traces of some of them in memory. It then becomes possible to think of
some sensations as repeating instances of past experiences, by associating the memory
and the present sensation, as if the latter is a recurrence of the former. It is here, and not,
for example, with the advent of words, that we first possess concepts of colors like red,
blue, and green, or other elements, as Mach sometimes calls these repeating element-
types. This concept or schema is much more "economical" than a separate record for
each individual occurrence:

When we look over a province of facts for the first time, it appears to us manifold, heterogeneous,
confused and contradictory. At first we only succeed in holding fixed every individual fact or
connection from the rest. The province is as we say unclear. Bit by bit we find the simple, self-
identical elements of the mosaic, from which the whole province permits of being reconstructed in
thought. If we get this far, to recognize the same facts again, we do not feel ourselves so foreign in
this province, we look over it without exertion, it is explained for us. 16

At best the modest beginnings of science are the uncovering of what are to us simple, self-identical
entities. Half aware and non arbitrarily, man obtains his first knowledge of Nature, in that he
pictures (nachbildet) the facts reproduced in thought, and projects them forward (vorbildet), in that
he completes ponderous experience through his more quickly moving thoughts, at first only to his
material advantage. 17

The original flux and its momentary functional complexes eventually give way to a
stage where we distinguish repeating individual elements that we recognize as after-
images of past experiences (Nachbilder), and fmally a stage where it is permissible to
discuss repeating complexes of the environment, such as bodies, and general causal
connections between bodies or their features.

Mach claimed that the very essence of science is "to complete a partially observed
fact in thought," 18 that is, to fill out the meager direct evidence of sensations by piecing
it together with our memories and imaginations into a complete space and time order
with a history. Mach says that in his work on falling bodies, Galileo 19 had to fill out the
shape of the parabola by adding to its present position and velocity all of its imagined
past and future positions in the curve. Mach calls these cases examples of the
"accommodation of thoughts to facts," as the missing parts of experience, such as stages
of the past as well as sections of space not currently in view, can be filled in by images

When we mentally add to those actions of a human being which we can perceive, sensations
and ideas like our own which we cannot perceive, the object of the idea we so form is
economical. .. Now this is exactly what we do when we imagine a moving body which has just
disappeared behind a pillar, or a comet at the moment invisible, as continuing its motion and
retaining its previously observed properties ... We fill out the gaps in experience by the ideas
that experience suggests. 20

I find this a very apt explanation of what it is like to understand a phenomenon

scientifically, which is to say having a solid imaginative schema that works in thought
the same way the phenomenon would work in reality if certain parameters are provided
by experience. In the imagination we can see in the falling body the maintenance of the

velocity it has acquired compounded in every moment with the velocity imparted by the
acceleration, although we never actually see this as a two-step process in reality. It
seems at times that when the workings are particularly well explained, the imagination
takes on an almost objective character; it will play back the events in the mind in this
way and no other and will even stop the thinker if he tries to substitute mere caprice for
the operation of laws. Mach remarked on this objective imagination as an instinctual
urge to complete the observed fact that confronts us "as if it were a force outside of


Of course, Mach just as quickly pointed out that we are only bound to complete a
partially observed fact to the degree of accuracy necessary for our present needs. The
fmal product is inevitably an abstraction:

In the reproduction of facts in thought we never reproduce the facts in full, but only that side of
them which is importaot to us, moved to this directly or indirectly by a practical interest. Our
reproductions are invariably abstractions. Here again is an economical tendency. 22

In his "economy" lecture of 1882, Mach presented his position that actual
repetitions of experiences were ideal fictions; he pointed out in particular that the
asymmetrical type of spatial or temporal cause and effect depends on a falsification:

If we intended to ascribe the property to nature, that under equal conditions she produces the same
effects, we would not know how to find these equal conditions. Nature is but once there. Only our
schernatizing imagination (Nachbilden) produces equal cases. The dependency of certain properties
on one another exists only in this. All of our toil, to mirror the world in thought were fruitless, were
it not possible to find something enduring in this brightly colored flux. Hence the craving for the
concept of substance, whose source is not different from that of modem ideas of the conservation of

There is no cause nor effect in nature; nature has but an individual existence; nature simply is.
Recurrence of like cases in which A is always connected with B, that is, like results under like
circumstances and the essence of the connection of cause and effect, exist only in the abstraction
which we perform for the purpose of mentally reproducing the facts? 4

The important office economy seems to serve is in the representation of events as if

they repeated themselves. A repetition of features is necessary to provide an enduring
backdrop against which relations of time can be observed. Thus, these examples make it
possible to pinpoint the first instances in which economical abstractions obtain a kind of
necessity from the need for an objective time representation.

Mach treated the question of time-representation at length in his lectures of 1898,

"Specialfragen der Psychologic," reflections that made their way into the 1900 second
edition of the Analysis of Sensations in a new chapter, "Sensation, Memory and
Association." In these notes, Mach hypothesized that both physical and psychical
processes left behind traces. He mentions the example of an old violin, which contains

traces of the pieces it has frequently played impressed upon the instrument,25 remarking
also that, like the violin, "a person that has had a certain experience is an essentially
different person. "26 In Mach's account, memory traces were not enduring
representations or energies that remained lively as part of a fixed ego-complex, but
instead resembled crags and pits in the nervous system in which a present energy of
attention could be nicked off course. 27 This leads to the reawakening of memories,
which are recognized as repetitions of past experiences, although just like current
sensations, they too are drawn from a present energy.

Mach's view does not seem to require that traces themselves endure-but rather
only that the disposition exists to associate a present trace with a present experience
taken to be its repetition. Operations with memories or other records occurring in the
present are offered as a substitute for enduring representations of the past. 28 We might
call this the Odysseus's Scar theory, recalling that the nurse who bathed Odysseus as a
child recognized him by his scar when he returned to Ithaca as a much-altered stranger.
Odysseus himself was no longer recognizable as the person he was, but the association
between him and the scar on his body remained as a means to decode his past identity.

Two of Mach's early critics, Buzello and Richard Honigswald, brought an

argument of Kant's to bear against Mach's notion of time representation by a changing
ego. They objected that a time series was only perceived as such if an unalterable ego
endured through every stage. But if the ego itself alters and is not even recognized as the
same later, then no change would be experienced either. For instance, Honigswald says:

What does Mach imagine by an economical "comprehension of the flux" through a changing
consciousness? He forgets that the perception of the flux would be precluded to such a

While Buzello suggests that:

Were time a sensation, I would have to asswne a time in which to perceive it changing. Further,
Mach himself describes the storing of memories, their connections, their reawakening by one
another, memory and association as the fundamental determinations of psychic life. But between
mere elements these relations are impossible, if one does not asswne memory-feelings or
"indicating marks" of memory representations. With this, naturally, the problem is just as little
resolved that a previously occurring experience is regarded at "another point in time. "30

Mach was, however, well aware of the difference between a changing awareness
and the awareness of change.31 In Knowledge and Error he even appears to take a
pseudo-K.antian position:

At the outset it is clear that a temporal course of mental elements, whether sensations or ideas, does
not as such include consciousness of such a course. If our mental field of vision were always
temporally confined to a sufficiently narrow present, the fact of change could not be perceived at
all. Therefore consciousness must always encompass a finite stretch of time in which there are
fading sensations and ideas at the same time as newly emerging ones, for the first to be viewed as
earlier and the second as later, if in addition we conceive of the relatively stable ego-<:omplex,
characterized by common feelings and so on, it constitutes a kind of rock with the temporally
ordered stream of change flowing past it.32

The difference is perhaps that the ego is only relatively permanent in Mach's treatment
and only needs to endure long enough for some sensations to pass away as others are

In sum, "On the Economical Nature of Physical Inquiry" laid down Mach's
teaching on economy in two overall stages: the normative stage, which explained the
emergence of general concepts and laws based on memory's need for permanence over
time; and the descriptive stage, explaining the sorting of the multitudinous patterns
generated in the first step, to maximize or economize a given lawlike feature. In order to
apply "descriptive economy" in the second stage, the set of all possible temporally
conserved objects in a domain of facts must be recognized first, and afterwards this or
that arrangement of facts can be chosen to maximize or minimize a given feature.
Normative economy, however, allows us to explain why repetitions in experience
appear to occur at all.

Mach generally held that scientific investigators make economical decisions as to

which properties to regard as enduring and which to consider evanescent. Unlike the
arrangements of "descriptive economy" discussed above, these conserved properties are
fundamental to science and are often chosen unconsciously or because of historical
convention. Mach pointed out in his Conservation of Energy33 that Riess's
thermoelectrometer offers a conception of electrical energy analogous to heat and
Coulomb's balance an electric force analogous to the pull of gravity. Which is regarded
as the "quantity of electricity" depends on convention. Mach tended to regard ideas of
conservation, such as the conservation of energy, as the most basic of all such repeating


The conception of Nature with which Mach perhaps identified most strongly was that of
Heraclitus's sun that is new each day. This conception poses a very old challenge to the
conservation of properties through time, which naturally suggests the question: What
about the conservation of energy? If, as Mach indicated, it has the same source as the
substance views that he rejected, must it not too be rejected? Mach's idealism about
substance views, in the form of mass or weight at least, was very pronounced in his

Indeed one is used to regarding the conservation of weight or mass as a direct proof of the
endurance of matter. Only this proof evaporates, when we get to the bottom of it, into such a mass
of instrumental and intellectual operations, that in a certain sense it only constitutes an equation
which our representations, picturing facts in thought, have to satisfy. The dark clump that we
involuntarily add in thought, we seek in vain outside our own thought. 34

Mach also added a footnote to his general discussion of energy conservation in the
Mechanics, saying: "When we reflect that the principles of science are all abstractions

that presuppose repetitions of similar cases, the absurd applications of the law of the
conservation of energy (Kriifte) to the universe as a whole fall to the ground. "35

But Mach went still further in his critique, as his thinking dwelled on the status of
energy conservation and the repetition of events in his 1884 remarks on the work of J.
R. Mayer. Mayer had suggested, in a passage Mach quoted with approval in his
Conservation of Energy, that heat disappears when mechanical work appears, but that
this mutual disappearance could either be taken as identity or its opposite.

Just as little as, from the connexion between the tendency to fall [Fallkrqft] and motion, we can
conclude that the essence of this tendency is motion, just so little does this conclusion hold for heat.
Rather we might conclude the opposite, that, in order to become heat, motion, whether simple or
vibrating, like light or radiant heat must cease to be motion. 36

In his Notebooks, Mach then began to question whether, in irreversible processes

where the ability to perform mechanical work is lost to heat, an identity of the same
energy is upheld, or whether it is simply an economical measure to assume the
constancy of something through change, having to do with our need to reproduce past
circumstances in thought.

The fact is, when things are made to go backwards, so far as can be, the same is again obtained, so
far as experience reaches.
Everything else, however, is just arbitrary determination. I need not conceive heat as equivalent. I
can however.
It is a pleasant, relaxing economical conception fmmded on the strongest, most visualizable image
of substance. The factual is related to reversible processes (Clausius). What is not reversible, about
this we can say nothing. The thought of equivalence is there completely arbitrary because it has no
practical worth.
An endurance for a certain type of reaction of our understanding. Not immediately visible. Time of
no influence. Only nobody can secure a guarantee from time. For irreversible processes the thought
of equivalence has no worth at all. It only becomes practical through reversibility.
Equivalence of transformations.
[Mayer) brings the factual into agreement with his concept and need. 37

This Gedankengang led Mach to doubt the conservation of energy for irreversible
processes because energy lost to heat is no longer completely recoverable as mechanical
work. He went further to argue that: "when however heat at a lower [potential] level
cannot be transformed, work is lost. There is thus no energy principle except in cases of
simple reversibility." 38 And that: "the energy principle has no unbounded validity." 39
And finally: "through irreversibility the whole energy principle is overthrown. That is a
special experience. 40

In his 1892 article "Zur Geschichte und Kritik des Carnot'schen W1irmegesetztes'141
and in his Theory of Heat, Mach converted himself to the idea that despite the
appearance of some reversibility in nature, in truth all processes (even mechanical ones)
were irreversible, or had at least one irreversible component. Of course, one
consequence of doing away with reversible processes, as Mach sees it in his notebook

fragments, is that any exact repetition of events becomes impossible. And this small
correction must therefore be added to all natural laws.

How serious is irrepeatability as a criticism of science, especially given an

acceptance that science always makes abstraction from the real world in order to
represent the largest number of cases?42 All sciences, with the possible exception of
cosmology, certainly depend on the repetition of events over different times, as the
performance of experiments depends on properties of space such as homogeneity (or
indifference to position) and isotropy (indifference to spatial direction). If the repetition
of events is put in doubt, scientific theories cannot be fully confirmed or refuted, since
testing requires that we set a present circumstance equal or unequal to a past
circumstance no longer extant. Both scientific confirmation and disconfirmation rest on
the simple constraint of the ability to recognize repetitive features. But the criteria for
repetition of an event, the comparison of a record and a present event, for example, are
themselves normative constraints isolated from testing, if only because they set the very
conditions for the test. As Mach observed in a notebook entry of 1904, these temporal
norms depend in no small way on our natural disposition to expect that the future will
resemble the past:

1. Disposition toward the vertical 43

2. Disposition toward the future. Practically, only the future is of interest. Biological disposition in
both cases. Not about knowledge in itself, rather about the biologically beneficial. Determined from
the present, past and future are equal, disturbances excluded. Space and time in reality not just
The present can determine past and future in part, with permissible inexactness. Past and future are
imprecisely known, hence the inference to the better known present is moot.
Conscious goals must already be known from the past, and suspected to occur in the future to
prepare for it. Time does not run backwards. Facts do not recur. Thought constructions can recur44

It is interesting to note that Mach's skepticism in regard to repeating forms concerns

some of the most exact principles of science based on symmetries and invariance. The
comparison is not inappropriate, since Mach himself had considered symmetries in
some detail, in connection with his work on the sensory manifolds in an 1871 article,
"Uber die physikalische Bedeutung der Gesetze der Symmetrie.'>'~ 5 Here, Mach defmed
a symmetry as the recognized "recurrence of a sensation," remarking that the most
pleasing ornamental and musical symmetries are produced when visual patterns and
notes, or their overtones, are repeated in a composition. Modem symmetry is defmed as
the invariance of practically any given feature through a coordinate transformation (or a
group of more abstract transformations). For example, a circle possesses centric
symmetry because it remains the same shape and size after a rotation of any angle about
the center.

With a similar definition, that of similar sensations after a transformation, Mach set
himself the problem of determining the symmetries of the space of tone, a one
dimensional manifold with two directions of lower and higher pitch. Based on his
definition, if a melody was recognized as the same after translating the pitches of its
individual notes up or down the scale, the Tonraum must be credited with possessing a

symmetry for those features the melody represented. Mach claimed that only the
transposition of major and minor key resulted in such a recognizable repetition of the
given feature, and thus that only this symmetry existed. 46 He closed his "Symmetrie"
paper with the warning that the properties of geometrical spaces (such as isotropy,
homogeneity, and three-dimensionality) could not be taken to apply to physical space
unless those properties were tested by certain criteria, probably also involving the
repetition offeatures. 47

Given Mach's claim that processes do not repeat themselves in time, basic
symmetries can be called into question to that small degree. However, if one were to
completely doubt repetition, conservation, and symmetry, practically no scientific
testing would be possible. Instead, we would be left with a sheer "experiencing" and
cataloging that stands at odds with scientific thought and might signify a return to the
primal chaos with which Mach believed every organism begins its mental life.

It is unclear whether Mach believed a science of pure experience was possible, or

whether he thought we were doomed to live with our own economical creations. He
suggests in the Mechanics 48 that in the best case of assuming infinite memories and an
ability to attend to all aspects of a phenomenon at once, science must be a mere
tabulation of experiences and should not involve laws at all. This idea is usually taken as
proof that Mach had succumbed to pure sense-data positivism, for which he was
criticized by Einstein, who in his 1922 Paris lectures called Mach's science "a catalogue
but not a system." 49 Moreover, a non repeating conception of science is not only
psychologically impossible, it almost fails to be science at all as we understand it. It
seems inevitable that if we are to practice scientific method, we must make abstraction
from the particulars of each individual event, even if it may still be true that no law
completely captures the peculiarities of each individual trial.


While in his earlier writings Mach was content to raise small economical doubts about
science, in his later years he seems to have envisioned a different natural science
altogether, one of "pure experience" in which individual, unique aspects of lived
experience are emphasized over its mnemonic, repeating forms. Mach closed his
economy lecture with the mysterious observation that physiological psychology would
light the way toward a future science:

Physiology, in a word, will reveal to us the true elements of the world. Physiological psychology
bears to physics in its widest sense a relation similar to that which chemistry bears to physics in its
narrowest sense. But far greater that the mutual support of physics and chemistry will be that which
natural science and psychology will render one another. And the results that shall spring from this
union will, in all likelihood, far outstrip those of the modem mechanical physics. 50

It was probably this underlying "physiological" view of nature that Mach was interested
in promoting, by calling the rest of science, from its Greek and seventeenth-century
foundations onward, an "economical abstraction" from events.

He seemed to have in mind a science of particular experiences that would emphasize

their quality and perhaps their uniqueness over and above the need for form. But as
Mach clearly recognized, science had proved impossible to do except by economical
means. In a sense, I think Mach was really looking ahead to a point still to be reached in
the development of knowledge, where the science of form had been perfected to such a
high degree that the individual differences suddenly become important, as surely they
would at some degree of precision.


Ernst Mach saw his position as part of a movement, including "a whole host of
philosophers-positivists, critical empiricists, adherents of the philosophy of
immanence, and certain scientists ... [who] have entered upon paths which ... converge
almost towards one point." Bertrand Russell, who stood at the end of that presaged
development, in turn claimed that Mach had "inaugurated the movement" toward
neutral monism. 1 Within the context of a neutral monist movement in Austria,
Britain, and America, some characteristic theses emerge such as the principle of
psychophysical parallelism, the necessity of a functional and causal presentation of
particulars or elements, sometimes events, the assumption of "unsensed sensibilia" in
matter and other minds and so forth.

The Muller-Helmholtz doctrine of specific energies was the essential

breakthrough to most of these developments. Sensation was treated in a naturalistic
way for the first time and expressed as the actual look of the physical energy of a
nerve or nerve fibre to its possessor. Meanwhile, natural science, through the work
of Joule and Mayer, developed a system of external relations between these energies.
In sensation one could guess that we catch the look of the energy that physics knows
only at a distance or through external relations. So an analogy could be extended
based upon the commonality of physical dimensions, viz., that what is related in the
physical and biological sciences is of the nature of sensation--even if of a radically
different quality in unobserved nature. As a work of the generation following Muller
and Helmholtz, Mach's position arose at the same time as the theory of reine
Erfahrung of Richard Avenarius and the "mind stuff' theory ofW.K. Clifford. Later,
the Analysis of Sensations itself led to the radical empiricism of William James and
to Russell's ambitious attempt to reconcile the theory with quantum physics and
relativity in a Principia Metaphysica.


Mach said that the affinity between his view and that of Richard Avenarius (1843-
1896) was "as great as can possibly be imagined where two writers have undergone a
different process of development, work in different fields and are completely
independent of one another. "2 Avenarius was professor of philosophy in Zurich, and
for many years a founding editor of the Vierteljahrsschrift for wissenschaftliche
Philosophie, which Mach later helped edit. Avenarius's "empirio-criticism" is
represented in a major work, Kritik der reinen Erfahrung (1888), the more readable
Der menschliche Weltbegrif.f(1891), 3 and a superb collection of articles all entitled

"Bemerkungen fiber den Gegenstand der Psychologie" (1894-95) for the

Vierteljahrsschrift. 4

Even in the title of the Kritik, Avenarius implicitly sets himself against Kant.
Where Kant found experience and sought transcendental foundations outside of
experience, such as the infamous Ding an sich, to explain how it was possible,
Avenarius boldly claims that he can accomplish everything Kant did, both a theory
of knowledge and being, all from within experience itself and employing only such
elements as are given. Avenarius thus begins with a totality of given elements
(colors, sounds, feelings, physical elements) and does not distinguish in kind among
them. In this fundamental order there is neither self nor outer objects, merely what
Avenarius called a more fundamental "third thing" (Ein Drittes).

Like Mach, Avenarius used the functional variations of these neutral elements
with one another to determine various subsystems within the totality. He separated
out the system E, consisting of psychical variations, and a system R, consisting of
the physical objects of the environment. Within R was a special system which
Avenarius called C, representing the physical elements of the central nervous
system. (This system C can itself be made part of the outside environment by
opening the brain to examination and by arraying a set of mirrors to bring it into the
visual field of its owner). 5

Within this overall framework, Avenarius isolates certain variations as bearing

out the physical connections between objects of the environment R and other
variations that establish a parallel correlation between psychical elements E and
elements of the central nervous system C, as studied by physiological psychology.
Thus, the whole apparatus of knowledge object, representation, and knower are all
constructions of the same pure experience.

One of the features Avenarius fmds immediately is that psychical states run in
parallel with their physiological correlates. This is contingent on experimental proof.
A physical R series consisting of R states runs in parallel with a psychical E series
consisting of E states. Many of these "vital series," as Avenarius calls them, are
triggered by a disturbance to the nervous system, with a concomitant feeling of
unpleasantness and a slow readjustment to a state of equilibrium and a parallel
feeling of relief.

Despite the closeness of the variations of the vital series between psychical and
physical events, Avenarius keeps these particulars separate, since they appear in an
environment as things external to one another. This led Avenarius to the
overstatement seized on by Lenin6 and other critics7 that "the brain is not the
dwelling-place, seat or producer of thought; it is not the instrument or organ, it is not
the vehicle or substratum of thought. .. Thought is not even a physiological function
of the brain," to which Mach replied that he was "not willing to subscribe to
everything Avenarius says or any interpretation of what he says, but still his
conception seems to approximate very nearly to my own. "8

Mach also separated the particulars of green colors and brain processes but did not
deny the tightest of functional relations between them, as Avenarius seems to. (What
Avenarius may really have meant was that the functional relation could in principle
be a different one, or that the substrate to mental events could be different from a
human brain.)

In his rnenschlichen Weltbegriff, Avenarius introduced and refuted a certain

fallacious argument he called the error of "introjection."9 In the development of
experience, he says, we come to view people like us in the proximity to objects. We
see our friend looking at a tree, for example, observing him perhaps down to the
level of his brain tissue. Because we see the tree as a colored object with green
leaves, Avenarius says, we assume that our friend must see something similar and
"introject" a little colored image somewhere into his brain. Yet, of course, if we
actually investigated the brain we would find nothing of the sort, only the
physiologico-chemical concomitants of the sensation. However, once the
introjectionist argument is made and believed, Avenarius says, we then tum it on
ourselves, assuming that the colored images in our own environment are nothing but
processes somewhere in our brains and not physical objects. This leads to the
abandoning of the realist standpoint and the adoption of the idea that all of the
elements are just processes in our ego. In particular, it leads to a representative
theory of perception and reference whereby an image in the mind is correlated to an
object completely outside and lacking any correlating link to it. 10

Mach commented on Avenarius's argument in the Analysis. He himself did not

think that the interpretation of other people and their perceptions was of such
fundamental importance. Rather, he urged, the differentiation of the ego and
environment could be done by one individual, Robinson Crusoe style, against the
surrounding objects. 11 Then, he says, it would not be necessary to exclude
introjection again. I imagine he meant that from the standpoint of the first division of
ego and environment in one consciousness it would already be self-evident that the
objects of the environment lie outside the ego, as, for example, he urged that the
sensations appear plainly "outside" the head and not inside. 12

Avenarius did not explicitly address the issue of world elements in matter or in
animal consciousness. For him, when we look over experience as a whole, we take
the elements as directly apprehended realities. We cannot restrict the whole totality
of elements to a delimited ego, or Absolute I, since a restriction is only the isolation
of certain variations within experience. But if we were to restrict the totality itself to
an ego, as the solipsist wants to do, other elements would have to be found outside
the ego against which to limit it (which would entail admitting the very mind-
independent elements the solipsist wants to exclude).


One writer who was not agnostic about world elements independent of the human
mind was W.K. Clifford, a marvelously eccentric English mathematician, whom
Mach credits with "an extremely close affmity to myself in the direction of his
thought." Clifford presented his doctrine of the "mind stuff'' in an 1878 article called
"On the Nature of Things in Themselves." 13 Clifford had arrived at his ideas with the
help of Muller's and Helmholtz's specific energies doctrine and by consideration of
the progress that the theory of sensation and perception had made in Germany. 14 He
had also read Mach's Conservation of Energy and was a defender of Machian views
in England. 15

Clifford begins his article with the familiar distinction of two orders, one
mental, one physical, into which the same elements may enter:

There is the internal or subjective order, in which sorrow succeeds the hearing of bad
news ... and there is the external or objective order in which the sensation of letting go is
followed by the sight of a falling object and the sound of its fall. 16

He then makes a distinction between these objects and what he calls "ejects" or
elements that are not experienced by me at all. The sensations of another
consciousness, for example, are annexed to the changes in the brain tissue of the
person who is exhibited as a physical object in my own consciousness. Another eject
is the actual thing in itself external to my mind.

Clifford says that, due to the work of Muller, Helmholtz, and Wundt, which he
had studied, it may be assumed that the ejective sensations of another consciousness
parallel the changes we observe in the subject's brain, 17 not the peripheral nerves,
and that this is an empirical fact. The parallel between brain tissue facts and
psychical facts, Clifford says, is one of structure or function, giving the mathematical
metaphor: "We should say a sentence spoken is the same function of the elementary
sounds as the same sentence written is of the corresponding letters." 18 The parallel
extends thoroughly to all details of the elements, and one can say that "to a simple
feeling corresponds a simple change of nerve matter." 19

Clifford makes a subtle point for assuming the eject, however, and that has to do
with what he dubs the "social object." What we perceive is not our own individual
object or private sensations, he says, but a complex that is the same for all observers,
"a table as an object in the minds of men." This social object thus has the missing
elements added in thought with the missing unobserved parts or perspectives given
to others filled in:

Now it is probable that the individual object, as such, never exists in the minds of men. For
there is every reason to believe that we were gregarious animals before we became men
properly so called. And a belief in the eject-some sort of kindred consciousness in one's

fellow beings-is clearly a condition of gregarious action among animals so highly developed
as to be called conscious at all. Language, even in its first beginnings, is impossible without
that belief and any sound which, becoming a sign to my neighbor, becomes thereby a mark to
myself, must by the nature of the case be a mark of the social object and not of the individual
object. 20

I do see a strong similarity between Clifford's social object and Mach's idea of
piecing out "partially observed facts in thought,'m by adding images to present
sensations. 22 When my friend and I observe a statue, I automatically fill in missing
elements in the back of it where my friend is situated. It is thus the social object I
mean when we talk about it and not my particular perspective of a backless statue.
The same goes for objects which existed some time ago, such as a house I once
inhabited that has long since been destroyed. Here a memory image serves in the
place of sensations that can no longer be had by anyone.

Clifford now makes his case for the existence of ejects from consciousness,
which is handled with great cleverness. Like Mach and A venarius, Clifford says that
for a sensation to be part of the mental order, that is, part of consciousness, it must
be linked externally to any number of psychical relations, in particular relations of
memory and association with mental images. These ties are made after the fact of a
sensation's appearance or perhaps even before if the sensation is anticipated, but in
any case they require some form of time orderings. But in the instant in which it
appears, the raw sensation has no such ties, and may be taken as an instantaneous
quality that is already part of a world external to consciousness:

A feeling at the instant when it exists, exists an undfor sich and not as my feeling; but when on
reflection I remember it as my feeling, there comes up not merely a faint repetition of the
feeling, but inextricably connected with it a whole set of connections with the general stream of
my consciousness. This memory, again, qua memory, is relative to the past feeling which it
partially recalls; but in so far as it is itself a feeling, it is absolute, Ding an sich.Z3

Clifford seems to be saying that my elementary feelings, of which I am

unaware, are already in some sense ejects and belong outside in the world-indeed
like Leibniz's petites perceptions, these apparently simple sense qualities may
decompose into even tinier unfelt elements-" of which the simplest feeling is built
up." These tiniest unfelt elements Clifford calls "mind-stuff. "24 Consciousness itself,
Clifford says, is nothing but a complex of these essentially ejective facts: "Reason,
intelligence and volition are properties of a complex which is made up of elements
themselves not rational, not intelligent, not conscious. "25 Clifford concludes from the
fact that we can gather from raw sensation a fragment of the object as it is an sich,
we may conclude that on its side the nature of the object in itself is at least partly
composed of mind-stuff.

Clifford makes an interesting remark that Kant "threw out a suggestion that the
Ding an sich might be of the nature of mind. "26 Clifford does not cite any passage
but I believe he means the one in the schematism section of the Kritik der reinen
Vernunfl where Kant pointed out that "as time is only the form of intuition, and thus

of objects, as appearances, so is that which corresponds to sensation in them, the

transcendental material of things in themselves. "27 Kant also says that what he calls
sensations an sich, raw intensive magnitudes, are neither in space nor time28 but
instantaneous in the same sense as Clifford's feelings an sich. Sensations an sich do
not have the schematized relations to mental images of the imagination and
associations which become evident upon later reflection and memory, and thus are
not ordered into objects. They are, as on Clifford's view, preconscious raw material
waiting to be organized into space and time series. Thus Kant does indeed seem to
suggest that sensations an sich really are parts of the non spatial, non temporal things
in themselves external to consciousness?9

But Clifford is still not finished. He says that we may even infer something as to
the structural makeup of the thing in itself. 30 The argument is an analogy in the
mathematical sense with the structure A:B :: B:C. Clifford starts by pointing out that
when we observe, for example, a man looking at a candlestick, observing even the
states of his brain, these can all be made states within my own consciousness as well
as being parts of immanent physical objects. Hence, Clifford says, because the
candlestick can be both my percept and a piece of the physical object affecting my
friend's brain, it can serve as the middle term of the following analogy. As the state
of the friend's brain (taken as physical object) is to the candlestick (taken as physical
object): so is the candlestick (taken this time as my sensation) to the unobserved
candlestick an sich (which causes my sensation). The unseen Ding an sich plays the
same structural role with respect to my sensation of the candlestick as the candlestick
plays to its image in my friend's brain. 31

The most interesting case of such analogies considered by Clifford is one we

observe in the nervous activity of less complex organisms and annex to them
sensations similar to our own. Here the unknown mind is the thing in itself we are
trying to divine. And here we are making analogy not based on the physical
variations in our own experiences but on the mental ones. Clifford claims the theory
of evolution requires us to assume a steady line of descent of conscious activity
(memory, association) from lower animals to simpler organisms and fmally
inorganic matter:

Consciousness is a complex of ejective facts~f elementary feelings or rather of those remoter

elements which cannot even be felt but of which the simplest feeling is built up. Such
elementary ejective facts go along with the action of every organism, however simple; but it is
only when the material organism has reached a certain complexity of nervous structure (not
now to be specified) that the complex of ejective facts reaches that mode of completion which
is called Consciousness. But as the line of ascent is unbroken, and must end at last in inorganic
matter, we have no choice but to admit that every motion of matter is simultaneous with some
ejective fact or event which might be part of a consciousness. 32

Thus, taking his views as a whole, Clifford held that "the reality external to our
minds which is represented in our minds as matter, is in itself, mind stuff." In various
sorts of functional relationships mind-stuff makes up both the physical objects of our

perceived environment as well as the reality behind these objects, organic creatures,
and the elements of their consciousness:

A moving molecule of inorganic matter does not possess mind, or consciousness; but it
possesses a small piece of mind-stuff. When molecules are so combined together as to form the
film on the underside of a jelly fish, the elements of mind stuff which go along with them are
so combined as to form the faint beginnings of sentience. 33

Here is essentially the same familiar parallelism of physical and psychical

elements to be found in Mach and A venarius. Clifford still appears to consider the
physical world to be made up of matter and motion with mind-stuff running parallel
to it, but it is fairly sure that he also intended for the matter and motion themselves to
be made up of sense like qualities. However, there is no indication in Clifford how
such quantitative features of molecular matter and motions are to be reduced to any
aspatial qualitative level.


Two correspondents and friends of Mach, the German-American editor of the Monist
Paul Carus in 1890, and the Austro-Marxist scientist Friedrich Adler in 1910 wrote
him about Clifford's article, asking if he too held the position that unsensed elements
could be assumed by analogy in animals and in matter. 34 Mach responded positively
to Adler, saying that the hypothesis of elements of other minds, animal
consciousness and "inorganic bodies" served to "round off the world view
provisionally and in hope of the future construction of biology." He then added: "I
have not cultivated all of these matters, for I always feared the nearness of the
metaphysical abyss, where there is no experiential foundation." Mach added his
dislike for Clifford's conception of elements as "things in themselves," as for Mach
this would have meant something that is not even possibly in a causal connection to
the senses and therefore completely unobservable. One would think that Clifford's
implication that a feeling could occur outside a complex, seemingly unconnected to
other elements would have conflicted with Mach's views. But Clifford did not seem
to believe in loose collections of particulars, a la Hume, the so-called Spencerian
"mind-dust," but rather in things embedded in functions and thus complexes.

Carus expressed to Mach a criticism of Clifford's conception that the universe

consisted of mind-stuff, saying that it was a swallowing of the object by the subject,
"as the nine lean cows devoured the fat in Pharaoh's dream." 35 This would later be a
criticism made by others of Mach himself, and of James and Russell. In an 1893
article in the Monist, cited earlier, Carus took back this view after realizing that
Mach understood by sensation "reality itself," not just a subjective side of it. 36


Mach's influence on the American psychologist and philosopher William James was
direct as could be. James and Mach worked in some of the same areas of sense
physiology and were thus familiar with each other's writings in a technical field long
before they became acquainted with each other's philosophical views. Joachim
Thiele reports that their correspondence probably began in the middle 1870s, when
Mach was occupied with the sensations of movement-specifically, the sensations of
innervation in which James did not believe. In 1882 James visited Prague, where he
met Hering, Stumpf, and Mach, recording his impressions in a letter to his wife:

Mach came to my hotel and I spent four hours walking and supping with him at his club, an
unforgettable conversation. I don't think anyone ever gave me so strong an impression of pure
intellectual genius. He apparently has read everything and thought about everything, and has an
absolute simplicity of manner and winningness of smile when his face lights up that are
charming. 37

Mach's impression of the meeting were conveyed in a letter of 1911 (in German) to
Anton Thomsen:

My personal memories of William James are very pleasant; he visited me while still in Prague
in 80 or 81. I remember no one with whom, despite the divergence of viewpoints, I could
discuss so well and fruitfully. He opposed me almost everywhere and yet I benefited almost
everywhere by his objections. Already at that time he avoided any drop of wine or coffee so
that I believed him more of a nervous hypochondriac than a really sick man. The center of his
work certainly lies in his excellent Psychology. I cannot quite come to terms with his
Pragmatism. "We cannot give up the concept of God because it promises too much." That is a
rather dangerous argument. 38

Mach, despite his own association with Fechner's pantheism, seems to have
interpreted James's Pragmatism as an attempt to squeeze in "Spiritualismus und
Schwiirmerei"39 alongside science. Whatever disagreements they may have had their
correspondence is courtly and mutually complimentary. Poignant are the letters the
two men exchanged after Mach's stroke and the beginnings of James's heart disease.
James wrote to Mach: "Les deux grandes restes consolaient entre eux" (one great
ruin consoles another). 40

James certainly read the first edition of the Analysis in 1886, which he much
admired. He also read the Mechanik, the Wiirmelehre, and the Popular Scientific
Lectures, which Mach dedicated to him. These books clearly had great influence, for
as James wrote Mach in 1902 about the development ofhis own views:

I am now trying to build up before my students a sort of elementary description of the

construction of the world as built up out of 'pure experiences' related to each other in various
ways, which are also defmite experiences in their turn. There is no logical difficulty in such a
description to my mind but the genetic questions concerning it are hard to answer. I wish you
could hear how frequently your name gets mentioned and your books referred to. 41

The fruits of these lectures were published as "Does 'Consciousness' Exist?" (1904),
"A World of Pure Experience" (1904), and ensuing articles in the Journal of
Philosophy, Psychology and Scientific Methods. 42

"Does Consciousness Exist?" covers much of the same ground as the Analysis of
Sensations twenty years earlier, i.e., the substitution of a functional connection for
the ego and the introduction of a neutral stuff which is the common constituent of
mental and physical things, which differ only in their variations. What is different is
James's concentration on the theory of knowledge:

My thesis is that if we start with the supposition that there is only one primal stuff or material
in the world, a stuff of which everything is composed, and if we call that stuff 'pure
experience,' then knowing can easily be explained as a particular sort of relation towards one
another into which portions of pure experience may enter. The relation itself is a part of pure
experience; one of the terms becomes the subject or bearer of the knowledge, the knower, the
other becomes the object known. 43

Like Mach, Avenarius, and Clifford, James allows pure experience to participate
in a mental ordering by psychological relations and a physical ordering by what he
calls "energetic" relations. 44 The same room in which one sits, James says, can be
part of the train of mental experiences and associations as well as parts of the history
of the room itself which existed before the observer. 45 James uses the fact that just as
a point may be located at the intersection of two different lines, so too can the very
same sensation partake of mental and physical histories. James seizes on this idea
that objects can be directly known merely by emphasizing now their psychical
variations, now their physical ones. James offers the new theory as a replacement for
the representative theory of knowing, the new theory being based squarely on the
identity of subject and object rather than on the classic three-way between an object
an sich, an intermediary image and the ego an sich:

If the reader will take his own experiences, he will see what I mean. Let him begin with a
perceptual experience, the 'presentation' so-called of a physical object, his actual field of
vision, the room he sits in, with the book he is reading as its centre; and let him for the present
treat this complex object in the common-sense way as being 'really' what it seems to be,
namely, a collection of physical things cut out from an environing world of other physical
things ... Now at the same time it is those self-same things which his mind, as we say,
perceives; and the whole of philosophy of perception from Democritus's time downwards has
been just one long wrangle over the paradox that what is evidently one reality should be in two
places at once, both in outer space and in a person's mind. 'Representative' theories of
perception avoid the logical paradox, but on the other hand they violate the reader's sense of
life, which knows no intervening mental image but seems to see the room and the book
immediately just as they physically exist. .u;

The solution, James says, is to regard the identical thing, the room, the book, as
part of two orders, two "takings," the first as a real object with purely physical
relations to other objects-both sensed and unsensed-and as part of knowledge and
accompanied by memories, images and associations that relate the object to the
elements of mental life.

James also resolves the elements of ideas or mental representations (called by Mach
the a,~,y ... elements) into physical objects. He speaks, for example, of being unable
to tell the difference between a real fire and a mental fire, except for the fact that real
fire will burn sticks, whereas mental fire may not. 47 This direct realism would
become the source of many misunderstandings and problems for James's readers.

The problem is one all the neutral monists had to face. If sensations had both
mental and physical variations, why not the ideas, memories, and other so-called
mental phenomena? Mach had addressed the same problem in the Analysis and he
too held that the a,~,y ... had reality an sich. Despite the "paleness" of ideas, their
equal claim to reality in dreams and sensory hallucinations serves as a proof of this.
But as for their physical variations, Mach seems to have taken them more at face
value than James, indicating an independent life of the sense organs, which need not
generate objects like fires, faces, or buildings, but are merely the specific energies of
central states considered as the blobs, flashes and outlines they are. For James a
mental fire is already recognizable as a fire, it even feels hot and so forth. But it is
doubtful that the blobs and colors of the mental image really look anything like a
fire, unless reference is made to external ways of determining what those shapes and
colors usually look like when sensations are pieced out into the rest of an object. In
themselves they look more like pure plays oflight and sound.

There is, of course, a very real, objective use of the imagination and memory in
experience, remarked on by Mach, which was mentioned above in connection with
the social object. Even when sensations are in play, they require mental images and
thought to fill out what is only "a partially observed fact" into genuine experience.48
As Mach says of Galileo,49 he was able to describe a falling body in a parabola by
adding to its present state other non present but imagined previous and future states.
Here it is very clear that there are objective, physical variations of the a,~,y...when
they form part of perceived objects.

In his "World of Pure Experience," James lays out his position more explicitly,
calling special attention to his belief in the reality of relations connecting bits of pure
experience. It is the belief in real, felt relations that James says distinguishes his
brand of empiricism from the British form. Similar to Mach's Spinozist belief in a
"whole of nature," James says that he, too, accepts a "universal conjunctive relation"
of all with all-involving "nothing whatever as to farther consequences."50 Within
this overall unity James allows for more specific conjunctive and disjunctive
(dissociating) relations, continuous transitions between experiences as well as
discrete ones-selves for James are continuous "streams of thought," whereas the
transition from one self to another is discrete. 51 Like Mach, James distances himself
from rationalists who would make experienced relations a priori forms of

Of all the felt relations, James was most interested in the "cognitive relation" of
knowing, which he divides into three sorts that may hold between two bits of pure
experience; either they are:
(1) the self-same piece of experience take twice over in different contexts;
(2) two pieces of actual experience belonging to the same subject with defmite
tracts of conjunctive transitional experience between them, or
(3) the known is a possible experience either of that subject or another, to which
the said conjunctive transitions would lead if sufficiently prolonged. 52

James divides these classes into knowledge by acquaintance in the case of (1)
and 'knowledge about' or knowledge by description in (2) and (3). However, James
is careful in the case of (2) and (3) to specify that, unlike (1) where the knowing
relation is more or less internal in the sense that the object is the same in two
"takings," in the other two cases knowing is strictly an external relation that may
turn out to lead to other bits of pure experience that are not the ones expected (error)
or to ones that serve to lead closely enough to what is desired or expected as to count
as success (truth). The relation between the bits of pure experience that serve as the
mental representations of the knower and those that serve as the elements known
must actually be run through and established as fact. The test here, James insists,
cannot be any prior resemblance or similarity between the mental image of an object
such as a house and its fulfillment in the perception of the house itself, but rather
whether certain operations can be carried out to actually establish the wanted
external connection between mental images and certain future sensations. No matter
how similar the image of the house in the mind and the actual house might seem to
be, the one has no intrinsic power to represent the other: 53

Suppose me to be sitting here in my library at Cambridge, at ten minutes' walk from

'Memorial Hall' and to be thinking truly of the latter object. My mind may have before it only
the name, or it may have a clear image, or it may have a very dim image of the hall, but such
intrinsic differences in the image make no difference in its cognitive function. Certain extrinsic
phenomena, special experiences of conjunction, are what impart to the image, be it what it
may, its knowing office.

For instance, if you ask me what hall I mean by my image and I can tell you nothing; or if I fail
to point or lead you toward the Harvard Delta; or if, being led by you, I am uncertain whether
the Hall I see be what I had in mind or not; you would rightly deny that I had 'meant' that
particular hall at all, even though my mental image might to some degree have resembled it.
The resemblance would count in that case as coincidental merely, for all sorts of things of a
kind resemble one another in this world without being held for that reason to take cognizance
of one another. 54

Mach and James treat concepts similarly, as being not general ideas or mental
presentations at all, or even words, for these cannot possibly include the vast set of
operations hypothetically needed to establish the connection between our
fragmentary thoughts and mental presentations and the public, physical experiences
that would fulfill them. As Mach insisted:

A concept is never simply a completed presentation. In using a word denoting a concept there
is nothing involved in the word but a simple impulse to perform some familiar sensory
operation, as the result of which a definite sensational element (the mark of the concept) is
obtained. 55

Mach also said that a concept is the result of "completing the sensations in
thought" by supplying the parts that are missing from a completed object. 56 Mach
gives the example that most chemists can reliably identify sodium on sight by its
color and softness, 57 but the scientific reference to sodium cannot be made on that
basis but only when a complex of chemical operations are performed. 58 As Mach
understood concepts, they are only really possessed by either specialists who have
taken the trouble to internalize all of these operations in thought, or they may be
possessed socially through the division of labor by multiple specialists. He says that
even words are no mark of real conceptual mastery:

I am of the opinion that the nature of concepts is necessarily much more clearly displayed in
scientific concepts, which have been consciously formed and applied, than in vulgar concepts.
The latter are so vague that they can scarcely be reckoned as proper concepts at all. The words
of ordinary speech are simply familiar signs, which occasion equally familiar habits of thought.
The conceptual content of such words, in so far as it has any form at all, is scarcely present to
consciousness ... 59

Mach gave further views on the relation of language and concepts in two
chapters of his Theory of Heat, "Language" and "The Concept," published in 1896
and translated in the American Open Court journal. Here he remarked that in essence
words are meaningless when they are not embedded in their customary background
of concrete extralinguistic facts:

Even in the most highly developed human languages it happens that the full meaning of some
utterance is determined entirely by the situation; while it is well known that languages in a low
stage of development very frequently have to have recourse to gestures to be understood, so
that when spoken in the dark they are partly unintelligible 60

Mach also regarded it as an exaggeration that language was indispensable for

thought and remarked that some operations may proceed wordlessly between ideas,
desires, or operations in thought and physical activity without any but the most
fleeting verbal accompaniments. 61 Mach considered language as just one in a
continuum of human prelinguistic conceptual abilities beginning with our senses; it
has no overarching normative or conceptual role to play vis-a-vis other disciplines
except as a thought-saving economical help. 62

Mach's and James's views have the interesting consequence (to be considered
more fully later) that very little of the traffic that travels through a human ego,
whether fragmentary half-thoughts, babble-like words, or images, qualifies as
conceptual knowledge. A figure like Newton may have had knowledge in that he
possessed both the concept of force and the most authoritative public means at the
time of demonstrating it, But such cases are rare. The puzzle is how all of these little
fragments in the minds of people who know practically nothing end up being
combined into knowledge held in trust by a community of such people.

The situation, Mach says, is exactly analogous to economic relations, where people
who could not possibly have amassed fortunes through individual effort are able to
do so by taking advantage and leveraging the labor of others:

Just as a single human being, restricted wholly to the fruits of his own labor, could never amass
a fortune, but on the contrary the accumulation of the labor of many men in the hands of one is
the foundation of wealth and power, so, also no knowledge worthy of the name can be gathered
up in a single human mind limited to the span of a human life and gifted only with finite
powers, except by the most exquisite economy of thought and by the careful amassing of the
economically ordered experience of thousands of co-workers. 63

Now, James does address one of the problems of his view, namely that it is hard
to see when the terminus of the chain of experiences can be reached to validate a
concept. According to Mach the chemist with his sodium reaches an answer when he
has conducted a battery of tests to determine the elemental properties. But how many
tests will be needed? One might object that it is not possible to test for the
application of a concept by pragmatic or positive means unless one knows what one
is testing for, or what will count for passing the test once it is conducted. Thus it
would seem that the representation of sodium that leads to the testing must have
some kind of "internal relation" to the results already in the sense of it. (As we will
see below, Bertrand Russell made such an argument against the theory.) James more
or less confesses that knowledge is "never nailed down." 64 We rather have on
Mach's and James's view a negative knowledge that extends not to terminal
experiences (which would have to be included in the definition of the term), but only
so far as the process of testing is not stopped somewhere by experience, an endless
chain. This has the effect of leaving concepts underdetermined by data. Yet both
Mach and James accepted these fallibilist consequences, as neither one believed in
knowledge pure and complete, anyway. Mach and James also had their doctrines of
economy of thought and the pragmatic method to avoid an infinite proliferation of
empirically equivalent concepts, since where the empirical consequences turned out
the same, or nearly the same, the more economical concept may be preferred.

Yet, as James realized, one case of knowledge is problematic even for the
pragmatic and economical view: namely those terminal experiences that lie outside
the ego of the knower-the sensations or ideas of another's ego or the constituent
elements of objects unobserved by anyone (case 3). Before tackling the existence of
such elements, James does essentially what Mach did with his "spatial identities,"
arguing that the variations of unseen elements might be inferred from those elements
that are present to a single person:

Where direct acquaintance is lacking, 'knowledge about' is the next best thing, and an
acquaintance with what actually lies about the object, and is most closely related to it, puts
such knowledge within our grasp. Ether waves and your anger, for example, are things in
which my thoughts will never perceptually terminate, but my concepts of them lead me to their
very brink, to the chromatic fringes and to the hurtful words and deeds which are their really
next effects. 65

It seems odd that James, who was so willing in one context to rule out the human
ego as the medium of knowledge, is so hard-pressed to keep near terminal
experiences within the ego here. James was far less direct than Mach, Clifford, or
Russell about the question of world elements, invoked to account for other minds
and objects unobserved. Eventually, of course, he does assume them, but after his
own fashion. He says, for example, that the mind of another consists of elements
inaccessible to me but in parallel causal relations to my sensations, as might occur in
my observation of the last nerve processes with which his sensations go:

In that perceptual part of my universe which I call your body, your mind and my mind meet
and may be called coterminous. Your mind actuates that body and mine sees it; my thoughts
pass into it as into their harmonious cognitive fulfillment; your emotions and volitions pass into
it as causes into their effects. 66


James also suggests, like Clifford, that objects be regarded as made of the same stuff
as minds, and he gives a new example of objects being found at the intersection of
several minds, with the vivid illustration of spatial perspectives. The same Memorial
Hall, James says, can be located at the spatial intersection of where it is located in all
of the individual mental perspectives:

If I ask where some object of yours is, our old Memorial Hall for example, you point to my
Memorial Hall with your hand which I see67 All of the relations whether geometrical or
causal of the Hall originate or terminate in that spot wherein our hands meet. 68

James calls this common meeting point "space" and holds that all observers
identify the same point in space as the location of the object. Observer A sees his
hand point straight ahead to Memorial Hall and sees Observer B's hand point at an
angle but still terminating at the hall. Observer B sees his own hand pointing straight
at the hall and sees A's hand at an angle but still terminating there. James says that
all the hands meet in the same space69 when what is really meant is not some
individual's space but a kind of super-space in which the private psychological
spaces and their perspectival differences are projections of a more general set of the
same particulars grouped instead by their causal laws. James comes up with what
Russell later called a causal "perspective."

James does not bother with the important fact stressed by Avenarius that the
images seemingly located in front of my eyes also have causal relations with events
behind my eyes in my brain. Thus the perspectives really terminate in two locations.
This second causal locus of the object appears in another person's space (or even my
own) at somewhere other than where he (or I) see Memorial Hall. Of course James
might reply that my friend and I also happen agree on the spot in my brain when our
fingers point there.

The problem is that this point in his brain will not affect my percept of Memorial
Hall, only his, and hence I must conclude inevitably that there are certain causal
locations of the hall different for him than for me.

James has a second problem in accounting for the difference in quality between
his perception of Memorial Hall and mine, given the fact that I may be color-blind
and he may not be. One quality and even one spatial perspective cannot be shared by
two minds. Thus, in order for James to say that these sensations in different minds or
causal perspectives belong to the same object, he has to say that the object is a
causal or perhaps (echoing Fechner) a functional unity of all of these qualities. That
causal unity has to be taken across the different observers and bundled into a
coherent trans-subjective object with the many perceptions as so many different
perspectives of it. The object is a common function of all individual percepts and
perspectives identified by its causal relations. The various trans-subjective
differences, such as the fact that my hall is in black and white and his is in color, or
the perspectival fact that my Memorial Hall is occluded by a tree, may be attributed
to the psychical variations of the object more causally connected with my sense
organs than with the physical variations on which I and other observers will agree.
These variations will have to be separated out. But James tries to make this go down
easier by pointing out that we already do this when we recognize diachronic objects
across times. The individual temporally evanescent aspects of a pen are different
even for the same ego, but the stages are recognized as the continuation of the same
object, "energetically" related to each other.

Now on unsensed sensations, such as those aspects around the hall where no
mind is, James was forced into a comer. These perspectives call for the world
elements of Mach or the unsensed sensibilia of Russell, but due to James's own
distaste for the Hegelian Geisterreich, he held aloof from the possibility of mind-
stuff that is gathered by similar causal laws but not into minds. However, after
considering the fact that sensations may have their own properties of"compounding"
and that Fechner could have been right in recommending the hypothesis of a matter
similar to sensation, James says:

We have now reached a point of view from which the self-compounding of mind in its smaller
and more accessible portions seems a certain fact, and in which the speculative assumption of a
similar but wider compounding in remoter regions must be reckoned with as a legitimate
hypothesis. The absolute is not the impossible being I once thought it. Mental facts do function
both singly and together at once, and we finite minds may simultaneously be co-conscious with
one another in a super-human intelligence. It is only the extravagant claims of coercive
necessity on the absolute's part that have to be denied by a priori logic. As an hypothesis trying
to make itself probable on analogical and inductive grounds, the absolute is entitled to a patient
hearing. 70

So, paradoxically enough, Mach and James in pursuing their neutral monism
forward into new reaches, both look back toward their point of departure with
Fechner. In their correspondence, Mach and James seem to have revised their own
estimation of Fechner's work as they came to wrestle with the same problems. 71 In
addition, the figure of Hegel appears in their correspondence (as James, was the first

to give Mach an understanding of the German philosopher). Mach was quite

interested in Hegel and made several efforts to understand him, to be sure only
through secondary works. 72 And yet both Mach and James were verging on
accepting one of Hegel's leading ideas, namely, that in the last analysis the agency
of knowledge was not the individual human ego at all, but a broadened Absolute,
through which the individual mind participates as a fragment. 73 Mach was quite
willing to regard knowledge as held socially, and all of his talk about a "broadened
ego," 74 extending even to class-consciousness and species-memory, seems to point
toward insights Hegel had promulgated one hundred years before. Mach seems to
have come to such ideas through his milieu or by himself.


This brings us to another ex-Hegelian, Bertrand Russell, who developed his own
neutral monist view from Mach and James. 75 But Russell started out as an opponent
of the view he eventually adopted. In his three articles for the Monist, "On the Nature
of Acquaintance I, II, III" later gathered into his unpublished 1914 "Theory of
Knowledge," 76 Russell accepted as a "service to philosophy" Mach's and James's
idea that "what is experienced may itself be part of the physical world and often is
so" 77 or that "constituents of the physical world can be immediately present to me." 78
What Russell could not abide in 1914 was the notion that other elements of mental
phenomena (ideas, memories, mental images, imaginings, and perhaps also
emotions) were also indifferently physical objects. His main problem is with the
view that these objects of thought such as images are realities in their own right,
specifically James's overbaked claim that mental fire as an object is indistinguishable
from a real fire. Russell employs an aspect of his theory of descriptions to combat
James, saying that the collection of images, memories, and representations making
up a belief about Memorial Hall cannot itself be the object Memorial Hall in a
realistic construal of these psychical elements. If this were the case, Russell says,
then even false beliefs could become the real objects corresponding to them:

This, which can be believed or disbelieved, is quite different from the actual entity (if any)
which does answer to the description. Thus the matter of belief, is in all cases, different in kind
from the matter of sensation or presentation, and error is in no way analogous to
hallucination. 79

Russell also claims that in the case mentioned by James where mental beliefs or
wishes lead by means of external relations to the real object answering to them, those
elements of a description must have an internal representing relation, a "logical
relation" even, to the object they represent "between what is believed in the earlier
stages and what is experienced in the fulfilment." 80 To believe that Memorial Hall is
reached by turning right, then left, then walking 200 yards is to have a representation
of psychical elements such that "there is (at most) one building called Memorial Hall
and it is to be reached from here by turning right then left ... etc." 81 Russell later said
he was at this time under the influence of the Austrian philosopher and psychologist

Franz Brentano who had maintained in a famous passage from his Psychologie vom
empirischen Standpunkt that all mental representations had internal, intentional
relations to an object. For Brentano, this thesis represented the incontrovertible logic
of belief and not just a psychological feature of human thinking.

Mach had also encountered the "intentionality" objection from Brentano's

student Antony Marty,82 who claimed that the tone and the hearing of a tone are two
different things. The tone quality itself is a content which may be either physical or
mental, but the hearing is an irreducibly psychical act with an internal reference to
that content whether it is present, remembered, or imagined. Mach's answer to
Marty, if he gave one, is not known to me, but I suspect that he could have pointed to
his own use of functions in the explanation of psychical life. The function a tone
enters into is not the tone, but a complex of it with other elements, such as memories
or imaginings. A function, being an external connection between elements, is
already a complex; and thus, if only certain parts of the complex are presented as
sensations, the rest may well be imagined as annexed to it and an intentional object
may be just one of these parts pieced out in the functional complex. One could even
say that assuming the fact of the complex, the missing elements can be added in
thought, hinzugedacht, as Mach said. But since every element belongs to a complex
both in psychical life and physical compounds,83 there is no need to attribute this fact
to a special power belonging only to the psychical realm.

James's related point was that psychical elements do not have the intrinsic
power to represent anything until they are brought into that connection externally. A
mental image of Memorial Hall, for James, is not that hall and does not represent
that object until an external relation is set up and carried through in thought or in
fact. Thus, when James and Mach interpret psychical objects as real existences, it
cannot be as objects of belief that may occur in a description, for their descriptive
office requires an external relation, but rather purely as the faint outlines, shapes, or
blobs that represent central energies an sich. In this physical interpretation, not only
are a collection of blobs not a hall, but such phenomena are not part of a judgment or
complex, have no reference beyond themselves, and in that sense refer neither truly
nor falsely.

Now, Russell also attacks James's notion that knowing is an external relation of
leading from psychical to physical elements. To use Russell's example, if I am
looking for my dog and believe he is in the park, I go there but on the wal fall down
into a coal cellar by accident, where the dog, by chance, has also fallen. 8 James and
Mach could handle cases like this with the economy of thought or the pragmatic
technique of emphasizing only those variations that are of use, leaving the residue of
chance connections as the background against which useful beliefs can be isolated.
That some of these beliefs do lead to error is far from a problem for this view, as
Mach wrote:

Knowledge and error flow from the same mental sources, only success can tell the one from the
other. A clearly recognized error, by way of corrective, can benefit knowledge just as a positive
piece of knowledge can. 85

For all his focus on knowledge by description and internal relations, Russell
finally reveals that he does not place much value on this sort of knowledge anyway,
for he holds in 1914 that the most immediate case of knowledge is one in which
Memorial Hall is simply perceived. In this case, Russell was willing to say that one
can have knowledge of the hall "even ifl do not know that that is its name, and even
if I make no propositions about it, I must be said to know it in some sense more
fundamental than any which can be constituted by the belief in true propositions
describing it." 86 Russell's own example comes very close to conflating knowledge
with mere being, which is in essence what neutral monism was also striving for, but
without the mind-matter dualism.


Russell later converted fully to neutral monism in the Analysis of Mind (1921),
which he begins by saying:

The stuff of which the world of our experience is composed is, in my belief, neither mind nor
matter, but something more primitive than either. Both mind and matter seem to be composite,
and the stuff of which they are compounded lies in a sense between the two, in a sense above
them both like a common ancestor. 87

Russell accepted the neutrality of sensation between psychical and physical

variations, but held that the mental images and other such phenomena obeyed
psychical laws only, claiming "images belong only to the mental world, while those
occurrences (if any) which do not form part of any experience belong only to the
physical world. 88 In particular, Russell says that mental phenomena are separated
from physical ones by laws of association and memory. He makes a point, however,
of not including intentionality, or the internal relation to an object, among the
properties of mental states. 89 But Russell is in the camp of those who believe that
mental and physical laws are somehow incompatible orders-sometimes called a
"type dualist." By contrast, Mach, Clifford, and Avenarius were committed to a
point-by-point parallelism between the realms, which necessitated an ultimate
agreement even on lawlike type variations. Mach stressed physical analogs to mental
processes such as memory in organized matter (not to mention his Schopenhauerean
analogy between will and force). Avenarius followed a similar principle of
psychophysical parallelism in his vital series, the parallel, for example, between
physiological disequilibrium and adjustment and the concomitant psychical feelings
of disturbance and relief. James, too, hinted that he regarded the division between
physical and psychical properties as provisional pending deeper analysis. 90

But Russell made significant advances of his own, such as his inclusion of
occurrences that "do not form part of any experience." These "unsensed sensibilia"
are Russell's notion of purely physical events that take place in nature unobserved
and not part of any consciousness:

The stuff of the world, so far as we have experience of it, consists, on the view that I am
advocating, of innumerable transient particulars such as occur in seeing, hearing, etc., together
with images more or less resembling these, of which I shall speak shortly. If physics is true,
there are, besides the particulars that we experience, others, probably equally (or almost
equally) transient, which make up that part of the material world that does not come into the
sort of contact with a living body that is required to tum it into sensation. 91

Thus a table 92 is perceived by all of the observers in the room, and in addition to
these elements which may be linked together into a public object, there are also
"certain other particulars," namely the elements of the table itself. These are not
Mill's "permanent possibilities of sensation," for the unsensed sensibilia actually
exist at all times whether or not the table is also being observed. Russell claims that a
closer look at physical particulars might reveal a concreteness to the physical world
that science has hitherto overlooked:

Physics, in itself, is exceedingly abstract and reveals only certain mathematical characteristics
of the material with which it deals. It does not tell us anything as to the intrinsic character of
this material. Psychology is preferable in this respect, but it is not causally autonomous ... But
by brining physics and perception together, we are able to include psychical events in the
material of physics and to give physics the greater concreteness which results from our more
intimate acquaintance with the subject matter of our own experience. 93

Like Mach, Russell also looks forward to a future psychophysical science in

which the particular qualities of physical things are also included in knowledge as
they are in psychology. He regards introspective psychology as currently being more
concretely realistic than physics because it deals with particulars as such and not
with systems of particulars collected in matter:

It is clear that psychology is concerned essentially with actual particulars, not merely with
systems of particulars. In this it differs from physics, which, broadly speaking, is concerned
with cases in which all of the particulars which make up one physical object can be treated as a
single causal unit. 94

As to intrinsic character, we do not know enough about it in the physical world to have a right
to say that it is very different from that of percepts; while as to structure we have reason to hold
that it is similar in the stimulus and the percept. 95

But whether Russell actually imagined people doing experiments to share or

determine the qualities of the physical world, I cannot say. Such possibilities were
more likely to occur to a free-ranging experimentalist in physiology like Mach or
James than to the asceptic mind of Russell, but I wouldn't rule it out.


Alongside the sensibilia, Russell's most significant contribution to neutral monism

was his theory of "perspectives." More than any other, Russell toughens the
functional presentation of Mach and the causal space of James into a solid, mentally
clarified manifold of what he calls "particulars" or "events." Russell works in this
general manifold by connecting the sensibilia either according to psychological laws,
in which case the serial ordering of events in time becomes that of an ego, or by
differential laws of physics, in which case the world-line is the temporal history of
an object. Russell's treatment of particulars and world-lines was deeply influenced
by the theory of relativity96 in which he says single events may be found at the
intersection of numerous world lines, numerous different spatia-temporal orderings
of those same Einzelerlebnisse, just as in neutral monism elements lie at the
crossroads of physical and psychical orderings. Russell introduces his idea of a
perspective by considering a set of observers looking at a table:

When several people simultaneously see the same table, they all see something different;
therefore "the" table which they are all supposed to see is either a hypothesis or a construction.
"The" table is to be neutral as between different observers: it does not favor the aspect seen by
one man at the expense of that seen by another. It was natural, though to my mind mistaken, to
regard the "real" table as the common cause of all the appearances which the table presents (as
we say) to different observers .. .Instead of looking for an impartial source we can secure
neutrality by the equal representation of all parties. Instead of supposing that there is some
unknown cause, the "real" table behind the different sensations of those who are said to be
looking at the table, we may take the whole set of these sensations (together possibly with
certain other particulars) as actually being the table. That is to say, the table which is neutral as
between different observers (actual and possible) is the set of all those particulars which would
naturally be called "aspects" of the table from different perspectives ... These closely similar
particulars are collected together by their similarity primarily and, more correctly, by the fact
that they are related to each other approximately according to the laws of perspective and of
reflection and diffraction of light. I suggest, as a first approximation, that these particulars,
together with such correlated others as are unperceived, jointly are the table; and that a similar
defmition applies to all physical objects."'

This is essentially the Machian insight that connects particulars functionally,

and which identifies the thing essentially with its so-called appearances or
interactions with other things rather than as effects of a thing in itself:

Like the different appearances of the table to a number of simultaneous observers, the different
particulars that belong to one physical object are to be collected together by continuity and
inherent laws of correlation; not by their supposed causal connection with an unknown
assumed existent called a piece of matter, which would be a mere unnecessary metaphysical
thing in itselC 8

Russell also considers these perceptual processes to apply to the non mental
example of a star surrounded by photographic plates, where the star leaves images
on film as its various perspectives. The object, he says, becomes identified with a set
of event-particulars that may be regarded as its effects:

According to the view that I am suggesting a physical object or piece of matter is the collection
of all those correlated particulars which would be regarded by common sense as its effects or
appearances in different places ... if photographs of the stars were taken in all points throughout
space, and in all such photographs, a certain star, say Sirius, were picked out whenever it
appeared, all the different appearances of Sirius, taken together, would represent Sirius. 99

Thus Russell pictures objects as spatially and temporally disparate systems of

event-particulars which can be projected to a causal origin that will serve as the
"location" of the star. If every event-particular is an effect or interaction, the next
question is likely to be: an effect of what on what? But this is the very question
Russell deems metaphysical and leading to the thing in itself. For it is the objects
that are drawn around the events, and not events that proceed from objects. This is
just like the location of individual events at the intersection of multiple worldlines,
objects, histories.

In graphics, a handy technique called an "exploded diagram" lays out the

function of a completed mechanical device by showing each of its parts in separate
places yet connected by a system of perspectives to show how they would work in
the whole. Likewise, if a system of two stars (A,B) is exploded into the sum
effects-thermal, gravitational, and so forth-<>n one another, what can be said to
exist is a vast set of interactions strewn about, containing at each point an element
consisting of an element that can be said to belong to each of them {a,b} .


By connecting these elements according to laws of perspective or causal laws, each

effect can be drawn back to a variety of"objects," some of which are more causally
localized to certain places. For example, suppose that A consists of various blue
aspects that are of a certain deep hue at the center and gradually lighter shades of the
same hue near the edges. Also assume that B is a deep red near the center with
lighter shades near the edges. Now, every aspect will have both blue and red
properties to it, but the blue perspective lines are easily separated from the red ones.
In particular, changes ofblue shade may be isolated against a fixed shade of red that

does not change at the same rate or at all. In a similar vein, Mach often pointed out
that in the proper scientific understanding, 100 everything has an effect on everything
else in its neighborhood and even mediately on spatially and temporally distant
objects by means of radiation or causal chains. Thus, objects cannot really be
isolated to just one place, but a tighter circle of effects can be chosen by the criteria
of provisionally closed systems. But how many systems? Russell believes that the
salient causal aspects of objects can be found by the method of variations. But can
we not arbitrarily "diagonalize" an object out of aspects by taking an aspect of A
here, one ofB there to make a composite object? I can see nothing to prevent this. In
particular, we can diagonalize a set of elements each selected from different egos.
For example, if someone standing in front of the room unexpectedly sets off a
firecracker, the sound of the explosion appears in every individual ego as a fragment
that can be linked up with the same object and which may produce similar effects,
such as those of a sudden shock or surprise. The question becomes why the aspects
of one ego seem so closely associated that they cannot be shared by others, not even
in principle. The answer of the neutral monist can only be that these elements are
very strongly associated with our bodies in nature.

As Russell himself points out, 101 the neutral monist system of perspectives is an
up-to-date Leibnizian monadology with the petites perceptions but without the
interior windowless monads. In Leibniz's system every monad had within it not only
its own interior states but also perceptions of the states of all the other monads in the
universe. Leibniz's universe was a vast system of interior perceivers. The Russellian-
Machian-Jamesian "objects" are, as it were, Leibnizian monads turned topologically
inside out but otherwise retaining all of their mutual relations and connectivity. A
thing may be regarded as a collection of its effects on everything else, and each of
the effects can be considered indifferently as part of the object itself or as parts of an
object that knows it.


Arthur Lovejoy's Revolt against Dualism was a systematic attack on neutral monism
with some original angles. Lovejoy called himself a natural dualist and said that
natural science had proceeded so far beyond naive perception that the physical world
could not be said to look at all like the world presented to our senses. Thus, Lovejoy
set himself against the direct realism that identifies the percept with the actual object.

One of Lovejoy's lead arguments (often repeated) is that spatially and

temporally separated perspectives such as a star and a sensation of a star produced in
a creature millions of years later (after the star has burned out) cannot really be part
of the same object, for they do not exist at the same time and in the same place. 102
But nothing about Russell's theory demands that objects cannot consist of their
spatially and temporally disparate effects. In fact, Russell's use of world-lines
indicates that various events that become part of objects will be simultaneous for one

observer but not for another. Granted the Russellian star-observer system is an
immense and diffuse object. It consists of incredibly many effects on distant things
included with the more proximate effects of extension into the human sensory field;
but it is a valid object of scientific consideration nonetheless. Lovejoy refuses to give
up spatial and temporal boundaries of objects for the really invariant laws of physics
that do determine matter. 103 But the Lovejoy criterion of spatio-temporal
individuation is dependent on the motion of the observer and thus is only ever
approximately valid, while the laws of physics that back Russell's perspectives have
the benefit of being valid for all observers.

Another objection by Lovejoy concerns the elements at the projective center of

the star. In Russell's examples of the photographic plates, as the pictures are taken
closer and closer to Sirius, the star gets bigger and more intense, and its mixture with
more distant objects becomes less. So what exists at the center of all the
perspectives? Here Russell's position is not easily made out, and Lovejoy claims that
on Russell's theory there is a star everywhere else in the universe except at the
location where we expect the star to be. 104 Instead, Lovejoy says, we find objects
with strewn perspectives and empty holes at their centers. Russell himself even says
this: "As a rule, even when the centre is occupied by a percipient, it nevertheless
contains no member of the group, not even an ideal member: 'the eye sees not itself.'
A group, that is to say, is hollow: when we get sufficiently near to its centre it ceases
to have members. This is a purely empirical observation." 105

This certainly seems damning, but is Russell describing his theory or is he

describing how his theory would appear to someone who still believed in the causal
theory of perception (the subject of the chapter where this quote occurs)? I think
Russell could have answered more simply that there is no difference between
elements at the "center" of the star and any others. What one calls the "interior"
states of the star are interactions among the interior states themselves, rather than
with observing devices outside such as photographic plates. These elements of the
object are still in causal relations to each other and to elements outside. So in essence
the "inside" of the star is no different from its "outside." Russell himself is guilty of
using spatial language of an "empty center" or "hole" where no such thing seems to

One such error that arises from a spatial, not a causal, interpretation of Russell is
his claim that a percept, say of a star, can be causally annexed as a perspective to two
objects ultimately located in different places, for example the brain and nervous
system at one perspective and the star itself in the other. (We saw the same problem
emerge for James.) The very same percept is shared as part of the brain and part of
the star. Lovejoy called this the "denial of simple location" 106 and said that it proved
the theory was inconsistent since the same percept had to be in two different places
at once. But this is just a mistake that Lovejoy typically makes by confusing
locations in perceptual space for those in causal space or super space. Besides, the
idea that a percept may belong simultaneously to the brain and to the star is no more

problematic than that events may belong to two perspective or world-lines at once.
On the Russellian view, as on the Machian and the Jamesian, one may draw objects
as overlapping some of the same percepts if causal relations so determine it.


Russell sometimes appears to differ with Mach and James over the relations between
his event-particulars. At first, Russell interprets these relations in an empiricist
manner and says he cannot answer Hume's problem ofinduction. 107 Mach and James
certainly held that all relations were empirical and both had spoken of a universal
associative relation, and James even of one element or sensation actually "leading
beyond itself to another." Mach also had a realistic interpretation of his functional
connections of elements in Herbartian fashion as pressing one another, like forces.

And yet Russell and Mach at least were very close on their conception of replacing
causal relations with differential laws (finite difference equations in the case of
discrete elements). Mach had said that differential equations were the most realistic
conception of natural causality, 108 and one that should replace philosophical accounts
of "a dose of effect follows on a dose of cause." 109 Russell, too, heaped scorn on the
philosophical analysis of sentences such as "lightning causes thunder."" 0 No matter
how many necessary and sufficient conditions are added, this can never work, he
says, for one can always imagine intervening circumstances that allow A to happen
but prevent B (and, he might have added, ways of bringing about B other than
through A). 111

The Russellian-Machian solution, as I read it, is to fuse the notions of event and law
continuously, as is done when a short stretch of a function is taken as its differential
element. One cannot isolate a stretch to call the event without also taking the law
into account. 112 This indicates a change in the notion of an event or a percept as a
static entity in time. What seems to be a static event separable from the functional
relationship is really just a result of treating very short periods of change as if they
were static, as pedestrians could be regarded from a train as if they were standing
still. In truth, says Russell, an event such as a red percept is a process and
corresponds in time to a steady physical process such as the continuous passage of
light waves from an observed surface. 113 A seemingly discrete event is a process
taken at a very short interval." 4 Russell and Mach seem to be sharing belatedly
Newton's great insight into natural processes, i.e., that natural events are never fixed
but are all in quicker or slower states of continuous change, whereby what is related
are the rates of change themselves.


Neutral monism was primarily an ontological theory designed to produce a common

foundation for physics and psychology. However, there are other aspects of mind
that are not mentioned by the theory at all. Rightly or wrongly, we speak in addition
to our sensations and presentations about inhabiting a world of meanings, intentions,
uses, norms, and the like. What is meant is a "third realm" (Karl Popper's term)
where public meanings exist, which are not reducible to psychological phenomena
(mental presentations, figures, images, words or the like). Nor are these entities part
of the physical world, as, for example, perfect triangles do not exist and a plate is
never a perfect circle.

Mach encountered this problem in the rival school of Brentano, who insisted on
a "third realm" as a new field for logico-metaphysical investigation. Husserl 115 and
Brentano 116 both related in letters to Mach that they were doing a reine Wissenschaft
of pure logical forms independent of empirical psychology. Mach did accept the idea
that there were logical norms of procedure in the sciences that did not have to reflect
the actual way human beings think, as he said: "I am perfectly able to distinguish
between psychological and logical questions." 117 But Mach also believed that logic
was essentially a science of organization for ordering factual knowledge already
attained. In his chapter "Deduction and Induction Psychologically Viewed" in
Knowledge and Error, Mach wrote of logic as a sterile field of separating out
propositions surreptitiously assumed already in the premises and providing no new
knowledge. Mach generally distrusted deduction as a mere device of presentation to
hide the empirical content of a science. For him, it seems to have been a sort of shell
game in which the pea is hidden under the shells, shuffled about, and then suddenly
revealed again in another place. He thus seems to have been unaware of the
constructive uses of mathematical logic to create abstract formal structures similar to
the way mathematics makes models of physical processes. Brentano's ontology and
Husserlian phenomenology are both attempts to formulate such constructions, and a
good deal of twentieth-century metaphysics followed that road.

Russell is an interesting case, for he seems to have rejected the "third realm"
after abandoning his earlier realist theory of universals. He did so after Ludwig
Wittgenstein reinterpreted his universals as tautologies, or purely syntactical truths
about symbols. Russell also rejected Frege's notion of a public "sense" for individual
words, which supposedly expressed a public trans-subjective meaning, which
designated objects by means of abstract functions or mappings. Russell's own uses
of mathematical logic became more concrete and targeted at solving real problems in
the sciences. In the Analysis of Matter he throws out an enlightening series of such
problems from relativity and other areas for solution.

In all, it might be said that neutral monism is a consistent ontological theory that
emerged from uniting the nineteenth-century sciences of psychophysics and physics,
but one that still has a long way to go in really explaining (or explaining away)
phenomena of mind (if those phenomena are to include so-called intentional beliefs
and or logical constructions).

But as for Muller's founding identity of sensation and energy, it may be rough but
"is it not life, is it not the thing?"


Although Ernst Mach's name is inextricably linked with Vienna and Viennese
philosophy, the scientist spent his most productive years in nearby Prague. Moving
there from Graz in 1867, where he had suffered from a lack of means to perform
laboratory experiments, 1 Mach took up the chair of experimental physics at the
University of Prague. The position promised research space and a budget for exactly
the sort of experimental work Mach enjoyed. His most signal professional
achievements in physics were in the area of experiment, for example in his work on
establishing the Doppler effect and in photographing the shock waves of supersonic
bullets, which gave us the term Mach I.

As Dieter Hoffmann describes, Mach inherited a rather shabby research facility

in the old university building, which reflected the rather somewhat backward state of
research in physics of that day. 2 However, in the course of the next ten years, Mach
obtained permission from Emperor Franz-Josef for a new physical institute, which
included a modem laboratory as well as living quarters for his wife and growing
family. In his lecture "On the Fundamental Principles of Electrostatics," Mach speaks
like a proud director of the great capacity of the battery housed in the institute. 3

The number of students interested in research careers in physics was not great.
During Mach's twenty-seven years in Prague, only a handful of doctoral theses were
written under him or with him as a chief advisor. 4 Yet by all accounts Mach's
lectures endeared him to his students the same way his books have delighted the
curious. He was an excellent teacher who reached for the deepest understanding of
his subject and who performed experiments with an assistant, which Hoffmann
reports was something of a novelty of the day. No doubt Mach's characteristic
historical treatment of material, which is so appealing in his books, also helped him
to communicate fundamental results in the simplest terms.

Already before Mach's arrival, Prague was on the boil from nationalistic
disputes between the German-speaking minority, which held sway in intellectual and
administrative life through the official use of German, and the Czech majority, which
was beginning to come into its own. Mach found himself in the middle of this dispute
as rector of the university from 1878 to 1880.

From the very beginning of his tenure, Mach urged privately for splitting the
university into Czech and German-speaking sides, a compromise which, when fmally
ordered by a "supreme decision" of Emperor Franz -Josef in 1881, was by no means
as straightforward a solution as it seemed. Through a misunderstanding, apparently
fostered by the German nationalists, Mach was identified by many Czech newspapers
as a nationalist himself, but, as Hoffmann reports, his Czech acquaintances knew
better. 5 As Hoffmann also stresses, the fact that Rector Mach was lumped with the
pro-German side by speaking on the same stage as E. Klebs, the dean of the medical
faculty, who made openly provocative remarks, an event which led to rioting by
Czech students,6 may have been what helped Mach to broker a compromise
acceptable to German-speakers, since he had accumulated political capital to spend
with them. National tensions reemerged when Mach was again elected as rector in
1883 and was expected to administer oaths to the largely Czech students of the still
undivided theology faculty. Mach resigned, citing ill health. 7 He complained more
candidly in an 1883 letter to his friend JosefPopper-Lynkeus:

The rectorate embarasses me a great deal. I cannot reflect in peace and must always defend
myself from becoming occupied with mere trifles. I deeply want to write the Analysis of

Mach often expressed his distaste for nationalism of any variety, which he
considered a primitive instinct held over from earlier times, and presciently
compared the results of national fervor to "the worst brutalities of the 17th century. "9
Mach also condemned the social order of his times in the strongest terms:

Let us remember what miseries our forebears had to endure under the brutality of their social
institutions, their laws and courts, their superstitions and fanaticism; let us consider how much
of these things remains as our own heritage and imagine how much of it we shall still
experience in our descendents. 10

I believe that, like many scientists, Mach saw in the best conduct of the scientific
community a pattern for a social order of reasoning individuals and hoped that the
political order might someday approximate to it. However, had more thinking people
of the nineteenth century overcome their evident loathing for public life, especially in
Vienna, instead of complaining of "the last days of humanity" and blithely "letting
things go to the dogs," they might have headed-off some of the worst disasters of the
twentieth century.

William James's impressions of Mach in 1882 were of a well-established man

with a favorite club and the leisure to indulge his voracious intellectual curiosity.
And yet Mach's life in Prague had its unhappy personal episodes. The theft of books
by an assistant" and the ingratitude of another12 depressed him deeply. But finally it
was the suicide of Mach's son Heinrich in 1894 that devastated the family and
hastened their departure for Vienna. Thereafter Mach embraced a Buddhistic
resignation and, as Hawthorne said of Melville, "had made up his mind to be

John Blackmore speculates that it caused Mach, who blamed himself for his
son's suicide, 13 to withdraw further and reject all public recognition for his work. 14
Blackmore relates a story by Wilhelm Jerusalem to this effect:

On February 18 in the year 1898 I came in and congratulated him on his sixtieth birthday, but
added at the same time that I hadn't told anyone else about it. "For that I am especially
thankful," answered Mach, and that this was his real opinion I discovered to my deep sorrow ten
years later. I had made all the arrangements in union with some of his friends and admirers for a
special Festschrift honoring his seventieth birthday, which would have had an
international character. The publisher had been found, contributions from William James,
from Harold Hiiffding and others had been promised, then Mach heard about the plans and
protested so vigorously that we had to abandon the whole thing. 15


Yet Mach did find a stable scientific home in Prague for many years. And here he
performed his famous experiments on the sensations of movement and photography
of supersonic bullets. In 1870 Mach was elected to the Lotos Society for the History
ofNature and in 1871 to the Royal Bohemian Society of the Sciences. 16 Some ofthe
papers Mach gave were reports on his ongoing work at the institute, where work
seems to have focused on acoustics, optics, and electricity. Mach's scientific
notebooks of these years contain page after page about optics, which played a very
central role in his physical thought. Hoffmann believes, however, that there was no
unified school ofMachian physics:

Mach ... was not able to found a school of experimental procedure in the real sense of the word-
not in Prague and even less so earlier in Vienna and Graz. His range of experimental interests
seems to have been too wide to allow the development of a tradition of research or laboratory
programs within the small group of his immediate students, though the work of Vincenz
Dvorak at Zagreb, Gustav Jaumann at Bmo and Anton Lampa at Vienna and later in Prague
should not be overlooked. 17

Thus, Hoffmann says, Mach's influence on science came mainly through his
philosophical-physical works, beginning with the Conservation of Energy in 1872.
Several journal notes that Mach published in Lotos should also be included in his call
for a "phenomenological" natural science. In fact, at least some of the work
performed by Mach in Prague (with his students) seems to have concerned finding
some expression equating the radiant energy of light and mechanical energy, which is
generally in line with Mach's ideal of a comparative physics. He spoke about this
general program in a section of the Mechanics:

The removal of notions whose foundations are historical, conventional or accidental, can best be
furthered by a comparison of the conceptions obtaining in the different departments, and by
finding for the conceptions of every department the corresponding conceptions of others. We
discover thus that temperatures and potential functions correspond to the velocities of mass-
motions. A single velocity-value a single temperature-value or a single value of potential
function, never changes alone. But whilst in the case of velocities and potential functions, so
far as we yet know, only differences come into consideration, the significance of temperature

is not only contained in its difference with respect to other temperatures. Thermal capacities
correspond to masses, the potential of an electric charge to quantity of heat, quantity of
electricity to entropy, and so on. The pursuit of such resemblances and differences lays the
foundation of a comparative physics, which shall ultimately render possible the concise
expression of extensive groups of facts, without arbitrary additions. We shall then possess a
homogeneous physics, unmingled with artificial atomic theories. 18

Like Julius Robert Mayer, the co-discoverer with Joule of the principle of the
conservation of energy, who had made an outline of the different sources of "natural
force," 19 Mach also intended to use analogies to develop a general science of energy
transformations that would cut diagonally across the different departments of physics.
He sometimes called this field the "chemical view" (in the sense that chemistry deals
with changes that draw on many physical domains at once). Thus, whereas a self-
contained theory of electricity or heat would restrict the scope of its laws to range
only over those phenomena, the general science aimed instead at framing laws of
transformation between electricity and heat. Some of Mach's Prague notebooks bear
out research in this direction as a working master plan for both Mach and his
students. In an entry of 187420 Mach coded a table of the different natural sources of
energy known to him:

M=Mechanical Work
W=Heat (Wiirme)
E=Electrical potential (Electricitiit)
Mg=Magnetism (Magnetismus)
L=Kinetic Energy (lebendige Kraft)

He then considered twenty combinations. Mechanics, for example, in the abstract, is

the study of energy transformations from M to L and L to M by the principle of least

It is possible to discover analogies for the principle of least action in the various departments of
physics without reaching them through the circuitous course of mechanics. I look upon
mechanics not as the ultimate explanatory foundation of all the other provinces, but rather
owing to its superior formal development, as an admirable prototype of such an explanation. In
this respect, my view differs little from the majority of physicists, but the difference is an
essential one after all. 21

The principle of least action in Lagrange's form states that, applied to a system of
masses, if a function T can represent the entire kinetic energy of the mechanical
system and -V can represent the entire potential energy, then we can form the
system's Lagrangian L=T-V. The positions and their time derivatives, the velocities,
appear in the Lagrangian as functions of time.

Now the system is free to move along any paths consistent with the connections
of its parts and satisfying the conservation of energy. But of all of these possible
paths, the mechanical system will actually take the one that makes the action integral
A= JLdt take a stationary value. In the Mechanics/ 2 Mach gives a non mathematical
illustration of this principle by attaching to the system a single external weight that
slowly accelerates up or down or finds a stationary position as the configuration of
the system attains a stable (or even unstable) equilibrium.

For all of the power of Lagrangian techniques, various conditions may still get in
the way of such a complete analysis of the problem. For example, the presence of
friction or other unanalyzed countervailing forces will interfere with the conservative
nature of mechanical systems by converting some of the excess energy to heat or to
charging of bodies by rubbing. Or the potential energy function may be explicitly
time dependent, in what is called a rheonomic system, as, for example, a charged
particle in a cyclotron returns to the same initial position with more and more
velocity on each revolution. 23 Also, it may occur that a system's position is not a
straightforward function of the positions of all of its parts; the example usually given
is of a rolling ball on a surface, which has only five degrees of freedom, and yet five
parameters are not sufficient to determine the system's position at a later time. The
system is said to be non holonomic. 24 Yet Mach's approach always assumed that
these difficulties represented incomplete knowledge and not insuperable conditions
once the problem is thoroughly analyzed. For Mach, the countervailing forces will
always themselves be derivable from potentials and be amenable to the same
treatment as other energy transformations:

Our attention is directed to an important property of the forces of nature, to the property that
the work of such forces may be expressed as a function of the coordinates. Whenever
exceptions to this principle are observed, we are disposed to regard them as apparent, and seek
to clear up the difficulties involved. 25

Thus, according to Mach, when we encounter a violation, we only need to expand the
system out to encompass the greatest variety of natural forces acting in it to apply a
generalized least action principle to all of them. Mach published a note about this in

The static and dynamical principles of mechanics may be expressed as isoperimetrical laws.
The anthropomorphic conception is, however, by no means essential, as may be seen, for
example, in the principle of virtual velocities. If we have once perceived that the work A
determines velocity, it will readily be seen that where work is not when the system passes into
all adjacent positions, no velocity can be acquired and consequently that equilibrium obtains.
The condition of equilibrium will therefore be oA=O; where A need not necessarily be exactly
a maximum or a minimum. These laws are not absolutely restricted to mechanics; they may be
of very general scope. If the change in the form of a phenomenon B be dependent on a
phenomenon A, the condition that B shall pass over into a definite form will be 1iA=0.26

In the case of the dynamical principle, the condition that the action A shall take a
stationary value is oA=O, or ofLdt=O, between two given times. This condition leads
to the Lagrangian equations of motion, one differential equation for each of the
generalized coordinates of the system, which can be solved for the accelerations of
the masses, i.e., the "definite form of the phenomenon B" as Mach says. 27

The extension of the Lagrangian methods to other sciences, as Mach indicates as

his goal, requires finding analogous terms to put into the Lagrangian formalism that
behave in an analogous way to the position (in a potential field) and the velocity (in
the kinetic energy) for mechanical problems. Hence there is a direct connection
between this program of general dynamics and the energetic analogies Mach sought.

The physicist Wilhelm Wien spoke in 189028 of the change in attitude toward the
energy in physics. Originally, for Lagrange and the other early investigators, it was
the mechanical system that was of primary importance, and energy was just one of
the integrals of the equations of motion. By objectifying the energy, one was still
referring to a property of an underlying system. However, with the observation that
energy was indestructible, Wien says, it obtained a special significance in its own
right. And in cases of the other sciences, dealing with thermal, chemical, or
electromagnetic properties, the underlying system is not observable, as it was in
mechanics, and thus the macroscopic parameters of the system such as the energy are
the only access to knowledge about it. If it is then found that equations relating these
parameters can be found in such a way that the future states of the system can be
determined, the thought arises that knowledge of the underlying system is

Hence Mach's overall willingness to ignore the details or mechanisms of

individual systems and go straight to his table of energies, which appears rather
simplistic at first sight. For example, all of the complexities of mechanics are allowed
to boil down to two directed transformations from potential to kinetic energy and
from kinetic to potential. As Mach notes:

There are twenty combinations to be tried systematically. Write down the fundamental laws
and concepts on individual sheets and then try to combine them. Electricity and heat are as
much material as masses. 29

In another place, Mach makes a kind of multiplication table, crossing each form with
each. 30 Mach also kept tables of the disanalogies between the forms, especially those
concerning the flow of heat between different temperatures (about which more later).
Physics like Mach's is called "phenomenological" for a special reason. Describing
energy changes often does not require an underlying knowledge of molecular
processes, transmission through a medium or by contact, or even whether the
mechanisms are continuous or discontinuous in nature. Thus the occurrences in

nature are constrained as to the energies of their initial and final states even if we
know nothing else about them, or even whether a process will occur or not. Erwin
Hiebert gives an instructive precis of the energy law thus:

If a one-to-one correspondence is maintained between any two forms of energy then there are
any number of paths along which the transformation may take place. For example if a certain
quantity of heat energy H is found to be equivalent to an amount of electrical energy E, being
compared along the same numerical scale, and this amount of electrical energy is also to be
equivalent to an amount of mechanical energy M along the same numerical scale, then since
H=E and E=M, then H=M also. Thus the same result obtains whether the transformation from
H to M is achieved directly or indirectly by way of E. This rule can be summarized by saying
that the energy changes for any given process are completely independent of the path or
mechanism for the process; the energy changes are solely dependent on the initial and the final
states. 31

It is important to add immediately that mechanism independence also limits the

application of the principle of energy. From the simple meeting of two energy forms,
electrical and gravitational for example, one cannot tell whether a transition from the
one to the other may take place, for example, if a weight is connected somehow to a
charged body, or whether these two natural potentials will have no effect on each
other and each will behave as if the other is not there. Mach was clear on this matter
and never tried to anticipate matters for empirical investigation a priori:

If different levels of energies of the same kind meet together, it depends wholly on special
physical circumstance whether and what transformations of energy occur. .. The principle of
energy only determines the amounts of the transformation, not the circumstances under which
the transformation takes place. It is a matter for the special domains of physics to determine
the circumstances. 32

Mach was especially at pains to point out that the peculiar irreversibility of heat flow
in comparison to the behavior of other energies was not a consequence of the energy

It is no accident that Mach's first book was entitled History and Root of the
Principle of the Conservation of Energy because it was exactly this phenomenon
which attracted his admiration. Although energy and action were the last concepts of
mechanics to be articulated, Mach showed that historically considered, Stevinus,
Galileo, Huygens, Lagrange, and Camot had all based their reasoning on the
principle of excluded perpetual motion, or the principle that energy cannot be created
from nothing. Hence, in many cases they could proceed by ruling out contradictions
with this fact-and not by the investigation of underlying material causes, which was
impossible for most of the pre-twentieth-century development of science. Such was
the prime illustration of what Mach meant by proceeding "by facts and not by
hypotheses," which he hoped would be pursued in the future.

Faithful adherence to the method that led the greatest investigators of nature Galileo, Newton,
Sadi Carnot, Faraday and J.R. Mayer, to their great results restricts physics to the expression of
actual facts, and forbids the construction of hypotheses behind the facts, where nothing
tangible and verifiable is found. If this is done only the simple connection of the motions of
masses of changes of temperature, of changes in the values of the potential function, of
chemical changes, and so forth is to be ascertained, and nothing is to be imagined along with
these elements except the physical attributes or characteristics directly or indirectly given by

It is often thought that Mach was hostile to physics that transcended observation
and that he regarded entities too small or too distant to observe as non existent.
However, this extreme form of positivism was never espoused by Mach. He seemed
rather to be calling attention to energy transformations as basic phenomena behind
which one could only expect to fmd further energy transformations all the way down,
or the same fact in a new form. The reason why Mach believed this has to do with an
argument he claimed to have learned from Newton about which laws run deepest, of
which more later.


Mach's phenomenological physics depended on analogies between different energy

forms, corresponding roughly to the two energy factors in thermodynamics: a
quantity of heat and a temperature. The analogy between mechanical energy and heat
is sometimes called Zeuner's analogy, after an illustration published in 1859?4
Zeuner compared producing energy by heat in Carnot's reversible cyclic process to
producing energy by raising and lowering a weight. In Carnot's process a theoretical
maximum of work is obtained from heat (1: 1) by forbidding exchanges of heat
between different temperatures and by forbidding the raising of weights by
accelerations. Thus, obeying the same conditions, Zeuner observed that a weight at a
height h possesses a certain potential energy, which it can transform to usable energy
through transfer (at constant velocity) from a greater height to a lower height. Of the
energy produced, some of it can be transformed into a free or usable form and taken
out of the system, while the rest of it remains in the same form but moves from a
higher to a lower "potential level" for effecting change.

According to Zeuner's analogy, heat possesses an energy proportional to the

quantity of heat Q and to its temperature T. This heat can be transformed into usable
work and leave the system if the remaining amount falls from a higher to a lower
temperature. Hence, in the analogy between potential energy and heat energy, the
mass is in some ways analogous to the capacity of a body for heat, and the height is
analogous to the temperature. The "energy value" of a weight is thus not W but Wlh,
analogous to what Zeuner called the "weight of heat" Q/T.

Georg Helm relates in his history of energetics that Zeuner also considered his
analogy to hold between heat flowing between unequal temperatures and weights that
are allowed to accelerate between changes of height. 35 Zeuner's idea here was that
just as dynamical heat flow always involves some waste of the working energy to be
obtained from the process, so too an accelerating body outruns the attempt to recoup
the working energy completely in a form such as impact, as some is lost to heat,
friction, or excess vibrations. If the weights are equilibrated by a support during their
fall, they will fall ideally with a constant velocity, not an acceleration, and, Zeuner
says, the maximum of their work can be converted over to the utilizable form.

Mach extended Zeuner's analogy, but only to reversible cyclic processes, not
dynamical ones, in an 1871 note in Lotos, "Eine Bermerkung tiber den zweiten
Hauptsatz der mechanischen Wiirmetheorie." Here Mach set up the analogy of the
reversible cyclic process for all energy forms, but, unlike Zeuner, restricted his
considerations to cases where the maximum of usable energy is obtainable and where
processes can all be returned to their initial states.

The second law of thermodynamics can, as is well known, be expressed for a simple case by
the equation:

-Q/T + Q'(l/T' - 1/T}=O,

where Q denotes the quantity of heat transformed into work, at the absolute temperature T, and
Q' the quantity of heat which simultaneously sunk from the higher temperature T to the
temperature T'. Now we have not far to seek for the observation that this theorem is not
limited to the phenomena of heat, but can be transferred to other natural phenomena if, instead
of the quantity of heat we put the potential of whatever is active in the phenomena, in the place
of the absolute temperature the potential function. Then the theorem may be expressed thus:

If a certain potential value P of an agent at the potential level V passes over into another
fonn--for example if the potential of an electric discharge is transferred into heat-then
another potential value of the same agent sinks simultaneously from the higher potential level
V to the lower one V'. And the said values are connected with one another by the equation:

-PN + P'(lN'- lN}=O.

In the application of the theorem, the only questions are, what is to be conceived as potential
(as the equivalent of mechanical work) and what as the potential function. In many cases this
is self-evident and long established, in others it can easily be found. If, for example, we wish to
apply the theorem to the impact of inert masses, obviously the vis viva of these masses is the
potential and their velocity is the potential function. Masses of equal velocity can
communicate no vis viva to one another-they are at the same potentiallevel. 36

Mach gave the equivalent reversible cyclic process for electrical charge and potential
difference in his "Fundamental Concepts of Electrostatics. "37 And he expanded on
the analogy of vis viva and heat in the Wiirmelehre. If two bodies with heat capacity
m and M equalize to a mean temperature U, then U is:

(MT + mt)/M+m

Whereas if two masses m and M collide at opposing velocities v and V and stick
together, the result is a mass Mm with a velocity:

(MV + mv)/M+m. 38

In fact, Mach seems to have arrived at his defmition of mass, which isolates a
capacity for acceleration from the chemical constitution of bodies, by a direct
comparison with Camot's procedure for isolating a body's working capacity for heat
from its chemical constitution. 39 In both cases an argument ruling out a perpetual
motion is used.

These analogies and others led Mach to the idea that every energy value could
be decomposed into what he called the potential of the process that acts and the
potential function, or sometimes the level value. Mach chose this terminology to
indicate that he wanted to deal with such energies that were derivable from potential
functions and that, in general, the ability of a potential source, a mass, an electrical
charge, to do work depended on a second factor of what the potential function is.
Two bodies can exchange vis viva by impact only through a difference in velocity,
masses can only do work by pulling another attracting body, the capacity of heat to
do work depends on the temperature difference, and so forth.

Hence, Mach is never interested in the source alone or the body alone, but
always in the energy value or, as he calls it, the Menge (WN), as a kind of primitive
term describing the physical situation. He is interested, for example, in a moving
body at a velocity, mass at a remove from other masses M/r, a quantity of heat at a
certain absolute temperature Q/T or Q(T). This reflects Mach's holistic outlook that
bodies possess physical properties and capacities for inducing physical changes as a
matter of their relation to one another and not in isolation. It is not the presence of
mass, charge or quantity of heat alone that should command attention in the attempt
to latch onto durable stuffs, but rather always the Menge in which the level value
appears as well:

Is there an equivalent of electricity as there is a mechanical equivalent of heat? Yes and no.
There is no mechauical equivalent of quantity of electricity as there is an equivalent of
quantity of heat, because the same quantity of electricity has a very different capacity for work
according to the circumstances in which it is placed; but there is a mechanical equivalent of
electrical energy. 40

Mach's terms "potential" and "potential function" correspond roughly to the terms
"capacity" and "intensity" used by Helm and Wilhelm Ostwald in their attempt to
found a science of energetics. What Mach calls the potential is called by Helm the
capacity M, and Mach's potential function is called the intensity factor J. Helm and
Ostwald then describe energy changes in nature as a result of the tendency of the

capacities to move from higher to lower levels of intensity whenever possible, as

Helm says:

These secure representations of individual forms of energy as products J dM show that energy
of a specific form passes from one body to another only when the bodies have different
intensities and it then passes from the higher to the lower intensity. 41

Helm credits Mach with recognizing a similar law, while Mach gave fairly generous
credit himself to the whole energetics movement and often seemed to stand with it:

The independent works of Popper [Die physikalischen Grundsiitze der elektrischen

Kraft!ibertragung] and Helm [Die Lehre von der Energie] are, in the aim they pursue, in
perfect accord, and they quite agree in this respect with my own researches, so much so in fact
that I have seldom read anything that,without the obliteration of individual differences,
appealed in an equal degree to my mind. These two authors especially meet in their attempt to
enunciate a general science of energetics and a suggestion of this kind is also found in a note
to my treatise Uber die Erhaltung der Arbeit. Since then "energetics" has been exhaustively
treated by Helm, Ostwald and others.

It is believed that Mach rushed his Wiirmelehre into print in 1896 to counter the
opposition to energetics that had emerged at the 1895 Liibeck Natur-
forschungsammlung due to the attacks of Ludwig Boltzmann and Max Planck on
Helm and Ostwald.42 The episode resulted in a four-way exchange in the Annalen der
Physik in 1896 that all but sealed the fate of the energeticists. As Robert Deltete
points out in his extensive study of the episode, the Liibeck debate and the
subsequent post mortem in the Annalen was the last public hearing energetics ever
got among physicists. Many, like Arnold Sommerfeld, who heard the discussion,
assumed that "behind both [Ostwald and Helm] stood the natural philosophy of Ernst
Mach, who was not present.'.43 And indeed, Mach did refer to Helm's and Ostwald's
work as the technical realization of his own analogies and said that "thermal
capacities correspond to masses, the potential of an electric charge to quantity of
heat, quantity of electricity to entropy, and so on," as if a definitive table of
correspondences had already been drawn up, perhaps in the works of others. But
what had actually been done in this line is unclear, as different tabulations were put
forward by different authors. And, as Robert Deltete and Matthias Neuber have
shown, the philosophical differences between Mach and the energeticists were also
significant. 44 Erwin Hiebert considers the differences so great, in fact, that he refuses
even to class Mach with the energetics movement at all:

For Mach the principle [of the Conservation of Energy] took the position of a maxim or
convention for organizing a large class of natural phenomena and was rooted in an
anthropomorphic sanction related to a biologically determined economy of effort conducive to
survival. "Energy" was for Mach no more than a plausible and powerful concept like force, space
or temperature. Thus it is wrong to include Mach among energeticists such as Wilhelm Ostwald
and Georg Helm---as is often done.45

As Deltete points out, however, Helm too believed energies were not a stuff but a
systematic set of useful relations among the phenomena and nothing more, an
eminently Machian position.

What cannot be denied, however, in this complex set of historical circumstances,

is that Mach, Helm, and Ostwald saw themselves as united against the mechanical
worldview and that each sought a level of description of natural processes deeper
than particles, forces, and motions. Energetics is identified with a tendency at the end
of the nineteenth century to move away from mechanical models and seek a new
basis for physics, perhaps from one of its extramechanical branches.
Thermodynamics inspired energetics for the most part, whereas the "electromagnetic
world view" of forces traveling at finite velocity and mediated by fields seems to
have been a close competitor. 46 Helm thought one of the chief benefits of his
intensity law was that "it liberates us from the compulsion to interpret everything that
happens mechanically.',47 Mach, however, was still more radical in calling for an end
to spatial and temporal representation of the natural processes, which Helm and
especially Ostwald did not pursue as a philosophy.

Mach's terminology actually stays quite close to classical potential theory in the
anticipation that all natural forces would yield to treatment in terms of potential
functions, and by extension to the powerful Lagrangian techniques that can be
applied to them. Mach's proposal, as he states it, was to find out for each form of
energy what the potential was and what the potential function was. For each energy
there is a potential source: the attracting mass as the source of a gravitational
potential for a smaller mass, the source of an electrical potential for a smaller charge,
and presumably, by analogy, a heat source at a given temperature as the source of a
"heat potential" for another body at a lower temperature and a moving mass as the
source of a kinetic energy potential for other moving masses at different velocity (or
squared velocity) levels. In each case, the potential function is the independent
variable determining the ability of the potential to act on others.

For example, if we think of a force emanating from a center and distances laid
off from the center in three dimensions, we can represent the potential energy of a
body moving under the influence of this force in terms of a single potential energy
function U(x,y,z). There are also vector potentials (such as static magnetic fields and
those generated by moving charges) in which the direction of the force is given by a
field of vectors. For a scalar potential, however, the values U takes are numbers, each
corresponding to concentric spherical shells of equal potential energy, but otherwise
indifferent to directions traversed on the surface of the shell. Unlike the force that
falls off as the square of the distance, the potential energy of a large mass M/r falls
off directly with distance.

When the forces that act on a system are central forces, whose intensity depends
only on the distance from a center, the forces in the system may be obtained by
differentiating the potential energy. By differentiating the gravitational potential energy
M/r with respect to the sepamtion r, we immediately get an inverse square force -m/r2
In geneml, force components in three directions appear as the partial derivatives of one
and the same potential function of the coordinates U(x,y,z):

8U/8x=Fx; 8U/8y=Fy; 8U/8z=Fz.

The potential energy difference for a sceleronomic (not explicitly time-dependent)

potential is independent of the path taken by a particle acted on. So, if a particle returns
to its original position the potential energy gained and lost in the trip is equal no matter
what path it takes, and for any closed path the energy is = 0.

The condition for U(x,y,z) to be a potential function exhibiting path independence

is that the function be continuous in all three directions simultaneously, and thus in all
mixed directions a particle might take in tracing out a motion. U's indifference to
direction or the equivalence of paths taking different directions is expressed by the
equality of mixed partials:

The path independence of the potential energy function yields up the conservation of
energy as an immediate consequence. If a point passes out of an initial position Po into a
new position P 1 and then back to Po by any path, the potential function resets to the
same initial value and the energy is conserved. The mathematical assumptions ensure
that mixed directions in the path will not matter to the result. Think of a stone rolling
down a hill and rising to the same distance above the ground on the next hill, or a
pendulum swinging forward and back to its initial position. As Mach emphasized so
often, the assumption that a particle's potential energy is a function of its position rests
on the principle of the excluded perpetuum mobile. For example, if the stone were able
to rise to a higher position using just the energy of falling and the shape of the path, it
could return to its original position with extra vis viva to spare and keep raising itself
higher still by repeating the process, which would genemte limitless energy from


Even though the conservation of energy does not say anything about the energy transfer
between initial and final states, Mach proposed a principle that holds for equalization of
energies that we do know are able to affect one another directly. He called this a "law of
differences" which runs: the equalization of an energy value as a function of time is
proportional to the excess of potential level above the average of neighboring potential

levels. Mach called this a general "phenomenological" law of physics that held over
different domains and even had an analog in nerve physiology.48 Like the energeticists,
who held to a Gesetz des Geschehens, or law of happenings, Mach called special
attention to natural inequalities in potential function as the factor that induced all
changes. Differences of level induced the natural processes in one direction only,
namely toward equalizing those differences by the flow of heat, the acceleration of
masses, and electrical discharge across a dielectric.

One of Max Planck's chief attacks on the energeticists and Mach was that changes
did occur in nature, such as the diffusion of two gasses at equal temperatures and
capacities for heat where a real happening takes place in one direction only but is not
demanded (apparently) by the law of differences or the law ofhappenings.49

Despite the fact that many of Mach's ideas about comparative physics really do
seem to reach their ultimate fulfillment in analytical Lagrangian and Hamiltonian
dynamics (as Planck pointed out), Mach's view of that discipline remained rather dim
and, like the energeticists, he regarded the law of differences as a more fundamental
consideration than even the law ofleast action:

Once we have recognized work as the factor detenninative of motion, once we have grasped the
meaning of the principle of virtual displacements to be, that motion can never take place except
where work can be performed, the following converse truth also will involve no difficulty, namely
that all the work that can be performed in an element of time actually is performed.50

Mach used two special examples to illustrate the law of differences, and unlike his
previous analogy these do not compare reversible cyclic processes but rather the
dynamical flow of heat and the acceleration of masses. The first is found in Mach's
Theory ofHeat during his discussion ofFourier.51

Fourier started from the principle of conduction that the velocity of heat flowing
through a surface is (all else being equal) proportional to the difference in temperatures
going in and coming out. For example, we may consider an infinite solid surface lying
between two parallel plates I (the heat source) and II (the cold sink), see the following
figure. If the temperatures are held constant at both ends, then the temperatures of the
individual points between will arrange themselves according to the law of a straight
line. The temperature of any point of the material is always an average of its neighbors
to the right and the left, and the heat will flow at a uniform velocity through the
material; as much heat flows to any point as flows away from it in unit time.

Cold Hot Source


If, on the other hand, the temperature at the source should suffer a rise, the temperatures
will pass into a curve, where the temperature at a midpoint A is suddenly below the
average of its neighbors. In this case, more heat will flow to the point A than flows from
it, and the temperature at A will rise with a velocity in time proportional to A's deviation
from the mean of neighboring temperatures.52 After this equalization, the temperatures
will pass into a straight line again.

Cold Hot Source

T emperatures

A second example of equilibration given by Mach is the acceleration of masses. 53

Masses in no special force relation to one another change their distances proportional to
the times. If a force relation holds, then the differences of distance between masses
causes them to accelerate and close this difference at a rate proportional to the square of
the time. This is unlike heat, where the equalization proceeds at a constant velocity in
time that is proportional to the temperature difference. In each case, however, Mach
emphasizes, the condition calling for equalization is the difference of energy level
(temperature, distance) above the average of neighboring energy levels, which Mach
expresses by saying that if u represents some property of temperature, of potential
function in short of energy level at a point, the conditions calling for equalization are

which represent the deviation of the energy level u from the mean energy level of
neighboring points in the three directions:

The equation:

thus only represents that u tends to the mean temperature of the surroundings with a velocity that
is proportional to the deviation from this mean. For the stationary state:

ctuldx.2 + ctu!d'/ + cfu1az' = o

that is to say, this state occurs if the above deviation from the mean is zero or if every point has
attained the mean temperature of the surroundings. The stationary (dynamical) state passes into a
complete (statical) state of equilibrium if the flow of heat vanishes, so that:

auJi)x. = away = awaz = 0, or u is constant 55

The stationary state of dynamic equilibrium given by Laplace's equation ((/u!()x2 +

f/u!iJ'/ + (/uJ()z? = 0) describes the mean distribution of temperatures and the constant
flow of heat, while the same equation also describes a body moving at a constant
velocity in three directions, the partial derivatives are in this case spatial accelerations
and not temperatures. Between these two cases only the rates of equalization are
different. Masses equalize by accelerated closing of the distance between them, which is
their potential level; temperatures equalize with velocities proportional to the
temperature differences. There is both analogy and disanalogy in the behavior of the
energies, and Mach believed both should be investigated.


Mach's views on energy and his phenomenological physics were actually of a piece
with his views on space and time. He felt that phenomenological physics would be more
abstract and would not require the use of physiological conceptions of space and time as
a supporting framework of variables. With the important disanalogies firmly fixed in
thought-chiefly those concerning the flow of heat-Mach still believed that
differences and equalizations of physical characteristics at points (U(x,y,z)) did point to
a universal phenomenological principle for all of the energies in what he called a
physical continuum:

If we conceive u as the characteristic of a physical state of a material point (such as temperature,

potential, concentration of a solution, velocity-potential, etc.) then every change of state, the
continuance of a stationary process, equilibrium, is determined by the differences of values of u at
the point (x,y,z) and the neighboring points. In a physical continuum, the behavior of every point
is determined by the deviation of the value of its physical characteristic from a certain mean value
of the characteristic of the neighboring points. 5 6

The physical manifold replaces space. Just as the various potential levels could be a
squared velocities, distances, temperatures, and so forth, the potential function need not
be expressed in spatial coordinates at all for Mach, but rather could be replaced by a
system of equal level-values. The fact that a body follows a straight line in space is for
Mach an expression of the fact that the body is moving on a "level surface" with
respect to all other attracting masses, which would otherwise cause it to accelerate.
Hence, those lines on which masses move inertially or remain at rest with respect to
other masses are a natural system of coordinates, a natural framework of lines and
level surfaces against which to chart the more complex accelerated motions. And
Mach emphasized that such an observation extended to other departments of
physics. 57 A "position" is identified in this manifold along several different directions,
different physical characteristics that characterize the point, presumably a set of
different potential values a point object might exhibit. Mach also called this a "chemical
conception" of space or a "chemical manifold."

For Mach, the process of tracing out distances from one point to another in this
generalized physical manifold is accomplished by action from one level (or set of
levels) to another. A quantity of heat or mass moves between potential levels r and r'
and traces out a swath of volume as it converts energy to motion or pressure. By
consequence, Mach could not accept that real physical properties like potential values
should be written as functions of spatial position. What would it mean to attribute
efficacious physical properties to a bare point of space?58 He seems to have thought it
made much more intuitive sense to derive positions and length in reverse from their
physical properties. This elimination of the spatial is parallel to Mach's elimination of
metaphysics from physics:
If we work over in a similar manner the entire domain of physics, we shall restrict ourselves
wholly to the quantitative conceptual expression of actual facts. All superfluous and futile
notions are eliminated, and the imaginary problems to which they have given rise forestalled. 59

In his later writings about a "chemical manifold" Mach sought a way to resolve
some of the difficulties of his view by attributing each direction in space to a different
physical energy. Thus, even if the value of U seems to require several spatial
directions-which it will especially if a vector potential is involved in electromagnetism
or in a rotating frame of reference-perhaps that is because the location of each point of
a physical manifold is really a combination of several different energies, each
representing the possibility of locating a point with respect to a different physical

Thus, Mach's overall goals, while similar to the energetics movement, were really
directed elsewhere, at the same target he had always sought as a philosopher and a sense
physiologist: the elimination of "sensory properties" of space from physical science and
the concentration on what were for Mach the raw facts of energy transformation.


Mach's lecture Die Geschichte und die Wurzel des Satzes von der Erhaltung der Arbeit
was held on November 15, 1871, at the Royal Bohemian Society of the Sciences in
Prague and published the following year. It is a work that had probably been forming in
Mach's mind all through the struggle of the 1860s, when, as he says, he developed the
ideas about space and time, including the first statement of Mach's Principle and Mach's
overall conception of space and time as applied measurement. Mach had finally
emerged from the wilderness, studying the physiology of the senses and investigating
the history of physics with original sources, especially during a period of illness in the
summer of 1870. 1 In the closing sections of the book he seems to have found his way
here to a solid psychophysical conception of his elements. Many of his views did
undergo changes later, in particular his ideas on thermodynamics, which caused him to
give up the principle of conservation of energy for heat. However even in 1909, Mach
continued to refer this short, fierce exposition in an unaltered edition, as his essential
profession de foi:

Here are to be found the fundamental ideas of the Mechanik of 1883, of the Analyse der
Empfindungen of 1886, which was primarily addressed to biologists, and of the Wiinnelehre of 1896,
and the Erkenntnis und !mum of 1905.2

The narrow thesis of Mach's booklet was that the law of the conservation of energy
was not an exclusively mechanical law. Rather, Mach hoped to show in the opening
historical section that mechanics had always assumed the conservation of energy in the
form of the excluded perpetuum mobile principle, and had never succeeded in proving
it in a noncircular way from more fundamental ideas, a critique that was carried through
relentlessly in the Mechanics.

Other goals of the book were to oppose the mechanical theory of heat, i.e., the
proposal that heat is nothing but a species of particle motion, 3 to oppose atomism in
chemistry in favor of pure energy transformations similar to heats of reaction,4 and to
argue against substance, space, and time as the basic scientific building blocks and in
favor of Mach's conception of energies in terms of their potentials and potential
functions. 5 At the end of the book Mach put forward his own theory of elements and
functions as the solution to these problems as an attempt to ground the excluded
perpetual motion in a dependence of all phenomena.

The Conservation of Energy was a thinly veiled roman a clef. Both in title and
content, the pamphlet was an attack on Helmholtz's famous treatise Ober die Erhaltung
der Krcift (1847). 6 Like all Streitschriften, its weakness was that Mach did not show in a

positive way how he intended to do science differently. Indeed, we are not even given a
clear picture of the view under attack, for which we must turn to Helmholtz himself.


Helmholtz had started from the principle of excluded perpetual motion as a principle for
unifYing natural phenomena, but he saw this principle as identical with the assumption
that all natural processes could be understood as motions of particulate matter pushed
and pulled by forces emanating from their centers, their intensities depending on spatial

The derivation of the proposed laws can be attacked from two points of departure, either from the
law that it is impossible to gain an unlimited amount of work from the actions of any combination
of natural bodies, or from the assumption that all activities in nature are to be traced back to
attractive and repulsive forces, whose intensity depends only on the distance of the two points
acting on one another. That both laws are identical is shown in the beginning of the treatise itself. 7

By this "identity" of the two laws Helmholtz also meant that wherever we can introduce
the phenomenological principle of excluded perpetual motion, we can also say that the
underlying mechanisms (i.e., of heat and electricity) are mechanical and involve
particulate matter and central forces. This was the thesis Mach sought to attack.

Helmholtz expressed the excluded perpetuum mobile principle with his customary

Let us think of a system of natural bodies, which stand in certain spatial relations to one another
and which, under the influence of their reciprocal forces, are set in motion until they arrive at
certain other positions: so can we consider their acquired velocities as a certain mechanical work,
and change them into such work. If we want now to make the same work over again, we must
reset the bodies in their original conditions, by applying other forces at our disposal. For this we
will need therefore a certain amount of work from the latter. In this case our principle demands
that the amount of work that is obtained when the bodies go over from the first position to the
second and is lost in going from the second back to the first is always the same, whatever the
manner, path or velocity of this transition may be. For if it were greater along any one path than
another, we could use the first to obtain the work and use the second to reset the positions, for
which we would only need a part of the work we had just obtained. Thus we would obtain an
indefinite amount of mechanical force and would have constructed a perpetuum mobile, which
could not only keep itself in motion, but would also be able to render up force to the outside. 8

The excluded perpetuum mobile principle is not well named, since it is not a
perpetual change of position that is ruled out but the creation of a perpetual (or merely
uncompensated) source of work. The periodic orbit of a planet and the uniform motion
of a body in space are perpetual changes of place, but not perpetual sources of work.
Mach also saw fit to make the distinction in his amendment to Stevinus's endlessly
moving chain, for example,9 and in his further instances of excluded perpetual motion. 10

Stevinus argued (with reference to the above figure) that the four weights on the steeper
incline must balance the three on the flatter. If not, the chain would be set in motion,
and once again four weights would ascend to a new position causing a new motion and
so on ad infinitum. As Mach remarks, we may well conceive of the chain in frictionless
perpetual motion as long as it is uniform. The instinctual objection is against an initial
self-acceleration of the chain from rest into motion and all of the additional
accelerations the chain gives itself by virtue of its position. 11 These are uncompensated
creations of vis viva from nothing and could be used as a generator of perpetual work.

Really any uncompensated element of work-energy a appearing isolated and alone,

once or repeatedly, is what constitutes a violation of the principle for Mach, as becomes
clear in the logical section of the book. This would also apply to a sudden appearance of
a quality like the pressure of a body, or even the creation of a sensation-energy in the
nerves from nothing. Hence, Mach's positive version of the excluded perpetual motion:
for every element of work a (a sensation, a pressure, an acceleration from rest), a
compensating element of work ~ is needed. Mach saw this as a consequence of the
principle that all elements occur together and as mutually compensating functions of
one another. This principle was the more fundamental for Mach, and it relates to his
idea that nature forms a whole.

"It is not possible to create work out of nothing." If a group of phenomena is to become the source
of continual work, it means it shall become the source of continual variation of another group of
phenomena. For by means of the general connexion of nature, all phenomena are also connected
with mechanical phenomena and therefore with the performance of work. Every source of
continual variation of phenomena is a source of work and vice versa. 12

This is the underlying rationale for Mach's fundamental equation for the work-energy

f( a,~,y,o, ... ,(l))=0.

He insisted that the excluded perpetuum mobile principle is just a "special case" of the
fact that no work elements go uncompensated. And it would seem that Mach believed
that elements could behave like little energy bits, like the differentials of the functions in

which they are bound together. He sometimes said that he regarded the elements as the
smallest divisions of which we are capable in current science, in which case they would
be differently sized at different periods in the history of science.

As for the functions, or variations, Mach pointed out carefully in this and all
subsequent expositions added to the Mechanics and the Theory of Heat that he was not
saying that every element depends rigidly on every other and thus that causal
connections could be predicted a priori from conservation laws. Boltzmann questioned
Mach on this point in correspondence and was satisfied with the answer he got,
although Mach's letters are missing. 13 Thus, there is room for a kind of natural
variability unpredictable from the mere fact of general functional dependence alone.
Mach says there may very well be islets of elements dependent on one another but
independent of the rest. 14 One suspects that these "subgroups shut up in themselves" are
tolerably closed systems like bodies or egos. And, as Mach says in his discussion of
frames of reference for the principle of inertia, we are extremely fortunate that nature
provides us with these examples of relatively closed systems to work with-the earth,
the solar system, the fixed stars-for they give us a chance to observe the conditions of a
miniature-universe and extend judgments about the universe itself

A second limitation, perhaps logically connected with the first, is that there are
fewer equations rigidly relating the changes of elements than there are elements; there is
a certain indeterminate variability built in by the very fact of change in nature. 15 This
argument seems to be directed against Laplacian determinism, i.e., if nature were
completely deterministic, time would grind to a halt and no change could be initiated.
Mach could have found this argument somewhere in Herbart with his processes of
Together and Not-Together. But Mach tells us that he wished to say that we only know
of processes where two energies of different levels meet and can equalize each other.
We do not know how to explain where the differences in level might come from in
nature or somehow be generated, for there is no law to explain it, only contingency. In
the last analysis, Mach attributed the course of time to "variables unknown to us,"
suggesting indeed that Mach may have been an indeterminist, but of an uncommon
sort. 16 He says that somehow change is to be attributed both to laws of the equalization
of energy differences and to the generation of those differences by an unknown
contingent, and perhaps even unlawlike, process:

Changes can only be determined by differences. Where there are no differences there is no
determination. The supervening change may increase the distinctions or it may diminish them.
But if the differences bad a tendency to increase, change would go on endlessly and aimlessly. The
only assumption compatible with a general representation of our own limited environment, is that
of a tendency, on the whole, to a diminution of differences. But if circumstances that set up
differences did not make themselves felt by forcing their way into our environment, a time would
soon come when nothing more would happen at all. 17

By contrast with Helmholtz, Mach's approach to excluded perpetual motion is based on

general principles of causality, an approach far more like that of J.R. Mayer, from

whom Mach seems to have got it. Notice that Mach, who was so hostile to metaphysics
otherwise believed that Mayer's use of general metaphysical principles about force and
the impossibility of obtaining net effect from nothing was continuous with a scientific
investigation, whereas Helmholtz's mechanical philosophy was deemed metaphysical.

Helmholtz's use of the perpetuum mobile principle in 1847 was grounded in the
tradition of mechanics. He said that one should imagine a mass m and a system A
exerting forces on it dependent on the position of m against A. 18 As m moves to various
positions against A the intensity of the forces the system A exerts against it change, and
m either picks up velocity (and thus vis viva) or loses it. According to Helmholtz, the
excluded perpetuum mobile principle demands that when m comes back to its initial
position against A where the intensity of forces is the same, m's velocity must be the
same too. That is m's vis viva is strictly a function of its position against A and thus the
intensity of forces as a function ofm's position against A. 19 So for a system of masses in
rectangular coordinates and central forces X, Y, Z with components dx dy dz measured
in the directions of the forces, Helmholtz could show that adding up all the elements of
the work and the kinetic energy of the different masses is

:E ](Xdx + Ydy + Zdz) = 112 :E mv 12 - 112 :E mv02

Mach said in the Mechanici 0 that if we knew that only the positions m takes up against
the system A are decisive but not the paths it takes, we must already possess the
concept of work, or even the work function itself, otherwise why would we feel justified
in ignoring the path in the first place? We must already know that only the forces acting
along the line of junction between m and A produce vis viva in m and not motions in
other directions (which is a logical possibility if one does not possess the concept of
work yet). The prior assumption of central forces thus manages the whole deduction.
Mach criticized this point directly in his lecture "Mechanik" of 1870-71:

If a force function exists, from which the forces may be derived, then is the work a function of the
coordinates. This is a circle, a tautology. The knowledge is important that in many natural forces
the work depends solely on the coordinates."

The purpose of Mach's criticisms was not to doubt the conservation of energy so
much as the otherwise unjustified assumption of special kinds of forces and mechanisms
as a guarantee of the existence of a potential function, when it was the potential function
that actually caused this assumption to be made in the first place. One sees this strategy
over and over in Mach's physical writings. In his famous critique of Archimedes' law of
the lever, Mach argues that even Archimedes' seemingly watertight deduction of
equilibrium for a lever with equal weights on equal lever arms cannot be based on pure
symmetry considerations since we must thereby know that a quantity called the "statical
moment" or weight times radius of lever arm (P.L) is the quantity with respect to which
the two sides of the lever are called "symmetrical. "22 Likewise, the assumption of
central forces, Mach is claiming, is made to justify the introduction of

the potential function, but central forces are just forces derived from a potential. Hence
the concept of potential function is what really allows us to ignore motion that is not in
the direction of the forces from the outset. And the deduction is no better than
introducing the potential function as a general concept of experience and deducing
consequences from it, a procedure which Mach thought was the most straightforward
and the least infected with what he called Archimedes' "Grecian mania for
demonstration." In the Mechanics Mach pursued this course ruthlessly, attempting to
destroy one deduction after another from Archimedes to Lagrange in order to reveal the
basic concepts of physics as concepts resting on experience alone and not a priori
elements of thought. The fact that we now hardly think this worth saying is a measure of
Mach's influence. In the period when he wrote, however, many thinkers still held that a
reine Naturwissenschaft resting on a priori principles was possible.

Of course, Helmholtz's hope in the "Erhaltung der Kraft" was not at all to prove the
law of vis viva a priori but to make plausible the assumption of central forces based on
further metaphysical assumptions of his underlying mechanical philosophy. For
instance, Helmholtz laid down the hypothesis that nature contains quality-less extended
inert matter in space--atoms-which can only suffer changes of position. Lacking any
other manner of influence, then, the only way one piece of matter can affect another is
through the space between them. The intensity of forces, having nothing else to depend
on, depends on the distance. Hence, Helmholtz seems to try to show that central forces
are the only ones mechanically possible.Z3

The mechanical worldview was thus a kind of reigning metaphysics in the mid-
nineteenth century, one which scientists, including Helmholtz himself, began to move
away from at the end of the century. A second representative of the mechanism was
Wundt, whom Mach also mentions in the Conservation of Energy. Wundt, like
Helmholtz, argued that motional changes of place were the only ones conceivable in
which an object could alter and yet remain identical to itself, a position that Mach
criticized as Eleatic and philosophically antique. Mach pointed out that qualitative
changes of state were just as admissible. 24

One gets the sense that Mach must have read the opening pages of Helmholtz's
treatise again and again, until he could maintain the very opposite of each of the theses
defended therein. First, matter, for Mach, is not inert quality less stuff; it has a
multiplicity of interior physical properties or capacities (for acceleration, for heat, for
electricity), not just mere extension. And, as we find out, matter for Mach is just a
functionally related complex, or tolerably closed system, of those more fundamental
energies or elements, not an inert material with different properties inhering in it. As
Mach had also made clear in his "Raumvorstellungen" paper, which he reprinted in the
Conservation ofEnergy, extension need not be primary in physics. Just as scientists had
learned to express the force as a function of the distance, scientists in the future might
well write distances as functions of the intensity of forces. Mach also suggested during

the 1860s that chemical changes might be interpreted as changes in fundamental

properties of matter not expressed through motion or change of position of atoms in a
molecule, but more like the changes of temperature among substances absorbing and
emitting energy.

To cap it all off, Mach presented the view that the law of excluded perpetual motion
could be interpreted as the deeper functional dependence of those intensive,
extensionless qualities on one another. Thus, causal relations are not necessarily to be
interpreted as motional changes or as extended in space at all, merely as abstract
variations. For Mach's antimechanical view then, the law of the functional dependence
of elements embraces mechanics and is more fundamental than the mechanical
metaphysics of HelmholU5 or even the perpetuum mobile principle used by Stevinus,
Huygens, and other investigators.

We have already seen how Mach could have become skeptical of spatial properties
in physics from his work in sense physiology and his reading of Herbart. During the
1870s Mach often wrote of his confidence that behind the phenomena of matter and
motion, there were processes that could not be thought of spatially. For example, in two
notebook passages contemporary with the Conservation ofEnergy:

The healthy fecund kernel, which is thought to derive from mechanics, can be carried within each
of the sciences on its own. There are few general fundamental laws. Behind the motional processes
that appear. Wherewith the old fear of ghosts [Gespensterforcht] is reintroduced. History. 26

Mach still looked to the theory of heat as an example of a non-spatial process:

"Temperatures are arrived at spatially without themselves being spatial." 27 Mach saw
the motion in place of particulate matter (locomotion) as a superficial metaphor for
change in physical science, but not as the actual nature of change, even if the motional
metaphor were actually pursued into infinitely small regions. He thought qualitative
changes ran deeper and could even explain spatial changes (perhaps through Herbartian
associative and dissociative mechanisms)?8

Mach's demand for a purely qualitative science and a construction of extended

matter and motion only intensified as time went on. It cannot be argued, for example,
that this project was only one of his youth. In his notebooks of the 1890s, Mach
compared the Wundtian metaphysics to the otiose Naturphilosophie ofOken:

Motion is an analogy like any other, which can be quite fruitful in special cases,... and can even
pass further than what has been assumed. The whole atomic theory. The general phrase that
everything is motion is no more fruitful than the phrases of Oken's Naturphilosophie, which for
example has it that the world arises with the greatest of ease through the division of zero into a+
and a-.
A Greek philosopher. Pressure and impulse. Much happens as if it rested upon pressure and
impulse. But that now we should remain with them, given our richer intuitions, whereby precisely
pressure and impulse are no longer so simple! Motion is in most cases a fruitless
naturphilosophische phrase. (As with Oken O=+a, -a,l9

Concerning his own counterproposal to spatial physics Mach was quite reticent in
print. He explained that the belief in the mechanical theory of all natural processes was
so ensconced in his colleagues' minds that he hesitated to oppose it directly? 0 In his
notebooks and lectures he spoke of the GespensteifUrcht that often greeted expositions
of his own theories. Thus, he worked his way critically through the whole body of
mechanics first before making his own point.

Now, to reverse the mechanical-metaphysical conception of the world is not, of

itself, to cease doing metaphysics, and Mach's views are quite as ontologically loaded as
Helmholtz's if not more so-a point he did not always appreciate. Indeed, Mach tended
to identifY his views with "the facts" in a way unbecoming of his critical intelligence. As
if Helmholtz did not think his views were the facts?


The general history of mechanics is for Mach a repeated illustration of how the
excluded perpetual motion was used to found the science, and as Mach openly suggests,
he believed the principle should continue to be the foundation of sciences of electricity,
heat, and chemical energy. Camot's use of the excluded perpetual motion in
thermodynamics is hailed by Mach as exactly the type of phenomenological science he

In the Conservation of Energy, and even more fully in the Mechanics, Mach
showed the role the excluded perpetual motion had played in the development of
mechanics. The endless chain of Stevinus could be used to determine the relationship
between the incline of the plane and the weight held in equilibrium on it, 31 Galileo's
principle that a body must rise to the same height from which it fell was used to deduce
the principle of inertia, 32 Huygens's principle of the center of oscillation for systems of
bodies and rigid connections,33 and the principle of virtual velocities set the general
conditions of equilibrium in statics.

In each case a deduction from symmetry is attempted, but as Mach patiently

showed, in every case the symmetry involved is that of the capacity of the system for
work in the two configurations of the system that are considered symmetrical, the
principle involved that a weight shall not rise of itself, or that it shall not be able to raise
another equal or greater weight as the mere result of falling. Thus, the great natural
investigators always applied the implicit principle (even if they did not articulate it) that
differences of work are the cause of motional forces, and that equilibrium subsists when
no work can be done in a mechanical system.

Work had unconsciously presented to the great investigators a general concept

determinative of motion or change in nature that, once assumed as a fact, could then be
used deductively with great power to rule out contradictions with it.

As Paul Feyerabend correctly pointed out, Mach believed that beyond the initial
assumption of these great facts, science proceeded in a deductive, not inductive, manner
to connect experiences with them. Einstein expressed his agreement with the deductive
use of general facts and even claimed that he had been led to regard the speed of light in
vacuuo as a constant of nature by an argument similar to the excluded perpetual motion:

By and by I despaired of the possibility of discovering the true laws by means of constructive efforts
based on known facts. The longer and the more despairingly I tried, the more I came to the
conviction that only the discovery of a universal formal principle could lead us to assured results.
The example I saw before me was thermodynamics. The general principle was there given in the
theorem: the laws of nature are such that it is impossible to construct a perpetuum mobile (of the
first and second kind). How then could such a universal principle be found? After ten years of
reflection such a principle resulted from a paradox upon which I had already hit at the age of
sixteen: If I pursue a beam of light with the velocity c (velocity of light in a vacuum), I should
observe such a beam of light as a spatially oscillatory electromagnetic field at rest. However, there
seems to be no such thing, whether on the basis of experience or according to Maxwell's
equations ... The universal principle of the special theory of relativity is contained in the postulate:
the laws of physics are invariant with respect to the Lorentz transformations. This is a restricting
principle for natural laws, comparable to the restricting principle of the non-existence of the
perpetuum mobile which underlies thermodynamics. 34

The classic perpetual motion principle does not make reference to the speed with
which processes occur, only to initial and fmal states, and in Einstein's example, he is
able to destroy the wave by adopting a parallel reference frame in which its electric and
magnetic fields are no longer seen as spatially oscillatory and mutually reinforcing. One
might also say that a different sort of impossibility proof based on action is suggested
by Einstein in the sense that the energy delivered in time cannot be allowed to depend
on the frame of reference. If a source of work could deliver the same energy faster in
one region (one frame of reference) than in another, one could move the source quickly
back and forth between the regions generating a cyclical creation of work in the slow
region, which could then be moved over to the fast region to do the same work more
quickly and brought back again to the slow region in time to build up an excess.

The most important application of the perpetual motion postulate for Mach was in
proving the principle of virtual velocities for a system of bodies and rigid connections.
The principle was enunciated by Stevinus35 for all machines (the inclined plane, the
lever, and the pulley) and formally perfected by Johann Bernoulli and Lagrange. It
states that if a machine is in a state of equilibrium-such as a system of weights held in
balance with one another through any arrangement of pullies, levers, inclined planes,
and connecting cords-the instantaneous distances that would be covered if the weights
moved in the direction of the forces that move them all add to zero. A simple example
is the inclined plane with weights held in equlibrium. If we suddenly move the parts of
the machine to an adjacent configuration such that weight N would have moved through
a vertical distance 2M, then the second weight 2N will have moved through the distance
M in that configuration.

The expression of the principle involves small displacements called virtual

displacements, 8s, which are carried out in thought immediately taking the system with
infmite velocity to an adjacent configuration consistent only with the constraints of the
system and its connections, but which are not required to conform to other aspects of
the real motion, except that energy be conserved. 36 Thus, what an application of the
principle shows is that if a system cannot possibly do any work consistent with those
constraints in any of its adjacent configurations, equilibrium will subsist instead.

In the Conservation ofEnergy Mach pays special attention to Lagrange's attempted

proof of the principle of virtual velocities in the Mechanique Analytique. Lagrange
started from Torricelli's idea that in a system in equilibrium if one weight falls, another
rises, in the proportion mgh = m'g'h' no matter how they are connected, thus the center
of mass of the connected system cannot descend. 37

Lagrange, as is known, conceived simple pulleys arranged in the directions of the forces of the
system, passed a cord through these pullies and appended to its free extremity a weight which is a
common measure of all the forces of the system. With no difficulty, now, the number of elements
of each pulley may be so chosen that the forces in question shall be replaced by them. It is then
clear that if the weight at the extremity cannot sink equilibrium subsists, because heavy bodies
cannot of themselves move upwards. If we do not go so far, but wish to abide by Torricelli's idea,
we may conceive every individual force of the system replaced by a special weight suspended
from a cord passing over a pulley in the direction of the force and attached at its point of
application. Equilibrium subsists then when the common center of gravity of all the weights
together cannot sink. The fundamental supposition of this demonstration is plainly the
impossibility of a perpetual motion. 38

Mach's conclusion is that the principle of the excluded perpetual motion 1s the
foundation for mechanics everywhere, but it is never proved, or provable, within
mechanics or even from mechanical assumptions such as the particulate extended
matter and central forces of Helmholtz:

The whole of mechanics is thus based upon an idea, which, though unequivocal, is yet unwonted
and not coequal with the other principles and axioms of mechanics. Every student of mechanics,
at some stage of his progress, feels the uncomfortableness of this state of affairs; everyone wishes
it removed; but seldom is the difficulty stated in words. 39


In the Mechanics, Mach extended his treatment to the dynamical laws with essentially
the same message: that work is still the concept that determines motion or change. Thus,
if no work can be accomplished, nothing will happen for systems in equilibrium, and if
work can be accomplished in mechanical systems that are not in equilibrium, the
extension reads that as much will happen as can happen. 40 As an example, heavy
weights fall whenever possible and never rise of themselves. We can view the
spontaneous falling of weights as the conversion of potential to vis viva whenever
possible. According to this principle, a machine is best viewed as a kind of energy-

conversion device (similar to a heat engine) for transferring potential energy of weights
into kinetic energy, by accelerating or back again by decelerating and rising. It is always
headed toward a configuration that is closest on the way to doing one or the other. Thus
in dynamics too, the accomplishment of work provides the essential guide to the
maximum or minimum or maximum-minimum action in the same constraints:

It will be seen that the principle of least action, like all other minimum principles in mechanics, is
a simple expression of the fact that in the instances in question precisely so much happens as
possibly can happen under the circumstances, or is detennined, viz. uniquely detennined by
them. 41

In mechanics the extension from statics to dynamics is accomplished in a very

elegant way by D'Alembert's principle. The principle, proposed in his Traite de
Dynamique, is a particularly brilliant philosophical application of the Newtonian
principle of action and reaction. According to the Newtonian principle, when a force
acts in nature, even on a particle free in space, there is a reaction, an actual
"counterforce" in the opposite direction. Thus, even when a force accelerates a free
mass, we write F=ma, but we could also write F -ma=O, and interpret the
accelerations of the masses as making up a new "negative" force of inertia that resists
and counterbalances the impressed forces on the system bringing the total back up to
zero, analogous to the principle of virtual velocities. 42

L (F- ma)8x + (F- ma)8y + (F- ma)8z =0. 43

Thus, when masses are accelerated by impressed forces, the reason why the forces
cannot move them any faster than they do is because in every moment of time they
are met by exactly resisting forces of inertia. Now, Mach saw here a principle of
"work and counterwork" which, as Thomson and Tait pointed out in their Treatise on
Natural Philosophy, could even have been what Newton originally meant by "actio"
in the principle of action and reaction, thus crediting Newton with the explicit
recognition of the energy as the action of a force impressed against an equal and
opposite resisting force in time. 44

In Mach's earliest preserved notebooks of the 1870s, he seems to have been

struggling toward a dynamical conception of least action in a Wechselwirkung of

In what relationship does action and reaction stand toward work? ... Follow it back to the reaction
and demonstrate the deeper grounds of the conservation of energy. At bottom this is the intent of
Jac. Bernoulli, which leads to D'Alembert's principle. The Principle of the Maximum of Work: if
they oppose one another so far. So far do they work against one another.45

Pressure and counterpressure. Work and counterwork.

Velocity by the Wechselwirkung of masses
Work by the Wechselwirkung of masses ...
Reaction. Different masses between one another...
General law of reaction is D'Alembert's principle.

Look over [Obersehen] the center of oscillation. Equality of action and reaction with consideration
ofD'Alernbert. ..
Principle of virtual velocities is actually the principle of wotk.
Principles of application. Lagrange.
D'Alembert's principle is actually the principle ofwork. 46

D'Alembert's principle does in fact correspond to the principle of the

conservation of energy in the case where the impressed forces are derivable from the
potential energy -V and the forces of inertia are derivable from a kinetic potential
energy T. Then the virtual displacements can all be replaced by the infinitesimal
elements of the actual motion, and the above equation reduces to what Mach calls the
"principle of work" for the mass-motions of the system:

~ f(Xdx + Ydy + Zdz) = 112 ~ mv 12 - 112 ~ mv02

What we observe even in cases of acceleration, Mach points out, is a transformation

of the impressed force -V to kinetic energy T with the total unchanged. This is why
Mach calls D' Alembert's principle a principle of the equality of "work and
counterwork" even in the case of accelerated motions.

Obtaining a result in terms of the motion of masses is useful as an end result,

given the relative hiddenness of other energies, and the mathematical methods for
describing motions in variables of space and time. But the motion of masses is just
another energy form for Mach among others and does not hold any privileged
position for him over the electrifying or heating of bodies as an end result of physical
processes. Hence, his insistence that other forms of energy transformation are just as
important to study by extending the mechanical principles. For Mach, energies like
heat and electricity also exhibit analogs to "inertial forces" with which they resist
changes against their own tendency to seek lower potential levels, and in this sense
they behave just like resistance to acceleration.

Thus, one of the uses of Mach's analogy between reversible cyclic processes of
different kinds appears to suggest an extension to dynamical processes along the
lines of D' Alembert's principle. 47 In a transformation of one form of energy WN to
some other, two things will happen. The natural potential source WN stands to drop
to some lower potential level V' and will produce an impressed force. But this is
balanced by the "inertial force" of the new form of energy that resists the change by
climbing from a lower to a higher potential level, for example in attaining a velocity.
Thus, the force employed in overcoming its inertia will balance the impressed force,
and the result will be that the amount raised up in level in the new form to equilibrate
the change will equal the amount that has fallen in the previous form.

This marks out an extension of the methods and principles of dynamics to other
non-mechanical sciences, but as Mach knew, 48 the exception is heat, which can fall in
its potential to do work simply by flowing to a body of greater capacity for heat
without transforming and raising up another form to an equivalent potential level.
This is a property Mach says marks out a special, quite puzzling nature for heat
among the other energies. One might even say with the D' Alembert analogy in mind
that heat offers "inertial" resistance to being made to flow backward like other
energies and can be made to flow forward against a resistance like a weight or an
electrical generator, but without a complete transformation into the resisting energy
form (making completely efficient heat engines impossible).

Overall, Mach conceived of the world as consisting of various energetic potentials

straining against one another, sometimes equalizing one another and doing work
through a transformation into an equivalent. Even free motions of accelerating bodies in
space are not unconstrained motions. If all the potentials are constrained by others, the
spatial concept of work as "force through a distance" could be replaced by the concept
of one potential in constrained action against another in an element of time.

The latter conception could be a means to reduce the spatial elements of mechanics
to a generalized dynamics, leaving only energy and time, very much in line with the
program Mach had announced in the Raumvorstellungen paper. With his own
"oppositional" view of elements-that variations of one sort are always met by
variations in an opposite direction or sense-Mach seems to be saying that
D'Alembert's principle in mechanics can be construed as a general law of dynamics for
expressing forces as the action taken against resisting force in a given time, or forces as
always making up the elements of energy in time. Although this is far from any kind of
complete argument, the way Mach chose to interpret mechanics seems to point at a
reduction of mechanics to the ideal dynamics Mach pursued in his philosophical
musings. If an element a changes to a', we expect to fmd a resistance, a countervariation
of some ~ to W. The reciprocal changes of these elements with respect to one another
are then the barest foundation for a measure of time, as his reduction in the
"Raurnvorstellungen" paper had indicated was his goal. 49


Mach's definition of mass-ratios was announced in his 1868 paper, "Uber die
Definition der Masse," which appeared in Carls Repertoriumfiir physikalische Technik
along with two other notes in 1868. It had been submitted one year earlier and rejected
by Poggendorf, the editor of the Annalen der Physik. 1 Mach claimed to have developed
his physical ideas much earlier in his lectures and private historical studies, but had
avoided presenting them to the stiff-necked community of physicists. He said in 1872,
reflecting back:

These thoughts, which, as the notes and quotations from my earlier writings show, are not of very
recent date, but which I have held since 1862, were not suited for discussion with my colleagues----I,
at least, soon tired of such discussions. With the exception of some short notices written on the
occasion of other works and in journals little read by physicists, but which may suffice to prove my
independence, I have published nothing about these thoughts. 2

His treatment of mass remained essentially unchanged in the 1883 first edition of his
Mechanik in ihrer Entwicklung. The definition soon became one of Mach's signature
achievements; Einstein, for example, utilized it in his lectures on elementary
mechanics at Zurich in 1909-1910 to call attention to the independence of
gravitational and inertial mass from the chemical composition ofbodies. 3

Mach's contemporaneous "Raumvorstellungen" paper of 1866 gives something

of a philosophical guide to the mass definition. Recall that Mach's program was to
replace both matter and spaces between matter with an underlying functional relation
of pressures. Pressures due to impact would be proportional to the masses and their
squared velocities, and pressures due to fundamental forces would be due to the
greatness of the masses and their distances from one another. Mach said that rather
than writing the forces as functions of the distance, one could construe distances by
means of the intensity of the forces. 4 There then follows a second stage where the
pressures are grouped into these more basic functional variations and as a function of
time, and then, finally in the third stage, the time is eliminated by substituting
variations of the elements with respect to one another, rather than to a standard
temporal process such as the rotation of the earth:

But what does time mean when we consider the universe? This or that "is a function of time"
means that it depends on the position of the oscillating pendulum, on the position of the
rotating earth, and so on. Thus "all positions are functions of time" means, for the universe,
that all positions depend upon one another. But since the positions in space of the material
parts can be recognized ony by their states, we can also say that all the states of the material
parts depend upon each other. The physical space which I have in mind-and which at the
same time contains time in itself-is nothing other than the dependence of phenomena on one

another. A complete physics, which would know this fundamental dependence, would have no
more need of special considerations of space and time, for these latter considerations would
already be included in the former knowledge. 5

Hence we fmally end up with elements and mutual functional relations, where we
had masses and motions before. This three-part reductive program was Mach's most
complete statement of how he intended to replace mechanical physics with a physics
of elements and functional dependencies.


Newton was the first to separate mass from weight and to determine by experiment that
a body's mass was independent ofits chemical constitution.6 For Newton, ifwe attempt
to accelerate a heavy object, it offers resistance, on earth or even in empty space with no
other bodies in the immediate neighborhood. Mass and weight are proportional because
objects twice as heavy offer twice as great a resistance to acceleration. The same mass
may have many weights on other planets, because of their different gravitational pulls,
while the inertial mass remains the same. Indeed, the proportionality between inertial
and gravitational mass explains why objects of different masses fall at the same rate in a
gravitational field, the extra pull of gravity on the heavier masses being offset by their
increased inertial resistance to acceleration. In addition to these observations, Newton
gave a completely independent definition of mass as the "quantity of matter" contained
in a body, of which more later.

Mach's 1868 definition of mass was not a sharpening or refinement of the stated
Newtonian "quantity of matter" but a new concept of mass ratios (m./llljl). Individual
mass values (a pair of a body and a number, for example) did not exist for Mach; even
the individual terms of mass ratios, like 01a, are themselves ratios. 7 This fact must be
kept firmly in mind. Mass values were the result of a system of chained comparisons
between one ratio and another and with a standard ratio, which could be verified
experimentally. Thus Mach was really defining values always for a system of masses.
To say that mass doesn't vary in a certain event means not that a certain lump of matter
is the same as it was before, but that a systematic relationship between all of the ratios is
upheld. Consequently, there are properties of Newtonian quantity of matter that mass
ratios lack. For example, there is no mass ratio for the entire universe. In addition,
because of the close relationship between Machian mass ratios and kinetic energy-
manifested most directly if the comparisons between masses are made by a system of
momentum-conserving collisions-they can be included within the special theory of
relativity with only a few modifications. 8

Mach's defmition appealed to an experimental procedure, which is left rather

vaguely to "experimental physics." We let two bodies {K.,,~) act on one another by
whatever forces they mutually exert-electrical, magnetic and gravitational
attractions-and forces of contact, such as the vis viva communicated in collisions.

The forces produced in the one body {Ka) cause accelemtions in the other ~). which
must be accepted as a fact of experience.

If we assign to the first body the mass value llla and the other a value lllp. then I11a
acquires an accelemtion !f>ap (which we read as "the accelemtion of Ka due to Kp").
Meanwhile, Kp acquires an acceleration q>13a (which we read as "the accelemtion of~
due to Ka"). Because the forces both bodies exert on each other are equal by action and
reaction, we write

(1) Fap =-Fila

(2) llla q>ap = - 1Djl !f>lla
(3) rnJlllp =- q>p./q>ap.

While the mass definition itself is far from an "elimination" of spatial intuition, it
may have formed a part of a larger antimetaphysical progmm with that goal. That
progmm began with the reduction of relations between bodies over spaces to a system
of reciprocally determining pressures-pressures of contact and pressures due to
fundamental laws of force, and thus with the notion of pressure due to mass as the
centml conception.

Now it is often assumed in the literature that Mach intended for his definition to
reduce "unobservable" pressures or forces to grossly observable spatial bodies and
accelemted motions. 9 That is an understandably stmightforward reading, of course,
since it looks in (3) as if the mass mtios (rnJlllp) are defined in terms of the ratio of their
accelemtions (-q>pa/q>ap). And it would be true if Mach thought space and time were
observable, or suitable primitive terms, which he did not. But since Mach's definition
ultimately refers to a system of reciprocally determining mtios, it is not a classical
defmition, made up of a definiendum and a definiens. In fact, Mach believed that forces
(or pressures) were the readily demonstrable data of experience. In this respect Mach
espoused a belief in the reality of forces, descending from Newton himself 10 Mach
even said in his criticism of Heinrich Hertz's Mechanik (which really did try to dispense
with forces):

To characterize forces as being frequently "empty running wheels" as being frequently not
demonstrable to the senses, can scarcely be pennissible. In any event, "forces" are decidedly in the
advantage on this score, as compared with "hidden masses" and "hidden motions." In the case of a
piece of iron lying at rest on a table, both the forces in equilibrium, the weight of the iron and the
elasticity of the table are very easily demonstrable. 11

In nature, we can immediately demonstrate pushes and pulls by all sorts of natural
potentials by fmding other standard pushes and pulls to oppose them. We can still feel
the pressure in the tension each offers the other without producing motion. For example,
without producing motion, we can slip a platform beneath the iron to replace the
pressure offered by the table, and we can switch the iron for an equivalent test weight.
Mach stated in his notebook that pressure should be measured with pressure [Drnck wie

d Druck messen]. 12 The experimental proposition that masses induce pressures in one
another would be verified by impressing test pressures. Since Druck is experimentally
fundamental, we need not even know yet that it can be further analyzed into the product
of mass and acceleration. Action and reaction are needed, however, and Mach assumed
Newton's third law (Fall = -Fpa) as an unanalyzed fact of experience. 13 I take it that Mach
held to his basic point that forces of nature are derived ultimately from potential
functions, but that perhaps since only differences of potential are observable, forces
would still be considered the basic observable term.

In the "Definition of Mass" article we find that pressures play the primary role.
Newton's second law defined force as a product of mass and acceleration, but pressure
was not a defined term for Mach. It was the common empirical datum from which both
the masses (m) and the spatial accelerations (a) were to be derived. Mach began his
article, in fact, not by mentioning Newton's quantity of matter at all, but rather by
attacking the logical circle in the concepts of mass, acceleration, and pressure due to
mass: 14 "m=p/g, p=mg This is either a very repugnant circle or we are required to
conceive of force as "pressure." 15

The reason there is a vicious circle, as he later observed in his Mechanics, was that
mass and pressure are always conjoined: wherever we have masses they attract one
another and induce mutually opposing pressures, and all pressures considered in
mechanics proper are due to the presence of masses. Nowhere do we find a pressure or a
mass alone. Mach thus proposed to put pressure at the head of mechanics along with
action and reaction:

In that principle of Newton, which is customarily placed at the head of mechanics and which runs
"Actioni contrariwn semper et aequalem esse reactionem: sive corporum duorum actiones in se
mutuo semper esse aequales et in partes contrarias dirigi," the "actio" is again a pressure or the
principle is quite unintelligible unless we already possess the conceptions of force and mass. But
pressure looks very strange at the head of the quite phoronomical mechanics of today. However this
can be avoided 16

He "avoided the problem" by making the fundamental law of mechanics a dynamical

principle: two qualitatively indistinguishable bodies always induce in one another equal
and opposite accelerations (Beschleunigungen):

Proposition of Experience: Bodies placed opposite one another communicate to each other
accelerations in opposite senses in the direction of their line of junction (the law of inertia is already
thereby included). 17

This proposition is tantamount to assuming Newton's law of universal gravitation,

because it includes the knowledge that all masses attract and accelerate toward one
another all the time. In fact, in his historical treatment, Mach claimed that Newton had
arrived at his definition of "quantity of matter" by conjoining masses and their
concomitant pressures. According to Mach, Newton should have been surprised by his
discovery that chemically different bodies possessed the same homogeneous properties
of mass and gravitational attraction, as if their differences in composition didn't count at
the deepest level. 18

If bodies were different all the way down, they shouldn't attract one another equally,
particle-by-particle; rather, the attraction of the first particle for the second should
always exceed or fall short of that of the second particle for the first, if they truly are
different in all respects. Why do not chemical differences among bodies make any
difference to their gravitational properties and lead to unequal attractions? Only if all
bodies were made of the same tiny homogeneous corpuscles should we expect them to
exert similar forces on one another. 19 Newton was thus led to regard every unit of
attractive force produced by matter as due to the attractions of each indistinguishable
corpuscle of matter for one another. Hence, the total attraction of the body is due to the
sum total of all its corpuscles' attractions, i.e., the "quantity of matter." According to
Mach, then, Newton was able to conceive of the quantity of matter as a pure number of
identical atoms in the body.Z0

Mach mentions not masses but bodies (K.,, KjJ) and their accelerations. The mental
picture it conjures up is of extended lumps of matter floating in space at distances that
change according to the accelerations. It would seem, then, that the existence of bodies
and accelerations in space were the primary facts. A possible reading of the mass
definition is thus one in which Mach began with bodies and their tendency to accelerate
one another and then merely defined:

1. What it was for two such bodies (K,K') to be equal in this tendency, and

2. What it was for a first body (K) to be equal in this tendency to a second
(K'), a second to a third (K") and the first to the third.

On that reading, Mach's definition is merely a way of standardizing a capacity of bodies

(K) in a system to accelerate one another and resist accelerations, a capacity we already
knew they had, and for which we simply needed a way of establishing relations of
measure. Ponderable bodies and the distances between them are not eliminated.

A deeper reading of Mach's definition follows the "Raumvorstellungen" program

and has textual support in Mach's other writings and notebooks. The first part of that
program was to reduce the spatial relations of extended bodies to the relations between
pressures, and Mach had declared his intention to make pressures (F) resulting from
potential differences fundamental to mechanics and not derived by definition (F=ma,
F=dp/dt). He added in a notebook entry that he intended to measure pressures with

There exists something delenninative of pressure independent of weight. But it increases and
decreases with weight. Proportionality p=mg p'=m'g' only by equivalent bodies, or one must already
posses the concept of mass, something delenninative of pressure independent of material
constitution. Measure pressure with pressure. What is pressure. Set it free of subjectivity. Through
weight. Mass is: what communicates and receives the m-fold acceleration to an assumed unit

Mach hid the distinction between acceleration [Beschleunigung] and pressure

[Druck] in his earlier 1868 article, switching tenns midstream. These tenns are not
synonymous: with the same Druck we do not know whether we are dealing with a small
mass with great acceleration or a great mass with a small acceleration. Mach seems to
have wanted to measure pressures, or the combination of mass and acceleration, with no
independent means to measure acceleration unconjoined with mass. But this may have
been justified, as all accelerations in mechanics proper are due to the pressures of
contact of masses or the pressures of fundamental forces due to mass. Nor is mass
measurable except through pressure due to mass, which is why Mach originally
complained of a logical circle in mass, acceleration, and pressure due to mass. That
logical circle existed simply because the definiendum and definiens were part of the
same reciprocally defining system of concepts. By accepting Druck as a primitive
concept, the circularity is avoided, as Mach promised, but the look of mechanics is
altered. The problem is now how to specialize certain determinations of the general
concept of pressure into more specific concepts corresponding to the mass and the


It is a delightful historical irony that just as Newton carried his own corpuscularian
views in his pocket, and believed that there could be only one fundamental sort of
matter, so too Mach held a deeper theory of mass standing behind his definition based
on his frequently used analogy of mechanical and thermodynamic concepts. Mach said
that his concept of mass was a direct result of his attempts to apply the principle of
excluded perpetual motion as it is used in thermodynamics to mechanics, 22 and
elsewhere the result of an attempt to connect the phenomena:

My definition is the outcome of an endeavor to establish the interdependence of phenomena and to

remove all metaphysical obscurity, without accomplishing on this account less than other
definitions have done. I have pursued exactly the same course with respect to the ideas of quantity
of electricity, quantity ofheat and so forth. 23

An explicit mention of the analogy occurred in his article, where Mach compared
thermal processes and the acceleration of masses. This mention occurs in an opening
section where Mach maintained that the principle of inertia was not a priori, for it could
always be imagined that masses did not induce accelerations in one another, but rather

If two masses opposite one another gave one another not accelerations, but rather perhaps velocities
dependent on the distances, then there would be no law of inertia. Only experience teaches whether
one or the other takes place. If we had only sensations of warmth, then there would only be
equalizing velocities, which would be zero when the temperature differences were themselves= 0. 24

Equalizations in temperature between a colder and a warmer body in contact occur at a

rate proportional to the temperature differences themselves, while gravitationally

attracting masses equalize their potential differences (their differences of position) by

closing the separation between them at an acce1emted mte, and indeed even the
accelemtion increases as the square of the distance. To a first approximation, however,
differences of position between two mass potentials induce changes of changes of those
initial levels, whereas heat tends to equalize tempemtures by flowing at a constant
velocity, faster or slower depending on the tempemture differences tmversed. In the
Mechanics, Mach indicated that he regarded the accelemtion of masses analogously:

That accelerations play a prominent role in the relations of the masses, must be accepted as a fact of
experience; which does not however exclude attempts to elucidate this fact by a comparison of it
with other facts, involving the discovery of new points of view. In all processes of nature the
differences of certain quantities u play a determinative role. Differences of temperature of potential
function and so forth, induce the natural processes which consist in the equalization of those
differences. The familiar expressions <fulilx2, <furiJ.j, <fuldi, which are determinative of the
character of the equalization may be regarded as the measure of the departure of the condition of
any point from the mean of the conditions of its environment-to which mean the point tends. The
acceleration of masses may be analogously conceived?5

In Mach's conception of the physical processes, the system of masses sitting at

distances was replaced by a set of capacities for attmction standing at their various
potential levels with respect to one another, and moving from one level to another to
equalize those differences. Thus, when Mach speaks of accelemtions in the above
passage of the Mechanics it is only in the sense of Ausgleich, an equalization of
potential level not of spaces per se. For him, this thermodynamic conception of
accelemtion had completely overtaken the Galilean concept of spaces covered in
squares of times. In his 1896 Theory ofHeat, Mach explicitly developed a definition of
the ca1:,acity for heat by an experimental procedure quite analogous to his defmition of

The disanalogy between equalizations of temperature and distance demonstmtes

the need for more than just an energy law in mechanics. We cannot know from the
energy principle alone the separate empirical fact that equalizations of level will be
proportional to the times in the case of tempemture and proportional to the squares of
the times in the case of equalization of distance between accelerating masses. Mach
noted the need for a separate law of motion or transfer in various places in his notebooks
of 1881 and 1882. He says there that the energy law alone would allow a body to float
in the sky so long as its potential and kinetic energy remained constant. The energy law
alone would not even predict any transfer of potential energy to vis viva as opposed to
some other effect, or predict whether tmnsfers through a fall in level would happen
simply with the time or according to the square of the time or some other law. Mach
thus held that an additional Beschleunigungsprincip must give the condition for a
change by stating that masses accelemte:

The energy principle would allow that a body could hang in the air. The principle of acceleration
gives the transformation its direction. It says more. Through a push, brings it out of position. 27

Thus we require a law in addition to the Arbeitsprincip, which could characterize the
action of forces. He wrote in a letter to Popper:
I believe that the concept of force and the concept of wmi<: have their uses. Why limit oneself to just
one... When masses without velocity face one another the relation of work and vis viva cannot be of
use. It does not tell us what will happen. It says only that when the pressure is so great, so is the
change of velocity as great. The force however tells us what happens. Marcus Marci and Mariotte
did not know what to say about the initial velocity of a freely falling body. They dido't know how a
body should begin to move, since it has no velocity as yet, and finally assumed a finite velocity,
which spoiled their investigations. This is a point that makes the force concept palpable.28

The Beschleunigungsprincip was Mach's fundamental law of motion of mechanics:

masses induce accelerations in one another and as much mass falls as can fall, given the
constraints, and becomes kinetic energy. The rise of vis viva is of course treated like a
result of the transformation, in the attainment of a separate (velocity-squared) potential


Mach's revisions to the Newtonian principle of inertia were also evident in the 1868
mass paper. For example, he made good on his claim to make the principle an empirical

Bodies placed opposite one another communicate to each other accelerations in opposite senses in
the direction of the line ofjunction. The law of inertia is included in this. 29

Why is the law of inertia "included" in the law of universal gravitation? One reason is
simply that the law of inertia is just the logical contrapositive ofF = rna (if accelerations
imply forces, then no forces implies no accelerations but either rest or rectilinear
motion). But really, Mach's implicit assumption is that all bodies are continually
accelerating one another when they can. In special circumstances, when all of the forces
on a body cancel in magnitude and direction, the body is at rest or moves at a constant
velocity in a straight line. Hence even the case of a body moving at a velocity in empty
space already implies specialized conditions of equilibrium that require a description of
the dynamical relations between a body and all surrounding masses that might tend to
accelerate it. As Mach wrote as early as in his 1872 notes to the Conservation of
Energy, 30 this can occur in local environments, such as the solar system, and in other
circumstances when a mass stands at a great enough distance from all of the masses of
the universe. However, should the universe "swarm in confusion," the law of inertia
would be revealed as the special equilibrium of forces that it is: a calm eye in the storm
of competing pulls. He also believed that invoking the masses of the universe as the
system of reference solved the paradox that a velocity can preserve its three-component
direction even when it is not evident toward what the directions and velocity are
reckoned. The direction of a velocity is threefold and requires three parametric
equations or a single vector equation since even motion at a constant velocity is an
implicit case of a threefold dynamical relationship. 31 A statement of Mach's dynamical
principle of inertia occurred already in the Conservation ofEnergy:

Now what share has every mass in the determination of direction and velocity in the law of inertia?
No definite answer can be given to this by our experiences. We only know that the share of the
nearest masses vanishes in comparison with that of the farthest. We wonld then be able completely
to make out the facts known to us if, for example, we were to make the simple assumption that all
bodies act in the way of determination proportionately to the masses and independently of the
distance, or proportionately to the distance, and so on. Another expression wonld be: In so far as
bodies are so distant from one another that they contribute no noticeable acceleration to one
another, all distances vary proportionately to one another. 32

In the Mechanics, Mach expressed the principle of inertia for a single mass J.1 and a
system of surrounding potentials (Mi/ri}, thus:

d2/df: ~ M;r;/Mi=O.

With this dynamic theory of inertia the implication is to assume that the resistances
masses offer to acceleration is a direct experimental manifestation of their dynamical
gravitational pull to other masses. But it is a well known fact that Mach resisted all
speculation about a mechanism, or even a need for one, that could convey gravitational
influences instantaneously, or at a velocity such as that of light, from the fixed stars to a
body in a local neighborhood like the solar system, thus stopping this theory in its
tracks. John Norton, who reviewed all of Mach's own pre-Einstein writings on the
subject of"Mach's principle," claims that:

The only unequivocal proposal is that we eliminate the odious notion of space by redescribing the
relevant experiment and law [of inertia] in a way that does not use the term 'space.' If there is a
suggestion of a new physical mechanism that would reach from the distant stars to cause the inertial
forces in Newton's bucket, then the proposal is made vaguely and we are left to wonder whether
Mach endorses it or condemns it as unscientific.33

The assessment is correct, but Norton completely overlooks Mach's

thermodynamic conception of the acceleration of masses-which is mentioned in the
conclusion to the very sections of the Mechanik he considers34 (also quoted above). The
thermodynamic conception was Mach's only endorsed view of the natural processes and
remained so throughout his life. 35 As he says above, Mach considered ''the facts" of
physics to be essentially described by the natural potentials, their potential functions and
laws governing the energy transformation; nothing else. Again we see here an
application of phenomenological physics as the science that does not seek to go behind
energy transformations and seek mechanisms-so Mach remained true to his roots

Robert DiSalle observes in a recent book that Mach's solution to the paradox of the
direction and magnitude of the "constant velocity" in the Newtonian law was a more
modest one? 6 More modest indeed than either of the competing solutions by Neumann
and Lange, which Mach compiled and criticized in later editions of the Mechanics. It is
true, Mach simply noted that Newton's fifth corollary identified an equivalence class of
inertial frames with respect to which the laws of motion were invariant. This
equivalence class is not necessarily the reference frame of the fixed stars or the invisible

absolute space Newton identified with the bucket experiment. So the answer to the
problem is simply that there is no preferred direction and velocity for the body in inertial
motion identified by Newton, but an equivalence class of them to which the other
Newtonian laws of F=ma and action and reaction are indifferent. I must emphasize,
however, that Mach brings this up simply as a way of quelling one sort of problem that
pestered him and his contemporaries, but not as an explanation of what is so special
about uniformly moving inertial frames to begin with. The principle of inertia is a
dynamical principle for Mach, and the inertial frames are cases of dynamical
equilibrium; that is what accounts for their specialness. Just as Mach's history of
mechanics made statics a special case of dynamics, so too Mach made the principle of
inertia a special case of more general dynamical principles that are active all the time,
even for bodies moving at constant velocities. One should not try to cover up this
connection by making claims that are too modest and limited.

An appreciation of Mach's non-Galilean theory of acceleration and the primacy of

dynamics appears to me to be crucial in any attempt to construe a Machian-style Mach's
principle. However, the hypothesis of spatially propagating actions (Wirkungen) from
one body to another, later invoked by Einstein, in a middle period, 37 to advance or
explain Mach's revision of Newtonian inertia, was probably seen by Mach as a step
backward into picture thinking and not an advance toward the future physics of pure
elements and functions that he favored. Even if the agency of inertial forces is due to the
stretching of a colloidal medium surrounding the bodies like an aether-a suggestion
Mach feinted towards38-still overall Mach clearly considered the explanatory style of
mechanism-seeking a mistake. Mach's low opinion of aether theories was given in his
lecture "Allgemeine Fragen der Naturwissenschaft" (Summer 1897):

Magnetic currents, gravitational currents. Only analogies. Pictures. Exactly like caloric. Electricity
as a fluid in which no one seriously believes. 39

Mach does acknowledge in the Mechanici 0 that some natural investigators "will feel
the need of further insight-of knowledge of the immediate connections, say, of the
masses of the universe. There will hover before him as an ideal an insight into the
principles of the whole matter from which accelerated and inertial motions result in the
same way." And he mentions the understanding of actions at a distance to be found in
field theories. But he does not endorse this course.

Indeed, Mach's critique of Newton's bucket experiment might best be summed up

as simply a criticism of space as a legitimate cause of a physical effect. Mach's
argument against Newton is really that accelerated reference frames (uniform and
variable) ought to be treated exactly like Newton's inertial reference frames. Just as we
can choose to see a body traveling at a constant velocity as being at rest with ourselves
traveling in the other direction, so too linearly accelerated and rotating bodies can be
viewed as if at rest by an observer accelerating or rotating in the opposite direction (or
from the point of view of a universe of masses in rotation about a body in the opposite
direction from the body's own rotation). So runs Mach's empirical proposal: when the

earth rotates with respect to the stars, there is no difference in the purely geometrical
spatial relations between that case and the case of the earth at rest with the whole shell
of the fixed stars spinning around it at an immense radius. The two cases are, for him,
one and the same case described twice, but given once. And if these cases are granted to
be spatially symmetrical then space cannot be a cause of any asymmetry in inertial
effects. From this position there are two immediate ways out that are still Machian in

1. Admit that there is an asymmetry in which inertial effects occur only when the
earth spins against the stars and not vice versa, but attribute the difference to some
fact other than spatial relations. In other words, make the presence of inertial
effects a further primitive property of the physical situation. This course was taken
by Sklar41 and recently by Dieks. 42

2. Extend the symmetry further to include the inertial effects, i.e., find some way to
allow for inertial effects to be produced also by the "Ptolemaic situation" of the
earth at rest and the shell of the distant masses in rotation around it. It does seem to
me that Mach indicated his preference for this way out when he said that the
Ptolemaic and Copernican systems were just conventionally different, and when he
hinted: "Try to fix Newton's bucket and rotate the heaven of fixed stars and then
prove the absence of centrifugal forces. "43

Einstein may have accepted from Mach an ideal goal of research that no inertial
system should be privileged over an equivalent set, even when uniformly accelerated
reference frames and rotations (non uniform accelerations) are involved.44 It appears
that Einstein went through an intermediate stage of development in which he attempted
to account for inertial effects physically by means of gravitational effects propagated
through an existing field, thus attributing inertia to a gravitational Wechselwirkung of
masses. 45 He also tried to show that piling up a shell of rotating masses in the
neighborhood of a body could induce tidal forces in a fluid body and encouraged other
workers to advance the analysis. 46 But in presenting his principle of equivalence,
Einstein finally departed completely from Mach in allowing that a uniform gravitational
field could be simulated by a uniformly accelerated reference frame without a source
mass (although not vice versa). As Einstein describes his discovery:

Now it carne to me: The fact of the equality of inert and heavy mass, i.e., the fact of the
independence of the gravitational acceleration from the nature of the falling substance, may be
expressed as follows: In a gravitational field (of small spatial extension) things behave as they do in
a space free of gravitation, if one introduces in it, in place of an "inertial system" a reference system
which is accelerated relative to an inertial system. If one then conceives of the behavior of a body,
in reference to the latter reference system, as caused by a "real" (not merely apparent) gravitational
field, it is possible to regard this reference system as an "inertial system" with as much justification
as the original reference system. So, if one regards as possible gravitational fields of arbitrary
extension which are not initially restricted by spatial limitations, the concept of the "inertial system"
becomes completely empty. The concept "acceleration relative to space" thus loses every meaning
and with it the principle of inertia and the paradox ofMach.47

According to Einstein, there should be no difference between a system of coordinates

centered on the earth at rest with the universe spinning around it and a system with the
universe at rest and the earth spinning in the opposite direction. As this case involves
non uniform accelerations for particles in the rotating mass there is not perfect
agreement, even among experts on this issue, whether this can be the case. The question
here for the general theory of relativity is not whether an arbitrary uniform acceleration
can be replaced by rest and a gravitational field, but whether any arbitrary non-uniform
accelerated motion due to the forces of nature can be similarly transformed away, i.e.,
into a non-uniform gravitational field.

The preferred view at present seems to be that the two cases are indeed
distinguishable by observers using inertial effects on the spinning earth which cannot be
transformed away simply by observing the earth in an accelerated reference frame
rotating in the opposite direction around the earth. One would have thought the opposite
was the case in Einstein's theory. 48 (And indeed Einstein is quoted by Norton as having
extended his principle of equivalence to rotating reference frames. 49) However, we are
not allowed to equalize an arbitrarily varying pull of nature by substituting an opposite
variable acceleration. The best we can do is break up a variable acceleration into a series
of short motions at constant acceleration and then apply the infinitesimal principle of
equivalence to each short segment. But that infinitesimal principle certainly does not
meet the Machian criterion we met with in the last chapter that the dynamically more
complex principle determines the simpler, and not vice versa, as Mach claimed in the
ideal treatment of mechanics that the dynamical D'Alembert's principle implies the
static principle of virtual velocities. As Julian Barbour has pointed out, relativity should
be interpreted the same way, with special relativity remaining merely as the static case
of the dynamical general theory. 50 The inertial frames of the special theory are specified
by dynamical equlibrium, not as a property of some underlying Minkowskian flat space.
Barbour has also been the leading defender of interpreting general relativity as a
perfectly Machian theory in which rotation is relativized, not via the more radical
Machian elimination of space, but rather through the limiting of the theory to relative
spaces or separations between particles and properties defined over these. 51

However, for linear accelerations, the presence of an inertial force manifested by an

accelerating body in one reference frame is replaceable by effects of a simulated
uniform gravitational field on a body at rest in a reference frame accelerated in the
opposite direction, so at least in this special case Einstein succeeded brilliantly in
relativizing the reference frames of accelerated motions. As Cornelius Lanczos explains
it, however/2 Einstein's principle of equivalence turns out to be a consequence of
generalizing D'Alembert's principle. The impressed forces (F) due to a gravitational
potential and the reverse accelerations (-rna) of the inertial forces are taken to be not
only equal but actually exchangeable for one another by imparting a (constant)
acceleration to the reference frame. In other words, by accelerating backward toward a
body affected by a force we "transform" the inertial forces back into greater impressed
forces derived from a "sourceless" gravitational potential, leaving the total the same as

What would Mach's way out of "Mach's Paradox" have been? It seems natural to
inquire because of the importance Einstein attributed to the problem. Mach's chief
overriding concern in treating questions of mass and inertia seems to have been the
elimination of spatial conceptions from physics, especially from the interactions of
masses, according to his "Raumvorstellung" program. Einstein, by contrast, did not
eliminate spatial conceptions from his physics; rather, as Bertrand Russell remarked, he
started with coordinates as numbers arbitrarily assigned in space and then proceeded to
remove all physical significance from them.

It is natural to ask: could we not dispense with coordinates altogether, since they have become little
more than conventional names systematically assigned? perhaps this will become possible in time
but at present the necessary mathematics is lacking. We wish, for example, to differentiate, and we
caonot differentiate a function unless its arguments and values are numbers ... To define such a
process is a problem in mathematical logic, probably soluble but unsolved hitherto. If it were
solved, it might become possible to avoid the elaborate and round-about process of assigning
coordinates and then treating all their properties as irrelevant, which is what is done when the
method of tensors is employed. 53

Russell himself makes a try at eliminating coordinates in terms of simpler qualitative

properties in the Analysis ofMatter, which I think would also have been Mach's way of
approaching the inertia problem, i.e., calling for the elimination of arbitrary or
mathematical coordinate systems altogether in favor of purely physical properties which
would then be used to determine position and distance as derived concepts. Dieks54 puts
forward a proposal like this in which Hamiltonian total energy functions and potential
differences are taken as the basic physical properties between points, which are then
used to lay down a coordinate system for determining the positions and the distances
between the point-potentials, a choice which offers several relationist alternatives and is
not committed to absolute space, since the spaces are defined over the physical events
such that they leave those events invariant. Never mind that positions and velocities
(their time derivatives) appear in the Hamiltonian as one can imagine individuating
them by their physical properties at points. This seems the most Machian approach of
all, especially because of the use of the potential energy functions to determine positions
and distances, as Mach himself had recommended. But Einstein, too, may have believed
in the elimination of prior coordinate systems, especially as he criticized his own earlier
use of ideal clocks and rigid meter sticks:

One is struck that the theory (except for the four-dimensional space) introduces two kinds of
physical things, i.e. (I) measuring rods and clocks, (2) all other things, e.g., the electromagnetic
field, the material point, etc. This, in a certain sense, is inconsistent; strictly speaking measuring
rods and clocks would have to be represented as solutions of the basic equations (objects consisting
of moving atomic configurations), not, as it were, as theoretically self-sufficient entities. However,
the procedure justifies itself because it was clear from the very beginning that the postulates of the
theory are not strong enough to deduce from them sufficiently complete eqnations for physical
events sufficiently free from arbitrariness, in order to base upon such a foundation a theory of
measuring rods and clocks. If one did not wish to forego a a physical interpretation of the co-
ordinates in general (something which, in itself, would be possible), it was better to permit such
inconsistency-with the obligation, however, of eliminating it at a later stage of the theory. 55

Mach and Einstein actually seem quite close here in sharing the goal of an
elimination of space from fundamental science, not just relativizing space to spatial
relations. One surviving (undated) letter from Einstein to Mach56 has him saying that
"for me, it is absurd to ascribe physical properties to 'space"' (Fiir mich ist es absurd,
dem "Raum" physikalische Eigenschaften zuzuschreiben). And he adds: "The reference
system is, so to speak, appropriately fitted to, or measured against, the existing world
with the help of matter-energy and loses its nebulous aprioristic existence." (Das
Bezugsystem ist der bestehenden Welt mit Hilfe der Energiestoffe sozusagen
angemessen und verliert seine nebulose aprioristische Existenz). 57

How would Mach have proposed to eliminate space so completely that even the
coordinate system was derived from physical properties? A likely reconstruction of
Mach's reasoning runs as follows, if we assume that the "Raumvorstellung" paper
indeed laid down the program to be followed in dealing with both masses and distances.
There is no such thing as a body that does not accelerate according to universal
gravitation; even inertial motions are just special cases of equilibrated forces. Thus, all
bodies will press one another according to their capacity and intensity factors for
inducing pressure. Wherever we find matter, we find pressure, whether at rest or
accelerating. We can feel the pressures of contact resident in the body by trying to
accelerate it, or collide with it, and we can feel the pressures due to fundamental forces
by bringing sufficiently large bodies together. Hence the reduction of bodies and
distances solely to considerations of pressure, a first stage of reduction. The constant
conjunction of force with matter thus leads us to believe that there is a capacity for
acceleration (or resisting it) resident in matter independent of its chemical composition
or other properties we discover in it, such as magnetism and electricity. Gravitational
and inertial mass exemplified but one of a number of capacities for energy possessed by
bodies. 58 Thus, we need to isolate this particular capacity from other capacities of a
body due to its peculiar chemical composition.

Once a capacity for generating (and resisting) acceleration has been isolated from
electrical and magnetic attractions, we discover the fact that if two bodies are equal in
their capacity to accelerate one another and one of them is equal to a third, the other is
also equal to this third. The separation of this capacity from the specific composition of
bodies is the first thing Mach's definition aimed to prove, thus the mass concept
represents a standardization and metricization of capacities that we already know they
have. Mach gives mass a unique meaning, even though it does not appear alone in
experience, independent of its intensity factor, the distance. This is why he insisted that
mass be defined first for two bodies m,m of identical chemical composition first, at a
fixed distance from one another, an experimental procedure that holds steady both the
capacity and intensity factors. Then we can consider extending the concept to bodies of
heterogeneous chemical compositions.

But for this extension we need an argument. Thus, Mach says that if bodies did not
press one another equally, either by attraction or by contact, then it would be possible to
construct a perpetuum mobile. Take the pressures produced by contact.

Say we set three bodies on a frictionless ring and start one in motion. 59 If the collision
produces more vis viva in the second body than was in the first, and more in the third
than was in the second, then when the machine goes through one complete cycle and the
vis viva is communicated back to the first body, there will be more kinetic energy in the
ring than there was to start with-a perpetuum mobile. Mach's proof mimicked exactly
(and is no doubt patterned on) Sadi Carnot's argument that the quantity of heat
exchanged between two bodies depends only on their temperatures and not their
chemical compositions.60

Mach assumed that two absolutely equal bodies at a distance from each another (a
fixed difference in potential level) will produce equal and opposite pressures, which he
will measure by means of other motionless test pressures:

Mach then assumed some separation between the experimentally fixed capacity and
intensity factors so that he could analyze pressure into masses and accelerations:

Hence the second real problem to be solved. There is no way to separate the mass
values and the pressures of accelerations completely. It is possible, however, to fix each
of them independently of the other by carrying out a set of systematic comparisons of
the pressures induced on a unit capacity for resisting acceleration at a set of various
distances (where, by the Beschleunigungsprincip, we know that the accelerations are
inversely proportional to the distances squared). This is the real reason why there are
mass ratios but no absolute masses. In these physical speculations, treating distance,
acceleration, and velocity in a very different way from the concepts used in Galilean
mechanics, Mach took certain concrete steps toward the realization of his 1866
"Raumvorstellungen" program. The mass definition, with the Besch/eunigungsprincip at
its head, though not the completion of the work, was a stop on the way. The "states of
matter" he referred to in the "Raumvorstellungen" paper are pressures (F, F', F" ... ) due
to contact (inertial masses) and fundamental forces (gravitational masses). These are the
basic elements in their reciprocal functional dependence on one another and are not
explicitly differentiated The functions that hold between these pressures must then be
specialized into two groupings: those that vary according to the mass factor, and those
that vary due to the distance factor. These two factors are fundamentally entangled in
the empirically observed pressures (potential differences), but it is possible to separate
them in thought by applying two different standards of comparison: a variety of
pressures due to mass capacities (m, m', m" ... ) at the same potential level (r); the same
mass capacity (m) at a variety of potential levels (r, r', r"). This last system of physical
constancies is probably where Mach would find his concept of a coordinate system
represented by a matrix of experimentally discovered equipotential levels and surfaces,

not numbers applied arbitrarily to space and then melted away in the further
development of the theory.

For a truly experimental definition of mass, a technique is required for fixing both
factors independently, which is provided by the methods of comparing test bodies in
"experimental physics." One could say that the same mass value is identified as the
same factor for pressure across numerous distances, and the same distance can be
identified as the same factor for inducing pressure between different masses, thus
deriving both of the spatial factors of the fixed bodies involved and the distances
between them from the pattern of pressures, holding one fixed and then the other. And
indeed Mach does say these are the only facts of the matter that "all bodies act in the
way of determination proportionately to the masses and independently of the distance,
or proportionately to the distance and so on [i.e., independently of the masses]. "61


Mach reflected on his mass definition in letters to his friends. Here he seemed to be
trying to convey the notion that the mass definition was similar to a concept of space. I
interpret this to mean that his definition standardizes the capacity factor of mass the
same way that a metric definition for space standardizes a certain length through
motions from place to place. In a letter to Josef Popper-Lynkeus of December 30, 1883
(right after the first edition of the Mechanik), Mach wrote: "Concerning mass I am
convinced that I have hit the mark in not making this a fundamental concept. It is not a
concept of experience like length, but rather a concept of space. "62 Mach also
corresponded extensively with Friedrich Adler about the latter's own mass-concept and
how it differed from Mach's:

For me the concept of mass is a dynamical concept of measure [Massbegr!ff], which as such
asswnes a relation to a comparison body and nothing else. When bodies dissolve, evaporate and so
forth, this can cause difficulties in the application of the concept but does not affect the concept
itsel Everywhere that the concept of mass or force presents itself in physics, only the mechanical
side of the process is meant; only ponderable masses come into consideration, and reciprocal action
against another ponderable body must take place. I have already said for 30 years that the validity of
the same measure of the mass for arbitrary physical circumstances is an hypothesis that has been
valid until now... When we speak of the conservation of"mass" or the conservation of"rnatter" that
can only mean one thing: in all changes of a closed system the swn of the mass-values of the
bodies, in relation to one determinate comparison-body remains constant. In "matter" I see nothing
but the proportionality of all physical capacities [Capacitiiten] to one another.63

What that last sentence seems to mean is that bodies also have thermal, electromagnetic,
and gravitational capacities, and that matter as such for Mach is nothing more than a
proportionality of these through transformations of the energy from one form to another.


Mach's Wiirmelehre, published in 1896, was an attempt to derive a "clean" version of the
theory of heat by freeing its concepts from the accumulation of historically
conditioned attitudes and metaphysics. Having written the history of mechanics in its
conceptual development, Mach then set out to do the same for the theory of heat, which
had a still more complicated history. His philosophical heroes in this quest for clarity
were Fourier, Kirchhoff, Joseph Black, Sadi Carnot, Robert Mayer, and William
Thomson. Rudolf Clausius would also be included on this list but for the fact that Mach
found his mathematical formulations artificial and misleading. As one reviewer pointed
out, Mach also, strangely, did not include the classic work of J. Willard Gibbs in his
review, which one would think belonged there.

The immediate cause of publication seems to have been the 1895 Lubeck
Natuiforscherversammlung, in which some of Mach's views had come under attack, in
particular his use of Zeuner's analogy and his analogs of potential and potential function
for the different forms of energy. A special section of the Wiirmelehre entitled "The
Opposition between Mechanical and Phenomenological Physics" contains Mach's very
paradoxical rebuttals to the critique at Lubeck (of which more later).

Mach clearly felt that the theory of heat was more ad hoc than mechanics and that it
lacked the economical and conceptual clarity of the other field. The Wiirmelehre begins
straight off with an unusually harsh critique of the concept of temperature. This opening
may have been confusing to his readers, as up until this point Mach had used
temperature as a model example of "potential level" in nature and actually used it to
interpret velocities and heights of weights as concepts of the same kind. And this view
did not change. But here, Mach complains that the correlation of volume expansion and
thermal state was originally just a straightforward convention. The expansion was taken
as a sign or indicator of thermal state, but not an identity whereby all properties of
thermal states would also be properties of volume expansions, such as a limit to

Mach critiques Amontons's concept of absolute zero expansion by saying that in a

sense the inference to such a state is no good. By tracing back the expansion coefficient
of a body for a difference in temperature an absolute zero of the body's thermal state is
obtained at the limit of zero volume expansion or a kind of collapse of a power to
expand. But the coefficient of expansion is just a mark or a correlate of the thermal
state, Mach says, not the state itself, and the properties of "a system of signs" should not
be extended to the properties of the things signified, since the correlation may
unexpectedly break down. Mach then says a very interesting thing. He notices that it is

not only the volume of bodies that alters with temperature but actually a whole cluster
of properties:

Concomitantly with the !henna! sensation which a body provokes in us, other properties of the
body also undergo alteration--, as for example, its electric resistance, its dielectric constant, its
thermoelectric motive force, its index of refraction, etc. And not only might these properties be
employed as indices of the thermal state, but they actually have found such employment. In the
preferment of volume, therefore, as a test of states of heat, there is involved, despite the manifest
practical advantages of the choice a certain caprice; and in the general adoption of this choice, a
convention. 1

He then maintains that although temperature does not manifest the characteristics of a
multiple manifold, it would not be out of the question logically if it were found that to
the thermal state corresponded rather a manifold of physical properties besides volume

If we were to take for example as our criterion of the state of a body K the pull exerted by K on an
iron ball suspended from a balance, these pulls, the aggregate of which as symbols likewise
constitute a simple manifold, could be determined indifferently by the electric, magnetic and
gravitational properties of K and would be the symbolic correspondent consequently of a threefold
manifold. Inquiry must determine in each case whether the symbolic system chosen is the
appropriate one?

Mach states that he is more amenable to an absolute zero expressed in terms of

force of expansion, rather than an expansion coefficient, where at least one does not
need to conceive of "the distressing idea of a body losing its volume when it loses its
heat." But Mach is actually no friendlier to this notion, as he remarks that at -273 C or
0 K, it is still conceivable to him that a body might still contain energy of heat, but not
such as could be communicated through a possible expansion of volume, which is its
usual conventional and reliable partner. 3 Thermoelectric currents, exchanges of heat to
electricity, are conceived by Mach as possible additional carriers of heat energy. Thus,
one of those other properties Mach speaks of as potential indicators of the thermal state,
is actually suggested by him as a candidate. Thus, although Mach saw temperature as a
genuine potential level in nature, he rejects the correlation between volume expansion
and temperature as "a matter of luck" and instead assimilates temperature to a more
general criterion for thermal state which may conceivably involve many different
properties besides volume expansion and thus might have been similar to the multiple
manifold of physical properties he called the "chemical potential level":

The concept of temperature is a concept of level, like the height of a heavy body, the velocity of a
moving mass, electrical and magnetic potential and chemical difference. Thennal action takes
place between bodies of different temperature, as electric action does between bodies of different
potential. But whilst the concept of potential was deliberately framed in perfect consciousness of
its advantages, in the case of the concept of temperature these advantages were a matter of good
luck and accident. In most departments of physics the differences alone of the level values play a
determinative part. But temperature appears to share in common with chemical level the property
that its level values are per se determinative. The fixed fusing points, melting points, boiling
points, critical temperatures, temperatures of combustion and dissociation are obvious instances.'

Mach reiterates his belief at the end of the book, where he says that the analogy between
heat and other energies is in part conventional and historical and does not explain the
fundamental asymmetries in behavior between heat and other energies.

It is quite a chance historical circumstance that Amontons, in papers of 1699 and 1702, fell upon
the idea of measuring temperature by the tension of a gas. He did not think of work performed by
heat. But from his method resulted that the numbers which express temperature are proportional
top the tensions of a gas, and therefore to the work performed by a gas with equal variations of
volume. Thus it happened that heights of temperature and heights of the level of work are
proportional to one another. This relation was first consciously established by William Thomson
in his scale of absolute temperature. If characteristics of thermal state which strongly deviate from
the tensions of gases had been chosen, this relation might have turned out to be very complicated
and the coincidence considered at the beginning between heat and other energies would not have
held. It is very instructive to reflect about this. There is no law of nature in the conformity in the
behavior of energies but this conformity is conditioned by the uniformity of our way of viewing
the facts and it is partly a matter ofluck. 5

I believe a hint of Mach's actual view of the matter is evident here; that it was
possible to correlate the work of heat to other mechanical equivalents, as if the capacity
of a body for heat and its temperature corresponded to the concepts of potential and
potential function established in mechanics and electricity. However, this is just
"lucky." In fact, for Mach, if experience justified it, heat could turn out to be a multiple
manifold of properties and more like chemical potential level actually is for him, in
which, as Mach says, a single chemical level value actually corresponds to a combined
manifold of level values of the other energies.

We need not shrink from the supposition of a multiple manifold of level values. The familiar
physical level values are it is true, level values of work, and as such simple manifolds. But though
the square of the velocity represents only a simple manifold, yet velocity, acceleration and so on,
since they are directed magnitudes, exhibit a three-fold manifoldness. These quantities are level
values although not those which are alone taken into consideration by the modem principle of
energy. But the principle of energy cannot settle all physical questions. Electricity and magnetism
stand to one another in much the same relation as real and imaginary quantities. If in this sense
level values of both domains are regarded as belonging together they form a system of six-fold
manifoldness. Much the same might be the case with the chemical potential; but the fixed points
would be wanting, to begin with, for determining the dimension number of the manifold. 6

Now, Mach often writes as if heat is just a peculiar natural potential among others
to be fitted into this general multiple manifold of level values, which he calls the
"chemical view." But there is just a hint in the above remark about thermal state as
indicated by multiple physical indicators that Mach was capable of thinking of
temperature the way he does of chemical potential level. Temperature, for Mach, is
merely historically taken to represent a simple continuum of level values, namely
temperature as correlated to volume expansion. It just so happens that in the case of
heat, for somewhat controlled circumstances of reversible cyclic processes for example,
this manifold property can be made to correspond to a simple continuum of level
values, of say mechanical energy. Otherwise, in dynamic processes, the heat's hidden
multiplicity of level values outruns the mechanical equivalent, as the volume or the
plane outruns the multiplicity of the line, or at least that is the straightforward

implication. But whether heat is absorbed into the manifold of other energies or whether
heat and chemical energy are the same, I think that the chemical manifold was the all-
encompassing Weltanschauung of Mach, for various reasons to be considered in due
course, and as he ends his book by relating thermal and chemical phenomena it seems
to me likely that he thought of them (or wished to think of them) in the same terms.

Mach treats the theory of the conduction of heat in very sympathetic terms,
beginning with Newton's law of cooling, viz., that the temperature difference between a
warmer body and a colder body, or colder surroundings, changes at a rate proportional
to the greatness of the temperature difference. Or sometimes Newton says that the loss
of heat to the surroundings is as the total heat present in the warmer body. As this total
heat increases, so does the rate of loss. There is no separation of the quantity of heat,
which behaves like a fluid, from the temperature. Indeed it is quite the opposite as a
temperature change is related instead to the intensity of heat flux through a surface.
Fourier's theory of conduction, which describes the flow of heat by conduction
according to the Newtonian principle of cooling is described by Mach as an "ideal
physical theory" and as the following of one great fact through a series of apparently
disconnected phenomena:

Fourier's theory is based upon the Newtonian principle of proportionality between difference of
temperature and velocity of equalization. The conducting powers ana capacities for heat
determined the factors in the proportions just as the masses do in the emchanical case. Distances
with bodies gravitating toward one another, and temperatures with bodies of unequal temperatures
tend to become equalized; only in the former case accelerations of equalization are determined by
the differences of distance, in the latter, velocities of equalization are determined by differences of
temperature. In saying that every material point tends to the mean temperature of the surrounding
points, the result of Fourier's theory is so expressed that it appears almost self-evident and very
close to our instinctive perception. It lies as close as the observation that heavy bodies left to
themselves sink ... In mechanics and in the theory of conduction of heat it is, really, only one great
fact in each domain which is ascertained. 7

If Newton and Fourier recognized the basic law of motion of heat according to
difference of temperature, corresponding to the Beschleunigungsprincip for mechanics,
there was still another analogy to be exploited between the excluded perpetual motion
of mechanics and the excluded perpetual motion of heat energy. The Camot engine is
the principal example formalizing the analogy. But Mach also mentions the application
of the principle in the case of the radiation of heat. It was discovered by numerous
investigators that bodies can also equalize their temperature differences by radiating
heat. For example, Pictet in 1790 published his discovery that if a body is put at the
focus of one spherical mirror and a blackened thermometer is put in the focus of the
other mirror, the body can raise the temperature of a thermometer without touching it.
The case of radiation of heat was generally recognized to be different than conduction
and convection in that heat rays behaved like light rays, for example in acting across a
vacuum, being reflected by mirrors, and absorbed for different wavelengths and colors
of surfaces. Now, this does not seem to have anything to do with perpetual motion on
the face of it.

But an application arises for bodies at the same temperature that are capable of
radiating one another. Intuitively, Mach says, it is seen that for all the differences in the
ability of the bodies to emit and absorb radiant heat, they cannot fail to irradiate each
other equally. If they did not, a body would become colder in emitting energy to another
body becoming warmer. Heat would thus flow from a colder to a warmer body against
its natural tendency, which is for Mach the conceptual equivalent of a body rising of
itself in mechanics. From this very simple premise it was derived that, in general, the
emissive power of a body is proportional to its power to absorb radiation. The body that
gains heat energy easily must lose it just as easily. A number of other results followed
for the energy exchanges, making equilibrium independent of (1) each orientation of
surface (Fourier, Lambert, Leslie), 8 (2) each wavelength of light exchanged and each
plane of polarization (Kirchhoff) 9, and (3) the velocities of propagation of the heat rays
through each medium (Clausius). 10 Mach compares the use of the equilibrium of
radiation ofheat to the endless chain ofStevinus:

The intensity of the radiation of different bodies may be very different without the equilibrium
being disturbed. The surface elements may have the most various orientation. Selective absorbtion
is different for different bodies and different wavelengths. It is likewise different in regards to the
kinds of polarization. It does not matter if the bodies taking part in the equilibrium of temperatures
are immersed in different media. From these facts, discovered by a particular observation, together
with the continuance of the equilibrium of temperatures arises a particular inference which
appears as a postulate of the supposed equilibrium and renders this equilibrium intelligible.
Perhaps in no other so small domain may the adaptation of ideas to the facts which they represent
and the adaptation of ideas to one another, be so beautifully observed as in the one just
considered. 11


Not all of the Wiirmelehre is so skeptical of the value of the theory of heat as an
autonomous science, albeit strongly influenced by mechanical analogy. Indeed, Mach
approved of the analogy and found in Carnot's reversible cyclic process the first
example of excluded perpetual motion argument extended outside mechanics. Carnot's
argument impressed Mach as the proper way to attack other branches of physics,
without annexing them as departments of the mechanics of particles and forces in space.
Mach was not against mechanics in its proper sphere, of course; he was simply
interested in showing that the use of perpetual motion arguments, potential functions,
and the law of least action were general forms of phenomena that applied to happenings
in nature in general:

Physical processes present numerous analogies with purely mechanical ones. Differences of
temperature and electric differences equilibrate themselves in a similar way to the differences of
positions of masses. Laws which correspond to the Newtonian principle of reaction, to the law of
conservation of the center of gravity, to the conservation of the quantity of motion, the principle of
least action and so on, may be set up in all physical domains. These analogies may be made to rest
upon the assumption which the physicist is fond of making, namely, that all physical processes are
in reality mechanical. But I have long been of the opinion that we can discover general
phenomenological laws under which the mechanical ones are to be classed as special cases.
Mechanics is not to serve for the explanation of these phenomenological laws but as a model in

form and as an indicator in searching for them. The chief value of mechanics seems to me to lie in
. 12

Thermodynamics-even today-is interpretable in two nonequivalent ways: (1) a

mechanical-statistical way based on the hypothesis that heat just is molecular vibration,
and (2) a phenomenological way in which only energy-transfers between heat and work
are considered as if these quantities were equal but not identical. In phenomenological
thermodynamics the perpetuum mobile principle (of which there are said to be two, one
for heat and one for the other energies) is a postulate of experience. The use of the word
"reversible" to describe a Carnot engine, or even other infmitesimal processes in
thermodynamic equilibrium, means that heat is not allowed to flow from warmer to
colder bodies but only between bodies of the same temperature, and bodies are raised
by heat expansion only very slowly, so that the pressure of the gas is always balanced
by the pressure of the weight.

Whenever work is performed by means of heat a certain quantity of heat passes from a warmer to a
colder body (supposing that a permanent alteration in the state of the acting body does not take
place). To the performance of work corresponds a transference of heat. Inversely, with the same
amount of work obtained, one can again transfer the heat from the cooler body to the warmer one.
Camot, now, found that the quantity of heat flowing from the temperature t to the temperature t1
for a definite performance of work, cannot depend upon the chemical natures of the bodies in
question, but only upon those temperatures. If not, a combination of bodies, which would
continually generate work out of nothing, could be imagined. Here, then, an important discovery is
founded on the principle of excluded perpetual motion. This is without doubt the first extra-
mechanical application of the theorem 13

The Carnot engine can be represented by a single piston capable of moving without
friction from a high temperature source of heat to a low temperature source in such a
way that the theoretical maximum of work done by heat is realized in a reversible way.
It is therefore an ideal process or limiting case that cannot be realized, as Mach adds. 14
For this it is necessary that

1. Bodies at unequal temperatures never come into thermal contact.

2. All the variations of temperature that occur are consequences of variations
of volume (the piston head is always pressed down on the gas with a
pressure equal to the pressure of expansion).

We consider a high temperature source T 1 and a low temperature sink T0 . Two

bodies with very great (practically infmite) capacities for heat are used. Between them
is an absolutely nonconducting insulator. Mach was well aware of the need for
idealization in physics 15 and encouraged it for clearly picturing and isolating facts in
thought. We start out the piston on the high temperature source T 1 The gas in the piston
is already at temperature T 1 and raises the piston heat uniformly by absorbing heat at
that temperature from the heat source. The gas would ordinarily cool off, but because it
takes a quantity of heat Qlf1 from the high temperature source, it remains at the same
temperature as heat flows through it. This is called an isothermic expansion. Now we
move the cylinder over the insulator.

The gas is allowed to expand further as the pressure on the piston head is relaxed, but
this time is allowed to fall in temperature to To without a change in volume (this is
called adiabatic expansion). Now, when the piston has reached the temperature of the
low temperature sink T0 we move it over there and press down the piston head until the
original volume is reached. Ordinarily this would cause an increase in temperature (as
pressing on a bicycle pump produces heat in the tube by increasing the pressure) but
since the heat is allowed to flow out through the bottom plate to the low temperature
sink, the temperature remains constant and a second quantity of heat Q0 is expelled
there. This is an isothermic compression. Finally, we move the cylinder back over the
insulator and keep the pressure on without decreasing the volume this time allowing the
temperature to increase. This is called an adiabatic compression. If everything goes
according to Camot's assumptions, the final temperature and volume reached are the
same as the initial temperature and volume with which the cycle began, thus the process
can be repeated.

Because all the stages of the cycle are states of equilibrium, thermal or mechanical,
the cycle can also run backwards, using work done on the piston to transfer heat from
the low temperature sink to the high temperature source. Camot proved that the heat
used to produce work in the Camot cycle did not depend on the chemical nature of the
bodies the heat flows through, but only on the temperatures. If this were not true and,
for example, the work done by heat in flowing through hydrogen could do more work
than the same quantity of heat flowing through air between the same temperatures, we
could use the hydrogen process to generate work in one piston (and reset itself) and then
use the work produced to run the air piston in the opposite sense (replacing all of the
heat back to its original source) and still have work left over. Thus, it must be the case
that the work done by heat between temperatures is independent of the composition of
the bodies it flows through. 16

Camot originally thought that the quantity of heat Q 1 that flows into the piston at
the high temperature source was the same as the quantity of heat Q0 expelled at the low
temperature sink, the way that 100 gallons of water does work by falling over the
paddles of a water mill while remaining the same quantity of water. There is a logical
gap in the argument on just this point, and indeed Camot privately changed his own
mind on the question, as Mach notes. Later, J.P. Joule and J.R. Mayer demonstrated
that a quantity of heat can actually disappear in becoming work. It was also noticed that
when a steam engine was actually driving a locomotive, less heat was passed to the
condenser (the low temperature source) than when the engine was idling. Clausius
fmally resolved the problem and gave thermodynamics its modem form. He said that of
the heat Q1 absorbed by the piston at the high temperature source T 1 some of it is
transformed into work and disappears in pushing up the piston-head, and what remains,
Q0, is expelled at the low temperature sink T0 . There are thus two parts of the heat (Q =
Q1+Q0) to consider in the reversible cyclic process of Camot: the heat transformed to
work, and the heat that "falls" from the high to the low temperature level. It turns out, as
Clausius discovered in formulating the second law of thermodynamics, that the amount
of heat converted to work in the Camot engine is proportional to the amount that flows

from the high to the low temperature. This fact is written:

Where Q~, Qo, T1. and To have the significance given above). Mach writes it thus:

While the "economical" coefficient: the ratio of the amount of heat converted to work to
the total quantity of heat is proportional to the difference of temperature between which
the engine operates is:
..Q1-=...Qo = I1..:....Io
Q1 +Qo T1


Mach's claim was that so long as a reversible cyclic process consisting only of
equilibrium states is concerned, there is no disanalogy between heat and other energies
and the manifold of level values is equivalent. The entropy, or the amount of heat that
flows from the high to the low temperature without performing work, the term Qo(l!f0-
l!f1) is equal to the amount of energy converted to work in the piston and thus by the
application of that work can again be made to flow, in a reversible process at least, back
from the low to the high temperature-which heat would never do spontaneously. The
difficulty stems from the fact that when heat flows between unequal temperatures
(unlike in the Camot cycle), as in most natural processes, the process cannot be
completely reversed without expending additional work from outside the system. As
Mach points out, other energies are not like this. Even when equilibrium is not
maintained, these energies can be returned to their initial states with the same amount of
work liberated by displacing them from those states, like the pendulum, which, without
resistance, would attain the same height it fell from.

Most crucially, Mach did not believe a perpetual motion of the second kind was
any explanation for the irreversible nature of heat flow. The analogy between the
reversible equilibrium process for heat and the other energies led Mach to criticize what
he called the "Carnot-Clausius principle" and the postulates of Clausius and Lord
Kelvin. He claimed these added nothing to our knowledge not already obtainable from
the behavior of other energies according to the classical excluded perpetual motion:

The mere exact knowledge of the conservation of energy is sufficient to obtain the theorem of
Carnot and Clausius; and because this theorem holds for the different fonns of energy a special
position for heat is not conditioned by this theorem. 18

The theorem of Carnot and Clausius, for Mach, is not yet a different law than the
first law, so long as all processes are reversible, and does not involve the special
irreversibility of the flow of heat from hotter to colder bodies, which Mach of course
acknowledged. Mach did not see that the postulates of Clausius and Lord Kelvin or the
Carnot-Clausius law as yet invoked such special irreversible properties. Lord Kelvin's
postulate states:

It is impossible, by means of inanimate material agency, to derive mechanical effect from any
portion of matter by cooling it below the temperature of the coldest of surrounding objects. If this
axiom be denied for all temperature it would have to be admitted that a self-acting machine might
be set to wmk and produce mechanical effect by cooling the sea or earth with no limit but the total
loss of heat from the earth or sea, or, in reality, from the whole material world. 19

Clausius's postulate states that heat cannot be transmitted from a colder to a hotter
body without adding additional work from the environment. 20 Mach remarked that "as
soon as we regard the equalization of heat connected with the difference of
temperatures as a source of work" these two postulates are equivalent and collapse into
Carnot's perpetual motion principle, itself identical with the argument that supports the
first law.

The real asymmetry, according to Mach, is that other energies fall in potential level
only when they can transform and raise others up. Thus the potential to do further work
is preserved in a different form. If electrical discharge lifts a weight, the electricity falls
in its potential to do further work, but the weight is at a height and can do further work
by falling. Thus the potential to do work is always preserved in another form. But heat
can fall in potential level, in temperature, without raising up any other potential as a

Every transformation of a kind of energy A is connected, in the case of heat as well, with a fall of
potential of the kind of energy considered. But while, for other kinds of energy, with the fall of
potential a transformation is inversely connected--and consequently a loss in energy of the kind
which falls in potential level--heat behaves in another way. Heat may suffer a fall in potential
without experiencing a loss of energy--at least according to the usual way of measuring it. If a
weight sinks, it must necessarily generate kinetic energy or heat or some other energy. An electric
charge,too, cannot undergo a fall of potential without a loss of energy, that is to say without a
transformation. Heat, on the other hand, may be transformed with a fall of temperature to a body of
greater capacity though the energy of heat remains the same? 1

Mach adds that the entropy mentioned in the second law is only capable of an increase
in the case of heat, since entropy increases denote the heat's lowered potential to work.

We have for the simplest reversible Camot's process, whatever the form of energy -WtNt +
W-;/V2 = O...While the energy value of a heavy mass, of an electric charge, and so on, sinks as the
height of level demands in proportion to this height, this is not the case with heat. Indeed to take
an extreme example, the level may sink on mere conduction and without variation of the energy
value so that W1=W2. Since V1>V2, the entropy increases in this process. Thus although an
analogue of entropy can be set up for every kind of energy, this quantity is only capable of
increase in the case ofheat.22

Mach had a surprising reaction to the special properties of heat. It is the customary
view that heat loses its potential to do work by increasing entropy of heat without losing
energy, which remains conserved. Mach, however, said that since heat energy loses its
potential to do work, in effect we must consider it irrevocably lost energy whenever the
entropy increases:

The measure of energy rests on the fact that we can make any physical reaction vanish and replace
it by mechanical work, and conversely. But is there any meaning in attributing a work value to a
quantity of heat, which cannot be transformed into work? Accordingly it seems that the principle
of energy, just like any other substance view, only holds for a limited domain of facts, and over
this limit we are only too prone to stray. I am certain that a doubt as to the unlimited validity of the
principle of energy will seem as surprising at the present time as a doubt of the constancy of the
quantity of heat would have seemed to the followers of Black. But we must reflect that every
dominating theory tends to extend its dominion beyond what is proper... The principle of energy is
a special form of viewing facts, but its domain of application is not unlimited?3

It's a very curious thing to say. After all, the heat doesn't seem to lose energy per se
although it may be scattered into various effects that cannot be rejoined as the original
energy source. Mach's objections have more to do with the usability of an energy source
compared to other sources, or as an issue of energy measurement. Mach sees these
questions as connected. He says very clearly that he does not agree with making the
entropy some function of the order or disorder of a system of particles, and that there is
"no true analog" of the increase of entropy in a purely mechanical system made up of
absolutely elastic atoms.Z4

But Mach's standpoint did not explain other uses of the entropy, for example in
describing any irreversible processes in which heat did not fall in temperature or
potential level, which was one of the crucial points that Max Planck urged against
Mach. Planck saw the expansion of a gas out of a tube as an irreversible process that did
not require any difference in tempemture between the tube and the air outside. The
entropy increase is thought of by Planck as more fundamental even than changes in
tempemture, which is just one of its manifestations. Entropy increase indicated for
Planck a preference of nature for certain fmal states of processes over their initial states.
According to Erwin Hiebert:

Planck stood firm in the view that there are irreversible processes in which there is no change in
temperature. He insisted that the essential meaning of the second law simply rests on the fact that
many different kinds of irreversible processes really do exist: the conduction of heat; the free
expansion of gasses; the freezing of supercooled water; the condensation of superheated steam;
explosive processes; chemical processes... in fact every change of a system moving toward a more
stable condition of equilibrium? 5

Mach remarked that the expansion of a gas could be interpreted as expansion

"under a resistance which is smaller than the force of expansion" and therefore a loss of
energy level to heat. The expanding gas overcomes the resistance and genemtes vis viva
which then passes over to heat, which then falls in level of itsel 26

He also cites Clausius on the overflow of gases: "Gas which overflows into empty
space without a variation of temperature cannot be brought back to its original state
without a decrease of entropy: its entropy has therefore increased on the overflowing. "27
Does he approve of this reasoning? And is this any less a subjective measure of the
entropy increase than disordered motions? Here again the entropy is imagined as a loss
of energy to do work, by invoking the energy it would have taken to restore an initial
state. Hence even diffusion and other irreversible phenomena do not involve a
temperature difference but they can be seen to involve a fall of level all the same if we
think about the work it takes to reset these processes to their original state along with
the environment of the system. Mach reflected in his notebook:

What happens thermodynamically when a gas expands isothemrically? Where is the fall ofheat/28
Where is the cause of change without a change of potential value? The potential value in a cycle. It
doesn't necessarily change through a fall but rather with time, so one is dependent on the other. 29

Despite Mach's talk about accounting for the course of time through a "law of
happenings" or a law by which the potential differences in nature always decrease, as
was his wont, Mach sometimes acknowledges that time is an irreversible variable even
for cyclical processes in which the initial states are restored. Here is a beginning to his
view that all processes are really irreversible.


In phenomenological thermodynamics the canonical view, under the influence of

Planck and Ostwald, became to assume two different perpetuum mobile postulates, the
first for reversible processes and the second governing irreversible processes. Thus, the
first is the classical principle that work and its equivalents cannot be created from
nothing. This principle says nothing about the direction of energetic processes from A
to B to C or from C to B to A. Therefore, energy transfers may run in any order and
there is no reason for a process to proceed in one direction as opposed to another.
Planck wrote:

The second law does not emerge from of the concept of energy, in no way is it exhausted by the
first law, in that one analyzes every natural process into a series of energy changes and now asks
after the direction of each individual change. Granted, one can mention by name the different
types of energy in each individual case which reciprocally replace one another; for indeed the
principle of energy must always be fulfilled. But there always remains a certain rubitrariness, how
the causal conditions of the transformations are expressed, and this aroitrariness cannot be
avoided through any general rule?0

The excluded perpetual motion of the second kind is, according to Planck, a logically
independent principle which forbids the uncompensated production of work from heat,
except in reversible cyclic processes, but not the uncompensated production of heat
from work. Thus, it picks a direction in the energy changes for which the first law only
demands the equality of initial and final states, but no order. Planck complained,
wrongly, that Mach had not understood this: 31

He speaks very often in his book about perpetual motion, but he does not attach any clear cut
physical meaning to this expression. He continually confuses perpetual motion of the first kind
(production of work from nothing) with perpetual motion of the second kind (uncompensated
production of work from heat) ... Mach does not even devote a syllable to the fact that the two basic
principles about the impossibility of perpetual motion are completely different from each other,
that the first is reversible (the impossibility of abolishing work) but the second not
(uncompensated production of heat from work being by no means impossible), that the energy
principle (the first law) rests on the first and that the Carnot-Clausius principle (the second law)
rests on the second, that the second is completely equivalent to the known Clausius principle of
the transfer of heat from a lower to higher temperature, and that this principle has as a
presupposition that without the assumption of a non-reversible process proof of the Carnot-
Clausius principle is not to be managed. 32

Mach wrote back in his 1911 "Leitgedanken" that he agreed with making a fundamental
distinction between reversible and irreversible processes? 3 And Mach even agreed with
(the early) Planck in making irreversible processes the most fundamental. This is not
generally known and it relates to Mach's theory of economy in that the schemata of
repeating events in nature are really economical abstractions serviceable to memory and
to the recognition of the same events in time. Mach went even further than Planck in
regarding even the reversible processes in nature as irreversible at bottom: "It is to be
remarked that even processes which can be reversed all contain an irreversible element
in the velocity, acceleration, and so on-namely time." 34

In the case of the isotherrnic expansion of gases, Mach's reaction to Boltzmann's

work on irreversibility as the outcome of a probabilistic process equivalent to
equilibrium-seeking seems to have been that the equipartitioned velocities and elastic
collisions of molecules are really not reversible as they seem but point at a deeper
irreversibility, which shows statistical mechanics to be too superficial a level of
description for thermodynamics as Mach saw the science. Mach said in the
"Leitgedanken"35 that even if the mechanical view succeeded a la Boltzmann, it would
still have to be replaced by a deeper view in time, so that it often seems that Mach was
simply trying to look past the reality of atoms and toward a new view of microphysical
properties. Remarkably, the gas molecules that were supposedly too small for Mach
were in this subject not small enough, as he said:

I agree entirely with F. Waid when he says "In my opinion the roots of the entropy theorem lie
much deeper and if the molecular hypothesis and entropy theorem be brought into harmony this is
a good piece ofluck for the hypothesis but not for the entropy theorem. "36

Mach said that he believed the deepest processes in nature, beneath molecules and
atoms, were probably thermal or chemical in nature? 7 Mach represented that level of
description in his own chemical manifold and its homogeneous volume element in
which the irreversibility of the natural processes would presumably be the rule and
reversibility would appear as an anomaly, as a result of economical description, or
perhaps a reversible result of two mutually offsetting irreversible processes working in
different directions. The increase in entropy was seen by Mach as the foundation for the
representation of time: "For me Time-Space questions are essentially physical

The riddle of time will be solved, I believe, through the correct conception of the
second law." 38

Mach went even farther than Planck in establishing irreversible processes as basic
to the reversible ones. This seems to have been Mach's fmal view of the second law:
that the time variable in all natural processes conceals an irreversible process or
chemical asymmetry, possibly pointing to the deeper nature of chemical processes as
opposed to more superficial mechanical processes of spatial bodies and motion. Mach
sometimes points to the second law as a kind of principle of sufficient reason for
guaranteeing that anything happens at all: were it the case that all processes and states
were reversible there would be no reason for a process to happen as opposed to not
happening at all.

Thus for Mach the Camot-Clausius theorem is the only point of contact between the
much more fundamental irreversible thermal and chemical processes and processes
analogous to the reversible changes of energy in mechanics, electricity, and other
departments of physics. Irreversibility does not enter into the comparison at all for
Mach and remains unexplained. "Heat energy" is nothing for Mach but a product of the
fact that the concept of mechanical work, which generated such fruitful analogies in the
other sciences, was applied here to the theory of heat. In reality, the flow of heat is a
different natural process that only came to be connected with energy changes by means
of an historical correlation. Mach closed the Warmelehre with what many probably
took at the time as a cynical remark, but which reveals how much Mach's view is guided
by the history of science rather than contemporaneous views:

The following are the principal results of the toregoing investigation. The various energies show in
their behavior a likeness which has its historical grounds in the fact that the heights of level were
measured in units of mechanical work. With respect to the energy of heat, however, this likeness is
owing to a historical chance. Besides this agreement the energy of heat deviates from other
energies in the fact that it can undergo a fall of potential without a decrease of energy, and that the
zero point of the level cannot be chosen arbitrarily. The principle of energy consists in a special
form of viewing facts, but its domain of application is not unlimited. 39

The lesson Mach wished to draw from thermodynamics is that the science had been
pursued from only one side-the interface with mechanical energy in reversible cyclic
transformations-but might be pursued from another side, the hidden side that energetic
conceptions did not reveal, perhaps because of their unsuitability for economical
picturing in thought like reversible processes: "The conceptions "quantity of heat"
"weight of heat" "equivalence value" and "entropy" must be carefully held apart just as
the domains of fact for which they were established." 40

Mach did not attempt an explanation of why heat energy cannot fully be
transformed into mechanical work or an equivalent except in the limiting case of the
Camot-Clausius cycle. According to Planck, this is what the "perpetual motion of the
second kind" really referred to, a full conversion of heat to another form of usable
energy in a dynamical process. I can only see a slight hint in Mach's musings that heat
and its temperature level for work may really represent a more complex, manifold

property than mechanical energy, a manifold property that may be evident in other
irreversible phenomena. Thus, if heat is to mechanical energy what a plane or a volume
is to a line, it may be that the equation of mechanical energy to heat or chemical effect
is dimensionally unsound and an historical accident, except where heat can be
artificially confined to only one of its sides. This would explain why the analogy
between quantities like "weight of heat" and "equivalence value" are carefully to be
held apart in Mach's view from the analogous terms in other fields to which they are
historically correlated, thus denying an analogy Mach previously had been at pains to
reinforce! This would be, admittedly, a rather extraordinary departure from Mach's
earlier views, which were based on the conservation of energy and which were very
influenced by J.R. Mayer's quest for a universal science. As of 1896, it does not appear
that Mach continued to hold such views, or if he did it was only in the context of his
chemical Weltanschauung.


As a matter of textual evidence, there is scant mention of the energetics movement as

such in the Wiirmelehre, and in the section on "Mechanical and Phenomenological
Physics" Mach makes his own critique of statistical mechanics, without mentioning
Helm or Ostwald. He also softens his assertion of analogies between the different
energies, saying that "analogy is not identity." In the text of the Wiirmelehre, Mach
mentions the analogies noticed by Maxwell between work, pressure, and volume and
quantity of heat, temperature, and entropy respectively. But he says that Artur Von
Oettingen had made the fullest study of these relations. 41 He does later give credit to
Popper, Helm, Wronsky, Meyerhoffer, and Ostwald with "ideas allied to mine in form
or matter." 42 But as noted above, Mach was generally concerned in the whole body of
his book to keep the theory of heat separate and to emphasize the disanalogies. As
Robert Deltete (the leading expert on the energetics episode) and Matthias Neuber have
argued,43 even at the time of the 1895 Lubeck Natwforschersammlung, Mach and the
other representatives of energetics were "uneasy allies," with numerous differences in
their philosophical aims.

Mach and Georg Helm were close to each other philosophically in their belief that
it is only the satisfaction of a fixed set of relations among the natural phenomena that
constitutes the conservation of energy. They did not see energy as a kind of Urstoffthat
persisted in nature underneath those changes:

Mach has repeatedly and justifiably warned of the mysticism associated with the word "transform"
that has sometimes tried to make its way into energetics. In the sense of its founder [Robert
Mayer], energetics is a pure system of relations and is not out to place a new absolute [i.e. energy]
in the world. When changes occur, this definite mathematical relationship still subsists between
them--that is the guiding formula of energetics and.... the only guiding formula of all true
knowledge of nature. What goes beyond it is fiction. 44

Ostwald, by constrast, regarded matter as an aggregate of its different energies:

"Matter," he said "is nothing but a complex of energy factors." 45 These energies are
really present in the body extended in space and time for Ostwald; thus, he held
something of a substance view, if not for the bodies themselves then for the energy
stuffs making up the bodies and migrating like restless spirits from one perishable body
to another.

Yes, people may say that matter is what is real, while energy is only a conceptual device. To this I
reply: It's just the reverse! Matter is the conceptual thing [Gedankending], which we have rather
imperfectly constructed for ourselves in an effort to represent what is permanent in the change of
phenomena. [But] as soon as we begin to comprehend that what is actual [dru Wirkliche] that is
what acts [wirkt] on us-is only energy, we have only to examine the relation in which the two
concepts [of matter and energy] stand. The result, is that the predicate of reality can be ascribed
only to energy...

Always we are dealing only with energy; and if we think further about the various kinds of
matter, ... there is nothing else ... Consequently matter is nothing but a spatially contiguous group
of different energies, and everything we assert of it we assert only of those energies.46

In Ostwald's view, the energetic relations give rules for these migrations as the
capacity factors of the energies were like the amounts of material moving47 through
space from one body to another, with the intensity factor controlling something like the
rate of motion. Capacity becomes something of a spatial factor and intensity becomes
something of a temporal factor for energy changes. Ostwald also said that the energy
was really "that which remains unaltered through all change,'"'8 suggesting a substance
view for each energy factor separately but not for the bodies they comprise.

Mach could well have accepted Helm's point that energy was just a fixed relation
among the phenomena. However, as Deltete points out, Mach's willingness to regard
energy as unconserved when lost to heat would have put him at odds with Helm, who
held no such view. In addition, Mach's breaking of the analogy between dynamic flows
of heat and the other energies (except in the case of Camot-Clausius) would have
alienated him from Helm and placed him closer to Max Planck-of all people.

In the case of Ostwald, I think it likely that Mach would have accepted features of
his view of matter as a manifold of energies, and even accepted some of Ostwald's ideas
about movable capacities as part of a comprehensive chemical manifold, but not in
space. The permanence of the underlying energy factors would not be absolute for
Mach, but rather a question of relations between those factors, and in the case of heat, a
loss. He does pay Ostwald a particular compliment when he describes the relations of
his elements in the 1910 essay "Sinnliche Elemente und naturwissenschaftliche

The inner connection of elements in a spatio-temporal place we call matter ... If we pay attention
not to the dependence of elements on the human body, on the sensations, but rather to the
reactions of the elements in general, we may say that the spatio-temporal points of attachment of
the reactions of elements may be called matter. The former expression is physiological or psycho-

physiological, the latter physical. To the extent that one conceives of all reactions as energetic, this
latter expression coincides with that of0stwald. 49

Mach seems to be associating himself with Ostwald while maintaining one difference;
he says that "to the extent" that one conceives the physical relations of the elements as
energetic, Mach's expression for matter will match that of Ostwald. Hence, there may be
relations of elements not described by the energetic principles of capacity and intensity
factors. It remains possible that the category of the elements and their variations is
wider than that of energetic relations, such that energies are cases of elements and
variations, but that the converse is not entailed. This would explain why Mach had
believed he could derive the principle of excluded perpetual motion from the functional
variation ofhis elements as the more general principle of nature.

There was one great difference that does not emerge here, but is clear from Mach's
late writings on space. Ostwald regarded space and time as fundamental to a scientific
worldview and thus represented his energetics as energy flows in space and time, and
his fundamental set of dimensions as (e,s,t). 50 Now, for Mach no theory could be
acceptable that did not reduce the spatial and temporal relations of elements to physical
reactions of elements on one another. Mach saw his chemical view as the means to
derive space and time from "chemical paths," 51 as he put it in the Wiirmelehre. I think it
likely that Mach could accept aspects of Ostwald's chemical view in the abstract, while
he certainly rejected the idea that space and time should be fundamental variables in
natural science.


In additional to their professional achievements, scientists often have a "natural

philosophy" to guide them in their choice of problems and theory construction. To
the degree that this natural philosophy raises general conceptual questions, often
outside the range of science to decide on them empirically, it is hardly possible to
separate philosophical and scientific work. Unlike a metaphysical system, however,
a natural philosophy may be only a working picture of the world standing opposite
the scientist's physics while, like metaphysics, it may not be subject to direct
experimental test. A natural philosophy may only touch physics at a high level of
generality or may represent an ideal point to which the most general physical
principles converge. In Mach's time, the mechanical worldview of particles and
action and a distance, the electromagnetic worldview of fields and action by contact,
and the thermodynamic view of pure energy transformations without regard to
mechanism were all natural philosophies that strove with one another for the

Ernst Mach called his overall natural philosophy the "chemical view"
(chemische Auffassung), although it had little to do with what chemistry means today
as a branch of applied physics. According to Mach, chemistry was more fundamental
and realistic than physics, since it described occurrences that cross-connected the
different departments of mechanics, electricity, and heat and had as its business the
laws of transformation and analogies between these different forms of energy in their
interaction. Mach ended his Mechanics with the remark that "purely mechanical
problems do not exist" except as an idealization:

The production of mutual accelerations in masses is, to all appearances, a purely dynamical
phenomenon. But with these dynamical results are always associated thermal, magnetic,
electrical and chemical phenomena, and the former are always modified in proportion as the
latter are asserted. On the other hand, thermal, magnetic, electrical and chemical conditions
also can produce motions. Purely mechanical phenomena, accordingly, are abstractions, made,
either intentionally or from necessity for facilitating our comprehension of things. The same
thing is true of the other classes of physical phenomena. Every event belongs, in a strict sense,
to all the departments of physics, the latter being separated only by an artifical classification,
which is partly conventional, partly physiological, and partly historical. 1

In his Wiirmelehre chapter "On the Relation of Physical and Chemical Processes"
Mach went into more detail about his chemical view of the natural processes, saying
that "an unprejudiced survey leads us to hold it possible that a chemistry of the future
should include physics-rather than that such a physics should include chemistry."2
Mach introduced his argument for this view by asking what a natural property is,
remarking that it is not at all obvious that iron at rest and iron in motion are the same
body, any more than hot iron and cold iron.

Mach says that in essence, physics ignores these differences because it isolates
different properties of bodies at a time. Chemistry, Mach imagines, would be a
science that describes how the "entire complex" of properties changes, and thus its
concept of matter could simply be expressed through the complex of properties itself
in a set of equations, rather than a quantity of matter like iron, which can be
investigated for its mass, velocity, thermal, magnetic, and electrical properties while
remaining the same iron.

Physical processes are subject to certain equations which present pennanences of combination
or relation of the elements entering into the equation. If a chemical change takes place, those
equations are replaced by entirely new ones. The rules which would determine completely the
transformation from one system of equations to the other would be the complete chemical
laws, and compared with physics, would represent pennanences of a higher order. 3

Mach remarks that masses often seemed to play the role of a carrier of many
physical properties, and indeed in his own energetic analogies, the mass in
mechanics is often proportional to the capacity factor in other energy
transformations. In fact, Mach said in a letter to Friedrich Adler that "in matter I see
nothing but the proportionality of all physical capacities to one another,"4 hence the
temptation to regard all energies as properties carried by tiny subsisting masses.
Mach more or less admitted that the principle of the conservation of the masses of
the reagents in chemistry, through changes in their other properties, argued in favor
of interpreting chemistry as applied physics with masses as the carriers of each
physical property independently. Mach also admitted that any other theory must
accomplish at least as much as the atomic theory in this line. 5

As a different theory, what Mach had in mind was a "chemical" potential level,
similar to his other level values for heat, velocity, electricity, and magnetism. When
compounds like NaCl are formed or decomposed into simpler elements, Mach
imagined that a general equilibration of level occurred that could have effects on all
of the different properties of the body, such as its conductivity, its mass (when
subtracted from that of the other reagents), its thermal capacity, and so on.

In physical processes, masses go over from one velocity level, thermal level, or electrical level,
to another; or, what amounts to the same thing, since at least two masses must take part in
reaction, level values travel from one mass to another. Cold iron and hot iron equalize their
temperatures; but cold iron and hot copper also behave in the same way. Here only one
property is involved in the equalization. May we not imagine that in the domain of chemistry
as well nothing further takes place than that masses--homogeneous throughout--of one
chemical level go over to another chemicallevel? 6

Mach remarks that unlike other level values, which are continuous manifolds,
the chemical potential level exhibits discrete grades. The reason for discrete steps in
chemical potential level is that chemical elements must combine in exact
proportions, leaving the rest of the components unchanged.

The fact [of discrete grades) obtrudes itself, so it seems to me, as such a striking difference
from what takes place in physical processes that to close our eyes to it and to deceive ourselves
by ingenious assumptions about its significance can hardly commend itself to us.

In addition, unlike the simple manifolds of the other potential levels, Mach held that
the chemical potential level was to be regarded as a "manifold level," since it
embraced a manifold of physical properties:

Such a [chemical] potential can hardly be conceived as a simple linear manifold like the
familiar physical potentials. The fact that in chemical processes (changes of potential) a whole
complex of properties varies makes this view difficult. The periodic properties ofMendelejeffs
series also point to a multiple manifold: these properties cannot be represented in a straight
line ....

We need not shrink from the supposition of a multiple manifold of level values. The familiar
physical level values are, it is true, level values of work; and, as such, simple manifolds. But
though the square of velocity represents only a simple manifold, yet velocity, acceleration and
so on, since they are directed magnitudes, exhibit a three-fold manifoldness. These quantities
are level values although not those which are alone taken into consideration by the modern
principle of energy. But the principle of energy certainly cannot settle all physical questions.
Electricity and magnetism stand to one another in much the same relation as real and imaginary
quantities. If in this sense level values of both domains are regarded as belonging together they
form a system of a six-fold manifoldness. Much the same might be the case with the chemical
potential; but the fixed points would be wanting, to begin with, for determining the dimension
number of the manifold. 7

Now, certainly Mach's chemical view seems excessively phenomenalistic and

reminds one of nothing so much as a nineteenth-century investigator who has no
access to the molecular and atomic (let alone nuclear) properties of a compound, but
who has instead a set of measurable properties of a galvanometer, a scale, and a
thermometer and seeks to relate these observational properties in such a way that he
can make sense of seemingly "qualitative" changes in the bodies he is investigating.
The remarkable success of thermochemistry in applying macroscopic parameters of
state (entropy, enthalpy, free energy) to chemical processes, by Gibbs and
Helmholtz, surely strengthened Mach's view that this phenomenological manner of
investigating chemical reactions would yield further results. 8

Another factor encouraging Mach in the direction of a chemical manifold was,

as he says, the promise of finally combining chemical processes with physiological
ones in one overall view that need not be changed in moving from sense physiology
to chemical physics. Mach includes these observations continuing from the quote

Reference has been made elsewhere to the fact that sensations are the true elements of our
world picture. Now we cannot doubt the extremely close connection of sensations with
chemical processes. If we have six fundamental color sensations, we assume that the albumen
of our bodies may be transformed by optical excitation in a six-fold manner. All sensations, as
also space sensations, would admit of an analogous view .. .If bodies so different as sugar,
permanganate of potassium and arsenic taste sweet, it naturally does not imply homogeneity of
these bodies but a similar transformation of the albumen brought into contact with them. There
will be as many sensations of taste as there are kinds of transformation of albumen by
immediate chemical action. For knowledge of the latter, rather than for characterizing the

compounds examined by taste, it would be worthwhile to work out a system of taste sensations
similar to that of color sensations. 9

Mach's chemical manifold thus brought together his conceptions of

physiological and physical manifolds and properties more closely than ever before.
He proposed a combined physical-physiological space, in which properties
immediately evident to us as our sensations and their qualitative dimensions (hue,
pitch, taste) would also have a chemical meaning as the manifold chemical
transformations of substances within the human body. Mach allowed, for example,
that physiological space sensations and anisotropy would have a correlate in physical
space, once it had been analyzed according to his view that each spatial direction
represented an essentially different "chemical direction":

Physical space also has three essentially different directions, which are most clearly manifested
in a triclinal medium, in the behavior of an electromagnetic element. The same physical
properties appear in our own body, which is the reason why our bodies can be used as reagents
in physical problems. If we had an exact physiological knowledge of an element of our bodies
we should thereby have laid, in all essentials, the foundation of our understanding of the
physical universe. 10

This is a difficult view to understand, particularly because of the misleading

impression that chemical processes would introduce a set of absolute directions in
space, as if, for example, all bodies would have to accelerate in the same direction,
which is absurd. Thus, we need to look at this more carefully. An example of what
Mach means by using the body as a reagent is the "right-hand rule" to determine the
lines of a magnetic field given the direction of the current by looking at one's right
hand. This is a deceptively trivial task for Mach since all one needs to be able to do
is tell right from left (not self-evidently easy for all of us!) and that ability, Mach
thought, depended on some still more basic physical difference in those directions
that is not evident. (Some students, like myself, may recall learning the laws of
electromagnetism with a sudden surprise that nature understands what a right angle
is in setting the changing electric and magnetic fields at right angles to each other,
like the real and the imaginary numbers.) In addition to the subtle anisotropy of the
human body, Mach was also fascinated by the unidirectional flow of memories and
thoughts 11 as another example of some basic chemical asymmetry manifested in
physiological processes.

Throughout most of his career, Mach could be described as an empiricist about

physical space and a nativist about physiological or psychological manifolds of
sensations. As an empiricist, Mach was skeptical about attributing any physical
efficacy to space itself, in his critique of Newton, or in his statements that space and
time are only concepts of measurement of certain physical events by other processes
assumed as a standard. But Mach regarded physiological space as a genuine
sensation, or as a manifold of sense elements such as a manifold of tones, of colors,
of orientations, of the musculature of the body or of the vestibular apparatus of the
inner ear. The framework of space sensations, as we saw in an earlier chapter, was
for Mach an innate register in the organism's body allowing for orientation against a
seemingly fixed background of directions.

Mach may have doubted absolute space for physics, but not for physiology.
What we have in Mach's late career is apparently a serious turn to a physical-
physiological view of space, connected with his idea of a chemical manifold.
Whether this was really a new view or a view he had held for many years I am
unable to decide, especially since Mach's utterances on chemistry and its relation to
spatial concepts such as locomotion and position really date back to his
"Raumvorstellungen" paper in 1866:

Many things that are still mysterious in physics, e.g., the Weber electrodynamic law, arranges
themselves more clearly under the influence of this point of view. When up until now one
thought of physical phenomena, with the inclusion of the chemical, preferably as motion,
position next to one another and so on, one might now, not without advantage, try to conceive
of the phenomena of motion chemically. Light is supposed to be an oscillation. We know that
chemical action can be induced by light. Why should we not think of light as rapidly changing
proto-chemical separation and connection? 12

Another similarity between the 1860s and the later writings is the reemergence
of Herbart's views of space in Mach's work, in particular in connection with the
chemical manifold:

Our intuitions of space and time form the most important foundations of our sensory view of
the world and as such cannot be eliminated. However this does not prevent us from trying to
reduce the manifold of qualities of place-sensations to a physiological-chemical manifold. We
might think of a system of mixtures in all proportions of a number of chemical qualities
(processes). If such an attempt were one day to succeed, it would lead also to the question
whether we might not give a physical sense to the speculations that Herbart, following Leibniz,
conducted as regards the constuction of intelligible space, so that we might reduce physical
space to concepts of quality and magnitude. There is of course much to be objected to in
Herbart's metaphysics. His tracking down of contradictions that are in part artificially contrived
and his eleatic tendencies are none too attractive, but he will hardly have produced nothing but
errors. 13

Recall that for Herbart the various qualities of the physical world were
something like pressures or energies and had both a direction and a magnitude. The
idea that space has "three different chemical directions" seems to indicate that each
direction in space is a different process that "goes" in a different direction,
independent of the others, or perhaps that each direction is really a different energy,
a different Herbartian quality, and chemical processes in space combine all six
possible directions, left right, up down, back front. If no isolated departments of
physics really exist for Mach then each and every process is a chemical one and
occurs in the full set of directions in space. This is the difficult aspect of his view:
that every process, even if it seems only a mechanical or electromagnetic one,
already involves the other processes, too. The only evidence of their involvement,
however, is the three-dimensionality of the space in which these events occur. In the
Warmelehre, Mach reiterated the analogy between the possible types of chemical
process and the concept of spatial direction:

Just as at present in stereochemistry we endeavor to explain chemical reactions by spatial

relations, it is quite possible that we may some day attain to the understanding of space, its
dimension number and so on by chemical paths. 14

And still more directly, Mach wrote in a notebook fragment of 1886 that "Threefold
chemical changes are space." 15

Besides the chance to develop Herbart's manifold further, the chemical view
with its manifold potential level also solved a difficulty in Mach's comparative
physics. This is the fact that Mach had been forced into interpreting all happenings
as essentially due to scalar potential sources and potential functions or levels,
whereas it was clear through electromagnetism and even mechanics in space that a
minimum of three dimensions, or six fundamental directions, were needed to
describe natural happenings. Mach wrote in 1870 that velocity was the potential
level for bodies exchanging energy of impact, but in the Conservation of Energy he
says he realized that this would not give a scalar value, like temperature, which he
could compare with other levels. 16 Thus, to make velocity a scalar, too, Mach
divided kinetic potential into the product of the mass as "the potential" and half the
squared velocity as the "potential function." Mach's chemical view appears to be a
different solution to this problem, allowing for manifold potential levels after all,
with the old scalar levels as just the separate energy components of a process in any
of the six fundamental directions.

To the objection that surely every energy need not be involved in every
problem, Mach could reply that in reality there are no physically isolated problems
of pure mechanics or pure electricity. These are all abstractions based on numerous
experiences, schematized and thought of in isolation from others. Hence, on this
view, every physical happening involves all of these factors, if only as the inert
dimensionality of space. Mach gave some thought to the idea of spatial dimensions
as a superposition of energy levels as its dimensions in the following remark of

Velocity is a three-fold level value. Collision just like temperature by conduction. Three-fold
manifold. The squared velocity some other characteristic ... How does it happen that.. .always
the same mass-relationships are the same in equivalents for every combination, so must
universal rules subsist for the potential levels, too. Think this through. All capacities go
proportional to the capacities for acceleration, all other things being equal. 17

This does indicate that Mach was thinking of a manifold in which the potential levels
and the capacities would be entered. Mach also indicates in Knowledge and Error
that he thought it possible to regard "position" in a physical manifold as a question of
possessing certain properties of state rather than a bare spatial position to which
properties affix: "Physics would be justified in considering an extended material
continuum, to each point of which a temperature, a magnetic, electric and
gravitational potential were ascribed as a portion or section of a multiple manifold."
18 Here it appears as if one uses the natural potentials, the directions, to locate a point

in space the way that a manifold mathematician locates a point by approaching it

from each separate direction of the manifold.


In his 1897 lecture "General Questions of Natural Science," Mach picked up the
question of space again, comparing the philosophical merits of the mechanical
atomic theory, the "aether theory" of W.K. Clifford and his own chemische

In this lecture Mach spoke frankly about the strengths of the atomic theory,
which he obviously appreciated. As Mach wrote in a letter to Josef Popper, 19 he saw
the strongest argument for atoms as one that he attributed to Newton, namely that the
exact scale-independent laws of nature, if they hold, must hold all the way down to
the tiniest particles. If the natural processes were not uniform in the smallest parts,
the smallest bodies exchanging heat, the smallest massy particles attracting one
another by gravitation and other fundamental forces, the laws themselves could not
be uniform. Some small parts would act on one another unequally and set up
different conditions and rates of equalization that would never settle. But this
unequal settling is never observed. This putative Newtonian argument trades on what
is often termed the scale invariance of a natural law. An informal rule of physics has
it that a natural law is not dependent on the size of the units of measure of its terms.
The law of gravitation, the conservation of energy, hold at whatever scale one is
considering. It follows from this, although the consequence is rarely drawn in
discussions of scale invariance, that one may also consider as small a unit as one
wishes, and these arbitrarily small units may then be chosen as the "atoms" of the
theory. As Mach puts it, the properties of mass must be taken to remain in the
smallest parts. He calls this an assumption [Voraussetzung] of the atomic theory.

That everywhere there should be magnitudes of the same order is to be expected from the
regularity of nature?
What is the assumption of the atomic theory?
That some properties remain in the smallest material parts, others not.
Nothing in principle to say against this. 20

But as we will see, Mach saw a way to use this argument in favor of his own
manifold "volume elements," so he felt it was not at all decisive for atoms. Indeed,
the conservation of energy being an exact law of nature, it is natural to apply the
same reasoning, saying that it too must hold at the smallest scale, and thus what
takes place there is energy transformation in the smallest parts and not energy as a
second-order property of a mechanical system, which may vanish at the last level of
description. To my knowledge Mach never expressed himself so, but it would seem
to be a natural way of thinking for him. However, Mach also gave the atomic devil
his due. In his lecture, Mach remarked on the "powerful" support the atomic theory
had received from thermodynamics, and in the combination by mass of reagents in

chemistry: "Wonderful proportions. Unexplained. Hence the outward appearance of

the atomic theory." 21

Strangely, Mach was very cognizant of the increasing power of the atomic
theory while feeling that these successes had no bearing on the metaphysical issues
of whether atoms existed or not. Mach still saw the issue in terms of natural
philosophy, where one had a choice what to believe, not science where the evidence
would soon be too strong to deny.

Next, Mach considered the aether theory ofW.K. Clifford. Now Clifford, whom
we have met previously for his neutral monist views, was the translator and a
defender of Bernhard Riemann's geometry in England. According to Clifford, it is
space itself that is the ultimate reality. Matter is to be interpreted as the curvature of
space, or, figuratively, as little hillocks in space. While Mach supported Riemann
and was obviously close to Clifford on many philosophical issues, he was not a
supporter of the independent existence of space as the ultimate substratum of events.
Mach remarks that: "The continuous aether gives a very obscure Shearing-
elasticity ... But what is the material of the aether? Doesn't it raise the same problem
over again?" 22 Mach added the criticism that he felt the idea of an aether, in which
the different forces are currents, was merely a mental picture, an analogy of sense
perception: "Magnetic currents. Gravitational currents. Just analogies. Pictures. Just
as much as matter of heat. Electricity as a fluid, in which no one seriously believes."

Mach blew hot and cold on aether theories. He could be admiring, for example,
of a paper he cited in a late edition of the Mechanics, Paul Gerber's "Ueber die
riiumliche und zeitliche Ausbreitung der Gravitation, "24 in which the advance of the
perihelion of mercury was explained by giving gravitation the velocity of light.
Mach said "this would speak for the Aether as the medium of gravity. "25 But he
generally opposed the ether, claiming that he could develop the needed concepts of
electricity on the basis of the potential and the potential function, which he gave,
saying that "no conception of fluid or medium" was contained in these fundamental
notions or those derived from them. 26 Generally, Mach's view seems to have been
that the extended fluid medium was a psychological image for visualizing events.

By contrast, Mach goes on to say in the "General Questions" that his own
chemical view does not try to assume a sensory space and then try to fit all the things
of thought into it: "We operate as if space and time, which are valid for our sensory
perception, also dominated and restricted all things of thought (Atom, etc.)." 27 What
Mach has in mind instead can be seen both in the "General Questions" and in the
chapter "Another View of Riemann's Manifold" in his Knowledge and Error, which
was discussed in a previous chapter on sensory manifolds. Mach acknowledges
Riemann's idea that three-dimensional flat space is only one of a number of possible
manifolds, all of which are possible for thought. However, Mach says that space and
the physical elements that make it up do share a correspondence, because the
dimension number of space has to represent the number of chemical elements, Stoffe,
a common chemical term. It appears that Mach was considering representing

chemical changes by a manifold of n dimensions, where each of the n was a

chemical property, or element, and the various laws of combination among elements
represented something of an invariant in this manifold:

Consider a series of elements (Stoffe)

a,b,c,d,e .. n
Arbitrarily many.
The complete continuous system of mixtures in all possible relations would be an n-fold
Can be treated analytically.
We cannot visualize all of the dimensions at once, but one after another.
The different points could have any kind of relations to one another.
xz+l+zz=r a2+b2+c2+dz+e2+ ... +i
That would be analogous to space
Another function fl:a,b,c ... )=const. Would be a completely different property.
There can be different ones. 28

Mach also discusses "mixtures" [Mischungen] of directions as if they were

analogous to mixtures of the basic chemical Stoffe. In the corresponding passage in
Knowledge and Error Mach calls the ingredients of the manifold "substances" or
"elements." It appears that Mach intended a particular physical manifold to represent
the result of analytical chemistry, as he says in a notebook entry of 1884:

[Eingeschachtelte] Manifold
Potential differentiation
As in analytical chemistry
Results. 29

One result Mach apparently wished to derive from his chemical manifold was
the problem of the constancy of mass values through various chemical reactions,
which gave so much support to the atomic theory. Mach seems to have wanted mass
to emerge as a "conserved property" or even perhaps a function, which remained
constant through the various mixtures of elements in chemical combinations.
Perhaps he wished for other properties (of which he said there could be several) like
quantity of heat or electricity to emerge from chemical reactions in the same way,
thus indeed following through on his claim to make laws of chemistry more basic
than physics, which would then simply deal with ideal types of properties remaining
constant over a large number of underlying chemical processes. In the Conservation
of Energy, Mach had said something similar about determining the heats of
combination of the chemical elements:

It is clear how we can study the nature of chemical combinations without giving ourselves up
to the conception mentioned, and how, indeed, people have now begun to study them. The heat
of combustion generated by a combination gives us a clearer idea of the stability and manner of
combination than any pictorial representation. If, then, it were possible, in any molecule
composed ofn parts to determine that n(n-1)/1.2 heats of combination of every two parts, the
nature of the combination would be characterized thereby. According to this view, we would
have to determine n(n-1 )/1.2 heats of combination, whereas if the molecules were thought
spatially, 3n-6 heats of combination suffice.

Perhaps, too, a more rational manner of writing chemical combinations can be founded on
this. We would write the components in a circle, draw a line from each to each and write on the
latter the respective heat of combination. 30

Of course, as opposed to the point made here, Mach's point in his later writings
is that chemical combinations do exhibit the same number of dimensions as physical
space because chemical paths are ultimately the basis for spatial ones. I am unable to
decide whether the chemical manifold (or the analytical chemistry) mentioned in the
"General Questions" and the physical continuum of different energies Mach
mentioned elsewhere ultimately represented the self-same conception, although it
would be convenient if they did.

The overall superstructure of the general manifold Mach speaks of comes from
the geometry of Riemann and Helmholtz, both of whom Mach mentions in the
"General Questions," but alongside them there is also prominent mention of
Herbart's "intelligible space," which may be even more important philosophically for
Mach's thinking about deriving space from spaceless "energies" or qualities:
"Herbart intelligibler Raum. Riemann. Helmholtz." 31

Mach places the manifold of colors next to this note as an illustration. It is

worthwhile noting that Mach could also have regarded the basic energy types as
analogous to "qualities" because of Johannes Muller's "specific energies" doctrine, in
which each sensation represented a different energy of the nerves. Of course, Mach's
elements were qualities themselves, and since at least some of the elements and their
variations were used to represent energy changes in the Conservation of Energy and
elsewhere in Mach's writings, the comparison is more than justified for him. Mach
may also have regarded the energies as fundamental directions as Herbart worked
with directed pressures.

It helps to compare Mach's views to other writers with similar ideas. The first to
propose a kind of manifold of energy-flow as the fundamental physical space was
probably Bernhard Riemann, but his ideas on natural philosophy were not well
known, nor is the connection between his views on geometry and physics well
understood. Wilhelm Wien, however, proposed a very straightforward rendering of
all physics as the flow of energy from place to place in time, as hydrodynamics
studies the flow of quantities of fluid at various rates. Heat and electrical energy had
been compared with fluids, and the extension of the thought that energy is a fluid
with a quantity flowing at a given rate across a boundary was only natural. In 1890,
Wien discussed the idea of treating energy flow in a space and time aether as the
fundamental process to be described in physics across its different departments. In
addition to mechanical energy Wien discussed radiation, electrodynamics, and
thermoelectricity, and summed up: "There is nothing in these reflections which
might prevent us from conceiving of the motion of energy as an appropriate remedy
(Erleichterung) for the treatment of natural phenomena." 32

Gustav Jaumann, who had been one of Mach's assistants in Prague, put forward
what he thought was an extension of Mach's "chemical theory," in which the
hydrodynamic analogy is fully developed. Mach did not accept Jaumann's extension
of his ideas of comparative physics, or the chemical view of natural processes, as he
wrote in a footnote:

The considerations under discussion [Mach's own view] have only the starting point of the
principle of comparison in common with Jaumann's "chemical theory" (Sitzungsberichte der
kaiserlichen Akademie Vol. CI, Ila May, 1892). I cannot regard Jaumann's attempt as a
solution of the questions referred to, though I am of the opinion that the points debated by him
are well worthy of discussion. 33

However, as Deltete and Neuber have shown, 34 Jaumann was proposing a theory
of energy flow that traded on the same analogies to which Mach had called attention.
In fact, Jaumann claimed to be carrying on Mach's "comparative physics" and
"chemical view" in doing so. Deltete and Neuber say that the most natural
reconstruction of Jaumann's view has him formulating equations of continuity for the
different types of energy flow, describing energy as a kind of stuff that flows across
a boundary and which increases on one side of the boundary exactly as much as it
decreases on the other, hence keeping the total amount constant. 35 For
electromagnetic processes, the Poynting vector describes the flow out of a point on a
surface in units of energy flux density, and its divergence gives the rate of that
energy flow. 36

If Mach knew about these approaches, he may have rejected them on

philosophical grounds as being too similar to an aether theory that made space
fundamental instead of deriving spatial properties from the type and behavior of the
energies, as his views would have demanded. It may have seemed to Mach that
Wien, Jaumann, and others were simply visualizing energy changes in space rather
than using the properties of energy to derive spatial concepts themselves through
underlying motions of energies. Riemann would have been a prominent exception,
but there is no sign Mach knew of his natural philosophy beyond the famous 1854
Probevorlesung on geometry. Riemann's Nach/ass, with his views on natural
philosophy and the unification of science under the concept of space that he had put
forward, was not edited and published until much later.

The relation between the dimension number of space and the independent types
of energy seems to have been fundamental for Mach. In these fragments he
expresses the fact that the dimension number of space is so determined:

For three dimensions. Four elements necessary. Four qualities. Chemical. When the aether
manages everything in a moment of action three specific directions. Measure everything with
the wave-lengths of light. Quality with magnitudes. 37

Space is a three-fold quality characteristic just like any other physical quality. Perhaps
chemical, perhaps electrical... the spatially closer relationship is also temporally closer.
Temporal propagation of the equalization to the mean."

We may infer from these remarks that Mach did envision a manifold in which
the different energy potentials would enter as the different possible directions for the
natural processes to take. This could be understood in the sense that different
potentials can act like independent directions, gravity and electromagnetism
normally being unable to affect each other as if they sat at right angles.

The proposal to treat each spatial direction as a different energy had, in fact,
been tried by Max Planck in 1897 in his Erhaltung der Energie. 39 Planck derived the
equations of motion of a point from the fact that energy is conserved in each
direction separately, and thus obtained the equations of motion from the energy
principle by splitting it into three separate scalar equations. Planck was criticized by
Boltzmann for this artificial assumption, because in general the motion of a point can
be described by arbitrarily stipulating the coordinate system, belying any such
underlying physical meaning to the directions. Mach never made this proposal for
mechanics, but rather for his generalized manifold, for which he denied the reality of
purely mechanical problems, anyway. In the mechanics of bodies and motions, as we
shall see below, Mach seems to have regarded the geometric independence of
directions, as well as their dependence, as expressing some underlying physical
dependence on which the facts of geometry supervened.

Mach does seem to have had in mind a physical manifold in which each point
would possess the potential values of a function U, which would then equalize
differences in potential level value with a neighboring point according to Mach's law
of differences. The processes of equalization of level values seem to have played a
significant role in the transfer of energy by moving waves in the manifold. Here light
was apparently Mach's standard energy carrier by radiation.

Analogy heat current. Temperature current. Potential function current. 40

Light as chemical oscillation... Light. The difference will diminish and approach a constant
limiting value. Difference-drop tubes. When the velocity falls the difference diminishes. When
it rises, the difference enlarges. Otherwise it remains constant.

Explain space by way of time. The lengths of lines by way of lengths of time. 41

First of all eliminate space. To get time pure. The smallness of the difference is the measure of
time. The diminishment of difference is a determinate goal. Middle ... Space reduced to the
temporal approach to the mean.42

Despite the darkness of these remarks, the connection of transfers of energy (the
equalizations of level) from place to place with the temporal tracing of spatial
regions is unmistakable. It is clear that Mach saw an identity of space and time-
tracing processes with the process of equalization of energy level. This impression is
confirmed in Mach's Knowledge and Error, where he describes replacing time as a
variable by the equalization of temperature between bodies in thermal contact. In the
case where intermediate bodies stand in the way of equalization of two bodies
situated at their extremes, Mach says the equalization is a "spatial" one. He also
describes using light waves as an appropriate measure of space and time.

We are clearly meant to infer the generalization that the idea of equalization of level
should be extended as a notion of temporal and spatial measurement in all areas of
physics. The idea of a "potential function current" or a current that propagates over
differences in potential level in general seems to be the appropriate generalization.
As such, Mach's oscillating equalizations of level are not all so different from
Herbart's space-tracing waves in his quality manifold taking advantage of differential

The differences of level value of the various potentials correspond to what the
energeticists called the intensity factor, or what Mach called the potential function.
The "capacity factor" of Ostwald also played a role in Mach's physical manifold.
Mach of course compared the quality of heat to mass as certain fixed factors in the
flux of changing qualities. His isolation of a mass factor from the play of forces is an
excellent example of Mach's procedure: "Because there are fixed capacities, the
product ct represents something constant in the W echselwirkung. As does m<p." 43

Mach adds to this an idea that seems to come from the metric geometry of
Riemann and Helmholtz. Helmholtz showed that the postulates of Euclidian
geometry could be replaced by a more general postulate of free mobility, or of the
conservation of volume through changes in position in spaces of constant curvature
(whether positive, negative, or zero). A volume remains the same when rotated in
place and moved between one coordinate position and another. Mach made much of
this idea in his treatment of the new geometry in a sequence of chapters in
Knowledge and Error, especially by saying that, as the Pythagorean theorem for a
plane triangle resulted from the conservation of areas swept out by the three sides, in
the three dimensional case the same held for the conservation of volume for rigid

If a triangle a,b,c be slid along a short distance in its own plane, it is assumed that the space
which it leaves behind is compensated for by the new space on which it enters. That is to say
the area swept out by two of the sides during the displacement is equal to the area swept out by
the third side. The basis of this conception is the assumption of the conservation of the area of
the triangle. If we consider a surface as a body of very minute but unvarying thickness of third
dimension (which for that reason is uninfluential in the present connection) we shall again have
the conservation of volume ofbodies as our fundamental assumption. 44

Mach expressed the hope that similar invariants would be discoverable in the
physical or physiological manifolds, such as invariants of tone or color. But he
seems to have found that what Ostwald called the "capacities" were suitably constant
factors to identify "the same" capacity after a passage from higher to lower potential
levels analogous to the preservation of a volume after a change of its position
coordinates. Mach says this only tentatively and in few places, never in published
Space-conservation plays the same role in geometry as the continuity and conservation of mass
in physics. 45

Transformations. N Dimensions. Nebengleichung.

Homogeneous volumes= Capacity
dx2+df+c dxdydz. 46

Mach's suggestion that capacity play the role of spatial constancy in the physical
manifold seems to be a direct analog of Helmholtz's principle of the constancy of
volume on displacement.

Thus, after all, the temptation is to interpret Mach as proposing a version of the
"movable stuffs" theory of Wien and Jaumann, if we regard the more fixed capacity
factors as the "stuffs" moving over differences of potential level according to Mach's
law of the equalization to the mean of the level differences. However, Mach never
spoke of space in those terms. The most he would say is that space should be
eliminated in favor of his elements. Yet there are various sorts of functional
dependence that hold among Mach's elements, which lend structure to his overall
view. We saw, for example, in the case of the "Raumvorstellungen" program that
Mach had thought of specializing from the reactions of the quality elements two
sorts of agency accounting for the greatness of the induced pressures in bodies: those
due to a capacity or greatness of an agent, and those due to the distance or the
intensity factor. It would be tempting to interpret these as the physical meanings of
space and time respectively.

Mach may have had something similar in mind for each of the various potentials
for energy and its potential function as part of a general manifold. The basis of the
construction would still be the elements, and as such nothing would be inherently
spatial about the construction at bottom. However, certain patterns could still be
isolated for their pragmatic interest. From among the pressures exerted by the
qualities on one another, one could isolate those pressures that can be attributed to a
constant potential (W) at varying potential levels (V',V",V"'), and one could isolate
other pressures due to varying potentials (W',W",W"') at a constant equipotential
level value (V).

As for the question of different "chemical directions," different potential sources

could be set apart if they are found to be capable of acting independently of one
another, as if the other is not there. Each energy type (WN) could thus be
represented as if situated in a different direction from the others, where
independence of direction is derived from this physical independence.

The above comparisons of capacity with volume in geometry would hold if the
capacities (W',W",W"') could be transformed in any of the various directions and
retain the same value in the exchange. This would be true if there were something
like a generalized multidimensional capacity of matter that transformed reliably
between energy forms like an object rotated through any of its coordinate axes in
geometry. Each axis here represents an independent potential that matter might have,
and all of the axes together constitute a manifold chemical potential level. This
conception of permanent matter, as signifying the "proportionality of all capacities to

one another" through transformations, was endorsed by Mach. What we have as the
overall result of these comparisons is Mach's general volume element standing in for
the concept of mass and his multidimensional chemical potential level as a manifold
of the other levels. Now indeed, the other transformations present themselves as
special cases of the chemical transformation of mass-equivalents from one chemical
potential level to another. In this way, too, generalized chemical notions would thus
serve as Machian replacements for the atomic theory of matter.

The tension between the elements and their ordering into this general manifold
of space, time, and matter is a general problem in Mach's philosophy of nature. The
divide falls between his heraclitean view that the elements are transitory unique
events, arising and vanishing and possessing always an individual existence, and his
view that space, time, and matter, however unreal they may be on a fundamental
level, represent for Mach economical permanencies that must be acknowledged as a
task of science. Space is real and properties of physiological space represent real
chemical properties for Mach. The three dimensions represent three separate
chemical directions. Yet space is unreal too, and represents an economical necessity
that vanishes when one considers that true isotropy and repetition of events, the
revisiting of places visited before, never truly occurs in nature but is always
something new. Indeed, if irreversibility is a fundamental physical fact, as Mach
thought it was, even chemical space must be something of an abstraction. Mach said
that he considered the real facts of nature to be the existence of "differences" or
inequalities as he says in one place in his notebook, he may not even have regarded
the potentials, the masses, the electrical charges, as fundamental, rather only the
differences set up and then attributed to the presence of these potential sources: "The
real world is not the potentials but the differences." 47

Mach's elements are the differences of state in the world and, by a careful
tracking of their effects on one another, the deteminations of the rates and
magnitudes of those effects, Mach thought one could deduce the existence of
independent potential sources and relations of intensity from this raw data by finding
orderings in it.

To borrow a set of terms from another philosopher, I think it likely that Mach
was an empirical idealist about space, when it is considered a property of empirical
elements and events, and a transcendental realist about chemical directions and other
properties of physico-physiological space that have genuine physical meaning, as yet
unknown to us, residing in the smallest parts of matter. I do not think such a
characterization misrepresents Mach's views as they must be taken.


Mach's late article "A Consideration of Time and Space" (1910) 48 presented a new
argument for considering the directions of space as having an underlying physical

meaning. Between 1905 and 1910, Mach had discovered an article by the
mathematician Kurt Geissler which is the basis for the consideration.

Mach discusses Kant's argument for the incongruent counterparts and takes
issue with him on what the example shows. In particular, Mach disputes Kant's
statement that how a right hand cannot be brought into congruence with a left hand
by a rigid motion "cannot be brought into clear concepts." Mach says that in fact the
two hands are not really counterparts at all but physically different objects. We
confuse a left hand and a right hand, or a left ear and a right ear, because the
symmetry of our own bodies and sense organs encourages us to see them as similar
and not as two physically different hands. 49 Mach also points out that it is possible to
bring two mirror hands into congruence through rigid motions, either by turning the
hand through a fourth dimension or, as Mobius showed, by remaining in three
dimensions and turning the surface inside out. A right-hand glove can be turned into
a left by turning it inside out, a topological transformation that respects all of the
local interior relations between points that Kant referred to. 50 (However, turning the
glove inside out does actually exchange the directions of all points reckoned from
infinity inward toward the center of the object.)

As Mach goes on to say, however, a rigid dependence between the directions of

space is more than a geometric fact: it is a physical fact. He seems to be appealing to
his chemical volume elements, saying, for example, that the three dimensions of
space are due to the fact that the smallest physical object is a volume element with
three essential chemical dimensions:

The smallest physical object is three dimensional, a volume element, a body. Surfaces, lines,
points are only mathematical fictions. The best argument that has been brought forward so far
against the arbitrary lessening or increasing of the number of dimensions appears to be that the
three dimensions are not independent of one another. (Kurt Geissler has put this forward.) 51

Mach suggests we take the example of our own bodies as a test case. We exist in
three dimensions, up and down, back and front, left and right. He says that if we
rotate on our vertical axis we can exchange our left and right with respect to the
environment. 52 But in order to do so we must also exchange our front and back with
respect to the environment (see the figure below). Likewise if we rotate about an axis
through our navels, exchanging right and left but keeping front and back constant,
we exchange up and down with respect to our environments and end up upside down
with our heads toward the ground. (I would just point out that the Mobius
transformation turning the glove inside-out is also an exchange of directions of
inward and outward (towards the infmitely large), and it too results in an exchange
of right and left, whether this indicates that "toward the inside" or "toward the
outside" is also a fundamental direction to reckon with could be dealt with

That this dependence of directions is a physical fact about the construction of rigid
bodies in the small appears to be Mach's own conclusion, as if the volume elements
he spoke of were small orientable objects that snapped together into a manifold. For
example, Mach goes on to say that the dependence of directions is not just due to the
fact that there are bodies in space, but rather that the structure of bodies in tum
depend upon the behavior of the various forces of nature situated along the different
dimensions, as if these dimensions are intrinsic to their being the energies they are.
They are the real reason for this behavior on the part of bodies and their

The dependence of the dimensions reaches beyond the rigid body. Among all others the
electromagnetic processes seem to reach the deepest in nature, and it is to be hoped that they
will someday for the foundation of a future unified physics. To give only one example of an
electromagnetic process, let us consider a positive electrical current running perpendicular to
the paper from above to below [from the facing side inward]. In the whole cylindrical space
around the path of the current, magnetic north will be turned in the direction of a clock's hands
for the observer of the paper. The process in space can be symbolized by the usual direction
that the corkscrew drills when it enters. Here we have an example of the physical dependence
of directions independent of rigid bodies. There are many such symmetrical counter-processes
in nature, right and left circulating light, right and left rotating Berg crystal and so on. Whether
nature has a symmetrical counterpart in all of its parts, or whether it is a one sided individual in
many relationships, whose counterpart does not exist, or is unknown, remains questionable.
Signs of the latter are not lacking.

Despite the fragmentary nature of our consideration, you will have received the impression that
time and space consist in ordering relations of physical objects, which are not merely carried
through us, but rather subsist in the inner relationship and the intimate reciprocal dependency
of phenomena. 53

Mach goes on to say that despite Kant's tendency to confuse psychologically

similar objects for one another in the small, perhaps psychophysiological differences
in right and left which are part of our experience do also have physical meaning,
such as our ability to tell our right sides from our left and our use of the "right-hand
rule." Mach has misunderstood Kant, however, since one of the points of the
incongruent counterparts argument is that we are required to bring in the
environment of surrounding space to explain the lack of congruence and cannot do

so merely by an examination of the interior relations of the parts of the body to one
another which will not reveal any difference between right and left-proof only that
the relation of a body to the environment of bodies outside of it is a crucial element
of spatial relations. Mach actually agrees with that point, and is careful to observe it
in his examples, so the question of his having "refuted" Kant really doesn't arise.

More interestingly, Mach's example of the magnetic field setting up at a right

angle to the direction of the current (i.e., the changing electric field of the moving
charges) appears to be his way of showing that indeed nature responds to changes in
direction with a physical reaction or response in another direction. The electric field
changes in the direction going into the paper causing a counter-reaction in the form
of a magnetic field perpendicular and clockwise in the plane of the paper. Here it is
the natural forces themselves whose directions are dependent on one another and not
just when those natural forces are combined in the constitution of bodies. I think in
essence that Mach's point is a generalization of an argument he had made earlier
about the law of least action. For example if a change of position is to take place
there must be a cause of motion, of an acceleration out of rest (or out of motion at a
constant velocity equivalent to rest) and energy must be expended to make a body
move. 54 Likewise, if we are to change the direction of a natural process with respect
to its environment, a velocity, a charge, an angular momentum acquired by a torque,
a force, or in this case a body's orientation, we must expend energy in making this
physical change too. To some degree that was already a Newtonian insight. But
Mach seems to see such changes of direction as physically meaningful and involving
a physical reaction between the body and the forces of the environment that extends
to the smallest parts of matter. The directions in which a body is situated are thus
inherently meaningful for him, because they consist of volume elements for which
the chemical directions are meaningful.

I do not see that Mach's argument is supposed to establish "absolute preferred

directions" in space-for example, a direction for electric forces, for magnetic
forces, for gravity, and for velocity. It is not only plainly false, since natural
processes may be oriented in any of three spatial directions, but indeed it would be a
most un-Machian result to his investigation to reestablish absolute space. An
absolute sense of direction would require not only that all electromagnetic changes
of direction occur at right angles to each other and other electromagnetic forces in
their environment, but also that all such processes would have to be oriented in the
same direction everywhere, i.e., in a set of absolute electromagnetic directions, an
absurdity like a set of tops that can only spin with their axes pointed at the north star
or at some other arbitrary direction in space. Mach's point is not this, but rather that
changes of direction between a volume element and the set of matching directions
that orient it with respect to the environment reveal the dependency between these
conditions. It may then tum out to be the case that some objects can only be oriented
in one way with respect to their environments. For example, it might have happened
that for some objects reversing left and right was only possible by reversing up and
down with respect to the environment, but not front and back.

In this case, in Machian parlance, we would conclude that the object only reacts
chemically with certain environmental directions and not others and thus that it has a
preferred orientation. In the case of the preferred left hand this would mean that the
exchange of directions by rotations about either of the two other axes has no
chemical effect on changing the direction of right and left, in the smallest parts of the
body as well as the largest. Mach regarded these ideas about the physical
significance of direction as important in his overall scheme of space and time. He
wrote to Adler about the central place of the Geissler argument in his thinking about

For me Time-Space questions are essentially physical questions. The question of time will be
solved, I believe, through the correct conception of the second law. As for space, it is perhaps
most intimately connected with the fact that one can orient oneself electrodynamically with
respect to one's own body; that right and left cannot be exchanged without exchanging either
up or down or front and back, in short with the inter-relation of the three dimensions 55


Mach's mention of Herbart in his later career and the connection of his own idea of a
general physical manifold with Herbart's "construction of space from notions of
quality (direction) and magnitude" does suggest that Mach, had he inclined to it,
might have pursued the construction of such a manifold for physics in which the
volume element would have played the role of a spatial element movable to different
positions and orientations in the manifold while retaining its overall "energy value"
or quantity and the directions of space would be replaced by qualities, or more
precisely, energy types. The Geissler argument then snaps into place as a bona-fide
physical dependence of those energies on one another in the smallest parts of matter.

To illustrate the concept ofHerbartian intelligible space as it might have worked

for Mach, consider the directions as if they were separate qualities like {blue,
yellow}, {red,green}, {black, white}. Let us symbolize them as {Up,Down},
{Left,Right}, {Front, Back}. Unlike the pressures considered by Herbart, let the fact
that two directions are paired with one another signifY the fact that a transfer of
energy can take place from one type of natural potential to another along a
continuum. Motion along a direction is to signifY that the energy is changing its form
from the type represented in one direction to the type represented in the other. If U
represents gravitational potential energy due to position in a potential field and D
represents vis viva or the energy of a mass with a velocity, for example, then the fact
of energy transfer from up to down is symbolized by the fact that down is the
opposite of up, or that these directions depend directly on one another, as Mach
described D'Alembert's principle as one of equality of work and the counterwork of
acceleration. If Left and Right represent the exchange of energies between electrical
potential and magnetic potential, they too are dependent on one another and
independent of Up and Down.

I take very seriously Mach's intention to find a "qualitative" or "chemical"

equivalent to phenomena of motion. How could this be done? We might represent
the potential function (or intensity level) of the energy considered as graded
coordinates numbered in ascending order in either the Up direction or the Down
direction. More Up would mean a higher intensity level in that direction, a higher
gravitational potential, a higher electrical potential, while more Down would mean
higher velocity, higher magnetic potential. A point moving at a constant level would
remain at the same coordinate with respect to the two directions, and no transfer of
level would occur, hence the point would not alter in direction. However, in motion
the point would exchange the one level for another as it passes in a definite direction,
although it might also reverse direction and head back up. What moves between
levels are the capacities, which would make them a kind of quantitative magnitude
describing the greatness of the energy changes or the greatness of the quantity of
stuffs moving in a definite direction, whereas the intensity differences, or the
coordinates between which the capacities move, would describe the rate at which the
capacities change their level.

The independence of the directions implies an independence of the energies

from one another, which can be demonstrated by allowing each type of potential to
obey a principle of superposition like that of the parallelogram of forces. Each set of
directions can act physically as if the other were not there. Conversely, the
dependence of the directions implies the transformation of one potential form to
another and involvement across different directions. For example, if the energy of a
falling body is channeled into electrifying another body or producing friction instead
of vis viva, the potential and the vis viva are no longer the directly dependent
directions. The cross-dependence of different directions may also involve several
natural sources of energy at once, which are normally independent of one another. If
a body accelerates down a hill, it may end up with energy of motion, but also
produce a quantity of heat, an electrical charge, and set up mechanical vibrations in
itself and in the materials it comes into contact with. In more violent action, complex
chemical or still more fundamental changes in matter may also result. There is thus,
in most cases, some change of energy along several different types of directions at

Suppose, then, that we consider Herbart's process of constructing lines by means

of associative and dissociative directions. If U and D are two directions, the simplest
thinkable motion for Herbart is expressed as an association ofD and U, followed by
a dissociation and a change of their initial values (for Herbart pressures but here
coordinates) followed by a new association. In the gaps of dissociation the D and the
U come into dependence on two further qualities, which we can call R and L. So
when D uncouples from U it comes into a new dependence on R at a different value.
Meanwhile U uncouples from D and comes into dependence on L at a new value.

Then we allow D and U to become associated with one another and reassociate
Rand L also. Both D and U and Rand L now fuse at new values. But these four
processes are still mutually dependent, especially if the overall balance of energies is
to be upheld. For example, if Up increases, Down decreases; but if Up is increasing,
Right is decreasing, since it is paired with Up, and Left is increasing because it is
paired with Down. Thus, the process in the {R,L} direction will run in the opposite
sense as the one in the {U,D} direction, and thus the required dissociation needed to
trace out an extended region of space is not yet present. Rather, we have the case
where we look head-on down a wire and are unable to turn it through an independent
dimension (or an independent natural process) in which it might appear extended.

One first gets to the notion of a space, working in Herbartian qualities, when one
brings in a third dimension, or a third set of directions like Front and Back {F ,B}.
Now we can dissociate the processes of change in the Up and Down dimension from
the process of change in the Right and Left dimension. We allow Up and Down also
to dissociate from the Right Left process by associating it instead with a Front Back
process. Likewise, we can allow any two processes to dissociate by associating them
with the third. And we can do this in a systematic way so that we allow all three
dimensions to associate and dissociate with one another in tracing out space by
motion among the qualities. For example, we may let Up and Down dissociate by
means of an association with Left and Right, and we may let Front and Back depend
on each another. But on the next round we may allow Up and Down to depend on
each another and let Right and Left depend on Front and Back, thus dissociating
them completely from Up and Down:

{U+D, L+R, F+B},

{U+L, D+R, F+B},
{U+D, L+F, R+B},
{U+F, L+R, D+B},
{U+D, L+R, F+B}.

The importance of the Geissler argument may be clearer, since what it says is
that if Up and Down are to exchange places, i.e., if they are to change direction by
allowing Up to start increasing and for Down to start decreasing, it must be the case
that some other process or set of directions reverses direction also. To see the cross-
dependency of directions, consider the intermediate stage of "dissociation" where
Right depends on Front (or Back), and Left depends on Back (or Front), before we
reassociate them with their usual partners at new magnitudes (see figure below)

Because the conservation of the volume-element is retained across the different

directions, only those component energies change. The Geissler observation is thus a
principle of physical geometry, as Mach had emphasized, though perhaps not of any
thinkable geometry. Thus, for making good on Mach's promises to give a qualitative
or chemical equivalent of motion, Herbart seems to provide the only constructive
model, and Mach's mention of Herbart as constructing a manifold of quality and
magnitude is thus quite a significant one.


Let me finally tum to some criticisms of Mach. It is often raised as an objection to

Mach's "philosophy" that as a scientist and not a philosopher he was not as
consistent as a Kant or a Leibniz. While he does not offer an articulated position like
those others and even dares to leave a few loose ends, I hope that I have shown that
his thinking does indeed merit the title of a philosophy and that many of the
commonplace criticisms leveled at his brand of empiricism are misplaced.

Of the criticisms I think do apply to Mach, I have already mentioned the first in
this chapter-the tension in his writings between a flux of non-repeating elements
and his attempts to construct stable manifolds of energy to replace space, time, and
matter with an abstract construction of qualities and chemical space. Mach would
like to say that one can be a realist about the quality flux and instantaneous
functional connections between its elements and believe at the same time that the
manifold of enduring, repeating events and objects is a necessary economical fiction.
But there appears to me to be a logical gap between the idea that there are
instantaneous functional relations between the elements, i.e., of action and reaction,
and the idea that these within-time relations imply the existence of enduring objects
over times, surviving from one snapshot of time to another. If there are no temporal
relations across those snapshots, binding one to the next, it does not seem possible to
explain the appearance of objects not to mention a "chemical manifold" by adding
together a series of snap-shots.

For example there is no reason why each snapshot might not contain stable
forms of objects with boundaries and yet wildly changing objects between one snap
shot and the next (if "next" has any meaning here). Objects really do become
economical illusions based on nothing but need. And while Mach's point is well
taken that what we ordinarily consider localized physical objects or egos are actually
already spread out into their effects and connections with environments of other
objects and cannot be completely localized, still there is a permanence to these
objects that seems to exceed pragmatic grounds alone, and which Mach cannot

Second, on his purported connection to energetics, we find Mach still arguing in

"A Consideration" that all of the changes in nature are due to differences in energy
level, or due to falls in level of the different energy values all in step with one
another in time. 56 Here he sounds like Helm and Ostwald, using their "law of
happenings," and he does not sufficiently attend to Planck's criticism that there are
real events that occur in nature, such as diffusion, mixing, and so forth, that are not
changes of temperature or other level values, except in a very indirect way of
speaking (as Clausius had suggested) about the "lost work" that would be needed to
restore the initial states. That Mach believed in the basic fact of irreversible
processes is clear, and it is even clear that he believed all processes in nature were of
that kind. But if so, the law of happenings, or his own law of differences, is far from
an explanation, since happenings without differences are not explained on that view
of things. And the law ofhappenings contains no grounds for irreversibility.

I also have a criticism of Mach's "antimetaphysical" strategy, not the

stereotypical "principle of verifiability" or other confusions between Mach and the
logical positivists, which I have largely avoided as unhelpful, but rather the subtle
physico-physiological argument analyzed in the earlier chapters. Antimetaphysics is
the ruling out of any extension of purely sensory properties to objects when they are
outside the sensory manifolds. In the most obvious case, Mach said he did not
believe physiological conceptions of space, time, and matter should be mirrored in
the concepts of physical theory. And yet perhaps the clearest case of a psychological
property bound to a sensory manifold is human sensation. But Mach was willing to
disobey his own stricture and follow Muller, granting that sensation indicated a real
property of nature, or came closer to it than any other aspect of human experience.
Space, time, permanent matter and objects, Mach claimed these were mere holdovers
from our primitive past and the influence of our senses. But not so our sensations,
through which we participate in a fragment of the real. Why this double standard and
does it not need a special argument, or was Mach hoping for experimental proof via
some future medical and psychophysical breakthrough that would allow human
beings to experience his world elements directly?

As in so many philosophies and natural philosophies, the crucial claim turns out
to be an existence claim. It was an article of faith for Mach that the elements of the
world consisted in a mass of sensations of human beings, animals, and even sense-

like qualities in material objects. As interesting as that idea is, it is, like so much of
his Mach's natural philosophy, an empirical proposal, open to test or refutation based
on the progress of a future science.

At the end of his famous 1882 Economy lecture, Mach remarked that in a future
science, human beings would confront the world of nature directly. To this effect he
quoted a Chinese philosopher, Licius (Lao-Tsu), who we may imagine speaks for
him, and who told his students when passing a heap of old bones: "These and I alone
have the knowledge that we neither live, nor are dead."

1 John Blackmore has catalogued Mach's influence

quite thoroughly in his Ernst Mach's Vienna.
Dordrecht: Kluwer, 2001 and Ernst Mach's Prague, Dordrecht: Kluwer 2003. For others see Barry Smith
"Austrian Origins of Logical Positivism" in Barry Gower, ed. Logical Positivism in Perspective. Totowa:
Barnes and Noble, 1987, pp. 35-68.
2 See the discussion of Blackmore and Wolters on atomism below.
3 Albert Einstein "Ernst Mach" Physikalische Zeitschrift
17 (1916): 101-104.
4 The concept of a gravitational field without a source mass,
but due only to bodies at rest in accelerated
reference frames, is a non-Machian idea due entirely to Einstein, see Jiirgen Renn "Mach, Einstein and
the Third Way to General Relativity" Max Planck Institutfor Wissenschaftsgeschichte reprints #9 1994;
John Norton "What was Einstein's Principle of Equivalence?" Studies in History and Philosophy of
Science 16 (3) Sept 1985: 203-246.
5 On Mach's attempts to understand
the special and general theories with the help of his collaborator
Phillip Frank see Gerald Holton "Mach, Einstein and the Search for Reality" pp. 182-185 in Robert S.
Cohen and Raymond J. Seeger, eds. Ernst Mach: Physicist and Philosopher. Dordrecht: D. Reidel, 1970
and "More on Mach and Einstein" in (DL), pp. 163-276. In addition to Friedrich Herneck's account of the
studies of relativity between Frank and Mach, "Die Beziehung zwischen Einstein und Mach
dokumentarisch dargestellt" Wissenschaftliche Zeitschrift der Friedrich-Schiller- Universitat Jena, Math.
naturwiss.- Reihe 15(1966), Arnold Koslow has told me in personal communication that Frank believed
Mach was favorably disposed towards the special theory.
6 Quoted in Holton (1970), p. 184.

7 (AS), p. 12.

8 (M), p. 579.

9 A good example of the "received view" of Mach is Peter

Alexander. Sensationalism and Scientific
Explanation. London: Routledge and Kegan Paul, 1963, pp. 4-22; For an all-encompassing sensationalist
reading of Mach and Avenarius see V.I. Lenin. Materialism and Empirio-Criticism. Moscow: World
Languages Publishing House, 1952.
10 Gerald Holton "Ernst Mach and the Fortunes
of Positivism in America" Isis 83 (1992) 27-60; Barry
Smith "Austrian Origins of Logical Positivism."
II (AS), p. 16.
12 Andy Hamilton "Ernst Mach
and the Elimination of Subjectivity" Ratio 3.2 (1990), pp. 128-131; see
also Moritz Schlick. General Theory ofKnowledge. La Salle, Illinois: Open Court, 1974.
13 See Manjulekha Bhattacharya "Ernst Mach:
Neutral Monism" Stud Int. Filosofia 4 (1972): 167-171.
14 Lenin Materialism and Empirocriticism,
pp. 48-49.
15 (AS), p. 27
16 Schlick General Theory of Knowledge,
17 (V), p.364-365.
18 Michael Heidelberger Die Innere Seite der Natur (Frankfurt
a.M. Vittorio Klostermann, 1993).
19 (V), p.364.
20 "Ober einige Hauptfragen der
Physik" Sommer 1872 (NL 174/11003), pp. 1-4; published in German in
Rudolf Haller und Friedrich Stadler eds. Ernst Mach: Werk und Wirkung. Wien: Holder Pichler Tempsky,
1988, p. 173 ff.
21 (AU), p. 68.
22 Ibid., pp. 71-72.
23 (PWV), p. 244.
24 (M), p. 559.
25 Gabriele Rabel "Ernst Mach und die Reali tat der Aussenwelt"
Physikalische Zeitschrift 21 ( 1920) 434.
26 Erich Becher "The Philosophical
Views of Ernst Mach" Philosophical Review 14.5 (1905), p. 547.
27 NotebookdatedAprill873 (NL 174/2/3).
28 Frank Kerkhof, ed. Symposium aus Anlass des 50.
Todestages von Ernst Mach. Freiburg im Breisgau:

Ernst Mach Institut, 1967-68; Robert S. Cohen and Raymond J. Seeger, eds. Ernst Mach: Physicist and
Philosopher. Dordrecht: Reidel, 1970; Jaako Hintikka, ed. An Ernst Mach Symposium, Synthese 18 (1968):
29 "Ernst-Mach-Bibliographie" Centaurus 8 (1963): 189-237.
30 "Die Beziehung zwischen Einstein und Mach dokumentarisch dargestellt", 45 ff.
31 Mach's Principle from Newton's Bucket to Quantum Gravity. Julian Barbour and Herbert Pfister eds.

Boston: Birkhauser, 1995.

32 John Blackmore. Ernst Mach: His Work Life and Irifluence. Berkeley: University of California Press, 1973,

Chapter 3 and especially Gerold Holton's "Mach, Einstein and the Search for Reality" in Cohen and Seeger
eds. Ernst Mach: Physicist and Philosopher (Dordrecht: D. Reidel, 1970).
33 Gereon Wolters. Mach/, Mach II, Einstein und die Relativitiitstheorie. Berlin: Walter de Gruyter, p. 9, see

also pp. 204-205.

34 Studies in the History and Philosophy ofScience 15 (1) 1984: 1-22.
35 "Philosophy of Science: A Subject with a Great Past" in Minnesota Studies in the Philosophy of Science

Vol. V{l970)pp.l79-181.
36 Feyerabend has in mind specifically Carl Hempel's nomological-deductive or "covering law" model of

explanation, his theory of inductive theory confmnation as well as Nelson Goodman's "grue" paradox of
inductive confirmation.
37 1bid, p. 181.
38 In the text of the Optik preface, which Blackmore says is written in "the pure Machian style" Mach

does not say he out-and-out rejects the theory. He merely says that it will form "an apers;u" in the broader
science of the future he envisions. Apers;us are insightful, if superficial, perceptions.
39 Gerald Holton "More on Mach and Einstein" in {DL) 163-276; See also Blackmore's comprehensive

marshaling of evidence for and against Wolters in "Mach iiber Atome und Relativitiit-neuste
Forschungsergebnisse" in Rudolf Haller und Friedrich Stadler, eds. Ernst Mach: Werk und Wirkung.
Wien: Holder Pichler Tempsky, 1988, pp. 463-483; also Gereon Wolters "Atome und Relativitiit: Was
meinte Mach?" in Haller und Stadler. Wolter's response is "Revisionismo nella filosofia e nella storia
della scienza. II caso di Mach e Ia teoria della relativita" Dianoia (2) 1997 155-172.
40 WoltersMach/, Mach lip. 205.
41 "Poetic hnagination and Economy: Ernst Mach as Theorist of Science" in {DL) pp. 215-228.
42 (DL), p. 240.

43 Barry Smith "Austrian Origins of Logical Positivism," p. 54.

44 "Our elements are only provisional, as those of alchemy were in the past and those of currently

accepted chemistry are now. Although for our purpose of eliminating sham problems reduction to these
elements seemed the best way it does not follow that every scientific inquiry must begin with them."
(KE), p.l2n.
45 Rudolf Carnap "Intellectual Autobiography" in Paul Arthur Schlipp, ed. The Philosophy of Rudolf

Carnap. La Salle, IL: Open Court, 1963, p. 50, Cf. pp. 16-18, p. 45.
46 1bid. p. 18.
47 "Carnap's Machist Sources" in Blackmore, Itagaki and Tanaka, eds. Ernst Mach's Vienna 1895-1930.

Dordrecht: Kluwer, 2001, pp.l59-186.

48 Rudolf Carnap. The Logical Structure of the World. Berkeley: University of California Press, 1969,

49 Michael Friedman, "Carnap's Aufbau Reconsidered" Nous 21.4 (1987), pp. 521-545.

5 For a critique of attempts to formalize similarity which seems directed at quasi-analysis, see W.V.

Quine's essay "Natural Kinds" in his Ontological Relativity and Other Essays. Columbia University
Press, 1960.
51 Mach believed similarity was just a way of expressing a partial identity. If two objects are similar to

one another it means they are identical in only one property among others. Hence, I think it could be
argued that anything is similar to anything if enough properties are brought into play. W.V. Quine, in his
own construction of science from stimuli relies instead on an "innate similarity standard."
52 (AS), pp.68-69; pp.262-309.
53 Nelson Goodman's Structure of Appearance, which is often thought to be an advance on Carnap's

Aufbau, contains essentially the same nominalistic conceptions of qualia and similarity.
54 Erwin N. Hiebert "The Genesis of Mach's Early Views on Atomism" in Cohen and Seeger, eds., Ernst

Mach: Physicist and Philosopher. Dordrecht: D. Reidel, 1970, pp. 79-106; Stephen G. Brush "Mach and
Synthese 18 (1968): 192-215; Laurens Laudan ''The Methodological Fowulations of Mach's Anti-Atomism
and Their Historical Roots" in P.K. Machamer and R.G. Turnbull eds., Motion and Time Space and Matter.
Ohio State University Press, 1976, pp. 390-417.
55 (DL), p. 151.
56 John Blackmore "An Historical Note on Ernst Mach" British Journal for the Philosophy of Science

36(1985), p. 301; John Blackmore "Ernst Mach Leaves 'The Church of Physics"' British Journal for the
Philosophy of Science 46(1989): 519-540; John Blackmore "Mach iiber Atome und Re1ativitiit-neuste
Forschungsergebnisse"; also Gereon Wolters "Atome und Relativitiit: Was meinte Mach?'' also in Haller
und Stadler.
57 (DL), pp. 152-153.
58 "Mach's Theory ofResearch ... " pp. 19-20.
59 (M), p.589.
60 (SG), p. 138.

62 (SG), p. I 03: "Physics would be justified in considering an extended material continuum, to each point

of which a temperature, a magnetic, electric and gravitational potential were ascribed, as a portion or
section of a multiple manifold."
63 "Mach's Theory ofResearch... " p. 11.
64 (DL), p. 139.
65 Quoted in Morris Kline, Mathematical Thought From Ancient to Modern Times Vol. I (Oxford: Oxford

University Press, 1972) p. 329.

66 "Ernst Mach war es, der in seiner Geschichte der Mechanik an diesem dogmatischen Glauben [d.h. dem

Glauben an die Mechanik als gesicherte Basis der Physik] riittelte; dies Buch hat gerade in dieser Beziehung
einen tiefen Einfhill auf mich als Student ausgeiibt." Einstein "Autobiographical Notes" in Schlipp ed. Albert
Einstein Philosopher-Scientist (Evanston, IL: Open Court, 1949), p. 7. "It was Mach, in his history of
Mechanics, who shook this dogmatic faith [i.e. the faith in mechanics as the sure basis of physics]; in this
very connection, this book exercised a deep influence on me as a student."
67 Einstein "Autobiographisches" in Paul Schlipp, ed., Albert Einstein Philosopher-Scientist, p. 53.


1 Hans Henning. Ernst Mach als Philosoph, Physiker und Psycholog. Leipzig 1915, Chapter 1.
2 K.D. Heller. Ernst Mach: Wegbereiter der modernen Physik. Wien: Springer-Verlag, 1964, p. 6.
3 Erwin Hiebert "The Influence of Mach's Thought on Science," Philosophia Natura/is 21 (1984), p.598.

4 This unhappy episode is told in John Blackmore. Ernst Mach: His Work, Life and Influence. Berkeley:

University of California Press, 1973, and in Dieter Hoffmann's "Ernst Mach in Prag" pp. 141-178 in
Dieter Hoffmann and Hubert Laitko eds. Ernst Mach: Studien und Dokumente zu Leben und Werk.
Berlin: Deutscher Verlag der Wissenschaften, 1991, (DL) pp. 29-47.
5 William Johnston. The Austrian Mind. Berkeley: University of California Press, 1972, p.49.
6 For the influence of the Exner-Bonitz reform on Mach, see his autobiographical fragment "Ernst Mach"

written in the third person (DL), p. 21, Henning. Ernst Mach als Philosoph. Chapter I and William
Johnston. The Austrian Mind. Chapter 19.
7 Wolfram Swoboda "The Thought and Work of the Young Ernst Mach and the Antecedents of his

Philosophy" doctoral dissertation. University of Pittsburgh, 1973, pp. 93-94.

8 Swoboda p. 90.
9 Hugo Dingler. Die Grundgedanken der Machschen Philosophie. Leipzig: J.A.Barth, 1921, pp. 19-20.

10 (V), p. 336.
11 Johnston. The Austrian Mind. p.230.
12 Die Innere Seite der Natur. Frankfurt a.M.: Vittorio Klosterrnarw, 1993, p. 214.
13 Dingler. Grundgedanken. pp. 28-29.
14 Ober die Erhaltung der Kraft. Leipzig: Wilhelm Engelmann, 1889, (origina\1847).
15 Helmholtz's ideas about the ultimacy of atoms may have changed over time. Heidelberger quotes

Helmholtz as saying in 1871 that "the foundation of theoretical physics should not be derived from the
hypothetical assumption of the atomic structure of natural bodies." Die innere Seite der Natur. p. 213.
Heidelberger claims Helmholtz's change of heart encouraged Mach to publish his anti-atomist views in
16 " the phenomena ofliving organisms are only complicated physical phenomena. The medicine of today

is long used to considering physiology as applied physics. A great number of phenomena which would
otherwise be attributed to vitalism are already traced back to physical laws and it becomes increasingly
probable that this will succeed more and more for all processes in organisms." (C), pp. 1-2. Cf. "Uber das
Sehen von Lagen und Winkeln durch die Bewegung des Auges" (SW) 41(1861) 215-222, p. 224: "Sollten
in Zukunft auch manche von den bier ausgesprochenen Ansichten modificiert, manche verworfen
werden, so scheint mir Eins klar, dass gerade nur diese Art von Untersuchungen, jene Bausteine zu einer
exacten Psychologie zu gewinnen sind, welche ganz ausserhalb der Seele, rein im kiirperlichen
Organismus liegt."
17 (V), p. 364.
18 (AS), p. I. See also Erwin N. Hiebert. "The Influence of Mach's Thought on Science," p. 606.

19 K.D. Heller. Ernst Mach: Wegbereiter der modernen Physik, pp. 11-12.

20 Blaserna-Mach-Peterin. "Uber elektrische Entladung und Induction" (SW) 37 (1859): 477-524; Mach

"Ober die Anderung des Tones und der Farbe durch Bewegung." (SW) 41 (1860): 543-560; "Ober das
Sehen von Lagen und Winkeln durch die Bewegung des Auges." (SW) 43 (1861): 215-224.
21 Swoboda pp. 20-74, Floyd Ratliff "On Mach's Contributions to the Analysis of Sensations." in Cohen

and Seeger, eds. Ernst Mach Physicist and Philosopher. Dordrecht: D. Reidel, 1970.
22 "0ber die Anderung des Tones und der Farbe durch Bewegung." (SW) 41 (1860): 543-560.

23 (SW) 43(1861): 215-224.

24 (AU), p. 1.
25 "Ober das Sehen von Lagen und Winkeln." p. 223.

26 John Blackmore, Ernst Mach: His Work, Life and Influence, p. 23.
27 K.D. Heller, p. 12.
28 See Frau Anna Karma Mach's remarks in Frank Kerkhof, ed. Symposium aus An/aft des 50. Todestages

von Ernst Mach. Freiburg: Ernst Mach Institut, 1966, p. 10.

30 "Zur Theorie der Pu1swellenzeichner." (SW)46(1862): 157-174; "Ubcr cine neue Einrichtung des

Pulswellenzeichners." (SW) 47(1863), 53-56.

31 "Zur Theorie des Gehiirorgans." (SW) 48 (1863): 283-300.
32 John Blackmore, Ernst Mach: His Work, Life and Influence. p. 14.

33 (DL), p. 24.

34 Swoboda, p. 136.

35 (CE), p. 87.

36 (CE), pp. 75-80.

37 (CE), p.85

38 (AS), p. 352.

39 Draft Foreword to the Analysis of Sensations, Russian edition translated in (DL ), p. 116.

40 (CE), pp.9-10.

41 (AS), p. 30 n.

42 Mach criticized Kant for thinking that mirror-image left and right hands were physically

indistinguishable. "Symmetric geometrical figures are, owing to our symmetric physiological

organization, very easily taken for the same, whereas metrically and physically they are completely
different. A screw with its spiral winding to the right and one with its spiral winding to the left, two
bodies rotating in contrary directions, appear very much alike to the eye. But we are for this reason not
permitted to regard them as geometrically or physically equivalent. Attention to this fact would avert
many paradoxical questions. Think only of the trouble such problems gave Kant!" (KE) p. 288. Some five
years later, however, Mach believed that differences in spatial properties often did point at underlying
physical differences. For example the fact that the dependence of electrical currents and magnetic fields
followed a right hand rule (when the fingers of the right hand curl in the direction of the field the thumb
points in the direction of the current) impressed Mach greatly as a case in which physiological asymmetry
could be used in a physical case. "Who knows," he said, "whether the Kantian "a priori" cannot find new

light along this indicated path?" "Eine Betrachtung iiber Raum und Zeit" (PWV), p. 508.
43 (DL), p. 115.
44 (CE), p. 16.

45 See my "Kant, Herbart and Riemann" Kant-Studien 2003.

46 Prolegomena to any Future Metaphysics. Carus and Ellington trans. Indianapolis: Hackett, 1977, Part I,

Remark III.
47 (AS), p.6
48 For all the differences between Berekeley and Mach, this antimetaphysical critique is indeed by all

rights attributable to Berkeley's critique of Locke on substance. The idea of a "support" for qualities
behind sensation is said by Berkeley to be itself a notion derived from sense, as are all of the primary
qualities of matter and figure said to explain the causes of sensation.
49 (DL), p. 116.
50 (AS), Preface to Fourth Edition 1902, xli

51 "Die Leitgedanken meiner naturwissenschaftlichen Erkenntnislehre und ihre Aufnahme durch die

Zeitgenossen" Physikalische Zeitschrifi 11 (1910), p. 604, partially translated in (DL).

52 (AS), pp. 361-362n.
53 See my "Mach, Hume and Functionalism" in John Blackmore and Setsuko Tanaka, eds. Ernst Mach's

Prague. Dordrecht: Kluwer, forthcoming.

54 (DL), p. 26.

55 (AS), p. 331.

56 (AS), p. 362.


I (CE),p. 86
2 (C), pp. 11-12.
3 (C) p. 15.

4 Wolfram Swoboda "The Thought and Work of the Young Ernst Mach and the Antecedents to His

Philosophy" doc. diss. University of Pittsburgh 1973, pp.l58-159. Sections published in German as
"Physik, Physiologie und Psychophysik-Die Wurzeln von Ernst Machs Empiriokritizismus" in Rudolf
Haller und Friedrich Stadler Ernst Mach-Werk und Wirkung (Wien: Verlag Holder Pichler Tempsky,
1988), pp. 356-403.
5 Erwin N. Hiebert "Mach's Early Views on Atomism" in Ernst Mach: Physicist and Philosopher Dordrecht:

D. Reidel, 1970, p. 95.

6 (C), vi.

7 (C), pp. 271-272.

8 (C), p. 256.

9 (CE), p.87.

10 (SG), pp.138n-139n.
11 Hiebert 1970, p 99.

12 Hiebert 1970, p!OO.

13 Ibid.
14 Joachim Thiele Wissenschafiliche Kommunikation: Die Korrespondenz Ernst Machs, (Kastellaun: A.

Henn Verlag, 1978, p. 32.

15 Die Fortbewegung d. Materie ist eine Welle. Eine Differenz die fortschreitet...Das Licht als periodische

Potential. Notebook dated 18 September 1874: (NL 174/02/05).

16 (R), p.232.
17 (DL), p.24.
18 (CE), pp. 86-87.

19 See for example David E. Leary "The Historical Foundation of Herbart's Mathematization of

Journal of the History of the Behavioral Sciences 16(1980): 154-155.
20 (CE), p. 87; (V), p364; (CE) p. 87: "In my lectures on psychophysics I already stated clearly that we are not

justified in thinking of atoms spatially."

21 (V), p. 364.
22 (V), p 364.

23 (V), p. 364.
24 G.T. Fechner. Elements of Psychophysics.
Helmut A. Adler, trans. (NY: Holt, Reinhart and Winston,
1965), pp.l -2.
25 (V), pp. 363-364
26 Andreas Laass "Yom Sinne des machschen Philosophierens"
in Ernst Mach: Studien und Dokumente
zu Leben und Werk (Berlin: Deutscher Verlag der Wissenschaften, 1991), p. 213.
27 (CE), pp.Sl-54.
28 (CE), pp. 53-54.

29 In connection with his late views of space

as a Riemannean chemical manifold, Mach reversed himself
later and decided that space was three-dimensional after all (CE), p. 94; (M), pp. 103-104. However,
behind this three dimensionality were still physically determined facts: three essentially different
"chemical directions" (AS), p.339 and a physical dependence of these directions on one another (PWV),
pp. 506-508.
30 "Uber die physikalische Bedeutung der Gesetze
der Symmetrie" Lotos 21, p. 146.
31 Symmetrie, p. 147; Cf. (CE), p.50.

32 Symmetrie, p. 147n.

33 (CE), p. 60.

34 (CE), p. 50.

35 "Mach and Ehrenfels" in B. Smith ed. Foundations of Gestalt

Theory (Munich and Vienna:
Philosophia, 1988) pp. 124-57. They call Mach's functions non-causal dependencies, but I think the
better word is causal, but aspatial and atemporal, in the sense that Mach's notion of functional
connections is prior to spatial and temporal properties usually included in the notion of cause.
36 It is true of course that Mach regarded
the physical and the psychical as two different orderings of the
same content (CE), p.91; (AS), p.16. So, for example, the orderings of the physical realm may be more
general than those elements presented to sensibility in manifolds of physiological space and time.
37 (AS), pp.38-39.

38 (M),p. 337.

39 (KE), pp.l3-14n, Mach criticizes G. Heymans by

saying: "One is curious why, if on Heyman's view the
method of metaphysics is exactly the same as that of natural science, though extended to a wider field, he
insists on calling it metaphysics, a term with so distasteful a flavor since Kant, and one that seems
contradicted by the addition "on the basis of experience."
40 Wo philosophische Fragen beriihrt
werden, was nicht ganz zu vermeiden war, hielt ich mich auf dem
Boden der Naturwissenschaft. Teils wei! dieser Standpunkt eine gewisse Berechtigung hat, teils auch in der
Oberzeugung, dass man von diesem Standpunkte einige Zeit die Aussicht miisse genossen haben, urn die
Notwendigkeit zu fiihlen, dass die Untersuchung auch von einer anderen Seite her gefiihrt werden miisse;
manche Menschen gelangen nie dahin.
41 (H), p. 233ff.

42 (KE), pp. 2-3.

43 (CE), pp. 88-89.

44 Ibid.

45 [Jourdain's note: For this purpose it would be necessary

also to know the velocities of the various parts at
that instant.]
46 (AS), pp. 362-363.

47 "Uber die Okonomische Natur der physikalischen Forschung"

(PWV), p. 234.
48 (CE), pp. 71-72.

49 Whether Mach expresses

the elements individually as pressures or in their variations as elements of
work, the two expressions may have been equivalent for him. This would be the case especially if
pressures, or forces, never act but in overcoming a given force in a given time, as for example in
dynamical applications of D' Alembert's principle. In such cases the elements of those processes are
expressed as work. Mach discusses this case in his notebooks of the early 1870s and calls D' Alembert's
Principle "the more generalized principle of action and reaction." (12 April 1873 NL 174/2/3).
50 (CE), pp. 61-62.

51 (CE), pp. 70-71.

52 (CE), p. 62.
53 (CE), p. 63.
54 (CE), p. 63.

55 (CE), pp. 62-63.

56 (CE), pp.42-43.
51 See P.M. Hannan The Natural Philosophy of James Clerk Maxwell (Cambridge Cambridge University

Press, 1998).
58 Ausgleich d. Differenzen

Nur wo Differenzen is ein Vorgang m6glich, wei! nur hier das Bestimrnende nicht fehlt Mit d. Ausgleich der
Differenzen nothwendig Gleichgewicht. Vergr6sserung d Differenzen. Vorgang bestimrnende
Vergr6sserung? Die Ver!inderung geht immer auf Beseitigung der Veriinderungsursache. Ein Zustand wird
nicht in einen anderen iibergehen der noch gr6ssere Ursachen der Ver!inderung in sich schliesst. Was hat bei
dem letzten Punkt die Zeit zu tun. Verkleinerung der einen Differenz, kann eine Vergr6sserung der anderen
nach sich ziehen. Aber alles in allem muss doch eine Verkleinerung eintreten... Notebook dated April 12
1874 (NL 174/02/03).
59 Gleichheiten k6nnen keinen Grund zu Ungleicheiten in sich tragen. Da liesse sich keine Regel angeben

nach welcher das vorginge. Ungleichheiten kllnnen nur kleiner werden.

60 (AS), p. 350.
61 (AS), p. 352.
62 Mach to Popper November 13, 1894 (AU), p. 19.
63 (KE), p. 359.
64 (M), p. 228.

65 A Treatise ofHuman Nature Part I. Sec. 4.

66 Verbindung auch gegeben (Notebook dated April 1882 NL 174/2/22).
67 Die Elemente und der Zusammenhang. Der Zusammenhang ist auch eine Tatsache. Was fiir eine? Eine

68 See the passages quoted in Paul Feyerabend "Mach's Theory of Research and its Relation to Einstein"

Studies in History and Philosophy ofScience 15(1) 1984, pp. 6-7.

69 (M), p. 356; see also Mach's discussion of Fourier's principle which allowed for a similar view of

phenomena involving the flow of heat (H), p. 113: "In saying that every material point tends to the mean
temperature of the surrounding points, the result of Fourier's theory is so expressed that it appears almost
self-evident, and is very close to our instinctive perception. It lies as close as the observation that all
heavy bodies left to themselves sink. Science confirms, in both cases, an obvious fact, only more exactly
and completely in all respects than involuntary and undisciplined observation is able to do. In mechanics
and in the theory of the conduction of heat it is, really only one great fact in each domain which is
70 William James "A World of Pure Experience" Journal of Philosophy, Psychology and Scientific

Methods Vol 1(20) 1904: "The relations that connect experiences must themselves be experienced
relations, and any kind of relation experienced must be accounted as 'real' as anything else in the system."
71 See James' profession de foi "Radical Empiricism" Writings of William James J.J. McDermott, ed.

Chicago 1977 p. 136.

72 Hans Henning Ernst Mach als Philosoph, Physiker und Psycholog (Leipzig: 1915), p. 19.
73 Henning pp. 18, 23.
74 Kurt Gerhards Machs Erkenntnistheorie und der Realismus (Stuttgart: W. Spemann, 1914), p. 166.

15 Herbart Allgemeine Metaphysik Sec. 201, p. 55; see also G. Boudewijnse, D. Murray and C. Bandomir

"The Fate of Herbart's Mathematical Psychology" History ofPsychology 4(2) 2001, p. 116.
76 See (AS) p. 46. see also Mach to Petzoldt Dec 12 1914 in (AU), letter #113, and Barry Smith and

Kevin Mulligan "Mach and Ehrenfels: Foundations of Gestalt Theory" in B. Smith Foundations of
Gestalt Theory Munich: Philosophia, Sec 8.


1 Wolfram Swoboda, "The Thought and Work of the Young Ernst Mach" doc. diss. (Pittsburgh, 1973), p.
2 Lott to Mach October 11 1864 (NL 174/5/1368); Fechner to Mach April 18, 1864 in (WK): "Ober die

Herbartsche Psychologic will ich mich nicht mit Ihnen streiten, verrniichte ich auch nicht, da es zu lange
her ist, daB ich umsonst versucht habe, mich mit ihr zu befreunden; und unstreiting ist in unserer
beiderseitigen Stellung dazu etwas Glaubensache. Ich will Ihr also einfach Gliick wiinschen, daB sie in
Ihnen einen neuen so gewiegten Vertreter gefunden hat. .. "
3 "On the Hypotheses that Lie at the Foundations of Geometry" in D.E. Smith, ed., A Source Book in

Mathematics, (New York: McGraw Hill, 1929).

4 See Max Jammer, Concepts of Space, Third Edition (New York, Dover, 1994) p. 177: "Certain ideas in

Herbart's philosophy seem to have had a great influence on Riemann and H. Grassmann in their
formulation of a manifold with an arbitrary number of dimensions."
5 "The Herbartian Psychology" Parts 1-2 Mind 13(51) July 1888: 321-338 and 13(52)0ctober 1888: 473-

6 J.F. Herbart, Allgemeine Metaphysik, (Konigsberg 1829) in Herbarts Werke 8. Band (Herausgegeben

von K. Kehrbach und 0. Flilgel), p. 12.

7 J.F. Herbart, Allgemeine Metaphysik (Konigsberg, 1829), Sec.l69.

8 Metaphysik, Sec. 171, p. 21.

9 Metaphysik, Sec. 170. p. 21.
10 Two assumptions are necessary for that to be true. First a standard quality must be chosen as a unit of

intensive magnitude and second those measures must be transitive for all other qualities for which the
standard is not in contact.
11 David E. Leary ''The Historical Foundation of Herbart's Mathematization of Psychology" Journal of

the History of the Behavioral Sciences 16(1980): 154.

12 Herbart Psychologie als Wissenschaft Zweyter Theil p. 193.
13 Psychologie als Wissenschaft Zweyter Theil p. 192.

1\V), p.l68ff. Beilage.

15 For this episode in Mach's Herbartian apprenticeship, see Andreas Laass "Vom Sinne des Machschen

Philosophierens" in Hoffmann und Laitko, eds. Ernst Mach: Studien und Dokumente zu Leben und Werk
(Berlin: Deutscher Verlag der Wissenschaften, 1991).
16 Andreas Laass "Vom Sinne des machschen Philosophierens" in Ernst Mach: Studien und Dokumente

zu Leben und Werk (Berlin: Deutscher Verlag der Wissenschaften, 1991), p. 213.
17 Metaphysik, Sec. 234, p. 103.
18 Gedanken von der wahren Schiitzung der lebendigen Kriifte in Kant's Gesammelte Schriften, Berlin,

Preussische Akademie der Wissenschaften, 1911.

19 K.d.r. V. A 266-A267, B 322-B 323.(Kant's Gesammelte Schriften, Berlin, Preussische Akademie der

Wissenschaften, 1911).
20 K.d.r. V., B 210-215.
21 K.d.r. V., A 166, B 207.
22 K.d.r. V. A 291-292, B 347-348.
23 Norman Kemp Smith. A Commentary on Kant's Critique ofPure Reason (London: MacMillan, p.35l).
24 Wayne Waxman has advanced this interpretation of Kant's transcendental idealism, i.e. that space and

time first arise in the imagination, and that raw sensation is aspatial and a temporal. See his "What are
Kant's Analogies About?" Review of Metaphysics 47(September 1993), pp. 81-82. Professor Alan
Ponikvar called this article to my attention.
25 K.d.r. V., A 101-102.
26 (V), p. 204

27 Psychologie als Wissenschqft Sec. ll 0.

28 Psychologie als Wissenschqft Sec. 112-113.
29 According to Geert-Jan Boudewijnse, David Murray and Christina Bandomir in their article "The Fate

of Herbart's Mathematical Psychology" History of Psychology 4(2) 2001, the famous "forgetting curve"

of Ebbinghaus is an example of the Herbartian process of inhibition and fusion in series. See pp. 112-13.
30 Mach "Vortrlige fiber Psychophysik" Oesterreiche Zeitschrift for praktische Heilkunde 9 (1863).

31 Psychologie als Wissenschaft 109-116.

32 For absolute accuracy, each of the dissociating series must be perpendicular to one another, which is

impossible to represent on the page, but may be readily imagined, especially if the dissociation is causal.
33 See G.F. Stout "The Psychological Theory of Extension" Mind Vol. 13 (51) 1888, pp. 418-424.
34 Gary Hatfield The Natural and the Normative: Theories of Spatial Perception from Kant to Helmholtz

(Cambridge: M.l.T. Press, 1990) p. 126.

35 K.d.r. V., A 173, B 215.
36 K.d.r. V., A 173, B 215- A 175, B 217.
37 K.d.r. V., A,B 182-183.
38 Metaphysik, Sec. 287.

39 Or to the magnitude in the Ding an sich, if the negation content of a sensation is meant to represent

positive intensive magnitude in the unknown thing in itself. Hegel may have been speaking of Kant or
Fichte in the "Kraft und Verstand" section of the Phiinomenologie des Geistes when he says that the
world beyond the senses is represented as an unknown negative force in an inverted world.
40 It is not a sufficient requirement to determine the objects of the community uniquely, for the action and

reaction pairings could still be done in any number of possible ways.

41 Metaphysik, Sec. 244, p. 118.
42 It would also be important to know in advance not only the absolute magnitude or greatness of the

force (5 newtons) but also the power delivered by the force, or the intensity with which it acts (which
shows up temporally as the difference between 5 newtons per second versus 5 newtons per hour). As
expressed by D'Alembert's principle, when an impressed force accelerates a stone, the reason why it does
not accelerate it faster than it does is that the stone's resistance succeeds in equilibrating the impressed
force over a very short time interval. The presence of weaker equilibrating forces that take time to
overcome is the reason why an acceleration should take time at all and not happen instantaneously.
43 Metaphysik Sec. 253. p. 136: "From the determinate beginning point A on, we find on the line AB two-

fold every conceivable distancing that allows itself to be thought precisely and without contradiction.
Namely right and left to indicate the opposing processes with well-known names."
44 Hermann Grassmann (while keeping the idea that natural number lines are traced out by travelling

processes of "joining and separation") significantly clarified Herbart's ideas in the Ausdehnungslehre by
introducing binary symbols u for joining(+) and n for separation(-). The properties of communitivity,
associativity, distribution, identity can then be added or taken away, in so far as the constructing process
shows or fails to show them. A New Branch of Mathematics: The Ausdehnungslehre of 1844 and Other
Works, Translated by Lloyd C. Kannenberg, (LaSalle, Illinois: Open Court, 1995), Introduction and
Chapter One to Sec. 15.
45 In Grassmann's treatment for example, the generating process is considered elementary when the

elements produced are always equal, that is if b is produced from a by the identical process by which c is
produced from b, (homogeneity or indifference to position), and if the reverse of the process produces
equal elements indifferent to direction (isotropy or indifference to direction). Sees. 13-14.
46 Metaphysik Sec. 258.
47 (V), p. 364.

48 (V), p. 338.

49 (V), p. 353.

50 (V), p. 353: If for example two sensations that arise from two different spots on the skin meet together

in the soul, spatial contiguity [Nebeneinander] is still not given.

51 (AS), p. 142.
52 (V), pp. 353-354.
53 Wilhelm Wundt Outlines of Psychology Translated by Chas. Hubbard Judd (New York: E. Stecher!,

1897), p. 127.
54 (SW)43 (1861): 215-224.

55 (V), p. 354.

56 (AS), p. 127-128.

57 (SG), p. 19

58 (SG), p. 103.


(CE), pp.87-88.
2 Translations of the Mach Bands articles and expansions of the law of inhibition can be found in Floyd
Ratliffs Mach Bands (San Francisco: Holden Day, 1966).
3 "Ernst Mach and the Episode of the Monocular
Depth Sensations" Journal of the History of the
Behavioral Sciences 37(4): 327-348.
4 (R), pp. 1-5.

5 Ibid., p. 3.
6 Ibid., pp. 4-5. Because of his investigation
of Gestaltensehen Edwin Boring gave Mach a place in the
history of Gestalt psychology, although he objected to Mach's calling space a "sensation" History of
Experimental Psychology (New York: Appleton-Century Crofts, 1950) p. 442.
7 (AS), pp. 62-63.

8 "Untersuchungen iiber den Zeitsinn des Ohres" (SW)

51: 146-147.
9 "Bemerkungen iiber die Lehre von raumlichen Sehen"
pp. 3-4.
10 (R), p. 228.

11 "Bemerkungen iiber monoculare Stereoscopie" (SW) 58 ( 1868)

12 (R), p. 229.

13 (R), p. 229.
14 (AS), p. 129.
15 Deutsche Literaturzeitung 7(1886): 947-948.

16 (AS), p. 166.
17 See Nicholas Wade Destined for Distingushed
Oblivion: The Scientific Vision of William Charles
Wells (Dordrecht: K.luwer Academic Publishers, 2002).
18 These motions have several uses, among them an adaptation
called Listing's Law which compensates
for rolling motions of the eye produced as a matter of geometric necessity.
19 The true number of cases is somewhat larger. As Mach
says (AS), p. 206, the eyes also carry out
compensatory movements to stay locked on the same point in the visual field, compensating for motions of
the head. This is due to the action of the otoliths and the semicircular canals which send messages to the
ocular muscles. In a compensatory movement the head is sensed to be in motion and yet the surrounding
objects are still seen at rest because the retinas manage to achieve rest through the motions of the head.
Finally we can will to inhibit some of these compensatory motions by moving our eyes voluntarily, but then
we still sense that our own bodies are in motion.
20 "We can conceive voluntary movement at any rate in principle
as reflex movement modified by reflexes."
(AS), p. 172.
21 (AS), pp. 128-129. Wells, too, was aware of this principle
in visual direction.
22 James Principles of Psychology (Great Books of
the Western World, Encyclopedia Brittanica, 1952),
23 (AS), pp. 178-179.

24 Erich Becher "The Philosophical

Views of Ernst Mach" Philosophical Review 14(5) September 1905:
25 See William James Principles of
Psychology Chapter 20 "The Perception of Space," pp.540-541. See
also William James and James Ward "The Psychological Theory of Space" Mind 14 (53) 1889:107-115.
26 (SG), pp. 13-14.


1 "Ober die Tatsachen die der Geometrie zum Grunde liegen," p. 619.
2 "Uber das Sehen von Lagen und Winkeln durch die Bewegung des Auges" (SW) 43 1861:215-224.
3 (R), p. 230.

4 Berhard Riemann "On the Hypotheses that Lie

at the Foundation of Geometry" in D.W. Smith, ed., A

Source Book in Mathematics, (New York: McGraw Hill, 1929), pp 415-416.

5 See Carl Ludwig's recommendation letter on behalf of Hankel (NL 174/5/1379): "Aile stimmen darin

iiberein, dass er ein Mann von reichen Bildung sei, und dass ihn bei seinem Studiengang das Gliick
geradezu begiinstigt babe so ziihlt er u.a. zu den wenigen welche Riemann in seinen besten Tagen gehiirt
haben; er hat einer der weniger Vorlesungen beigewohnt in denen jener hervorragende Mathematiker
seine neuen Ansichten auf das erschiipfendste behandelt hat."
6 Philosophische Fakultat der Universitat Graz, Commissions Bericht von E. Mach den 21. Dec. 1865.

The original document is in Graz, but Deutsches Museum possesses a photocopy, numbered (NL
7 There are notes by Riemann on Herbartian psychology from his Nachlass and reprinted in Erhard

Scholtz's article "Herbart's Influence on Bernhard Riemann" Historia Mathematica 9 (1982): 432-433.
Among these notes are reflections on the force-like, opposing nature ofHerbart's ideas.
8 Psychologie als Wissenschaft vol. I Sec. 39-40.
9 "On the Hypotheses which lie at the Foundations of Geometry" in D.W. Smith, ed. a Source Book in

Mathematics (New York: Dover, 1929), p. 412.

10 Gesammelte Werke und Wissenschaftlicher Nachlass Herausgegeben von Heinrich Weber und Richard

Dedekind 2"d Auflage (New York: Dover 1953), pp. 507-508.

11 Bertrand Russell An Essay on the Foundations of Geometry (London: Routledge 1996 orig. 1897),

12 Roberto Torretti The Philosophy of Geometry from Riemann to Poincare (Dordrecht: D. Reidel, 1978),

pp. 107-108; Gregory Nowack "Riemann's Habilitationsvortrag and the Synthetic a Priori Status of
Geometry" in David Rowe and John McCleary eds. The History of Modern Mathematics Vol I. (Boston:
Academic Press, 1989).
13 Erhard Schatz "Herbart's Influence on Bernhard Riemann" Historia Mathematica 9(1982): 413-440.

14 Luciano Boi "Die Beziehungen zwischen Raum, Kontinuum und Materie im Denken Riemanns ... "

Philosophia Natura/is 31(2) 1994: 174.

15 Riemann "On the Hypotheses" ... p. 411.

16 Ibid., p. 412.

17 Ibid.
18 Erhard Scholz "Herbart's Influence on Bernhard Riemann," p.436 note 12: "Herbart discussed the one-

dimensional character of the line of sound by stressing the fact that in it "only a single transition"
between two states (nur ein einziger Uebergang durch die sammtlichen dazwischenliegenden) is possible,
an explanation which greatly resembles Riemann's own approach to the idea of a one-dimensional
manifold; see also Herbart's "Psychologische Bemerkungen zur Tonlehre" 1812, in Siimmtliche Werke (K.
Kehrbach and 0. Fliigel eds.) Vol. 3: 98-118.
19 Erhard Scholz "Riemann's fiiihe Notizen zum Mannigfaltigkeitsbegriff und zu den Grundlagen der

Geometrie" Archive for History ofExact Sciences 27(1982): p. 222. The original German is:
Wenn von einer Bestimmungsweise eines veranderlichen Gegenstandes zu jeder anderen ein stetiger
Obergang moglich ist, so bilden sammtliche Bestimmungsweisen (Wenn unter einer Menge von
verschiedenen Bestimmungsweisen eines veranderlichen Gegenstandes von jeder zu jeder anderen ein
stetiger Obergang moglich ist, so bildet die Gesammtheit dieser) eine stetig ausgedehte Mannigfaltigkeit;
jede einzelne heisst ein Punkt dieser Mannigfaltigkeit. Eine stetig ausgedehnte Mannigfaltigkeit ist eine
Grosse aber nicht messbar und darf daher nicht als eine benannte veranderliche Zahl dargestellt werden.
(Ein veranderliches Stiick einer Mannigfaltigkeit ist daher eine veranderliche Grosse deren Werthe nicht
benannte Zahlen sind.)
20 Ibid. p. 414.
21 The Nature and Growth of Modern Mathematics (Princeton, 1982), p. 455.

22 "The question of the validity of the postulates of geometry in the indefinitey small is involved in the

question concerning the ultimate basis of relations of size in space. In connection with this question,
which may well be assigned to the philosophy of space, the above remark is applicable, namely that while
in a discrete manifold the principle of metric relations is implicit in the notion of this manifold, it must
; ome from somewhere else in the case of a continuous manifold." Ipid., pp. 424-425.
An Essay on the Foundations of Geometry, p.73.
24 Ibid. pp.69-70.

25 Ibid. pp.70-71.

26 (AS), p. 294; "Zur Analyse der Tonempfindungen" SW92 (December 1885): 1283-1289.
27Another benefit of using qualities is that the fixed points can be specified by looking, or inunediate
qualitative identity, and without a cascading regress of relations to all other points.
28 Mach suggests further down ((SG), p. 105) in his exposition of the color manifold that the function log

(n'/n) can be used as a means of determining coordinates, and he probably has in mind the tone-row when
he says this.
29 Where n' is higher in pitch than n.
30 (SG), p. I 0 I.
31 (SG), p. 10 I.
32 Mach of course follows Hering's "opposing processes" color theory and not the Helmholtz-Young

33 (SG), p. 99.
34 (SG), p. I 05. The third coordinate expression seems to suggest that Mach regarded the coordinates of

the tone-row as constructed by the same logarithmic rule that determines the invariant distance. However
he does not say it must be so.
35 It will always be the case that with the second to last coefficient chosen the last is determined. If only

one coefficient can be picked arbitrarily the manifold is one-fold extended. If three can be picked
arbitrarily and the fourth determined it is three fold extended and so on. The extension always refers to
the number of arbitrary coefficients needed to determine a point there.
36 Sophus Lie eventnally generalized Helmholtz's ideas using groups of transformations. Each

transformation ("motion" of a point) is a function taking one coordinate-system to another. The

transformations form a group if there is an identity transformation, inverses and associativity, the
replacement of two transformations with a third single transformation. See Max Jammer, Concepts of
Space, Third Edition (New York: Dover, 1994), p. 157.
37 Wissenschaftliche Abhandlungen p. 618-619.
38 (AS), p. 331.
39 Helmholtz, Wissenschaftliche Abhandlungen p. 622.
40 (AS), p. 29.
41 This case is described in my "Ernst Mach and the Episode of the Monocular Depth Sensations"

Journal of the History of the Behavioral Sciences 37(4) October 2001: 338-339.
42 (AS) pp. 233-234. Ecological psychology, based on the work of American psychologist J.J. Gibson,

holds that invariants are what is perceived in the visual field, and that individual sensations tend to alter
in order to accommodate object-level perception. This view only seems anti-Machian, so long as one is
ignorant of Mach's own championing of the similar position for functions and form-qualities.
43 (AS), pp. 191,331-332
44 Helmholtz's statement that in the end spatial properties rest on physical laws makes me incline toward

that interpretation.
45 In some cases distances alone do not distinguish between objects that are mirror images of one another,

a right and a left hand for example. In this cases we have to draw a plane between them and give further
information on which side of the plane the object is to be produced.
46 (CE), pp. 51-52.
47 (CE), pp. 53-54.
48 (CE), p.54.
49 Wie kommt es dass die Diagonale Hinger, als die heiden Quadratseite